Ludwig Wittgenstein (Auth.), Jaakko Hintikka (Eds.)-Wittgenstein in Florida_ Proceedings of the Colloquium on the Philosophy of Ludwig Wittgenstein, Florida State University, 7–8 August 1989-Springer

WITTGENSTEIN
INFLORIDA
Proceedings of the Colloquium on
the Philosophy of Ludwig Wittgenstein,
Florida State University, 7-8 August 1989
Editedby
JAAKKO HINTIKKA
Reprinted from
Synthese
Volume 87, Nos. 1-2, 1991
SPRINGER SCIENCE+BUSINESS MEDIA, B.V.
ISBN 978-94-010-5573-4 ISBN 978-94-011-3552-8 (eBook)
DOI 10.1007/978-94-011-3552-8
Printed on acid-free paper
AlI Rights Reserved
1991 Springer Science+Business Media Dordrecht
Originally published by Kluwer Academic Publishers in 1991
Softcover reprint ofthe hardcover 1st edition 1991
No part of the material protected by this copyright notice may be reproduced or
utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retrieval system, without written permission from the copyright owner.
SYNTHESE / Volume 87 Nos.1-2 May 1991
SPECIAL ISSUE ON WITIGENSTEIN
Part I
Preface 1
LUDWIG WITIGENSTEIN (HEIKKI NYMAN, editor; C. G.
LUCKHARDT and M. A. E. AUE, translators) / Philosophy:
Sections 86-93 (pp. 405-35) of the so-called "Big Typescript"
(Catalog Number 213) 3
BURTON DREBEN and JULIET FLOYD / Tautology: How not to
Use a Word 23
DAVID CHARLES McCARTY / The Philosophy of Logical Wholism 51
STEVE GERRARD / Wittgenstein's Philosophies of Mathematics 125
JULIET FLOYD / Wittgenstein on 2, 2, 2 ... : The Opening of Remarks
on the Foundations of Mathematics 143
Partll
JAAKKO HINTIKKA / An Impatient Man and his Papers 183
DA VID STERN / The "Middle Wittgenstein": From Logical Atomism
to Practical Holism 203
M. R. M. TER HARK / The Development of Wittgenstein's Views
about the Other Minds Problem 227
C. GRANT LUCKHARDT / Philosophy in the Big Typescript:
Philosophy as Trivial 255
DA VID PEARS / Wittgenstein' s Account of Rule-Following 273
MICHAEL LEE KELLY / Wittgenstein and "Mad Pain" 285
KENT LINVILLE and MERRILL RING / Moore's Paradox Revisited 295
THEODORE R. SCHATZKl / Elements of a Wittgensteinian Philoso-
phy of the Human Sciences 311
PREFACE
Most of the papers appearing in volume 87 numbers, 1-2 are based
on papers presented at the Colloquium on the Philosophy of Ludwig
Wittgenstein held at the Department of Philosophy at Florida State
University on 7-8 April 1989. We owe warm thanks to Florida State
University for generously supporting this colloquium. The English
translation of the chapter entitled 'Philosophie', from Wittgenstein's
typescript number 213 (von Wright), appears here with permission of
Wittgenstein's literary heirs, without affecting existing copyrights. The
original German version of this chapter was edited by Heikki Nyman
and appeared in Revue Internationale de Philosophie 43 (1989), pp.
175-203. Jaakko Hintikka's article (87, No.2) first appeared in a shorter
form in The Times Literary Supplement No. 4565 (28 September to 4
October 1990, p. 1030). The present version appears with the permis-
sion of The Times Literary Supplement, which is gratefully acknowl-
edged. Our thanks are due to all the participants of the colloquium and
the contributors to these special numbers.
JAAKKO HINTIKKA
Synthese 87: 1, 1991.
LUDWIG WITTGENSTEIN
PHILOSOPHY
SECTIONS 86-93 (pp. 405-35) OF THE SO-CALLED
"BIG TYPESCRIPT" (CATALOG NUMBER 213)
Edited by HEIKKI NYMAN
Translated by c. G. LUCKHARDT AND M. A. E. AUE
EDITORIAL NOTE
One of the most interesting writings by Wittgenstein is the typescript
no. 213 in von Wright's catalogue (see G. H. von Wright, Wittgenstein,
Basil Blackwell, Oxford, 1982, pp. 43-57). This work is colloquially
known as The Big Typescript. It is unlike Wittgenstein's other writings
in that it is organized in an almost conventional way into chapters and
sections. Parts of The Big Typescript were included by Rush Rhees in
the volume called Philosophical Grammar, but much of the text has so
far remained unpublished.
What appears here (in the original German) is the chapter entitled
'Philosophie' (pp. 405-35 of the original). This chapter is Wittgenstein's
most extensive statement of his conception of philosophy. Even if we
make due allowance for the development of Wittgenstein's views, this
chapter is of great importance for the understanding of Wittgenstein's
entire philosophy. It appears here with the kind permission of Professor
G. E. M. Anscombe and Professor G. H. von Wright.
The typescript form of TS 213 suggests that Wittgenstein might have
considered it for publication. If so, he quickly changed his mind, and
began to make handwritten changes in the typescript. They have been
indicated here as fully as possible. The actual editing has been carried
out by Heikki Nyman.
I hope that the publication of this important chapter helps our readers
to appreciate more and more the scope and subtlety of Wittgenstein's
thought.
JAAKKO HINTIKKA
Synthese 87: 3-22, 1991.
This translation 1991 Kluwer Academic Publishers.
4 LUDWIG WITTGENSTEIN
EDITOR'S NOTE
The text of the typescript is rendered here as accurately as possible.
All alternatives which have been typed into the typescript, both those
indicated by the symbols / / ... / /, as well as those written in over the
line, are rendered in the edited text. On the other hand, all handwritten
additions by Wittgenstein (alternatives, corrections and marginal notes)
are given in footnotes. '
In the margins of the pages of the typescript there are also numerous
diagonal lines and crosses entered in handwriting. These notations,
whose meaning or weight cannot be determined unambiguously, have
not been included in this edition. However, the question marks above
the lines as well as the broken underlinings - both indications of discon-
tent or uncertainty - were retained.
I did not want to alter Wittgenstein's orthography or his notation,
and without noting it I have corrected only a couple of obviously
erronemis parts of the text.
HEIKKI NYMAN
TRANSLATORS NOTE
We wish to thank Professor G. E. M. Anscombe for going over our
translation with us, and Mr. D. Hudson Mulder for his helpful com-
ments on a draft of the translation.
C. G. LUCKHARDT
M. A. E. AUE
PHILOSOPHY 5
86
DIFFICULTY OF PHILOSOPHY NOT THE INTELLECTUAL DIFFICULTY OF THE
SCIENCES, BUT THE DIFFICULTY OF A CHANGE OF ATTITUDE. RESISTANCES
OF THE WI L L MUST BE OVERCOME.
As I have often said, philosophy does not lead me to any renunci-
ation, since I do not abstain from saying something, but rather abandon
a certain combination of words as senseless. In another sense, however,
philosophy requires a resignation, but one of feeling and not of intellect.
And maybe that is what makes it so difficult for many. It can be difficult
not to use an expression, just as it is difficult to hold back tears, or an
outburst of anger I /ragel I.
I(Tolstoy: the meaning (meaningfulness) of a subject lies in its being
generally understandable. - That is true and false. What makes a
subject difficult to understand - if it is significant, important - is not
that some special instruction about abstruse things is necessary to under-
stand it. Rather it is the contrast between the understanding of the
subject and what most people wan t to see. Because of this the very
things that are most obvious can become the most difficult to under-
stand. What has to be overcome is not a difficulty of the intellect, but
of the will.)1
Work on philosophy is - as work in architecture frequently is -
actually more of alia kind of! I work on oneself. On one's own concep-
tion. On the way one sees things. (And what one demands of them.)
Roughly speaking, in Ilaccording toll the old conception - for in-
stance that of the (great) western philosophers - there have been two
kinds of problems in fields of knowledge Iitwofold kinds of prob-
lems ... .I I: essential, great, universal, and inessential, quasi-accidental
problems. And against this stands our conception, that there is no such
thing as a g rea t, essential problem in the sense of "problem" in the
field of knowledge.
87
PHILOSOPHY SHOWS THE MISLEADING ANALOGIES IN THE USE OF LANGUAGE.
Is grammar, ~ ~ ~ ~ _ ~ ~ ~ ~ _ ~ , , : , ) _ ~ ~ , only the description of the actual
handling of language I /languagesl I? So that its propositions could actu-
ally be understood as the propositions of a natural science?
That could be called descriptive science of speaking, in contrast
to that of thinking.
Indeed, the rules ?f could be taken as propositions from the
natural history of man. (As the games of animals are described in books
on natural history.)
If I correct a philosophical mistake and say that this is the way it has
always been conceived, but this is not the way it is, I always point to
an analogy I II must always point to ... .I I that was followed, and show
that this analogy is I I . ... I must always point to an analogy
according to which one had been thinking, but which one did not
recognize as an analogy.! I
The effect of a false analogy taken up into language: it a
constant battle and uneasiness (as it were, a constant stimulus). It is as
if a thing seemed to be a human being from a distance, because we
don't perceive anything but from close up we see that it is a
tree stump. The moment we move away a little and lose sight of the
explanations, 0 n e figure appears to us; if after that we look more
closely, we see a different figure; now we move away again, etc., etc.
(The irritating character of grammatical unclarity.)
Philosophizing is: rejecting false arguments.
The philosopher strives to find the liberating word, that is, the word
that finally permits us to grasp what up until now
l
has intangibly
weighed down our consciousness.
(It is as if one had a hair on one's tongue; one feels it, but cannot
grasp I Iseizel I it, and therefore cannot get rid of it.)
The philosopher delivers the word to us with which one I III I can
express the thing and render it harmless.
(The choice of our words is so important, because the point is to hit
upon the physiognomy of the thing because only the exactly
aimed thought can lead to the correct track. The car must be placed
on the tracks so that it can keep rolling correctly.)
PHILOSOPHY 7
One of the most important tasks is to express all false thought pro-
cesses so characteristically that the reader says, "Yes, that's exactly the
way I meant it". To make a tracing of the physiognomy of every error.
Indeed we can only convict someone else of a mistake if he acknowl-
edges that this really is the expression of his feeling. / /. . .. if he
(really) acknowledges this expression as the correct expression of his
feeling.! /
For only if he acknowledges it as such, is it the correct expression.
(Psychoanalysis. )
What the other person acknowledges is the analogy I am proposing
to him as the source of his thought.
88
WHERE DOES THE FEELING THAT OUR GRAMMATICAL INVESTIGATIONS
ARE FUNDAMENTAL COME FROM?
(Questions of different kinds occupy us, for instance "What is the
specific weight of this body", "Will the weather stay nice today", "Wh02
will come through the door next", etc. But among our questions there
are those of a special kind. Here we have a different experience. The
questions seem to be more fundamental than the others. And now I
say: if we have this experience, then we have arrived at the limits of
language. )3
Where does our investigation get its importance from, since it seems
only to destroy everything interesting, that is, all that is great and
important? (As it were all the buildings, leaving behind only bits of
stone and rubble.)
Whence does this observation derive its importance:
4
the one that
points out to us that a table can be used in more than 0 n e way, that
one can think up a table that instructs one as to the use of a table?
The observation'that one can also conceive of an arrow as pointing from
the tip to the tail, that I can use a model as a model in different ways?
What we do is to bring words back from their metaphysical to their
correctS (normal) use in language.
(The man who said that one cannot step into the same river twice
said something wrong; one can step into the same river twice.)
And this is what the solution of all philosophical difficulties looks
like. Their
6
answers, if they are correct, must be homemade and
ordinary.7 But one must look at them in the proper splik-aocfihen it
doesnS- matter. 8 ------ --
Where do / /did/ / the old philosophical problems get their importance
from?
The law of identity, for example, seemed to be of fundamental
importance. But now the proposition that this "law" is nonsense has
taken over this importance.
I could. ask: why do I sense a grammatical joke as being in a certain
sense dee'p? (And that of course is what the depth of philosophy is.)
Why do we sense the investigation of grammar as being fundamental?
(When it has a meaning at all, the work "fundamental" can also
mean something that is not metalogical, or philosophical.)9
The investigation of grammar is fundamental in the same sense in
which we may call language fundamental - say its own foundation.
Our grammatical investigation differs from that of a philologist, etc.:
what interests us, for instance, is the translation from one language'
into other languages we have invented. In general the rules that the
philologist totally ignores are the ones that interest us. Thus we are
justified in emphasizing this difference.
On the other hand it would be misleading to say that we deal with
the essentials of grammar (he, with the accidentals).
"But that is only an external differentiation / Ian external differ-
ence/ /." I believe there is no other.
Rather we could say that we are calling something else grammar than
he is. Even as we differentiate kinds of words where for him there is
no difference (present).
----------
The importance of grammar is the importance of language.
PHILOSOPHY 9
One could also call a word, for instance 'red', important insofar as
it is used frequently and for important things, in contrast, for instance,
to the word 'pipe-lid'. And then the grammar of the word 'red' is
important because it describes the meaning of the word 'red'.
(All that philosophy can do is to destroy idols. And that means not
creating a new one - for instance as in "absence of ar idol".)
89
THE METHOD OF PHILOSOPHY: THE PERSPICUOUS REPRESENTATION
OF GRAMMATICAL / /UNGUISTIci / FACTS.
THE GOAL: THE TRANSPARENCY OF ARGUMENTS. JUSTICE.
lO
Someone has heard that the anchor of a ship is hauled up by a steam
engine. He only thinks of the one that powers the ship (and because
of which it is called a steamship) and cannot explain to himself what
he has heard. (Perhaps the difficulty doesn't occur to him until later.)
Now we tell him: No, it is not t hat steam engine, but besides it a
number of other ones are on board, and one of these hoists the an-
chor. - Was his problem a philosophical one? Was it a philosophical
one if he had already heard of the existence of other steam engines on
the ship and only had to be reminded of it? - I believe his confusion
has two parts: what, the explainer tells him as fact the questioner could
easily have conceived as a possibility by himself, and he could have
posed his question in a definite form instead of in the form of a mere
admission of confusion. He could have removed this part of his doubt
by himself; however, reflection could not have instructed him about
the facts. Or: the uneasiness that comes from not having known the
truth was not removable by any ordering of his concepts;
The other uneasiness and confusion is characterized by the words
"Something's wrong here" and the solution is characterized by
"Oh, you don't mean t hat steam engine" or - in another
case - " .... By 'steam engine' you don't mean just a piston engine."
The work of the philosopher consists in assembling reminders for a
particular purpose.
A philosophical question is similar to one about the constitution of
a particular society. - And it would be as if a society came together
without clearly written rules, but with a need for them; indeed also
with an instinct following which they observed / /followed/ / certain rules
at their meetings; but this is made difficult by the fact that nothing is
clearly expressed and no arrangement is made which clarifies
/ /brings out clearly/ / the rules. Thus they in fact view one of them as
president, but he doesn't sit at the head of the table and has no
distinguishing marks, and that makes doing busine,ss difficult. Therefore
we come along and create a clear order: we seat the president in a
clearly identifiable spot, seat his secretary next to him at a little table
of his own, and seat the other full members in two rows on both sides
of the table, etc., etc.
H one asks philosophy: "W hat is - for instance - substance?" then
one is asking for a rule. A general rule, which is val i d for the word
"substance", i.e., a rule according to which I have decided to play. -
I want to say: the question "What is .... " doesn't refer to a particular-
practical - case, but we ask it sitting at our desks. Just remember the
case of the law of identity in order to see that taking care of a philosoph-
ical problem is not a matter of pronouncing new truths about the subject
of the investigation (identity).
The difficulty lies onlyll in understanding how establishing a rule
helps us. Why it calms us after we have been so profoundly12 uneasy.
Obviously what calms us is that we see a system whlc-h-(s-ystematically)
------------------
excludes those structures that have always made us those we
were unable to do anything with, and which we still thought we had to
respect. Isn't the establishment of such a grammatical rule similar in
this respect to the discovery of an explanation in physics, for instance,
of the Copernican system? A similarity exists. - The strange thing about
philosophical uneasiness and its resolution that it is
like the of an ascetic who stood raising a heavy ball, amid
groans, and whom someone released by telling him: "Drop it." One
wonders: if these sentences make you uneasy and you didn't know what
to do with them, why didn't you drop them earlier, what stopped you
from doing it? Well, I believe it was the false system that he thought
he had to accommodate himself to, etc.
13
(The particular peace of mind that occurs when we can place other
similar cases next to a case that we thought was unique, occurs again
and again in our investigations when we show that a word doesn't have
PHILOSOPHY 11
just 0 n e meaning (or just two), but is used in five or six different ways
(meanings). )
Philosophical problems can be compared to locks on safes, which can
be opened by dialing a certain word or number, so that no force can
open the door until just this word has been hit upon, and once it is hit
upon any child can open it. I I . ... and if it is hit upon, no effort at all
is necessary to open the doorl I itl I. I I
The concept of a perspicuous representation is of fundamental sig-
nificance for us. It earmarks the form of account we give, the way we
look at things. (A kind of 'Weltanschauung', as is apparently typical of
our time. Spengler.)
This perspicuous representation produces just that comprehension
I lunderstandingl I which consists in "seeing connections". Hence the
importance of i n t e r m e d i ate cas e sl I of finding i n t e r m e d i -
ate cases.11
A sentence is completely logically analyzed when its grammar is laid
out completely clearly. It might be written down or spoken in any
number of ways.
Above all, our grammar is lacking in per s pic u i t y.
Philosophy may in no way interfere with the real I lactual! I use of lan-
guage I I . ... with what is really saidl I; it can in the end only describe it.
For it cannot give it any foundation either.
It leaves everything as it is.
It also leaves mathematics as it is (is now), and no mathematical
discovery can advance it.
A "leading problem of mathematical logic" (Ramsey) is a problem
of mathematics I ike any 0 the r.
(A simile is part of our ~ ~ _ i ~ ~ ; but we cannot draw any conclusions
from it either; it doesn't lead us beyond itself, but must remain standing
as a simile. We can draw no inferences from it. As when we compare
a sentence to a picture (in which case, what we understand by 'picture'
must already have been established in us earlier I Ibeforel I) or when I
compare the application of language with, for instance, that of the
calculus of multiplication.
Philosophy simply puts everything before us, and neither explains
nor deduces anything.)
Since everything lies open to view there is nothing to explain either.
For what might not lie open to view is of no interest to us. I I . ... , for
what is hidden, for example, is ... .I I
The answer to the request for an explanation of negation is really:
don't you understand it? WefC ir yo-li--unders'tand it, there
what business is there left for an explanation?
We must know what ex pIa nat ion means. There is a constant
danger of wanting to use this word in logic in a sense that is derived
from physics.
When
14
methodology talks about measurement, it does not say which
would be the most advantageous to make the measuring stick
of in order to achieve this or that result: even though this too, after
all, is part of the method of measuring. Rather this investigation is only
interested in the circumstances under which we say that a length, the
strength of a current (etc.) is measured. It wants to tabulate the methods
which we already used and are familiar with, in order to determine the
meaning of the words "length", "strength of current", etc.)
If one tried to advance the s e s in philosophy, it would never be
possible to debate them, because everyone would agree to them.
Learning philosophy is rea 11 y recollecting. We remember that we
really used words in this way.15
The aspects of things I lof language II which are philosophically most
important are hidden because of their simplicity and familiarity.
(One is unable to notice something because it is always (openly)
before one's eyes.) --------
The real foundations of his inquiry do not strike a man at all. Unless
t hat fact has at sometime struck him Ilhe has become aware off I.
(Fraser, etc., etc.)
And this means he fails to be struck by what is most striking (power-
ful).
(One of the greatest impediments for philosophy is the expectation
of new, deep Ilunheard of! I elucidations.)
PHILOSOPHY 13
One might also give the name philosophy to what is possible Ilpre-
senti I b e for e all new discoveries and inventions.
This must also relate to the fact that I can't give any explanations of the
variable "sentence". It is clear that this logical concept, this variable, must
belong to the same order as the concept "reality" or "world".
If someone believes he has found the solution to the. 'problem of life'
and tried to tell himself that now everything is simple; then in order to
refute himself he would only have to remember that there was a time
when this 'solution' had not been found; but at t hat time too one had
to be able to live, and in reference to this time the new solution appears
like Ilasll a coincidence. And that's what happens to us in logic. If
there were a 'solution' of logical (philosophical) problems then we
would only have to call to mind that at one time they had not been
solved (and then too one had to be able to live and think). ---
All reflections can be carried out in a much more homemade manner
than I used to do. And therefore no new words have to be used in
philosophy, but rather the old common words of language are sufficient.
lithe old ones are sufficientl I
(Our only task is to be just. That is, we must only point out and
resolve the injustices of philosophy, and not posit new parties - and
creeds.)
(It is difficult not to exaggerate in philosophy.)
(The philosopher exaggerates, shouts, as it were, in his helplessness,
so long as he hasn't yet discovered the core of his confusion.)
The philosophical problem is an awareness of disorder in our con-
cepts, and can be solved by ordering them.
A philosophical problem always has the form: "I simply don't know
my way about."
As I do philosophy, its entire task consists in expressing myself in
such a way that certain troubles I Iprotilemsl I disappear. ((Hertz.
------------
If I am correct, then philosophical problems must be completely
solvable, in contrast to all others.
If I say: here we are at the limits of language, then it always seems
/ /sounds/ / as if resignation were necessary, whereas on the contrary
complete satisfaction comes, since no question remains.
The problems are dissolved in the actual sense of the word - like a
lump of sugar in water.
/People who have no need for transparency in their argumentation
are lost to philosophy./
90
PHILOSOPHY.
THE CLARIFICATION OF THE USE OF LANGUAGE. TRAPS OF LANGUAGE.
How is it that philosophy is such a complicated building / /structure/ /.
After all, it should be completely simple if it is that ultimate thing,
independent of all experience, that it claims to be. - Philosophy un-
ravels the knots in our thinking; hence its results must be simple, but
its activity is as complicated as the knots that it unravels.
Lichtenberg: "Our entire philosophy is correction of the use of lan-
guage, and therefore the correction of a philosophy, and indeed of the
most general philosophy."
(The capacity16 for philosophy consists
17
in the ability 18 to receive a
strong lasting impression from a grammatical
Why are grammatical problems so tough and seemingly ineradi-
cable? - Because they are with the oldest thought habits,
i.e., with the oldest images that are engraved into our language itself.
((Lichtenberg.) )
/Teaching philosophy involves the same immense difficulty as instruc-
tion in geography would have if a pupil brought with him a mass of
false and far too simple / / and falsely simplified/ / ideas about the course
and connections of the rmites of rivers / /rivers/ / and mountain chains
---------------
/ /mountains/ /.I
/People are deeply imbedded in philosophical, i.e., grammatical con-
fusions. And to free them from these presupposes pulling them out of
the immensely manifold connections the-Y-are-caught up in. One must
PHILOSOPHY 15
SO to speak regroup their entire language. - But this language came
about I I developed I I as it did because people had - and have - the
inclination to think i nth i s way. Therefore pulling them out only
works with those who live in an instinctive state of rebellion against
IIdissatisfaction language. Not with those who following all of
their instincts live within the herd that has created this language as
its expression.!
Language contains the same traps for everyone; the immense network
of well-kept I Ipassablel I false paths. And thus we see one person after
another walking the same paths and we know already where he will
make a turn, where he will keep on going straight ahead without
noticing the turn, etc., etc. Therefore wherever false paths branch off
I should put up signs which help one get by the dangerous places.
One keeps hearing the that philosophy really makes no pro-
gress, that the same philosophical problems that had occupied the
Greeks are still occupying us. But those who that don't under-
stand the reason it is Ilmust bel I so. The reason is that our language
has remained the same and seduces us into asking the same questions
over and over. As long as there is a verb 'to be' which seems to function
like 'to eat' and 'to drink', as long as there are adjectives like 'identical',
'true', 'false', 'possible', as long as one talks about a flow of time and
an expanse of space, etc., etc., humans will continue to bump up against
the same mysterious difficulties, and stare at something that no explana-
-------------- -------------
tion seems able to remove.
And this by the way satisfies a longing for the supra-natural I !tran-
scendental! I, for in believing that they see the "limits of human under-
standing" of course they believe that they can see beyond it.
I read" .... philosophers are no nearer the meaning of 'Reality' than
Plato got .... ". What a strange state of affairs. How strange in that
case that Plato could get that far at all! Or, that we were not able to
get farther! Was it because Plato was so smart?
The conflict in which we constantly find ourselves when we undertake
logical investigations is like the conflict of two people who have con-
cluded a contract with each other, the last formulations of which are
expressed in easily misunderstand able whereas the explanations
of these formulations explain everything unmistakably. Now one of the
two people has a short memory, constantly forgets the explanations,
misinterprets the conditions of the contract, and continually gets
Iitherefore continually runsl I into difficulties. The other one constantly
has to remind him of the explanations in the contract and remove the
difficulty.
Remember what a hard time children have believing (or accepting)
that a word really has Ilcan havel I two completely differe-nt me-1in{ngs.
The aim of philosophy is to erect a wall at the point where language
stops anyway.
The results of philosophy are the uncovering of one or another piece
of plain nonsense, and are the bumps that the understanding has got
by running its head up against the limits I !the endl I of language. These
bumps let us I Irecognizel I the value of the discovery.
What kind of investigation are we carrying out? Am I investigating
the probability of cases that I give as examples, or am I investigating
their actuality? No, I'm just citing what is possible and am therefore
----------
giving grammatical examples.
Philosophy is not laid down in sentences but in a language.
Just as laws only become interesting when there is an inclination to
transgress them Ilwhen they are transgressedl I grammatical
rules are only interesting when philosophers want to transgress them.
Savages have games (that's what we call them, anyway) for which
they have no written rules, no inventory of rules. Let's now imagine
the activity of an explorer, who travels through the countries of these
peoples and -takes -1in-{nventory of their rules. This is completely anal-
ogous to what the philosopher does. But why don't I say: savages
have languages (that's what we .... ) .... without a written
grammar?) )20
PHILOSOPHY 17
91
WE DON'T ENCOUNTER PHILOSOPHICAL PROBLEMS AT ALL IN PRACTICAL
LIFE (AS WE DO, FOR EXAMPLE, THOSE OF NATURAL SCIENCE). WE EN-
COUNTER THEM ONLY WHEN WE ARE GUIDED NOT BY PRACTICAL PURPOSE
IN FORMING OUR SENTENCES, BUT BY CERTAIN ANALOGIES WITHIN OUR
LANGUAGE.
Language cannot express what belongs to the essence of the world.
Therefore it cannot say that everything flows. Language can only say
what we could also imagine differently.
That everything flows must lie in how language touches reality. Or
better: that everything flows must lie in the nature of language. And,
let's remember: in everyday life we don't notice that - as little as we
notice the blurred edges of our visual field ("because we are so used
to it", some will say). How, on what occasion, do we think we start
noticing it? Isn't it when we want to form sentences in opposition to
the grammar of time?
When someone says 'everything flows', we feel that we are hindered
in pinning down the actual, actual reality. What goes on on the screen
escapes us precisely because it is something going on. But we are
describing something; and is that something else that is going on? The
description is obviously linked to the very picture on the screen. There
must be a false picture at the bottom of our feeling of helplessness.
For what we want to describe we can describe.
Isn't this false picture that of a strip of film that runs by so quickly
that we don't have any time to perceive a picture?
For in this case we would be inclined to chase after the picture. But
in the course of something going on there is nothing analogous to that.
It is remarkable that in everyday life we never have the feeling
that the phenomenon is getting away from us, that appearances are
continually flowing, but only when we philosophize. This points to the
fact that we are dealing here with a thought that is suggested to us
through a wrong use of our language.
For the feeling is that the present vanishes into the past without our
being able to stop it. And here we are obviously using the picture of a
strip that constantly moves past us and that we can't stop. But of course
it's just as clear that the picture is being misused. That one cannot say
"time flows" if by "time" one means the possibility of change.
That we don't notice anything when we look around, look around in
space, feel our own bodies, etc.', etc., shows how natural these very
things are to us. We don't perceive that we see space perspectivally or
that the visual image is in some sense blurred near its edge. We don't
notice this, and can never notice it, because it is the mode of percep-
tion. We never think about it, and it is impossible, because the form
of our world has no contrary.
I wanted to say that it is odd that those who ascribe reality only to
things and not to our mental images move so self-confidently in the
world of imagination and never long to escape from it.
I.e., how self-evident is the given. Things would have to have come
to a pretty pass for that to be just a tiny photograph taken from an
oblique angle.
What is self-evident, Ii f e, is supposed to be something accidental,
unimportant; by contrast something that normally I never worry my
head about is what is real!
I.e., what one neither can nor wants to go beyond would not be the
world.
Again and again there is the attempt to define the world in language
and to display it - but that doesn't work. The self-evidence of the world
is expressed in the very fact that language means only it, and can only
mean it.
As language gets its way of meaning from what it means, from the
world, no language is thinkable which doesn't represent this world.
In the theories and battles of philosophy we find words whose mean-
ings are well-known to us from everyday life used in an ultraphysical
sense.
When philosophers use a word and search for its meaning, one must
always ask: is this word ever really used this way in the language which
created it! Ifor which it is createdl/?
Usually one will then find that it is not so, and that the word is
used against Ilcontrary tol I its normal grammar. ("Knowing", "Being",
"Thing" .)
(Philosophers are often like little 21 who first scribble
PHILOSOPHY 19
lines on a piece of paper with their pencils, and now I !thenl I .
ask an adult "What is that?" - Here's how this happened: now and
then the adult had drawn something for the child and said: "That's a
man", "That's a house", etc. And then the child draws lines too, and
asks: now what's t hat ?)
92
METHOD OF PHILOSOPHY.
THE POSSIBILITY OF QUIET PROGRESS.
The real discovery is the one that makes me capable of stopping
doing philosophy when I want to.
The one that gives philosophy peace, so that it is no longer I Ibeingl I
tormented by questions which bring its elf in question.
Instead, we now demonstrate a method by examples; and one can
break off the series of examples I land the series of examples can be
broken offl I.
But more correctly, one should say: Problems are solved (uneasiness
I Idifficultiesl I eliminated), not a sin g I e problem.
Unrest in philosophy comes from philosophers looking at, seeing,
philosophy all wrong, i.e., cut up into (infinite) vertical strips, as it were,
rather than (finite) horizontal strips. This reordering of understanding
creates the g rea t est difficulty. They want to grasp the infinite strip,
as it were, and complain that it I Ithisl I is not possible piece by piece.
Of course it isn't, if by 'a piece' one understands an endless vertical
strip. But it is, if one sees a horizontal strip as a piece Iia whole,
definite piecell. - But then we'll never get finished with our work! Of
course I Icertainlyl I not, because it doesn't have an end.
----------
(Instead of turbulent conjectures and explanations, we want to give
quiet demonstrations
23
I Istatementsl I of linguistic facts II[bout linguis-
tic the of linguistic facts'! I
We must plow though the whole of language.
(When most people ought t0
25
engage in a philosophical investi-
gation, they act like someone who is looking for an object in a drawer
very nervously. He throws papers out of the drawer - what he's looking
for may be among them - leafs through the others hastily and sloppily.
Throws some back into the drawer, mixes them up with the others,
and so on. Then one can only tell him: Stop, if you look in t hat way,
I can't help you look. First you have to start to examine one thing after
another methodically, and in peace and quiet; then I am willing to look
with you and to direct myself with you as model in the method.
J -- ---- --------
93
THE METHODOLOGY IN THE FORMS OF OUR LANGUAGE. PAUL ERNST.
In ancient rites we find the use of an extremely well-developed lan-
guage of gestures.
And when I read Frazer, I would like to say again and again: All
these processes, these changes of meaning, we have right in front of us
even in our language of words. If what is hidden in the last sheaf is
called the 'Cornwolf', as well as the sheaf itself, and also the man who
binds it, then we recognize in this a linguistic process we know well.
The scapegoat, on which one lays one's sins, and who runs away into
the desert with them - a false picture, similar to those that cause errors
in philosophy.
I would like to say: nothing shows our kinship with those savages
better, than that Frazer has at hand a word like "ghost" or "shade",
which is so familiar to him and to us, to describe the views of these
people.
(This is quite different than if he were to relate, for instance, that
the savages imagined / /imagine/ / that their head falls off when they
have slain an enemy. Here 0 u r des c rip t ion would contain nothing
superstitious or magical.)
Indeed, this oddity refers not only to the expressions "ghost" and
"shade", and much too little is made of it that we include the words
"soul" and "spirit" in our own educated vocabulary. this
it is insignificant that we do not believe that our soul eats and drinks.
------.---------
An entire mythology is laid down in our language.
PHILOSOPHY 21
Driving out death or killing death; but on the other hand it is portrayed
as a skeleton, and therefore as dead itself, in a certain sense. "As dead
as death." 'Nothing is as dead as death; nothing as beautiful as beauty
itself!' The picture according to which realitv is thought of here
. ilTe pure (concentr,ted) sUbstances
IS that beauty, death, etc., IS the pure concentrated) substance,
h
. b 'f I b' ate . d d .
w ereas m a eaut! u 0 Ject It IS contame as an a mIxture. - And
don't I recognize here my own observations about 'object' and 'com-
plex'? (Plato.)
The primitive forms of our language: noun, adjective and verb, show
the simple picture whose form tries to force everything.
So long as one imagines the soul as a t h i n g , abo d y , which is
in our head, this hypothesis is not dangerous. The danger of our
models does not lie in their imperfection and roughness, but in their
unclarity (fogginess).
The danger sets in when we notice that the old model is not sufficient
but then we don't change it, but ony sublimate it, as it were. So long
as I say the thought is in my head, everything is all right; things get
dangerous when we say that the thought is not in my head, but in my
spirit.
NOTES
1 Handwritten alternative: then.
2 In the original typescript: he. The initial letter "w" is a handwritten addendum.
3 Handwritten marginal note: belongs to "must", "can".
4 The typescript has: its importance:, the.
5 Handwritten alternative: normal. There is a handwritten wavy line under "correct":
6 Handwritten alternative: our.
7 Handwritten alternative: ordinary and trivial.
8 At the end of the remark there is the handwriting: < ["plain nonsense"].
9 The parentheses are a handwritten addition.
10 Under the title, in handwriting: V p. 40/3? This is a reference to page 40 of the
typescript. Wittgenstein had apparently wanted to include a remark or a part of one
from page 40 in page 414.
11 Handwritten alternative: now.
12 Handwritten alternative: deeply.
13 At the end of the remark there is a handwritten addition: hen and chalk trick.
14 Before the remark, in handwriting in the margin: VII 7.
15 In handwriting, in the margin: VII 164.
16 Handwritten alternative (with an unbroken wavy line under the original word): talent.
17 Handwritten alternative (with an unbroken wavy line under the original word): lies.
18 Handwritten alternative (elevated, with broken underlining): susceptibility.
19 Handwritten marginal remark: for 'humor', 'depth'.
20 In the typescript the parentheses are missing at the end of the remark.
21 Handwritten alternative: (Philosophers} often behave like little children ....
22 Handwritten alternative: some.
23 Handwritten alternative: reflection.
24 Originally in the manuscript: establishment.
25 Unclear textual point in the manuscript. The typewriting gives the impression that the
original words "want to" were overstruck to produce "ought to".
Luckhardt:
Dept. of Philosophy
Georgia State University
Atlanta, GA 30303-3083
U.S.A.
Aue:
Dept. of German
Emory University
Atlanta, GA 30322
U.S.A.
BURTON DREBEN AND JULIET FLOYD
TAUTOLOGY: HOW NOT TO USE A WORD
In 1923 C. I. Lewis wrote to F. J. Woodbridge, editor of The Journal
of Philosophy:
Have you looked at Wittgenstein's new book yet? I am much discouraged by Russell's
foolishness in writing the introduction to such nonsense. I fear it will be looked upon as
what symbolic logic leads to. If so, it will be the death of the subject. 1
Nonsense or no, Lewis would, nine years later, pay homage to
Wittgenstein's account of logic. In the Introduction to Lewis and Lang-
ford's Symbolic Logic, published in 1932, Lewis wrote:
the nature of logical truth itself has become more definitely understood, largely through
the discussions of Wittgenstein. It is 'tautological' - such that any law of logic is equivalent
to some statement which exhausts the possibilities; is affirmed in logic is a truth
to which no alternative is conceivable. From the relation of mathematics to logic, it
follows that mathematical truth is similarly tautological. 2
No dictionary in 1932 would have glossed "tautological" as "exhaust-
ing the possibilities". For example, in the second edition (1934) of the
Merriam-Webster New International Unabridged Dictionary (for which
Lewis was a consultant) we find the adjective "tautological" under
"tautology", derived from the Greek tautologia, "speaking the same":
tautology .. 1. Rhet. Repetition of the same words or use of synonymous words in close
succession; also, an instance of this. 2. Repetition of a statement, of acts, experiences,
etc., esp. when superfluous.
Syn. and Ant. - See REDUNDANCY.
But Lewis was not alone in misspeaking. Russell, Ramsey and Carnap
also took Wittengstein as giving them license to misspeak. This paper
will discuss the (linguistic) misdeeds of Russell, then briefly of Ramsey
and Lewis. (Carnap is for another time and another place.)
l.
First, a bit of prehistory. In the Jasche Logic, after defining analytic
propositions as "those propositions whose certainty rests on identity of
Synthese 87: 23-49, 1991.
1991 Kluwer Academic Publishers.
24 BURTON DREBEN ANI7JULIET FLOYD
concepts (of the predicate with the notion of the subject),,,3 Kant goes
on, in a section entitled "Tautological Propositions," to further divide
analytic propositions into two kinds: 'explicit' and 'non-explicit'. He
says:
The identity of concepts in analytic judgments can be either explicit (explicita) or non-ex-
plicit (implicita). In the former case analytic propositions are tautological.
Note 1. Tautological propositions are virtualiter empty or raid of consequences, for
they are of no avail or use. Such is, for example, the tautological proposition, Man is
man. For if I know nothing else of man than that he is man, I know nothing else of him
at all.
Implicitly identical propositions, on the contrary, are not void of consequences or
fruitless, for they clarify the predicate which lay undeveloped (imp/icite) in the concept
of the subject through development (explicatio).
4
In the Prolegomena to any Future Metaphysics, Kant emphasizes that
all analytic judgments depend wholly upon the law of contradiction .... For the predicate
of an affirmative analytical judgment is already contained in the concept of the subject,
of which it cannot be denied without contradiction. Such is the nature of the judgments
"All bodies are extended," and "No bodies are unextended (that is, simple,,).5
Kant's insistence that a proposition such as "Man is man" is tautologi-
cal and "void of consequences" is presumably directed against Leibniz,
who, in 'On the General Characteristic' (c. 1679) wrote:
The first of the true propositions are those which are commonly called identical; such as
A is A, non-A is non-A, and if the proposition L is true, it follows that the proposition
L is true. And however much useless 'coccysm,6 there seems to be in these judgments,
they nevertheless give rise to useful axioms by a slight change. Thus from the fact that
A is A, or for example, that three-legged is three-legged, it is obvious that anything is
as much as it is or is equal to itself. Hence (to show how useful identities are by an
example) philosophers have long ago demonstrated that a part is less than the whole by
assuming only this definition.
7
For Leibniz, in the sense of being resolvable to identity, all truth is
from God's point of view what Kant called 'analytic'; that is:
An affirmative truth is one whose predicate is in the subject; and so in every true
affirmative proposition, necessary or contingent, universal or particular, the notion of
the predicate is in some way contained in the notion of the subject, in such a way that
if anyone were to understand perfectly each of the two notions just as God understands
it, he would by that very fact perceive that the predicate is in the subject. 8
Leibniz would never have called instances of the law of identity
'tautologous'. Kant did. But Kant still cites the law of identity in his
TAUTOLOGY: HOW NOT TO USE A WORD 25
Logic
9
as one of the fundamental laws, along with the other two tra-
ditional "laws of thought", the principle of contradiction and the law
of the excluded middle. 10 The point is tha_t he groups it (as does Leibniz)
with the law of contradiction, 11 because instances of the law of identity
such as "Man is man", although tautologous and therefore "void of
consequences", still have a legislative or negative role to play. Though
nothing follows from "Man is man", to say that "Man is not man"
violates the law of contradiction. Indeed all of formal logic, that is,
what Kant calls "general logic", as opposed to "transcendental logic",
plays this negative role. In the Critique of Pure Reason Kant writes:
General logic resolves the whole formal procedure of the understanding and reason into
its elements, and exhibits them as principles of all logical criticism of our knowledge.
This part of logic, which may therefore be entitled analytic, yields what is at least the
negative touchstone of truth. Its rules must be applied in the examination and appraising
of the form of all knowledge before we proceed to determine whether their content
contains positive truth in respect to their object. But since the mere form of knowledge,
however completely it may be in agreement with logical laws, is far from being sufficient
to determine the material (objective) truth of knowledge, no one can venture with the
help of logic alone to judge regarding objects, or to make any assertion. 12
Here Kant is being quite traditional with respect to formal logic. Gen-
eral logic deals with the "form of thinking" as he says in his Logic,!3
and not the matter. Formal deductive inference, which for Kant is
essentially syllogistic, makes explicit in the conclusion what is implicit
in the premises. It is analytic, but not tautologous. Clearly Kant does
not view formal logic as pointless or trifling, as did Locke.
14
Still, he
warns us against the tendency to exaggerate the scope of formal logic.
He wrote,
There is, however, something so tempting in the possession of an art so specious, through
which we give to all our knowledge, however uninstructed we may be in regard to its
content, the form of understanding, that general logic, which is merely a canon of
judgment, has been employed as if it were an organon for the actual production of at
least the semblance of objective assertions, and has thus been misapplied. General logic,
when thus treated as an organon, is called dialectic. 15
The Idealists pressed Kant's restriction of the scope of formal logic
to the point of denigrating it. For Hegel the fundamental laws of
thought are "opposed to each other, they contradict one another". 16
Moreover,
the Law of Identity ... in its positive expression A = A is, in the first instance, nothing
26 BURTON DREBEN AND JULIET FLOYD
more than the expression of an empty tautology. It has therefore been rightly remarked
that this law of thought has no content and leads no further. 17
Similarly for the law of contradiction ... "A cannot at the same time
be A and not-A"; this is merely "a negative form" of the Law of
Identity. 18
Bradley, whose Principles of Logic (1883) is the main source for the
Idealist conception of logic which the young Russell embraced in his
early book, Foundations of Geometry (1897), says:
The principle of Identity is often stated in the fQrm of a tautology, "A is A." If this really
means that no difference exists on the two sides of the judgment, we may dismiss it at
once. It is no judgment at all. As Hegel tells us, it sins against the very form of judgment;
for, while professing to say something, it really says nothing .... We never at any time
wish to use tautologies. No one is so foolish in ordinary life as to try to assert without
some difference. We say indeed "I am myself" and "Man is man and master of his fate."
But such sayings as these are no tautologies .... Every judgment is essentially synthetic. 19
and also, in a footnote:
Every possible judgment, we shall see hereafter, is both analytic and synthetic.
20
For Bradley, then, "Whatever is, is" must be interpreted, if it is to be
a judgment at all. "The real axiom of identity," he says,
is this: What is true in one context is true in another. 21
Bradley also feels compelled to interpret the other two traditional laws
of thought - the law of contradiction and the law of the excluded
middle - in order to prevent them from collapsing into mere tauto-
logies.
22
Russell, in The Foundations of Geometry (1897), reiterated the Ideal-
ist position that "every judgment is both synthetic and analytic. ,,23 And
in his transitional Philosophy of Leibniz (1900), Russell argues that all
analytic judgments are tautologies, "and so not properly propositions
at all. ,,24 This latter position is stated in more detail in Moore's 1900
paper 'Necessity'. Says Moore,
there is much doubt whether any truths are analytic. Any proposition, it would seem,
must contain at least two different terms and their relation; and, this being so, the relation
may always be denied of the two terms without a contradiction. It takes two propositions
to make a contradiction: the law of contradiction itself excludes the possibility of any
single proposition being both true and false, or self-contradictory. And hence the defi-
nition of an analytic proposition as a proposition, the contradictory of which is self-
contradictory can apply to nothing ....
Moreover, the law of contradiction itself, than which nothing is commonly supposed
to be more plainly analytic, is certainly synthetic. For suppose some one to hold that not
every proposition is either true or false. You cannot deny that this is a proposition, unless
you are also willing to allow that the law which it contradicts is not a proposition; and
he may perfectly well maintain that this is one of those propositions which is true, and
the contradictory of which, your law, is false, although this is not the case with every
proposition. Whereas, if you urge that it is included in the notion of a proposition that
it should be either true or false, either your law becomes a pure tautology and not a
proposition, or else there is something else in the notion of a proposition beside the
property that it is either true or false, and then you are asserting a synthetic connexion
between this property and those others.
25
Thus the (newly-minted) anti-Idealists, Moore and Russell, reject
Kant's class of non-tautologous analytic judgments by exploiting the
Idealist argument against purely analytic judgments. For Moore and
Russell, all purported analytic judgments are mere tautologies, and
hence not judgments at all.
26
Now what of Frege? In the Grundlagen, Frege proposes his cel-
ebrated definition of 'analytic':
these distinctions between a priori and a posteriori, synthetic and analytic, concern, as I
see it, not the content of the judgement but the justification for making the judge-
ment .... When a proposition is called a posteriori or analytic in my sense, this is not a
judgement about the conditions, psychological, physiological and physical, which have
made it possible to form the content of the proposition in our consciousness; ... rather,
it is a judgement about the ultimate ground upon which rests the justification for holding
it to be true.
This means that the question is removed from the sphere of psychology, and assigned.
if the truth concerned is a mathematical one, to the sphere of mathematics. The problem
becomes, in fact, that of finding the proof of the proposition, and of following it up right
back to the primitive truths. If, in carrying out this process, we come only on general
logical laws and on definitions, then the truth is an analytic one.27
Hence a proposition is analytic for Frege if and only if it is derivable
from purely logical laws. Therefore, presumably, purely logical proposi-
tions are analytic also. Having defined the primitive notions of arithme-
tic in what for him counted as purely logical terms, arithmetic was for
Frege, as it had not been for Kant, analytic. Pure logic, given in the
Begriffsschrift, constituted for Frege the set of maximally general laws
governing any scientific or rational discourse whatsoever. 28 So for
Frege, as for Leibniz, logic, although 'analytic', was surely not empty
of content. 'Tautologous' was the last adjective Frege would have ap-
plied to a logical truth.
29
Thus Frege's notion of 'analytic' is a strength-
ening and extension of Kant's non-tautological analytic.
2.
In the Tractatus, when at 6.1 and 6.11 Wittgenstein asserts that the
propositions of logic are tautologies, say nothing, and are the analytical
propositions, he is continuing the Idealist and Moore-Russell traditions
of identifying the class of analytic propositions with tautologies
30
-
indeed, Wittgenstein is denying that they are genuine propositions
3
! -
and is directly attacking Frege's conception of logic.
At the same time, Wittengstein is attacking Russell's longstanding
claim that logic is synthetic. From 1900 through 1913 Russell took his
(and Frege's) "reduction" of mathematics to logic to show that logic
was synthetic, since mathematics was of course synthetic, as Kant,
according to Russell, had demonstrated. In The Principles of Mathemat-
ics (1903) Russell writes:
Kant never doubted for a moment that the propositions of logic are analytic, whereas he
rightly perceived that those of mathematics are synthetic. It has since appeared that logic
is just as synthetic as all other kinds of truth.
32
In his debate with Poincare, whose Science and Hypothesis he re-
viewed in 1905, Russell summarily dismissed Poincare's characterization
of formal logic as tautologous. Chapter I of Science and Hypothesis
(1905) opens thus:
The very possibility of mathematical science seems an insoluble contradiction. If this
science is only deductive in appearance, from whence is derived that perfect rigour which
is challenged by none? If, on the contrary, all the propositions which it enunciates may
be derived in order by the rules of formal logic, how is it that mathematics is not reduced
to a gigantic tautology?33
Poincare continues:
it must be granted that mathematical reasoning has of itself a kind of creative virtue, and
is therefore to be distinguished from the syllogism. The difference must be profound.
We shall not, for instance, find the key to the mystery in the frequent use of the rule by
which the same uniform operation applied to two equal numbers will give identical
results. All these modes of reasoning, whether or not reducible to the syllogism, properly
so called, retain the analytical character, and ipso facto, lose their power.
34
So for Poincare, formal logic is empty, tautologous, uninformative, but
mathematics is not. Hence he searches for a mode of reasoning that
explains the creativity of mathematics, that protects mathematics from
the sterility of logic. And Poincare finds this mode in mathematical
induction, a form of inference which contains within itself, or so he
says, an infinite number of syllogisms, and is therefore synthetic. He
thus returns, at least verbally, to Kant's view that mathematical truths
are synthetic a priori. (Goldfarb has convincingly shown that Poincare's
use of Kantian terminology has little to do with K--ant. 35)
Poincare's argument directly clashes with Russell's own for logicism.
Recall that the key to. the Frege-Russell reduction of arithmetic to
logic is its treatment of mathematical induction, resting on the notion
of the ancestral of a relation. Hence, in his review, Russell writes:
M. Poincare gives uq reasons for the view that deduction can never give new truths. The
fact is that the general principles of deduction are in this respect, to what he
conceives mathematical induction to be; that is to say, they lead to conclusions which
are other than themselves, so that in this sense they are synthetic. We shall conclude,
therefore, that mathematics does not, as M. Poincare affirms (p. 24), contain an inductive
element, and yet is not 'a vast tautology'. 36
Seven years later, in his Problems of Philosophy (1912), Russeil writes:
[Kant 1 perceived that not only the connexion of cause and effect, but all the propositions
of arithmetic and geometry, are 'synthetic', i.e., not analytic: in all these propositions,
no analysis of the subject will reveal the predicate. His stock instance was the proposition
7 + 5 = 12. He pointed out, quite truly, that 7 and 5 have to be put together to give 12:
the idea of 12 is not contained in them, nor even in the idea qf adding them together.
37
Russell also discussed the traditional "laws of thought" iIi Problems of
Philosophy. Here we find a defense of the view that logic is obvious,
apparently trivial, a priori, yet fruitful and informative. The truth of a
principle of logic says Russell, echoing Leibniz,
is impossible to doubt, and its obviousness is so great that at first sight it seems almost
trivial. Such principles, however, are not trivial to the philosopher, for they show that
we may have indubitable knowledge which is in no way derived from objects.
38
And again:
The name "laws of thought" is also misleading, for what is important is not the fact that
we think in accordance with these laws, but the fact that things behave in accordance
with them; in other words, the fact that when we think in accordance with them we think
truly. 39
In the same year (1912) Coffey published his neo-scholastic work,
The Science of Logic.
40
This is the one logic textbook we can assume
Wittengstein read, since he reviewed it in the Cambridge Review in
1913.
41
Coffey's version of the principle of identity is
simply the self-evident truth that Everything is identical with itself; Everything is its own
nature. It is in,Volved in every judgment - more directly in every affirmative judgment -
and that throughout every thought-process the objects represented by our
concepts and by O\lr terms remain identical with themselves. It thus expresses
the unambiguity of the judgment the immutable character of truth. It does not give
us any positive information about a thing, beyond what we possess by thinking of the
thing. But we cannot think definitely about anything without mentally marking it off
from all that is not itself. Hence the principle is not a bare tautology, capable of being
expressed by the statement that A is A. 42
Like Russell, Coffey stresses that the laws of thought are self-evident
truths grounded in the nature of things, not merely of thought.
Nowhere yet, at all events, has any valid reason been advanced why we should doubt
the soundness of man's spontaneous convictions that the necessary truth of those self-
evident first principles is rooted in the nature of things no less than in the nature of
thought. They not merely assure us that we cannot think that a thing can be other than
itself, or that we cannot conceive a thing being and not being at the same time and in
the same respect, or that we are forced to think that a thing must either possess a certain
attribute or not possess it: they assure us that the things themselves are so, as we think
them and that it is not merely a matter of how we must think about things, but also a
matter of how things really are. 43
Wittgenstein's review of Coffey is scathing:
The author's Logic is that of the scholastic philosophers, and he makes all their mistakes -
of course with the usual references to Aristotle. (Aristotle, whose name is so much taken
in vain by oUf logiciaps, would turn in his grave if he knew that so many logicians know
no more about logic today than he did 2,000 years ago.) The author has not. taken the
slightest notice of the great work of the modern mathematical 10gicians - work which
has brought about an advance in logic comparable only to that which made Astronomy
out of Astrology, and Chemistry out of Alchemy .
. . . The worst of such books as this is that they prejudice sensible people against the
study of Logic.
44
[Written ten years before C. 1. Lewis's letter!)
Not a word, either in Russell's twelve years of proclaiming the syn-
theticity of logic nor in Wittgenstein'spaean to the new logic, gives us
the slightest hint that the pejorative term "tautology" could in any way
be appropriately applied to "the great work of the modern mathema-
tical logicians". Yet in his eight lectures, The Philosophy of Logical
Atomism, given in the early spring of 1918, Russell professes a new
faith:
Everything that is a proposition of logic has got to be in some sense or other like a
tautology. It has got to be something that has some peculiar quality, which I do not know
how to define, that belongs to logical propositions and not to others.
45
This totally unexpected use of "tautology" is given no justification
whatsoever in these lectures. Russell's concern is to find an adequate
characterization of logic, since total generality will not do, that is,
not all "propositions which can be expressed in the language of pure
variables ... are ... propositions of logic". 46 A worthy goal, but in no
way achieved by a label which is unexplained and indeed which Russell
"do[ es] not know how to define".
In his prison document of 1918, Introduction to Mathematical Philos-
ophy (published in 1919), Russell further professes his new-found faith
in the tautologousness of logic:
All the propositions of logic have a characteristic which used to be expressed by saying
that they were analytic, or that their contradictories were self-contradictory. This mode
of statement, however, is not satisfactory. The law of contradiction is merely one among
logical propositions; it has no special pre-eminence; and the proof that the contradictory
of some proposition is self-contradictory is likely to require other principles of deduction
besides the law of contradiction. Nevertheless, the characteristic of logical propositions
that we are in search of is the one which was felt, and intended to be defined, by those
who said that it consisted in deducibility from the law of contradiction. This characteristic,
... for the moment, we may call tauto[ogy.47
If we have the temerity to ask, Why should we call this characteristic
"tautology"?, Russell has only the following to say:
It is clear that the definition of "logic" or "mathematics" must be sought by trying to
give a new definition of the old notion of "analytic" propositions. Although we can no
longer be satisfied to define logical propositions as those that follow from the law of
contradiction, we can and must still admit that they are a wholly different class of
propositions from those that we come to know empirically. They all have the characteristic
which, a moment ago, we agreed to call "tautology" .... For the moment, I do not
know how to define "tautology". [FN: The importance of "tautology" for a definition of
mathematics was pointed out to me by my former pupil Ludwig Wittgenstein, who was
working on the problem. I do not know whether he has solved it, or even whether he is
alive or dead.) It would be easy to offer a definition which might seem satisfactory for a
while; but I know of none that I feel to be satisfactory, in spite of feeling thoroughly
familiar with the characteristic of which a definition is wanted.
48
So there is no argument in Russell for his use of the word
"tautology"; there is just the footnote referring to Wittgenstein.
When and how did Wittgenstein first point out to Russell the "impor-
tance of 'tautology'''? Presumably, in a November 1913 letter from
Norway. Wittgenstein writes:
Lieber Russell,
... a logical prop [osition] is one the special cases of which are either tautologous - and
then the prop [osition] is true - or "self-contradictory" (as I shall call it) and then it is
false.
49
Most remarkably, Wittgenstein bases his new characterization of
logic on the claim that there is a decision procedure for all of logic. He
begins with what is probably the first statement of a decision procedure
for truth-functional logic. What he calls his "ab-Notation" is essentially
equivalent to the truth-table notation, hence his "ONE symbolic rule"
is essentially equivalent to what today would be the standard truth-
table method for showing truth-functional validity, contravalidity, or
consistency:5o
Now listen: I will first talk about those logical prop[osition]s which are or might be
contained in the first 8 Chapters of Princ[ipia] Math[ ematica] .... ONE symbolic rule is
sufficient to recognize each of them as true or false. And this is the one symbolic rule:
write the prop [osition] down in the ab-Notation, trace all Connections (of Poles) from
the outside to the inside Poles: Then if the b-Pole is connected to such groups of inside
Poles ONLY as contain opposite poles of ONE prop[osition], then the whole prop [osition]
is a true, logical prop[osition]. If on the other hand this is the case with the a-Pole the
prop[ osition] is false and logical. If finally neither is the case the prop [ osition] may be
true or false but is in no case logical. 51
Wittgenstein then claims that his decision procedure can be extended:
52
Of course the rule I have given applies first of all only for what you called elementary
prop[ osition ]s. But it is easy to see that it must also apply to all others. For consider
your two Pps in the Theory of app[arent] var[iable]s *9.1 and *9.11. Put there instead
of >x, (3y). >y.y = x and it becomes obvious that the special cases of these two Pps like
those of all the previous ones becomes tautologous if you apply the ab-Notation. The
ab-Notation for Identity is not yet clear enough to show this clearly but it is obvious that
such a Notation can be made up ... there is one Method of proving or disproving all
logical prop[osition]s and this is: writing them down in the ab-Notation and looking at
the connections and applying the above rule. 53
The crucial question for us is, Why did Wittgenstein think that a
general decision procedure would show that the propositions of logic
are generalizations of tautologies? An answer begins to emerge in his
very next letter to Russell. This letter (R. 23), of late November or
early December 1913, from Norway opens thus:
I want to repeat again, in a different form, what I wrote about logic in my last letter.
All the propositions of logic are generalizations of tautologies and all generalizations of
tautologies are propositions of logic. There are no other logical propositions. (I regard
this as definitive. )54
And it concludes:
The big question now is, how must a system of signs be constituted in order to make
every tautology recognizable as such IN ONE AND THE SAME WAY? This is the
fundamental problem of logic [Grundproblem der Logik]!55
But in the middle of this letter we not only read his admission that a
general decision procedure is not quite at hand,56 but, more impor-
tantly, why Wittgenstein's use of the term "tautology'l is not arbitrary-
as Russell's 1918-19 usage appears to be - given the history of its use
we sketched from Kant onward.
As to what tautologies really are, however, I myself am not yet able to say quite clearly
but I will try to give a rough explanation. It is the peculiar (and most important [hOchst
wichtigeJ) mark of non-logical propositions that one is not able to recognize their truth
from the propositional sign alone. If I say, for example, "Meier is stupid", you cannot
tell by looking at this proposition whether it is true or false. But the propositions of
logic - and only they - have the property that their truth or falsity, as the case may be,
finds its expression in the very sign for the proposition. I have not yet succeeded in
finding a notation for identity that satisfies this condition; but I have NO doubt that it
must be possible to find such a notation. 57 For compound propositions ("elementary
propositiorts") the ab-notation is sufficient. 58
It therefore seems that the picture gripping Wittgenstein is this: if
truth can be discerned from the propositional sign alone, then no claim
is made upon reality, there is nothing corresponding to the proposition
that makes it true or false, and hence, intuitively, the proposition is
uniformative, superfluous, empty, perhaps not even a genuine proposi-
tion: in short, tautologous.
59
This suggestion of what moved Witt-
genstein - and what led him to think that Russell would accept his
characterization of logic - gains some support from various passages in
the 'Notes on Logic', especially when we keep in mind that for more
than a year the two questions absolutely central to Russell and Witt-
genstein were: What makes a proposition true? (i.e., What corresponds
to a proposition when it is true?) and, What is it to understand a
proposition, whether true or false?60 We must also keep in mind that
in the 'Notes' Wittgenstein is building on but drastically restructuring
Frege's Sinn/Bedeutung (sense/meaning61) distinction.
First:
The meaning of a proposition is the fact which actually corresponds to it.
62
Second:
A [molecular] function is like a line dividing points of a plane into right and left ones;
then "p or not-p" has no meaning because it does not divide the plane.
63
Third:
A proposition is a standard to which facts behave [sich verhalten, are related, edd.], with
names it is otherwise; it is thus bi-polarity and sense comes in; just as one arrow behaves
[sich verhiilt, is related, edd.] to another arrow by being in the same sense or the opposite,
so a fact behaves to a proposition. 64
Fourth:
Every proposition is essentially true-false: to understand it, we must know both what
must be the case if it is true, and what must be the case if it is false. Thus a proposition
has two poles, corresponding to the case of its truth and the case of its falsehood. We
call this the sense of a proposition.
65
Finally:
Signs of the form "p v - p" are senseless, but not the proposition "(p).p v - p". If I
know that this rose is either red or not red, I know nothing. The same holds of all
ab-functions. 66
Further substantiation is gained from the only other source
67
of
Russell's acquaintance with Wittgenstein's views on logic and tautology
before 1919: 'Notes Dictated to G. E. Moore in Norway in April
1914,.68 Indeed, the first two remarks in these 'Moore Notes' exploit
what was not explicit in the "Notes on Logic", namely, the show/say
distinction, to state that logical propositions say nothing, and hence
that they are not genuine propositions:
Logical so-called propositions shew [the ] logical properties of language and therefore of
[the] Universe, but say nothing. [Cf. 6.12.]
This means that by merely looking at them [the so-called logical propositions] you can
see these properties; whereas, in a proposition proper, you cannot see what is true by
looking at it. [Cf. 6.113.]69
But now there is an all-important shift of emphasis; there is no
longer tHe insistence on a general decision procedure for seeing (i.e.,
recognizing) a true logical proposition:
We want to say ... what properties a symbol must have, in order to be a tautology.
Many ways of saying this are possible:
One way is to give certain symbols; then to give a set of rules for combining them; and
then to say: any symbol formed from those symbols, by combining them according to
one of the given rules, is a tautology. This obviously says something about the kind of
symbol you can get in this way.
This is the actual procedure of [the] old Logic: 70 it gives so-called primitive propositions;
so-called rules of deduction; and then says that what you get by applying the rules to the
propositions is a logical proposition that you have proved. The truth is, it tells you
something about the kind of proposition you have got, viz. that it can be derived from
the first symbols by these rules of combination (= is a tautology).
Therefore, if we say one logical proposition follows logically from another, this means
something quite different from saying that a real proposition follows logically from
another. For so-called proof of a logical proposition does not prove its truth (logical
propositions are neither true nor false) but proves that it is a logicfll proposition = is a
tautology. [Cf. 6.1263.fl
The die has been cast: henceforth the tautologousness of logic will
be insisted upon come what may;56 it will simultaneously play the role
both of informing and following from the basic Tractatus doctrines -
almost all of which are at least adumbrated in the 'Moore Notes' -
limning the nature of language (thought) and its relation to reality, to
the world. "Tautology" is no longer defined or specified in terms of
"one Method of proving or disproving all logical prop[osition]s" (R.
22). In the Tractatus, a decision procedure, when applicable, is just an
aid to recognizing
a tautology as such ... [For example] in cases in which no sign of generality occurs in
the tautology. (Tractatus 6.1203.)
And as we have already seen in the "Moore Notes",
Proof in logic is only a mechanical expedient to facilitate the recognition of tautology,
where it is complicated. (Tractatus 6.1262.)
No wonder there is not even an attempt at an argument in Russell's
calling logic "tautologous". He did not accept, or perhaps even under-
stand, Wittgenstein's basic doctrines. In a letter received by Russell in
January 1915 Wittgenstein writes:
I find it inconceivable that Moore wasn't able to explain my ideas to you. Were you able
to get anything at all out of his notes? I'm afraid the answer is, No.72
And in a letter to Russell dated 22 May 1915:
I'm extremely sorry that you weren't able to understand Moore's notes. I feel that they're
very hard to understand without further explanation, but I regard them essentially as
definitive. 73
3.
Russell, of course, was not alone in adopting Wittgenstein's word.
Ramsey, Carnap and Lewis followed suit - though we find virtually no
argument in their work for the application of (the term) 'tautology' to
logical truth (except to cite Russell and Wittgenstein), and when we do
we see none of them agreed with each other about why application of
the term to logic is warranted. They did agree" however, in rejecting
the Tractatus distinction between logic and mathematics; each accepted
the logistic reduction in some form or other and applied 'tautologous'
to the propositions of mathematics as well. That is, each ignored
Wittgenstein's Tractarian criticisms of the logistic reduction, in parti-
cular, his criticism of the Frege-Russell definition of the ancestral as
circular. 74 More importantly, each rejected the core of the Tractatus
account of logic, namely, the say/show distinction and the Sinnlos/Un-
sinnig (lacking in sense/nonsensical) distinction. Logical propositions
in the Tractatus are Sinnlos, tautologous, having no subject matter. In
the Tractatus Wittgenstein is refusing to justify the idea that deduction
is informative in the traditional way, that is, by viewing a given deduc-
tion as making explicit in the conclusion what is implicit in the premises.
He rejects, in other words, the Kantian notion of the non-tautologous,
analytic character of formal logic.
Kant and others had seen in the law of identity, "A is A", or,
"Whatever is, is", a fundamental law of logic, an underlying boundary
point beyond which thought and the world cannot go, which, despite
its fundamental character, is expressed in a tautology; an empty
proposition, redundant and "void of consequences". Such are all the
(pseudo-) propositions of logic in the Tractatus. "We can get on without
logical propositions,,,75 as Wittgenstein says, though they show us the
"scaffolding of the world".
Ramsey dismissed the say/show distinction with his famous quip,
"But what we can't say we can't say, and we can't whistle it either". 76
In his 'Foundations of Mathematics' Ramsey invokes the passage from
Russell's Introduction to Mathematical Philosophy (quoted above),
where Russell insists that the propositions of logic are to be called
'tautologies,.77 Then, after introducing Wittgenstein's truth-table analy-
sis, Ramsey writes:
We have here, thanks to Mr. Wittgenstein, to whom the whole of this analysis is due, a
clearly defined sense of tautology; but is this, it may be asked, the sense in which we
found tautology to be an essential characteristic of the propositions of mathematics and
symbolic logic? The question must be decided by comparison. Are the propositions of
symbolic logic and mathematics tautologies in Mr. Wittgenstein's sense?78
But the "sense in which we found" logic and mathematics to be tautol-
ogous is simply Russell's words in Introduction to Mathematical Philos-
ophy. Russell, however, as we have seen, had offered 'tautology' as a
mere label for that characteristic of logic which he confessed he could
not define, although he felt "thoroughly familiar" with it. But now
Ramsey asks whether the Tractatus captures the characteristic with
which Russell was grappling! With the help of Russell, Wittgenstein
had created a climate in which it was taken to be already demonstrated
that logic is tautologous! The only task remaining, according to Ramsey,
is to give a precise definition of it.
'Tautology' and its cognates occur throughout Lewis and Langford's
Symbolic Logic (1932). For Lewis, as we saw on page 23 above, "tautol-
ogous" means "exhausts all possibilities". 79 Although this use of
"tautology" was inspired by Wittgenstein's truth-table analysis, Lewis
in no way restricted his use to the applicability of truth-tables. In fact,
he labelled all the theorems of his various systems of modal logic
'tautologies'. Lewis's basic explanation of 'tautology' rests on the notion
of 'analytic'. He writes:
any logical principle (and, in fact, any other truth which can be certified by logic alone)
is tautological in the sense that it is an analytic proposition.
80
And 'analytic' for Lewis is akin to 'analytic' for Kant. Lewis writes:
The only truth which logic requires, or can state, is that which is contained in our own
conceptual meanings - what our language or our symbolism represents. Or to put it
otherwise: there are no laws of logic, in the sense that there are laws of physics or
biology; there are only certain analytic propositions, explicative of 'logical' meanings,
and these serve as the 'principles' which thought or inference which involves these
meanings must, in consistency, adhere to.
8
!
Lewis is exploiting the traditional account that logic informs by making
the implicit explicit. For Lewis, a tautology is true in virtue of meaning.
And this directly undercuts the Tractatus notion of tautologies as
Sinnlos. In fact, the claim that the propositions of logic are Sinnlos was
strongly criticized by Lewis in spite of his insistence that they are
tautologous.
There has always been some confusion about the nature of analytic propositions, which
are true by definition or follow from the meanings of terms; and there has been a
tendency to say that such propositions 'make no assertion' or 'are not significant,' or
even that they 'are not propositions.' They are not of course, 'assertions' or 'significant'
in the sense that synthetic propositions are, such as generalizations from experience, like
the physical law v = gt. Such natural laws state something which could quite conceivably
be false. Whether what could not conceivably be false - such as logical principles and
arithmetical sums - is 'significant' or not, or 'makes an assertion' or not, is merely a
question of propriety in the use of language, and is not worth arguing. We can only ask
that those who deny significance or the quality of assertion to analytic propositions should
be a little more explicit as to what they mean by 'significant I or 'an assertion.'82
(In fairness to Lewis, we must point out that he added a footnote to
his homage to Wittgenstein with which we began this paper. The foot-
note reads:
With the further detail of Wittgenstein's conceptions the present authors would not
completely agree.
83
)
4.
In the third edition (1961) of the Merriam-Webster New International
Unabridged Dictionary (for which Max Black was a consultant) we find:
tautologous 1: TAUTOLOGICAL 2a: ANALYTIC b: true in terms of the sentential
connectives of a truth table c: true purely by virtue of the meanings of component terms.
And:
analytic ... logic: of or relating to a truth, a proposition, or a statement that is true in
all possible worlds, that is true independently of any facts by reference to meanings
alone, or that is logically true or definitionally reducible to logical truth.
Finally, in the second edition of the Oxford English Dictionary (1989):
tautological l.a. Pertaining to, characterized by, involving, or using tautology; repeating
the same word, or the same notion in different words ... b. Mod. Logic. Characterized
by or involving tautology (in sense f.) 1922 tr Wittgenstein's Tractatus 97 In the one case
the proposition is true for all the truth-possibilities of the elementary propositions. We
say that the truth-conditions are tautological 1926. F. P. RAMSEY in Proc. London
Math. Soc. XXV 341. The idea to be defined is one of the essential sides of mathematical
propositions, their content, and their form. Their content must be completely generalized,
and their form tautological ... 1950 R. CARNAP Logical Found. Probability iv 289 With
respect to the tautological evidence 't'.
tautologous ... repeating what has been said ... 1940 W. V. QUINE Mathematical Logic
i 50 Statements which are true by virtue solely of the truth-functional modes of composi-
tion will be called tautologous. 84
tautology ... a. A repetition of the same statement. b. The repetition (esp. in the immedi-
ate context) of the same word or phrase, or of the same idea or statement in other
words: usually as a fault of style ... f. Mod. Logic. A compound proposition which is
unconditionally true for all the truth-possibilities of its elementary propositions and by
virtue of its logical form.
1919 B. RUSSELL, Introd. Math. Philos. xviii 203 The characteristic of logical proposi-
tions that we are in search of is the one which was felt ... by those who said that it
consisted in deducibility from the law of contradiction. This characteristic we may call
tautology. Ibid. 205 The importance of 'tautology' for a definition of mathematics was
pointed out to me by ... Ludwig Wittgenstein, who was working on the problem. 1922: tr.
Wittgenstein's Tractatus 97 The tautology ... is unconditionally true ... 1964 M. BLACK
Compan. Wittgenstein's Tractatus xliii 231 Johnson's ... 'formal truth' and 'formal fals-
ity' ... seem to correspond exactly to W.'s 'tautology' and 'contradiction'.
And (for fun):
analyticity ... Philos. The property, in propOSItIons or statements, of being ana-
lytic ... 1953 W. V. O. QUINE From a Logical Point of View ii. 21 Kant's intent, evident
more from the use he makes of the notion of analyticity than from his definition of it,
can be restated.
One way to avoid the charge of misspeaking is to change the language. 85
NOTES
I From Lewis's cover letter for his 'A Pragmatic Conception of the A Priori', in The
Journal of Philosophy 20 (1923): 169-77.
2 C. I. Lewis and C. H. Langford, Symbolic Logic (New York: Century Company, 1932),
p.24.
3 Immanuel Kant, Logic eds. and trans. Robert S. Hartman and Wolfgang Schwarz (New
York: Dover Publications, 1988), I Section 36 (p. 117). (Page references, in parentheses,
are to this translation.)
4 Ibid., I Section 37 (p. 118).
5 Immanuel Kant, Prolegomena to Any Future Metaphysics (1783), ed. and trans. Lewis
White Beck, (Indianapolis: Bobbs-Merrill Company), Preamble, Section 2 (p. 14).
6 In Leibniz's Philosophical Papers and Letters, ed. Leroy E. Loemker. The editor notes
(p. 227) that
"coccysm" was proverbial for wordiness or redundancy, from the reputation of John
Cocceius, Professor of Theology at Leyden, and his followers.
7 Leibniz, 'On the General Characteristic' (c. 1679) (in Philosophical Papers and Letters,
(hereafter L) ed. Leroy E. Loemker (second edition, Chicago: University of Chicago
Press, 1976): pp. 221-8) pp. 225-6.
8 'Necessary and Contingent Truths' (c. 1686) (in Leibniz, Philosophical Writings (here-
after LPW) ed. G. H. R. Parkinson (London: J. M. Dent & Sons, 1973): pp. 96-105),
p. 96. In 'First Truths' (c. 1680-84) Leibniz calls the principle of identity a "First Truth"
(L, p. 267). In the second letter to Samuel Clarke (1715-16) (L, p. 677) he claims it
underlies all of mathematics. In the Monadology (1714) Sections 31-2 (L, p. 646) Leibniz
says:
31. Our reasonings are based on two great principles: the principle of contradiction,
by virtue of which we judge to be false that which involves a contradiction, and to
be true that which is opposed or contradictory to the false;
32. and the principle of sufficient reason, by virtue of which we consider that no fact
can be real or existing and no proposition can be true unless there is a sufficient
reason, why it should be thus and not otherwise, even though in most cases these
reasons cannot be known to us.
See also 'On the General Characteristic' (1679) (L, p. 226).
9 Kant, Logic, Introduction VIIB (p. 58).
10 Ibid. Kant, following Leibniz, also cites the principle of sufficient reason as a funda-
mental law. In the quotation from Leibniz's Monadology in footnote 8 above, note his
claim that the principle of contradiction and the principle of sufficient reason underlie
all our reasoning. Kant echoes this idea in Logic, Introduction VIIB (p. 57).
11 See Leibniz's remarks at L, p. 385, 633, and Benson Mates, The Philosophy of Leibniz
(Oxford University Press, 1986) p. 153.
12 Kant, Critique of Pure Reason, trans. Norman Kemp Smith (New York: St. Martin's
Press, 1967) (hereafter KRV) A60/B85.
13 Kant, Logic, p. 15.
14 See Locke, An Essay Concerning Human Understanding (ed. Peter H. Nidditch (Ox-
ford University Press, 1975) Book IV, Ch. xvii, s. 4.) Locke says (p. 493ff) of the abuse
of language, which he calls "affected obscurity", that
To this abuse, and the mischiefs of confounding the signification of Words, Logick,
and the liberal Sciences, as they have been handled in the Schools, have given
Reputation; and the admired Art of Disputing, hath added much to the natural
imperfection of Languages, whilst it has been made use of, and fitted, to perplex
the signification of Words, more than to discover the knowledge and Truth of Things:
And he that will look into that sort of learned Writings, will find the Words there
much more obscure, uncertain, and undetermined in their Meaning, than they are
in ordinary Conversation.
15 KRV A60-61/B85-86.
16 Hegel, The Science of Logic (ed. H. D. Lewis, trans. A. V. Miller (Atlantic Highlands:
Humanities Press International, 1989, p. 411 (Vol 1, Bk 2, Ch 2, A).
17 Ibid., p. 413 (Vol. 1, Bk. 2, Ch 2, Remark 2).
18 Ibid., p. 416 (Vol. 1, Bk. 2, Ch 2 A, Remark 2).
19 F. H. Bradley, The Principles of Logic (London: Kegan Paul, Trench & Co., 1883),
pp. 131-2.
20 Ibid., p. 48, fn.
21 Ibid., p. 133.
22 Ibid., Book I, Ch V, Sections 11, 12, 18,24.
23 Russell, An Essay on the Foundations of Geometry (Cambridge University Press,
1897), p. 58.
24 Russell, A Critical Exposition of the Philosophy of Leibniz: With an Appendix of
Leading Passages (Cambridge University Press, 1900), pp. 16--17.
25 G. E. Moore, 'Necessity', Mind (1900), 289-304. Quotation is from p. 295.
26 For Moore's and Russell's reactions to Idealism, see Peter Hylton, Russell, Idealism,
and the Emergence of Analytic Philosophy (Oxford University Press, 1990).
27 Gottlob Frege, The Foundations of Arithmetic (trans. J. L. Austin, second edition,
Evanston: Northwestern University Press, 1980), pp. 3-4. ,
28 For an account of Frege's conception of logic as maximally general laws see Thomas
G. Ricketts.' triad of papers 'Objectivity and Objecthood: Frege's Metaphysics of Judg-
ment', (L. Haaparanta and J. Hintikka, eds., Frege Synthesized (D. Reidel, 1986,
'Generality, Meaning and Sense in Frege', (Pacific Philosophical Quarterly 67 (1986):
172-95) and 'Frege, the Tractatus, and the Logocentric Predicament' (Nous, XIX, No.1
March 1985: 3-15). For Frege's 'logistic reduction' see George Boolos, 'Reading the
Begriffsschrift' (Mind 94 (July 1985): 331-4) and Charles Parsons's 'Frege's Theory of
Number' (in Parsons's Mathematics in Philosophy, Selected Essays (Ithaca: Cornell Uni-
versity Press, 1983): pp. 150-75). For Frege's relation to Kant (and as a precursor of
Wittgenstein) see Joan Weiner, Frege in Perspective (Cornell University Press, 1990).
29 In Section 73 of The Foundations of Arithmetic (p. 85) Frege writes:
Our next aim must be to show that the Number which belongs to the concept F is
identical with the Number which belongs to the concept G if the concept F is equal
[gleichzahlig] to the concept G. This sounds, of course, like a tautology. But it is
not; the meaning of the word "equal" ["g/eichzahlig"] is not to be inferred from its
etymology, but taken to be as I defined it above.
30 Fritz Mauthner, named in the Tractatus at 4.0031 - where Wittgenstein distinguishes
his "Sprachkritik" from Mauthner's - pursued the rare, if not unique, philosophical
program of combining Locke with Schopenhauer. As a result, he held not only that the
Laws of Thought and all analytic propositions are tautologies, but even that all truths,
once known, become tautologies. See Mauthner, Beitriige zu einer Kritik der Sprache
(first edition, Stuttgart: J. G. Cotta'sche, 1902), Vol. 3, pp. 326--7 and p. 364ff, especially
p. 378. We thank Professor Ignacio A. Angelelli for calling our attention to the discussion
of tautology in Mauthner.
31 In Wittgenstein's Notebooks 1914-1916 (eds. G. H. von Wright and G. E. M. An-
scombe, trans. G. E. M. Anscombe (Second edition, Oxford: Basil Blackwell, 1979);
hereafter NB), in the entry for 29 October, 1914 (NB, p. 21) we read:
There are no such things as analytic propositions.
32 Bertrand Russell, The Principles of Mathematics (Cambridge University Press, 1903;
second edition with a new introduction (pp. v-xiv) 1938), p. 457. Where Russell differs
from Kant and agrees with Frege is in denying that (Kantian) intuition is needed for
mathematics; hence for both Frege and Russell mathematics is of the same character as
logic. Still, it would be an overstatement to equate Russell's use of "synthetic" with
Frege's "analytic", since Frege would hardly say "logic is just as analytic as all the other
kinds of truth". !lor a discussion of Kantian intuition see Jaakko Hintikka's 'Kantian
Intuitions' (Inquiry 15 (1972): 341-5); his 'On Kant's Notion of Intuition (Anschauung),
(in Terence Penelhum and J. J. MacIntosh, eds., Kant's First Critique (Belmont: Wad-
sworth, 1969): 38-53); as well as Parsons's 'Kant's Philosophy of Arithmetic' (Mathematics
and Philosophy, Selected Essays: pp. 110-49).
33 Henri Poincare, Science and Hypothesis (Walter Scott Publishing Company, 1904;
republished by Dover Publications, 1952), pp. 1-2, Dover reprint. Russell's references
to this work refer to the original edition.
34 Ibid., p. 3.
35 Warren Goldfarb, 'Poincare against the Logicists' Studies in the Philosophy
of Science, Vol. XI, History and Philosophy of Modern Mathematics, William Aspray
and Philip Kitcher, eds., (University of Minnesota Press, 1988), pp. 61-81).
36 Russell's review appeared in Mind, 1905 and is reprinted in his Philosophical Essays
(London: George Allen & Unwin, 1966). The quotation is from p. 73 of Philosophical
Essays.
37 Russell, Problems of Philosophy (London: William & Norgate, 1912), Chapter 8,
Paragraph 3.
38 Ibid., Chapter 7, Paragraph 5.
39 Ibid., Chapter 7, Paragraph 8.
40 P. Coffey, The Science of Logic (London: William and Norgate, 1912).
41 Wittgenstein's Review, 'On Logic and How Not to Do It' was printed in The Cambridge
Review 34 (1912-13), p. 351. It is reprinted in Brian McGuinness's Wittgenstein, A Life
(Berkeley: University of Calfornia Press, 1988), pp. 169-70.
42 Coffey, The Science of Logic, Vol. I, p. 23.
43 Ibid., Vol. I, p. 26.
44 McGuinness, Wittgenstein, A Life, p. 170.
45 Russell, Philosophy of Logical Atomism (first published in the Monist 28, October
1918, pp. 495-527; then again at La Salle, IL: Open Court Publishing Company, 1985).
References are to the Open Court edition. The quotation is from pp. 107-8.
46 Ibid., p. 107:
There are ... two propositions that one is used to in mathematical logic, namely,
the multiplicative axiom and the axiom of infinity. These can be expressed in logical
terms but cannot be proved or disproved by logic.
47 Russell, Introduction to Mathematical Philosophy (New York, The Macmillan Co.,
1919), Chapter XVIII, Paragraph 19. As is often the case, remnants of the old faith
persist in the new. In Chapter XVI, Paragraph 7 we find
Logic, I should maintain, must no more admit a unicorn than zoology can; for logic
is concerned with the real world just as truly as zoology though with its more abstract
and general features.
48 Ibid., Chapter XVIII, Paragraph 21.
49 Ludwig Wittgenstein: Letters to Russell, Keynes and Moore, ed. G. H. von Wright,
English trans. by Brian McGuinness (Ithaca: Cornell University Press, 1974), hereafter
LRKM, p. 37 (letter R.22 according to von Wright's numbering scheme).
50 Wittgenstein sketched his ab-Notation in a short document, now called 'Notes on
Logic', whose mostly German Ur-text(s), unfortunately lost, were obtained by Russell
from Wittgenstein in early October 1913. (Wittgenstein did not see or speak to Russell
from mid-October 1913 until December 1919.) For the provenance of these 'Notes' see
Brian McGuinness's 'Bertrand Russell and Ludwig Wittgenstein's "Notes on Logic'"
(Revue Internationale de Philosophie (1972): 444--60). ('Notes on Logic' (hereafter NL)
is printed as Appendix I in the second edition of Wittgenstein's Notebooks 1914-1916.)
In two letters (R.20, R.21) of November 1913 that immediately precede the present
letter (R.22), Wittgenstein peremptorily answered some questions of Russell about the
Notation. Wittgenstein chided Russell in letter R.20:
If you had only remembered the WF scheme of -p you would never have asked
this question (I think). In fact all rules of the ab symbolism follow directly from the
essence of the WF scheme. (LRKM, p. 33).
The phrases "WF scheme" and "TF scheme", accompanied with very brief discussion,
appear in 'Notes on Logic' (pp. 94-9); according to McGuinness (Wittgenstein: A Life,
p. 160), Wittgenstein showed Russell the truth-table notation (i.e., "the WF scheme")
in the winter of 1912.
51 LRKM, p. 36 (letter R.22).
52 Post also tried to extend his truth-table decision procedure to all of logic. But by the
end of 1921 he (correctly) conjectured that there was no such method, and that "a
complete symbolic logic is impossible". (See E. L. Post, 'Absolutely Unsolvable Problems
and Relatively Undecidable Propositions: Account of an Anticipation', unpublished paper
printed in Martin Davis, ed., The Undecidable (Raven Press Books, 1965), pp. 348, 397,
and 416. The phrase quoted is from page 416.) Unlike the young Wittgenstein, however,
the failure to extend his method leads the young Post to view logic (and mathematics)
as essentially creative, not empty (tautologous).
53 LRKM, p. 37 (R.22).
54 LRKM p. 41 (English translation of letter R.23; German original, p. 39).
55 LRKM, p. 43 (English translation of letter R.23; German original, p. 41).
56 One cannot help but be struck by an enormous historic irony: Turing and Wittgenstein
discussed in the late 1930s the nature of logical and mathematical rules, indeed of
algorithms. And even more remarkably, they each continued to call the formulae of
Principia Mathematica (at least, those in Volume I) "tautologies". See Wittgenstein's 1939
Lectures on the Foundations of Mathematics, ed. Cora Diamond (Chicago: University of
Chicago Press, 1975).
57 Wittgenstein had previously written to Russell (29 October 1913):
Identity is the very Devil and immensely important; very much more so than I
thought. It hangs - like everything else - directly together with the most fundamental
questions, especially with the questions concerning the occurrence of the SAME
argument in different places of a function (LRKM, p. 31 (R.19)).
To understand why Wittgenstein never would find his "notation for identity" see the
discussions of "occurrence of the same argument in different places of a function [quanti-
ficational formula)" in Burton Dreben and Warren Goldfarb, The Decision Problem:
Solvable Classes of Quantificational Formulas (Addison-Wesley, 1979) and Harry R.
Lewis, Unsolvable Classes of Quantificational Formulas (Addison-Wesley, 1979). To
appreciate the full magnitude of the question see Hao Wang, 'Dominoes and The AEA
case of the Decision Problem,' Proc. Symposium on the Mathematical Theory of Automata
(New York: Polytechnic Press, 1963) and Stal O. Aanderaa and Harry R. Lewis, 'Linear
Sampling and the V3V-case of the Decision Problem', Journal of Symbolic Logic 39:
519-47. Finally, to see how truly difficult is the problem of identity see Warren Goldfarb,
'The Unsolvability of the Godel Class with Identity', The Journal of Symbolic Logic 49:
1237-52.
58 LRKM, p. 42 (English translation of letter R.23; German original, pp. 39-40).
59 Even Quine says:
Ontology is internally indifferent also, I think, to any;theory that is complete and
decidable. Where we can always settle truth values mechanically, there is no evident
internal reason for interest in the theory of quantifiers nor, therefore, in values of
variables. (Ontological Relativity and Other Essays (New York: Columbia University
Press, 1969), p. 63).
60 See McGuinness, Wittgenstein, A Life, pp. 89-91, 162ff; and David Pears's two papers
'The Relation between Wittgenstein's Picture Theory of Propositions and Russell's Theo-
ries of Judgment', (Philosophical Review (April 1977): 177-96) and 'Russell's 1913 Theory
of Knowledge Manuscript' (in E. Wade Savage and C. Anthony Anderson (eds.), Reread-
ing Russell: Essays on Bertrand Russell's Metaphysics and Epistemology (Minnesota
Studies in the Philosophy of Science, XII; Minneapolis: University of Minnesota Press,
1989: pp. 169-82)).
61 The justification for reading "meaning" as "Bedeutung" and "sense" as "Sinn" in the
'Notes on Logic' rests on the fact that the original source of the 'Notes' was mostly in
German (see footnote 50 above) and on Wittgenstein's letter to Russell of November
1913 (R.21), LRKM, p. 35.
62 NL, p. 94.
63 NL, p. 94.
64 NL, p. 95.
65 NL, p. 98-99.
66 NL, p. 104.
67 Pedantry - or scholarly integrity - demands that we note letters R.24, M.3 and MA
where "tautology" occurs:
15 December 1913 (to Russell):
The question as to the nature of identity cannot be answered until the nature of
tautology has been explained. But that question is fundamental to the whole of logic
list die Grundfrage aller Logik]. (R.24, LRKM p. 45.)
30 January 1914 (to Moore):
Have you ever thought about the nature of a tautology? That's what I am now
bothered with. (M.3, LRKM p. 146.)
18 February 1914 (to Moore):
It ... all turns on the question as to the nature of deduction. And - I think - the
clue to it all lies in the fact that cPx :Jx IjJX only then expresses the deductive relation
when this prop[osition] is the generalization of a tautology. (MA, LRKM p. 147.)
68 These 'Moore Notes' are printed as Appendix II to Wittgenstein's Notebooks
1914-1916: pp. 108-19. Hereafter we will refer to them (by page .number in the Note-
books) as MN. Unlike the 'Notes on Logic', where the word "tautology" never appears,
and only one occurrence of a cognate (i.e., "tautologous") in a context offering little
support for Wittgenstein's characterization of logic, here "tautology" and its cognates
frequent Wittgenstein's discussion. Also, unlike the letters to Russell of November and
December 1913, the so-called propositions of logic are no longer generalizations of
tautologies, but are tautologies themselves.
69 MN, p. 108. The bracketed references are to the Tractatus.
70 The logic of Frege and Russell.
71 MN, p. 109. The citation in brackets is to the Tractatus.
72 LRKM, p. 59 (R.30).
73 LRKM, p. 62 (R.31). [February 1991. Two letters from Russell to Wittgenstein, which
were recently found in Vienna, are most instructive. In the first, dated 10 May 1915,
Russell writes, "I have got from Moore everything he had to report about tautologies
etc., but it was intelligible to me only in very small measure". In the second, dated 13
August 1919, Russell writes:
I have now read your book [manuscript of the Tractatus] twice carefully. - There
are still points I don't understand - some of them important ones - I send you some
queries on separate sheets. I am convinced you are right in your main contention,
that logical props are tautologies, which are not true in the sense that substantial
props are true. (Russell: The Journal of the Bertrand Russell Archives, New Series
10,2, Winter 1990-91, pp. 103 and 107.)]
74 Tractatus 4.1273.
75 Ibid., 6.122:
Whence it follows that we can get on without logical propOSItions, for we can
recognize in an adequate notation the formal properties of the propositions by mere
inspection.
76 Ramsey, The Foundations of Mathematics and Other Logical Essays, ed. R. B.
Braithwaite (Paterson: Littlefield, Adams & Co., 1960), 'General Propositions and Caus-
ality. p. 238.
Professor Nickolas Pappas has pointed out to us the following passage from Act II of
Shaw's Man and Superman:
TANNER: ... Let me remind you that Voltaire said that what was too silly to be
said could be sung.
STRAKER: It wasn't Voltaire: it was Bo Mar Shay.
TANNER: I stand corrected: Beaumarchais of course. Now you seem to think that
what is too delicate to be said can be whistled. Unfortunately your whistling, though
melodious, is unintelligible. (Man and Superman, Penguin Books, New York, 1946,
p. 106.)
77 Ramsey, The Foundations of Mathematics, p. 4.
78 Ibid., p. 11.
79 G. H. von Wright and particularly Jaakko Hintikka have explored and interestingly
developed this new sense of "tautologous". See von Wright, 'Form and Content in Logic'
and 'On the Idea of Logical Truth (I)', both reprinted in his Logical Studies (London:
Routledge & Kegan Paul, 1949) and his "On the Idea of Logical Truth (II)" (Societas
Scientiarum Fennica. Commentationes Physico-Mathematicae XV No. 10 (1950): 1-45).
See also Jaakko Hintikka, Distributive Normal Forms in the Calculus of Predicates (Acta
Philosophica Fennica, 6, Helsinki: 1953); Logic, Language-Games and Information
(Oxford: Clarendon Press, 1973), especially chapter VII; '0. H. von Wright on Logical
Truth and Distributive Nonnal Fonns' (in The Philosophy ofG. H. von Wright (La Salle:
Open Court).
80 Lewis and Langford, Symbolic Logic, p. 211.
81 Ibid., p. 211.
82 Ibid., p. 212.
83 Ibid., p. 24.
84 Lest any reader be misled:
The term "tautology" is taken from Wittgenstein. The present notion of tautologous
statements, as those true by virtue solely of truth-functional composition, seems to
agree with his usage; he contrives to make the term cover truths which involve also
quantification, but this is consequent only upon an effort to explain quantification
as a sort of infinite mode of truth-functional composition. A broader use of the term
"tautologous" has arisen in subsequent literature, because of Wittgenstein's doctrine
that all mathematics and logic is tautologous. This doctrine was intended by Wittg-
enstein as a thesis, not as a definition of tautology; and indeed it is a difficult thesis
to defend. But some who do not maintain the thesis in any such form, and who
regard the inferences of logic and mathematics as "merely verbal transformations"
or "disguised repetitions" only in some much broader sense, have been led thus to
transfer the term "tautologous" to this broader sense. It is not clear just what this
broader sense is (cf. my 'Truth by Convention'); but, whatever it is, there is already
a term of long standing ready at hand for it - Kant's term "analytic". Hence,
following a suggestion of Carnap's, I am confining the term "tautologous" to the
narrower sense - though in abstraction from Wittgenstein's theories. (W. V. Quine,
Mathematical Logic (Norton, 1940) Section 10, p. 55)
85 We are indebted to John Rawls for helpful comments on an earlier version of this
paper. We are also indebted to the members of the Centennial Wittgenstein Conference
held in April 1989 at the Florida State University, Tallahassee - especially to Jaakko
Hintikka, David Pears and David Stern.
REFERENCES
Aanderaa, S. O. and H. R. Lewis: 1974, 'Linear Sampling and the V3V-case of the
Decision Problem', Journal of Symbolic Logic 39, 519-47.
Boolos, George: 'Reading the Begriffsschrift', Mind 94, 331-34.
Bradley, F. H.: 1883, The Principles of Logic, Kegan Paul, Trench & Co., London.
Coffey, P.: 1912, The Science of Logic, William and Norgate, London.
Dreben, B. and W. Goldfarb: 1979, The Decision Problem: Solvable Classes of Quantifi-
cational Formulas, Addison-Wesley, Reading, MA.
Frege, G.: 1884, The Foundations of Arithmetic, (originally published as Die Grundlagen
der Arithmetic, by Wilhelm Koebner, Breslau; English trans. J. L. Austin, second
edition, Northwestern University Press, Evanston, IL, 1980).
Goldfarb, W.: 1984, 'The Un solvability of the G6del Class with Identity', The Journal
of Symbolic Logic 49, 1237-52.
Goldfarb, W.: 1988, 'Poincare against the Logicists', Minnesota Studies in the Philosophy
of Science, Vol. XI, History and Philosophy of Modern Mathematics, William Aspray
and Philip Kitcher (eds.), University of Minnesota Press, Minneapolis, pp. 61-81.
Hegel, G. W. F.: 1812-13, The Science of Logic, ed. H. D. Lewis, trans. A. V. Miller,
Humanities Press International, Atlantic Highlands, NJ, 1989.
Hintikka, J.: 1953, Distributive Normal Forms in the Calculus of Predicates, Acta Philo-
sophical Fennica, 6, Helsinki.
Hintikka, J.: 1969, 'On Kant's Notion of Intuition (Anschauungj', in T. Penelhum and
J. J. MacIntosh (eds.), Kant's First Critique, Wadsworth, Belmont, CA, 38-53.
Hintikka, J.: 1972, 'Kantian Intuitions', Inquiry 15, 341--45.
Hintikka, J.: 1973, Logic, Language-Games and Information, Clarendon Press, Oxford.
Hintikka, J.: 1990, 'G. H. von Wright on Logical Truth and Distributive Normal Forms',
in The Philosophy of G. H. von Wright, Open Court Press, La Salle, IL, pp. 517-37.
Hylton, P.: 1990, Russell, Idealism and the Emergence of Analytic Philosophy, Oxford
University Press.
Kant, I.: 1783, Prolegomena to Any Future Metaphysics, Lewis White Beck ed. and
trans., The Bobbs-Merrill Company, Indianapolis, IN, 1950.
Kant, 1.: 1787, The Critique of Pure Reason, trans. Norman Kemp Smith, St. Martin's
Press, New York, 1967.
Kant, 1.: 1800, Logic, trans Robers S. Hartman and Wolfgang Schwarz, Dover Publi-
cations, New York, 1988.
Leibniz, G. W.: 1679, 'On the General Characteristic', in Leibniz: Philosophical Papers
and Letters, ed. L. E. Loemker, second edition, University of Chicago Press, 1976,
pp.225-6.
Leibniz, G. W.: 1680-84, 'First Truths', in Leibniz: Philosophical Papers and Letters,
pp. 267-71.
Leibniz, G. W.: 1686, 'Necessary and Contingent Truths', in Leibniz, Philosophical
Writings, ed. G. H. R. Parkinson, J. M. Dent & Sons, London, 1973, pp. 96-105.
Leibniz, G. W.: 1714, Monadology, in Leibniz: Philosophical Papers and Letters, pp.
643-53.
Leibniz, G. W.: 1715-16, 'Second Letter to Samuel Clarke', in Leibniz: Philosophical
Papers and Letters, pp. 677-79.
Lewis, C. 1.: 1923, 'A Pragmatic Conception of the A Priori', The Journal of Philosophy
20, 169-77.
Lewis, C. 1. and C. H. Langford: 1932, Symbolic Logic, The Century Company, New
York.
Lewis, H. R.: 1979, Unsolvable Classes of Quantificational Formulas, Addison-Wesley,
Reading, MA.
Locke, J.: 1689, An Essay Concerning Human Understanding, ed. Peter Nidditch, Oxford
University Press, 1975.
Mates, B.: 1986, The Philosophy of Leibniz, Oxford University Press.
Mauthner, F.: 1902, Beitriige zu einer Kritik der Sprache, first edition, J. G. Cotta'sche,
Stuttgart.
McGuinness, B.: 1972, 'Bertrand Russell and Ludwig Wittgenstein's 'Notes on
Logic", Revue Internationale de Philosophie, 444-60.
McGuinness, B.: 1988, Wittgenstein, A Life, University of California Press, Berkeley.
Moore, G. E.: 1900, 'Necessity', Mind, 289-304.
Parsons, c.: 1983a, 'Frege's Theory of Number', in Parsons, Mathematics in Philosophy,
Selected Essays, Cornell University Press, Ithaca, NY, pp. 150-75.
Parsons, C.: 1983b, 'Kant's Philosophy of Arithmetic', in Parsons, Mathematics and
Philosophy, Selected Essays, pp. 110-49.
Pears, D.: 1977, 'The Relation between Wittgenstein's Picture Theory of Propositions
and Russell's Theories of Judgment', Philosophical Review, 177-9.
Pears, D.: 1989, 'Russell's 1913 Theory of Knowledge Manuscript', in E. W. Savage and
C. A. Anderson (eds.), Rereading Russell: Essays on Bertrand Russell's Metaphysics
and Epistemology, Minnesota Studies in the Philosophy of Science, XII, University of
Minnesota Press, Minneapolis, 169-82.
Poincare, H.: 1904, Science and Hypothesis, Walter Scott Publishing Company; repub-
lished by Dover Publications, New York, 1952.
Post, E. L.: 1941, 'Absolutely Unsolvable Problems and Relatively Undecidable Proposi-
tions: Account of an Anticipation', unpublished paper printed in M. Davis (ed.), The
Undecidable, Raven Press Books, Hewlett, NY, pp. 338-433.
Quine, W. V. 0.: 1940, Mathematical Logic, Norton & Co.; second revised edition 1951,
Harvard University Press, Cambridge, MA.
Quine, W. V. 0.: 1969, Ontological Relativity and Other Essays, Columbia University
Press, New York.
Ramsey, F. P.: 1960, The Foundations of Mathematics and Other Logical Essays, R. B.
Braithwaite (ed.), Littlefield, Adams & Co., Paterson, NJ.
Ricketts, T. G.: 1985, 'Frege, the Tractatus, and the Logocentric Predicament', Nous
XIX, No.1, 3-15.
Ricketts, T. G.: 1986a, 'Generality, Meaning and Sense in Frege', Pacific Philosophical
Quarterly 67, 172-95.
Ricketts, T. G.: 1986b, 'Objectivity and Objecthood: Frege's Metaphysics of Judgment',
in L. Haaparanta and J. Hintikka (eds.), Frege Synthesized, D. Reidel, pp. 65-95.
Russell, B.: 1897, An Essay on the Foundations of Geometry, Cambridge University
Press.
Russell, B.: 1900, A Critical Exposition of the Philosophy of Leibniz: With an Appendix
of Leading Passages, Cambridge University Press.
Russell, B.: 1905, 'Review of Henri Poincare's Science and Hypothesis', Mind; reprinted
in Russell's Philosophical Essays, George Allen & Unwin, London, 1966.
Russell, B.: 1912, Problems of Philosophy, William & Norgate, London.
Russell, B.: 1918, Philosophy of Logical Atomism (first published in the Monist 28,495-
527) Open Court Publishing Company, La Salle, IL.
Russell, B.: 1919, Introduction to Mathematical Philosophy, The MacMillan Co., New
York.
Russell, B.: 1938, The Principles of Mathematics, Cambridge University Press, first edition
1903, second edition (with a new introduction), 1938.
Russell: The Journal of the Bertrand Russell Archives, New Series, Vol. 10, No.2, Winter
1990-91.
Shaw, G. B.: Man and Superman, Penguin Books, New York, 1946.
Von Wright, G. H.: 1949, Logical Studies, Routledge & Kegan Paul, London.
Von Wright, G. H.: 1950, 'On the Idea of Logical Truth (II)', Societas Scientiarum
Fennica. Commentationes Physico-Mathematicae XV No. 10, pp. 1-45.
Wang, H.: 1963, 'Dominoes and The AEA Case of the Decision Problem', Proc. Sympos-
ium on the Mathematical Theory of Automata, Polytechnic Press, New York.
Wittgenstein, L.: 1912-13, 'On Logic and How Not to Do It', The Cambridge Review
34, 351. (Reprinted in Brian McGuinness, Wittgenstein, A Life, pp. 169-70.)
Wittgenstein, L.: 1921, Logische-Philosophische Abhandlung, final chapter, Ostwald,
Annalen der Naturphilosophie; first published in English, 1922, under title Tractatus
Logico-Philosophicus, C. K. Ogden (trans.) Routledge & Kegan Paul, London; re-
printed with corrections 1933.
Wittgenstein, L.: 1974, Ludwig Wittgenstein: Letters to Russell, Keynes and Moore, G.
H. von Wright (ed.), B. McGuinness (trans.), Cornell University Press, Ithaca, NY.
Wittgenstein, L.: 1975, Wittgenstein's 1939 Lectures on the Foundations of Mathematics,
C. Diamond (ed.), University of Chicago Press, Chicago.
Wittgenstein, L.: 1979, Notebooks 1914-1916, G. H. von Wright and G. E. M. Anscombe
(eds.), G. E. M. Anscombe (trans.), second edition, Basil Blackwell, Oxford.
Weiner, J.: 1990, Frege in Perspective, Cornell University Press, Ithaca, NY.
Dreben: Dept. of Philosophy
Harvard University
Cambridge, MA 02138
Floyd:
Dept. of Philosophy
Boston University
745 Commonwealth Ave.
Boston, MA 02215
U.S.A.
Dept. of Philosophy
The City College of New York
New York, NY 10031
U.S.A.
DAVID CHARLES MCCARTY
THE PHILOSOPHY OF LOGICAL WHOLISM*
Words are like a filin on deep water
LUDWIG WITTGENSTEIN
Notebooks 1914-1916
ABSTRACT. The present paper is one installment in a lengthy task, the replacement
of atomistic interpretations of Wittgenstein's Tractatus by a wholistic interpretation on
which the world-in-logical-space is not constructed out of objects but objects are ab-
stracted from out of that space. Here, general arguments against atomism are directed
toward a specific target, the four aspects of the atomistic reading of Tractatus given in
the Hintikkas' Investigating Wittgenstein (Hintikka & Hintikka 1986). The aspects in
question are called the semantical, metaphysical, epistemological and formal.
What follows a precis of the Hintikkas' rendering of Wittgenstein's perspective is a
characterization of the wholistic interpretation, comparing Wittgenstein's world and the
transcendental conditions it sets upon possible notation to a blank page and the conditions
it sets upon what is about to be written there. There will not be occasion to bring
arguments against each plank in the atomist's platform or in support of each facet of
wholism. But there is an extended treatment of the first two aspects - the seman tical and
metaphysical - which takes off from Wittgenstein's determination that, in his hands,
"logic must take care of itself".
The second half of the paper contains a negative assessment of the support the atomistic
reading can glean from the texts of Tractatus and Notebooks. From a detailed look into
a range of relevant textual and translational issues, we find little there to encourage that
interpretation and much to discourage it.
The paper closes on a preliminary consideration of one segment of the formal aspect
of the Hintikkas' atomism, the idea that the analysis of Tractatus is the analysis of Russell
or is, at worst, a near relative. Examination shows that Wittgenstein would have little
reason to model his analysis on that of Russell. The fundamentally wholistic vision
expressed in Tractatus requires a distinctively non-Russellian, decompositional version of
analysis.
1. INTRODUCTION
If there was a specifically "early" Wittgenstein, he was not a logical
atomist but a logical wholist. By this, I mean to break with a long
interpretative tradition and to turn a standard view of Wittgenstein's
Tractatus and contemporaneous writings on its very head. By my lights,
Synthese 87: 51-123, 1991.
52 DAVID CHARLES MCCARTY
Wittgenstein's Welt, the world-in-Iogical-space, was not the great logical
fabrication, built up by iterable semantic constructions out of indepen-
dent units of meaning, the logical atoms. To read Wittgenstein's treatise
is not to read a mild revision of Russell's semantics, or of any semantics,
provided that a semantics is codified as a set of principles of meaning,
a knowledge of which would constitute an understanding of language.
Wittgenstein's logic, his perspective on the logical, was not an extension
of a line first drawn by Frege and extended into our century by Russell,
but a determined attempt to rub that line out.
The tradition of logical atomism and of atomistic interpretations
of Wittgenstein was established in Russell's troubled Introduction to
Tractatus, where he wrote "the naming of simples is shown to be what
is logically first in logic" (Russell 1981, p. xiii). It was extended through
The Philosophy of Logical Atomism (Russell 1985), and commentaries
such as Black's A Companion to Wittgenstein's Tractatus (Black 1964).
More recently, it is upheld in Merrill and laakko Hintikkas' Investigat-
ing Wittgenstein (Hintikka & Hintikka 1986). Toward this book are my
remarks principally directed. However, it is clear that many apply far
more widely.
Banishing logical atomism and welcoming logical wholism is setting
up an opposition between two ways of reading Tractatus or, better, of
picturing to ourselves Wittgenstein's world. For the atomists - those
who favor the attribution to Wittgenstein of logical atomism - the world
of Tractatus is semantically structured. Its image is absolutely clear,
preternaturally sharp: a logical mosaic tiled with purely referential
units, each lying at the lower limit of conceptual and perceptual reso-
lution. For the wholists, Wittgenstein's world has no philosophical
image at all, no picture of a logically primeval landscape which later
gets reported in the grammar of our language. The world there is not
one which is represented to us. It is not there to be an object of our
intelligence but emerges to make intellection possible. Certainly, it is
not a world constructed from seman tical ephemera.
I caution the reader that this opposition is not one between philosoph-
ical views comparable with the outposts of warring tribes which over-
look a neutral terrain. We ought to refuse the temptation to borrow a
metaphor from the information-processing trade and call atomism a
"bottom up" version of Wittgenstein while wholism is "top down".
There is danger in this metaphor's false suggestion: that the two read-
ings might go variously up and down the same ladder and then meet
THE PHILOSOPHY OF LOGICAL WHOLISM 53
in the middle. There is no ladder. Hermeneutic evil awaits us should
we think that there is, in Tractatus, a single idea or set of ideas, some
fixed content on which, if it were only formulable in a fair fashion,
atomist and wholism might both agree. It is not as if Tractatus bears a
message which is interpretation ally untainted, as if Wittgenstein's words
formed a plain brown wrapper which, once removed, reveals a Tract-
arian world - either simple or conglomerate - within a single logical
space, a world which one party - of which the Hintikkas are subscrib-
ers - prefers to survey from the soil up while the other, somewhat
perversely, wishes to take in from the stratosphere and on down. Atom-
ists and who lists do not have a world in common.
The distance between the two readings, wholism and atomism, is not
like that between two painted dots on a single canvas; it is much more
like the distance between two so-called styles of painting. Indeed, it
bears marked similarity to the distance between impressionism and
expressionism. The extent of the distance is measurable from the extent
that we disagree over seemingly simple terms, words such as 'philoso-
pher' or 'logician.' When the Hintikkas wrote that Wittgenstein was a
"philosopher", they meant that he was a twentieth century analytic
philosopher. As such, he would have shared some intellectual concerns,
a standard of discourse and a certain method with philosophers of the
Hintikkas' milieu. As it happens, these sentiments of solidarity would
be much more appropriate to the eighteenth century philosopher Hume
than to the twentieth century Wittgenstein. Hume shares with his latter
day sympathizers a commitment to semantical atomism, a philosophical
aesthetic according to which the world comes readymade to be talked
about by us. It is constructed or built up - in just the way our language
is supposed to be - out of discrete units of ideation, each containing a
precise measure of thought. Hume is also, and self-consciously so, a
scientist-philosopher. He declared himself a full partner in the great
firm of empirical science. He saw philosophical inquiry as apiece with
the scientific (or what he took to be the scientific) redescription of
everyday life, a redescription ratified by a hidden reality, an intelligible
structure behind our words, structure invisible to ordinary mortals but
open to philosopher-scientists.
Wittgenstein was never a philosopher of this Humean, or analytical,
persuasion. He shunned the comforts both of semantical atomism and
of the scientifically-based worldview. In both the early Notebooks and
Tractatus, Wittgenstein wrote:
At bottom the whole Weltanschauung of the moderns involves the illusion that the so-
called laws of nature are explanations of natural phenomena. In this way, they stop short
at the laws of nature as at something impregnable as men of former times did at God
and fate. And both are right and wrong. The older ones are indeed clearer in the sense
that they acknowledge a clear terminus, while with the new system it is supposed to look
as if everything had a foundation. (N, 72)1
Wittgenstein is also called "logician". This is accurate insofar as,
throughout his life, he experimented in anti-canbnical notations, rede-
signing forms of concept writing so that crooked lines of thought come
out straight. He did not hanker, as do contemporary logicians, after a
complete theoretical redescription of the abstract form of correct infer-
ence; he thought such description unattainable. Wittgenstein's logic was
far more praxis than theoria; like medieval logic, it was a science of
personal mental hygiene or, better, a talisman against pernicious and
subtle fallacies. Today's logic is in its middle-age; in certain neighbor-
hoods, its business seems firmly established. It is as if our logic had
clearcut a vast region of linguistic jungle, exposing a granite landscape
of Boolean connectives, quantifiers, incompleteness phenomena and
model theory over which it exerts eternal claim. By contrast, Wittgen-
stein's logic could be a young Turk - an anarchist, in fact - never
seeking the establishment of a particular notation but always a disestab-
lishment, a permanent logical revolution. As Wittgenstein wrote of
language games in paragraphs 130 and 131 of Philosophical Investi-
gations, his logical way with notations was one of Ahnlichkeiten und
Uniihnlichkeiten, similarities and dissimilarities: notations "are much
more set up as comparative objects which, through similarities and
dissimilarities, are meant to throw light upon the status of our language"
(Wittgenstein 1953).
Nor is there in Tractatus the familiar vision of language as mathema-
tical object. In contemporary logic, language is a free semigroup of
simple forms, generated in a rule-governed way from a tiny base of
linguistic surds. The direction of generative construction is a reconstruc-
tion of understanding: we are thought to grasp sentences by cognitive
building, first putting semiotic hands on the members of the base and
building up, from them, a whole algebra of knowledge by iteration of
operations. However, the mathematical language is conceived as, in
itself, mute or uninterpreted; its elements cannot speak of their own
accord. Only in another language, a semantic language, does the al-
gebra of signs get a voice. This is the distance between syntax and
semantics and there is no greater distance in mathematical logic than
that. The gap is so wide it calls for the seeming strangeness of inten-
tionality to bridge it across. For Wittgenstein, all this is changed. If we
have a knowledge of meaning, then the logician cannot reconstruct it.
The elements of language, the simple forms - Wittgensteinian objects -
are not cognitive starting-points but, rather, endings. There is no gap
between syntax and semantics; there is no Wittgensteinian syntax with-
out reading and no reading without a world of interpret'ation. Wittgen-
stein's world is just the perspective on signs we naturally take when we
see them as signs. There is no gap to bridge between name and Bede-
utung; use a sign as a name, its Bedeutung forms automatically - not
as a matter for thought but as a condition of it.
In the sequel, we look to see what kind of philosopher and what
kind of logician Wittgenstein was. We look over and, then, overlook
the kind of philosopher and logician he was not, an atomistic one.
2. THE PARTS OF T O M I S ~
For many, Tractatus and Notebooks are the twentieth century Urquellen
for the semantico-metaphysical doctrine of logical atomism. In place of
the emptiness of Wittgenstein's Welt, they put a colored picture-book
world, an epistemically crafted and intelligible whole afloat in a limpid
logical space. At its base lie logical atoms; riding so low on the logical
hierarchy that we might say they are mind-bogglingly simple. The atoms
are monadic - themselves miniature worlds. They are classically sub-
stantial, existing independently of their fellows, conceivable and -
above all - nameable without concern for any other. Atoms are made
to combine and, when they do, they give rise to logical molecules,
Sachverhalte, atoms fit together by propositional modes of construction.
Sachverhalte then bond into even larger units, Tatsachen and these, in
their turn, into yet more comprehensive Tatsachen. In this way is the
world of atomism built up.
In this world, grammatology recapitulates ontology; the world's struc-
ture tells us how to speak. The order of exposition is the order of
cognition. The constructional history of the world is semantical, an
ontological chronicle of our personal histories in understanding sen-
tences. The world tells us how sentences have to be grasped: first, we
come upon atoms as elements in perception and, repeating Adam's
work with the animals, assign them names. In the field of our semantic
cognition, they are then joined into sentences by description operators,
truth functions, quantifiers and the glue that sticks subjects onto predi-
cates. These are the syntactic shadows of the bonds that create Tat-
sachen. The latter are, in turn, joined truth-functionally into yet more
complex sentences, ones with further Tatsachen as their referents. Par-
allelism is complete: sentences - properly understood - are constructed
out of simple names just as their propositional contents, Wittgenstein's
Tatsachen, are constructed out of sensory primitives, their referents.
3
Atomism is resolvable into a constellation of four attributions. In
college catalogue style, I call them the semantical, ontological, epistemo-
logical and formal. The semantical is the focal point for atomism's
vision: as just described, that world is in itself grammatically structured
and the grammar is Russell's logical grammar. As Russell explained, it
is "the view that you can get down in theory, if not in practice, to
ultimate simples, out of which the world is built, and that those simples
have a kind of reality not belonging to anything else" (Russell 1956, p.
58). In Tractatus, the simples are Tractarian objects. Available modes
of construction - of propositions and complexes from objects - makes
for an object its logical form, the weight of its logical contribution. For
the atomist, the bulk of logic is the sum of these individual weights.
So, the Hintikkas towed the same line as Russell when they wrote "the
logic of Tractatus is a logic of simple objects" (Hintikka & Hintikka
1986, p. 100).
In the atomist's semantical household, objects "wear the pants";
there, "the crucial subject matter in trying to understand the Tractatus
is obviously Wittgenstein's conception of object" (Hintikka & Hintikka
1986, p. 30). The objects take pride of place because the project of
atomism is a project of denotational justification, without which the
image of constructed world is empty slogan. The project idea is this:
language, in its true character, is fully rational. Although it is possible
to mistake that character, logic can discover it. Once that is done, the
rationality of language will be visible in the presumptive parallelism
between speech and ontology: one can then explain and justify the
inner character of language by pointing to the way the world is. What
makes Tractarian objects a fit subject for analytical philosophy rather
than for science fiction is their putative role in this justificatory project,
in a science of atomistic semantics. Objects are semantical quarks,
ultimate theoretical entities in the story of meaning, one that begins
when names get attached directly to objects. Because the homomor-
phism of syntax to semantics is defined "up from the bottom", the
account of sentential combination, Satzge!ilge, is reduced to a theory
of objects and their bonds. Ultimately, therefore, logical licenses are
granted and grammatical prohibitions issued by the natures of the
objects. For instance, if 'a' denotes Terry and ' P ~ ' denotes the function
~ i s a pirate", and this is known, then we are justified in asserting that
'Pa' is meaningful while 'aP' and 'PP' are not. A justification for this
claim is meant to flow from knowing the natures of the ;denotata of 'a'
and of 'P' - in Fregean terms, from knowing that Terry is saturated
while ~ i s a pirate" is not. Atomism gives us facts of meaning which
facts of metaphysics explain.
At the ground floor of atomism's explanatory structure lie theoretical
relations among objects. They establish the second or ontological aspect
of atomism in Tractatus. It is divided into three subaspects. First,
Tractarian objects are said to fall into diverse metaphysical categories,
some of which are distinguished, using Frege's terminology, by satura-
tedness or by its opposite. Second, as the Hintikkas would have it, the
metaphysical categories are split off from one another along traditional
lines, those drawn by the ancient semanticists who distinguished parti-
cular from universal. Talk of names in Tractatus is interpreted so that
they denote both individuals such as Terry and properties such as
"being a pirate". (I am not proposing that humans like Terry and
properties like "being a pirate" are actual samples of the sorts of things
which atomists have in mind when they speak of objects. They are
intended as expository models for the kinds of differences thought to
obtain.) Finally, objects manage the division into metaphysical types
on their own. As a child might say, "they just come like that". Without
interference from language or mind, objects sort themselves into the
medieval categories just as protons and electrons are sorted by their
relative mass.
For epistemological reasons, Terry and "being a pirate" are unlikely
to satisfy the atomist as echt Tractarian objects. They are not epistemic
primitives, but are instead - or so we are told - composites of primitives.
Even though objects in Tractatus fall into a range of metaphysical
categories, epistemologically they come as one. They are all Russellian
objects of acquaintance. Objects which are particulars are sensory parti-
culars; their esse is really percipi. Objects which are universals include
directly intuited predicables - perhaps not blueness or triangularity, but
something along those lines. Russell was willing to count logical con-
stants among the objects and to allow us an. acquaintance with such
things as truth-functional negation. Via the avenue of acquaintance,
the atomist can arrive at a story about how we first come to a knowledge
of the basic semantic items, a knowledge from which all other semantic
knowledge is constructed.
Finally, there is the formal, the parasitism between logical atomism
and symbolic logic. Logic puts down the theoretical routes along which
atomistic explanations flow from ontological explanans to linguistic ex-
plananda. Also, logic is an instrument of discovery, the eye in which
the inner shape of language is reflected. The attribution to Wittgenstein
of logical atomism is ringed with formal techniques and logical theories.
Most prominent in the Hintikkas' exposition is Tarski's theory of truth,
formal semantics for a language with regimented syntax, one specified
by rules linking denotation to truth such as
(cpAI/!) is true if and only if both cp and I/! are true.
The Hintikkas asserted not only that Wittgenstein would have approved
the details of denotational semantics but also that the picture "theory"
of Tractatus was nothing less than a lyrical rendering of what would
come out more prosaically from Tarski's pen.
4
Prevalent also is the
notion that Wittgenstein took over Russell's analysis of belief contexts
as sketched in The Problems of Philosophy. Analysis is viewed as
primary means to logical discovery, playing for semantics the role the
telescope plays for astronomy. Through it, we see a need for objects,
or so it is said. Lastly, Wittgenstein is portrayed not as in revolt against
the logical precedents set by Bolzano, Frege and Russell, not as reject-
ing the unholy union of logic, metaphysics and a foundational mathe-
matics represented by Russell's theory of propositional functions and
Frege's hierarchy of concepts, but as the good steward in Frege's vine-
yard. The Hintikkas wrote:
Wittgenstein is in many respects an integral part of th[e] Fregean tradition, to the extent
that his overall philosophy in the Tractatus could almost be described as the metaphysics
that naturally accompanies the Fregean conception of logical language. (Hintikka &
Hintikka 1986, p. 88)
The section to follow marks the beginning of a critique of the four
aspects of atomism and the start of a lengthy demonstration that Witt-
genstein held none of the views ascribed to him there. When that is said,
there will be occasion to review some few of the remarks of Tractatus
which seem to offer solace to atomism and to confront the semantical
aspect of atomism with Wittgenstein's Tractarian anti-semantics. Com-
plete investigation of Wittgenstein's ideas on the internal relations be-
tween language and its readings and his logical insights, ones which sur-
round his concept of das eigentliche Zeichen, the real sign, will have to
wait for another time. The next section contains a summary statement of
the view on language actually contained in Wittgenstein's first writings; I
labor under no illusion that it is adequately documented here. I do not
doubt, however, that this can be done.
3. THE WHOLE OF LOGICAL WHOLISM
3.1. Four Broken Parts
In Tractatus, there is no semantics of the form which Frege and Russell
envisioned; the sense of a sentence is not something thought out by us
from principles of sense which a philosopher can discover. Meaning
has no philosophical doctrine for a logician to codify. Propositions,
although vehicles of knowledge, are not, semantically, objects of it. As
Wittgenstein would have it, we "understand the proposition without
having its sense explained to us" (T 4.021). The wellsprings of language,
the empirical details of its workings, the "deep water" of Wittgenstein's
Notebooks, are concealed from our explanatory abilities. What philoso-
phers mistakenly erect as a metaphysical theory which explains our
grasp of language is merely a projection of conventional aspects of
grammar.
In consequence, it is unimportant that language reflect a world struc-
ture rationalizing our knowledge of meaning; even less is it important
that this nonexistent knowledge be built up from primitives. In Trac-
tatus, the proprietary relation between Tatsache and Gegenstand is not
one of construction but, rather, of abstraction. What makes objects,
die Gegenstiinde, stand gegen is an irreversibly destructive process, one
bearing a similarity to destructive analysis in Frege (cf. Dummett 1981).
Hence, Tractarian objects are not starting points, but endpoints in a
process of analysis and (as we shall see) of a transcendental deduction,
by Wittgenstein, from the possibility of sense to a need for objects.
The precise nature of the objects becomes, not merely a question of
no particular import, given that objects no longer afford the ultimate
explanans, but, in truth, a foolish question. Their "natures" (were this
term still apposite) are exhausted in their logical locations as analytic
termini, which are positions in an essentially symbolic arrangement.
Tractarian objects are neither substantial nor monadic in a classical
way; they are not metaphysical characters proper to objects in them-
selves. Nor are there categories into which objects come already sorted.
The primitivity of objects is written into the character of their symbols
and is revealed in what Wittgenstein called their 'real signs'. The ques-
tions of logical atomism, of the justificatory project, need no answers.
Questions remaining for logical wholism do not concern the natures of
objects but the irreducible structure of the whole.
As for epistemology in Tractatus, something is indeed hidden. It is
the direction, path and target of the referential arrow. As Wittgenstein
lectured Russell in his letter of August 1919,
the kind of relation of the constituents of the thought and of the pictured fact is irrelevant.
It would be a matter of psychology to find out. (N, p. 129)
Hence, if the referential arrow lands on sensory primitives, it is of no
concern to Wittgenstein the logician. There is no pressure from the
reductive force of atomism to explain our ken of atoms in terms of
acquaintance or description, since there is, in Tractatus, no need to
explain understanding cognitively at all. In the sequel, we find evidence
that Wittgenstein rejected the epistemic plank of atomism outright.
The place of logic in Tractatus is some way off from that marked on
the Hintikkas' map. For example, contemporary logicians enforce and
exploit the gap between syntax and semantics; it is the joint which
bends to allow interpretational variation. And variation is nowadays
the principal moving part in accounts of validity: a logical scheme is
said to be valid when it is true under every interpretational variation.
Wittgenstein's logic would admit no such gap - not even a gap with a
bridge - between syntax and semantics or, more appropriately, between
real sign and Bedeutung. The first completely determines the second.
Examine the following passage from Notebooks:
If a name signifies an object, then it stands thereby in a relation with it, which is wholly
determined by the logical kind of the object and which signalizes that logical kind. (N,
p. 70)
Recall that the logical kind of an object is fully incorporated in the
character of its real sign. Meanings are not the products of casual or
incidental ways of addressing language or of standing in attendance to
it. Logical validity is not the fact remaining constant across all variation
in possible fact; Wittgenstein's validity is an expression of a possible
view, of the prospect that language be viewed as transcendental.
To put it another way, the relevant relation between language and
the world is internal (cf. T 4.014). It is contained in the symbol which
is created with the sign and is revealed by what is essential to the
symbol. That essentiality is, in turn, laid bare by logical investigation,
the method of search for the Wittgensteinian real sign.. Passage 3.3411
is a short tutorial on this method:
One could, therefore, say: the real name is that which all symbols, which signify the
object, have in common. Thus, successively, it could be shown that all kinds of composi-
tion are inessential for the name.
Like the faces in our dreams, our signs would rather hide their true
significance. The logician is not, however, utterly powerless; we are
able to elaborate the character of the symbols which our signs create.
A Tractarian symbol is not an object and, especially, not an abstract
object. It is a distillation of use, the essential core of the use of the
sign or the sign taken together with its logical role. Here is what
Wittgenstein said of it at 3.326: "In order to recognize a symbol by its
sign we must pay attention to its meaningful use".
What counts for logic in the sign, what succeeds in signifying, is the
essential in the symbol. This is available to the logician as a notational
"least common denominator". It is the commonality of all symbols
which serve the same purpose that encapsulates the precise logical role
of the sign. "What signifies in a symbol is what is common to all the
symbols that, according to the rules of logical syntax, may be substituted
for it" (T 3.344). What is inessential, what is not part of the real sign,
appears as that which can vary under notational change. The hunt for
the real sign, then, is just a method of variation. Wittgenstein, there-
fore, conducted a series of syntactic experiments. Each move from
notation to notation put into relief features of the signs which vary
from one to the other and, hence, could be ignored. In the words of
the schoolboys, notation is changed "just to see what we can get away
with". The real sign is that which remains, what we cannot get away
without.
If there is alternative notation which succeeds in expressing a parti-
cular notion but which does not contain notational feature X, then the
adequacy of the alternative shows X to reflect no part of the real sign
for that notion. To take a trivial example, it is no fixed feature of
logical notation for formal derivations that it be linear - that it be
run out along the horizontal. Both Gentzen and Frege gave us two
dimensional signs for derivations. Less trivially, that logical notation
be separable into parts which can be repeated is essential to it; no
successful notation will be constructed from a single undifferentiated
black inkblot. Hence, what Wittgenstein called "articulation" is essen-
tial to sentential notation and was part of the real sign for Tatsachen.
Wittgenstein's logic was anything but acquiescence in what had gone
on in logic before, what he stigmatized as "the old notation". The
method of variation relies entirely upon putting old notation into ques-
tion and finding alternatives for it. This points up but one inaccuracy
in the Hintikkas' picture of the formal side of Tractatus. Wittgenstein
rejected the proposed identification between what is today called
"Tarskian semantics" and the metaphor of propositions as pictures. In
Wittgenstein's works, semantical pronouncements such as "'cats' de-
notes cats" fail to make connections between word and world. Nothing
can point from language to something outside of it; there is nothing in
Tractatus which is outside language. Hence, such pronouncements can
succeed only in reinforcing connections within the system of signs.
5
Moreover, there is evidence that the analysis of which Wittgenstein
wrote could not have been Russellian. The material cited in our final
section will present some of this evidence.
3.2. The World in Tractatus
From the very first, Tractatus is a rejection of the traditional in philos-
ophy. Since Parmenides, the study of ontology has been a study in
world lists or world inventories. The world was something we compre-
hended by building up to it, following from below an order which we
took to lie entirely in it. The first step is to run through all the sorts of
things in the world. The world becomes a stockroom whose contents
we debate. You say, "Perhaps there are only elementary particles" and
I might reply "No, everything is pure spirit or neutral monad". This is
the comfortingly familiar approach to philosophy.
In ontology, it would be wholly alien to start the other way alto-
gether - from the world as undifferentiated whole. This would defeat
our ontological expectation for world overview by world list. This is
why Wittgenstein's first sentence
Die Welt ist alles was der Fall ist.
ought to shock. Neither Russell nor Frege would have taken the whole
world as philosophical starting point. Frege dreamt a world intelligible
through its regions. As in Caesar's Gaul, there were in Frege drei
Reiche: the realm of reference, the realm of ideas and the realm of
sense. As in Der Gedanke, nothing remains for a unitary world; it is
no more than the sovereign union of the Reiche. Unlike Wittgenstein,
there is no seman tical juristic necessary to forge the Reiche into a single
domain; structure in Frege is not first in language, but first in the
Reiche. Language then waits on that structure.
Kraus wrote "Der Kunstler ist ein Diener am Wort" (Kraus 1987, p.
14): the artist waits on the word. In Tractatus, the world waits upon
language. World comes not by assembly or abstraction from an inven-
tory but by taking a stance relative to (all of) language.
6
Language in
Wittgenstein is the vehicle of the world. Conceived as divorced from
language, Wittgenstein's world is, in itself, empty, as is a blank page
waiting for someone to write. What structure it has - either transcen-
dental or empirical - it gets in language. The structure is emergent,
springing into existence as we live with new signs. The structure to
which logic pertains is transcendental. It is this structure that prohibits
the possibility that objects might be, as the atomists aver, of fundamen-
tal importance. This world is not come at from below via its individual
things because, first, even things (here, objects) are only come at from
above; they are only conceivable, even specifiable, from the wholistic
perspective. Second, logic is not a matter of things at all. Particular
things, even classes of them, are no concern to the logician. In Witt-
genstein, logic is first philosophy and "not only from a concern with
specific things must logic prescind; but it is as little permitted to occupy
itself with predicates and relations" (Wittgenstein 1984, p. 195). On
the same page appears: "logic cannot deal in any special groups of
things".
How does the world's transcendental structure manifest itself? Witt-
genstein did not write 'Die Welt ist alles' so that we could continue to
think of it as a collection of all things. He wrote that the world is "alles
was der Fall ist". Even a casual acquaintance with 'Fall' (translated
'case' in English, although not everywhere substitutable for it) reveals
its link to certain sorts of sentential complements. Germans say "Es
kann der Fall sein, dass ... " wherein 'der Fall' refers us to the intrinsic-
ally sentential content of the words to follow 'dass'. Wittgenstein is
giving notice: structure comes on dressed in the form of a sentence.
Second - and importantly - what is called 'der Fall' is not just any old
thing that can go into a sentence. Isolated events or even the events
one undergoes during the course of a day will not do as an answer to
the question 'Was war der Fall?' They are related as answers to a
question such as 'Was ist passiert?' - "What happened?" Appropriate
answers to the former are not, then, mere co'llections of reports -
one hour's output from a news agency teletype. Nor are they usually
narratives, forms of rendition like the plots of suspense novels, ex-
tending in time in the same way as the matters they convey.
"Was ist der Fall?" is, rather, a confluence of influences, a pattern
of considerations which the speaker assembles in order to explain or
to make sense of a complex issue, say, the course of a treaty negotiation
or the passage of legislation. The difference marked by 'der Fall' is one
of exposition; it lies not in the sort of event told but in the style of the
telling. It is an organization we set. Sometimes, we set it by convention;
more often, it comes by what we naturally find intelligible. World as
alles was der Fall ist is, then, world as structured not in itself but in the
arrangement of the story. In this case, it is the whole story. It is,
therefore, our contention that one can see the error in atomism from
the very first sentence of Tractatus.
Logical investigation reveals the transcendental features of the whole
story. Here, that investigation is not one of surveying individual verdicts
conveyed in or by formulae of particular logical systems. It is not one
of specific formal rules. It is a gauging of the potential in all signs and
their symbols, the potential for the battery of signs to become a logical
medium of thought. It is a search after the origin of life, of what gives
life to the sign. (In a moment, we shall pursue the search, at least
metaphorically. The idea of the blank page suggests a kind of creation
myth for the transcendental in language.) Before the origin of life,
nothing is living. The potential for language sets down absolutely no
requirements on the world. There is nothing statable to which we are
responsible and to which the content of our sentences must conform.
There is nothing which we consciously need to do in order to fulfill
such responsibility, since the abstract structure of the world is formed
within language and not the other way round. Wittgenstein wrote that
"[t]here is no a priori order of things" (T 5.634).
The contrast with Frege is marked. The Fregean logician must obey
the dictates of the laws of thought, records of what exists objectively
and for all time in the behavior of special functions, truth functions,
and in special objects, the truth values. Frege sought to explain and
justify the pretensions of systems of logic by checking that they fulfill
requirements for our script set down in the form of a metaphysics. In
Tractatus, there is no checking of signs against the features of a seman-
tically structured world. Wittgenstein, like Lewis Carroll, recognized
the circularity inherent in any attempt at such checking;. The "metaphys-
ics check" is itself an instance of thought and, as such, calls for logic.
In the course of the check, one is supposed to deduce consequences
from rules of formation or from semantical principles such as composi-
tionality. One needs to read these rules and principles and, so, must
parse them. In the semantics of Frege and Russell, I must think my
way from a parsing to a comprehension of them and thinking must be
done in accord with "laws of thought" .
Wittgenstein shunned rules and principles as aids to shoring up logic
as he did the entire justificatory project which gives to "shoring up" a
sense. Also, the world is my perspective on signs; as alles was der Fall
ist, it is the whole structure of explanatory narrative. Just as I cannot
escape my perspective, I cannot, in trying to explain language, escape
the whole structure of explanatory narrative. We cannot, by Wittgenste-
in's lights, get outside of language by this route. Next, as Wittgenstein
wrote, every word is a new symbol. In part, this delphic utterance
means that the content of a sign is never fixed exclusively by what the
Hintikkas call "a vertical linkage", that is, by the bare fact of its
standing in a denotational relation to some bit of the world. The only
"fixing" here runs in the opposite direction; the reader will recall that
it is the logical role of the sign which fixes the tenor of any relation
which strikes us as word-to-world. Every use of a sign is, thus, new
because the determining factor, the logical role, consists in the play of
the sign against the background afforded by the system of all signs. This
is a background in flux; there is "nothing fixed there" by a denotational
relation which coerces the logical role. Meaning, then, is a product of
the system of signs and there can be no meaning outside that system.
No claims stand apart from the system and get their meanings, once
and for all, by passing judgments on it from without.
The world without language is a philosophically featureless back-
ground against which sense will be made. But, as Wittgenstein would
write later (in Wittgenstein 1953, remark 304), there is "not nothing
there"; there is just nothing there for philosophers to read, nothing for
them to study with an eye to the rationalization of language. In this
regard, it is helpful to compare the world without language to a blank
page. It is certainly not the case that the paper has no properties. I
could list some which allow me to write words on it, that make it a fit
foundation for content. It is semipermanent - it will not evaporate at
the touch - and somewhat but not completely porous - so that the
letters do not run together. The ink soaks in but not through. My
letters, when I make them, must be separate on the page; the paper
will not allow me to stack them, one atop another. All these are
foundational for writing, but they are not part of a cognized foundation
for what gets written. First, the learning of writing is not the learning
of these properties of paper. Nor do I check the properties before I set
out to write. Next, reference to them is requisite neither for the justifi-
cation of word formation, grammar and usage nor for the justification
of what is written. I cannot write the properties of paper down on the
page and then deduce the technology of writing from them. Third,
these are not properties I take on mentally when I write; nothing in
the morphology of a word or the syntax of a sentence must be settled
in my mind to correspond with features of the empty page. All this is
to repeat, over and over, that paper is just the start or field of potential
for writing; it is not writing itself. And, as the reader will recall from
grammar school, the only standard against which to check one's pen-
manship is writing itself. Staring at the paper will not help.
In ordinary light, the world, as the paper, seems bright, but, before
the philosophial ultraviolet of justification, it is the dark water beneath
the film of words. The nature of world, as with paper, is not that of a
metaphysical medium clear to the philosopher's eye. The writing paper
is not something which I read. Rather, I rely silently upon its properties
in all that I write. Paper's properties will only shore up my writing in
the physical sense, not in the semantical. Also, the subsistence of
those properties is revealed in the success of the writing. As with
Wittgenstein, signs are themselves silent on the subject of the transcen-
dental features of the world. The latter makes thought possible and is
revealed in the most economical means for expression.
As it happens, this address to paper as a logical medium is just the
sort of address to which Wittgenstein returned in his Last Writings
(Wittgenstein 1982) and in Philosophical Investigations, Part II (Witt-
genstein 1953). On page 120 of the former appears the following:
Do I want to say something like 'the certainty of mathematics rests upon the reliability
of ink and paper?' No. (That would be a vicious circle.) ... It is completely clear that
one could not calculate with certain sorts of ink and paper, if they were, for instance,
under the sway of certain queer changes, however, that they do alter themselves can
only be determined through memory and the comparison with other means of calculation.
And these cannot, indeed, be tested by comparison against something else.
The world is a logical medium as paper is a medium for writing. As
such, media set limits. They are not limits which might, in the course
of the medium mediating, actually be broken, however. We need no
guarantee that what we put on the page with a pen will not abuse the
paper itself, perhaps cause it to vanish or turn into a pool of water.
The paper is a frame around my written message but I do not shatter
the status of the frame by writing clear off the paper and onto the
tabletop. Should I do so, paper plus tabletop provide the frame; we
might say that, like time, the notional page expands to fill the needs
of the project. Should I choose to "write vertically" and, instead of
spacing my letters across the sheet, try to trace one letter's outline right
atop another, I shall not have violated the limits so much as ceased to
write at all.
In several respects, paper resembles other media, the physical space
exploited by the sculptor, for instance. The sculptor need not check
that the laws of geometry are still valid before taking a first swing at
the chisel. Nor need the sculptor be concerned that the final product
will violate those laws. Speculations of string theorists aside, a sculpture
cannot require nineteen independent dimensions for its execution. The
logical world of Wittgenstein is also a medium and, as such, its being
in place (and our being placed relative to it) are all that is important.
It does not call for questioning. It sets inviolable limits, even though
they go without saying. However, they do not, as Wittgenstein would
have it, go without showing.
These analogies also have limits, ones that do call for questioning.
The world features Wittgenstein called 'transcendental' are, unlike the
permeability of paper, not material properties or any sort of Fregean
concepts. Nor are they to be housed within a special realm of being, a
Fregean Reich, a nebulous land of items waiting for us to speak and to
write them. Also, we do not conceive the permeability of paper as
inherent in writing or in the prospect of it but as existing in the paper.
By contrast, it is only their standing relative to the prospect of thought
that makes the world's features transcendent; they do not exist apart
from their manifestations in signs. In the latter respect are Wittgen-
stein's transcendental features immanent; they must appear in the con-
crete lives of signs just as committees inhere in their members, sym-
phonies in their scores and sizes in shoes. Wittgenstein would, in
general, allow no abstracta which are not immanent. Even the Wittg-
ensteinian numbers (and by that we mean numbers and not numerals)
are indices of linguistic operations and those, in turn, are inherent in
the concrete prospect that signs be repeated. Numbers - and all that is
transcendental - exist only in the prospects of the system of signs.
Systematicity is itself transcendental. There is no writing which is
nothing but a single sign; to extend the metaphor, a single dot on a
single page made by one person on a single occasion is not writing. We
make do with nothing less than a realm of signs, the geography of
which is not exhausted in any list of individual signs. It is only exhausted
by drawing a map of the signs which displays their relative logical
locations. In general relativity, matter becomes a modulation in space-
time. In the hands of Wittgenstein, signs are vortices or modulations
in logical space, inseparable from their relations with one another.
Logical space, the background of the map on which those relations
are displayed, is a field of comparison. As will be detailed anon, the
determinacy of these relations require that signs create symbols of quite
determinate articulation. These articulations resolve themselves into a
need for objects. This chapter in the creation myth, "How there came
to be objects", waits on my later explanation of how logic can "take
care of itself". There I tell how the features of world without language
become a world whose only features are language. The storyline now
bends to follow the fault lines dividing atomism from wholism.
Need I now say that the wholist sees Wittgenstein as a linguistic
idealist, as someone wanting what we call the world to be in language?
Wittgenstein's logical world is neither the empirical world, nor the
metaphysical world, nor the everyday world. In abstraction from every-
thing linguistic, it is without form and void. As signs are formed,
as the hand begins to write, the structure it gains is written as an
epiphenomenon of signs. This structure Wittgenstein also calls Gerust,
scaffolding. Everything that comprises physical world and everyday
world must fit within it. (In the case of paper, one might consider the
spaces between words and that around the perimeter of the message
as essential scaffolding or support for the writing.) It only emerges as
a feature of the signs; scaffolding only arises together with the building
and in the process of its creation. It is something we construct. Witt-
genstein wrote at passage T 4.023 that a
proposition constructs a world with the help of a logical scaffolding and this is why one
can see, on the proposition, how everything logical stands, if the proposition is true.
(For future reference, note that, although Wittgenstein speaks here of
a kind of construction, it is not construction out of given materials.
Konstruieren, the word used for "construction:', applies, in everyday
German, to abstract geometrical construction and not to physical con-
struction or literal "building up". Words for that would be 'bauen' or
'aufbauen'.) To steal a metaphor from Dummett (1976), the world's
structure comes into existence as we begin to probe linguistically. Struc-
ture only becomes visible by reflection, as it were, against the symbolic
aspect of the world. This is part of what Wittgenstein sought to convey
by describing objects, facts and propositions as formal concepts at and
following T 4.126.
Be warned, though. To imbibe this sort of tipple is not to drink the
cup of idealism to the dregs. Wittgenstein is not committed to main-
taining that, since Geriist is coeval with language, there were, for
instance, no real dinosaurs before there were humans to speak. This
confused commitment results from collapsing two distinct attitudes or
ways of addressing language, the empirical and the transcendental. As
empirical objects, objects of the languages of science and the languages
of sensation, there were real dinosaurs and there are actual languages.
The latter are attenuated entities, but no more so than Roman law or
Greek democracy; scholars trace their roots and chart their spread.
Dinosaurs and languages can be dated and traced. They are located
upon a temporal continuum; time is, in Wittgenstein, a feature of the
real signs of empirical language. Qua empirical, dinosaurs certainly
antedate language. Qua transcendental, as objects (or complexes) in
the Tractarian sense, dinosaurs and languages are not so much timeless
as untimed. It is only relative to language as the medium of thought
that we make sense of dinosaurs or of languages such as Urdu as objects
at all, as subjects of predication, as substantial and not hallucinatory,
as timed or timeless, as admitting of identity or difference. It is only
with an eye to logical relations between dinosaurs and the system of
signs, an eye to the transcendental peephole, that we conceive a dino-
saur in this way. In this sense - that dinosaur as object is not conceivable
apart from the medium - are they all coeval: individual languages,
dinosaurs and the medium of language. As Tractarian objects, lan-
guages and dinosaurs share a birthday.7
This is the principal error of atomism: that the world is ordered in
itself and the sign must obey its order; this is its diagnosis: that the
atomists ignore the transcendental, the respect in which there is a world
of logic, a world which is servant of the sign. There are three more
errors and three diagnoses in the short Sections 3.3, 3.4 and 3.5 to
follow. The first is the error of asking after the metaphysical natures
of objects, the second of adopting for Wittgenstein the goal of analytical
philosophy - of retelling the genesis of comprehension - and the third
of seeing a retelling in a theory of truth or in a logical grammar.
3.3. The Nature of Objects: Working Within the System
In language, as in the world, we start from the whole. What impresses
immediately about a calculus such as Frege's is its interconnectedness;
it is only by working within the whole system that we express content
and trace inferential paths. Signs do not come singly but arrive tightly
interlocked; without the connections, they lose their powers. There is
a thicket of signs but one can still pick out from it logical traces. There
is repetition in the system and, according to Wittgenstein, we select
out logical relations by focusing on constancies and variations in those
repetitions. Two sorts of constancies are Tatsachen and Gegenstande;
they stand out as lulls among changes. For instance, the object we think
of as a is better symbolized as
~ l a ~ 2
a pattern we can discern around a and repeated in Fa and in Rab. An
object is the condensation of this kind of pattern. One might say that,
from propositions, we abstract their simpler elements. This is hardly a
process of building up. After all, we started from the very top level,
that of the system. It is better described as a process of breaking down.
Confirmation for this suggestion is at T 1.2: "The world decomposes
into facts". (Later on, this passage becomes a matter of great concern.)
The idea of destructive analysis in Wittgenstein had its precursor in
the purely decompositional analysis Frege created for Begriffsschrift.
The parts, pieces or patterns which come to light by this analysis Frege
thought to carry no claim to independent existence; they cannot be
prized away from the decomposition process. In other words, the parts
into which a proposition decomposes are functional parts; they derive
their modus vivendi from service to the analytical whole. By means of
decompositional analysis, Frege separated function and argument out
from what he then called 'judgment'. This was his preferred route to a
correct assessment of a judgment's contribution to the validity (or
invalidity) of an argument. When measured in terms of progress in
linguistic knowledge, the analysis was unidirectional: it is always "top
down". In the unpublished Boole's Logical Calculus, Ftege wrote apro-
pos destructive analysis:
As opposed to this, I start out from judgements and their contents, and not from concepts.
I only allow the formation of concepts to proceed from judgements .... The content of
possible judgement is thus split into a constant and a variable part .... And so instead
of putting a judgement together out of an individual as subject and an already previously
formed concept as predicate, we do the opposite and arrive at a concept splitting up the
content of possible judgement. (Frege 1979, pp. 16-17).
The products are wholly dependent upon entire judgments for their
existence and intelligibility. In Begriffsschift, Frege viewed functions
and arguments as artifacts of logical analysis, as generalized parts of
linguistic expressions which correspond to no extralinguistic reality and
which exist solely to serve analytical purposes. On this subject, Frege
wrote:
This distinction (between function and argument) has nothing to do with conceptual
content; it concerns only our way of looking at it. (Frege 1977, p. 12).
In Tractatus, Bedeutungen of signs - facts and objects - are individu-
ated functionally. To a certain extent, it is not wrong to think of
Wittgenstein's logical world on the model of the social studies' descrip-
tion of a corporation. The corporation is first introduced as an indissolu-
ble whole, the functional means for the creation of wealth and the
provision of goods and services. No understanding of the corporation
comes via a separate grasp of each of its parts, as if the parts were to
be assembled like a jigsaw puzzle. As with Wittgenstein's world, the
corporation is "articulated" - it is naturally organized into parts but it
is not constructed out of its elements. Presidents, vice-presidents, board
chairmen and marketing managers are distinguished, but not by their
heights or races. They are presented in term of their respective offices.
To describe the "typical corporation president" is just to describe that
office. The offices are distinguished by reference to their varied services
to the corporation. In the same way, facts are logical offices and are
distinguished from each other and from objects or complexes by their
logical services and not via their constituents. A Tractarian object is a
Tractarian object in virtue of its functional role as the Bedeutung of a
name. One could say that an object is nothing more than a role.
The metaphor of corporation bears certain dialectical benefits. To
the extent that the realm of facts and objects really is like a corporation,
I am absolved of facing metaphysical questions about the true nature
of objects, questions that bear down so hard onlthe atomist. Corporate
executives do not comprise a natural kind nor do they constitute a
separate blood group. We do not ask after their intrinsic natures or
their individual subatomic makeups. We do not ask "What colors are
they?" or "How do they interact with the other political parties,
namely, Republicans and Democrats?" To ask such questions about
corporate executives is to confound functional with natural or other
nonfunctional comparisons. In the same way, to ask certain metaphys-
ical questions about objects is to confound transcendental features of
language with material or medieval comparisons. Objects do not com-
prise a natural or perceptual kind nor do they constitute a traditional
metaphysical group. This is because objects too are functional, more
like doorstops and executives than tigers. There is no more call to
answer the questions "What colors are Tractarian objects?" or "Are
they particulars like Terry or universals like 'being a pirate'?" or "What
are they in themselves?" than one has to answer the same questions
about executives. The only sure reply to the question "What is an
executive really?" is in terms of the corporate organization. In this way
is "executive" a "formal concept" in the world of economics, just
as "o'clock" is a "formal concept" in the world of telling time. The
only guaranteed reply to the question "What is a Tractarian object
really?" is in terms of transcendental organization. In this way is "ob-
ject" a formal concept in the world of Tractatus.
3.4. Understanding as Standing
In Frege and Russell, understanding is epistemic contact and cognitive
control, just the talents and materials required to build up an accurate
representation in mind: one must grab hold of basic semantic materials
and marshal them. Wittgenstein will have none of this. In his sight,
understanding is much more like standing. In standing, one takes a
position or adopts a perspective within an objective setting. The opera-
tive setting is what I choose to call "the mechanism of language". It is
a mechanism over which we exert little control, especially of an intellec-
tual sort. As mechanism, language is an automaton; like the vast subter-
ranean machines of Lang's Metropolis, a major part of its working is
in operating us. The language mechanism is infinite - boundless and
all-embracing. It is beyond intelligibility. We are linked to
it and in it but we have no real notion of the details 'of its workings.
Certainly, we do not work it by grasping it or by lodging its governing
principles in our minds.
Mankind possesses the ability to build languages with which to give expression to any
sense, without having any notion how and what each word signifies. - Just as one also
speaks, without knowing how individual sounds are brought about.
Colloquial language is a part of the human organism and is no less complicated. (T 4.002)
The notion that language works with us, rather than the other way
around, appears at 6.124:
in logic it is not that we express, by means of signs, what we wish, but rather in logic it
is the nature of the naturally necessary signs which itself speaks out.
Wittgenstein's language is truly a mirror; it reflects but is not transpar-
ent.
To grasp Wittgenstein's approach to understanding, it helps to set
understanding language off sharply from understanding a code, a lan-
guage-surrogate whose mere existence is parasitic upon that of lan-
guage. A code is certainly a system of representation which we construct
and over which we hold complete cognitive sway. Codes we master by
internalizing the rules and principles of their representation. Code is
transparent because, by thinking, we determine everything that is
relevant to the job of a code. But language in Tractatus is unlike code:
it waits on nothing for its meaning, we hold no cognitive sway over it
and its working is largely opaque to us.
If analogy is required, language is, in this respect, more like music
or like a monetary system. We create music by setting down notes; this
is one way in which we respond to the influence music has over us. As
parents of teenagers know, there is little chance of establishing control
over that power. Humans make music( al notes); they do not make
music's power. Nor have they any real understanding of the means by
which music takes effect. Language is also akin to a monetary system.
Monetary units are not codes or replacements for goods. 'One dollar'
is not, for instance, a code word for "two loaves of bread". The
workings of the monetary system are little understood (at least by the
vast majority of us). What understanding we have is not from studying
its inner arrangements or underlying principles. As with language, we
learn to live within it. We merely grow accustomed to doing our busi-
ness with slips of colored paper; we do not have to fit an entire system,
somehow, into our heads. The system itself is transformed and renewed
every day through monetary transactions but few individuals exercise
any measure of control over it. To a frightening degree, it controls us.
To that degree is it similar to language.
3.5. Through the Great Looking Glass
It was for Wittgenstein crucial that we stand in no need of a mastery
of language that comes from semantic knowledge; we do not need to
pick up what "comes naturally". As he was to ask in Last Writings,
"What do you need to know to find a smell repulsive?" (Wittgenstein
1982). We should not, as Frege did, conceive of Sinne as objects of
knowledge rather than, merely, as vehicles of it. There is, therefore,
no knowledge of meaning or of language which is a subject of a philo-
sophical inquiry that ends in principles we know in virtue of a mastery
of language. How, then, are we to treat the sayings which semanticists
present as expressive of such mastery, putative rules of denotation such
as
"'cats' denotes cats"?
What would be a suitably Wittgensteinian reaction to a pretended
semantic pronouncement? Consider the claim to have discovered that
Names denote both particulars and univerals.
No longer is it to be counted as a candidate for universal truth about
language. Nor is it to be construed as commentary on the real semantic
structure of an elementary proposition. These have no significant logical
syntax but are mere combinations of names. As a specifically philosoph-
ical pronouncement, Wittgenstein would have thought it an expression
of a linguistic fantasy which is parochial - which is only at home among
incidental features of language. It is viewed as a failed attempt to see
an inessential feature of English, that its sentences divide into subject
and predicate, as an essential aspect of the world. Wittgenstein tells us
that language is a mirror, a looking glass; when we look into it, what
we see in it can be nothing more than our own faces in reverse. The
distinction between particulars and universals is made with respect to
the idiosyncracies of the locatives and temporal particles of English,
parts of the inessential and not, according to Wittgenstein, of the real
sign. (Or, if they turn out to be parts of a real sign, this is not now
known to be so, and, hence, is not now within the province of logic. I
do not believe that Wittgenstein ever shared with Frege the presump-
tion that the real sign might be presented once and for all time to come.
Logic, in Wittgenstein, is a course in perennial linguistic investigation.)
It would seem that a knowledge of the features that make for the
division into particulars and universals is no more tightly bound to a
knowledge of meaning than is knowing American paper money is green
central to knowing the bases of monetary exchange. To put it another
way, the arrow of denotation in Wittgenstein always points back at us
and at our language.
What of the other trappings of denotational semantics and conven-
tional philosophical logic: unsaturatedness, arity or logical typing? As
they are usually addressed - as elements in the conditions which the
world sets down for meaningfulness - they fail. At best, they are
reflections on the character of the signs, either essential or incidental.
There is also no place in the world for indeterminacy, vagueness and
generality. At T 5.1311, you see
That one can conclude from (x).fx to fa shows that generality is present in the symbol
(x).fx.
There is nothing in the world of Gegenstande and Tatsachen that corre-
sponds to these. As we shall see, that world has, as its raison d'tre,
complete determinacy and complete specificity; in fact, it is merely the
objective guarantee of the determinacy of sense. It is correct to hold
that Wittgenstein's world resolves itself into a juxtaposition of objects
and, as such, the need for vagueness, generality, propositional functions
and classes must be referred back to ourselves and to our signs.
4. WHY THERE IS NO TRACTARIAN SEMANTICS OR WHY LOGIC
MUST TAKE CARE OF ITSELF
There is no Tractarian semantics. There is no doctrine in Tractatus of
the actual mechanisms by which our signs are interconnected. Witt-
genstein thought the mechanisms to be hidden from us. In our dream
life we shall not find a scientific account of the true relations between
dream images and the neuronal firings which, I presumably, give rise
to those images. So Wittgenstein took the incidental details of the
Umgangssprache, like dream images, to offer no sure guide to the
workings of language. Philosophers' efforts to come at those workings
through the dream images of metaphysics was counted by him a failure.
To carry the analogy with the dream further, Wittgenstein's efforts in
logic are not directed toward those inner workings; his writings are not
prolegomena to a science of language or a science of dreams. They
take their start from expressions such as 'dream world', the very pros-
pect that there might be such a thing as a world of dreams, as there is
a world of logic. And also from the fact that, in logic as in dreaming,
certain types of errors do not arise. We do not make mistakes in
reporting our dreams. One might say that "dreams take care of them-
selves" .
Therefore, the semantical aspect of logical atomism, the nub of the
attribution of atomism to Wittgenstein, is false and the whole of atom-
ism, as an interpretation, is falsified with it. Still, matters of delicacy
remain; there are things to be explained and to be debated. For exam-
ple, the epistemological part of the Hintikkas' description of Tractatus
is plainly and simply false; it stands in direct contradiction to the text.
It is just false to say that Wittgenstein's Tractarian names were intended
to denote Russellian objects of acquaintance. By contrast, it is not just
false to say that, in Wittgenstein's "fully analyzed propositions", names
will denote items of different logical types, among them properties and
relations. Wittgenstein allowed that there were readings of 'names
denote properties and relations, too' under which it makes a sensible
claim. One of these will be as a none-too-useful grammatical com-
mentary, something of the same order as "nouns and verbs are different
parts of speech". However, these are not readings pleasing to the
semanticist or ontologist. With 'names denote properties and relations',
the ontologist wants (Wittgenstein) to say: "despite what you may have
heard from formalistic and nominalistic philosophers, Plato was right
in thinking that there are special classes of semantical entities such as
properties and that the expressions of our language denote them and
that is how the relevant parts of the language function". Those who
see Wittgenstein as an atomist want Tractatus to include this ontological
proposal.
When the ontologist speaks of properties and their roles, he is draw-
ing a caption on a picture, the icon of the grammatical world. In the
picture is a pre structured world, standing ready across the intentional
gap from language. It is made up of a realm of items with which
words can collaborate and, so, get their semantic powers. Through
collaborations here and there, individual words or names get their
meanings and, thereby, semantic licenses are issued. Some strings of
words receive legitimacy while the world forces us to treat others as
illegitimate. In the grammatical picture, all this collaborating and licen-
sing must go on before logic, since logic presupposes that our names
name and our sentences have meanings. Metaphysics (such as Russell's
doctrine of types) is prologue to logic.
Accordingly, the logician is thought to wait on the grammatical world
to serve up judgment on language. First, there must be an ontology
check. We can, in fact, see Frege carrying out such a check for the
logic of his Grundgesetze. The three sections 29, 30 and 31 together
comprise what we might call the "proof of a metatheorem", in this
case, that announced in the title of Section 32: every proposition of
Begriffsschrift expresses a thought. Proof is by structural induction.
Since Sinn is the mode of presentation of a denotation, Frege starts
with a proof that every name in his concept writing has a unique
denotation. Frege has already divided the class of names up into simple
and complex names, the latter formed by applying various means of
combination on the former. He checks for unique denotation in the
case of each of the simple names. Then, in Section 31, he demonstrates
that the property of bearing unique denotation is preserved through
legitimate combination. It follows from this that every name has a
unique denotation and, given Frege's understanding of Sinn, that an
unambiguous sense has been assigned to each properly formed ex-
pression of the system. Clearly, Frege thought these considerations to
constitute a quasi-mathematical foundation to the later creation of the
logical edifice. He conceived this as a necessary metaphysical check
and only begins to derive logical laws in earnest after the material on
definitions (of which Sections 29 through 32 form a crucial part) is
complete. Frege's definitions and the metalogical strictures on them are
conceived so as to forge the connections between word and structured
world. From these, thought Frege, his logic derived its legitimation.
In Tractatus, logic is not metaphysically legitimated; the play of logic
needs no prologue. As Wittgenstein wrote in the first entry of Note-
books and, in very much the same context, at T 5.473,
Logic must take care of itself. (N, p. 2)
Just as Notebooks begins from this point, one can start here and follow
Wittgenstein's course to the anti-semantical conclusion that there are
no seman tical principles which shore up laws of logic. If we care to
speak of 'legitimation', this is completely reverse. At the bottom, the
world is a semiotic realm and gets its legitimation from the sign. The
world of Tractatus is just a perspective on the sign, a particular way of
seeing otherwise mute items, alphabet letters, footprints or musical
notations as signs. In this case, the perspective is not merely an inner
or mental orientation. The word 'perspective', like its cousins 'point of
view' and 'position' (in German, 'Einstellung') embraces in its sense a
fruitful duality. A perspective involves both conditions we contribute
of our own individual accord and enabling conditions which are there
regardless of what we do. For example, my perspective on a scene may
be that of a view from a high hill. I have that perspective because I
walked up there. I also have it because the hill is itself high. Or a local
official could take the perspective of the village mayor. He can adopt
this outlook because he ran for office and won and, also, because the
village is governed so that there is such an office as mayor. This is a
prefiguring of the duality which Wittgenstein clearly exploits, that be-
tween transcendental and empirical attitudes toward language. It is the
unpacking of this duality of perspective that results in the words of
Tractatus. What follows here is a truncated wholistic rereading of the
first six sections of the book, presenting its main conclusions as those
of a transcendental deduction.
Transcendental conditions in Tractatus are those which need be in
place if a mark is to be a sign in any language whatsoever. Those other
conditions which actually obtain when the mark is a sign of some
particular language(s) comprise the empirical. For instance, for a mark
to be a sign in any language whatsoever, it must be comparable differ-
entially with other signs; we must be able to trace lines of similarity
between it and the others of the system. Signs cannot be lonely but must
form a determinate system. As with Saussure, so with Wittgenstein: the
individual item is sign insofar as it is one cog in a formal or differential
mechanism, a working together of signs.
For the sake of comparison, each sign need be composite (in a quite
specific fashion) or, as Wittgenstein wrote, it must be "articuliert". At
T 5.5261, we are told that comparison is the real sign for articulation.
Characteristic sign of composite symbols: it has something in common with other symbols.
Such articulation, a prospect for comparison, yields immediately divi-
sion into function and argument. Because, if I compare two signs, I
select points of similarity around which there can be planes of variation.
Or, to use a metaphor appropriate to Wittgenstein (who took details
of the vector calculus to be especially revelatory of transcendental
signific conditions), if I compare signs, I must select axes of similarity,
dimensions along which features of the signs can vary in an assessible
fashion. To do so is to separate the sign up into function and argument;
each of which is a field of variation around a point of constancy; they
are both Form and Inhalt or content. Recall that the expression glag2
may be more real as a sign for an object than a; in the former, decompo-
sition into form and content is plain. There are many different ways in
which the sign can decompose into function and argument; none is
prima inter pares. There is no more preeminent division than there is,
to extend the vectorial metaphor, a special scale on the x axis which is
demanded by the structure of space itself.
The ultimate role of Hauptsatz 6 in Tractatus, the claim that
The general form of truth function is [p, g, N(t)].
This is the general form of a proposition.
is to yield, in concise form, the delicate balance between precision and
generality required for the prospect of logic, in other words, the signific
conditions which comparison among signs will - of their own accord -
satisfy. When Wittgenstein wrote of naturnotwendige Zeichen, the nat-
urally necessary sign, the "nature" at issue is in full display in the
Wittgensteinian variable-expression,
[p,t,N(O]
The fields of the expression indicate, by exemplification, the sorts of
articulation required for the transcendental aspect of the signific per-
spective. First, starting at the right, there is logical articulation; that is
the particular office of the Sheffer-like logical sign N. As Wittgenstein
explained, the N acts, in the real sign for a proposition, as punctuation
mark. Here, it reminds us that signs are punctuated so as to admit
arbitrary truth-functional comparisons with the other signs of the sys-
tem. Second, t is a propositional variable and, hence, a sign of gen-
erality. In Tractatus, a sign of generality is a specification of a range of
propositions by means of the similarity conveyed by the form of the
variable itself.
As Wittgenstein liked to remind us, sense must be determinate. This
will assure, first, that the comparisons by similarity and difference which
make for the form of a proposition must be wholly unambiguous.
One way to guarantee this is that each proposition admit of unique
readability; it must be uniquely parsable in terms of a single recursive
mode of combination. As far as truth-functional logic is concerned,
each proposition must come uniquely punctuated in terms of N. But,
this is insufficient guarantee of the determinacy of sense; famously, two
distinct propositions may have identical truth-functional forms. Truth-
functional punctuation cannot, then, exhaust the sense of a proposition.
What remains of its sense is the contribution of Elementarsiitze, proposi-
tions which admit only of a trivial or normalized parsing in terms of N.
Sense is determinate provided that, further, each proposition be re-
solved into a unique selection among elementary propositions. To say
that these comparisons are required for the determinacy of proposi-
tional sense is to say that there must be Tatsachen, facts.
8
Of course, the determinacy of sense or of reading, the unambigu-
ousness of the comparison, must extend to elementary propositions as
well. Such determinacy would be insured if elementary propositions
were, in a sense, digitally rather than analogically individuated. This
will be the case if, first, there is a specific, delimited array of axes or
avenues of comparison among them. Further, each elementary proposi-
tion has a unique resolution in terms of a specific number of such axes
and a single point or vector determined by units measured along the
axes. So, I am thinking of the comparisons between elementary proposi-
tions accomplished by viewing each as a vector in a vector space over
a field which is not continuous (such as the real numbers) but discrete,
such as the naturals. Then, a pair of elementary propositions can be
compared by lining them up along the relevant axes and juxtaposing
vectorial magnitudes, their modes of combination, if you will.
The suggestion implicit in the terminology of 'resolution', 'axes' and
'combination' is not accidental; Wittgenstein took quite seriously the
proposal that the real sign of an elementary proposition is a vector, an
element of a vector space. The image of the real sign as a vector
pervades his early writings. The following quotation from Notebooks
(pp. 20-21), shows that the vector idea was not mere metaphor for
Wittgenstein.
The internal relation between the proposition and its Bedeutung, the way of signifying-
is the system of coordinates that pictures the state-of-things in the proposition. The
proposition corresponds to the basis coordinates. One could consider two coordinates ab
and bp as a proposition which asserts that the material point P fipds itself at the place
(ab).
When, in solving a physical problem, I apply the vector calculus, I
"coordinatize" the given space, the space of the problem, by comparing
it, as a sign, with an abstract vector space (usually R
3
) and with the
latter's coordinate axes. This form of comparison takes place by select-
ing, in the space of the problem, what might be called "instantiations"
of the abstract coordinate axes of the vector space. I "locate" coordi-
nate axes in the space of the problem. Now, I am ready to apply
features or results of the abstract space and the calculus inherent in it
to the presented problem space. The instantiations or locations of
axes are Wittgenstein's Bedeutungen, the objects of Tractatus. The
digitalized axes along which elementary situations are compared are
Tractarian objects. The ambient space coordinatized by the array of
objects is logical space. The magnitude of a vector within that space
reflects the precise way in which a proposition or situation, a vector, is
to be resolved into its objects; this is its pictorial form.
There are at least two things to note about this retelling of the genesis
of objects. First, the object is not some new kind of thing, something
for which we were not prepared by the name in its logical setting. After
all, the application of the vector calculus, the fact that we might apply
it, was not a happy discovery of Descartes or of Newton; application
and calculus evolved as one. The coordinatization of the problem space
does not occur when I draw some malleable or incidental connection
between a name and a particular thing, recognized independently of
the name. I am, instead, confirming what Wittgenstein would call an
"internal" relation between two signs or, equivalently, I am reading
the space of the problem as an instantiation of a vector space. In other
words, I am turning it into a vector sign (or space of such signs) itself.
Second, the comparison between coordinate axes and names casts
light, from a new direction, on the metaphysical aspect of atomism. It
would be indicative of a failure of comprehension on the part of the
student of vector geometry to ask, after the first lesson, "Well, what is
an axis really? Could our rockets run aground on them as they travel
through outer space?" or "Do axes have colors?" Perhaps all the
instructor could do in reply is to repeat her presentation of the vector
space concept; a coordinate axis has no more to its nature than its place
as a gear in the mathematical mechanism of vector calculus. To ask
after the true nature of Tractarian objects or to ask whether they are
particulars or universals is to ask questions of a 'similar kind. Objects
are exhausted by their logical roles.
Transcendental conditions establish the possibility of comparisons.
Empirically, in the case of any particular human language, actual com-
parisons will be in force. These will determine the particular class of
objects of that language by setting the referential relations in which
names stand to them. But that there be objects - the condition that
there must be some objects or other - is transcendental. It is the demand
that there be simple names such that, in virtue of their functioning as
simple names, we say that Bedeutungen are associated with them. By
Wittgenstein's lights, this is implicit in the very nature (again, 'natur-
notwendige Zeichen') of signing. All of this comes from a reflection
upon the differential comparisons that confirm the mark as the sign;
the comparisons are not wholly arbitrary but give the mark a definite
position within the system of signs. The sense of a proposition is deter-
mined by its position in the system, in the space of signs. The Bedeutung
of a name is wholly determined by its position. Facts and objects - or
sentences and names - are ways of addressing the definiteness of those
positions.
There is yet more to analysis; there is the dissolution of complexes.
To see the need for it, one returns to the general form of a proposition
and to Hauptsatz 6. The need derives not from the appearance of
Sheffer's N but from the presence of a notation for generality, t. To
carry on a decomposition analysis of a complex is to expose the precise
ways in which lines of similarity are drawn between the elements of
propositions. As the generality notation indicates, suitable proposi-
tional signs will include variables, indices of generality. These are speci-
fications of ranges of propositions, ways of collapsing, in a single sign
such as ~ a whole panoply of signs. The general signal supplies no-
tational economy at the price of some notational indefiniteness. Witt-
genstein wrote, at 3.24,
[t]hat a propositional element indicates a complex one can see from an indeterminacy in
the propositions in which enters. We know that everything is not yet determined through
this proposition. (The generality sign contains a prototype.)
A sign containing generality index g, say Fg, does manifest indefi-
niteness. If ~ is the collapse or synopsis of this range of propositional
signs:
Fa, Fb, Fe,
then, with respect to negation, Fg admits an ambiguity of scope. It is
exposed by asking what the denial of Fg could involve. One can deny
Fg - or put forward its negation - by denying Fa or by denying Fb or
by denying Fe or by denying any positive combination of these. Hence,
is the sense of the negation of ~ indefinite and indefiniteness is some-
thing Wittgenstein's world will not allow.
By having a determinate sense, each proposition fixes the world on
either 'yes' or 'no' in an unambiguous fashion: determinacy of sense
applies to proposition and negation equally. Wittgenstein's world is
simply the transcendental rationale for the determinacy of sense. There-
fore, such indeterminacy we find in ~ must be a phantasm of the sign,
a shadow in the thicket and not what the thicket hides. By analysis is
it shown to be mere appearance. By analysis, all indefiniteness, all
generality in the sign, must be expanded out; each variable must ulti-
mately vanish and be replaced by the range of propositions for which
it is a collapse. As we shall see, this makes a marked contrast between
Wittgenstein's and Russell's concepts of analysis. Russell's analysis via
definite and indefinite descriptions is the logical tyranny of generality:
naming is traded away for quantifiers. In Wittgenstein, things are just
the opposite: analysis is the elimination of all generality. (Cf. the final
section of the present paper.)
In sum, Wittgenstein's world is a perspective on the system of signs.
The exegetical world of the present writing is a perspective on the signs
of Tractatus. From this perspective, the first part of Tractatus reappears
as a transcendental deduction. It takes its start from the observation
that we calculate successfully with signs - and with signs alone - using
the techniques of propositional logic. It has as its goal a satisfying
answer to the question "What makes pure logic and mathematics pos-
sible?" That which is deduced is Wittgenstein's Hauptsatz 6 and what
comes from its unpacking: the system of signs falling apart into truth
functions, elementary propositions, functions, arguments and names.
Objects and facts are, then, fields of logical similarity between proposi-
tions and between names, respectively.
To say that the world falls apart into facts and facts into objects is
not, belatedly and reluctantly, to give credence to the Hintikkas' logical
atomism. To emphasize Zerfall, either in the form of the logical falling-
apart of sign into function and argument or in the form of decomposi-
tional analysis, is not to embrace the bottom-up cognitive process of
semantic composition. The sense of a proposition is not a cognitive
value which I win in a game of semantic calculation. That game has
no rules here; there are no principles of semantic composition which
legitimize the sign and guarantee for it a sense. I do not build Wittgen-
stein's world of meaning out of a mindedness of atoms. Just as the role
of the executive in the corporation is not a mere composite of individual
tasks but a systematic array of them, so also the role of the propositional
sign in a system of signs is not a mere composite of individual, atomic
roles, each of which independent of the others.
I freely admit that there are simple objects or Tractarian atoms. But
I insist that there is no more to them than their functional role, a role
which we abbreviate by speaking of them as the Bedeutungen of simple
names. Atoms are functional units, such as the electromotive force and
the resistance in an electric circuit. Analysis of circuity is conducted by
a division - into force and resistance; the points of division are deter-
mined only by the limits set by laws such as Ohm's. Atoms are not
independent substances. They serve the purposes of the means by which
they are seen, the need for precision in the logical comparisons of signs.
For these reasons, Wittgenstein's logic must be "top down" and not
"bottom up". Logic cannot begin, as in the icon of atomism, by the
examination of the logical atoms. Put another way, speaking or reading
cannot begin with naming. In this regard, one should compare T 5.526:
"One can describe the world completely through entirely generalized
propositions, that means, therefore, without correlating, from the be-
ginning, any name with a definite object".
Logical properties do not "flow upward" through language from the
properties of the elements of the putative foundation. Language and
its logic do not stand against the world, mirroring it by copying its
grammatical structure atom by atom. This is just another way of paint-
ing the picture of the certification of language by ontology, a picture
which must be rejected if we are to start reading Tractatus properly.
As Wittgenstein wrote at 5.634, "there is no a priori order of things"
and the intended order of atoms is an a priori order if anything is.
But what has all this to do with logic as a process of calculation and
assessment? How do comparison, articulation and Zerfall prepare for
its success, for the fact that we can succeed? First, there must be
articulation in logic; it is essential to representing the validity of modus
ponens (and invalidity of asserting the consequent) that signs are repe-
ated and their locations compared. We look to displays such as
~ Q
P
Therefore, Q.
Second, logic goes on without further ado. We do not check the logic
as Frege did; we do not try to stare outside the penumbra of light cast
by the signs to see that, once we have set down the symbols of the
calculus and their relations to each other, once we have indicated how
the logical tool is to be put to work, we can apply it. We can then
judge, with perfect accuracy, of the validity of contentual inferences in
everyday discourse. We use the signs just as they are to give logical
"proofs". Of this feature of the logical tool, Wittgenstein said:
And we do this when we "prove" a logical proposition. For, without bothering ourselves
a whit over sense or Bedeutung, we form abstractly rbi/den] the logical proposition from
others according to mere rules for signs. (T, 6.126)
This is a kind of aprioricity for logic. The question remains "What
makes this property of logic possible?" A Wittgensteinian answer is
that the signs of logic need no special license, once they are constituted
as signs. That means: once they are incorporated into the perspective
on which they are viewed as signs. By the attainment of such perspective
(with its dual aspect), certain conditions are au\omatically imposed -
without further ado. For logic, these are the transcendental conditions
for what I called "the mechanism of language", not a pseudolanguage
or scheme of representation living in our minds but a language setting
in which we live. From the aprioricity of logic (and the transcendental
deduction), we appreciate that the conditions are in place. Matters such
as unique parsing are not conditions which the logician qua ontologist
has to check before he can proceed; there is no call for him to look over
his shoulder. Transcendental conditions are emergent characteristics of
the signific medium; they evolve automatically as the mark becomes
sign. Here is a fragment of the Notebooks entry from 15 December
1914.
It is obvious, we can introduce whatever we want as the written sign of the ab-functions
[truth functions) the real sign will form itself abstractly [bilden sieh) and automatically.
And which properties will form themselves thereby?
Again, the logician need no more look over his shoulder than the
sculptor has to check that the laws of spacetimy are still operative in
his studio or than the mathematician needs to check that pencil and
paper are still suitable media for the conveyance of mathematics.
The philosopher's icon of naming - an inexplicable triple liaison
between the epistemic subject, a mute, structureless tag and some
isolated item - has no role to play here. Naming is no prerequisite for
concept writing. Nor is there a place for the philosopher's icon of
meaning: special act of thinking or intending. Again, the page is blank;
there can be no writing while the page is still empty. There can be no
thought before thought can begin. The beginning of thought is coeval
(v.s.) with logic. Any slack left by the banishment of philosophers'
meaning and naming is fully taken up by comparative similarity. Allow
me to repeat sections from T 5.526 and 5.5261.
One can describe the world completely through entirely generalized propositions, that
means, therefore, without correlating, from the beginning, any name with a definite
object. (5.526)
An entirely generalized proposition is, as with other propositions, composite. (This is
revealed in the fact that we, in '(3x, q,). q,x', mention 'q,' and 'x' separately. Both stand
independently in signifying relations to the world, as in an ungeneralized proposition.)
The characteristic mark of a composite symbol: it has something in common with other
symbols. (5.5261)
If there is a legitimate need for a "connection between word and
world", comparative similarity is all that is required, as the possibility
for fully generalized description shows.
4.1. The Ineffability of (what remains to) Semantics
Wittgenstein wrote that reality is itself determined as 'yes' or 'no'
through the proposition. The ultimate locus of this determination and
of the space of propositions is language transcendental, "the" language,
where "the" language is neither one or another particular language nor
some conjectured ideal language. It is semiotic language, language
as all-embracing scheme of presentation. As is clear from the early
Notebooks, what we consider experience - Erfahrung - is already in-
cluded within language. Experiential episodes are themselves read.
Experience is itself a panoply of signs and, as such, becomes a facet of
the language. Experience does not stand outside of language and op-
posed to it. A fortiori, it is not a field to which I can refer for advice
on the construction of an appropriate language, as the atomists would
like.'
What, in the first place, the proposition determines in reality is a
truth-conditional arrangement of such 'yes's' and 'no's'. One might
think of this as a tagging of states-of-things. (Or, though there is some
danger of confusion, one could conceive of the arrangement as the
graph of the (possibly infinitary) Boolean function defined by the logical
form of the proposition.) This arrangement in logical space serves as
the content of a thought in Tractatus. Wittgenstein held that the con-
ditions under which a sign is well-formed, under which it is capable of
expressing a sense, do not themselves constitute the content of a
thought. They do not themselves admit of formulation which in turn
admits of truth or falsehood; these are things which go unsaid. Indeed,
this unsayability or ineffability (to use the Hintikkas' expression) must
apply to at least some propositions if certain assumptions are imposed.
First, assume that every statable condition requires there to be some
proposition which determines it. Second, assume that no proposition
contributes to the determination of its own well-formedness. And,
lastly, assume that a determination of well-formedness is always well-
founded: there is no infinite w-chain of propositions each of which
serves, if only in part, to determine the well-formedness of the preced-
ing proposition in the chain.9 The latter two conditions will follow from
the assumptions that everything going into the determination of the
meaningfulness of a sign is, when properly understood, part of the
symbol attached to that sign and that the process of analysis, the
subdivision of the symbol which occurs naturally in the creation of
sense, is well-founded. As Wittgenstein wrote at T 3.332, no proposi-
tional sign can be contained in itself and, if the process of analysis were
not well-founded, the prospect of self-containment would open up. It
is clear that the well-foundedness of analysis is required for the determi-
nacy of sense.
Ineffability, the fact that those conditions which go into making a sign
fit to express a sense are themselves not statable, was for Wittgenstein
characteristic of language transcendental. Everything statable is spoken
within and in accord with the transcendental conditions of language,
conditions which are themselves unconditioned. The transcendental
perspective is a tectonic plate on which Wittgenstein's exposition is
founded and which breaks through and surfaces at crucial junctures.
One of these is T 2.0211. Wittgenstein here insisted that the very
possibility of sense depends upon a radical independence of conditions
on meaningfulness from conditions on truth. Whether one proposition
has meaning cannot, he said, depend upon whether another is true.
This draws a firm line across philosophical logic; Wittgenstein's i n v s t i ~
gations of the sign lie on one side and semantics - both Russell's theory
of types and Frege's theory of function and object - on the other. On
the far side of the line, with Russell and Frege, lies the notion that
meaninglessness is simply a species of falsehood, a kind of "superfalse-
hood" if you will. Here, it is assumed that, as the truth of a proposition
certifies it as a faithful report on the factual dispositions of objects, so
also the meaningfulness of a proposition certifies it as a faithful report
upon their metaphysical dispositions.
To see a doctrine of meaninglessness as falsehood at work, consider
a Frege-style explanation of the fact that 'Bill is tall', as ordinarily
understood, expresses a sense while 'is tall Bill' does not. (I ignore, for
the moment, Frege's reaction to the vagueness of predicates such as
' ... is tall'.) Of the many possible destructive analyses or "dissolutions"
of 'Bill is tall' into arrays of logically significant fragments, one was of
crucial import for Frege. It is the one that presents the true semantical
Bausteine or building blocks of the sentence. Unquestionably, it would
be the dissolution into a saturated term 'Bill' and an unsaturated ex-
pression (written in a highly un-Fregean style) ' ... is tall'. No account
of our semantical ken of 'Bill is tall' could be adequate (for Frege) and
exclude a description of this division and its denotational congruence
to two "items", one an individual named by 'Bill' and the other an
unsaturated nonindividual, a function or concept, the Bedeutung of
' ... is tall'. To expose the sense of 'Bill' is to expose the cognitive
means by which it attaches to the individual Bill and, mutatis mutandis,
the same applies to ' ... is tall'.
It is in terms of these extralinguistic attachments and, in particular,
of the metaphysical properties of and relations among items of these
sorts - whether they be saturated or unsaturated, of which level they
belong and of what number of arguments they can take - that Frege
would care to offer an explanation of the capacities of signs to make
sense. For instance, it is the unsaturatedness of the function signified
by ' ... is tall' which will, supposedly, account for the facts that atomic
sentences exhibit a logical grammar. 'Bill is tall' can express a sense
while 'Bill Bill', 'is tall is tall' and 'tall is Bill' do not. 'Is tall is tall' is
senseless since the unsaturatedness in ' ... is tall' is not properly com-
pleted by an unsaturated nonobject such as ' ... is taU' itself. Notice,
however, that, strange as it may seem, the ability ascribed to us by this
explanation requires that we be able to do something akin to reading
the supposedly senseless inscription 'is tall is tall'. We must, in con-
fronting 'is tall is tall', discern two occurrences of the' ... is tall' radical,
associate with each the Bedeutung which is the " ... is tall" function
and, then, note that the attempted congruence or fit fails, due to the
logical dispositions, the unsaturatedness, of the "is tall" function. We
thereby record, in a statable condition, the situation which corresponds
to a particular failure of a sign. In Russell, flaunting of a typing rule
will generate a precisely analogous explanation of and report upon the
failure of a sign. It will also require that, per impossible, we read a sign
which the theory says cannot be read. In this way do Frege and Russell
treat the blank page as if it were a highly subtle form of writing, rather
than a voiceless medium. This is a reason for Wittgenstein's rejection
of semantics.
This is not merely a long-winded way of saying that Wittgenstein
refused the Fregean metaphysics of sense. Wittgenstein certainly re-
fused to take Sinne as real but nonactual objects, strange citizens of das
dritte Reich. He did not do so after the fashion of some contemporary
philosophers who reject the "ontological commitment" but continue to
do homage to the cognitive role of senses in some other, "less com-
mitted" fashion. A certain sort of knowledge does flow through Trac-
tatus from a grasp of sense (as we see in 4.024), but, if there is something
called 'knowledge' which is prerequisite to it, it is not knowledge of a
liaison among objects. Nor is it a knowledge of a set of constructions
by which we might build a seman tical whole out of epistemic parts.
Wittgenstein did not refrain from all talk of Sinn. He clearly recognized
the need for a (a kind of) sense-reference distinction. Complete propo-
sitions, however, were the only sorts of linguistic expressions to have
both. He also worked the notion of sense in a crucial way to explicate
what he intended by 'an expression', ein Ausdruck. However, it is
important to note that, unlike Frege, Wittgenstein holds that interpre-
ted propositions do not contain their sense (cf. 3.13).
4.2. Plato's Problem and Other Matters Metaphysical
To preserve the radical independence of sense from truth, to keep
language transcendental from slipping into language empirical, requires
philosophical effort. Wittgenstein had to push hard against various
supposed truisms. As a result, he maintained the following four points:
(1) there is no distinction in fact between a sentence and a mere
list of words;
(2) the familiar Fregean explanation for the "creativity of lan-
guage" is unacceptable;
(3) there is no unsaturatedness, no arity, no generality in the
extralinguistic world; and that
(4) de notational semantics, as an attempt to explicate the logical
fit between word and world, is a failure.
Point (1): In Sophist, Plato asked after the distinction in fact between
a sentence and a mere list of words. In Notebooks, the answer to
Socrates's question is plain. On 28 May 1915, Wittgenstein wrote that
"'composite sign' and 'proposition' are coextensive". We find, in the
entry two days later and following after the words cited at the very
start of this article, "It is clear that it comes to the same thing to ask
what a sentence is as it does to ask what a complex is".
Wittgenstein is here adverting to the commonplace that, in explaining
what a list of words is, we point to an articulate sign. In explaining
what a sentence is, we point to an articulate sign as well - perhaps even
to the very same sign. What Wittgenstein took to be the telling differ-
ence between the explanatory episodes, as subsequent considerations
in Notebooks indicate, lies not in the arrangement of the signs on the
page or in the air. Nor does it lie "below them", in any special semantic
arrangement among the denotations of the portions of the signs.
Rather, the differences, if any, lie in the analyses of the signs - in
the ways in which comparisons are drawn and in which the signs are
marshalled together with others. Differences lie in the ways we draw
the lines of similarity and that is a matter of our history rather than of
our metaphysics. And to say that it is a 'matter of history' is not to
propose a reduction of semantics to history or a replacement of seman-
tics by history. The facts of history cannot stand in a justificatory
relation to our play with signs.
The distance here between atomism and wholism can be measured
using the question
What distinguishes a(n idle) mark from a (meaningful) sign?
The atomists have a ready answer. When answered in the style of
Russell, it is:
A mark is a sign either when it is a logically proper name, so that its meaning is just its
referent, or when it admits, in context, of a uniform analysis so that any significance it
has reduces ultimately to that of naming.
To answer in a Fregean voice, we would say:
A mark is a sign when it has been assigned a reference; in the process of assignment, it
comes to express a sense and, so, to contribute in a regular fashion to the expression of
judgable content, namely, thought.
As we have seen, Wittgenstein's commitment to the absolute indepen-
dence, the absolute priority, of a (thoughtless, mechanical) logic has
him refuse such answers. On either of these alternatives, the meaning-
fulness of a sign becomes a sort of fact in which the sign can participate
or, equivalently, the meaningfulness of a sign is a kind of property of
it, one which is itself an object of thought. It is a feature of such
accounts to presume that what we call language and what we call
grammar are written already into the world itself - that it is a plain
fact that 'cat is the mat on the' is not meaningful and that this fact is
of much the same sort as the fact that London is on the Thames.
On the views of Russell and Frege, there is a very good sense on
which a logical knowledge is posterior to an ability to read these sorts
of facts off the world. First, there is given to us a collection of items
or quasi-items (objects, functions or objects of acquaintance) which
are, in the philosophically explanatory order, prior to language. Then,
the mark becomes sign through its attachment by thought to these
objects. Once attached, the sort of object it is determines the sorts of
linguistic relations in which the mark can sensibly stand. In this way,
there come to be various "principles of proper attachment" such as 'a
first level concept cannot be predicated of another first level concept'
or 'two individuals alone cannot constitute a content of judgment'.
These are conditions which the mark is intended to satisfy in order that
it could be a sign. It is by a process of thought that I supposedly
determine the mark's qualifications for such a lofty task. Here, both
extralinguistic world and thought are suitably prior to language.
To the question "What distinguishes a(n idle) mark from a (meaning-
ful) sign?", one of the answers Wittgenstein might have preferred to
give was "Nothing in the world: there is no real difference in fact
between a Shakespearean sonnet and any other kind of mark, say, a
muddy footprint". There is nothing metaphysical in the world or in the
mark per se that qualifies it to be a bearer of meaning. The only
difference is a contingent one, an historical difference: that the muddy
footprint is yet to be included into the whole structure of signs. Or,
simply, that I have yet to read it. And that inclusion is, in turn, nothing
philosophically special; we might say that it is nothing more than this:
we do not see it as language. It is yet to be linked with the mechanism.
Point (2): The metaphysics of function and object plays a starring role
in the Fregean story about what Chomsky has popularized under the
title 'the creativity of language'. In that story, we can be creative with
language as long as we do not exceed semantic limits. The story starts
with the putative datum that we understand sentences which are, to
some extent, novel in that we have never encountered this particular
sentence before. I suppose one of these might be "My llama has never
encountered a lama". The delimited extent to which it is novel is a
depiction of the limits of our creativity. It is novel in that it is a new
instantiation of an old, recognizable form. It is a combination of known
words, such as 'lama', into a form whose principle I have already
recognized, at least implicitly.
It is a new linguistic house in a recognized style. We can cope with
it because we can build it up semantically, from materials which are
familiar and on a familiar plan or so the story goes. We can arrange
the old materials - the ones given as basic function and argument - in
new and surprising but recognizable ways. For example, even if no one
had ever seen the sentence (form) 'Bill is tall and so is Fred' before,
anyone who grasped 'Bill', 'Fred', ' ... is tall' and 'and so' should be
able to grasp its significance. The occurrence of 'Bill' there continues
to stand to its object - Bill - in the same relation as before. As does
the expression ' ... is tall'; it refers to the same function as before and
one can realize as much. Also, one can attribute a real content to 'and
so' along parallel lines: it too denotes a function, but of an attenuated
kind. Our knowledge of these relations and interrelations conspire
together to produce a knowledge of the meaning of 'Bill is tall and so
is Fred'. In summary, we defer to the objective relations between words
and items and to the permanency of our ken of these relations when
in comes to the governance of our "creativity" in moving from simple
sentences to complex ones. In other words, we can be creative with
language as long as we still know what we are doing.
Wittgenstein is refusing to items in the world the powers
which the story of linguistic creativity attributes to them. They now
exert no control over expressive capabilities of signs. Insofar as the
purported explanation of that creativity depends upon our remaining
within the bounds set by those powers, that explanation must be re-
jected by him as well. In Wittgenstein's writings, he wipes the explana-
tory slate clean and declares the creativity of language to be absolute.
We are not limited by any "knowing what we're doing" of a semantical
kind.
Absolute creativity is a recurrent theme, set as early as the second
sentence of the Zweites Manuskript of Aufzeichnungen uber Logik:
"Every sentence is a new symbol indeed". The freedom we exercise
under the guise of "creativity of language" is - as far as principles of
meaning are concerned - absolute. It is literally but trivially true that
when we write down a wholly new sentence such as "My pet ferret
loves Brahms", we are creating a wholly new sign - what Wittgenstein
would have called the 'perceptual part of a symbol'. New relations of
similarity are marked out. Simultaneous with this creation of the physi-
cal sign is the coming-in to-existence of an immanent nonphysical "ob-
ject", a symbol, which we can conceive as a convenient condensation
of the logical properties of the sign as determineti by the way in which
the sign features in logical manipulations. In creating the symbol, our
sign is fit into the pictorial mechanism of Tractatus and, thereby, it
comes to express sense; it is thereby subject to logical treatment.
The creativity of language, like the creativity of the plastic artist, is
absolute. There need be no "world checks" before we begin to speak
as there need be no "space checks" before we begin to sculpt. The
sculptor's ability to refrain from constructing pieces which violate the
laws of three-dimensional geometry is not a function of the sculptor's
knowledge but of the sculptor's medium. In the same way, our ability
to work logically with the new sign is not a function of semantic knowl-
edge. It is a feature of our standing within the logical medium. There
is no worry; we need no guardian angel to enforce the "law" that our
signs continue to carry meaning; the mechanism of language, of symbol,
does that for us. Since there are really no laws, there is nothing for us
to violate, no prohibitions - of a logical sort - to break. We cannot
break out of the logical medium in working with signs. All of the
logically (and, hence, philosophically) relevant properties of the in-
terpreted sign, the Ausdruck, must be aspects of the logical medium of
the sign, of the objective setting. Hence, the onJy requirement on the
fitness of sign to give rise to a symbol, to start to cover a region in the
medium, is that it be capable of appearing on the map at all, that it
be, in Wittgenstein's terms, articulate. Hence, as far as properties of
the sign subject to a logician's enquiry matter, there can be no difference
between 'Bill is tall' and 'tall is Bill'. Both of these cover an area in
logical space, both can function as maps. In the case of the latter, we
just do not know what the scale of the map is yet. And that, according
to Wittgenstein, is a relative triviality.
This very point - this very reaction to the creativity of language -
surfaces on a number of occasions, each time under the banner headline
"Logic must care for itself". The first is at the start of Notebooks 1914-
1916; the other is at Tractatus 5.473ff. The idea expressed in both is
that, in relation to the rest of philosophy, including semantics, episte-
mology and metaphysics, logic is foundational. The only conditions set
on my creativity are ones I cannot check, the transcendental conditions
which make for the prospect of logic. Hence, Wittgenstein is rejecting
the Fregean story of the sensibility of signs told in terms of the meta-
physical properties of the things which signs name. In each of its appear-
ances, the headline is followed by a reassertion of the absolute creativity
of language in the form of the claim that, in contemporary terms, there
is no such thing as a logical syntax of atomic sentences. Wittgenstein
wrote in Notebooks, "and I say: every possible sentence is well-for-
med".
Of course, Wittgenstein has to make room somewhere for our strong
feelings for grammar and ungrammaticality. After all, there just is
something wrong with 'is tall is tall' which is not wrong with 'Bill is
tall'. Plato was right, at least, to worry about the problem. Wittgenstein
seems to credit our feelings by recognizing a subject of (nonlogical)
grammar. He wrote (in the second paragraph of Notebooks):
Let us remind ourselves of the explanation why 'Socrates is Plato' is meaningless. Namely,
because we have not yet come to an arbitrary determination, and NOT because the sign
might be, in and of itself, illegitimate.
This sentiment is repeated - and as emphatically - at 5.473ff in Trac-
tatus. The presence of the italics is indicative of Wittgenstein's attitude.
The purported grammatic explanation is simply the citation of a fact -
and a contingent one - about us: that we, at present, find certain
combinations agreeable, others disagreeable but that, in future, we
might make determinations so as to change that situation. For Witt-
genstein, the study of grammar - logical or otherwise - is an investiga-
tive path that always leads us back' to parts of ourselves. Sometimes,
as in the case of a study of those particular objects to which we refer
in language, our study leads us to essential features of ourselves. In the
case of a study of ordinary grammar, it leads - Wittgenstein thought - to
certain accidental features of our nomenclature. This is merely the
grammatical partner to Wittgenstein's response to denotational seman-
tics: ordinary grammar, like the arrow of denotation, always points
back to ourselves.
Point (3): In Tractatus, there are no necessities except the logical, those
which are generated by the very prospect of signing, those which limit
us as does the blankness of the page. As we have seen, there can be
nothing in the nature of a denotation which sets a limit upon the
admissible grammar of a sign, or else our creativity would be limited.
Also, at T 5.453, we see
All numbers in logic must justify themselves. Or, more to the point, it must be apparent
that there are no numbers in logic. There are no privileged numbers.
Start with any of these lines of thought and you arrive at the same
place: if there are properties and relations, it is mistaken to think that,
in and of themselves, they have arities. Were there to be in the world
"out there" a property named by 'left of' and if it, as a binary relation,
had the number two as its arity, then it would be a nonlogical necessary
truth that the property of "leftness" has arity two. In the same way,
arities would set objective, extralinguistic limits to the completions
which the fragment 'Charlie is to the left of ... ' can meaningfully take.
These would limit the absolute creativity of language. And, if there
were arities, then there would certainly be privileged numbers; we
might say, as do many logicians, that the possession of an arity of 0
(zero) is the trademark of a logical individual. Passage T 5.553 makes
explicit the connection - in Wittgenstein's thought - between "privi-
leged numbers" and arities:
Russell said there are simple relations among various numbers of things (individuals).
But between what numbers? And how is that supposed to be decided? - Through
experience? (There is no privileged number.)
Arities are privileged numbers and must fall by the wayside. Unsatu-
ratedness, universality and generality must do so FlS well, at least where
these are meant to be semantically explanatory features of a realm of
reference. If functions are, as Frege held, in themselves unsaturated,
then they will each possess some index of unsaturatedness, an arity, a
number which will tell us how many names can be used to fill up the
distinguishable failures in saturation. If we feel constrained to mix the
metaphors of Tractatus with those of traditional philosophy (a constraint
which, I suggest, ought to be avoidable), we might say that everything
in the Tractarian world is particular and universality or generality can
only be lodged in the character of the sign. But, in truth, it would be
better to say that the distinction particular-universal does not apply at
all.
Point (4): If denotational semantics is a discipline which purports to
expose philosophically important structural connections between sign
and signified, then, in Tractatus, there is no denotational semantics.
Such a semantics presupposes that there is a semantically explanatory
structure in the world itself. According to Wittgenstein, there is no
such structure; anything which we, as logicians and philosophers, think
to see in the signified has been misplaced from the signifier - or so goes
Tractatus. In consequence, the exposition of Tractatus is hardly to find
much solace in, say, a model-theoretic rendering of Tarski's "theory of
truth" .
Admittedly, there are the occasional passages where Wittgenstein
might be read as subscribing to a "theory of truth". Here is one from
the early Notes on Logic in which Wittgenstein seems to be identifying
the process of setting out the (pictorial) form of an atomic proposition
xRy with the process of specifying (what we would call) the satisfaction
conditions of the atomic relational predicate uRr. The latter conditions
are determined by describing (what we might call) a suitable set of
ordered pairs to serve as the extension of R.
The form of a proposition may be symbolized in the following way: Let us consider
symbols of the form "xRy", to which correspond primarily pairs of objects of which one
has the name "x", the other the name "y". The x's and y's stand in various relations to
each other, and among other relations the relation R holds between some and not
between others. I now determine the sense of "xRy" by laying down the rule: when the
facts behave in regard to "xRy" so that the meaning (or reference) of "x" stands in the
relation R to the meaning (or reference) of "y", then I say that these facts are "of
like sense" with the proposition "xRy"; otherwise, "of the opposite sense" .... Thus I
understand the form "xRy" when I know that it discriminates the behaviour of x and y
according as these stand in the relation R or not. (pp. 98-99)
The Hintikkas seem to have fixed upon passages like these as providing
proof positive of the logical aspect of their atomistic interpretation.
They seemed particularly enamored of the idea that there is at most a
rhetorical distinction between the so called "picture theory" of Trac-
tatus and Tarskian truth theory for languages in standard formalization.
In fact, Wittgenstein came to make a sharp separation between the
specification of the form of a proposition and a specification of its
conditions of truth. In other words, he seems to have abandoned, in
the general case, the sort of identification which is perhaps suggested
by the preceding citation. (I happen to think that Wittgenstein never
meant to endorse, and did mean to repudiate, an equation between his
picture theory and our denotational semantics.) By the time he wrote
in his notebooks on 1 November 1914, he was distinguishing between
the pictorial relation and the relation of truth in no uncertain terms:
Very close lies the mix-up between the representing relation of a proposition to its
Bedeutung and the truth relation. The former is different for different propositions, the
latter is one and for all propositions the same. (p. 22)
Further on in Notebooks, we find equally definitive repudiations of the
denotational view - at least as applied to what we would take to be
paradigmatic examples of unary and relational predications. Witt-
genstein is refusing outright to allow that my knowledge of the meaning
of sentences such as 'Socrates is mortal' and 'The watch is lying on the
table' is to be accounted for on the basis of a knowledge of the deno-
tations of expressions such as the particularizer 'Socrates' and the gen-
eral expression 'is lying'. In the midst of the longest concerted discussion
in Notebooks - the examination of the claim "objects are simple" -
Wittgenstein wrote:
However the logic, roughly as it stands in Principia Mathematica, allows itself to be
applied perfectly well to our ordinary propositions. For example, from "All men are
mortal" and "Socrates is a man" it follows, according to this logic, that "Socrates is
mortal", which is obviously correct, even though I, just as obviously, do not know what
structure the thing Socrates or the property of mortality'has. These function, even here,
as simple objects. (p. 69)
Further remarks, recorded on the succeeding day, reinforce the point
in other terms.
I say to someone "the watch is lying on the table", and now he says "yes, but if the
watch were to lie in such-and-such a way, would you still definitely say 'it is lying on the
table"'? And I would become uncertain. That shows that I'do not know what I meant
by 'lying' in general. (p. 70)
Surely, if I thought my knowledge of the meanings of 'Socrates', 'is
mortal' and 'lying' to be encapsulated in a knowledge of standard
definitions of the forms, for all x and y,
x is the denotation of 'Socrates' iff <I>(x)
x satisfies 'is mortal' iff'l'(x) and
x and y satisfy 'is lying on' iff 8(x, y),
for suitable <1>, 'I' and 8, then it would be strange to claim that I
obviously do not know the semantically relevant structure of either
Socrates or mortality. It would be equally unusual to claim that I do
not know the meaning of 'is lying on' in general, since, ex hypothesi,
to know the meaning in general is just to be apprised of a suitable
statement of satisfaction conditions.
As should be clear from the text which surrounds the passages just
cited, Wittgenstein has come to think that there is a good deal more
to the "semantic structure" of statements such as 'the watch is lying on
the table' and 'Socrates is mortal' than meets the eye of ordinary logical
grammar. It is also clear that this is unavailable to the logician and to
the denotational semanticist. It is a contingent part of our organism
and no less complicated than it. Wittgenstein believed that there may
be a further investigation of the relations of reference - an investigation
of the actual means by which terms of ordinary English are to be
analyzed. But this investigation is the province of the psychologist and
not of the logician.
5. TEXT AND TRANSLATION
For the logical atomist, some form of reasonably conventional grammar
is of paramount importance. The atomist believes that grammar must
be certified by something outside it, that the world will vouch for
what Wittgenstein called the spezielle Haken of grammar. Grammar is
elevated to the level of a special description of the true structure of the
world. Of course, individual words are not grammatically structured.
And, as the stock of grammar goes up, that of individual words must
fall in relation to it. Word meaning becomes a subject beneath mention
or, if there is a question of the meaning of a word, it is supposed to
be settled by gesturing toward an item. There comes to be little in the
world for words.
In the face of the text, it is as if those who favor an atomistic reading
of Wittgenstein practice what they would have Wittgenstein preach. At
their pens, the budget of Wittgenstein's philosophical power has been
drained from the flow of words and channeled into motionless back-
waters of semantic principles and doctrines. But, in truth, the details
of his expression, the particular words chosen by Wittgenstein, come
neither before nor behind, but are wholly inseparable from the "con-
tent" or message of Wittgenstein's words. This is an ideal to be served
in reading Wittgenstein and, once served, an ideal which directs us
away from atomism.
5.1. Numbering Tractatus
Objects are of first importance; this is the semantical plank of logical
atomism. In Investigating Wittgenstein, the Hintikkas wrote the "crucial
subject matter in trying to understand the Tractatus is obviously
Wittgenstein's conception of object". By contrast, those concepts which
the who list would take to stand equally among the real working parts
of the machinery: logischer Raum, Tatsache and Sinn are virtually
shunned. I believe that there is at most one mention (and that in a
quotation from Wittgenstein) of the concept of logical space in In-
vestigating Wittgenstein. The consecration of objects (and relative de-
nigration of other concepts) stands at odds with the very structure of
the text.
In the footnote to page Tl, Wittgenstein troubled to tell us that the
numbers on his remarks accord with their relative import. If so, then
the Hauptsiitze, whole-numbered propositions 1 through 7, have first
claim to attention. After that in importance come remarks whose num-
bers are single decimals, remarks such as 3.1. Next are the double
decimal digit claims, and so on. The notion of object - for which
Wittgenstein often used the logical term 'Gegenstand' - features neither
in the Hauptsiitze nor in any proposition with single decimal digit
number. In fact, Gegenstiinde receive only one concerted treatment -
and that in a relatively brief series of remarks following items 2.01 and
2.02. Rated on the same score card - in terms of numbered appearance,
logical space rates much higher than object. Reference to 'logischer
Raum' appears, and not for the first time, in the single-digit proposition
3.4,
A proposition determines a place in logical space.
On this scale, the term which is accorded greatest import is 'the world',
'die Welt'. 10 Here, all the points go to the wholist.
The atomist cannot reply that the levels of importance reflected in
Wittgenstein's numbering are not semantic or metaphysical but merely
thematic. To offer such a reply is already to have swallowed a good
dose of atomistic interpretation. On a wholistic view, Wittgenstein does
not believe in (the usual sort of) explanatory semantics. Nor does he
believe in a metaphysics - or at least a metaphysics which extends
beyond the thematic. For the wholistic reader, there is nothing to being
a Tractarian object beyond being a theme, an object of conversation.
There is no latching on to a difference between order of explanation
and order of exposition, since Tractatus, on the preferred reading, does
not purport to explain anything.
5.2. Gegenstand, (Ding, Sache)
The very casualness of Wittgenstein's usage discourages the atomist
foundational pretensions of objects in Tractatus. Consider 2.01:
A circumstance is a linkage of objects (items, things).l1
The parenthesis (and a similar one at 4.1272) foster the idea that
the author meant to make no substantial distinction among the terms
translated as 'object' (Gegenstand), 'items' (Sachen) and 'things'
(Dingen). One impression is that he wished to use them inter-
changeably. This is the impression the Hintikkas endorsed when they
wrote:
Instead of 'objects' (Gegenstlinde), Wittgenstein sometimes says 'things' (Dinge), cf.,
e.g., 1.01. Nothing seems to hang on this terminological variation, however. (Hintikka
& Hintikka 1986, p. 30)
By my lights, this is not mere "terminological variation" and something
does indeed hang on this fact. Wittgenstein was a philosopher of extra-
ordinary care in published writing. He chose terminology scrupulously;
he reworked it endlessly. It would be natural to think that Wittgenstein
handled fundamental terminology with extreme care. In this regard,
there is great difference between the supposed "terminological varia-
tion" with respect to 'Gegenstand' and Wittgenstein's treatment of
terms for situations and their cognates. He established and respected
the distinctions among Sachverhalte (states-of-things), Tatsachen (facts)
and Sachlagen (situations) (which, incidentally, play no crucial role for
the Hintikkas). Each term has its special position in Tractatus, one it
retains throughout. There is no impression that these terms are every-
where interchangeable.
If we look to the German language, we fail to see 'Gegenstand',
'Ding' and 'Sache' comprising a range of terminological variation.
Wittgenstein's supposed nonchalance over their uses looks all the more
extraordinary, since they are, in no sense, simple "variants" of each
other. First, surrounding 'Gegenstand' there is an air of educated re-
finement, an air lacking in the matter of a (mere) "Sache". As in
English, philosophers are said to seek an inventory for the category of
physikalische Gegenstande - "physical objects" but worry much less
about Sachen - things or, better, "its". Second, 'Gegenstand', even at
its most colloquial, as in the question 'Siehst du den Gegenstand auf
dem Tisch?' - "Do you see the object on the table?" - carries a certain
specificity which 'Sache' and 'Ding' do not. It makes sense to ask
"Welche Sachen hast du heute erledigt?" ("What things have you ac-
complished today?"). You may hear in reply "Zwei Sachen. Ich habe
meinen Vortrag gescrieben und dann ging ich zur Post" ("Two things.
I wrote my lecture and then I went to the post office"). Therefore,
'Sache' is not ordinarily replaceable by 'Gegenstand'; no circumstances
would license the reply 'Zwei Gegenstande' to the previous question.
'Sache' - as with 'Ding' in some of its uses - is wondrously unspecific.
Its scope is tremendous, covering not merely particulars but also activi-
ties and events. The atomist ought to treat objects as comprising a
category distinguished by the intrinsic properties of its members. How-
ever, if one treats 'Gegenstande', 'Sachen' and 'Dinge' as marking such
categories, then the three categories so marked would be wholly dis-
tinct.
Moreover, Wittgenstein's supposed nonchalance is tough to square
with the specificity of the epistemological tenet of atomism, that Tract-
arian objects are the bearers of Russellian proper names. Wittgenstein
would be guilty of criminal nonchalance were he to have encouraged -
not just once, but on two occasions - the glossing of the supposedly
crucial (and quite refined) notion of 'object of acquaintance' as a mere
(and quite unrefined) Sache. The occasions in question are truly and
equally important; the first is the initial introduction of Gegenstiinde,
the second is part of the page-long explanation 'of the important idea
of formal concept.
What seem to be the intrinsic differences between Gegenstand and
Sache offer no difficulty to the wholist. For the wholist, nothing is
intrinsically an object; to treat something as an object is to adopt a
certain logical perspective toward it, a perspective that includes repre-
sentation as a certain sort of sign, a name. To put it crudely, to be a
Tractarian object is to be a topic, a topic of conversation. In itself,
nothing is a conversational topic per se. Anything becomes so describ-
able, as topic or as instance, when it plays a functional role with respect
to talk. So, there is nothing on earth - short of language itself - which
"picks something out" as a topic of conversation. This is very much a
suggestion of Wittgenstein's terminology. In German, one can describe
a topic as an object, as in den Gegenstand des Gespriichs. This is not
the "object" or point of the conversation, but instead, its topic, what
one is conversing about. In much the same vein, one can speak of den
Gegenstand des Satzes, of the (grammatical) object of a sentence. This
is a use which would come naturally to the logician Wittgenstein just
as it did to Frege in his" Begriff und Gegenstand".
Either of these uses has better chance of a match with the words
'Ding' and 'Sache'. Although 'Gegenstand', as topic of conversation, is
not interchangeable with 'Sache' and 'Ding', it shares with them a vast.
generality of application. Anything which one can denote with a noun
phrase, anything that one can talk about is a Sache. But it is also ein
Gegenstand des Gespriichs, a topic; this holds true even be it an event,
an activity, a nation-state or a planetary system. Now, on this treatment
of 'Gegenstand', what is it to say that "the world is constructed entirely
out of objects?" Read it one way and it comes out a banality: there
isn't anything we cannot talk about. But it can be read more deeply;
it may say that world is only a construction from out of c{j)nversation
and that the conversation is ultimate: it is only within the form of
speech that we can conceive of reality as having "units".
5.3. Atomism in Tractatus
Those passages which do seem to lend credence to the atomistic reading
need to be reconstrued. Once that is done, those same passages serve
not as support for atomism but as arguments against it. One might go
so far as to deny that atomism remains as a reasonable alternative
reading of Tractatus.
Consider Tractatus 1.2. In the Pears and McGuinness translation, we
find what strikes one as a clear statement of logical atomism - at least
as regards facts. It is:
The world divides into facts.
Max Black reads in this an emphatic expression of atomism. For this
passage, he writes that, in Wittgenstein, "[t]he world is a mosaic of
facts" (Black 1964, p. 37). Mosaics, like the atomists' world, are con-
structed out of their parts. They can and do fall apart into their constitu-
ent tiles and they can, in principle, be rebuilt out of them. To an extent,
this is because tesserae are real separable parts of mosaics. They can
exist independently, outside the mosaic. To reassemble the mosaic from
its parts is a task for our intelligence; we must use our wits. So, the
mosaic image is one the who list must refuse. Wittgenstein's world
does not come apart into independent items out of which it can be
reassembled. Nor can thought be the means to reassembly. Thought is
just one linguistic aspect of the whole world. And nothing less.
In the German version of the same passage, the mosaic disappears.
The verb translated by Pears and McGuinness as 'divides' is 'zerfallen',
a word whose logic is inseparable from wholism. In everyday use,
'zerfallen' means decomposition and normally denotes a falling apart
with destructive finality. And, short of godly or Frankensteinian inter-
vention, that which decomposes is not reassembled out of its parts.
When 'zerfallen' appears in Part II, Section ii, of Philosophical Investi-
gations (Wittgenstein 1953, p. 175), Elizabeth Anscombe translates it
as 'disintegration'. Colloquially, 'zerfallen' applies to the process which
bodies undergo in the grave; we find "Der Mensch zerfiillt in Staub und
Asche" or "humans fall apart into dust and ashes". The elements into
which something zerfiillt are not, in general, those out of which it can
be reassembled, even in thought. They need not be what we would call
'ordinary parts' of the thing. Nor need they be units which we would
choose in explaining the natural behavior of the original thing; we
would not come to an understanding of the nature and function of
human bodies by understanding, first, ashes and dust and, then, prin-
ciples by which these might be combined to form bodies. There is no
solace for the atomist's "construction" in the idea of disintegration.
(,ZerJallen' is also an echo in Wittgenstein of analysis in Frege, an
analysis which is, as we shall see, disintegrative.)
Passages which look to give wholism trouble include T 2.021. In
Pears and McGuinness, it is:
Objects make up the substance of the world.
What clearer expression of atomism could be desired? It seems to say
that, in its heart, the world is made up of or built out of objects. Black's
commentary echoes the natural atomistic reading; he wrote as gloss on
2.021 that objects "are the materials of which atomic facts are con-
structed, the substance of the world" (Black 1964, p. 57).
Again, the German original dispels atomistic impressions. There we
find:
Die GegensHinde bilden die Substanz der Welt.
First, the statement neither says nor implies that objects are substances
or that the objects are the substances. In the German, the word 'Sub-
stanz' is not in the plural. Hence, the passage does not concern sub-
stances. Rather, it appears to concern a Spinozistic sort of "Sub-
stance" - an abstract and unitary feature of the world, if anything. And
here first appearances are not deceiving: use of the singular and the
association between Substanz and an irreducibly abstract feature of the
world carries through all of Wittgenstein's remarks on the topic. For
instance, Wittgenstein tells us outright, at 2.0231, that the world's
Substanz establishes only a form and not any material properties.
Now for the verb of the statement, 'bilden'. The tie with abstraction
which the noun introduces is carried further by the verb. Die Gegen-
stiinde bilden die Substanz der Welt: 'bilden' is an echo of the Spinozistic
associations of "Substance". Given the English rendering and the con-
structional harmonics of words such as 'make up' and 'construct',
'bilden' is not the verb you should expect. One should expect verbs
such as 'bauen' or 'konstruieren', verbs reproducing constructional over-
tones in the German. 'Bilden', unlike 'bauen', does not stand for a
concrete sort of agglomerative activity, the kind of "putting together"
which the words "make up" suggest. Instead, 'bilden' connotes a thor-
oughly abstract operation for which the English verb 'form' is a fair
rendering. 'Bilden' can, for example, refer to the establishment of
a company or to the passage from an active sentence to its passive
equivalent.
The point here is level of abstractness - 'bilden' is neither for house
building nor for applying cosmetics. Wittgenstein's use of 'bilden'
throughout Tractatus bears this out. Most often it refers to the forma-
tion of propositional signs generally or the formation of a particular
sign. For instance, it is applied to the formation of the sign of a negated
proposition out of that for the unnegated proposition. And it appears
crucially at 5.5151, where one sees:
Muss das Zeichen des negativen Satzes mit dem Zeichen des positiven gebildet werden?
"Need the sign of a negative proposition be formed (gebildet) with that
of the positive?" In short, 'bilden' and its cognates replay the main
theme, which is the thoroughly abstract - and that means linguistic -
nature of what is mistakenly called "Tractarian ontology". To put it
another way, the world disintegrates into particulars, voiceless logical
surds; any abstraction, any generality, any indefiniteness, Wittgenstein
thinks to lie in language. Indeed, insofar as a world requires abstraction,
there is no world without language.
Verbs such as 'bauen' and (perhaps) 'konstruieren' naturally carry
with them the idea of simple and rather concrete construction out of
proper parts. Were the foundational idea of atomism truly at work in
the early sections of Tractatus, we might look to find these verbs there.
Interestingly, they make their first appearances only on the far side of
Hauptsatz 4. With one exception, they appear when we are described
as engaging in some relatively ordinary sort of construction, usually of
or via signs. Wittgenstein speaks of us as constructing languages and
also constructing gramophone records and musical scores. Only once
does he speak of the world as possessed of ein Bau, a manner of
construction or an architecture (5.5262). But this Bau seems well re-
moved from the concept of Gegenstand. In fact, b a u e ~ never seems
to stand in close proximity with that word for object. In summary, if
we ask what kind of expression of atomism is T 2.01, "Objects make
up the substance of the world", then we must reply that it is no
expression at all.
5.4. The Force of'Durchgreifen'
There is one more passage I would like to consider in the Pears and
McGuinness translation. Even though it does not seem to lend positive
support to atomism, it is at least comfortable with it. But this too is
appearance; upon inspection, it takes on the aspect of an argument
against atomism. The proposition in question is 3.42. In part, their
translation reads:
The force of a proposition reaches through the whole of logical space.
In the German, we see
Der Satz durchgreift den ganzen logischen Raum.
One notes immediately that the words 'the force' are wholly absent
from the German. It is the proposition itself, if anything, that durchgreift
the whole of logical space - and not its "force".
And what is this durchgreifen? First, it is not the separable form
'greifen . .. durch'. This latter denotes reaching through and grasping
hold of a relatively concrete sort. For instance,
Der Lehrer greift in der Klasse durch.
means that the teacher reaches through and exerts control over the
class. The connection here between what reaches through and what is
reached through is a contingent one; as we know, there is at mQst a
whimsical and sometime relation between the classes and the control
of their teachers. This form, the separable, is indeed the more common.
But what of the other, less common, inseparable forms, 'durchgreifen'
and 'durchziehen?' Here, proximity in word is suggestive of a stronger
conditioning among ideas. When X durchgreift Y, X surely reaches
through Y but, this time, not as a sometime interloper but as a unifying
principle. The relation is clearer with durchziehen. When I write
Profitdenken durchzieht die gesamte amerikanische Wirtschaft.
I mean not only that thought of profit runs through the American
economy as water through a pipe but also that it unites and brings the
economy a systemic unity. It is much the same with 'durchgreifen'. The
proposition which runs through the whole of logical space does so in a
unifying and systemic way.
The logical cosmology of atomism, wherein lies a logical space which
is a semantic patchwork, made up of elementary propositions which
are, in turn, built up from simple objects, is at ease with the suggestion
that it is the force of each proposition which reaches through logical
space. After all, we might be told, the atoms, the names, out of which
one Elementarsatz is produced appear simultaneously in an unlimited
number of others. In its atoms, the force of a single proposition might
be said to percolate through all the others. On the other hand, there
is "inferential force"; like the magnetic force of a single molecule, it is
a force that extends without a priori limits throughout physical space.
We might say that the inferential force of a single proposition runs
through the whole of logical space. After all, infinitely many distinct
propositions will follow from it.
These suggestions are readily dispatched. First, Wittgenstein thought
that our ordinary concepts of inferential force (and also mathematical
proof) are philosophically defective. In fact, he believed that a true
view of symbology would result in our giving up altogether on the
notion that logical inference is a tracing-out of relations between propo-
sitional contents. Second, if force is inferential, then the force of a
single elementary proposition cannot reach through the whole of logical
space, since any two distinct elementary propositions are logically inde-
pendent.
In truth, 3.42 strains the mosaic picture of atomism beyond the
breaking point. In that picture, Tractarian objects are tesserae and
elementary propositions are simple arrangements of a number of such
tesserae. But if the tesserae and their arrangements are independently
intelligible substantial units, it cannot come about that trivial compos-
ites of them "run through" the matrix of all the others. On the one
hand, the atomist insists that logical space, and the world in it, are built
up out of independently intelligible parts. On the other hand, 3.42
constrains him to admit that the "independently intelligible parts" can-
not be prized apart logically, seeing that each one runs through the
matrix composed of all the others. Atomism, as a doctrine of under-
standing, as a purported explanation of the way we come to appreciate
senses of individual propositions, requires that I can grasp propositions
one-at-a-time or, at most, several-at-a-time - just as independent tiles
in the mosaic ought to be assembled into independent groups, each
containing a reasonable number of tiles. But the idea that I might
bootstrap my way up semantically, launching semantic cognition from
atomic propositions and springing up through the truth functions, comes
to naught if propositions such as 'my dog has fleas' truly run through
all of logical space and hence are semantically attached to propositions
as seemingly divorced as 'butter can undergo Cambridge changes'.
I admit, the force of this word 'durchgreifen' is, at first, difficult to
swallow. One wonders how any single propositions can run through
and unite the whole logical world. I would like to think that this
difficulty comes because Wittgenstein's picture of the proposition, once
quite vivid, has been darkened by interpretative overlay. I think it is
common to conceive of Wittgenstein's propositions as logical shadow
puppets, that words of the sign act like cardboard cutouts and the
process of interpretation, the grasping of sense, is thought of as the
casting of a shadow from the cutouts onto the world. The "facts" which
the sign then reports are just the shadows that get drawn. Perhaps
remarks such as 3.12, "a proposition is a propositional sign in its
projective relationship to the world" , have encouraged the conventional
idea. A reappraisal of the concept of "projective relation" does a good
deal to dissolve the overlay of conventionality. Unlike the relation
of the cutout with its shadow, the projective relation in which the
propositional sign stands is an internal or structural one. First, that
means that it is a necessary or essential relation; it could not be the
same proposition and yet bear a different projectum. Second, an in-
ternal relation is only apparent from the structure of the relevant lan-
guage and, in the case of projective relations, the structure is entirely
intralinguistic. Wittgenstein's comments on the projective relations
which obtain between musical scores, their renditions and their re-
cordings (at 4.0141) remind us of this.
As Wittgenstein tells us, we think of the proposition as bipartite; it
is perceptible sign and projective relation. And, since the sign is hardly
a candidate for durchgreifen, it is the projective relation which runs
through and serves as a unitary principle in logical space. Again, propo-
sitional signs are not isolated objects of acts of interpretation. Rather,
they seem to have much the same ontological status in Tractatus as
complex predicates did in the destructive analysis of Frege. They are
not real parts of anything but rather Muster, patterns which we can
discern running through the whole of the system. The individual propo-
sition which is p is factored out of an infinite range of propositions
which include ,p, ~ q pl(qlr), that is, all the propositions which
we would normally consider its logical composites. In the appropriate
conceptual notation, the patterns which are propositions become appar-
ent in the fact that the sign for any individual proposition will itself be
a propositional variable. This, in turn, is nothing more than a stipu-
lation, a pointing-out, of the course of the pattern as it runs through
and unites p with its negations, double negations, implications, and
conjunctions and, ultimately, with every other proposition.
Context confirms this construal of the durchgreifende Siitze; in fact,
it does this twice. I quote the remainder of 3.42; it precedes the line
already quoted in the text.
As much as the proposition permits the determination of only one'spot in logical space,
so must the whole logical space be already given through it.
(Otherwise, would, through negation, logical sum, logical product, etc., new elements -
in coordination - always be brought into play.)
(The logical scaffolding surrounding the picture determines logical space.)
It is in its "logical scaffolding" that the proposition reaches through
and grasps all of logical space and, as the passage indicates, the welds
in this scaffolding are joins made by the logical operations. Once again,
the Siitz is a slot within a system, a sequence of coordinates which
manages to pick out a place only in relation to all the other possible
sequences, only in relation to a grasp of the system of coordinates as
a whole.
Brilliant and beautiful reconfirmation of this reading comes in an
entry from Wittgenstein's early Notebooks, dated 15 December 1914.
Wittgenstein refers to the durchgreifender Siitz in even more emphatic
terms after a brief consideration of the role of the sentence within a
system of logic. It is worth citing at length, since it also affords a plain
statement of what we might call Wittgenstein's logical transcendental
idealism: that, "before language", the world is unstructured and that,
"with language" a logical structure emerges. Moreover, the logical
structure is revealed in the conjectural linguistic domain of real signs.
It is obvious: we can introduce whatever we wish as the written sign of the ab-functions
(i.e., truth functions), the real sign will form itself automatically. And what properties
will be formed of themselves in this way?
The logical scaffolding around the picture (of the proposition) determines the logical
space.
The proposition must reach through and grasp (durchgreifen) the entire logical space.
In the world according to logical wholism, propositions are patterns
and objects are patterns; the mutuality in their natures is best conveyed
by an appropriate understanding of certain sorts of variables. At this
level, my claim is of a parity of sorts between objects and propositions;
and the parity is borne out in the text. The durchgreifen proposition,
3.42, is the propositional analogue to 2.0123, which is to be read in the
same spirit.
If I am acquainted with the object. then I am also acquainted with the collective possibili-
ties of its entering into states of things. (Every such possibility must lie in the nature of
the object.)
It also has a companion passage in Notebooks, one which appears on
page 83 of the English edition and page 178 of the German. Here I
present the German in translation:
Every thing determines the entire logical world, so to say, the entire logical spacc.
5.5. There are no Special Groups of Objects
Let it not be thought that logical wholism stands entirely upon textual
considerations which are negative. There are any number of passages
which are difficult to reconcile with one or another of the aspects
of the Hintikkas' vision of Wittgenstein's presumptive atomism. For
example, there are straightforward denials of the metaphysical facet of
logical atomism, the one on which the category of objects is thought
to embrace not only particulars but the more expansive denizens of
other logical types: properties, classes, functions and relations. I begin
with the earliest antiatomistic statements, those from Notebooks.
In the German edition of Notes on Logic, 1913, in the Zweites
Manuskript, we see:
Not only from a concern with [singular 1 things must logic prescind; it is as little permitted
to occupy itself with predicates and relations. (Wittgenstein 1984, p. 195)
The sentiment is repeated again on the same page:
In the same way, '(1/ in 'cPx' appears to be a substantive without being one .... One
reason which speaks against it(s being a substantive) is the generality of logic: logic
cannot deal in any special groups of objects.
If the atomists were right about Tractatus, then Wittgenstein counted
properties and relations among the "Tractarian" objects. But as proper-
ties and relations comprise special groups (in the original, "spezielle
Menge") if anything does, logic, according to Wittgenstein, cannot treat
of them. But it was Wittgenstein's study of logic which was supposed
to create a legitimate demand for such spezielle Menge. Therefore, the
metaphysical plank of atomism must be rejected.
The English version of the same, Notes on Logic, September 1913,
contains a parallel remark. Wittgenstein is here diagnosing an "error"
in Russell's approach:
It is easy to suppose that "individual", "particular", "complex", etc., are primitive ideas
of logic. RusselI, e.g., says "individual" and "matrix' are "primitive ideas". This error
is presumably to be explained by ... (N, p. 105)
In both passages, Wittgenstein is taking exception to Russell's efforts
in foundations of logic. First, Wittgenstein was accusing Russell of
reversing the right philosophical order of things. For Wittgenstein,
Logik (with a capital 'L') was first philosophy, not only in terms of
temporal or intellectual priority, but also in terms of what is philosoph-
ically definitive. Within the confines of an independent Logic were all
the problems of philosophy to be resolved. And that independence was
a freedom from a prior metaphysics such as Russell's theory of types
or Frege's conceptual hierarchy. To Wittgenstein, Russell sought always
to turn this picture around and make logic wait upon the verdicts of
Russell's neo-Cartesian metaphysics. Second, Russell certainly wanted
to "turn back the clock" as Wittgenstein saw it: the traditional meta-
physical concepts of particular, universal, property, object, function,
proposition, and symbol were absolutely defunct. Little trace of them
can remain in a logical investigation. Russell believed that a version of
traditional metaphysics, carried on in traditional terms, could proceed
under the imprimatur of his mathematical logic. Wittgenstein believed
that the only imprimatur which logic offers is interdiction.
5.6. Expression of Idealism
Tractatus 4.023 is a straightforward denial of the principal plank of
atomism, that the world is, independently of actual or potential ex-
pressions of thought, a construction out of real atoms. Like atoms as
envisioned in the time of Newton, these are antecedent to any speech,
they are fit together in one way or another - that is how actuality is
determined. Then, once we do get around to enunciating propositions,
they either accord or fail to do so with a pre determinate logical cosmos,
a great confabulation of atoms; concordance yields truth and dis-
cordance falsehood or, in the worst case, meaninglessness. But this is
far from the way in which Wittgenstein describes it here:
Through the proposition must reality be fixed as to "yes" or "no".
Contrary to the doctrine of logical atomism, Wittgenstein seems here
to be making the factual pleasure of the world wait on the prospect of
the sentence for its determination. It is through the proposition that the
determination occurs. Hence, it is only by creating signs that symbols
are created and symbols, in turn, are the relations we demarcate among
the signs of a system. It is only through the mechanism of symbols that
reality is determined one way or another.
6. ANALYSIS IN TRACTATUS
Analysis in Tractatus is not Russellian. There is plenty of evidence to
suggest that Wittgenstein did not endorse Russell's theory of descrip-
tions. This conclusion stands in opposition to the Hintikkas' contention,
according to which Russellian analysis is a linchpin in the formal aspect
of atomism, that is, in the technical means for carrying out the logic of
objects. Analysis for them forges the essential connection between an
arbitrary sign and the primitive sign and, hence, brings logic and lan-
guage in general into touch with the all-important objects. Ultimately,
for the Hintikkas, Russell's analysis points us out of language alto-
gether; it points us toward the objects. Analysis in Tractatus, however,
aids and abets the program of linguistic idealism: we are never pointed
out of the language. (In the regard, Wittgenstein's analysis is compa-
rable to one of Frege's, a decompositional analysis which uncovers the
formal patterns we need to note in order to assess inference.) To be
seen as language in Tractatus is to appear to us as part of a logically-
enabling structure or pattern. The boundaries of such a pattern corre-
spond to the linguistic area covered by the real or proper sign of which
the mark (as sign) is just one manifestation. The tracing of such a
pattern, or the demarcation of the pattern, is the job of one of the
processes we may call "analysis" in Tractatus.
It has been taken as given that Wittgenstein's analysis is some nebu-
lous version of Russell's. There are, however, a number of textual and
doctrinal indications that this is not so. First, we have Wittgenstein's
explicit endorsement of Frege's strictures on the adequacy of attempted
definitions. This appears in the parenthesis of T 5.451:
In short, what Frege has said concerning the introduction of signs via definitions (in
Grundgesetze) also applies, mutatis mutandis, in the same way to the introduction of
primitive signs.
Wittgenstein is endorsing material that includes Grundgesetze (Volume
II, Section 66), wherein Frege is explicitly refusing legitimacy to contex-
tual definition. According to the Geach and Black translation, Frege
wrote:
So we may not define a symbol or word by defining an expression in which it occurs,
whose remaining terms are known .... Rather, the definition must have the character
of an equation which is solved for the unknown, and on the other side of which nothing
unknown occurs any longer. (Frege 1977, pp. 170-71) ,
Professor Dummett characterizes Frege's attitude toward contextual
definition in his The Interpretation of Frege's Philosophy as a form of
"hostility" toward them (Dummett 1981, p. 36).
Passage T 3.3442 may be construed as a reassertion of Frege's senti-
ments:
The sign of a complex does not dissolve via analysis arbitrarily, so that, somehow, it
would have a distinct dissolution in every sentential nexus.
It is the very character of contextual definition that the resolution of
the definiens depends upon context and, therefore, differs according to
the sentential nexus in which it appears. Needless to say, replacement
of definite descriptions is the very soul of Russell's analysis and this is
a matter entirely of contextual definition.
Second, Wittgenstein's ideas on the nature of a proper Begriffsschrift
allow him to skirt the semantical pitfalls to which Russell would have
referred in explaining the need for a Russellian analysis, one set atop
his theory of descriptions. First, Wittgenstein, unlike Russell, meets
no special difficulty in accounting for the informativeness of identity
statements; he avoids Frege's problem of the morning star and the
evening star. This is because, for Wittgenstein, there are no identity
statements at all. According to T S.S3ff, what we think of as statements
of identity are not true statements but philosophically dangerous and
logically superfluous commentary on the sorts of signs which we are
using. As Wittgenstein wrote at T 4.242,
Expressions of the form "a = b" are, therefore, merely aids to representation; they assert
nothing about the significance of the signs 'a' and 'b'.
Naturally, without identity statements, there is no worry over identity
substitutions into belief contexts, the sort of problem exemplified by
Russell's 'Scott is the author of Waverly'. Moreover, there is enough
in Wittgenstein's cursory discussion of the logical form of belief state-
ments (T 5.542) to show that, even if he had allowed significant identi-
ties, the use of explicit quotation in the analysans for belief blocks
possible substitutions.
Third, Wittgenstein seems to have banished the problem of empty
names - the issue of the "the golden mountain" which Russell associ-
ated with Meinong - outright. As we said, the. wholist believes that
the meaningfulness of an expression sets no condition to which the
extralinguistic world would have to conform. As Wittgenstein wrote in
T 2.0211,
Had the world no substance, then whether a sentence had sense would depend upon
whether another sentence were true.
From context, it is clear that Wittgenstein intended this claim to feature
in a brief modus toll ens argument with conclusion "The world has
indeed substance". He also intended that the reader supply the missing
premise, "Whether a sentence has sense does not depend upon whether
another is true". It follows that, on a suitable view of language, 'The
golden mountain does not exist' does not require for its meaningfulness
that there actually be a golden mountain; according to Wittgenstein,
no sentence can lay down such a requirement.
There is one caution, however. It should not be thought that Wittgen-
stein's extraordinary attitude toward (apparent) statements of identity
and toward the independence of conditions of meaningfulness from
conditions of truth is only possible or is only made plausible (to the
extent that it is) by the prospect of Russellian definite descriptions
analysis. It might be thought that Wittgenstein intended his ideas on
identity, belief and empty names to apply only to those expressions
which have been suitably analysed, which already appear in fully per-
spicuous form in the notation of his own Begriffsschrift. Hence, it could
be (or so the putative reply would have it) that it is only by means of
the familiar process of the contextual replacement of definite descrip-
tions that we bring a statement from a linguistic state of nature into a
form to which Wittgenstein's strictures will apply. This may be the only
way we can get ourselves into the contortions requisite to make such
ordinary statements as
Mark Twain is Samuel Clemens
conform to Wittgenstein's seemingly procrustean views on identity.
Hence - or so it might be argued - Wittgenstein needed descriptions
analysis to ensure his view and, thereby, to replace statements such as
the above by statements of some other form, in particular, by series of
statements not containing names for Mark Twain but only "logically
proper names", ones for whom Wittgenstein's mysterious views on
identity and existence can be imagined to be a bit less mysterious.
A moment's thought shows that this reply is unsatisfactory. In order
to apply descriptions analysis, we assume that "Mark Twain is Samuel
Clemens" is to be construed as equivalent to and analyzable in terms
of the descriptive statement
The author of Pudd'nhead Wilson is the same as the author
of Innocents Abroad.
Now, let Px stand for 'x authored Pudd'nhead Wilson' and let Ix
stand for 'x authored Innocents Abroad'. From "Mark Twain is Samuel
Clemens", analysis would lead us to
3x(Px A Ix A 'VyPy v Iy) y = x.
Now, if we put this into conformity with Wittgenstein's suggestions on
identity - that identity of Bedeutungen be indicated by identity of
variable names - we obtain
3x(Px A Ix) A 'Vy -, (Py v Iy).
Here, the difference in the sign for the variable - 'y' as opposed to
'x' - indicates (what we would call) disjointness in the range.
Famously, Wittgenstein did not view 3x and 'Vx as logical signs but
as confused composites of generality and truth-functionality. When
properly understood, they serve, in association with the remainder of
the quantified statement, to stipulate a truth-function of a sequence of
propositions. Therefore, when we move to a suitable notation, Witt-
genstein would prefer to see the first conjunct of the above formula as
replaced by an infinitary disjunction, something akin to
ViPa Ala)
where the latter is a shorthand for an actual infinitary disjunction
(Pa A Ia) v (Pb A Ib) v (Pc A Ic) v ...
wherein 'a', 'b', 'c' et aI., comprise a suitable collection of names, the
extent of the collection indicated by the form of the 'x' in the original
expression.
Now, our point is clear. If you look, the move from names such as
'Mark Twain' has come full circle. On Wittgenstein's approach to the
quantifiers, there is no net gain, if we measure gain in terms of the
designata of names which are called for in expressing a given proposi-
tion. We may not require lexically distinct names, one for the author
of Pudd'nhead Wilson and another for the author of Innocents Abroad,
but we shall need some name or other which denotes Sam Clemens,
the human being who wrote both works.
The contrast with Russell's views ought to be plain. For Russell, a
need for ordinary names fades away into the work of the quantifiers. In
their turn, quantifiers have the job of naming properties of propositional
functions and make no claim about the individuals which are the desig-
nata of names. Russell says as much in the fifth of his 1917-1918 lectures
on logical atomism:
Existence is essentially a property of a propositional function .... Existence-propositions
do not say anything about the actual individual but only about the class or function.
(Russell 1985)
In other (Russellian) words, the proposition expressed by 'an author
of Innocents Abroad exists' contains no vestige of the historical individ-
ual Mark Twain. (Incidentally, Wittgenstein could hardly have accepted
such a characterization of the logic of "existence-propositions" on which
existence is essentially a property of propositional functions. Functions,
like objects, are formal concepts; hence, there are no property state-
ments asserting the existence of one or more of them.)
For Wittgenstein, the commitment to the name - and to its object -
is inescapable. To use the jargon of the Introduction to the second
edition of Principia Mathematica, we can say that Wittgenstein believed
that propositional functions only enter into propositions through their
values. Among other things, this means that, if we say that quantifiers
do indicate second-level concepts or properties of first-level proposi-
tional functions, then all we mean by that is that quantifiers indicate
patterns in (possibly infinitary) truth functions of elementary proposi-
tions, propositions whose signs are combinations of names. Notice,
however, that the inescapability of naming does not arise because an
appropriate theory of understanding takes its start from the relation of
denotation. On Wittgenstein's view, we do not need names because,
through them, we "attach language to the world". It is not because
names have to be correlated with items of a certain sort if we are to
have any expression of sense. It is simply because, in order to be signs,
even marks such as footprints need to be read. Wittgenstein takes
whatever agreement reading requires to be captured, in large part, in
our explanations of the Bedeutungen of names. As we saw earlier,
names are required for the very subsistence of sense.
These considerations on the relation in Tractatus between signs of
generality and names expose a third reason why Wittgenstein's concep-
tion of analysis differed from that of Russell. As we have seen, Witt-
genstein would allow nothing in an ontological way, nothing in terms
of metaphysical commitment, to be gained through the elimination of
names by replacement with some quantified concoction. If there is a
privileged class of "true" or "real" names, then they are not reached,
it would seem, by some mere iteration of the procedures which Russell
first set out in On Denoting. Some other form of analysis is required
to make the connections between what is normally called 'a name' and
whatever items form the conjectural collection of "real names".
As suggested by the remarks at T 3.23 and following, analysis in
Wittgenstein is a unidirectional process of resolution (and here you
may, with considerable propriety, think of the resolving of vectors). The
process is triggered by the recognition of indeterminacy or generality in
a sign, a generality inherent in the equivalence which Wittgenstein
noted between fact and complex. As Wittgenstein indicated in his
Notebooks for 28 May 1915, he thought of the terms 'complex' and
'fact' as coextensive (N, p. 52). Consider the fact (or complex) which
consists in the layout of a desktop with pen lying to the left of pencil.
Faced with his fact (or complex), which we might call 'pen-Ieft-of-
pencil', we may take a propositional perspective on it, we may read it
as having a sense - we may read it as a proposition or as comparable
with a proposition. This is to treat it as a fact, as the fact that pen is
to the left of pencil. When we think of it as Bedeutung, as something
comparable with a name, it is treated as a complex, as, roughly "pen-
left-of-pencil" as is indicated by the use of this complex name. Now,
when I predicate something of the complex, say "pen-left-of-pencil is
to the right of the blotter", there is an indeterminacy. In fact, it is just
the sort of indeterminacy which Wittgenstein examined so thoroughly
in the case of "the watch is lying on the table" (N, p. 69ff). It seems
perfectly clear that 'pen-Ieft-of-pencil is to the right of the blotter' is
true if the pen remains to the left of the pencil on the desktop and the
whole complex lies to the right of the entire blotter. Things are not so
clear, however, if, say, the pencil is to the right of the blotter but some
part of the pen is not. Perhaps the pen is partially atop the blotter. Is
it then true to say that the complex is to the right of the blotter?
According to Tractatus and Notebooks, such an indeterminacy is
merely apparent, an artifact of our profound ignorance about the me-
chanism of meaning - words are like a film on deep water. To think
that the logical world itself contains indeterminacy is to confuse, as the
denotational semanticists did, the character of the sign with the charac-
ter of its Bedeutung. First, whether we recognize it or not, the determi-
nacy of sense requires that a reading of 'pen-Ieft-of-pencil is to the right
of the blotter' associates that sentence somehow with a quite specific
fact, one containing none of the indeterminacy which lodged in the sign.
Second, an analysis of the sentence would eliminate the presumptive
generality or indeterminacy in favor of a range of propositions. In
this case, the process might result in the replacement of our original
sentence,
pen-Ieft-of-pencil is to the right of the blotter
by the sentences,
pen is to the right of the blotter
pencil is to the right of the blotter
pen is to the left of the pencil
the last of which results from changing our perspective upon the original
description of the complex and reading it as a proposition. I presume
that this is, in truncated and outline form, the sort of manuever which
Wittgenstein described at T 2.0201:
Every statement about complexes permits itself to be pulled apart [zerlegenJ into a
statement about their constituent parts and into propositions which describe the com-
plexes completely.
Third, the details of the analysis do not become available as the result
of any investigation in pure logic; the precise nature of the situation to
which 'pen-Ieft-of-pencil is to the right of the blotter' has been attached
cannot be discovered by a transcendental investigation of the sign in
general. Rather, Wittgenstein thinks that an investigation into psychol-
ogy would be required.
One final point: as Wittgenstein makes plain at 3.24, we can also
conceive of the process of analysis as the clearing-away of generality.
Since a propositional variable is nothing more than a specification of a
range of propositions in virtue of their structural similarities, we can
think of the original sentence, 'pen-Ieft-of-pencil is to the right of the
blotter', as a propositional variable for the sentences into which it
is resolved. Viewed in this light then, Wittgensteinian analysis is an
elimination of generality in terms of specificity.
In summary, the number of points at which Wittgenstein's analysis
differs from Russell's ought to be emphasized. First, Wittgenstein's
analysis is part of the empirical investigation of the sign; it is a process
of displaying a sign's inner nature. Analysis reveals the way in which
signs are actually formed in the mechanism which underlies reading.
As it exposes logical connections, analysis stays entirely within the
bounds of language. It never takes us to a place where language
"points" out of itself, perhaps into a putative intentional gap between
word and object. 12 All we can find before us are more words. Second,
analysis proceeds not by the introduction of ever greater generality, as
in Russell, but via its complete elimination. Third, as mentioned, analy-
sis lies not in the province of the logician, but in that of the psychologist.
Finally, as the word 'zerlegen' ['pull apart'] suggests, Wittgenstein's
analysis is destructive in ways that Russell's is not. For one thing,
analysis is not a cognitive, semantical recipe for the grasp of a sense,
the indication of a process of thought by which someone (perhaps even
implicitly) would come to an understanding of the analyzed expression.
In Tractatus, there are no such recipes and no need for them: what we
need to have before us in order to stand among signs as a reader is not
constructed from below, from a semantics of simples. At T 5.526,
Wittgenstein put it quite plainly: "One can describe the world com-
pletely through fully generalized propositions, that means, therefore,
without correlating in advance any name whatsoever with a specific
object". There is no need for a "bottom up" construction.
The distance we have marked out here between Wittgenstein's analy-
sis and Russell's seems the same conceptual distance as that marked
out by Wittgenstein himself in a passage of Notebooks:
Let us assume that every spatial object consists of endlessly many points, then it is clear
that I cannot introduce all these by name if I speak of such an object. Therefore, here
would be a case in which I cannot arrive at a complete analysis in the old sense at all;
and perhaps this is just the normal case. (N, p. 62)
Provided that the "analysis in the old sense" is Russell's, Wittgenstein
is pointing out that it could hardly be a verdict of logic, mathematics
or an aprioristic conceptual investigation that every propositions admits
of a complete Russellian analysis. The claim that every proposition
admits of a complete Russellian analysis implies, or so Wittgenstein
believed, that every object consists of at most finitely many simples.
The latter could hardly be a verdict of a purely conceptual enquiry.
Wittgenstein's response to this character of the "old" analysis is re-
corded next:
This is indeed clear: that the propositions which humanity exclusively uses will have a
sense just as they stand and do not wait on a future analysis in order to acquire a sense.
So, our grasp of a sense is not constructed from cognitive units such as
the objects of acquaintance which were to be revealed in Russellian
analysis. The standing of understanding is always "from above", from
a perspective on the whole world of signs. Nothing less will ever do.
ACKNOWLEDGMENTS
I am grateful to the members of the Logic, Language and Science
Colloquium at the Ohio State University, especially Stewart Shapiro,
for their comments on an earlier version of this article. Also, my
thanks go to the organizers of the XIVth International Wittgenstein
Symposium, Wittgenstein Centenary Celebration, especially Professor
Rudolf Haller. Ellen Klein made it possible for me to present my ideas,
first at Rollins College, and later at Washington College, where my
claims were subjected to thorough examination by the members of her
Wittgenstein seminar. The article would never have come to exist but
for the persistent and insightful criticisms of members of the onetime
Florida State Wittgenstein group, especially C. J. B. Macmillan, Eman-
uel Shargel and Anton Mikel. Finally, Jaakko Hintikka and Luise Prior
McCarty deserve my special thanks for many stimulating discussions
on Wittgenstein.
NOTES
* A highly condensed version of the ideas here presented will appear in the article
Hintikka's Tractatus in Proceedings of the XIVth International Wittgenstein Symposium,
Wittgenstein Centenary Celebration, 1991.
1 Throughout the sequel, 'T', followed by a numeral, will refer to a passage in Wittgen-
stein's Tractatus. References to Blackwell's edition of Notebooks, 1914-1916, will be by
'N' followed by a page number. Unless otherwise noted, translations from the German
are my own.
2 My description of logical atomism is not intended to be historically accurate restatement
of Russell's one-time view; principally, it is meant to fit the exposition of Investigating
Wittgenstein (Hintikka & Hintikka 1986).
3 After seeing the world according to logical atomism, one is moved to exclaim "Gee,
it's a good thing for us that our language fits onto the world!". Atomism does make it
seem that interpreted language and the world into which it is interpreted are wholly
independent fields and that their perfect logical liaison is fortuitous;. That the language-
world connection is incidental - indeed, that there is need for such connection at all - is
one of the fundamental flaws in the atomistic interpretation. In Tractatus, there is an
internal relation between sign and significatum; name could not be what it is without
Bedeutung.
4 Since co-authoring Investigating Wittgenstein, Jaakko Hintikka has come to doubt that
Tarski's theory of truth is truly a recreation of Wittgenstein's picture theory.
5 Here a "system" ought to be conceived as a network or space - on the model of a
vector space - that is, a structured arrangement of articulated nodes each of which has
a precise, formally determined, position relative to the others. I should not be read as
relying upon the contemporary notion of a logical or formal system which is related
symbiotically to the syntax/semantics distinction. A Wittgensteinian system of language
would encompass far more than any conventional formal system. It will include a sign
for 'the watch is lying on the table' which is the concrete circumstance in which I illustrate
the words 'the watch is lying on the table' by pointing out a particular watch x lying on
a particular table y. Claims such as '''the watch' refers to x" are, then, interpreted as
features of such situational signs.
6 If the world in Wittgenstein is a perspective on signs, then there is, in truth, no world
without language. It is here an expository device or the first rung on a ladder we ought
to kick away once we have climbed it.
7 Far be it from me to suggest that particular languages such as Chinese or Frege's
Begriffsschrift are extensionally distinct from the language, the system of signs, of which
Wittgenstein wrote in Tractatus. We might say that they are the same things - just
addressed with speech in different ways. Similarly, Bernard Ortcutt the college president
and Bernie the spymaster are the same man, but get addressed in quite different ways.
This distinction is to be compared with the one drawn by Wittgenstein (and redrawn by
us in a section to come) between complex and proposition.
8 To say that each proposition resolves into elementary propositions or, equivalently, to
say that each Tatsache resolves or divides naturally into Sachverhalten is, first, not to say
the reverse, that, for example, propositions are built up or composed out of elementary
propositions. Second, it is not to say that we either do or must resolve them ourselves
or that, in our "cognitive processing" of propositions, we carry out such a manuever.
Rather, it is Wittgenstein's contention that propositions resolve themselves. Also, it does
not follow from the claim that every proposition resolves into elementary propositions
that every proposition contains only finitely many elementary constituents. It is only
required that the resolution process be well-founded.
9 Those familiar with Zermelo-Franekel set theory and its close relatives will notice the
resemblance between this reconstruction of Wittgenstein's views on well-formedness and
attempts to justify the axiom of foundation. It is unlikely that the resemblance is acciden-
tal - at least from the mathematical standpoint. (I am not asserting that there was
historical coincidence - perhaps a chance meeting between Wittgenstein and Mirimanoff
on the Russian front.) In my mind, they are linked through Wittgenstein's remarks on
mathematics at T 6.2ff. And the connection lies in this: the category of Wittgensteinian
real signs contains an isomorphic embedding of Cantor's second number class.
10 It is noteworthy that the nature of "the world" at large and in itself never seemed to
be a first concern of either Frege or Russell. The world was nothing more, in the case
of the author of Der Gedanke, than the sum of the drei Reiche. The world was, however,
of first concern to nineteenth century logicians of the idealist persuasion.
11 The word translated as 'linkage' here is 'Verbindung'. Wittgenstein has chosen to use
it rather than 'Zusammenhang' or 'Verkettung', words which he employs on other oc-
casions to describe the inner structures of facts. The choice is especially interesting since
'Verbindung' connotes a connection or linkage which is dependent for its existence upon
our activities, in particular, some activity of bond creation or ratification. By contrast,
this connotation is absent from a word such as 'Zusammenhang'. "Eine Verbindung" is
a linking or binding together which has to be established or set up before it can exist.
The marital bond is a sample "Verbindung" which comes readily to mind.
12 Should the reader be in any doubt about this, it might allay skepticism to consult T
3.263, where Wittgenstein makes it clear that, even if analysis were to bring us to a
statement whose constituent signs were all primitive, we are not freed of the bonds of
speech. He wrote the following:
The Bedeutungen of primitive signs can be clarified through elucidations. Elucidations
are propositions which contain the primitive signs. Therefore, they can only be under-
stood if the Bedeutungen of the signs are already familiar.
He thereby refused the icon of the bare ostensive association of primitive sign with object.
One should note that it is not the possibility of ostensive definition which is under
discussion; the main issue is its bareness, the prospect that, at the point of ostension, we
leap the intentional gap between word and object. (Compare, in this regard, N, p. 70.)
REFERENCES
Anscombe, G. E. M.: 1959, An Introduction to Wittgenstein's Tractatus, 2nd ed., Harper
& Row, New York.
Black, M.: 1964, A Companion to Wittgenstein's Tractatus, Cornell University Press, New
York.
Dummett, M.: 1978, Preface to Truth and Other Enigmas, Harvard University Press,
Cambridge, Massachusetts.
Dummett, M.: 1981, The Interpretation of Frege's Philosophy, Harvard University Press,
Cambridge, Massachusetts.
Frege, G.: 1977, Translations from the Philosophical Writings, P. Geach and M. Black
(trans.), Cornell University Press, New York.
Frege, G.: 1979, Posthumous Writings, Basil Blackwell, Oxford, England.
G6de\, K.: 1983, 'Russell's Mathematical Logic', in P. Benacerraf and H. Putnam (eds.),
Philosophy of Mathematics. Selected Readings, 2nd edn., Cambridge University Press,
pp.447-85.
Hintikka, Merrill B. and Jaakko Hintikka: 1986, Investigating Wittgenstein, Basil
Blackwell, Oxford, England.
Kraus, Karl: 1987, Ober die Sprache, Suhrkamp, Frankfurt.
McGuinness, B. (ed.): 1979, Wittgenstein and the Vienna Circle. Conversations Recorded
by Friedrich Waismann, J. Schulte and B. McGuinness (trans.), Basil Blackwell, Ox-
ford, England.
McGuinness, B. (ed.): 1984, Wittgenstein und der Weiner Kreis. Gesprtiche, aufgezeichnet
von Friedrich Waismann, Ludwig Wittgenstein, Werkausgabe, Band 3, Suhrkamp.
Quine, W V. 0.: 1971, Set Theory and its Logic, rev. edn., Harvard University Press,
Cambridge, Massachusetts. '
Russell, B.: 1956, Logic and Knowledge. Essays 1901-1950, R. C. Marsh (ed.), London.
Russell, B.: 1961, Introduction Ludwig Wittgenstein Tractatus Logico-Philosophicus,
Routledge and Kegan Paul, London.
Russell, B.: 1985, The Philosophy of Logical Atomism, Open Court.
Russell, B. and A. N. Whitehead: 1967, Principia Mathematica to *56, Cambridge Univer-
sity Press.
Wittgenstein, L.: 1953, Philosophical Investigations, G. E. M. Anscombe (trans.), Mac-
millan, London and New York.
Wittgenstein, L.: 1961, Notebooks 1914-1916, G. H. von Wright and G. E. M. Anscombe
(ed.), Harper and Row, London and New York.
Wittgenstein, L.: 1961, Tractatus Logico-Philosophicus, D. F. Pears and B. F. McGuin-
ness (trans.), Routledge and Kegan Paul, London.
Wittgenstein, L.: 1982, Last Writings, G. H. von Wright and H. Nyman (eds.), University
of Chicago Press, Chicago, Illinois.
Wittgenstein, L.: 1984, 'Aufzeichnungen tiber Logik, 1913', in Wittgenstein: Werkausgabe
Band I, Suhrkamp.
STEVE GERRARD
WITTGENSTEIN'S PHILOSOPHIES OF
MATHEMATICS*
ABSTRACf. Wittgenstein's philosophy of mathematics has long been notorious. Part
of the problem is that it has not been recognized that Wittgenstein, in fact, had two chief
post-Tractatus conceptions of mathematics. I have labelled these the calculus conception
and the language-game conception. The calculus conception forms a distinct middle
period. The goal of my article is to provide a new framework for examining Wittgenstein's
philosophies of mathematics and the evolution of his career as a whole. I posit the
Hardyian Picture, modelled on the Augustinian Picture, to provide a structure for
Wittgenstein's work on the philosophy of mathematics. Wittgenstein's calculus period
has not been properly recognized, so I give a detailed account of the tenets of that stage
in Wittgenstein's career. Wittgenstein's notorious remarks on contradiction are the test
case for my theory of his transition. I show that the bizarreness of those remarks is largely
due to the calculus conception, but that Wittgenstein's later language-game account of
mathematics keeps the rejection of the Hardyian Picture while correcting the calculus
conception's mistakes.
1.
The philosophy of mathematics dominated Wittgenstein's philosophical
middle years - from the late 1920s when he returned to philosophy
until 1944 when he began revising material that eventually became Part
1 of the Philosophical Investigations. Wittgenstein's work on this subject
marks his transition from the Tractatus's world of logical form and
definiteness of sense to the Investigation's language-games and forms
of life.
Michael Dummett represents a standard view of this work when he
writes: "Wittgenstein's vision of mathematics cannot, I believe, be
sustained; it was a radically faulty vision".1 Since Dummett elsewhere
acknowledges an enormous debt to Wittgenstein, the standard view in
effect comes down to this: Wittgenstein made great contributions to
philosophy in the Tractatus and the Investigations, but in between he
worked for one and one-half decades on a faulty (and even perverse)
philosophy of mathematics. In other words, there is a black hole pre-
cisely in the center of Wittgenstein's career.2
The black hole theory involves looking at Wittgenstein's philosophy
of mathematics as an unchanging doctrine. Remarks composed in the
Synthese 87: 125-142, 1991.
126 STEVE GERRARD
early 1930s are treated in the same way as remarks from the early
1940s. This is looking at Wittgenstein's career as if there were no
transition: there is the earlier Wittgenstein of the Tractatus, the later
Wittgenstein of the Investigations, and in between only the unchanging
darkness of his philosophy of mathematics. This view, however, cannot
survive a careful examination of the texts.
Wittgenstein, in fact, had two chief post- Tractatus accounts of mathe-
matics. I have labelled these the calculus coraception and the lan-
guage-game conception. The calculus conception dominated Wittgen-
stein's thought from 1929 through the early 1930s, although in some
areas (such as contradiction) its influence lasted longer. In the middle
1930s, his views began to change to the language-game conception, and
by the early 1940s, the view of mathematical language as a nexus of
language-games had completely overturned the calculus view.
In the transitional (calculus) period Wittgenstein saw mathematics as
a closed, self-contained system. The rules (construed extremely nar-
rowly) alone determine meaning, and thus become the final and only
court of appeal. In the more mature (language-game) period, meaning
and truth can be accounted for only in the context of a practice, and
mathematics is examined by seeing what special role it plays in our
lives and its special relationship to other language-games.
But not everything changed; throughout all the stages of Wittgen-
stein's work on the philosophy of mathematics, he remained opposed
to and tried to undermine what he considered to be a misleading picture
of the nature of mathematics. According to this opposing picture, math-
ematics is somehow transcendental: a mathematkal proposition has
truth and meaning regardless of human rules or use. According to this
picture there is an underlying mathematical reality which is independent
of our mathematical practice and language and which adjudicates the
correctness of that practice and language. This plays a similar (negative)
role for Wittgenstein's philosophy of mathematics as does the Augustin-
ian Picture for his later philosophy of language. For reasons given in
the next section, I call this the "Hardyian Picture" after the mathemati-
cian G. H. Hardy.
The Hardyian Picture helps to give a structure to Wittgenstein's work
on the philosophy of mathematics and to unite seemingly disparate
discussions. Regardless of what else he was doing, Wittgenstein always
kept this picture in mind and tried to distance himself from its temp-
tations and confusions.
WITTGENSTEIN'S PHILOSOPHIES OF MATHEMATICS 127
The goal of this article is to challenge the black hole view, giving a
new framework for examining Wittgenstein's philosophies of mathemat-
ics and the evolution of his career as a whole. The idea is not to claim
that Wittgenstein never made a m i s t ~ k - the calculus conception is
certainly flawed - rather, it is to show the motivation behind the mis-
takes, and to demonstrate that, at least in the paradigm case of contra-
diction, these mistakes were made during the transitional Wittgenstein
period and were corrected during the later Wittgenstein period.
The first step is to provide a structure and motivation for Wittgen-
stein's work on the philosophy of mathematics. This is what the Hardy-
ian Picture does, and is the subject of Section 2. Wittgenstein's calculus
period has not been properly recognized, so Section 3 gives a detailed
account of the tenets of that stage in Wittgenstein's career. Wittgen-
stein's notorious remarks on contradiction may be taken as repre-
sentative of the bizarreness of the calculus period and are often given
as the main evidence for the black hole view. Section 4, concerning
contradiction, is the test case for my theory of Wittgenstein's transition.
There it will be shown that the bizarreness of those remarks is largely
due to the calculus conception, but that Wittgenstein's later language-
game account of mathematics keeps the rejection of the Hardyian
Picture while correcting the calculus conception's mistakes.
2.
A notable feature of Wittgenstein's post- Tractatus work is that he seems
to spend far more time attacking opposing views than he does asserting
his own. Thus, in imposing a structure on Wittgenstein's philosophies
of mathematics the natural place to look is to the view Wittgenstein is
rejecting. Most commentators call that view "Platonism", but this is not
a happy choice. What is commonly labelled as Platonism in mathematics
includes many trivialities - even truisms - and attacking these misses
the force of Wittgenstein's real objections.
3
The Philosophical Investigations begins with a quotation from St.
Augustine describing how he learned to speak. Wittgenstein does not
call this a theory, but a picture: "These words, it seems to me, give us
a particular picture of the essence of human language" (PI: 1). A great
deal of the Investigations concentrates on trying to undermine this
picture. In addition, this picture ties together seemingly isolated dis-
cussions. On closer look, for example, we can see that what in pre-
128 STEVE GERRARD
Kripke days was considered to be the Private Language Argument,
especially the S-game of a sensations diary, is really the Augustinian
Picture all over again. The diary keeper is naming sensations the way
the stylized Augustine named physical objects: without any background
context or language.
Wittgenstein's contemporary, the Cambridge mathematician G. H.
Hardy, plays a similar role for Wittgenstein's philosophies of mathemat-
ics as Augustine does for the Investigations, and ;it is helpful to posit a
Hardyian Picture on the model of the Augustinian one.
4
Wittgenstein spoke of "a false idea of the role which mathematical
and logical propositions play", and then identified this view with Hardy:
Consider Professor Hardy's article ('Mathematical Proof') and his remark that "to mathe-
matical propositions there corresponds - in some sense, however sophisticated - a reali-
ty." (The fact that he said it does not matter; what is important is that it is a thing which
lots of people would like to say.) (LFM [1939]: 239)
The last remark in parentheses is important. Just as it has been
pointed out that St. Augustine was more sophisticated than the Augus-
tinian Picture, it is undoubtedly true that G. H. Hardy was more
sophisticated than the Hardyian Picture. What is important in each case
is that in the words of Augustine and Hardy Wittgenstein found a
model for the most serious temptations that he believed philosophers
could fall prey to. To Wittgenstein, the most important way a philo-
sopher of language can go wrong is by being misled by the Augustinian
Picture; while the most important way a philosopher of mathematics
can go wrong is by being misled by the Hardyian Picture.
What is so misleading about the Hardyian Picture? Wittgenstein is
not objecting to the idea of a mathematical reality, as naive objections
to Platonism do. What Wittgenstein is objecting to is a conception of
mathematical reality that is independent of our practice and language
and that adjudicates the correctness of that practice and language. The
faulty conception is of a mathematical reality that is capable of over-
ruling how we actually do mathematics.
If Wittgenstein had written a book on the philosophy of mathematics
with a structure similar to the Investigations, that book might have
begun with the following quotation from Hardy's A Mathematician's
Apology:
I believe that mathematical reality lies outside us, that our function is to discover or
WITTG ENSTEIN' S PHILOSOPHIES OF MATHEMATICS 129
observe it, and that the theorems which we prove, and which we describe grandiloquently
as our "creations", are simply our notes of our observations.
s
Wittgenstein might have replied to this as he did in a meeting with
Rush Rhees in 1938:
Certainly we may say that in our arithmetic it is a property of 3 that 3 + 3 = 6; and so
that this proposition states a certain property of 3 in our arithmetic. But this would be
a misleading way of speaking if it suggested that we are dealing ,here with a property
which 3 has independently of any proposition in arithmetic; or that it is because of a
certain property which 3 has that we say in arithmetic that 3 + 3 = 6.
6
Wittgenstein's point goes beyond the context principle. What can be
misleading in Hardy's words is that it makes it look as if there were
two criteria for whether 3 + 3 = 6: (1) the computation or proof; and
(2) the further checking of whether the results of this computation or
proof match something else ("mathematical reality"). The Hardyian
Picture is not only that there are two criteria, but that the second is
the real one; proofs are only psychological devices to get us to see
mathematical reality more clearly. The real warrant for the truth of our
mathematical propositions comes not from proof, but from this reality.
In fact, G. H. Hardy himself, deliberately exaggerating, comes close
to saying this very thing. After drawing an analogy between a mathema-
tician and "a man who gazes at a distant range of mountains and notes
down his observations", Hardy writes:
The analogy is a rough one, but I am sure that it is not altogether misleading. If we were
to push it to its extreme we should be led to a rather paradoxical conclusion; that there
is, strictly, no such thing as mathematical proof; that we can, in the last analysis, do
nothing but point; that proofs are what Littlewood and I call gas, rhetorical flourishes
designed to affect psychology, pictures on the board in the lecture, devices to stimulate
the imagination of pupils.
7
Thus, just as in the Augustinian Picture the sentence "that is a desk"
points to a desk, in the Hardyian Picture a proof points to a mathema-
tical reality. If that object is not a desk, then the sentence is wrong;
if mathematical reality is different, then the proof is wrong. But to
Wittgenstein this account is no more satisfactory for mathematics than
it is for our ordinary talk of desks and tables.
The issue here is whether sense can be made of anything overruling
our mathematical language and practice, our proofs and computations.
The picture is that contrary to our computation, it could be false that
3 + 3 = 6, or that contrary to our proof it could be false that there is
130 STEVE GERRARD
not a largest prime - false not in the sense of a slip of the pen or an
arithmetical error or the use of an invalid inference, but false in the
sense that the computation or proof has failed to accurately represent
mathematical reality.
This might be considered a skepticism about the truth of mathematics
on the analogy of the familiar skepticism about the truth of our ordinary
talk of objects.
8
We believe there is a desk here, but physical reality is
different (perhaps there is only a Cartesian demon making us believe
there is a desk); we believe that 3 + 3 = 6, but mathematical reality is
different (perhaps a demon has misled us about our mathematical
rules).
But such skepticism about the truth of our mathematical propositions
would consistently lead to skepticism about the meaning of our mathe-
matical propositions. It is part of the meaning of our mathematical
terms '3' and '+' that 3 + 3 = 6. We would not have good reason to
translate foreign terms as "3", "6', "+", and "=" unless such a transla-
tion kept the truth of the equation. But how could a demon mislead
us about the meaning of our terms?
Let us review the situation. The Hardyian Picture holds that the real
criterion for the truth of our mathematical propositions is the nature
of a mathematical reality independent from our practice and language,
and that our proofs only point to that reality. (Notice that unlike some
traditional attacks on Platonism, the issue here is not whether we can
perceive this reality, but whether this reality can serve as a warrant for
our proofs.) But this implies that the meaning of our mathematical
terms is also independent of our practice and language. But, contra the
Hardyian Picture, our terms and proofs are only meaningful (in other
words, are only terms and proofs as opposed to ink marks or grunts)
in the context of our language.
Much of the attack on the Augustinian Picture in the Investigations
seeks to establish the importance of the point that expressions are
expressions only in the context of a language. In this case there is
nothing special about mathematical language: just as "3 + 3 = 6' is
meaningful only in the context of our language; my pointing to a desk
is only pointing to a desk within our language-game; and my statement
"the cat is on the desk" is only about animals and furniture in the
background of our language.
If meaning is a function of practice, then there is no room for any
determinations of meaning that are not part of our practice. The mis-
take is not in talking of a mathematical reality, the mistake is to think:
that the meaning of this reality can be independent of our practice;
that we can say something more than the proof; and that mathematical
reality can be opposed to proof. Thus the Hardyian Picture of a mathe-
matical reality adjudicating our practice has become, in a phrase Witt-
genstein sometimes used, a wheel turning idly. As Wittgenstein wrote:
"Even God can determine something mathematical only by mathemat-
ics" (RFM [1941]: VII.41). .
The Hardyian Picture is the constant in Wittgenstein's philosophies
of mathematics; as different as the calculus and language-game concep-
tions are, they are both motivated by an opposition to the Hardyian
Picture's view that there is an underlying mathematical reality which
our language and practice must mirror or be responsible to. In the
calculus period Wittgenstein's point can be put this way: there are no
criteria for mathematical correctness outside of the rules of individual
calculi. In the language-game period the point becomes: there are no
critieria for mathematical correctness outside of mathematical practice.
It is one thing, however, for Wittgenstein to say that the Hardyian
Picture is wrong; it is another to give a positive account of mathematics
in terms of language and language-games. This is what the calculus and
language-game conceptions are each designed to do, and we turn to
these very different positive accounts next.
3.
As is well known, the earlier Wittgenstein looked at meaning and
language in terms of logical form and definiteness of sense, while the
later Wittgenstein looked to language-games and forms of life. In be-
tween, however, there was a distinct transitional theory of language -
the calculus conception.
The transitional Wittgenstein wrote:
Grammar is not accountable to any reality. It is grammatical rules that determine meaning
(constitute it) and so they themselves are not answerable to any meaning and to that
extent are arbitrary. (PG [1932-34): 184)9
This is obviously part of Wittgenstein's rejection of the Hardyian Pic-
ture's external criteria which judge language. However, the calculus
conception is itself an incorrect account, failing to capture the open-
ended character of mathematics. Godel's theorem, by showing that any
132 STEVE GERRARD
formal axiomization of arithmetic will be incomplete, demonstrates
conclusively that any conception of mathematics as simply a calculus
fails. In addition to this problem, the calculus conception fails to give
a correct account of mathematical language. The conception unneces-
sarily restricts the scope of mathematical language, reducing it to mere
syntactical rules. Furthermore, as heavily as the conception depends
on the notion of rules, there is here no analysis of that notion: any
account of the application of a calculus or a critique of what it means
to follow a rule is pushed aside only to be ignored. No account of the
change or growth of mathematics can be given; the rules are all we
have and that is all we can say. The result is a conception of mathema-
tical language that is, whatever Wittgenstein's intentions, divorced from
human use or purpose.
The language-game conception overturns this. There meaning and
truth can be accounted for only in the context of a practice, and
mathematics is examined by seeing what special role it plays in our
lives and its special relationship to our other language-games. The
calculus conception was unable to account for the change and growth
of mathematics, while the language-game conception emphasizes this.
The bizarre views that have earned many commentators' derision drop
out.
It is not that Wittgenstein's calculus period was a complete disaster,
nor is our point simply that at one time Wittgenstein's views were
confused but he later corrected them. Rather, the language-game con-
ception is an extension and revision of the earlier view and not a
complete rejection of it. The mistakes of the calculus period were not
haphazard, but had a point; a point that could not reach complete
fruition until the language-game period.
Wittgenstein said at the beginning of the transition that "[c]ontradic-
tion is between one rule and another, not between rule and reality"
(CAM I [1931-32]: 92). Wittgenstein's transition is characterized in
many ways by his working out of this theme. What he is doing is trying
to account for the meaning and objectivity of mathematical language
(and indeed of all language) without the Hardyian Picture's misleading
notion of a transcendental mirroring of reality. In seeing that Witt-
genstein is expounding this theme, we can rationalize many of the
mistakes of the calculus period. Many of those mistakes come from
reaching too far and from unnecessarily restricting the scope of lan-
guage.
From the texts of the transitional period we can abstract four interre-
lated principles that characterize Wittgenstein's calculus conception:
a) the autonomy principle: each individual calculus is a closed, self-
contained system, having no external critique;
b) there is a sharp separation between a calculus and its application
or use;
c) any revision of the rules, no matter how minor, means an entirely
new calculus; and
d) we can only discuss individual calculi, we can assert no require-
ments of calculi in general.
These four principles (and their corollaries) ripple throughout the
transitional works. Wittgenstein's changing philosophies of mathematics
is the story of how each of these principles is modified or discarded as
the notions of language-games and forms of life come to the fore.
The autonomy principle (a) is never completely discarded; properly
modified it will play an important part in Wittgenstein's later philos-
ophy. At first look it seems little more than a restatement of the context
principle.
A name has meaning, a proposition has sense in the calculus to which it belongs. The
calculus is as it were autonomous. - Language must speak for itself ....
The meaning is the role of the word in the calculus. (PG [1932-34]: 63)
There is more here than the context principle, however. At this
stage the autonomy principle denies the possibility of any critique of
language. Wittgenstein often makes this point rather bluntly:
A mathematical system, e.g., the system of ordinary multiplication, is completely closed.
I can look for something only within a given system, not for the system. (WWK [1929]:
35).
I might as well question the laws of logic as the laws of chess. If I change the rules it
is a different game and there is an end of it. (CAM I [1930]: 19)10
This is part of Wittgenstein's opposition to the Hardyian Picture, a
preliminary step in the rejection of the view that in some transcendental
sense language must mirror the world.
The emphasis at this stage is on rules, but although the rules bear
the entire brunt of meaning there is no critique of them here. Rules
are seen as meaningful (indeed as the sole criterion for meaning) totally
apart from their use or a background practice. This will obviously
change in the language-game conception.
Part of the problem of a lack of critique of rule following involves
134 STEVE GERRARD
principle (b) where there is a sharp separation between the calculus
and its application. In his transitional years, what Wittgenstein usually
means by "calculus" is an uninterpreted logical or mathematical lan-
guage - the syntax alone
ll
(see WWK [1930]: 106). "If I am asked,
then", Wittgenstein says, "what it is that distinguishes the syntax of a
language from the game of chess, I answer: It is its application and
nothing else" (WWK [1930]: 104). He continues later: "Detached from
its applications and considered by itself it [syntax] is a game, just like
chess" (WWK [1930]: 105).
It is not very clear (and that is a part of the problem with the calculus
conception) what Wittgenstein means by application here. As it is
commonly used, "application" can have two different meanings in the
context of mathematics; the distinction is between the application of
the calculus and the application of the rules. In the former, call it (1),
we can speak of applying a calculus to something else. An example
would be using arithmetic to say that if there are two apples in the
refrigerator and two apples on the table, then there are four apples in
the kitchen. In the latter, call it (2), it is an application of the rules of
arithmetic that the series n + 2 is continued 1000, 1002, 1004. It appears
that at this stage of his thought Wittgenstein was not concerned with
the difference between (1) and (2), but usually thought of application
as meaning practical use (the application of a calculus as a whole, not
the application of the rules).
One reason for the lack of clarity here is that it is not just that
calculus and application (in any sense) are completely separated, but
just as in the Tractatus; where problems of epistemology and application
are shunted aside as not germane to philosophy, here application is
separated from the calculus only to be ignored. "Mathematics is always
a machine, a calculus", Wittgenstein said, "[t]he calculus does not
describe anything. It can be applied to everything that allows of its
application" (WWK [1930]: 106).12 The motivation for this view is the
rejection of the Hardyian Picture's correspondence of mathematics with
reality. But the (unnecessary) price Wittgenstein pays here for that
rejection is to separate mathematical language not only from this reality,
but from human use or practice as well. Here we see the separation
without any further account of application or the relationship between
calculus and application. Both the examination of the application of a
calculus and the examination of following a rule belong to a later
Wittgensteinean stage. Indeed, once Wittgenstein saw that rules have
no meaning apart from a practice he had to reject the calculus concep-
tion.
Principle (c) causes more trouble. It claims that any revision of
the rules means an entirely new calculus.
13
This peculiar view can be
explained with reference to Wittgenstein's favorite example of chess.
At one time the rules of chess were different from the modern ones.
Pawns could only be advanced one square at a time - even the first time
they were moved. This slowed the opening of the game considerably, as
usually in the opening a player advanced a pawn once and then on the
very next move advanced it again. In order to remedy this a new rule
was added: the first time a pawn was moved it could be moved either
one or two squares. The en passant rule was then added so no pawn
could be promoted without the possibility of it being captured by an-
other pawn. All this seems very reasonable, as do the following judg-
ments: modern chess is a faster game than old chess; modern chess is
a better game than old chess; and modern chess without the en passant
rule would result in some "cheap" victories. In addition, certain middle
and endgame positions from old chess are the same as in modern chess,
and we can compare how different players handled similar situations.
Many themes present in old chess are useful for modern chess. Again,
all this seems reasonable, but under Wittgenstein's calculus view the
above judgments would be completely senseless. There can be no such
thing as the evolution of a game; a game is completely characterized
by its rules alone, and if any rule is different, then all that can be said
is that they are different games. There can be no such thing as an
improved chess; there is only chessl and chess2 and chess3' Since all
we can appeal to is the rules, when the rules are at all different there
is no more basis for saying chessl and chess2 have anything in common
than saying that cat and cattle have anything in common.
Wittgenstein explicitly compares the situation in chess with the situ-
ation in mathematics.
We cannot reduce mathematics; we can only make a new one. The size of a proof can
be reduced, but not the body of mathematics. The same point can be made about chess.
Suppose chess is defined by the way we move pieces, and that a new way of producing
a certain move is discovered. This is not to reduce the old game; it is to make a new
game. (CAM II [1933-34]: 71)
Again, this is the autonomy principle gone too far. The number system
with irrationals is obviously different from the number system with only
136 STEVE GERRARD
rationals. But it is an extension of the old system, not something
completely new and different. This is what the transitional Wittgenstein
explicitly denies (see WWK [1929]: 35-36) and is what the later Witt-
genstein explicitly asserts in the Investigations (see PI: 67).
Principle (d) emphasizes this sharp separation between different cal-
culi. 14 Since there can be no reference to anything outside of the rules
(including their purpose), and since there is no basis to assert that
calculi have anything in common, then all that we can do is discuss
individual calculi: we can assert nothing about calculi in general. Under
this principle it makes no sense to require that for something to be a
calculus at all it must be consistent, nor do we have a basis for saying
that one calculus is more complete than another.
4.
The test case for my theory of Wittgenstein's transition is his notorious
remarks on contradiction. Those remarks have puzzled and even irri-
tated many commentators. In 1931 Wittgenstein said:
I want to object to the bugbear of contradiction, the superstititious fear that takes the
discovery of a contradiction to mean the destruction of the calculus. (WWK [1931]: 196)
This typical quotation seems to justify one commentator's statement
that "[n]ow on the face of it, the views Wittgenstein puts forward here
are absurd. ,.15 In remark after remark from 1929 through the 1930s,
Wittgenstein seems to be doubting the harm of contradiction.
16
It is
almost as if he was following the Walt Whitman school of contradiction:
"Do I contradict myself? / Very well, then, I contradict myself. / (I
am large, I contain multitudes).,,17
But more than a decade later, what Wittgenstein decides to include
on the subject of contradiction in the Investigations is perfectly reason-
able.
The fundamental fact here is that we lay down rules, a technique, for a game, and then
when we follow the rules, things do not turn out as we had assumed. That we are
therefore as it were entangled in our own rules.
This entanglement in our rules is what we want to understand (i.e. get a clear view
of) ....
The civil status of a contradiction, or its status in civil life: there is the philosophical
problem. (PI: 125)
Here, far from debunking the problem of contradiction, Wittgenstein
is trying to get a clear view of why a contradiction does mean the
destruction of a calculus, of what it means to be entangled in our own
rules.
What we see here is Wittgenstein's two different philosophies of
mathematics, both partly motivated by an opposition to the Hardyian
Picture. The crucial fact is that Wittgenstein's Walt Whitman theory of
contradiction follows directly from the calculus conception's principles,
but the language-game conception rejects that theory' by overturning
those principles.
The autonomy principle (a) quickly leads to the transitional Wittgen-
stein's views on the harmlessness of a contradiction in a calculus. Our
syntactical rules may allow us to produce a sentence of the form 'F &
- F', but since under the calculus conception the rules of the calculus
itself are the only criteria for the meaning of the calculus's expressions,
then there is no standpoint from which to judge such sentences a
problem. The rules are not about anything, so they cannot come into
conflict with anything. ("[M]athematics is a calculus and hence isn't
really about anything" (PG [1932-34]: 290).) It cannot even be claimed
that the rules come into conflict with each other, for if all we have is
the syntax (the rules of the calculus), then the rules themselves are the
final court of appeal. The rules determine what the calculus is, and if
the rules produce F & -F then that is all there is to it: we have a
calculus with F & -F just as there are calculi with P v -Po
If you change the rules to disallow sentences of the form F & -F,
then according to principle (c) you have changed the meaning and
moved to a new calculus; you have not corrected the old one (just as
chessz is a different - not a better - game than chessl)' And principle
(d) rejects the requirement that all calculi be consistent, and indeed,
holds that any such requirement for all calculi would be senseless.
The calculus principles, while opposing the Hardyian Picture, have
restricted too much and resulted in seeing the fear of contradiction as
a superstition.
But when considered from the language-game conception consistency
is not an optional characteristic of calculi, nor is it a superstitution.
The point, once again, can be made with reference to chess. In the
Investigations Wittgenstein writes: "Let us say that the meaning of a
piece is its role in the game" (PI: 563). (Note that without further
interpretation of "role" this sentence would be accurate for both the
calculus and language-game periods.) But Wittgenstein then imagines
138 STEVE GERRARD
that the rules of chess state that the king will be used at the beginning
of the game to determine which player gets white: "To this end one
player holds a king in each closed fist while the other chooses one of
the two hands at random." Is this rule on the same level as the rule
that the chess game is won when the opponent's king cannot escape
capture? Wittgenstein implicitly denies this, and then writes: "The
game, one would like to say, has not only rules but also a point" (PI:
564). This last sentence is crucial. It keeps the calculus conception's
emphasis on rules, but adds the new theme that the rules must be
placed in a context of human activity.
Take three games. The first is our chess, the second is chess with
Wittgenstein's white-choosing rule, and the third is like our chess except
that the winner is the one who first moves her queen rook. Compare
this to three calculi. The first is a standard predicate calculus. The
second is like the first, but all variables must be underlined. The third,
like Frege's, can produce Russell's paradox. Assuming standard human
characteristics and goals, one would say that the white-choosing rule is
an inessential modification of our chess, while the queen-rook rule
destroys the point of the game, preventing it from serving all the
purposes we now use chess for. (White would always win in two moves.)
Similarly, the variable underlining rule changes nothing essential,
whereas the inconsistency of the third calculus destroys the point of the
calculus. For, under Wittgenstein's later conception, calculi are not
seen as isolated systems of rules, but are systems of rules which have
a role - and a point - in our lives.
This returns us to principle (b). In Wittgenstein's later conception
one cannot separate a calculus from at least one sense of application.
Being a calculus already implies an application, namely, to calculate.
We may get truths from falsehoods or falsehoods from falsehoods, but
the absolute minimum requirement of a calculus is that it not yield
falsehoods from true premises. IS If we had such a calculus we might
just as well consult the entrails of a bull. And this, of course, is what
happens with inconsistent calculi: from true premises one can deduce
falsehoods; indeed, one can deduce anything at all. 19
Looked at from the calculus conception, an inconsistent calculus is
just another calculus. In order to oppose the Hardyian Picture Witt-
genstein had stripped away any constraint on meaning not due to
syntactical rules. In so doing Wittgenstein denied the harm of contradic-
tion and was led to bizarre bugbear statements. But under the language-
game conception the opposition to the Hardyian Picture remains with-
out paying the calculus conception's unnecessary price. The new goal
is to describe the harm without falling prey to the misleading picture.
What is wrong with an inconsistent calculus is not that it somehow
misrepresents mathematical reality (as in the Hardyian Picture), but
that in producing falsehoods from truths and thus allowing anything to
follow, it fails to do what we want our calculi to do, namely, to calculate.
The situation is like the strange tribe
who used money in transactions; that is to say coins, looking like our coins, which are
made of gold and silver and stamped and are also handed over for goods - but each
person gives just what he pleases for the goods, and the merchant does not give the
customer more or less according to what he pays. In short this money, or what looks like
money, has among them a quite different role from among us. (RFM [1937-38]: 1.153;
emphasis added)
Money has a special role in our lives, and this is what justifies translating
foreign terms as "money". Mathematics has a special role in our lives
and special characteristics, 20 and these are what justify calling alien
practices "mathematics." An inconsistent calculus cannot serve this
role. Wittgenstein's mature philosophy of mathematics saves objectivity
without paying the Hardyian Picture's price of meaningless external
criteria or paying the calculus conception's price of too restricted in-
ternal criteria. Mathematical objectivity and necessity are due to the
special role of the mathematical language-game.
The general lesson of this specific look at contradiction is that an
account of Wittgenstein's mature philosophy of mathematics should not
saddle the language-game conception with the mistakes due to the
calculus conception's principles. The Investigations (and some later
sections of Remarks on the Foundations of Mathematics) emphasize and
examine the change and growth in the language-games of mathematics.
Rather than neglecting application, they concentrate on it. Instead of
seeing mathematics as isolated calculi, they conceive it as a nexus of
language-games, related to each other and embedded in our form of
life. No longer is it forbidden to say anything about calculi other than
give the rules - calculi have roles, and they can serve them more or
less successfully.
140 STEVE GERRARD
5.
The conception of language in general as a nexus of language-games is
the view of the Investigations, and is, to say the least, well known. For
too long, however, it has not been recognized that Wittgenstein has a
coherent language-game conception of mathematics. This is partly due
to the failure to recognize the calculus period and its demise. Most
commentators have either felt obligated to defend the bizarre remarks
of the calculus period, or alternatively, have ~ t t c k e d those remarks
without recognizing that Wittgenstein himself later saw the problems.
Coupled with the Hardyian Picture, the recognition of Wittgenstein's
two philosophies of mathematics should help overturn the black hole
view of Wittgenstein's career. Stripped of the calculus conception,
Wittgenstein's language-game conception of mathematics can now be
examined unencumbered by the earlier conception's mistakes.
21
NOTES
* The following abbreviations are used in this article to refer to Wittgenstein's works:
WWK: Ludwig Wittgenstein and the Vienna Circle: Conversations Recorded by Friedrich
Waismann, ed. B. F. McGuinness, trans. J. Schulte and B. F. McGuinness, Oxford:
Basil Blackwell, 1979; CAM I: Wittgenstein's Lectures: Cambridge, 1930-32, ed. D. Lee,
Chicago: University of Chicago Press, 1982; CAM II: Wittgenstein's Lectures: Cambridge,
1932-35; ed. A. Ambrose, Chicago: University of Chicago Press, 1982; PG: Philosophical
Grammar, ed. R. Rhees, trans. A. Kenny, Oxford: Basil Blackwell, 1974; BIB: The Blue
and Brown Books, Oxford: Basil Blackwell, 1958; LFM: Wittgenstein's Lectures on the
Foundations of Mathematics: Cambridge, 1939, ed. C. Diamond, Ithaca: Cornell Univer-
sity Press, 1976; RFM: Remarks on the Foundations of Mathematics, ed. G. H. von
Wright, R. Rhees, G. E. M. Anscombe, trans. G. E. M. Anscombe, revised ed.,
Cambridge: MIT Press, 1978; PI: Philosophical Investigations, ed. G. E. M. Anscombe,
R. Rhees, trans. G. E. M. Anscombe, New York: Macmillan Company, 1953; Z: Zettel,
ed. G. E. M. Anscombe, G. H. von Wright, trans. G. E. M. Anscombe, Berkeley and
Los Angeles: University of California Press, 1970.
References to PI and Z are to remark number; references to RFM are to part number
(Roman numerals) and remark number (Arabic numerals); and references to the other
works are to page numbers. As the evolutionary nature of Wittgenstein's work is an
important theme of this article, following the abbreviation for the book in the text I have
put in brackets the date of the book or the part of the book from which the quotation
comes.
1 Michael Dummett, 'Wittgenstein on Mathematics', Encounter 50, (March 1978), 68.
2 Calling the black hole view a "standard view" does not, of course, mean it is the only
view. There are those who, like Dummett, read Wittgenstein as a radical in the philosophy
of mathematics, but, unlike Dummett, think Wittgenstein is right. They dissent from the
black hole theory for different reasons than I do. Two notable examples are Crispin
Wright, Wittgenstein on the Foundations of Mathematics, Cambridge: Harvard University
Press, 1980; and S. G. Shanker, Wittgenstein and the Turning-Point in the Philosophy of
Mathematics, State University of New York Press, Albany, 1987.
3 W. W. Tait writes that, properly examined and freed of confusions, "Platonism will
appear, not as a substantive philosophy or foundation of mathematics, but as a truism"
('Truth and Proof: The Platonism of Mathematics', Synthese 68, (1986), 342). In this
section I am heavily indebted to this article.
4 There is ample evidence that Wittgenstein was familiar with both Hardy and his works.
See John King, 'Recollections of Wittgenstein', in Rush Rhees (ed.), Recollections of
Wittgenstein, Oxford University Press, Oxford, 1984, p. 73; G. H. 'von Wright, A Bio-
graphical Sketch, in Norman Malcom, Ludwig Wittgenstein: A Memoir, Oxford University
Press, Oxford, 1958, p. 6; Wolfe Mays, 'Recollections of Wittgenstein', in K. T. Fann
(ed.), Ludwig Wittgenstein: The Man and the Philosopher, Humanities Press, New Jersey,
1967, p. 82; and Garth Hallett, A Companion to Wittgenstein's 'Philosophical Investi-
gations', Cornell University Press, Ithaca, 1977, p. 766. References to Hardy in LFM are
listed in that book's index. Wittgenstein also mentions Hardy in CAM II [1932-33]: 215-
20, 222, and 224-25; in BIB [1933-34]: 11; and quotes from 'Mathematical Proof' in Z
[1945-48]: 273.
5 G. H. Hardy, A Mathematician's Apology, Cambridge University Press, Cambridge,
1967, pp. 123-24.
6 Rush Rhees, 'On Continuity: Wittgenstein's Ideas, 1938', in Discussions of Witt-
genstein, Routledge & Kegan Paul, London, 1970, p. 109.
7 G. H. Hardy, 'Mathematical Proof', Mind 38, (January 1929), 18. Hardy continues:
"This is plainly not the whole truth, but there is a good deal in it."
8 See W. W. Tait, op. cit., 343-45. Once again I am heavily indebted to this article.
9 See also PG [1932-34]: 89: "[L]anguage is not something that is first given a structure
and then fitted onto reality", and even more simply PG [1932-34]: 143: "It is in language
that it's all done."
10 For more on the early transition's comparison between the rules of chess and language
see WWK [1930]: 134: "[I]t is only the rules of the game that define this [chess] piece.
A pawn is the sum of the rules according to which it moves (a square is a piece too),
just as in language the rules of syntax define the logical element of a word."
11 There are troubles here from the very beginning. How is something logical, mathema-
tical, or even a language if it is uninterpreted?
12 See also WWK [1930]: 114: "The calculus can be applied to anything that admits of
such application. (And you cannot say anything that goes beyond this.)"
13 For representative statements see PG [1932-34]: 374-75 and WWK [1929]: 35-36.
14 See WWK [1930]: 133; WWK [1931]: 149 and 202; and PG [1932-34]: 369.
15 Charles S. Chihara, 'Wittgenstein's Analysis of the Paradoxes in his Lectures on the
Foundations of Mathematics', Philosophical Review 86 (1977), 369.
16 See also WWK [1930]: 120, 131; WWK [1931]: 174,208; PG [1932-34]: 303-05; CAM
II [1933-34]: 71; LFM [1939]: 206-07, 211, 224-25; and RFM [1937-38]: App. III. 17.
17 Leaves of Grass, "Song of Myself", verse 51.
18 Wittgenstein was late in seeing the implications of this point. (The calculus conception
held sway the longest in the area of contradiction.) In LFM [1939]: p. 230, Wittgenstein
writes: "You could say that if we allow a contradiction, in the sense that we allow
anything to follow from it, then we have given up any idea of a calculus at all." This is
right (and, indeed, is my point), but at the time Wittgenstein mistakenly thought that it
142 STEVE GERRARD
was false that one may be able to derive any sentence from a calculus which contains a
contradiction without actually going through the contradiction. (Chihara, op. cit., takes
Wittgenstein to task for this.) Thus he thought he could have an ad hoc solution to
contradiction: when a contradiction is discovered, simply add an ad hoc rule prohibiting
the drawing of any conclusions from the contradiction. Even this serious mistake was
partly motivated by opposition to the Hardyian Picture. Wittgenstein was trying to say
that it is we who are in charge llere; we are not being pushed from the outside. The
community of language users makes the rules, and if things do not turn out as they
intended (e.g., they get F & -F), then they can simply change the rules.
19 My colleague Sahotra Sarkar advises me that I need only to assume that the logical
system is of such a type and has such a notion of derivability that it permits all that goes
on in classical mathematics.
20 Chief among these special characteristics is that the statements of mathematics are
non-revisable by sensory experience. This partly separates a proof from an experiment.
A full account of Wittgenstein's language-game conception of mathematics would exam-
ine this more fully.
21 It would be impossible to exaggerate the help I received on an ancestor of this article
from Leonard Linsky and especially W. W. Tait. I am also grateful to Lydia Goehr for
her helpful criticisms.
Dept. of Philosophy
Boston University
Boston, MA 02215
U.S.A.
JULIET FLOYD
WITTGENSTEIN ON 2,2,2 ... : THE OPENING OF
REMARKS ON THE FOUNDATIONS
OF MA THEMA TICS!
l.
If we consider Wittgenstein's career as a whole, it is clear that he wrote
more on philosophy of logic and mathematics than on any other subject.
Yet his writings on these topics have exerted little influence. Indeed,
the tide of response to Remarks on the Foundations of Mathematics -
which contains the bulk of his later views of mathematics - has been
for the most part negative, and many able readers have concluded that
Wittgenstein was simply out of his depth in the fields of mathematics
and logic.
2
Thus the most pressing question with regard to Wittgen-
stein's discussion of mathematics - addressed as yet by relatively few
commentators - is, Does it offer anything at all?
The opening sections of Remarks on the Foundations of Mathematics
are, I shall argue, the beginnings of a powerful and total reworking of
a traditional philosophical picture of logic and mathematics. Witt-
genstein is out to question whether the hallowed earmarks of logical
and mathematical truth - its necessity, a prioricity, self-evidence and
certainty - may serve as the appropriate starting point for philosophiz-
ing, or even whether these notions, in the end, make sense. Even more,
he will question the very applicability of the notion of truth to so-
called logical and mathematical "statements". The later Wittgenstein's
conception of philosophy has long been a stumbling block for his read-
ers. But the very strangeness of the opening questions of Remarks on
the Foundations of Mathematics is highly revealing of Wittgenstein's
aims: these questions are intended to slow up the too facile reasoning
of the philosopher who fastens on the necessity of mathematical and
logical inference as a peculiar philosophical datum. He challenges his
interlocutor to explain precisely what it is he is trying to account for
and what it is he is seeking. Wittgenstein's ultimate aim is to test the
limits of idealized conceptions of logic in the doing of philosophy. He
attacks the construal of formal logic as the overarching standard for
clarity, as the basic model for the functioning of language in general,
and as the key tool in the resolution of philosophical problems. (Such
Synthese 87: 143-180, 1991.
144 JULIET FLOYD
a construal was central to the work of Frege, Russell and Wittgenstein,
himself, in Tractatus.)
Wittgenstein is out to undercut the notion that particular claims must
be made from within an over arching general philosophical conception
or systematic logical structure. This means that he will continually
question the urge to construe his own statements in general philo-
sophical terms, that is, as part of a systematic a priori theorizing about
language, or logic, or thought. Wittgenstein writes in dialectical form
in order to expose the character of philosophical debate over the nature
of logic and mathematics. The philosopher's task, according to Witt-
genstein, consists in "assembling reminders for a particular purpose,,3 -
that is, at a particular juncture in the philosophical discussion - in order
to give us insight into the workings of our language. The profundity
and impressiveness of philosophy retreats from purported answers to
the nature and character of the questions themselves.
4
There is an impression, held by many readers of Philosophical Investi-
gations, that Wittgenstein's interlocutor is a straw character whose
queries, assertions and claims are those of no serious philosopher. I
believe that no amount of argument or methodological clarity can
totally lay such objections to rest. However, Wittgenstein seems to
have been acutely aware of this. As I read him, the dialectical style of
his later writing - and its seemingly interminable character - reflects
his conception of the problems under discussion. We are left in the
end with dispensable comments, for purposes of linguistics, empirical
psychology or mathematical practice, but the point of the investigation
is to show that what the essentialist says about language, thought and
mathematics is equally dispensable.
s
This is not supposed to be demon-
strated by a linear argument from fixed premises. The tone and manner
in which Wittgenstein's own considerations emerge - which so many of
his readers find frustrating and tortuous - is internal to the claims being
made.
Wittgenstein's interlocutor reacts in an unguarded way, without re-
finement. He does not begin with a metaphysical conception. But as
the interlocutor refuses, over and over again, to accept Wittgenstein's
countersuggestions about the uses of various expressions, his words
about "justification", "knowledge", "compulsion", "truth" and
"proof" begin to take on a metaphysical cast. This is Wittgenstein's
way of attempting to depict the move into a philosophical worry, i.e.,
WITTGENSTEIN ON 2, 2, ... 145
the origin of both particular philosophical accounts and the need for
them.
6
Clearly, if there is such a stage in philosophizing to be un-
earthed, it cannot itself be made out in systematic or general terms; if
someone (including himself) tried to do so, Wittgenstein would no
doubt have subjected these terms themselves to the same sort of scru-
tiny. The roots of philosophy's demands for systematic accounts are
divergent and multifarious. Wittgenstein's continued sparring with the
interlocutor is meant to show this, and at the same time to portray the
task of clarification as never definitively completed. The interlocutory
voice is not meant to be silenced; illumination comes from the shifts
and modulations of its tone and through the connections drawn between
questions in the course of the conversation. The simplicity and naiVete
of the interlocutor (whose remarks, at the same time, are never foolish
or off the point) is also intended to get us to ask the question, How
much more sophisticated and explanatory are the full-blown theories
which so attract us (including Wittgenstein himself)? It is an open
question how much distance lies between a given, more refined view
and the utterances of the interlocutor. And Wittgenstein intends to get
us to see that this distance is always less than we might have initially
supposed.
The force of Wittgenstein's method - insofar as it attempts to depict
the roots and seeds of philosophical puzzlement -lies in the details, the
inner gropings of the exchanges between the interlocutor and Witt-
genstein. Too few of Wittgenstein's commentators (whether interpreters
or critics) have been willing to offer detailed exegesis of his interlocutory
debates (especially in the case of Remarks on the Foundations of Mathe-
matics). But engagement with the text is the only way to see what the
method in the end amounts to. The remainder of this paper will attempt
to show, by way of example, the value of such interpretive scrutiny. My
focus, after a brief discussion of the textual context of Part I of Remarks
on the Foundations of Mathematics, will be the first three sections of that
book. In these sections, Wittgenstein attempts to unearth the roots of
philosophical wonder about the nature and necessity oflogical and mathe-
matical inference. He constantly challenges the interlocutor to rethink his
(pre )conceptions of necessity, knowledge and truth so as to reflect our
actual use of these notions. The intended result is to shock us into asking
ourselves whether we are as clear as we might have supposed about what
inference and necessity really involve.
146 JULIET FLOYD
2.
The first sections of Remarks on the Foundations of Mathematics were
not originally intended by Wittgenstein to stand on their own. They
form the initial paragraphs of the second part of the so-called "early
version" of Philosophical Investigations, composed between 1936 and
1937. This "Friih version" is a typescript consisting of two parts.
7
The
first part corresponds to a substantial draft of th,e first 188 sections of
the final published version of Philosophical Investigations, while the
second focuses on the philosophy of logic and mathematics. In 1938
Wittgenstein wrote a preface to the entire typescript and submitted it
to the Cambridge University Press. Though the Press agreed to its
publication, Wittgenstein never allowed the book to come to print.
8
Von Wright hypothesizes that Wittgenstein's hesitation "was connected
with his continued work on the second part of the book dealing with
the philosophy of mathematics".
9
This work Wittgenstein never lived
to complete. He set aside the second part of the manuscript, reworked
it,1O and finally abandoned it in 1944. It has been published posthu-
mously as Part I of Remarks on the Foundations of Mathematics. 11 Thus
it is clear that the text of Remarks on the Foundations of
lacks some of the historical authority and integrity of Philosophical
Investigations and must be used with care. Nevertheless, I believe it can
fruitfully sustain detailed scrutiny. Its first three sections are especially
noteworthy because they form the branching point of Wittgenstein's
earlier and later drafts of Philosophical Investigations. We may think
of Remarks on the Foundations of Mathematics, Part I, Section 1, as
initiating an alternate trajectory for the first 188 remarks of Philosoph-
ical Investigations. 12
The first two sections of Remarks on the Foundations of Mathematics
are almost identical with Sections 189-90 of Philosophical Investi-
gations, and, as I have just said, the whole of part I of Remarks on the
Foundations of Mathematics originally followed (a draft of) Philosoph-
ical Investigations Sections 1-188. This explains its rather abrupt begin-
ning, which examines, as if out of nowhere, the notion of the
"steps ... determined" by an algebraic formula.
13
In the framing dis-
cussion (i.e., the earlier sections of Philosophical Investigations and the
"Fruhversion") Wittgenstein had embarked on a lengthy investigation
into the notions of understanding, meaning and thinking in connection
with our command of language. At issue is the nature of the logic of
our language, and in what sense our use of language may be regarded
as the operation of a calculus of definite rules. Wittgenstein writes:
F. P. Ramsey once emphasized in conversation with me that logic was a 'normative
science'. I do not know exactly what he had in mind, but it was doubtless closely related
to what only dawned on me later: namely, that in philosophy we often compare the use
of words with games, calculi which have fixed rules, but cannot say that someone who
is using language must be playing such a game. - But if one says that our linguistic
expression only approximates to such calculi, one thereby immediately stands on the very
brink of a misunderstanding. For then it may look as if what we were talking about in
logic were an ideal language. As if our logic were, so to speak, a logic for a vacuum. -
Whereas logic does not treat of language - or of thought - in the sense in which a natural
science treats of a natural phenomenon, and the most that can be said is that we construct
ideal languages. But here the word "ideal" would be misleading, for it sounds as if these
languages were better, more perfect, than our everyday language; and as if the logician
were needed to show people at last what a proper sentence looked like.
All this, however, can only appear in the right light when one has attained greater
clarity about the concepts of understanding, meaning, and thinking. For it will then also
become clear what can lead us (and did lead me)14 to think that if anyone utters a
sentence and means or understands it he is thereby operating a calculus according to
definite rules. 15
Among central questions examined in Philosophical Investigations are:
What is it that we learn when we learn language?; What is it to under-
stand a word, or a sentence?; How is it that words, sentences, have
meaning at all?; And, is the command of language itself a special kind
of knowledge? These are for Wittgenstein simultaneously part of an
investigation of the scope and nature of what he views as logic, that is,
the logic of our language. Evidently, in Philosophical Investigations
Section 81, Wittgenstein is revealing a plan to investigate the limits and
presuppositions of philosophical uses of "ideal languages", including
formalizations of what we have come to think of as logic, i.e., for-
malized languages whose grammatical structures are the truth-func-
tional and quantificational structures developed by Frege and Russell.
In Wittgenstein's later philosophy, logical structure is not conceived of
as lying behind language awaiting discovery through analysis of the
proposition, the concept, or the nature of logical inference. Instead,
structure is to be exhibited and elicited by "comparing" our uses of
language to games with fixed rules (the simplified "language games"
of, e.g., Philosophical Investigations). But this indicates that for Witt-
genstein our language is not itself a game with fixed rules; that is, our
language is not itself a "language game". 16
148 JULIET FLOYD
The example of an algebraic formula is brought into Wittgenstein's
discussion in Philosophical Investigations Section 146 as a paradigm of
what it is to operate in accordance with a definite rule. In the simplified
"language game" of Section 143, we are asked to imagine that "when
A gives an order, B has to write down series of signs according to a
certain formation rule [Bildungsgesetz]".17 And now we imagine teach-
ing a pupil to write out the series of numbers from zero to nine. But
then "the possibility of getting him to understllfld will depend on his
going on to write it down independently,,18 - that is, in his being able
to follow a rule. What enables him to follow the rule? We can imagine
that the pupil is unable to grasp the rule of the series; that no matter
how we encourage, blame, or play with him, he simply continues on
differently, all the while calling his way "going on in the same way". 19
Wittgenstein emphasizes that we may imagine this,20 but imagining a
scenario is not for Wittgenstein the same as understanding or making
sense of it. Having a picture is not the same as using or applying it.
Hence it is not clear that Wittgenstein wishes to say that we could make
sense of such a case. Still, our capacity to picture it may allow us to
reflect on the extent to which understanding may be fruitfully viewed
as an activity governed by fixed rules - or even on the notion of what
it is to operate with a definite rule in mind.
The rule of an arithmetic series expressed by an algebraic formula is
a paradigmatic case of something we feel we need to grasp before we
are to be capable of carrying out the correct development of a series.
By analogy, the understanding of a word may be thought of as the
grasping of a rule for its use. Is the rule itself then a Platonic entity?
How do we reach or follow out its "extension", i.e., the things or
situations to which it correctly applies? According to Wittgenstein's
figure of speech, the rule for use of a word must be grasped before
correct responses and utterances are possible?1 What the recalcitrant
pupil shows, however, is that the framing of such rules alone will not
make sense of our uses of language. For even in very simple cases (like
0,1,2,3,4,5,6,7,8,9) it is clear that an algebraic formula will not carry
within itself the understanding of its application.
22
We may be tempted
at this point to draw a distinction between the expression of a rule (the
algebraic formula) and the rule itself.23 But then, having drawn a
systematic distinction between the two, we may begin to suspect that
the rule, not the rule's expression, compels and makes possible an
individual's correct responses. The analogy with understanding a
linguistic expression may then incline us to think that the understanding
of the word, the rule, or concept grasped, is something other and more
fundamental than any expression of the rule we can provide. The rule
itself - or the understanding - determines something ahead of our
applications of particular linguistic expressions. But since it is difficult
to imagine what a self-interpreting rule would be, we are faced with a
puzzle about how and in what sense a rule may determine its appli-
cations in advance.
24
Wittgenstein's framing of the problem of rule-following in terms of
an analogy between the use of an algebraic formula and the use of a
linguistic expression is no accident, as Investigations Section 81 shows.
But the analogy is more complex than either the background of Trac-
tatus or the text of Investigations straightforwardly indicate. For Witt-
genstein pushes the analogy in several different directions. In Investi-
gations the question of a rule's "determining the steps" ahead of time
is examined as part of an effort to show the limits of conceiving of the
meaningfulness of language as constituted by a set of definite rules.
Various purported criteria of understanding are put forward as possible
candidates for the role of "that which constitutes grasping a rule (or
understanding a word)": an image (Vorstellung] (PI Sections 6,396), a
picture [Bild] (PI Sections 6, 425-6), a text (PI Section 156ff), a sound
(PI Sections 6, 156ff), a sensation (PI Sections 159-60, 243ff), a feeling
(PI Section 184), an act or way of meaning (PI Sections 188, 190), an
act of intuition or insight (PI Sections 186, 213-4), an intention (PI
Sections 197, 337). None of these seem to be able to constitute or
explain the application of a linguistic expression on their own.
But the analogy between a rule for the use of a word and the
rule expressed by an algebraic formula has two sides, as the alternate
trajectory of Investigations sections 1-188, in Remarks on the
Foundations of Mathematics Part I, indicates. And in fact Wittgen-
stein's investigation of the notion of what it is to grasp a rule for the
use of a word is meant to shed light in the other direction: on our
understanding of what it is for an algebraic formula to "determine" the
algebraic "steps". The Tractatus notion that the necessity of logic and
mathematics could be in some way illuminated by means of the (linguis-
tic) notion of grammar ("logical syntax") is still in play, though it has
been substantially transformed. Now the notion of grammar is itself to
be given content by way of an examination of our uses of language in
particular contexts - including logical and mathematical contexts. How
do we actually operate with algebraic formulas?
In Remarks on the Foundations of Mathematics the determination of
150 JULIET FLOYD
algebraic steps by a formula is an example of the necessity of mathema-
tical truth and logical inference. (In Tractatus it was intrinsic to Wittgen-
stein's attack on Frege's and Russell's conceptions of logical truth to
view "proof" in logic in an algebraic way, as a purely mechanical
transformation of symbols.
25
) In Remarks on the Foundations of Mathe-
matics Wittgenstein views the "steps" determined by an algebraic for-
mula, the necessity governing the development of a mathematical ser-
ies, and the inferential transitions determined bo/ the rules of logic as
exactly parallel. (The same German word "Ubergang" is used for all
three cases.) In Part I, Section 4, Wittgenstein writes:
"But then what does the peculiar inexorability of mathematics consist in?" - Would not
the inexorability with which two follows one and three two be a good example?
and in Section 5:
"But doesn't it follow with logical necessity that you get two when you add one to one,
and three when you add one to two? and isn't this inexorability the same as that of
logical inference?" - Yes! it is the same ...
In the initial sections of Remarks on the Foundations of Mathematics,
Wittgenstein seeks to "get clear what inferring consists in". 26 But he
seeks to undermine the idea of locating necessity in one place, that is,
in a fixed structure or grammar that may be delineated a priori. Remarks
on the Foundations of Mathematics begins with an examination of our
uses of the notion of algebraic determination in order to question the
idea that we can make any general sense of what following an algebraic
rule consists in. By analogy, Wittgenstein is thereby questioning
whether we can make general sense of what inferring, understanding,
or using a word in accordance with its grammar or meaning consist in.
His attack on a fixed sense of "determination" is an attack on the idea
of necessity as a clear phenomenon, and so, a fortiori, a challenge to
the idea that logic has an interesting and unique characterization.
3.
In the original manuscript (the Fruhversion) and in Philosophical Inves-
tigations's version of the first section of Remarks on the Foundations of
Mathematics Part I, Wittgenstein is responding to the interlocutor's
charge that in the preceding discussion Wittgenstein has in some way
wished to deny, in a skeptical vein, that there is an appropriate sense
in which an algebraic formula may be said to "determine the steps".
The interlocutor objects,
"[But] are the steps then not determined by the algebraic
formula?,,27
to which Wittgenstein responds,
- The question contains a mistake.
28
Clearly Wittgenstein wishes to fend off the interlocutor's attribution of
skepticism. We need to ask what sort of "mistake" the interlocutor has
made. Wittgenstein's response (with which Remarks on the Foundations
of Mathematics opens) makes this only partially clear.
We use the expression: "The steps are determined by the
formula ... " How is it used?,,23
Wittgenstein's question insinuates that the interlocutor's misunder-
standing is about this, i.e., the use of an expression. How is calling
attention to the use of an expression intended to unmask the interlocu-
tor's mistake? In a sense, my entire reading of Section 1 will be required
to show how this might happen. There is no simple "mistake" or misuse
of language involved in the interlocutor's question, and Wittgenstein
invokes no theory of meaning (e.g., of meaning as use) to refute some
independent interlocutory thesis. Instead, Wittgenstein immediately
goes on to respond to his own question in what I shall view as six
different ways. He fully knows that much of what he says will seem off
the point to those who would ask the interlocutor's question. But once
these responses to his own question are assembled Wittgenstein has
succeeded, at least for the moment, in shifting the burden of explication
onto the interlocutor, in effect making a (reasonable?) demand that
the interlocutor specify what more he wants from the notion of determi-
nation and what more is left out to be accounted for.
The first response Wittgenstein offers to the question "How is the
expression used?" sounds quite behavioristic:
We may perhaps refer to the fact that people are brought by their education (training)
so to use the formula y = x
2
, that they all work out the same value for y when they
substitute the same number for X.
3D
Wittgenstein's response suggests that the interlocutor who took
Wittgenstein's earlier words to deny that the steps are determined by
152 JULIET FLOYD
the formula has misunderstood; for Wittgenstein can certainly find a
use (so he seems to be saying) for the expression "the steps are deter-
mined by the algebraic formula". But was that all that was in question?
Surely what bothered the interlocutor was not merely finding a sense
for this series of eight words, but a concern that something essential to
the notion of determination (perhaps its "compulsion", its "necessity",
or its "rule-governedness") was being denied. Simply finding a use for
an expression - particularly a use appealing so prominently to the
notions of training, education and sameness of response - cannot touch
that worry we may feel. So the interlocutor is bound to find Wittgen-
stein's answer here grossly irrelevant. Of course, this objection cannot
rule out Wittgenstein's answer to the question about how the expression
is used as false, or even inappropriate; and Wittgenstein's suggestion
certainly raises a question about what more the interlocutor is after.
Nevertheless, there are grave difficulties facing any attempt to align
Wittgenstein's answer with the interlocutor's concerns. Viewed as a
direct response, it seems crudely behaviorist. In fact, Wittgenstein's use
of "Abrichtung", translated as "training", is bound to invite this read-
ing; "abrichten" is the verb for training or breaking in an animal, or
coaching or drilling someone, as if by rote?1 Is Wittgenstein then
denying that there is any sense of what Kant called "inward necessity"
for the people computing the square of x? No. (See my discussion of
this point in Section 4 below.)
The behavioristic reading of Wittgenstein's suggestion is equally well
invited by his second response:
Or we may say: "These people are so trained that they all take the same step at the
same point when they receive the order "Add 3"." We might express this by saying: for
these people the order "Add 3" completely determines every step from one number to
the next. (In contrast with other people who do not know what they are to do on receiving
this order, or who react to it with perfect certainty, but each one in a different way. )32
Here Wittgenstein puts the notion of a command or order in the place
of a formula.
33
The comparison or shift from "y = x
2
" to the imperative
"Add 3" is reminiscent of the language game of Investigations Section
143, and is intended to offer the interlocutor a kind of anthropomorphic
source of compulsion which might give countenance to the idea of
determination the interlocutor wishes to secure. As in Investigations
Section 143, it also places the notion of truth in the background, since
imperatives (like rules) are either followed or not followed, but are
WITTG ENSTEIN ON 2, 2, ... 153
neither true nor false. "Determines" now means that educated people
use the formula (and react to the order) as they were taught to, thereby
always reacting in the same way, "with perfect certainty" at each stage.
As Wittgenstein says, however, "we might" have been inclined to
express the idea differently, in terms of the notion of "complete deter-
mination". Why might this way seem to better express the necessity or
compulsion the interlocutor feels in the steps taken with an algebraic
formula? Since this use of "complete determination" goes along with
a reference to "every step from one number to the next", we may at
first be tempted to read "complete determination" as making implicit
appeal to the notion of a function. The generality of the function (i.e.,
its being defined for all natural numbers), coupled with its being a
function (i.e., its uniquely "associating" a single "output" with a
"given" "input"), then seems to provide us with an explanation of what
lies behind the "determination" in the cases described. It would be
natural to take the first formula, y = x
2
, to express the squaring func-
tion, and to use it as a representation of the computation of a function
at a (given) point. For someone educated in elementary arithmetic, the
ability to respond correctly to arbitrary inputs is then a test of their
grasp of the function. Similarly, the command to "Add 3" determines
our steps because the function expressed by, e.g., y = x + 3 determines
a unique answer as correct. The root of our idea would be this: since
a function "determines" a value for a given input (in virtue of being a
function), may we not better explain how the formula "determines the
steps" by saying that it expresses or denotes a function, which then
does the work of "determining the steps"?
In fact the word "function" [Funktion] appears nowhere in the pas-
sage.
34
Wittgenstein's parenthetical remark wards off this Platonistic
idea of "complete" determination, making clear that he means by it
"complete determination of behavior", i.e., all people in this group
react in the same way when asked to "Add 3,,?5 Although we may
feel deprived of an essential concept, namely, that of a function, Witt-
genstein is not recommending that we substitute the notion of "having
received appropriate training" for the idea of "grasping a function".
Rather, he is suggesting that an ambiguity lies in the notion of determi-
nation. In order to be justified in postulating the existence of a function
we need more than a mere formula: we need a determinate mathema-
tical system within which to work.
This interpretation is confirmed in the remainder of the section. In
154 JULIET FLOYD
Wittgenstein's third response, he suggests an alternative way to find a
use for the expression "The steps are determined by the formula ... ":
On the other hand we can contrast different kinds of formula, and the different kinds of
use (different kinds of training) appropriate to them. Then we call formulae of a particular
kind (with the appropriate methods of use) "formulae which determine a number y for
a given value of x", and formulae of another kind, ones which "do not determine the
number y for a given value of x". (y = x
2
+ 1 would be of the first kind, y > x
2
+ 1, y =
x
2
1, y = x
2
+ z of the second.) The proposition "The formula ... determines a number
y" will then be a statement about the form of the formulae.
36
Once again, Wittgenstein resists appeal to the notion of a "function".
The italicized "Then we call" is meant to shift our focus from thinking
about what the formulas denote (those of the first kind, as we might
say, denote or express total functions on the integers; those of the
second kind denote non-functional relations) to what we do, viz., define
a notion of "determination" by way of our classificatory system. "The
formula y = x
2
determines the number y for a given value of x" becomes
an expression of how we have decided to classify the formula; it is a
statement about the "form" of the formula. Not, however, literally its
mere shape. The activity of "sorting shapes" usually requires something
like putting circles with circles and squares with squares, or like outlines
with like outlines. But there is little, spatiotemporally speaking, which
the shapes of formulas of the same kind in this system of classification
have in common.
37
The "form" of a formula is not for Wittgenstein
merely its syntactic form; it is instead its kind relative to a practice in
which we are trained. In this case, assuming our training in the use of
particular formulas (as, e.g., in the first two of Wittgenstein's respon-
ses), we can reflect, in turn, on this use of these formulas and thereby
be able to further sort given formulas into kinds, according to the sort
of reactions trained persons (including, presumably, ourselves) have to
them.
Wittgenstein's fourth response, at first difficult to fathom, strikingly
reveals the character and motives of his criticisms. Wittgenstein writes:
- and now distinguish such a proposition as 'The formula which I have written down
determines y", or "Here is a formula which determines y", from one of the following
kind: "The formula y = x
2
determines the number y for a given value of x". The question
"Is the formula written down there one that determines y?" will then mean the same as
"Is what is there a formula of this kind or that?" - but it is not clear off-hand what we
are to make of the question "Is y = x
2
a formula which determines y for a given value
of x?" .38
In Wittgenstein's example of a classification system, particular state-
ments such as "This formula is of the first kind" represent elements of
our (use of the) classificatory system itself. But now Wittgenstein wants
us to distinguish such sentences from those such as "The formula y =
x
2
determines the number y for a given value of x". But on what basis
are we to distinguish them? In the first case I point or call attention to
a formula written somewhere; in the second I write "The formula y =
x
2
determines the number y for a given value of x'
l
- But this does
not seem to involve anything logically or semantically different. Both
sentences are, as we would say, "metalinguistic", about a formula ("y =
x
2
,,). The suspicion may arise here that Wittgenstein is involved in a
use/mention confusion. But I think the distinction he asks us to draw,
though subtle, is real. (It also illustrates the sort of distinction Witt-
genstein often urges us to draw in philosophy.) Wittgenstein is still
bearing in mind the interlocutor's "mistake" about determination. The
problem with the second sentence ("The formula y = x
2
determines the
number y for a given value of x") is that it is more likely to lead to
the kind of philosophical worries about necessity which spawned the
interlocutor's objection in the first place.
39
Wittgenstein says that it is
"not clear off-hand" what we are to make of the question "Is y = x
2
a
formula which determines y for a given value of x?". He has given us
several ways in which to respond to his question about use (and there
are others, e.g., appealing to the notion of a function). These ways
differ in emphasizing, on the one hand, the determination (through
training) of a group of people, hence their uniformity of response; and,
on the other, their training in a practice of sorting formulas into those
which "determine" and those which do not. In the first cases, it is the
people who are (said to be) determined; in the second, the formulas
(are said to) determine values; but in each of Wittgenstein's examples
education (or rote training) in a practice bears crucially on the "determi-
nation". Thus, if the question "Is y = x
2
a formula which determines
y for a given value of x?" is unclear without something further ["ohne
Weiteres"], namely, a choice among alternative scenarios, how much
more equivocal must be the expression, "The steps are determined by
the formula ... "! And this ambiguity bears crucially on the interlocu-
tory question which originally prompted the whole remark, namely,
"But are the steps then not determined by the algebraic formula?".
Wittgenstein has not denied that there is any sense in which a formula
can be said to determine the steps. (By analogy, he has not denied that
156 JULIET FLOYD
we are able to grasp rules for the use of a word.) He has instead
suggested that there is not a single notion of determination (or rule-
governing, or rule-grasping) one could negate (or, presumably, affirm)
unambiguously. Without something further the question "Is y = x
2
a
formula which determines y for a given value of x?" is simply unclear;
it has no reference to anyone essential thing. And this indicates the
root of what was "mistaken" in the objection the interlocutor voiced.
The last two of Wittgenstein's six responses 'are meant to indicate
that although the sentence "The formula y = x
2
determines the number
y for a given value of x" is (as we have just seen) off-hand misleading
and unclear, it is not meaningless. For, in a particular setting, the
question may have an answer, e.g.,
[0 ]ne might address this question to a pupil in order to test whether he understands the
use of the word "to determine"; or it might be a mathematical problem to work out
whether there was only one variable on the right-hand side of the formula, as e.g. in the
case: y = (x
2
+ Z)2 - z(2x
2
+ Z).40
In the first example we are to imagine a pupil who has been trained,
like the people in Wittgenstein's first two responses, to calculate with
individual formulas, and now we are trying to teach the pupil to use
the system of classification outlined in Wittgenstein's third response.
We may "test" whether he understands how to use the system of
classification, and in this case there would be nothing misleading in
asking the question, "Is y = x
2
a formula which determines y for a
given value of x?". We would simply be asking the student to classify
the formula according the system we are trying to teach.
Wittgenstein's final response sets a "mathematical problem" to work
out. A calculation shows that the variable "z" can be eliminated from
the (simplified) fromula.
41
In carrying out this calculation (or indeed,
in using formulas like "y = x
2
" , or classifying formulas in the previously
described way) we do not need to be clear about what the notions of
necessity, compulsion, or determination consist in. Rather, we know
what to do when asked.
The entire Section 1, in examining the expression "the formula deter-
mines the steps", investigates the concept of "determination" ["Bestim-
mung"]. But I have interpreted Wittgenstein's use of the expression,
and the examples he gives, to indicate an indeterminateness in the
interlocutor's use of this notion. So there is a tremendous irony here,
a play on the word "Bestimmung".42 On its own, Wittgenstein suggests,
the notion of "determination" makes no sense, and so offhand it is not
clear what to do with either the question "Is y = x
2
a formula which
determines y for a given value of x?", or the question "What sort of
determination is the determination of steps by a formula?". To sum
up, the interlocutor's "mistake" was to think that we have a determinate
sense of determination (or a determinate absolute notion of a function).
But then the idea that their is an univocal (metaphysical?) quality,
logical or mathematical necessity, which may be said, to compel us to
assent to a mathematical sentence (such as "4 = 2
2
,,) on penalty of
absurdity - a quality whose character is independent of our ability to
partake in particular mathematical practices - begins to crumble. The
example of an algebraic formula fails to exhibit a special kind of deter-
mination except in particular mathematical contexts. Similarly, revert-
ing to Wittgenstein's governing analogy between an algebraic formula
and a rule for the use of an expression, we may begin to suspect that
the "determination" of our linguistic steps by our understanding, if
conceived of as the grasping of a rule for the use of an expression, may
not be the monolithic or clear phenomenon we are tempted to treat it
as being when we wonder how it is that we can follow any rule at all.
The logical necessity we think of as inhering, in traditional terminology,
in the relation between a concept and its instances, and indeed, in the
unity of a judgment itself, begins to slip from our grasp.43
4.
Neither of Wittgenstein's final examples of uses of the expression "the
formula determines the steps" seems to secure what the interlocutor
was after. So, taken collectively, all six of Wittgenstein's responses to
his own question about use fail to give content to the interlocutor's
objection. This is Wittgenstein's point: the problem of necessity does
not seem to lie in either our uses of the word "determines" or the
carrying out of a solution to a particular mathematical problem. So
there is a difficulty in getting a general problem about logical compul-
sion or necessity off the ground. But the interlocutor tries again in
Section 2. Note, however, that in the opening question of Section 2
the interlocutor now uses "meinen", translated here as "to mean"; that
is, how a speaker intends or means to use an expression. This contrasts
with the discussion in Section 1, where one was tempted to talk about
the meaning (in the Fregean sense of Bedeutung, or Sinn) of the formula
158 JULIET FLOYD
itself. In German a formula [Formel] cannot, grammatically speaking,
"mean" [meinen], just as in English a formula cannot, literally "intend".
The interlocutor responds to Wittgenstein's play on the ambiguity of
"determines" as follows:
'The way the formula is meant [gemeintJ determines which steps are to be taken,.44
The interlocutor's idea here is this: the determination of each of my
particular steps, and, indeed, of the "mathematical" steps from input
to output, is secured or constituted by the particular intention I have
to use the formula and the particular meaning I endow or associate the
formula with. This is a natural suggestion. As a description or (begin-
ning) explanation of how the formula "determines the steps", it has the
appeal of seeming less crudely behaviorist than Wittgenstein's previous
suggestions about how we are trained to use the expression "The for-
mula determines the steps". "Rule" is, apparently, an intentional no-
tion insofar as it plays the role of a guide for action. But Wittgenstein
quickly presses the interlocutor on what it is to mean or intend the
formula "in a particular way". He asks
What is the criterion for the way the formula is meant? Presumably the way we always
use it, the way we were taught to use it. 45
The suggestion is that "meaning" or "intending" a formula "in a parti-
cular way" is itself an activity or feat accomplished by being able to
use the formula, having been trained in its use. Our "criterion" for
determining whether or not someone means the formula in a particular
way is what they do. Wittgenstein makes his point by looking at how
we incorporate into our language a sign unknown to us, but used by
someone who speaks our language. Thus he says:
We say, for instance, to someone who uses a sign unknown to us: "If by "x!2" you mean
x
2
, then you get this value for y, if you mean x, that one". - Now ask yourself: how
does one mean [meinenJ the one thing or the other by "X!2,,?46
Since the natural home of the notion of intending a formula in a
particular way is just the intention to use it in a particular way, the
answer to Wittgenstein's question presumably lies, so far as my view
of the person using the symbol "x!2" goes, in what he does, and for me,
in how I go on to use the symbol myself. I have no special experience of
"meaning 'x!2' the first way" when I do mean it; and I do not attribute
WITTG EN STEIN ON 2, 2, ... 159
such experiences to the other simply in virtue of interpreting his or her
actions in one way or the other. Wittgenstein emphatically declares:
That is how meaning it [das Meinen] can determine the steps in advance.
47
So once again the use, the training, come to the fore. Wittgenstein is
happy to give the interlocutor, at least for the moment, the notion
of "intention" or meaning. Insofar as he compares the "truths" of
mathematics and logic with rules of grammar, Wi'ttgenstein himself
assimilates questions about the special necessity of mathematics and
logic to broader questions involving the relation of intention, expec-
tation, and desire to their (particular) fulfillments. This is indeed part
of his challenge to the traditional picture of the a priori as a category
of knowledge designed to account for our knowledge of necessary truth,
for traditionally the relation of intention to action was hardly deemed
one of a priori knowledge.
48
So in addition to attempting to show the
unclarity of the notions of logical and mathematical necessity, Wittgen-
stein's discussion restructures and broadens the field of concepts to be
investigated in connection with the apparent necessity and full-blooded
truthfulness of logic. This shifts the traditional epistemological frame-
work for discussing the character of mathematics and logic. For Witt-
genstein, Platonism in mathematics makes no more and no less sense
than Platonism about meaning, desire, or intention. This is borne out
in Investigations Section 437, where he writes:
A wish seems already to know what will or would satisfy it; a proposition, a thought,
what makes it true - even when that thing is not there at all! Whence this determining
[Bestimmen] of what is not yet there? This despotic demand? ("The hardness of the
logical must. ,,)49
My reading of the first two sections of Remarks on the Foundations of
Mathematics indicates at least the beginnings of an answer to the ques-
tion whether Wittgenstein is ultimately offering a behaviorist or conven-
tionalist account of logical inference and mathematical necessity. 50 The
answer he would give to such a question would be another question,
viz., "Convention (or behavior) as opposed to what?". The opposed
notion ("nature", "mind", "inner" "feeling", etc.) may be seen, with
appropriate scrutiny, to suffer from the same unclarity as the notions of
"conventional" and "behavior" themselves. The other side of absolute
arbitrariness is absolute necessity, absolute determination or compul-
sion. And these are the notions I have argued Wittgenstein is trying to
160 JULIET FLOYD
unmask as unclear in the first two sections of Remarks on the Foun-
dations of Mathematics.
5.
Wittgenstein's emphasis on our criteria for ascribing intentions and his
example of the introduction of an unknown symbol into our language
fail to unearth what has struck the interlocutor. Obviously we can
define or find a meaning for the sign "x!2" in particular cases like the
one described. But this does not get at the peculiar compulsion, the
sense of what must happen, the idea that something must be fixed
ahead of our particular steps in virtue of the ways in which we mean,
or intend to mean, the formulas (or, for that matter, the words) we
are now operating with.
51
Wittgenstein's emphatic remark that "That
is how meaning it can determine the steps in advance" was intended
to call attention to the fact that our criteria for determining "the way"
someone intends or means an expression lie in what they (and we) do.
But now a natural response to this is to ask How is it that he (or J)
know what to do? Is Wittgenstein denying that we may say that someone
who understands an expression knows what to do with it ahead of time?
This is the kind of worry which prompts Wittgenstein to suggest, in
Section 3, that the interlocutor pose the following question to himself:
How do I know [Wie weiss ich] that in working out the series +2 I must write
and not
"20004,20006"
"20004,20008,,?52
Unlike the questions Wittgenstein posed to the interlocutor in Section
2, this first-person question is inflicted on a sign I already know how
to use ("+2"). Thus it may seem to demand a different sort of answer
than the one about "x!2". Indeed, in the course of the ensuing passage
Wittgenstein will try to elicit from his reader a sense that this is not
the straightforward question it might appear to be: construed in the
most natural way it is empty, that is, "How do I know ... " is out
of place here. What sort of "emptiness" is in question? In this case
Wittgenstein does not, nor do I for him, appeal to an absolute standard
of emptiness. (Wittgenstein had invoked something like such a standard
in Tractatus in connection with the doctrine of a "formal", or "in-
ternal", relation holding between the members of a "formal" series. )53
Still, the relation(s) between the members of a series such as +2 is
thought of by the later Wittgenstein as in some sense or another "gram-
matical".54 Our question is, What does this idea amount to in Remarks
on the Foundations of Mathematics?
One can, upon reflection, imagine circumstances in which "How do
I know to write ... ?" might have a point. For example, it might arise
in the writing of a step in an algebraic series whose generating principle
was especially complex or difficult (e.g., if I found myself computing
differently each time I tried to determine the correct next step). In this
case, "How do I know that in working out this series that this is the
correct next step?" might be answered with several calculations worked
out on paper. Similarly, a child learning arithmetic might reasonably
pose this question. On the other hand, the rule or law of the series
considered here (+2) is familiar and easy. I seem to be so certain that
"20004,20006" is the correct thing to write that a calculation would
seem superfluous, ad hoc - hardly informative about this particular
step. This is part of Wittgenstein's point.
The interlocutor continues to feel that Wittgenstein is denying some-
thing. He responds to the question "How do I know that in working
out the series +2 I must write "20004 ,20006" ?" , as if Wittgenstein is
endorsing a skeptical view.
But you surely know for example that you must always write the same sequence of
numbers in the digits: 2,4,6,8,0,2,4, etc.
The interlocutor seeks to find some basis or justification for writing
"20006" at the 1O,003d step of the series. He locates his confidence in
the repeating pattern of right-most digits of numerals in the series, thus
hoping to circumvent worries about the infinite length of the series
and his (apparently) unlimited capacity to continue writing different
numerals in the correct order. Wittgenstein accepts this response as
"Quite true!" ("Ganz richtig!", "Completely correct!"), and then,
pushing the interlocutor's idea one step further, turns the purported
answer against him:
The problem must already appear in this sequence, and even in this one: 2,2,2,2, etc. -
For how do I know that I am to write "2" after the five hundredth "2"? And if I know
it in advance, what use is this knowledge to me later on? I mean: how do I know what
to do with this earlier knowledge when the step actually has to be taken?
What is "the problem"? A regress arises because the interlocutor
162 JULIET FLOYD
supposes that he knows something in advance of the step's being taken,
something different from the taking of the step itself. He supposes that
he holds a belief which leads him to act. The case of the recalcitrant
pupil (PI Sections 143, 185) is brought home to the first person:
Wittgenstein's question places a seeming gap between my knowledge
of a general principle and its application, thus inviting a "Third Man"
argument which asks, each time I try to apply my knowledge, how I
know that this application is correct. At each particular step, when I
try to say how I know that this is the next correct step in the series,
my answers seem thinner than my conviction. And if there is trouble
with the series 2,2,2 ... , then will there not be trouble, i.e., trouble
in saying how I know, in every case? Wittgenstein's new question
("How do I know that I am to write "2" after the SOOth "2"?") seems
to raise a general wonder about how knowing the principle of any series
tells me what to do in the particular case. How can I follow any rule
at all?55
In fact, however, the example seems to just as easily cut the other
way. Just because there is no problem (other than boredom or exhaus-
tion or pointlessness) with my writing the SOOth "2" in the series
2, 2, 2 ... , it seems foolish to allow such worries to creep in about the
1O,003d step in the series +2 - or at least, to view such questions as
carrying the same force with them as grammatically analogous "How
do I know ... ?" questions raised about matters of genuine arithmetic
or empirical difficulty. Admittedly, Wittgenstein's reduction of "the
problem" to the case of 2, 2, 2 . .. draws a somewhat daring analogy
between the series +2 and the series +0 - if that is how we want to
view my writing of 500 "2s". The series + 0 is a limiting case, requiring,
as we should ordinarily say, no calculation at all from step to step. In
fact, it appears that nothing specifically mathematical is required to
continue; i.e., no special kind of knowledge that, say, my writing out
of a series of repeating color hues, or shapes, or repeating letter a's
(a, a, a, a ... ) does not require. I have an idea in such cases that I write
"the simplest" or "the most natural" pattern,56 but this does not pro-
vide me with knowledge of something specific which compels me at
this particular step. The interlocutor's attempt to answer the original
question about "20004,20006" directly by insisting that I ought to do
"the same" (writing digits in the pattern 2,4,6,8,2,4 ... ) evidently
presupposes the knowledge or capacity in question. Thus it yields no
independently informative answer. This is because knowing that
"20006" is the numeral to write in order to "go on in the same way"
at the 1O,003d place is just (part of) what it is to know that this is the
series "+2" starting with 0.
57
Suppose 1 exhibited confusion at the
10,003d step. To get this far successfully and then hesitate would proba-
bly look most like exhaustion, not error. But if it did not seem to be
a mental seizure of some kind (e.g., 1 cheerfully chew on my pencil,
giving every indication of thinking hard) 1 would, barring special cir-
cumstances, (be said to) not understand the series. ;So far as other
people's views of me goes, my knowledge of the general and knowledge
of the particular go hand in hand here. And even for myself, my
wondering how 1 know what to do in the particular case is inseparable
from wondering whether 1 grasp the series itself. But the "How do 1
know ... ?" question, as phrased, makes it look like my knowledge of
the series has to hook up with, or fit into, my knowledge of the parti-
cular step. It is as if what the rule actually consists in is two things: the
general principle and the atomic points in reality which fit its demands.
However, while 1 can memorize a formula and call it "the formula of
a series", 1 am not without further ado said to understand the formula
unless 1 can operate with it. And if 1 operate with the series well enough
to convince myself, my teachers and other speakers that 1 have grasped
the principle, what more is needed? The question about what to do at
the 1O,003d (or the SOOth) step, if viewed as opening up a doubt of
some kind, begins to look absurd, like the question "How do 1 know
that this thing is identical with itself?". One is tempted to say that this
is simply what working out the series +2 (or what identity) is, amounts
to, or comes to. This will not give me an informative answer to the
question or a justification for taking this particular step in the way 1
do - much less a general defense of the practice of calculating with
ruleS.
58
Nothing (in particular) does. But so far, nothing is needed.
"How do 1 know that in writing out 500 "2s" 1 should write "2" at
the SOOth step?". The question is empty - unless it is taken as a
neurophysiological question, or a question about purported "mental
machinery". (I shall discuss this way of giving the question content
below.) Hence Wittgenstein's parenthetical remark,
(If intuition is needed to continue the series + 1, then it is also needed to continue the
series +0.)59
Wittgenstein wants to undermine both the idea that a justification is
required for each particular step in the number series and that in the
164 JULIET FLOYD
absence of such justification a special appeal to intuition is the only
recourse.
60
But the interlocutor reacts to Wittgenstein's move to 2,2,2 ... as if
Wittgenstein is denying that he knows what to write at the 1O,OO3d step
of the series + 2:
But do you mean to say that the expression "+2" leaves you in doubt what you are to
do e.g. after 2004?61
Wittgenstein reacts sharply:
- No; I answer "2006" without hesitation. But just for that reason [aber darum] it is
superfluous [ilbellilssig] to suppose that this was determined earlier on. My having no
doubt in face of the question does not mean that it has been answered in advance.
We need to get to the bottom of this "superfluousness". Wittgenstein-
says that he answers "2006" "without hesitation", "ahne Bedenken",
without thinking over or considering anything. The analogy between
the step to "2004" from "2002" and the SOOth 2 in the series 2, 2, 2 ...
helps make Wittgenstein's point clearer. In the case of 500 2s, there is
a real sense in which no thought can be brought to bear at the SOOth
step. Something rote governs the process. Unlike an initial segment of
a series whose principle requires some kind of calculation, 2, 2, 2 ...
requires no guess, no inference, no thought at all between steps. My
actions may become mechanical and they may remain mechanical. So
Wittgenstein seems to be saying, with the writing of "20006" at the
1O,OO3d step of working out the series +2.
Of course my certainty that "20004" is the correct thing to write at
this point is supposed by the interlocutor to reflect the fact that some-
thing was "answered in advance" of the steps actually being taken. He
objects, still supposing that Wittgenstein is denying that he knows what
to write:
But I surely also know that whatever number I am given I shall be able, straight off and
with certainty, to give the next one.
to which Wittgenstein responds,
- Certainly my dying first is excluded, and a lot of other things too. But my being so
certain that I will be able to go on is naturally very important.
62
Wittgenstein does not deem it irrelevant, much less false, for the inter-
locutor to fix on the certainty with which I develop the series under
WITTG ENSTEIN ON 2, 2, ... 165
certain circumstances.
63
But the certainty, Wittgenstein says, is a cer-
tainty that I shall be able to do something.
64
I have been trained to
react with confidence, without doubt, to such rules as +2. This is indeed
of the utmost importance, this rote ness of response, but the need
for further attempts to justify it on independent and specific grounds
("grounds" more specific than "I have learned arithmetic") has not yet
been made clear. 65
The interlocutor's idea is that something is known; "in advance" of
the actual writing of the step. His reaction to Wittgenstein's remarks
is that something is being denied, some kind of knowledge it seems to
make sense to say we possess, knowledge that explains how it is we
know what to do. However, it is not yet clear what the notion of "in
advance" adds to the idea of being "determined". If our steps are
"determined", it is natural to suppose something has fixed them ahead
of time; this is just what being "determined", presumably, comes to.
Wittgenstein is interested in the idea that something is "known ahead
of time" insofar as it expresses the a priori. But Wittgenstein's earlier
investigations of "determination" and his suggestion that "knowledge"
is out of place (or anyway playing an idiosyncratic role) in the opening
question of Section 3 serve to undermine the traditional category of
the a priori entirely - at least insofar as it is viewed, in Kant's manner,
as the epistemic companion to necessity. If "knowing" is out of place
at the 1O,003d step of working out the series +2, then (as the reductio
ad absurdum of Section 3 shows) the urge to say that I (know to) write
"20006" because it is a mathematical fact (or: mathematically true) that
20006 follows 20004 in the series + 2 is undercut. There is no question
of anything known here - it is rather a question of what is done.
66
Having no doubt and being able to say how I know in this sort of
case pull in opposite directions, in part because my having no doubt is
seen (by the interlocutor) to be a matter of my being "determined".
The "determination" must be determination, in this case, of a very
particular sort. Let us try to make out a sense of genuine determination
in advance.
I have said that something "mechanical" emerges in "2, 2, 2 ... ". In
what sense might we think of a machine determining the steps of this
series in advance? Consider, by way of analogy, a computer, pro-
grammed to both add 2 and to write a sequence of repeating "2s"
indefinitely. In a sense, the program (plus the machine's hardware)
"determines" the steps it will take "in advance". Now we could ask,
166 JULIET FLOYD
"How does the machine know to write "2" at the SOOth step?" An
answer would presumably specify the program and the details of how
the hardware implements the program. (The computer might even be
programmed to "answer" the (first person) question "How do I know
to "write" ... ?") But granted the program and the specification of the
hardware's implementation of it, there is no further question to ask
about this machine's taking this step at this point. The analogy with
Wittgenstein's suggestion in Section 3 is this: once my program (my
training in arithmetic) and the implementation of my hardware (by my
neurophysiological nature) have been specified, it seems absurd to think
that there could be more to ask about why I write "20006" (or "2")
here. Thus, although it may be an interesting neurophysiological or
psychological question what it is, e.g., that allows me to write "2" at
the 1st, the 10th, and the SOOth step of 2, 2, 2, ... , or what the state
of my retina and brain are when I watch a series of "2s" go by, one
after the other, first in this sort of light, then after food deprivation,
etc., these are still very particular and difficult questions with (if they
make scientific sense) very particular answers, answers which would
yield no general conclusions about the kind of determination the inter-
locutor is interested in.67
Wittgenstein offers the interlocutor other examples of epistemic com-
pulsion, or genuine knowing in advance in Sections 21 and 22 of Re-
marks on the Foundations of Mathematics. These sections are especially
concerned with the Platonism of Frege and Russell, as the remark of
Section 21 makes clear.
21. In his fundamental law
68
Russell seems to be saying of a proposition: "It already
follows - all I still have to do is, to infer it". Thus Frege somewhere says that the straight
line which connects any two points is really already there before we draw it; and it is the
same when we say that the transitions, say in the series + 2, have really already been
made before we make them orally or in writing - as it were tracing them.
69
Wittgenstein now replies to this patent Platonism by giving several
further examples of "determination".
22. One might reply to someone who said this: Here you are using a picture. One can
determine [bestimmen] the transitions [Ubergiinge] which someone is to make in a series,
by doing them for him first. E.g. by writing down in another notation the series which
he is to write, so that all that remains for him to do is to translate it; or by actually
writing it down very faint, and he has to trace it. In the first case we can also say that
we don't write down the series that he has to write, and so that we do not ourselves
make the transitions of that series; but in the second case we shall certainly say that the
series which he is to write is already there. We should also say this if we dictate what he
has to write down, although then we are producing a series of sounds and he a series of
written signs. It is at any rate a sure way of determining [bestimmen] the transitions
[Obergiinge] that someone has to make, if we in some sense make them first.
70
Thus Wittgenstein has given us several ways to determine the transitions
"in advance". But he immediately returns to the lessons of Sections 1,
2 and 3, writing:
,
If, therefore, we determine [bestimmen] these transitions [Obergiinge] in a quite different
sense, namely, by subjecting our pupil to such a transition as e.g. children get in the
multiplication tables and in multiplying, so that all who are so trained do random
multiplications (not previously done in the course of being taught) in the same way and
with results that agree - if, that is, the transitions which someone is to make on the order
"Add 2! " are so determined by training that we can predict with certainty how he will
go, even when he has never up to now taken this step [Obergang]- then it may be natural
to us to use this as a picture of the situation: the steps [Obergiinge] are all already taken
and he is just writing them down.71
But now our picture - the Platonist picture - is no longer literally true.
Relative to these examples, it is only a metaphor. In the case of our
actually writing out the series faintly, and then asking the pupil to trace
what we have already marked out, there is a perfectly good sense in
which the pupil, may make a mistake (perhaps by a slip of the pen).
As the pupil copies, we can see the transitions or steps that are made,
from faint numeral to darker numeral. If we ask the pupil to ask him
or herself, "How do I know to write "20004,20006"?" there is a per-
fectly good answer the pupil can give: simply point to the faint marks
that look just like this ("20004,20006"). But in the cases we were
imagining earlier in Section 3, there was nothing to point to. The
"picture" the interlocutor is operating with is that the case we described
is just like the pupil's tracing of the faint numerals, only the numerals
are very, very faint - in fact so faint they cannot be seen. But now the
notion of a "transition" or "step" is metaphorical, a Platonic entity of
some kind, or a whispering of the entity to us to take this step. 72
Wittgenstein indicates in Section 3 that in order to be "answered in
advance" there must be some question, some possible mistake, some
particular way of getting it wrong. But rote learning - like the kind of
drills lying behind our "training" in arithmetic - is a kind of memori-
zation ahead of time. This is not a matter of "answering questions in
advance", but of practicing so as to not have to think. As we have
seen, something "mechanical" is happening by the SOOth step of
168 JULIET FLOYD
2, 2, 2 ... This is what leads to our not being sure whether to call the
step at that point calculation, or inference, or even whether I may be
said to have "intended" this step (or not intended it). It is no longer
clear what sort of "necessity" is in question, or how we are to demarcate
the role and scope of the notions of logical or mathematical inference.
Let me back up for a moment and discuss a parenthetical analogy
Wittgenstein draws in Section 3. After the posing of the original ques-
tion ("How do I know that in working out the series +2 I must write
"20004,20006" and not "20004,20008"?"), Wittgenstein writes:
- (The question: "How do I know [Wie weiss ich] that this color is "red"?" is similar.)
Now in Philosophical Investigations Section 381 we read the following:
How do I recognize [erkenne] that this color is red? - It would be an answer to say: "I
have learned English". 73
Wittgenstein's response in Philosophical Investigations that "I have
learned English" is, it may seem, hardly an answer at all. In fact, it is
tantamount to a rejection of the question, at least as an intended
articulation of the interlocutor's sense of the mystery and uniqueness
of the meaning of "red". For the interlocutor did not doubt that he
spoke English. If he had, he would have pictured himself as a foreigner
or a child; but this is part of Wittgenstein's point: what the foreigner
and the child lack is the training, the practice that we have engaged in.
Posing the question in the way the interlocutor does can make it seem
as if there is something special going on when I recognize that this
color is "red", some particular understanding that is manifested, some
knowledge of a particular and unique meaning relation between the
word and the thing. Yet when I ask myself how I know ... , there
seems nothing particular to pinpoint in virtue of which I may claim to
know. The result may be a kind of skeptical giddiness about meaning,
but Wittgenstein offers instead the fact that I speak English. This
squares with our untutored reactions: if I ask myself the question,
"How do I now this color is "red"?" I am tempted to respond, "Because
it's red!,,74
These may seem to be irrelevant and impoverished responses if taken
on their own. Certainly they are "circular" if taken as direct answers
to the questions, but viciously so? From what standpoint could we find
a better sort of response?
Thus, using Wittgenstein's parenthetical analogy, we may rethink the
opening question of Section 3. An analogous response to the question
about my compulsion to write "20004,20006" would not be that I speak
English; it would presumably be something like "I have learned basic
arithmetic" or "I have learned to count". This response hardly informs
me about this particular step (from "20004" to "20006"). In a real
sense, it rejects the question. But it serves as a kind of reminder (cf. PI
Section 127). Wittgenstein wishes to call attention to the interlocutor's
inclination to view as irrelevant the fact that I have learned arithmetic.
Why does this seem irrelevant? What more could the interlocutor want
by way of a justification?75
NOTES
1 Eds. G. H. von Wright, R. Rhees and G. E. M. Anscombe; trans. G. E. M. Anscombe
(revised edition, Cambridge: M.LT. Press, 1978). Hereafter cited as RFM. Subsequent
references to Wittgenstein's other works will be as follows: Tractatus Logico-Philosoph-
icus, trans. C. K. Ogden (London: Routledge & Kegan Paul, 1933) is cited as TLP;
Philosophical Investigations, trans. G. E. M. Anscombe (Oxford, Basil Blackwell, 1958)
as PI; the "Frilhversion" of Philosophical Investigations (discussed below in Section III)
as FV; the "Mittelversion" of Philosophical Investigations (discussed below in Section
III) as MV; Wittgenstein's 1939 Lectures on the Foundations of Mathematics, ed. Cora
Diamond (Ithaca: Cornell University Press, 1976) as LFM; Remarks on Color, ed. G. E.
M. Anscombe, trans. Linda L. McAlister and Margarete Schiittle (Berkeley: University of
California Press, 1977) as ROC; On Certainty, eds. G. E. M. Anscombe and G. H. von
Wright, trans. G. E. M. Anscombe and Denis Paul (New York: Harper & Row, 1972)
as OC; Philosophical Grammar, ed. Rush Rhees, trans. Anthony Kenny (Oxford: Basil
Blackwell, 1974) as PG; Wittgenstein and the Vienna Circle, ed. Brian McGuinness,
trans. Joachim Schulte and Brian McGuinness, (Oxford: Basil Blackwell, 1979) as WVC;
and Zettel, eds. G. E. M. Anscombe and G. H. von Wright, trans. G. E. M. Anscombe
(Berkeley: University of California Press 1967) as Z.
2 The early reviews of RFM were particularly scathing. See, e.g., A. R. Anderson,
'Mathematics and the "Language Game''', Review of Metaphysics 11, 446-58; G. Kreisel,
'Wittgenstein's Remarks on the Foundations of Mathematics', British Journal of the Philos-
ophy of Science 9, 135-58; Paul Bernays, 'Comments on Ludwig Wittgenstein's Remarks
on the Foundations of Mathematics', Ratio 2, 1-22.
More recently there has been more willingness to countenance Wittgenstein's views,
though little consensus on what those views are or even how ignorant Wittgenstein was
of certain central technical results. See, e.g., Michael Dummett, 'Wittgenstein's Philos-
ophy of Mathematics', in Truth and Other Enigmas (Cambridge, MA: Harvard University
Press, 1978), pp. 166-85; Charles Chihara, 'Wittgenstein's Analysis of the Paradoxes in
his Lectures on the Foundations of Mathematics', Review 86, 365-81;
Crispin Wright, Wittgenstein on the Foundations of Mathe'inatics (London: Duckworth,
1980); V. H. Klenk, Wittgenstein's Philosophy of Mathematics (The Hague: Martinus
Nijhoff, 1976); Saul Kripke, Wittgenstein on Rules and Private Language (Cambridge,
170 JULIET FLOYD
MA: Harvard University Press, 1982); S. G. Shanker, Wittgenstein and the Turning Point
in the Philosophy of Mathematics (Albany, NY: State University of New York Press,
1987); and G. P. Baker and P. M. S. Hacker, Wittgenstein: Rules, Grammar and Necessity,
Vol. 2 of an Analytical Commentary on the Philosophical Investigations (Oxford: Basil
Blackwell, 1985). I have learned from all these authors, but none of them seems to see
Wittgenstein's later dialectical style of writing as internal to his conception of logic and
grammar in the way I outline in this essay.
3 PI Section 127.
4 PI Section 110:
"Language (or thought) is something unique" - this proves to be a superstition (not
a mistake!), itself produced by grammatical illusions.
And now the impressiveness retreats to these illusions, to the problems.
5 See LFM, Lecture I (p. 14):
Mathematicians tend to think that interpretations of mathematical symbols are a lot
of jaw - some kind of gas which surrounds the real process, the essential mathematical
kernel. A philosopher provides gas, or decoration - like squiggles on a wall of a
room.
I may occasionally produce new interpretations, not in order to suggest they are
right, but in order to show that the old interpretation and the new are equally
arbitrary. I will only invent a new interpretation to put side by side with an old one
and say "Here, choose, take your pick." I will only make gas to expel old gas.
6 Cf. Warren Goldfarb, 'I Want You to Bring Me A Slab: Remarks on the Opening
Sections of the Philosophical Investigations', Synthese 56 (1983): 265-82, especially pp.
266-7.
7 A critical edition of both parts of this so-called "Friihversion" (hereafter FV), has been
prepared by G. H. von Wright and Heikki Nyman. The first part of FV was composed
during 1936; the second during the fall of 1937 (see pp. 113-4 of von Wright, 'The Origin
and Composition of Philosophical Investigations' (in his Wittgenstein, (University of
Minnesota Press, 1983, pp. 111-36. The parts are of equal page length and form
a consecutive typescript (see Ibid., p. 117). The so-called "intermediate version", or
"Mittelversion" of PI (MV) was compiled by von Wright, who conjectures that it was
completed by Wittgenstein some time in 1945 (see von Wright, The Origin and Composi-
tion of Philosophical Investigations', pp. 125ff).
8 von Wright, The Origin and Composition of Philosophical Investigations', pp. 120ff.
9 von Wright, 'The Origin and Composition of the Investigations', p. 121.
10 In the typescript from which RFM, I was printed (TS 222) approximately 50 passages
were deleted and the order of others rearranged from the FV version (TS 221). Anhang
III of von Wright's critical edition gives a list of the correspondences and alterations
between TS 221 and TS 222.
11 See von Wright, 'The Origin and Composition of the Investigations', pp. 114-19.
According to von Wright, there were originally three copies of the second part of the
typescript. One contained a few revisions in Wittgenstein's hand; the second was entirely
clean; the third Wittgenstein cut up into pieces which he "clipped together in a large
number of bunches" many of which are annotated and changed (Ibid., p. 118). Part I of
RFM was printed from this third copy, entitled TS 222 in von Wright's catalogue. In the
editors' preface to the revised edition of RFM, p. 30, we read that "in a notebook as
late as 1944 [Wittgenstein) proposed a few alterations to this typescript".
12 In Volume II of their Analytical Commentary on the Philosophical Investigations
(Chapter I), Baker and Hacker emphasize the manuscript basis of PI and view PI Sections
1-189 as a tree which bore "two fruits": in the original sequel (FV) Wittgenstein focussed
on the philosophy of mathematics and logic; in the intermediate and final versions (MV
and PI) on the philosophy of mind. Developing this metaphor, Baker and Hacker suggest
several interesting similarities between Wittgenstein's discussions of mathematics and his
discussions of mind. But the complexity of the i'llterweaving I find among Wittgenstein's
remarks in RFM I and later sections of PI does not seem to me to bear out the metaphor
of "two fruits", which suggests that there are two distinct yet comparable realms of
Wittgenstein's later philosophy, distinguished by rali\ge of topic.
13 RFM I Section 1.
14 In an earlier draft of PI Section 81 (FV Section 78.(80., Wittgenstein explicitly
mentions Tractatus, writing in the final sentence:
Denn dan wird auch klar werden, was dazu verleiten kann - und mich verleitet hat
(Log.Phil.Abh.) - zu denken, dass, wer einen Satz ausspricht und meint, oder versteht,
damit einen Kalkiil betreibt, nach bestimmten Regeln.
15 PI Section 81. I have altered Anscombe's translation in the fifth, ninth and final lines
of the published English text. Her translation suggests that games and calculi which have
fixed rules must be different, and leaves out the word "logic" at one point altogether.
16 Despite occasional remarks which suggest otherwise (e.g., PI Section 7, RFM VI
Section 28, RFM VII Section 35), my interpretation is reinforced by PI Sections 130-1:
130. Our clear and simple language-games are not preparatory studies for a future
regularization of language - as it were first approximations, ignoring friction and air-
resistance. The language-games are rather set up as objects of comparison which are
meant to throw light on the facts of our language by way not only of similarities,
but also of dissimilarities.
131. For we can avoid ineptness or emptiness in our assertions only by presenting
the model as what it is, as an object of comparison - as, so to speak, a measuting-
rod; not as a preconceived idea to which reality must correspond. (The dogmatism
into which we fall so easily in doing philosophy.)
17 PI Section 143.
18 Ibid.
19 See PI Sections 143, 185.
20 PI Section 144:
144. What do I mean when I say "the pupil's capacity to learn may come to an
end here"? Do I say this from my own experience? Of course not. (Even 'if I have
had such experience.) Then what am I doing with that proposition? Well, I should
like you to say: "Yes, it's true, you can imagine that too, that might happen
too!" - But was I trying to draw someone's attention to the fact that he is 'capable
of imagining that? - I wanted to put that picture befote hiIJI, and his acceptance of
172 JULIET FLOYD
the picture consists in his now being inclined to regard a given case differently: that
is, to compare it with this rather than that set of pictures. I have changed his way of
looking at things. (Indian mathematicians: "Look at this".)
The FV basis for this section (FV Section 126 (128, identical with Z Section 461, draws
a connection between geometrical figure construction, Wittgenstein's own philosophical
method, and discussions of the convincingness of proof in RFM.
21 Stanley Cavell emphasizes the quality of the analogy as a "figure of speech" in The
Claim of Reason (New York: Oxford University Press, 1979). In discussing the function
of mathematical examples in PI he writes (pp. 121-2):
Their general background is an idea that the primitive abilities of mathematics (e.g.,
counting, grouping, adding, continuing a series, finding quantities equal or smaller)
are as natural as any (other) region of a natural tongue, and as natural as the primitive
abilities of logic (e.g., drawing an inference, following a rule of substitution). The
implication is that ordinary language no more needs a foundation in logic than
mathematics does. More specifically, he uses the picture of "continuing a series" as
a kind of figure of speech for an idea of the meaning of a word, or rather an idea
of the possession of a concept: to know the meaning of a word, to have the concept
titled by the word, is to be able to go on with it into new contexts - ones we accept
as correct for it; and you can do this without knowing, so to speak, the formula
which determines the fresh occurrence, i.e., without being able to articulate the
criteria in terms of which it is applied. If somebody could actually produce a formula,
or a form for one, which generated the schematism of a word's occurrences, then
Wittgenstein's idea here would be more than a figure of speech; it would be replaced
by, or summarize, something we might wish to call the science of semantics.
22 PI Sections 146, 152.
23 Elsewhere, however, Wittgenstein plays on the fact that it is usual in mathematical
contexts to speak of the formula (e.g., (a + b) + c = a + (b + c and the rule (e.g., of
associativity of addition) interchangeably. This comes out strikingly in his discussion of
recursive proof, Skolem primitive recursive arithmetic and mathematical induction in PG
Part II, Chapter IV.
24 Viewing concepts as rules, Kant raised a similar problem in the Critique of Pure
Reason (A 133/B 172-A 136/B 175) and attempted to resolve it in the schematism and
in the introduction to the Critique of Judgment.
2S See Tractatus 6.126ff. Compare Frege's critical discussion of Schroder in his review of
Schroder's Algebra der Logik, translated in Translations from the Philosophical Writings
of Gottlob Frege, eds. Peter Geach and Max Black (Totowna, NJ: Rowman & Littlefield,
1980) pp. 86-106. See also Russell's implicit criticisms of the algebra of logic tradition
in his Principles of Mathematics (New York: W. W. Norton & Co., 1938), Chapter II.
26 RFM I Section 6.
27 See both the Friihversion (FV, p. 162.(168. and PI Section 189.
28 Ibid.
29 This portion of the passage is identical in FV Section 162.(168.) and RFM I, 1. In PI
Section 189 the first word of the question is italicized: "How is it used?".
30 RFM I Section 1.
31 Wittgenstein's play on the ideas of instruction ("unterrichten"), training ("abrichten"),
teaching ("lehren") and education, or up-bringing ("erziehen") in. connection with the
learning of language are of course familiar themes in the opening sections of PI, especially
Sections 6, 7 and 9.
32 RFM I Section 1.
33 PI Section 206 uses a similar analogy in a different context to make a somewhat
different point:
206. Following a rule is analogous to obeying an order. We are trained to do so;
we react to an order in a particular way. But what if one person reacts in one way
and another in another to the order and the training? Which ooe is right?
Suppose you came as an explorer into an unknown country with a language quite
strange to you. In what circumstances would you say that the people there gave
orders, understood them, obeyed them, rebelled against them, and so on?
The common behavior of mankind [Die gemeinsame menschliche Handlungsweise]
is the system of reference [das Bezugssystem] by means of which we interpret an
unknown language.
34 Nor does it appear in the final version PI Section 189.
35 All people, every time? Human frailty, such as it is, would eventually bring errors or
exhaustion to falsify this claim. The possibility of mistakes in calculation versus uses of
alternative "arithmetics" is a related and complex topic investigated by Wittgenstein at
length in RFM and elsewhere. See, e.g., RFM I Section 136; RFM III Section 81; RFM
IV Section 26, and ROC III, Section 293.
Kripke offers an especially incisive criticism of dispositionalist accounts of meaning by
focussing on the difficulty of stating non-question-begging ceteris paribus clauses to pin-
point an explanation of behavior. See Kripke, Wittgenstein on Rules and Private Lan-
guage, pp. 25ff.
36 RFM I Section 1.
37 One might have dismissed the conversation between the interlocutor and Wittgenstein
from the beginning by dismissing the interlocutor's initial objection as rooted in a (rather
obvious) confusion of use and mention. But such a dismissal would skirt the philosophical
interest of Wittgenstein's writing: to plumb the sources of the interlocutor's interest in
the necessity of logical inference.
38 RFM I Section 1. I have altered Anscombe's translation of the first clause quoted.
39 Here I am differing in point of emphasis with both Garth Hallett (A Companion to
Wittgenstein's Philosophical Investigations, (Ithaca: Cornell University Press, 1977), pp.
278-9), and Baker and Hacker (Wittgenstein, Rules, Grammar and Necessity, An Analyti-
cal Commentary on the Philosophical Investigations, Vol. II, pp. 107ff). These com-
mentators label the latter kind of statement "grammatical".
40 RFM I Section 1.
41 Strangely, the simplified formula is y = X4, rather than y = x
2
, which would have made
better sense. Perhaps Wittgenstein made an algebraic error, or there was a typographical
error in transcnptlOn. If the formula to be simplified were instead y =
(x + Z)2 - z(2x + z), the simplification would be y = x
2
. Compare PI Section 189, where
the altered passage reads in part:
or it might be a mathematical problem to prove in a particular [bestimmten] system
that x has only one square.
174 JULIET FLOYD
Note the substitution of the notion of proof for calculation in PI Section 189.
42 This is borne out in the revision of the section's final sentence Wittgenstein made in
PI, quoted in the previous footnote.
43 RFM VII Section 42:
When I said that the propositions of mathematics determine rbi/den] concepts,
that is vague; for "2 + 2 = 4" forms a concept in a different sense from "p J p".
"(x).fx :::> fa", or Dedekind's Theorem. The point is, there is a family of cases.
And again, at RFM VII Section 45:
The word "concept" is too vague by far.
Cf. RFM VII Section 70 and RFM VII Section 71: "'Concept' is a vague concept".
Wittgenstein's discussion in RFM I Section 1 should be compared with the "indetermi-
nateness" [Unbestimmtheit] of our notion of sameness of color in ROC I Sections 17, 56
and 59; and ROC III Section 251.
44 RFM I Section 2. See the corresponding PI Section 190.
45 RFM I Section 2.
46 Ibid.
47 Ibid. I have altered Anscombe's translation of this sentence.
48 The tradition, of course, tended not to think of the relation as primarily epistemic. If
it is thought of as an epistemic relation at all, such a relation would not presumably be
known by means of the senses. (Compare G. E. M. Anscombe's idea of "non-obser-
vational knowledge" in her Intention (Ithaca: Cornell University Press, 1963) pp. 12ff.)
Kant, who viewed the will as a kind of causality, would have discerned a (causally)
"necessary connection" between an intention and an action, hence a relation with an a
priori element. But the intention itself would not have been deemed knowledge a priori
of the effect (the action); rather, the (regulative) category of causality would give us a
priori knowledge that the effect (the movement of my body, say) had some cause or
other in the realm of appearance.
49 The entire surrounding discussion in PI is relevant, especially Sections 435 and 437
where Wittgenstein says:
457. Yes: meaning [meinen] something is like going up to someone.
(This section was drafted much earlier in Wittgenstein's life. See PG p. 157.) In PI
Section 439 Wittgenstein explicitly speaks in terms of "metaphor".
50 Wittgenstein does his best to explicitly avoid being read this way. See, e.g., RFM II
Section 61, LFM, p. 111 and PI Sections 307-8:
307. "Are you not really a behaviorist in disguise? Aren't you at bottom really
saying that everything except human behavior is a fiction?" - if I do speak of a
fiction, then it is of a grammatical fiction.
308. How does the philosophical problem about mental processes and states and
about behaviorism arise? - The first step is the one that altogether escapes notice.
We talk of processes and states and leave their nature undecided. Sometime perhaps
we shall know more about them - we think. But that is just what commits us to a
particular way of looking at the matter. For we have a definite concept of what it
means to learn to know a process better. (The decisive movement in the conjuring
trick has been made, and it was the very one that we thought quite innocent.) -
And now the analogy which was to make us understand our thoughts falls to pieces.
So we have to deny the yet uncomprehended process in the yet unexplored medium.
And now it looks as if we had denied mental processes. And naturally we don't want
to deny them.
51 Cf. PI Section 188:
Here I should first of all like to say; your idea was that that act of meaning the order
[Meinen des Befehls] had in its own way already traversed all 'those steps: that when
you meant it your mind as it were flew ahead and took all the steps before you
physically arrived at this or that one.
Thus you were inclined to use such expressions as: "The steps are really already
taken, even before I take them in writing or orally or in thought". And it seemed
as if they were in some unique way predetermined, anticipated - as only the act of
meaning [das Meinen] can anticipate reality.
52 RFM I Section 3. At PI Section 211 the question finds an analog:
211. How can he know how he is to continue a pattern by himself - whatever
instruction you give him? - Well how do I know? - if that means "Have I
reasons?" the answer is: my reasons will soon give out. And then I shall act, without
reasons.
53 See TLP 4.1252,4.1273, 6.02ff.
54 See, e.g., RFM I Section 128, RFM 1I Section 27, RFM III Sections 26,39, RFM VI
Section 27.
55 This is, essentially, the interpretive route taken by Kripke, whose much discussed
Wittgenstein on Rules and Private Language outlines an interpretation of Wittgenstein as
a skeptic about the notion of meaning a word in a particular way. Kripke regards RFM
I Section 3 as a vivid statement of the paradox he attributes to Wittgenstein. Says Kripke
(p. 20):
The passage strikingly illustrates a central thesis of this [i.e., Kripke's] essay: that
Wittgenstein regards the fundamental problems of the philosophy of mathematics
and of the "private language argument" - the problem of sensation language - as
at root identical, stemming from his paradox. The whole of Section 3 is a succinct
and beautiful statement of the Wittgensteinian paradox; indeed the whole initial
section of Part I of Remarks on the Foundations of Mathematics is a development
of the problem with special reference to mathematics and logical inference.
I cannot take up my systematic differences with Kripke in this essay. Reasonable doubts
about his argument as interpretation of Wittgenstein seem to me to have already been
raised by several commentators, including G. E. M. Anscombe, 'Critical Notice of Saul
Kripke's Wittgenstein on Rules and Private Language', Canadian Journal of Philosophy,
15 (March 1985), 103-9; G. E. M. Anscombe, 'Review of Wittgenstein on Rules and
Private Language', Ethics 95 (January 1985), 342-52; Baker and Hacker, Scepticism,
Rules and Language, (Oxford: Basil Blackwell, 1984); Stanley Cavell's second Carus
Lecture, 1988; Warren Goldfarb, 'Kripke on Wittgenstein on Rules', Journal of Philo-
176 JULIET FLOYD
sophy 82 (1985), 471-8; and William Tait, 'Wittgenstein and the "Skeptical Paradoxes"',
Journal of Philosophy 81 (1984), 759-78.
56 RFM IV Section 9 suggests that the appeal to "same" occurs most naturally when
trying to explain a rule to someone:
How could one explain to anybody what you have to do if you are to follow a rule?
One is tempted to explain [Man ist versucht, zu erkliiren]: first and foremost do
the simplest thing (if the rule e.g. is always to repeat the same thing). And there is
of course something in this. It is significant that we can say that it is simpler to write
down a sequence of numbers in which each number is full same as its predecessor
than a sequence in which each number is greater by 1 than its predecessor. And
again that this is a simpler law than that of alternately adding 1 and 2.
57 Compare Baker and Hacker's discussion in their Scepticism, Rules & Language,
(Oxford: Basil Blackwell, 1984), pp. 82ff. I have some hesitations about Baker and
Hacker's rather free use of the notion of an "internal relation" to expound the later
Wittgenstein, hesitations I elaborate in my dissertation, 'The Rule of the Mathematical:
Wittgenstein's Later Discussions' (Harvard PhD Dissertation, 1990).
58 Compare PI Sections 215-6.
59 RFM I Section 3. See also PI Sections 213-4 where we read:
213. "But this initial segment of a series obviously admitted of various interpreta-
tions (e.g. by means of algebraic expressions) and so you must first have chosen one
such interpretation". - Not at all. A doubt was possible in certain circumstances.
But that is not to say that I did doubt, or even could doubt. (There is something to
be said, which is connected with this, about the psychological 'atmosphere' of a
process.)
So it must have been intuition that removed this doubt? - If intuition is an inner
voice - how do I know how I am to obey it? And how do I know that it doesn't
mislead me? For if it can guide me right, it can also guide me wrong.
Intuition an unnecessary shuffle.
214. If you have to have an intuition in order to develop the series 1 2 3 4 ...
you must also have one in order to develop the series 2 2 2 ...
60 Cf. Frege's Gmndgesetze der Arithmetik, p. vii:
Of course the pronouncement is often made that arithmetic is merely a more highly
developed logic; yet that remains disputable so long as transitions occur in the proofs
that are not made according to acknowledged laws of logic, but seem rather to be
based upon something known by intuition.
Wittgenstein is also indirectly attempting to undermine the kind of debate which took
place between Russell and Poincare over "intuition". See Russell's review of Science and
Hypothesis, reprinted from Mind, July 1905, in Russell's Philosophical Essays (New
York: George Allen and Unwin, 1966).
61 RFM I Section 3. Wittgenstein moved the final two paragraphs from a different place
in the manuscript, which may explain his switch to "2006" from "20006". See FV, Section
310.(316.).
62 RFM I Section 3. I have altered Anscombe's translation of the last sentence.
63 Cf. RFM IV Section 52:
Now can it be said that the concepts which mathematics produces are a convenience,
that essentially we could do without them?
First and foremost the adoption of these concepts expresses the sure expectation
of certain experiences.
We do not accept e.g. a multiplication's not yielding the same result every time.
And what we expect with certainty is essential to our whole life.
64 See PI Section 211.
65 Cf. DC Sections 448-9.
66 PI p. 211:
"I know what I want, wish, believe, feel, .... " (and so on through all the psychologi-
cal verbs) is either philosophers' nonsense, or any rate not a judgment a priori.
67 What would count as a relevant specification of my physiological state - whether, e.g.,
it would have to be similar in different environments for the same computation, whether
it could vary from step to step of the same series, whether there would have to be any
similarity between my state for the SOOth step of 2,2,2 ... and the 20003d step of +2
beginning with 0, whether it would be similar to other humans' physiologies for the same
computation, or even "related" computations, etc. - all this is, at present, unknown. An
informative physiological specification of my state at the 20003d step of the series + 2
would be considerably more difficult than our explanation of the computer's hardware,
since after all, we built the computer. Warren Goldfarb, 'Wittgenstein, Mind, and Scien-
tism', (Journal of Philosophy 86 (Nov. 1989),635-42) explores such difficulties in connec-
tion with Wittgenstein's discussions in PI of intentional mental notions (understanding,
believing, remembering, thinking, etc.) and the question of whether Wittgenstein is
standing in the way of empirical inquiry on the basis of an a priori argument. The care
with which Wittgenstein asks us to treat (and distinguish) specific questions serves, it
seems to me, as a reasonable philosophical lesson for the psychologist, who may jump
too quickly into glossing the aim of particular experiments to be the support of a general
psychological hypothesis about, say, inference in general.
68 Wittgenstein is referring to the law of Russell and Whitehead expressed in Principia
Mathematica (Press Syndicate of the University of Cambridge, 1910; reprinted paperback
version 1962), pp. 131-2:
We have next two propositions concerned with inference to or from propositions
containing apparent variables, as opposed to implication. First, we have, for the
new meaning of implication resulting from the above definitions of negation and
disjunction, the analogue of *11, namely
*912. What is implied by a true premiss is true. Pp.
That is to say, given ~ . p and ~ . p : J q", we may proceed to ~ . q", even when
the propositions p and q are not elementary. Also, as in *111, we may proceed from
~ . 4>x" and ~ . 4>x:J !/Ix" to ~ . 1jJx", where x is a real variable, and 4> and !/I are not
necessarily elementary functions. It is in this latter form that the axiom is usually needed.
It is to be assumed for functions of several variables as well as for functions of one
variable.
178 JULIET FLOYD
See RFM I Section 19ff.
69 RFM I Section 21.
70 RFM I Section 22.
71 Ibid.
72 Cf. PI Section 223:
One does not feel that one has always got to wait upon the nod (the whisper) next,
but it always tells us the same, and we do what it tells us.
One might say to the person one was training: "Look I; always do the same thing:
I ... "
73 PI Section 381:
Wie erkenne ich, dass diese Farbe Rot ist? - Eine Antwort ware: "Ich habe Deutsch
gelernt".
In the above translation, I have altered Anscombe's translation of "erkenne". The context
of this remark is different from that in RFM I, 3. In PI Wittgenstein is engaged in a
discussion of the notion of a private ostensive definition and interlocutor's notion that
such a definition might be secured by way of an image. The primary focus of the discussion
in the opening sections of RFM I is, by contrast, the character of logical inference. Cf.
Baker and Hacker, Scepticism, Rules & Language, p. 26.
74 Compare RFM VI Section 28:
Someone asks me: What is the color of this flower? I answer: "Red". - Are you
absolutely sure? Yes, absolutely sure! But may I not have been deceived and called
the wrong color "red"? No. The certainty with which I call the color "red" is the
rigidity of my measuring-rod, it is the rigidity from which I start. When I give
descriptions, that is not to be brought into doubt. This simply characterizes what we
call describing.
(I may of course even here assume a slip of the tongue, but nothing else.)
75 I have many intellectual debts. My emphasis on the importance of Wittgenstein's style
of writing derives from several years of courses, lectures, and reading of the works of
Stanley Cavell, Burton Dreben, Warren Goldfarb and Hilary Putnam. My first debt is
to Dreben, who urged me to work on Wittgenstein's discussions of mathematics and logic
and whose wit and acuity have been a constant source of encouragement and stimulation.
Conversation with Cavell on an earlier draft of Part 3 helped me to articulate my outlook
on Wittgenstein's notions of grammar and logic and to rethink Baker and Hacker's
metaphorical description of the manuscript basis of Wittgenstein's text. Goldfarb offered
numerous helpful criticisms of earlier drafts of Parts 1-4. Thomas Scanlon's detailed
comments on my characterization of Platonism helped me to clarify what I see at stake
in Wittgenstein's writing. William Flesch's reading of an early draft of Parts 4-6 kept me
focussed on Wittgenstein's text, while the thoughtful responses of Gary Ebbs and Charles
Parsons made me better appreciate the importance of placing an emphasis on Wittgenste-
in's philosophical method in defending my interpretation. Susan Austrian's willing ear
provided me with strong support in the early phrases of writing, and David Stern was
kind enough to provide me with a copy of the critical edition of FV and MV compiled
by G. H. von Wright and Heikki Nyman. Judson Webb and Cora Diamond offered
supportive criticism of the paper's final draft.
REFERENCES
Anderson, A. R.: 1958, 'Mathematics and the 'Language Game", Review of Metaphysics
11,446-58.
Anscombe, G. E. M.: 1985a, 'Review of Wittgenstein on Rules and Private Language',
Ethics 95, 342-52. '
Anscombe, G. E. M.: 1985b, 'Critical Notice of Saul Kripke's Wittgenstein on Rules and
Private Language', Canadian Journal of Philosophy 15, 103-9.
Baker, G. P. and P. M. S. Hacker: 1984, Scepticism, Rules and Language, Basil
Blackwell, Oxford.
Baker, G. P. and P. M. S. Hacker: 1985, Wittgenstein: Rules, Grammar and Necessity:
(An Analytical Commentary on the Philosophical Investigations, Volume II), Blackwell,
Oxford.
Bernays, P.: 1959, 'Comments on Ludwig Wittgenstein's Remarks on the Foundations of
Mathematics', Ratio 2, 1-22.
Cavell, S.: 1979, The Claim of Reason: Wittgenstein, Skepticism, Morality, and Tragedy,
Oxford University Press.
Cavell, S.: 1988, The Carus Lectures, 1988.
Chihara, C.: 1977, 'Wittgenstein's Analysis of the Paradoxes in his Lectures on the
Foundations of Mathematics', Philosophical Review 86, 365,..81.
Dummett, M.: 1959, 'Wittgenstein's Philosophy of Mathematics', in M. Dummett: 1978,
Truth and Other Enigmas, Harvard University Press, Cambridge, MA, pp. 166-85.
Floyd, J.: 1990, The Rule of The Mathematical: Wittgenstein's Later Discussions, PhD
Dissertation, Harvard University.
Frege, G.: 1893, Grundgesetze der Arithmetik, originally published at Jena; republished
by George Olms Verlagsbuchhandlung, 1966, Hildesheim; English translation (which
I cite), 1969, The Basic Laws of Arithmetic: Exposition of the System, Montgomery
Furth (trans.), University of California Press.
Frege, G.: 1980, Translations from the Philosophical Writings of Gottlob Frege, P. Geach
and M. Black (eds.), third edition, Rowman & Littlefield, Totowna, NJ.
Goldfarb, W.: 1983, 'I Want You to Bring Me a Slab',. Synthese 56, 26S-82.
Goldfarb, W.: 1989, 'Wittgenstein, Mind, and Scientism', Journal of Philosophy 86,635-
42.
Kant, 1.: 1790, The Critique of Judgment, James Creed Meredith (trans.), Oxford Univer-
sity Press, 1952.
Klenk, V.: 1976, Wittgenstein's Philosophy of Mathematics, Martinus Nihjoff, The Hague.
Kreisel, G.: 1958, 'Wittgenstein's Remarks on the Foundations of Mathematics', British
Journal for the Philosophy of Science, 9, 135-58.
Kripke, S.: 1982, Wittgenstein on Rules and Private Language, Harvard University Press,
Cambridge, MA.
Kripke, S.: 1985, 'Kripke on Wittgenstein on Rules', Journal of Philosophy 82, 471-8.
Russell, B.: 1903, The Principles of Mathematics, Cambridge University Press, Cam-
bridge; reprinted with a new introduction, 1938.
180 JULIET FLOYD
Russell, B.: 1905, 'Review of Science and Hypothesis', Mind; reprinted in Russell's
Philosophical Essays, George Allen and Unwin, New York, 1966: 70-8.
Russell, B. and A. N. Whitehead: 1910, Principia Mathematica, Cambridge University
Press, Cambridge; second ed. (1927) reprinted in paperback to 56, 1962.
Shanker, S. G.: 1987, Wittgenstein and the Turning Point in the Philosophy of Mathemat-
ics, State University Press of New York, Albany, NY.
Tait, W. W.: 1984, 'Wittgenstein and the "Skeptical Paradoxes''', Journal of Philosophy
81,759-78.
Von Wright, G. H.: 1969, The Wittgenstein Papers', in vol\ Wright, (1983), pp. 35-62.
Von Wright, G. H.: 1979, 'The Origin and Composition of the Philosophical Investi-
gations', in von Wright (1983), pp. 111-36.
Von Wright, G. H.: 1983, Wittgenstein, University of Minnesota Press, Minneapolis.
Wittgenstein, L.: 1921. Tractatus Logico-Philosophicus, originally published in the final
number of Ostwald's Annalen der Naturphilosophie under the title Logische-Philosophi-
sche Abhandlung (1921); English translation, C. K. Ogden (trans.) Routledge & Kegan
Paul, London, 1933 (first published in English, without corrections, 1922).
Wittgenstein, L.: 1937, The "Friihversion", or Early Version of Philosophical Investi-
gations, compiled by G. H. von Wright and Heikki Nyman (unpublished).
Wittgenstein, L.: 1945, The "Mittelversion", or Intermediate Version of Philosophical
Investigations, compiled by G. H. von Wright and Heikki Nyman (unpublished).
Wittgenstein, L.: 1958, Philosophical Investigations, G. E. M. Anscombe and R. Rhees
(eds.), G. E. M. Anscombe (trans.), second edition, Blackwell, Oxford.
Wittgenstein, L.: 1967, Zettel, G. E. M. Anscombe (ed.) and G. H. von Wright (trans.),
Blackwell, Oxford.
Wittgenstein, L.: 1969, On Certainty, G. E. M. Anscombe (ed.), Blackwell, Oxford.
Wittgenstein, L.: 1974, Philosophical Grammar, R. Rhees (ed.), A. J. P. Kenny (trans.),
Blackwell, Oxford.
Wittgenstein, L.: 1976, Wittgenstein's Lectures 011 the Foundations of Mathematics, Cam-
bridge 1939, Cora Diamond (ed.), University of Chicago Press; originally published
by Cornell University Press, Ithaca, NY. 1976.
Wittgenstein, L.: 1977, Remarks on Color, G. E. M. Anscombe (ed.), L. L. McAlister
and M. Schiittle (trans.), University of California Press.
Wittgenstein, L.: 1978, Remarks 011 the Foundations of Mathematics, G. H. von Wright
(ed.), R. Rhees and G. E. M. Anscombe (trans.), revised edition, Blackwell, Oxford.
Wittgenstein, L.: 1979, Wittgellstein and the Vienna Circle, Brian McGuinness (ed.),
Joachim Schulte and Brian McGuinness (trans.), Basil Blackwell, Oxford.
Wright, C.: 1980, Wittgellstein on the Foulldations of Mathematics, Duckworth, London.
Dept. of Philosophy
The City College of New York
New York, New York 10031
U.S.A.
ANNOUNCEMENT
THE LONDON SCHOOL OF ECONOMUCS
M.SC. IN THE PHILOSOPIDCAL FOUNDATIONS
OF PHYSICS
The Department of Philosophy, Logic and Scientific Method announces a
new one-year M.Sc. degree. The degree is designed specifically for students
with either a first degree in philosophy, or a science degree, who are inter-
ested in philosophical questions that arise in scientific inquiry in general, and
physics in particular. Applications from students for October 1991 are now
welcome.
The core course is Philosophical Foundations of Physics:- An introduction to
some fundamental ideas in physics and the methodological and philosophical
issues they raise.
For further details and application forms contact: The Graduate
Admissions Office, The London School of Economics, Houghton Street,
London WC2A 2AE. Telephone 071 405 7686. Fax 071 2420392.
Synthese 87: 181, 1991.
ANNOUNCEMENT
THE LONDON SCHOOL OF ECONOMICS
AND POLITICAL SCIENCE
LAKATOS AWARD
No Lakatos Award was awarded in 1990. (Professor John Earrnan won it in
1989, not 1990 as appeared in the last announcement.)
Synthese 87: 182, 1991.
JAAKKO HINTIKKA
AN IMPATIENT MAN AND HIS PAPERS*
Ludwig Wittgenstein died in 1951. At the time, the i,deas he had de-
veloped after his return to philosophy in 1928 were known only to a
handful of people through his lectures and conversations and through
the typescripts which he had dictated and which were circulating among
friends and colleagues. The wider philosophical community was eagerly
waiting for Wittgenstein's ideas to become effectively available.
Since 1953, no fewer than twelve volumes of Wittgenstein's writings
dealing with his later philosophy have seen the light of printer's ink.
In addition to them, posthumous material has been published in several
jourrials, and at least six volumes of Wittgenstein's recorded lectures
and conversations as well as five volumes of letters have found their
way into print. This might very well seem to constitute an adequate
basis for understanding Wittgenstein's later philosophy. Indeed, the
great majority of philosophers writing on him have obviously been
working on that assumption. Why, in view of these prima facie reasons
to the contrary, is there an urgent need to make all of Wittgenstein's
Nachlass available in print?
The reasons are deeply rooted in Wittgenstein's character and tem-
perament. A revealing glimpse is offered by Fania Pascal's perceptive
reminiscences in "Wittgenstein: A Personal Memoir" (in Rhees, R.
(ed.): 1981, Ludwig Wittgenstein: Personal Recollections, Basil
Blackwell, Oxford, pp. 26-62, esp. pp. 48, 60). She recounts the strange
incident of Wittgenstein's confession in 1937.
"I have come to make a confession". He had just been to Professor Moore for the same
purpose. "What did Professor Moore say?" He smiled. "He said, 'You are an impatient
man, Wittgenstein' .... " "Well, did you not know you were?" Wittgenstein, with dis-
dain: "I did not know".
Later in her memoir Fania Pascal sums up some of her observations as
follows:
He was an aggressive and explosive man, but this too in a very peculiar, naive way of
his own. At 48 he did not know the simplest thing about himself, namely, that he was
impatient. 1 have several times mentioned the forbidding severity he directed at himself.
Synthese 87: 183-201, 1991.
184 JAAKKO HINTIKKA
But he never saw himself through the eyes of others, and he had no other standards than
his own.
No one who ever came in contact with Wittgenstein is likely to
challenge Ms. Pascal's observations. Not long ago, I was asked by a
younger philosopher whether I had some insight into, or knowledge
of, why Wittgenstein and Carnap could not along. The implied
expectation obviously was of some dark moral secret casting its shadow
over the relationship of the two men. For a brief moment I was as-
tounded, but only until I realized that the questioner was too young to
have met either philosopher in person. If he had, I doubt that he would
have raised the question. Wittgenstein was the most impatient of human
beings; Carnap was one of the most patient of men. Carnap was wont
to ponder what this or that philosopher meant by what he had said,
and to ask repeatedly questions about them. Wittgenstein hated having
to explain himself. Carnap's patient and persistent questioning must
have driven him up the wall.
Wittgenstein's impatience offers a clue to the significance of his Nach-
lass. As a philosophical expositor, he exhibited the same qualities as
he did as a person. He directed a "forbidding severity" at Wittgenstein
the philosophical writer; but he did not see his writings through the
eyes of others, only in the light of his tremendous impatience.
What this implies for the reader of Wittgenstein's writings is clear.
He was struggling heroically to reach the clarity which would have
satisfied his own high standards, and anxious to express the ideas that
he had been able to reach. But he was frequently too impatient to
explain adequately what the problems were that his ideas were supposed
to be solutions to, partly because he did not realize that his problem
background was not always familiar to his readers.
Of course his problems were not traditional "philosophical prob-
lems", such as the reality of the world, the mind-body prob-
lem, etc. Such questions were regarded by him as hopeless muddles.
But this does not affect the fact that Wittgenstein's entire thought was
problem-driven, even though the solution to his problems consisted
typically of the elimination of a confusion or some other "mental
cramp".
Sometimes Wittgenstein was impatient even in presenting his own
solutions. After having moved on to new problems, he no longer felt
the need of expounding his solutions to the old ones again at any length,
AN IMPATIENT MAN AND HIS PAPERS 185
and expressed them only through shorthand references. For instance,
in Philosophical Investigations the name-object relation is explained in
a couple of sentences:
What is the relation between name and thing named? - Well, what is it? Look at
language-game or at another one; there you see what his relation can consist in.
(My reasons for departing from the standard translatipn are indicated
in Merrill B. Hintikka and Jaakko Hintikka: 1986, Investigating Witt-
genstein, Basil Blackwell, p. 209.) Yet when Wittgenstein had first come
upon the idea that language-games mediate the name-object relation,
viz., in the Brown Book, pp. 171-73, he had spent two entire pages
for the purpose of explaining the idea, and also had signalled unmistak-
ably its connection to his other problems.
This is an excellent illustration of what Fania Pascal meant by saying
that Wittgenstein never saw himself (or in this case, his ideas and
writings) through the eyes of others. This impatience affects different
parts of Wittenstein's oeuvre differently. His pre-World War I problems
were by and large the same as those of Russell and the other Cambridge
philosophers. Yet even there, in understanding Tractatus, Wittgen-
stein's impatience has made the task of later interpreters needlessly
difficult. As David Pears has convincingly shown, an important part
of Wittgenstein's background was Russell's theory of acquaintance.
Tractatus was in effect an Aufhebung of certain aspects of Russell's
theory, especially of the idea that logical forms can be objects of ac-
quaintance. Yet Russell's theory, which Wittgenstein takes to be known
to the reader, was never spelled out fully in print in Russell's lifetime.
It is developed most explicitly in the 1913 manuscript Theory of Knowl-
edge whose relevant parts were first published only in 1984 in vol. VII
of Russell's Collected Papers (Russell, B.: 1984, Collected Papers, ed.
by Elizabeth R. Eames in collaboration with Kenneth Blackwell, Allen
& Unwin, London).
Later, the members of the Vienna Circle shared most of Wittgen-
stein's central problems. However, subsequent generations of philoso-
phers, beginning with Wittgenstein's own students in the thirties, have
not always been aware of where Wittgenstein was coming from and
what path his work on his problems took. For this reason, Wittgen-
stein's former students are not necessarily in a better position to under-
stand his writings than others. Not all of them have realized that they,
too, can reach an adequate understanding of Wittgenstein's problems
186 JAAKKO HINTIKKA
and hence a deeper understanding of Wittgenstein's philosophy onl
through a careful study of the posthumous materials, especially of th
notebooks.
These insights into Wittgenstein as an expositor of his own ideas pl
the entire project of publishing his Nachlass into a new
The early publication efforts seem to have aimed at publishing Wittger
stein's later philosophical ideas in the way he himself intended. Thi
aim, understandable though it is, is not the appropriate one, for tw
different reasons. First, Wittgenstein's ideas were constantly
It is far from clear that there is anything like the later philosophy c
Ludwig Wittgenstein, much less that it is codified in anyone manuscrir
or typescript (or sets thereof) that he left behind. It is for instance
monstrous oversimplification to try to view The Blue and the Brow,
Books as "Preliminary Studies of the Philosophical Investigations", t,
the subtitle Rush Rhees gave to the volume.
Second, in view of what has been said, the relatively finished writing
of Wittgenstein's require as a necessary complement, needed to enabl,
us to understand their problem background, the availability of th,
notebook materials.
The emphasis in Wittgenstein editing has in fact shifted, at leas
ostensively, from the publication of individual volumes to the produc
tion of a Gesamtausgabe.
Before discussing the complete works projects, it is in order to notl
the serious problems that beset some of the piecemeal publications
For instance, the only halfway conventional book Wittgenstein lef
behind is TS 213 in von Wright's catalogue, commonly referred to a
The Big Typescript. (It even has chapters, chapter titles, sections, sec
tion titles, etc.) Rhees was supposed to edit it, but as Anthony Kenn:
has shown, he ended up doing something quite different. (See Kenny
1984, 'From the Big Typescript to the Philosophical Grammar', in Th,
Legacy of Wittgenstein, Basil Blackwell.) Rhees assembled a medley 0
materials, from different sources, which was never intended by Witt
genstein to go together and which are sometimes lifted out of an impor
tant context.
Likewise, Rhees's 1968 edition of Wittgenstein's highly import an
"Notes for Lectures on 'Private Experience' and 'Sense Data'" (1968
Philosophical Review 77, 274-320) omits without any indications abou
30 percent of Wittgenstein's actual text, including what I consider somt
of the most important passages Wittgenstein ever wrote. The
warning Rush Rhees issues is to say that "(f)or special reasons, I have
not included the sections on mathematics". This is entirely misleading,
because most of the omitted material has nothing to do with mathema-
tics. Rhees's omissions frequently take place in the midst of Wittgen-
stein's text, thus destroying any hope of grasping the continuity of
Wittgenstein's thought on the basis of his "edition".
Furthermore, the volume entitled Remarks on the, Foundations of
Mathematics is a mixture of materials from ten different manuscripts
of typescripts, of which five are merely excerpted. These MSS were
written over a period of several years. There is little reason to expect
that the different parts of the volume reflect one and the same set of
views or that these parts show faithfully Wittgenstein's line of thought.
Most importantly, there are serious questions also about Philosoph-
ical Investigations, especially about Part II. It is supposed to have been
intended by Wittgenstein to be a part of the same work as Part I, but
no documentary evidence to that effect has ever been made public.
This problem needs a separate study, but in view of its importance, a
few words may be in order here, even though I have little to add to
the facts of the case as presented by von Wright. The final version of
Part I of Philosophical Investigations was essentially finished in 1946,
even though Wittgenstein kept on polishing it. The preface to the
published book was dated in January 1945. Between 1946 and 1949
Wittgenstein wrote down his further ideas in notebooks 130-138. On
their basis he dictated two TSS, nos. 229 and 232, most of which were
subsequently published as Remarks on the Philosophy of Psychology
I-II. He also used these notebooks to compose carefully and selectively
a further MS, viz., MS 144. Out of this MS, somehow a typed version
(TS 234) was produced. This TS is the text Part II of Philosophical
Investigations was actually printed from. Professor Rhees has claimed
that Wittgenstein dictated it. There is some - admittedly inconclusive
- evidence to suggest that this claim is mistaken. By and large, the
published version (Philosophical Investigations II) and MS 144 are very
close together. There are a number of discrepancies, however, and in
a few of them MS 144 is apparently better in line with Wittgenstein's
intentions than the printed text. Moreover, five fragments which are
included in MS 144 are missing in Philosophical Investigations. It would
be of some interest to know precisely why these remarks were omitted
from the printed version, especially as some of the omissions seem to
exhibit a philosophical bias. Unfortunately, the situation is made even
188 JAAKKO HINTIKKA
more difficult by the fact that TS 234 has meanwhile disappeared and
no copy of it is known to survive. (Further information is nonetheless
forthcoming in a new, so far unpublished paper by G. H. von Wright,
whose perspective differs somewhat from my own.)
Whatever the details are or may be, it is clear that Philosophical
Investigations cannot automatically be assumed to be Wittgenstein's
definitive statement of his own settled philosopi;lical views.
The upshot of this list of problems is the importance of the rest of
Wittgenstein's literary remains. They provide, in fact, an ideal antidote
to philosophers' exclusive and misdirected preoccupation with the pos-
tulated end product of Wittgenstein's literary and philosophical labors.
In order to see this, it is useful to recall Wittgenstein's working method.
He kept a series of notebooks of which about thirty have survived from
the post-1928 period. In them, he wrote his ideas as they came to
him, sometimes dating the entries. These notebooks constitute a highly
unusual and fascinating document. Historically, psychologically, and
philosophically they offer a rare opportunity to witness a major philo-
sopher in the very act of coming upon new ideas, developing them,
revising them, and so on. Sometimes Wittgenstein even comments on
his own changes of mind. Reading the notebooks attentively and in
awareness of their ancestors and descendants almost gives one a feeling
of peeking over his shoulder.
When Wittgenstein had reached something like an equilibrium in his
thinking, he typically hired a typist and, using the notebooks as raw
material, dictated a typescript which he often hoped would eventually
become a publishable book. In dictating the typescript, he of course
omitted liberally old material and added new text. Thus, if the note-
books show what Wittgenstein's spontaneous new ideas were and how
he developed them before he was satisfied with them, the typescripts
show what solutions eventually satisfied him, if only for the time being.
Above all, the notebooks show what Wittgenstein's problems were,
what questions he was raising and where he was looking for answers
to them.
Publishing only the typescripts, or some of them, as Wittgenstein's
editors have mostly done so far, will give an incomplete picture of the
development of his ideas and consequently of these ideas themselves.
It would perhaps be an exaggeration to say that Wittgenstein's later
philosophy cannot be understood without a thorough knowledge of his
development such as it is reflected in his notebooks. It is extremely
difficult to do so, however, and it can only be done by someone who
is attuned to the philosophical problems which occupied Wittgenstein.
And even so, I doubt that an interpretation which is not based on the
notebooks can be fully detailed and accurate. Even more, I doubt
whether it can be convincingly proved to be correct. More than once
I have myself (alone, or in cooperation with Merrill B. Hintikka) first
came upon an interpretation on the basis of Wittgenstein's published
writings only to find subsequently the "smoking gun" 'that clinched the
case in Wittgenstein's notebooks or in other unpublished materials.
One of the many invaluable clues which the unpublished materials
yield to Wittgenstein's thought and to its development are the several
explicit statements that Wittgenstein makes as to what his own previous
views had been on this or that matter. Thus in MS 105, p. 251, Witt-
genstein offers an explanation of what he had earlier thought of the
color incompatibility problem. And in TS 213, as well as in MS 166,
he makes an extraordinary statement which shows to what extent his
discussion of rule-following in Philosophical Investigations was directed
against his own earlier views of what it takes to understand a rule.
And even apart from their role in showing us the development of
Wittgenstein's thought, his notebooks are a rare document in the entire
field of human thought. I doubt that there is any other major thinker
who has provided us with a more direct access to his own thought
processes.
Hence there are ample reasons for making Wittgenstein's Nachlass
available to the philosophical community in its entirety and in a form
in which it can be used by active philosophers who want to understand
his thought, evaluate it, and use his insights in their own work.
Thus we are led to the crucial question: What has been done to make
available the full story of Wittgenstein's philosophical efforts, such as
they are reflected in his literary remains?
The stage was set for these efforts by Wittgenstein's will. Its relevant
parts read as follows (I follow the original spelling and punctuation):
I give to Mr. R. Rhees, Miss G. E. M. Anscombe and Professor G. H. von Wright of
Trinity College Cambridge All the copyright in all my unpublished writings; and also the
manuscripts and typescripts thereof to dispose of as they think best but subject to any
claim by anybody else to the custody of the manuscripts and typescripts.
I intend and desire that Mr. Rhees, Miss Anscombe and Professor von Wright shall
publish as many of my unpublished writings as they think fit ...
190 JAAKKO HINTIKKA
The frequently used term "literary executors" for the three original
copyright owners does not occur in Wittgenstein's will. For con-
venience, I shall nevertheless use it here.
The three literary heirs mentioned in the will donated in 1969 all the
MSS and TSS to Trinity College, Cambridge. According to the terms
of the agreement between the literary heirs and Trinity College, two
committees were created to serve the two functions mentioned in
Wittgenstein's will. There is a Board of Trustees who administer the
copyrights and a Committee of Editors managing the publication of the
posthumous material. So far, both of these legally separate bodies have
had the same members. Originally, both of them consisted of the three
literary heirs. After Rush Rhees's death in 1989, Professor Peter Winch
was co-opted to serve on both committees, and in January 1990 Dr.
Anthony Kenny was likewise added to both boards.
There is another curious fact about the copyrights to the bulk of
Wittgenstein's Nachlass. In a sense, the bulk of Wittgenstein's literary
remains has been made available to the general philosophical public.
Since 1967, it has been possible to purchase a microfilm copy of most
of the unpublished materials through the Cornell University library
system. This has greatly facilitated scholars' work on Wittgenstein.
However, it has complicated the legal situation. In U.S. copyright law,
making these microfilms available for purchase (with the consent of the
copyright owners) constitutes publication. Therefore, the microfilmed
material is subject to the same provisions of the law as ordinary pub-
lished material, including the normal fair use provisions. Among other
things, this means that bona fide scholars have the right to quote short
passages from the Cornell material without an explicit permission from
the copyright owners.
Furthermore, it appears that the copyrights to the entire Cornell
material, which amounts to the bulk of Wittgenstein's Nachlass, are in
the public domain. The reason is that U.S. copyright law, at the time
of the Cornell "publication", required that, in order to retain exclusive
rights, the publisher register the copyrights and indicate them on the
publication itself. Apparently, neither of these things was done.
The availability of the Cornell microfilms means that most of Wittgen-
stein's literary remains are in a sense accessible to interested scholars.
This does not close the issue, however. For all practical purpose the
microfilm version is unusable, unless one is willing to devote one's
entire professional life to the study of Wittgenstein's development. This
is because much of the material exists only in the form of Wittgenstein's
longhand untranslated, unedited and untranscribed. Admittedly, as
German longhand goes, Wittgenstein's is mostly relatively easy to read.
(Personally I find myself temporarily stymied every couple of pages or
so, but only until I realize that Wittgenstein, in his impatience, has run
two German words together.) But what is impossible is to have any
quick overview of what Wittgenstein is saying on anyone page. In
order to see what he is as much as writing about, you have to read his
words one by one. This makes it extremely hard to get any sense of
the continuity of Wittgenstein's thought. For he seldom pursues the
same line of thought uninterruptedly for very long. He is usually think-
ing about several different topics at the same time, and records his
ideas on all of them one after the other. After dropping temporarily
one line of thought, it may be hours, days, weeks, months or years
before he takes it up again. Hence the use of his notebooks requires a
constant series of comparisons between different pages of the same
notebook and between different notebooks, which is agonizingly diffi-
cult on a microfilm machine. Furthermore, the recognition of the conti-
nuity of Wittgenstein's thought is made difficult by the habit of his that
Waismann complained so bitterly about, viz., his as it were constantly
beginning from the beginning, as if he had never thought about the
topic before.
Hence the Cornell microfilms cannot serve the purpose of making
Wittgenstein's literary remains effectively accessible to serious profes-
sional philosophers trying to understand his thought. In view of the
complexity of Wittgenstein's surviving writings and of their interre-
lations with each other, it is tempting to try to use computer technology
for the purpose. If an accurate, readable machine database is created
of the Nachlass, one is tempted to think that it could be used as a basis
of a printed edition of some or all of the materials, and, in any case,
could be used by scholars to follow the continuity of Wittgenstein's
thought (with the help of suitable global search programs) in less time
than the months and years of hard work presently required. It is there-
fore not surprising that the use of computers has been the leading
idea in the three major projects aimed at making Wittgenstein's entire
Nachlass available to scholars.
Before discussing these projects, it is in order to spell out what is
involved. When Wittgenstein died, no one, including his literary execu-
tors, had a realistic idea of what he was leaving behind. Gradually,
192 JAAKKO HINTIKKA
a surprisingly large and complex Nachlass has emerged, compnsmg
somewhere between twenty and thirty thousand manuscript or type-
script pages. Initially, it was virtually impossible to find one's way
among them. The philosophical community owes an enormous debt to
Professor G. H. von Wright whose work now enables us to have an
overview of the Nachlass. He has written an annotated, critical cata-
logue of the posthumous material, whose of the MSS and
TSS I am following in this article. He has also related the Nachlass
to Wittgenstein's most influential works, Tractatus and Philosophical
Investigations. Professor von Wright's papers on all these subjects are
most easily available in his volume Wittgenstein (Basil Blackwell, 1982).
The first attempt to edit a complete edition of Wittgenstein's writings
was launched around 1975. It involved a team working at the University
of Tiibingen in what was known as the Tiibingen Wittgenstein Archive.
The team was led by Mr. Michael Nedo and Professor H. J. Heringer.
The other scholars involved were Drs. R. Nowak, M. Rosso and J.
Schulte. The legal basis of their work was an agreement between the
three literary executors and the Wittgenstein-Archiv Tiibingen which
was signed on 19 October 1974. The project was supported mostly by
Fritz Thyssen Stiftung.
For a variety of reasons, the work on this project was never carried
out to its intended conclusion. The team members quarrelled with each
other, and work on this project came to an end sometime in 1980-
81. Not a single volume of Wittgenstein's posthumous manuscripts or
typescripts was published as a result of this project or has been subse-
quently published. The Tiibingen Wittgenstein-Archive which was to
host the project was also dissolved. Some insight into the reasons for
the collapse of the Tiibingen project is perhaps provided by the closing
report of the project. In it Professor Heringer complained that Mr.
Nedo "was incapable of directing such a project in an organizationally
serious or personally responsible manner" ("der ein solches Projekt
weder organisatorisch serios noch menschlich verantwortlich leiten
konnte") and that eventually "there arose with all the collaborators
considerable doubts concerning Mr. Nedo's scholarly competence".
On a later occasion, Mr. Nedo listed as his contributions to this
project only the organization (together with H. J. Heringer) of two
international symposia whose proceedings were published by Suhrkamp
and the editing (together with Michele Ranchetti) of a pictorial biog-
raphy of Wittgenstein. Neither enterprise furthered in any way the
Gesamtausgabe project.
In reality the Tiibingen project had nevertheless accomplished a fair
amount. Among other things, more than a half of the Nachlass was
transcribed into a database; a search program for textual similarities
was developed; and a prototype text segment was produced to serve
(it was hoped) as a basis for further development and for scholarly
discussion of the different forms which a definitive ect'ition might take.
Because of disagreement between the different persons involved, these
results could not be utilized by the second editing project, however.
After the Tiibingen project had come to an end, a new one was
launched. On 16 October 1981, the three literary executors applied for
support from Fonds zur Forderung der wissenschaftlichen Forschung,
which is the main Austrian governmental research foundation. The
project was to be directed this time by Mr. Nedo alone, and the project
description was written by him. The project was to involve Mr. Nedo
as Projektleiter, Ms. Isabelle Weiss plus short-term technical and secret-
arial help. The aim of the project included a "complete transcription
of the posthumous writings into a database and especially developed
computer programs". As far as the timetable was concerned, Mr. Nedo
wrote that
given appropriate working conditions a complete edition of Ludwig Wittgenstein's works
can be available by his hundredth birthday in April 1989.
Curiously, there is no indication in the grant application that the
proposed work would, to a considerable extent, merely repeat what
had already been accomplished by the Tiibingen group, or that Mr.
Nedo would not have access to those fruits of the Tiibingen group's
labors, even though such information would clearly have been highly
pertinent to the evaluation of the scholarly significance of the new
project.
The Fonds (as I shall here abbreviate the name) undertook to support
the project. On 27 September 1982, its Kuratorium decided to fund its
first stage as a pilot program which was to last twelve months and was
to be continued if successful. The Fonds awarded 633,000 Austrian
shillings for the purpose. At the same time, IBM Wien donated the
necessary computer time to the project.
The subsequent history of the project is not easy to chronicle, one
194 JAAKKO HINTIKKA
reason being Mr. Nedo's failure to keep his own sponsors apprised of
what he had done and what he had not done. The following is at least
part of the story. On 5 November 1983, Nedo submitted an annual
report to the Fonds. This report was approved of by Professor
Anscombe on the behalf of the literary executors without actually
consulting all of them. On the basis of her Aussage, the grant from the
Fonds was renewed in March 1984. It was apparently renewed again
in 1985.
In spite of the renewals, the project was not proceeding very well.
Mr. Nedo had promised the literary executors to produce for their
inspection a transcript of MSS 105-108 and TSS 208-210 as evidence
of his progress by October 1984. Because Nedo had not fulfilled any
of this promise by November 1987, Professor G. H. von Wright infor-
med the Fonds that he was withdrawing any further support of Mr.
Nedo) Without his knowledge, Professor Anscombe had meanwhile
submitted a renewal application on behalf of all the literary executors,
dated on 19 September 1987. As a consequence of Professor von
Wright's letter, the Fonds postponed its decision on the renewal appli-
cation.
Early in 1988 Mr. Nedo produced a transcript of MSS 105-106.
Although this represents only about one quarter of his initial task, a
compromise was reached among the literary executors in October 1988
to the effect that a new application should be made to the Fonds for a
grant which would enable Nedo to complete the assignment. On 24
April 1989, the Kuratorium of the Fonds awarded the sum of 1,690,000
Austrian shillings for the project, which is the sum applied for in 1987.
What all this amounts to is a nearly total failure of the project so
far. Mr. Nedo has yet to fulfill most of the work he promised to have
ready in 1984. In general, the documented progress of the project has
been minimal. Several disturbing questions are prompted by Mr. Nedo's
record. For instance, in a report of his activities in 1983, Nedo claimed
that the first notebook volume had been typeset in 1983. If this infor-
mation is correct, much of the material which Nedo finally produced
as evidence of his activities in 1988 was in effect ready in 1983. If so,
there is precious little that the project accomplished between 1983 and
1986. The very same picture book that Nedo had, in 1981, listed as a
product of the Tiibingen project is now listed as a result of the 1981
project. In reality, this book has nothing to do with the collected works
project. In the earlier project descriptions, several interesting results
were promised by Nedo, including the following: In the 1981 project
description, Nedo claimed to have reconstructed two MSS, MS 126 and
MS 127 (which have disappeared), and promised to edit and publish
them. Not only has this promise remained unfulfilled, it has been
dropped completely from the new project. Likewise, in the original
project Nedo undertook to publish "an extensive bibliographical de-
scription of the entire Wittgenstein Nachlass". Nothing has come about
this plan either. In 1981 Mr. Nedo was thinking about having the entire
Nachlass published by the Wittgenstein centennial in April 1989. In
1987 he claimed that, with appropriate support, he would have four
volumes in print by April 1989 containing MSS 105-108 and TSS 208-
210. Needless to say, nothing like that was ever accomplished.
In view of this string of unfulfilled promises, it does not seem advis-
able to let Mr. Nedo continue the editing of Wittgenstein's Nachlass.
His track record shows amply that he is not a suitable person to carry
out the Gesamtausgabe project.
Why has the Gesamtausgabe project failed so far? How should it be
continued, if it will be? There is not enough evidence available to give
a definitive answer to these questions, but a few educated guesses are
possible. According to the information supplied by Mr. Nedo to his
sponsors, by 1985 he and his collaborators had 15,000 pages, that is,
about a half of the Nachlass, transcribed in a database. How is it
conceivable, assuming that this information is correct, that by 1990 no
single volume has appeared and that a decent hard copy is available
only of a tiny fragment of the total material? Several explanations are
possible, but they all point to the same direction. The transcribed
material might have been coded in a way that makes its utilization for
editing purposes difficult. Furthermore, it seems obvious that Mr. Nedo
and his aides have not managed to develop software that would enable
them to move effectively from the transcribed text material to an edited
text. This is undoubtedly the gist of the problem. The alleged coding
problems are also likely to be but consequences of the same fact: it is
hard to know what kind of coding is appropriate when you do not have
an adequate idea of the software to be used to handle the inputted
text.
It is in fact embarassingly obvious that the entire editing project was
launched before the people responsible for it had a realistic conception
of the software needed for the purpose, or even a conception of how
such software could be developed. What is amazing is nevertheless not
196 JAAKKO HINTIKKA
this overoptimism which the project undoubtedly shares with several
others. What is astounding is the failure, as far as the available record
above, of Mr. Nedo to use any of the available resources for the
purpose of software development. The industrial companies he has
tried to befriend have been big computer manufacturers like IBM and
Olivetti, not software companies, even though such companies have
meanwhile actually developed all kinds of software for the very purpose
of text editing. There is no evidence available that Mr. Nedo has
consulted any software expert, and no evidence that he has availed
himself of the experience of other computer-based editing projects.
His excuse has been the alleged special difficulty of the Wittgenstein
manuscripts. Yet there are successful computer-based editing projects
where the technical problems, however different, are of an even greater
complexity, such as the Leibniz project. Mr. Nedo's comments on other
recent editing projects, such as the new Holderlin edition, merely show
that he has no grasp of where his real problems lie. It looks as if Mr.
Nedo has been waiting for a solution to fall into his lap as a gift
from the heavens, perhaps simply in the form of increasingly powerful
computers, without realizing that you need appropriate software for
them, too.
I for one do not see any reason for continuing the Gesamtausgabe
project before a suitable software has been developed and successfully
tested. And the most efficient way of doing so is probably a close
cooperation with a suitable ambitious software company.
Other important questions concern the rationale of the entire project.
The edition of MSS 105-106 which Mr. Nedo has finally made available
to others does not yield to a philosophical reader any information which
a conventional critical edition would not. Yet such a conventional
edition, even the most ambitious editio diplomatica, could have been
produced in a few months at a fraction of the cost. It is also hard to
see what the decisive advantages are supposed to be of Mr. Nedo's so
far unpublished edition of the two notebooks over the prototype
printout which was produced by the Ttibingen project.
It is hard to avoid the impression that the total plan of the Gesamtaus-
gabe project has never been clearly thought out. If the aim has been
simply to edit Wittgenstein's complete writings, it could have probably
been done in fifteen years with the resources that have actually been
available for the purpose, while maintaining the highest editorial stan-
dards.
The advantage which can be claimed for a computer-based edition
is that, after a database has been created out of the total material,
suitable software will enable scholars do research on Wittgenstein in a
way which is now either impossible or else extremely tedious. However,
it is obvious that the fundamental questions have not been asked here.
How important would, e.g., the availability of computerized search
programs be for Wittgenstein research? Having u ~ e unpublished
Wittgenstein materials in my own work, I am keenly aware of the
tremendous convenience promised by search programs and other simi-
lar automated research tools. I am also keenly aware how very impor-
tant the precise form of Wittgenstein's words can be, including the
variants he himself penned in the text. Yet I cannot help asking: How
crucial are such automated research tools as compared with the avail-
ability of a decent critical text? If, for instance, I look back on the
insights into Wittgenstein's terminology and choice of words that have
recently been achieved, such as the force of his term Aspekt or the
crucial contrast between Name and Bezeichnung, would they have been
easier to reach by means of computerized methods of dealing with the
database? Scarcely. What is required for such insights is a sensitivity
to the philosophical and conceptual issues involved such more than
extensive comparative evidence, whose main role perhaps lies in verify-
ing results rather than in reaching them. There is, in my judgment,
even a clear danger that Wittgenstein research, or part of it, will be
directed by the increased reliance on computers into philosophically
unimportant directions.
Be such fears justified or not, the fact remains that decisions concern-
ing editorial priorities cannot be detached from judgments concerning
the philosophical interest of the different aspects of Wittgenstein's liter-
ary remains. I am very uneasy to see such decisions entrusted to Mr.
Nedo, who has no formal philosophical training and whose expressed
views on the content of Wittgenstein's Nachlass seem to me superficial
and arbitrary. For instance, in the 1981 grant application to the Fonds,
Mr. Nedo discusses Wittgenstein's ideas about mathematics in a way
which shows that he has no sense of what is essential and inessential
in Wittgenstein's views.
And even if I am underestimating the importance of computerized
text processing, a very serious priority question arises here. Which
would have been easier, quicker, and more economical, producing a
conventional critical text first and then inputting it into a database or
198 JAAKKO HINTIKKA
the other way round, as in the two Gesamtausgabe projects? Would
anything have been lost in the former procedure? Maybe the answer is
that something would have been lost, but it seems to me that the
question has not been raised. And the moral of the story of the two
Gesamtausgaben projects surely is that the editing could have been
done much more quickly in the old-fashioned way.
It may also turn out that the bottom line on tpe balance sheet of the
Gesamtausgabe project is in a sense not even zero, but negative. Not
only is it the case that the project has not produced any publications
so far. Because of Mr. Nedo's work, the majority of the literary execu-
tors has disallowed certain other Wittgenstein editions. For instance,
there exist several fully edited volumes, virtually ready for the printer,
in Helsinki which have been edited by Professor von Wright and his
assistants. They include most of the unpublished materials that are
crucially important for our understanding of the genesis of Philosophical
Investigations, including MSS 116 and 144 as well as TSS 220, 221, 227,
and 239.
Ironically, Mr. Nedo's partial monopoly of editing Wittgenstein
seems to have, at this moment, no legal basis, if it ever had one. In
1979, the literary executors wrote a letter to the Tiibingen group promis-
ing that they would not give permission for any competing edition in
the next ten years. However, this promise was given to the Tiibingen
group, which came to an end in 1981, not to Mr. Nedo. In any case,
the monopoly expired in 1989. But even before it, Nedo had little claim
to any preferential treatment, for he had consistently failed to fulfill
the terms of the 1981 application to the Austrian Fonds. In particular,
he had failed to keep all the literary executors informed about his work.
The efforts to publish a Gesamtausgabe are thus after some fifteen
years without any definitive product, even though there is some hope
for the future. This has created an unfortunate situation in several
respects. There is a veritable scholarly industry of books and papers on
Wittgenstein going on unremittingly, oblivious to the critical importance
of the notebooks and other unpublished materials for the interpretation
of Wittgenstein, which will be subject to a sharp re-evaluation in the
light of the literary remains.
What may be worse, an unhealthy climate has been created among
those who are aware of the importance of the Nachlass but are without
easy access to them. The main reasons for the failure of the major
editing projects have been the judgments and decisions of the literary
executors. Yet some of the very same persons responsible for the
editing of Wittgenstein have also been engaged in interpreting his
philosophy. They have therefore placed themselves in the precarious
position of being in control of other scholars' access to materials in the
light of which their own interpretations are to be judged and which
could conceivably prove some of these interpretations wrong. For their
own sake, it is to be hoped that the present untenaple situation will
soon be resolved so as to clear up unnecessary suspicions and rumors
of their motivation and comportment.
The third project designated to make Wittgenstein's Nachlass more
easily available to scholars is of an entirely different kind. There is a
detailed account of this project available in the form of mimeoed report
by Claus Huitfeldt and Viggo Rossvaer entitled The Norwegian Witt-
genstein Project Report 1988 (The Norwegian Computing Centre for
the Humanities, October 1989, 282 pp., ISBN 82-7283-052-3). Because
of the public availability of this report, I shall restrict my description
of this Norwegian Wittgenstein Project to a minimum.
The project began in 1981 as a joint venture of a number of Norwe-
gian Wittgenstein scholars who formed a committee representing all
the philosophy departments of Norwegian universities. The aim was
not a published edition, but a computer-readable text which would be
able to yield to scholars all the necessary information about Wittgenste-
in's own textual changes, corrections, alternatives, etc. The project was
funded exclusively from Norwegian sources, to wit, by the Norwegian
Research Council, the Nansen Foundation, and different Norwegian
universities.
Approximately 3250 pages had been transcribed under the auspices
of the Norwegian project by the time the report was written. Over
and above that, transcribed material from the Tilbingen project was
transferred to Norway, resulting in machine-readable text totalling
somewhere between 700 and 8400 pages. Software to facilitate machine
reading was also developed. At least one non-Norwegian scholar has
actually used the material in question and found it useful. The Tilbingen
material has not yet been properly converted, however, because of
disputes concerning its legal status. These disputes go back to the
quarrels within the Tilbingen group.
The Norwegians unfortunately did not realize that in order to make
the Nachlass generally available even in a machine-readable form they
needed the permission of Wittgenstein's literary executors. After a long
200 JAAKKO HINTIKKA
series of negotiations, that permission turned out not to be available.
In the absence of a permission, no continued financial support was
available, and effective work on the project came to a standstill in
December 1987.
Recently, however, the Norwegian Wittgenstein Project has been re-
organized and revived. The University of Bergen has assumed the
responsibility for the project. On 3 May 1990, a, decision was made by
the committee controlling the Nachlass to grant the Norwegian project
the permission to create a machine-readable database Gesamtausgabe
for the use by Wittgenstein scholars. I do not know whether or not this
welcome development will actually help the production of a printed
Gesamtausgabe.
A secondary, but not unimportant subplot in the story of Wittgen-
stein's literary remains, concerns the publishers of the remains. The
original text material published so far in book form has been mostly
brought out by Suhrkamp or by Basil Blackwell. Alas, Mr. Nedo has
had a bitter quarrel with Suhrkamp, and there has been friction between
Professor Anscombe and Basil Blackwell. Perhaps because of these
problems, probably for other reasons as well, the Gesamtausgabe pro-
ject does not have a publisher for the products of their labors.
Another subordinate issue concerning Wittgenstein's notebooks is
due to his double use of them. Besides writing in them his philosophical
ideas as they came to him, Wittgenstein used the very same notebooks
to record a number of other observations. In order to make it impossible
for any casual reader to have access to his private thoughts, he used in
these observations a code, which is, nevertheless, easy to decipher. The
coded passages have in fact been deciphered. In 1985, a pirated edition
of the coded parts of the 1914-1916 notebooks was published in Span-
ish, in the magazine Saber in Barcelona.
In principle, however, access to the coded passage is strictly con-
trolled by the literary executors. They are supposed to be censored from
the microfilms distributed hy the Cornell University library, although at
least one major university library has mistakenly received an unexpur-
gated copy.
Once again, the issues surrounding the coded passages are too com-
plex to be resolved in a single essay. I cannot pass judgment here on
what the proper attitude ought to be to these passages of Wittgenstein's
notebooks. On a deeper level, in any case, one can scarcely dismiss
the facts of Wittgenstein's personal life as being irrelevant to a full
understanding of his thought. To think that the way his personality is
reflected in his life and in his moral struggles is unrelated to his philo-
sophical thought is to take a shallow view of the complexity of a major
philosopher's thought and its roots in his personality.
There is a more direct reason, however, why some of the coded
passages should be made available to philosophers. Many of the coded
comments have nothing to do with Wittgenstein's private life. Instead,
they are his comments on his own philosophical activity. Even though
they do not tell anything directly about the content of his philosophical
thoughts, they are quite revealing about the dynamics of his thinking
and about his relations to other thinkers. For this reason, they offer us
vital testimony concerning the development of his ideas. This is amply
shown by the selection of coded passages published in Wittgenstein's
Vermischte Bemerkungen in 1977. However, other philosophically
relevant passages are still waiting to be made public.
NOTES
* A shorter version of this article appeared in The Times Literary Supplement N. 4565,
28 September-4 October 1990, p. 1030. The reprint of the previously published material
is by permission of The Times Literary Supplement.
I am also grateful to Professors G. H. von Wright, Rudolf Haller, and David Stern as
well as to Mr. Heikki Nyman for information and for other kinds of help.
Dept. of Philosophy
Boston University
745 Commonwealth Ave.
Boston, MA 02215
U.S.A.
DAVID STERN
THE "MIDDLE WITTGENSTEIN":
FROM LOGICAL ATOMISM TO PRACTICAL HOLISM*
1. THE "MIDDLE WITTGENSTEIN"
Wittgenstein arranged the Tractatus in its final form during the summer
of 1918; Part I of the Philosophical Investigations was put into the form
in which we now have it during the mid 1940s. The Tractatus was
published in 1922; the Philosophical Investigations in 1953, two years
after Wittgenstein's death. Because the two books were widely studied
and interpreted at a time when the rest of his writing was unavailable,
it became common practice to speak of the author of the first book as
"Early Wittgenstein", and the author of the second as the "Late
Wittgenstein". As it gradually became clear that his writing during the
intervening years was not only voluminous, but also could not simply
be understood as a rejection of one view and the adoption of another,
it was natural to speak of the author of this further body of writing, or
at least those parts of it which could not be regarded as "Early", or
"Late", as "Middle Wittgenstein".
While it is certainly convenient to use this talk of an early, middle
and late Wittgenstein as conversational shorthand for lengthier and
more cautious ways of discussing the development of his philosophy, it
also has its dangers. One of the main assumptions which commonly
accompanies this periodisation is the belief that we can neatly divide the
development of Wittgenstein's thought into several sharply demarcated
phases. While there is some truth to this, it is only a half-truth', and
for this reason it is particularly dangerous. First, it can lead to an
interpretive strategy which cuts Wittgenstein up into a number of inde-
pendent time-slices and so loses sight of the unity of his philosophy.
Such an outlook lends itself to schematic summaries of what Witt-
genstein really meant, summaries which turn his writing into just the
kind of philosophical theories which he so vigorously opposed. I think
it should be clear by now that we ought to take Wittgenstein's rejection
of philosophical theorising, his lifelong conviction that philosophy is
"not a theory but an activity" seriously, even if we don't ultimately
take him at his word.
1
Synthese 87: 203-226, 1991.
204 DAVID STERN
But perhaps the greatest danger in this talk of early, middle and late
Wittgenstein is that we read Wittgenstein in the light of currently
fashionable interpretations, rather than evaluating those interpretations
against a careful reading of what Wittgenstein actually wrote. Much of
the recent work on "Middle Wittgenstein" has been particularly valu-
able precisely because it is based on close examination of the full range
of Wittgenstein's writing. While I don't think we should give up trying
to provide interpretive schemata when studying Wittgenstein, such
schemata cannot be based on an analysis of isolated fragments selected
from his writing. Rather, we have to start by asking what Wittgenstein's
problems were, and need to understand the "problem background"
against which he worked.
2
To do this, we need to read his writing as
a whole, rather than concentrating on isolated fragments. For this
reason, I begin by looking at a passage from the Blue Book which
highlights these issues. The first forty or so pages of the Blue Book are
about the nature of meaning, the last thirty about solipsism and the
nature of experience. In between, there is a short passage in which he
compares different ways of making progress in philosophy with different
ways of arranging books scattered across a library floor. Wittgenstein
contrasts the straightforward case in which one can pick up each book
and put it in its final place with a situation where one might have to
start by putting several books in the right order, simply to show that
they belong together in that order. This leads him to the following
reflections:
In this case, in fact, it is pretty obvious that having put together books which belong
together was a definite achievement, even though the whole row of them had to be
shifted. But some of the greatest achievements in philosophy could only be compared
with taking up some books which seemed to belong together, and putting them on
different shelves; nothing more being final about their positions than that they no longer
lie side by side. The onlooker who doesn't know the difficulty of the task might well
think in such a case that nothing at all has been achieved?
In its place in the Blue Book, in between the discussion of meaning in
the first part of the book and the discussion of experience in the second,
the point of the simile is to explain the overall relationship between
these two parts. The first part of the book argues that meaning does
not consist in the occurrence of mental processes, separating two prob-
lems which he had previously thought of as belonging together. But
the results achieved in the first part of the book are only provisional,
and their final location is not yet determined, for "every new problem
LOGICAL ATOMISM TO PRACTICAL HOLISM 205
which arises may put in question the position which our previous partial
results are to occupy in the final picture".4
Perhaps this image of organising a library, of working out what goes
where, is also an apposite description of our present relationship to
Wittgenstein's writings as a whole: we need to sort out which parts
belong together, and why. In this paper, I offer a selective outline of
my interpretation of the development of Wittgenstein's work, one which
concentrates on the so-called "Middle period". 5 I do this by discussing
a number of crucial passages in his writing, passages in which he deci-
sively changes his conception of the nature of mind and language,
moving away from the Tractatus and toward the Philosophical Investi-
gations. I identify a train of thought in Wittgenstein's writings which
leads from Tractarian logical atomism, through the logical holism of
the late 1920s and early 1930s, and from there to his later practical
holism. In outline, my reading can be summarised as follows.
Wittgenstein's initial break with the Tractatus, which probably oc-
curred in the late 1920s, consisted in the realisation that he had been
wrong to think that all logic could be reduced to the truth-functional
logic of the Tractatus. Consequently, he gave up logical atomism, the
doctrine that all meaningful discourse can be analysed into logically
independent elementary propositions, for logical holism, the thesis that
analysis leads to systems of logically related propositions. At first, he
retained the Tractarian conviction that language is grounded on refer-
ence to objects, which he now identified with the contents of experience.
This project of analysing the structure of the experientially given is
briefly articulated in the paper, 'Some Remarks on Logical Form'. 6 At
this point, in the early months of 1929, he conceived of the project as
a matter of articulating a "phenomenological language" , a language for
the description of immediate experience. Later that year, he gave up
the idea that philosophy ought to start from a description of the immedi-
ately given, motivated by the conviction that philosophy must begin
with the language we ordinarily speak. This led him to reject the
idea that language is animated by intrinsically representational mental
processes, in favour of a conception of a language on which the meaning
of a sentence is determined by the rules for its use. Finally, he replaced
the view that each case in which one applies a rule requires an act of
insight with the view that the application of a rule is a matter of deciding
to apply the rule in a certain way.
These changes are followed by a transitional period during the first
206 DAVID STERN
half of the 1930s in which Wittgenstein explored their implications. The
transitional period comes to end in the mid 1930s, with the first expo-
sition of what has since become known as "the private language argu-
ment", and the construction of the 'first part of what von Wright has
called the "Early Investigations", which roughly corresponds to the
first 188 sections of the published Philosophical Investigations. During
this period, he moves away from a conception of; language as constitut-
ing a formal system of rules, embracing the view that mastery of rules
is dependent on a background of shared practices. As he stresses in
the Investigations, his use of "the term 'language-game' is meant to
bring into prominence the fact that the speaking of language is part of
an activity, or of a form of life".
7
Though these forms of life, the taken-
for-granted ways of acting which make language possible, are social
and practical, not individual and incorrigible, they occupy an analogous
role in Wittgenstein's later philosophy to the role he gave to the experi-
entially given in 1929 or objects in the Tractatus, for they are the point
at which the later Wittgenstein acknowledges the limits of language.
Thus the Philosophical Investigations also states that "what has to be
accepted, the given, is - so one could say - forms of life". 8
Consequently, he no longer thinks of the rules of our language as a
matter for decision or convention, though he does not return to his
earlier essentialism, either; rather, he tries to achieve a clear view of
the relation between logic and practice by exploring the attractions and
shortcomings of both the conventionalist and essentialist positions. This
practical holism is sharply opposed to the logical holism which it re-
places. I take this term from Dreyfus's analysis of Heidegger and late
Wittgenstein in 'Holism and Hermeneutics', where he draws a parallel
distinction between theoretical holism, the view that all interpretation
is a matter of trap slating between theories, and practical holism, the
view that while everyday coping with things and people "involves ex-
plicit beliefs and hypotheses, these can only be meaningful in specific
contexts and against a background of shared practices".9 Because this
inherited background involves skills, habits and customs, it cannot be
spelled out in a theory. In the remainder of this paper, I shall flesh out
the line of development I have just indicated, from logical atomism to
logical holism, and from logical holism to practical holism. 10 I shall also
argue that a certain conception of the primacy of immediate experience
lies behind the earlier view, and that the rejection of this conception of
experience is crucial for the subsequent development of Wittgenstein's
philosophy.
2. LOGICAL ATOMISM
At the heart of the Tractarian system is an argument for the necessary
existence of simple objects. That argument begins from, the observation
that I speak and understand language, and leads to the conclusion that
language must be analysable into elementary propositions, "logical
atoms". This argument is extremely abstract, for it turns on establishing
that it is only possible for language to be composite if there are simples
out of which it is composed. As a result, Wittgenstein concluded that
every significant statement must be composed of logical atoms, yet was
unable to give any examples. Indeed, it was the rationalistic and abstract
character of this train of thought which Wittgenstein most vehemently
criticised after he rejected logical atomism in the late 1920s. Shortly
afterwards, he branded his earlier work as "dogmatic" - a term with
strongly Kantian overtones - precisely because he had placed so much
weight on the claim that it was possible to carry out an analysis of
our language yet had failed to do SO.l1 The Tractatus is an abstract
metaphysical framework, a product of trying to think through a number
of key issues in the philosophy of mathematics and language which
Wittgenstein took from Frege and Russell. Because of this, it is mislead-
ing to read a well worked out position on the nature of experience and
knowledge back into it, although that is just what Wittgenstein did in
the early 1930s, when he identified the simple objects with sense-data.
Instead, I want to stress the way Wittgenstein's "dogmatism" gives the
Tractatus its schematic character, which lends itself to any number of
interpretations of "what Wittgenstein really meant".
But the main lines of Wittgenstein's conclusions about logical atoms
in the Tractatus can be summed up in a few sentences. These atoms, the
so-called elementary propositions, cannot be decomposed into smaller
units, but they are nevertheless made up of names, which play the role
of sub-atomic particles: they refer to the simple objects which make up
the world.
12
The meaning of what we ordinarily say is the logical
product of the combination of these sub-atomic particles into the mol-
ecules of ordinary speech. A proposition is a picture, or a model, of
reality: the fact that the names of which it is composed are related to
208 DAVID STERN
one another in a determinate way represents that things are related to
one another in the same way.13 A proposition is true if things are
arranged as it says they are, false if they are not. Even though Witt-
genstein had found no examples of the elementary propositions, the
basic components of his theory of meaning, he thought he had shown
that they must be logically independent of each other. By this, he
meant that the truth or falsity of each atomic proposition did not depend
on the truth or falsity of any other atomic proposition. 14
Of course, the signs from which our propositions are constructed -
the words of everyday language, and the corresponding psychical
constituents which make up our thoughts - do not clearly display the
underlying structure which Wittgenstein had deduced. IS However, he
believed that any meaningful proposition of ordinary language must
have an analysis which does clearly display its underlying structure.
This would be a truth-functionally related ensemble of atomic proposi-
tions whose surface structure is identical with the structure of the
facts it represents. Wittgenstein distinguished between the sign, the
perceptible part of the proposition, and the symbol, the sign together
with its application, its representational relation to the world.
16
Ordin-
ary language consists of relatively simple signs, which stand in a very
complex representational relationship with the world; a fully analysed
language would consist of a much more complex arrangement of signs
in a very simple representational relationship with the world. But no
sign, considered in isolation, is intrinsically meaningful: one always has
to see how the signs are to be applied, to grasp the way they represent.
This leads Wittgenstein to distinguish two ways of conceiving of a
picture, corresponding to his distinction between sign and symbol: as
a fact, a determinate arrangement of objects, and that fact, together
with the representational relation which makes it into a picture.
17
This ability to apply language, to see signs as symbols, as intrinsically
related to their objects, is not itself a further fact in the world, but is
a matter of our establishing a projective relationship between certain
facts. The projective relation, the sign's meaning, cannot itself be a
fact, for all facts are logically independent of one another, and there
is a logical connection between a picture qua symbol and what it pic-
tures. Thus this connection cannot itself be stated, for only facts can
be stated. But it can be shown by the structure of our language, much
as the projective connection between two depictions of the same thing
in different perspectives can be shown. Any combination of signs is
only meaningful due to our applying them in a certain way; everyday
language depends on the existence of extremely complex conventions
connecting the words we use with the objects they stand for. Similarly,
even the signs of a fully analysed language would have to be explained
before we could understand them. Like the signs that make up the
propositions we presently speak, utter and think, they would be conven-
tionally meaningful: their meaning would be a prqduct of linguistic
conventions, just as the meaning of the words we presently use is
conventional. In both cases, we would also have to know how to use
the words in question, how to apply them. But not all signs can be
given a meaning in this way: some must be non-conventionally meaning-
ful and so not susceptible to being either interpreted or misinterpreted.
In the Tractatus, "thoughts" play this role: a thought is an applied
propositional sign, a picture together with the method of projection
which gives it its significance.
18
A Tractarian thought is not simply an
inner monologue or image, for these are facts on a par with physical
facts composed of words and pictures. Indeed, Wittgenstein explicitly
told Russell that a thought consists of "psychical constituents that have
the same sort of relation to reality as words. What those constituents
are I don't know" .19 However, a thought is not just a concatenation
of signs, it is the signs together with their application, the projective
relation which gives them their life. This is the original private language,
the inner symbolic processes which give public language its significance.
Unlike any physical language, thoughts are essentially meaningful and
so cannot be misunderstood. They encompass the "meaning-locus"
which provides the ground for our use of language, the projective
relationship we apply to the signs.
20
A thought, an applied and thought
out propositional sign, is intrinsically related to its object. There is a
logical connection between the thought and its object. Our language is
like paper currency; it is only meaningful insofar as it is backed up by
gold in the bank - both are intrinsically representational processes.
Wittgenstein's view of the irreducibly mental and intentional charac-
ter of meaning was diametrically opposed to Russell's Humean view of
meaning, on which it is nothing more than a feeling which accompanies
certain mental processes.
21
But this view is only implicit in the Trac-
tatus; it is protected from explicit formulation by the Tractarian doctrine
that the relation between language and world cannot be stated, but can
only be "shown". However, I believe it was motivated by a picture of
the relationship between language, experience and world, a picture that
210 DAVID STERN
is best described as Cartesian, though it also has affinities with post-
Kantian thought. While this picture is hinted at in the Tractatus, in the
central role Wittgenstein gives to thought in his account of propositions,
in his sympathy for solipsism and his identification of the world with
life, it only emerges clearly in his subsequent work.22 For it was only
in the late 1920s, when Wittgenstein realised that the Tractatus's dogma-
tism would not do, that he returned to grappliqg with the problems
raised by his essentialism about meaning and the need for an account
of the relationship between language and experience.
3. FROM LOGICAL ATOMISM TO LOGICAL HOLISM
By the late 1920s, Wittgenstein had come to think of the task of arriving
at a characterisation of the ultimate level of analysis as a matter of
formulating what he called a "phenomenological language". 23 How-
ever, in the first section of the Philosophical Remarks, drafted in Oc-
tober 1929, he wrote that he no longer had
phenomenological language, or "primary language" as I used to call it, in mind as my
goal. I no longer hold it to be necessary. All that is possible and necessary is to separate
what is essential from what is inessential in our language.
24
The original manuscript volume entry reads instead: "I no longer hold
it to be possible".zs Some commentators regard the replacement of this
impossibility claim by the seemingly weaker claim that a phenomenolog-
ical language is not necessary as a qualification which implies that
Wittgenstein had recognized that he could not yet prove that phe-
nomenological language is impossible. Hintikka and Hintikka, for in-
stance, argue that this lacuna would only be filled much later by the
private language argument. 26 However, the continuation of the passage
in question suggests a much simpler explanation for the rewording:
Wittgenstein sometimes uses the term "phenomenological language" in
a restricted sense, to mean a canonical analysis of the experience of
the present moment. In this sense, he consistently maintained after
1929 that such a language was indeed impossible. But he also spoke of
"phenomenological language" in a looser sense, meaning by it any way
of talking about the content of experience, and in this sense of the
term, he holds that a phenomenological language is possible but not
necessary. Thus, he turned away from phenomenological language in
the narrow sense, the construction of an artificial philosophical language
that would be capable of fully describing present experience, in favour
of a study of the structure of the language we ordinarily speak, which
still included a study of phenomenological language in the looser sense
of the term.
Wittgenstein's clearest explanation of what he had meant by the term
"phenomenological language" occurs in the first section of the chapter
entitled 'Idealism' in the Big Typescript, assemble<;i during 1932-33;
the section is entitled 'The Representation of Immediate Experience'.
There, he writes: "Phenomenological language: the description of im-
mediate sense perception, without hypotheticaf addition". 27 Unlike our
ordinary ways of describing sense perception, in terms of persisting
objects, a phenomenological language would make no such hypothetical
commitments; instead, it would restrict itself to representing whatever
is immediately experienced. But his objection to this conception is
already implicit in the broken underlining he put under "language",
Wittgenstein's way of indicating that he was not satisfied witlijiis choice
of wofds. For he had realised that the motives which led him to seek
a phenomenological language ensured that no language could be ad-
equate. What one really wants to do, he suggests, is to directly present
what is experienced:
If anything, then depiction by means of a painted picture or something like it must surely
be such a description of immediate experience. As when we, e.g., look in a telescope
and record or paint the constellation seen.
28
But even the most direct language cannot be more than a re-pre-
sentation, and so the very idea of a phenomenological language turns
out to be incoherent. The desire for a transparent intermediary, present
to consciousness, which guarantees that my thoughts are about their
objects, and which provides an ultimate basis for language, is the prod-
uct of philosophical confusion.
Wittgenstein sets out his objection to phenomenological language by
asking us to consider how we might go about actually reproducing sense
perception; he suggests that we imagine constructing a mechanical scale
model of what is seen. Seen from the correct point, the model produces
the appropriate perception:
the model could be set in the right motions by a crank-drive and we could by turning
the crank read off the description. (An approximation to this would be a representation
in film.)
If that is not a representation of the immediate - what would be one? - Anything
212 DAVID STERN
which tried to be more immediate still, would have to give up being a description. Instead
of a description what would then corne out would, rather, be that inarticulate sound with
which some writers would like to begin philosophy. ("I am, knowing of my knowledge,
conscious of something" Driesch. )29
The very generality of this criticism, the fact that it does not turn on
any specific formulation of a sense-datum theory, implies that it would
be a mistake to construe Wittgenstein's "phenoIl;lenological language"
too narrowly, as expressing a commitment to a specific philosophical
theory about how to analyse ordinary language into observation state-
ments. Rather, phenomenological language inherits much of the gen-
erality of the Tractarian notion of the ultimate level of analysis: it is
the language, whatever it is, that enables one to describe immediate
experience without any hypothetical additions. But the work I have
just been discussing was motivated by an overall conception of the
relation between experience and the external world. During this period,
Wittgenstein frequently explained his conception of the relation be-
tween experience and the world in terms of a comparison with the
relation between the picture one sees on the screen at the movies and
the pictures on the reel of film in the movie projector. Although this
analogy is never fully elaborated in any of the published works, there
are a number of references to it in Wittgenstein's manuscripts, in the
Philosophical Remarks and Big Typescript, and in the notes of his
lectures and conversations in the years following 1929.
Talking to Bouwsma in 1949, Wittgenstein mentioned that the "figure
of the cinema lainp" had first struck him when he was talking to Frege
in 1911.
30
Bouwsma recalls a meeting of Malcolm's discussion group in
Cornell in which Bouwsma began the conversation by briefly talking
about Descartes's Cogito, ergo sum. Wittgenstein responded by saying
that the real question was "How did Descartes come to do this?"
Bouwsma asked whether he meant to ask what led up to the cogito in
Descartes's own thought, to which Wittgenstein replied
No. One must do this for oneself ... I always think of it as like the cinema. You see
before you the picture on the screen, but behind you is the operator, and he has a roll
here on this side from which he is winding and another on that side into which he is
winding. The present is the picture which is before the light, but the future is still on the
roll to pass, and the past is on that roll. It's gone through already. Now imagine that
there is only the present. There is no future roll, and no past roll. And now further
imagine what language there could be in such a situation. One could just gape. This!31
The difference between the picture on the film-reel which is in front of
the projector and the picture on the screen is emphasised by contrasting
the way in which the picture on the reel is part of a sequence of
neighbouring pictures which either have been or will be projected,
while the picture on the screen has no such neighbours. If we are
serious about trying to represent the picture on the screen without any
hypothetical additions, we must take it by itself, excluding the sequence
of past, present and future pictures on the film reel; in fact, we must
exclude the whole physical world. But any description of the picture
would be a sequence of symbols which was another part of the physical
world. Furthermore, our ordinary language is temporal; how could
we use it to talk about an atemporal, self-contained world without
misdescribing it?
If we restrict ourselves to the case of describing presently given
experience, it may be hard to see the full force of the problem: Why
worry about how to represent it when it is already present? But even
if we were to accept, for the sake of argument, that representing the
immediately given is unproblematic, we would still have to account for
our ability to remember and think about phenomena which are not
part of occurrent sensory experience. Wittgenstein's discussions of the
representation of experience often employ examples in which the in-
tended object is not given in current experience. While the events
depicted on the roll of film seemed unproblematic when considered by
themselves, our ability to think about those events, to grasp that certain
signs mean those events, became highly puzzling. Thus, Wittgenstein
wrote in late 1929:
The application of words, considered as extended in time, is easy to understand; in
contrast, I find it infinitely difficult to understand the sense in the moment of application.
What does it mean e.g. to understand a sentence as a member of a system of
sentences?32
Wittgenstein tried to express his conception of a categorical distinc-
tion between experience and the world by saying that the picture on
the screen, unlike the picture on the film-reel, "has no neighbours".
This is not a matter of saying that my experiential field has a certain
location, and that there is nothing next to it. Rather, it is to say that
it makes no sense to think of my immediate experience as adjoining
anything else - that it is neighbourless: self-contained, complete. In
lectures given in the early 1930s, Wittgenstein said that
214 DAVID STERN
... the pictures in the lantern are all "on the same level" but that the picture which is
at any given time on the screen is not "on the same level" with any of them, and that if
we were to use "conscious" to say of one of the pictures in the lantern that it was at that
time being thrown on the screen, it would be meaningless to say of the picture on the
screen that it was "conscious". The pictures on the film, he said, "have neighbours",
whereas that on the screen has none.
33
This is closely connected with the idea that it makes no sense to speak
of experience as "present": just as Wittgenstein' denies that experience
has a spatial location, he also denies it has a temporal location. In both
cases, he wants to resist the temptation to project our everyday gram-
mar onto the phenomena. For he holds that there is a "grammatical"
difference between the two; the picture on the film and the picture on
the screen are not "'on the same level'''. The pictures are not two
related objects which share a common space; instead, they represent
two different spatialities.
34
If we try to bring the two together, we shall
run up against the limits of language.
In the early 1930s, Wittgenstein came to think that the mistake which
leads to solipsism of the present moment, the view that only my present
experience is real, can be explained in terms of the movie metaphor. On
this account, the solipsist mistakenly compares "present" experience to
the frame of the film which is currently being projected, in front of
those past frames which have been projected and behind those future
frames which will be projected. Wittgenstein's reply is to contend that
the correct analogy is with the picture on the screen, which does not
lie in this order at all, and so "present" - and for that matter, "my",
"experience", and "real" - is inapplicable:
The present we are talking about here is not the picture in front of the projector's lens
at precisely this moment, as opposed to the pictures before and after it, which have
already been there or are yet to come; but the picture on the screen which would
illegitimately be called present, since "present" would not be used here to distinguish it
from past and future. And so it is a meaningless epithet.
35
During the period when he still found the movie metaphor a compel-
ling illustration of our predicament, Wittgenstein must have wished he
could get around the need to use language to convey what cannot be
said, to find a way of getting his readers to simply see the nature of
the phenomena. One aspect of this problem is his recognition that
language itself is part of the physical world and so is on a different
level from the phenomena. Thus, early in 1929 he raised the following
question about the relation between physical language and phenomeno-
logical language:
Language itself belongs to the second system. If I describe a language, I am essentially
describing something that belongs to physics. But how can a physical language describe
the phenomenal?36
Shortly afterwards, he began to realise that it was this very conception
of present experience which was responsible for his difficulties:
The way of looking which leads into a magic valley, as it were, from which there is no
way into the open countryside is taking the present as the only reality. This present,
constantly flowing or, rather, constantly changing, cannot be caught hold of. It disappears
before we can think of grasping it. We stay stuck in this valley, bewitched, in a whirl of
thoughts ....
What I may not think, language can't express. That is our comfort.
But if.ione says: the philosopher must step down into this encircled valley and grasp
the pure reality itself and bring it to the light of day, then the answer runs that he would
have to in so doing leave language behind and so come back without having achieved
anything.
And yet there can be a phenomenological language. (Where must it stop?)37
Soon after, in the same notebook, Wittgenstein decisively rejected this
conception of a phenomenological language.
38
In short, what Witt-
genstein realised, late in 1929, is that there cannot be a phenomeno-
logical language of the kind he had hoped for - a complete analysis of
experience - and that we must start from our everyday language' in
describing experience. In the Tractatus, he had held that while truth-
functional propositional logic limits what can be said, its logical form
shows the common structure of language and world: "Logic must take
care of itself". 39 Once he gave up the logical atomist doctrine that
all relations between propositions are ultimately truth-functional, it is
language as a whole which takes its place: "Language has to speak for
itself".40 The logical holist conceives of everyday language as a system
of rules, such as the rules of a formal calculus, or a scientific theory.41
In the final section of this paper, I summarise why Wittgenstein replaced
this with a practical holist conception of everyday language on which
even formal rules must be understood in terms of their practical back-
ground: "rules leave loop-holes open, and the practice has to speak for
itself".42
216 DAVID STERN
4. FROM LOGICAL HOLISM TO PRACTICAL HOLISM
In 1929, Wittgenstein rejected logical atomism for a logical holist con-
ception of language as a system of calculi, formal systems characterised
by their constitutive rules. But by the mid 1930s he came to see that
the rules of our language are more like the rules of a game than a
calculus, for they concern actions within a social context. This context,
our practices and the 'forms of life' they embody, on the one hand,
and the facts of nature on which those practices depend, on the other,
are the background against which rule-following is possible. It is this
emphasis on both the social and natural context of rule-following which
is characteristic of Wittgenstein's later conception of language as a
practice.
Both the logical holist conception of language as a calculus or a
theory and the later practical holist notion of a language-game stress
the paramount role of context, the notion that the context in which an
utterance belongs is the whole of language. This is why both conceptions
are holistic: the meaning of an utterance, an inscription or a thought
is not an entity independent of the rest of our language, but rather
consists in its relation to the rest of language. There is no sharp transi-
tion from the calculus model to the language-game model, from logical
holism to practical holism: Wittgenstein does not give up the idea that
our linguistic practices are rule-governed, but rather comes to see that
rule-governed behaviour is only possible against a background of prac-
tices which cannot themselves be explicitly formulated as rules. Alterna-
tively, one might say that his conception of language becomes increas-
ingly broad, until it includes the whole range of human activity.
Despite these continuities, Wittgenstein's conception of language
changed radically during the 1930s. During the early 1930s, Witt-
genstein frequently compares language to a calculus, a formal system
of rules. While he retains this analogy in subsequent writing, it is usually
as an object of comparison, as a way of bringing out the disanalogies
between ordinary language and a calculus. Thus, he dismisses the idea
that we can think of the meaning of a word as an entity, a "meaning-
body", which lies behind the use of our words, in favour of a description
of the rules which we accept for the use of the word.
43
In the Philosoph-
ical Investigations, Wittgenstein implies that the analogy between a
calculus and ordinary language had been responsible for his conception
of language as governed by a system of rules:
F. P. Ramsey once emphasized in conversation with me that logic was a "normative
science". I do not know exactly what he had in mind, but it was doubtless closely related
to what only dawned on me later: namely, that in philosophy we often compare the use
of words with games and calculi which have fixed rules, but cannot say that someone
who is using language must be playing such a game ....
All this can only appear in the right light when one has attained greater clarity about
the concepts of understanding, meaning and thinking. For it will then also become clear
what can lead us (and did lead me) to think that if anyone utters a sentence and means
or understands it he is operating a calculus according to definite r u l e ~ 44
By the mid 1930s, Wittgenstein plays down the role of explicit rules
in explaining what a word means, and instead stresses that in practice,
we usually explain the meaning of words by giving paradigmatic ex-
amples, not by ostensive definition or stating necessary and sufficient
conditions. The examples need not be given in such a way that they
are protected against all possibilities of misunderstanding; it is enough
that they usually work. In discussing the Tractarian theory of the nature
of the proposition, Wittgenstein replies in section 135 of the Philosoph-
ical Investigations: "Asked what a proposition is ... we shall give exam-
pIes". Most of the next one hundred sections of the Philosophical
Investigations are occupied with exploring and undermining the two
main avenues of reply to this position: the view that meaning consists
in mental processes and the view that meaning consists in implicit rules,
views which Wittgenstein had previously been attracted to. The aim of
this discussion is to discredit the idea of hidden processes which underlie
meaning, whether they be subjective mental processes or objective
rules.
A key argument which motivates this train of thought is that just as
no mental content is intrinsically meaningful, so no strict rule by itself
can determine how we go on, as all determination of meaning is depen-
dent on interpretation. Given any mental process or any formulation
of a rule, it is always, in principle, open to a further, deviant, interpreta-
tion.
45
No occurrent act of meaning or intending, or of grasping an
essence or of deciding to go on in a certain way, can give a rule the
power to determine our future actions, because there is always the
question of how that act is to be interpreted. As a result, the idea that
a rule can determine all its future applications turns out to be misguided.
Only if we ignore the context can we think that some isolated act or
event can have a determinate meaning regardless of its context. A
change in the context of application can yield a change in meaning,
and therefore meaning cannot be identified with anything independent
of context.
218 DAVID STERN
We can see a parallel development in Wittgenstein's conception of
following a rule. In the Philosophical Remarks, he still thinks that the
application of a rule always depends on an act of insight, the insight
that the rule can be applied in this case:
Is it like this: I need a new insight at each step in a proof? ... Something of the
following sort: Supposing there to be a certain general rule (therefore one containing a
variable), I must recognise afresh that this rule may be applied here. No act of foresight
can absolve me from this act of insight. Since the form to which the rule is applied is in
fact different at every step:6
But there is a marginal note in the original copy of the Philosophical
Remarks which encapsulates the next phase of Wittgenstein's thought.
Next to "No act of foresight can absolve me from this act of insight"
he wrote in the margin: "Act of decision, not insight". That is, the
application of the rule is no longer treated as a matter of my seeing
that it must apply, but rather in terms of my deciding to apply it here.
The issue arises again in the Brown Book in an early version of the
rule-following discussion in Philosophical Investigations section 185ff.
Wittgenstein discusses the possibility that someone might learn how
to add one as we do with small numbers, but does what we would call
adding two when asked to add one to numbers between one hundred
and three hundred, adds three when asked to add one to larger num-
bers, and persists in regarding this procedure as a correct application
of the rule he or she was taught. This leads Wittgenstein's interlocutor
to ask whether he thinks an act of intuition will always be needed to
protect us against the possibility of a deviant interpretation of the rule:
"I suppose what you say comes to this, that in order to follow the rule 'Add l' correctly
a new insight, intuition is needed at every step. ,,47
Instead of giving a direct answer to this question, Wittgenstein replies
by asking the interlocutor to explain the notion of following the rule
correctly, a notion which he has simply taken for granted in his ques-
tion:
- But what does it mean to follow the rule correctly? How and when is it to be decided
which at a particular point is the correct step to take?48
The interlocutor responds by appealing to the rule-giver's intentions:
"The correct step at every point is that which is in accordance with the rule as it was
meant, intended.,,49
In the ensuing dialogue, the interlocutor tries to specify what the rule-
giver's intentions consist in, and Wittgenstein repeatedly undermines
him by asking how the application of the intention, meaning, mental
act or whatever other candidate, is offered can be guaranteed in ad-
vance: the same deviant possibilities can always be raised. A page later,
Wittgenstein sums up:
If the mere words of the rule could not anticipate a future transition, no more could any
mental act accompanying these words.
We meet again and again with this curious superstition, as one might be inclined to
call it, that the mental act is capable of crossing a bridge before we've got to it. This
trouble crops up whenever we try to think about the ideas of thinking, wishing, expecting,
believing, knowing, trying to solve a mathematical problem, mathematical induction, and
so forth.
It is no act of insight, intuition, which makes us use the rule as we do at the particular
point of the series. It would be less confusing to call it an act of decision, though this
too is misleading, for nothing like an act of decision must take place, but possibly just
an act Of writing or speaking. so
Here, there is no suggestion that there must be an act of decision; only
that if any mental process at all is involved, it is a decision, not an
intuition. The search for a self-interpreting interpretation only arises if
one treats the words and actions in which the rule was expressed as an
interpretation, a construal of the rule which still needs to be made
completely determinate. In some cases, of course, our words or actions
will be ambiguous. But in others, we can just get on with it and write
or say the next term in the series. Wittgenstein is proposing that we
look at these cases, the cases in which we do not do any interpreting,
but simply grasp the rule in practice, such as everyday conversation
and arithmetic, as prototypical instances of rule-following. As he puts
it in On Certainty:
As if giving grounds did not come to an end sometime. But the end is not an ungrounded
presupposition: it is an ungrounded way of acting.
s1
In pursuing these issues in Philosophical Investigations section 138ff.,
Wittgenstein focusses on the question: What is it to understand a word
in a flash or, as he had previously put it, "to understand the sense in
the moment of application,,?52 For, on the one hand, understanding a
sentence is a matter of being able to use those words correctly, in
applying them "in the course of time"; on the other hand, we may
grasp the meaning of a sentence in a flash, when it "comes before our
220 DAVID STERN
mind in an instant".53 In other words, there is both a subjective and
an objective aspect to understanding and we need to understand how
they hang together. The problem is directly descended from Wittgen-
stein's earlier worries about the relationship between the words I use
in describing my experiences and the experiences themselves, and the
way in which the words seemed to be unable to describe "that which
goes on in the reading of the description" ,54 the mental processes which
animate our words. Certainly we do, on occasion, understand words in
a flash, but the danger here is to think that this can give us an insight
into the essence of understanding. The same image can mean different
things in different contexts:
What is essential is to see that the same thing can come before our minds when we
hear the words and the application still be different. Has it the same meaning both times?
I think we shall say not. 55
Wittgenstein holds that it is' the circumstances in which the experiences
occur, not the experience taken by itself, to which we should look if
we want to get at what justifies someone in saying that he or she has
understood. The therapy he proposes is that we consider relatively
simple cases of mental activity, such as what goes on when we read out
loud, in order to see how the temptation of thinking that ineffable
mental processes are involved arises.
56
In other words, he asks us to
consider "the activity of rendering out loud what is written or printed;
and also of writing from dictation, writing out something printed, play-
ing from a score, and so on". 57 The first point Wittgenstein makes in
this connection is that although we are all very familiar with such
activities, we would find it difficult to describe the part which they play
in our life "even in rough outline". The observations that follow suggest
that this is due to the differences between what goes on when a skilled
reader reads, where all kinds of things may go on, and often nothing
at all over and above the successful completion of the task, with what
goes on with a beginner, who makes a conscious effort to read. And if
we concentrate on the case of the beginner we shall indeed be inclined
to say that it is "a special conscious activity of the mind". 58 The idea
is that there must be some special conscious act of reading, "the act of
reading the sounds off from the letters". 59 Here Wittgenstein replies
by thinking of cases in which the experience goes on, but the "reader"
does not really understand (perhaps he or she has been drugged), and
cases in which the reader does understand, but nothing or something
else goes on. Whether something is reading or not is a matter of
whether the activity is successful, not a matter of whether I consciously
apply a rule or feel guided or whatever. One wants to say that in
reading "the words come in a special way" or that "I experience the
because", but in practice, the words don't have to come in a special
way and we needn't experience anything in particular. 60
In most cases of proficient reading, we just get on with it and do it.
Here we have an example of what Heidegger calls "readiness-to-hand"
[Zuhandenheit]: our ordinary use of everyday things does not call for
reflective awareness of what we are doing. We only become aware
of these things when something goes wrong or some other unusual
circumstance draws our attention to them, making them "present-at-
hand" [vorhanden].61 Thus, under normal circumstances, reading a
familiar language, "can we say anything but that ... this sound comes
automatically when we look at the mark?". 62 But if we try to under-
stand the cases where we are proficient on the model of what goes on
in abnormal or problematic cases, we shall inevitably look at what goes
on when we make an effort to read as revealing the essence of being
influenced, an essence that is concealed in normal usage.
63
We think
of particular cases of being guided or being influenced, and think that
the experience we have in this or that case is paradigmatic, what goes
on in every case.
Ultimately, our explicit beliefs and interpretations are only meaning-
ful in specific contexts and against a background of shared practices -
these practices are the skills and customs which we have learnt, ways
of acting which were not acquired as beliefs, even though we may
express them in beliefs. It is this "way of grasping a rule which is not
an interpretation, but which is exhibited in what we call 'obeying the
rule' and 'going against it'" which ultimately ends the regress of inter-
pretations.
64
In other words, "It is our acting, which lies at the bottom
of the language game". 65
NOTES
* Earlier versions of this paper were presented at the University of Iowa, at a symposium
on the "Middle Wittgenstein" at an APA meeting in Oakland, California, and at the
Wittgenstein centennial conference at the Florida State University, Tallahassee, Florida.
I am indebted to members of those audiences for their constructive comments - especially
to Jaakko Hintikka, David Pears and Hans Sluga.
1 L. Wittgenstein: 1933, Tractatus Logico-Philosophicus, translated by C. K. Ogden
222 DAVID STERN
(London: Routledge and Kegan Paul), 4.112. Cf. Philosophical Investigations: 1967,
translated by G. E. M. Anscombe (Oxford: Blackwell) # 109ff. , #124ff. The German text
and English translation are on adjacent pages. While the quotations from Wittgenstein's
published writing in this paper are based on toe available translations, they are sometimes
rather more literal.
2 J. Hintikka: 1988, '''Die Wende der Philosophie': Wittgenstein's New Logic of 1928",
Proceedings of the Twelfth International Wittgenstein Symposium (Vienna: Holder-
Pichler-Tempsky), p. 380.
3 L. Wittgenstein: 1975, The Blue and Brown Books [hereafter either Blue Book or
Brown Book] (Oxford: Blackwell), pp. 44-45. He introduces the analogy as an illustration
of the way in which philosophical problems are interdependent.
4 Blue Book, p. 44.
5 While it is clearly impossible to do justice to the development of Wittgenstein's philos-
ophy during this period in a paper of this length, I believe it is possible to indicate some
of its main characteristics. I explore these themes in greater depth in Wittgenstein on
Mind and Language (Oxford: Oxford University Press, forthcoming).
6 L. Wittgenstein: 1929, 'Some Remarks on Logical Form', Proceedings of the Aristotelian
Society, supplementary volume, pp. 162-71.
7 L. Wittgenstein, Philosophical Investigations, #23.
8 Philosophical Investigations, p. 226.
9 H. Dreyfus: 1980, 'Holism and Hermeneutics', The Review of Metaphysics 34, p. 9.
Cf. the discussion of the background in chapter 5 of J. Searle: 1983, Intentionality
(Cambridge: Cambridge University Press).
10 While this way of dividing up the development of Wittgenstein's thought amounts to
qualified support for the notion of a "Middle Wittgenstein", these are certainly not the
only important turning points in Wittgenstein's work. For instance, his criticism of his
earlier work culminates in the completion of Part I of the Philosophical Investigations in
the mid 1940s, after which he made a fresh start. This work resists classification but can
loosely be described as being on philosophical psychology, the inner and the outer,
knowledge, certainty and relativism.
11 See F. Waismann: 1979, 'On Dogmatism', in Ludwig Wittgenstein and the Vienna
Circle: Conversations Recorded by Friedrich Waismann, edited by B. F. McGuinness,
translated by J. Schulte and B. F. McGuinness (Oxford: Blackwell), 9 December 1931,
p. 182 ff. German edition, with the same pagination: 1967, Ludwig Wittgenstein und
der Wiener Kreis: Gesprdche aufgezeichnet von Friedrich Waismann, edited by B. F.
McGuinness (Oxford: Blackwell).
12 Tractatus, 3.2 ff.
13 Tractatus, 2.15.
14 Tractatus, 1.21,2.062,4.211,5.134-5.135.
15 Tractatus, 4.002.
16 Tractatus, 3.262, 3.31, 3.32.
17 Tractatus, 2.14, 2.1513.
18 Tractatus, 3.5; see also 3 ff., and 4.
19 L. Wittgenstein: 1974, Letters to Russell, Keynes, and Moore, edited by G. H. von
Wright (Ithaca, NY: Cornell University Press), 19 August 1919, p. 72.
20 B. Goldberg: 1968, 'The Correspondence Hypothesis', Philosophical Review 77,438-
54.
21 See: R. M. McDonough: 1986, The Argument of the Tractatus (Buffalo, NY: SUNY
Press), Ch. VI.1; and S. S. Hilmy: 1987, The Later Wittgenstein (Oxford: Blackwell), pp.
109 ff.
22 See Tractatus, 3-3.12, 5.6-5.641, and 6.431-6.4311.
23 See MSS 105-107 and 'Some Remarks on Logical Form'. I discuss this material in
greater detail in Wittgenstein on Mind and Language.
References to Wittgenstein's typescripts and manuscripts use the numbering system in
G. H. von Wright's catalogue of the Wittgenstein papers, originally published in the
Philosophical Review 78 (1969). The latest revisions can be found in the version published
in S. Shankar: 1986, Ludwig Wittgenstein: Critical Assessments (Wolfeboro, NH: Croom
Helm), volume 5, pp. 1-21.
24 L. Wittgenstein: 1975, Philosophical Remarks, 2nd edition, edited by Rush Rhees,
translated by R. Hargreaves and R. White (Oxford: Blackwell), #1. German edition, with
the same pagination: 1964, Philosophische Bemerkungen, edited by R. Rhees (Oxford:
Blackwell) .
25 MS 107, p. 205.
26 M. Hintikka and 1. Hintikka: 1986, Investigating Wittgenstein (Oxford: Blackwell),
pp. 137 ff., 172, 241 ff.
27 Big Typescript, #101, p. 491. The German reads:
Phan6menologische Sprache: Die Beschreibung der unmittelbaren Sinneswahrneh-
mung, ohne hypothetische-Zutat.
The Big Typescript, TS 213 in von Wright's catalogue, is a book-length draft which
Wittgenstein assembled in 1932-33, based on material which he had written since his
return to Cambridge in 1929. For a discussion of the place of the Big Typescript in the
Wittgenstein oeuvre, see A. Kenny: 1976, 'From the Big Typescript to the Philosophical
Grammar', edited by 1. Hintikka, Essays on Wittgenstein in Honour of G. H. von Wright,
Acta Philosophica Fennica 28; reprinted in Kenny: 1984, The Legacy of Wittgenstein
(Oxford: Blackwell).
28 Big Typescript, #101, pp. 491-92. The German reads:
Wenn etwas, dann muss doch wohl die Abbildung durch ein gemaltes Bild oder
dergleichen eine solche Beschreibung der unmittelbaren Erfahrung sein. Wenn wir
also z.B. in ein Fernrohr sehen und die gesehene Konstellation aufzeichnen oder
malen.
29 Big Typescript #101, p. 492. The German reads:
Denken wir uns sogar unsere Sinneswahrnehmung dadurch reproduziert, dass zu
ihrer Beschreibung ein Modell erzeugt wird, welches, von einem bestimmten Punkt
gesehen, diese Wahrnehmungen erzeugt; das Modell konnte mit einem Kurbel-
antrieb in die richtige Bewegung gesetzt werden und wir konnten durch Drehen
der Kurbel die Beschreibung herunterlesen. (Eine Annaherung hierzu ware eine
Darstellung im Film.)
Ist d a s keine Darstellung des Unmittelbaren - was sollte eine sein? - Was noch
unmittelbarer sein wollte, mtisste es aufgeben, eine Beschreibung zu sein. Es kommt
dann vielmehr statt einer Beschreibung jener unartikulierte Laut heraus, mit dem
manche Autoren die Philosophie gerne anfangen mochten. CIch habe, un mein
Wissen wissend, bewusst etwas" Driesch.)
224 DAVID STERN
Cf. Philosophical Remarks, #67-68; Philosophical Investigations, #261.
The quotation from Driesch is taken from page 19 of his Ordnungslehre, where he
calls it "our first philosophical proposition, the primal philosophical proposition. It still
points to the birthplace of all philosophy, 'experience', [Erleben] but it nevertheless lifts
itself out of the everyday toward language" (Jena: Eugen Diederichs, 2nd edition, 1923).
Cf. his Wirklichkeitslehre (Leipzig: Emmanuel Reinicke, 1922), p. 8.
30 O. K. Bouwsma: 1986, Wittgenstein: Conversations 1949-1951, edited by J. L. Craft
and R. Hustwit (Hackett, Indianapolis), 5 August 1949, p. 10.
31 Bouwsma, op. cit., 7 August 1949, p. 13. See also Ludwig Wittgenstein and the Vienna
Circle, p. 50; and Big Typescript, #102.
32 MS 107, p. 233. The German reads:
In der Zeit ausgedehnt betrachtet ist die Anwendung der Worter leicht zu ver-
stehen; dagegen finde ich es unendlich schwierig den Sinn in Moment der Anwendung
zu verstehen.
Was heiBt es z.B. einen Satz als ein Glied eines Satzsystems zu verstehen?
33 G. E. Moore: 1959, 'Wittgenstein's Lectures in 1930-33', Philosophical Papers (Lon-
don: Allen & Unwin), p. 310. Originally published in three parts in Mind (1954-55). Cf.
Wittgenstein's use of the term "neighbour" in the "Notes for Lectures on 'Sense Data'
and 'Private Experience"', Philosophical Review 77 (1968), p. 297, and in the Blue Book,
p.72.
34 Cf. Zettel, #648: "One language-game analogous to a fragment of another. One space
projected into a limited extent of another. A 'gappy' space. (For 'inner and outer' .)" L.
Wittgenstein: 1967, Zettel, edited by G. E. M. Anscombe and G. H. von Wright,
translated by G. E. M. Anscombe (Berkeley and Los Angeles: University of California
Press). The first two sentences are from material published as the Remarks on the
Philosophy of Psychology I, edited by G. E. M. Anscombe and G. H. von Wright,
translated by G. E. M. Anscombe (Oxford: Blackwell, 1980), #936.
35 MS 108, p. 3; Philosophical Remarks, #54.
36 MS 107, p. 114; Philosophical Remarks, #69.
37 MS 107, pp. 1,2-3 (1929).
Die Betrachtungsweise die gleichsam in einen Talkessel des Magischen fiihrt aus
dem kein Weg in die freie Landschaft fiihrt ist die Betrachtung der Gegenwart als
des einzig Realen. Diese Gegenwart in standigem FluB oder vielmehr in standiger
Veranderung begriffen laBt sich nicht fassen. Sie verschwindet ehe wir daran den ken
konnen sie zu erfassen. In dies em Kessel bleiben wir in einem Wirbel von Gedanken
verzaubert stecken ....
Was ich nicht den ken darf, kann die Sprache nicht ausdriicken. Das ist un sere
Beruhigung.
Wenn man aber sagt: Der Philosoph muB aber in diesen Kessel hinuntersteigen
und die reine Realitat selbst erfassen und ans Tageslicht ziehen, so lautet die Antwort
daB er dabei die Sprache hinten lassen miiBte und daher unverrichteter Dinge wieder
heraufkommt.
Und doch kann es eine phanomenologische Sprache geben. (Wo muB diese Halt
machen?)
The translation of the first paragraph is based on the one in Hallett: 1977, Companion
to Wittgenstein's Philosophical Investigations (Ithaca, NY: Cornell University Press), p.
471.
38 See the discussion of this passage, MS 107, p. 205, at the beginning of this section.
39 Tractatus, 5.473.
40 Philosophical Grammar, p. 40.
41 There is thus a close affinity between Wittgenstein's logical holism, once he had given
up the goal of a phenomenological language, and what Dreyfus calls "theoretical holism",
the view that all understanding is a matter of formulating a theory.
42 On Certainty, #139. There is a valuable discussion of the tryptich formed by this
quotation and the preceding two in K. S. Johannessen: 1988, 'The Concept of Practice
in Wittgenstein's Later Philosophy', Inquiry 31, 357-69.
43 See Philosophical Grammar, #16, where Wittgenstein discusses how one might explain
the difference between the "is" of identity and the "is" of predication in these terms
[Edited by R. Rhees, translated by A. J. P. Kenny (Oxford, Blackwell, 1974)]. Cf.
Philosophical Investigations, #558-60.
44 Philosophical Investigations, #81. In both the early and intermediate versions of the
Investigations, the parenthetical phrase "(and did lead me)" read "(and did lead me
(Tractatus" .
45 See, e.g., the discussion of a schematic leaf (#73-74), following arrows (#86), the
drawing of a cube (# 139 ff.).
46 Philosophical Remarks, #149 (1930). Cf. #104, #107, #164. See also A. Ambrose:
1979, Wittgenstein's Lectures, Cambridge, 1932-1935 (Chicago: University of Chicago
Press), pp. 131-34.
47 Brown Book, p. 141 (1934-35). Cf. Philosophical Investigations, #185ff., and L.
Wittgenstein: 1970, Eine Philosophische Betrachtung, edited by Rush Rhees (Frankfurt
am Main: Suhrkamp), pp. 214-16.
48 Brown Book, p. 142.
51 L. Wittgenstein: 1969, On Certainty, edited by G. E. M. Anscombe and G. H. von
Wright, translated by G. E. M. Anscombe and D. Paul (Oxford: Blackwell), #110.
52 MS 107, p. 233 (cited at greater length above, p. 213.)
53 Philosophical Investigations, #141, #139.
54 Big Typescript, #102, p. 496. The German reads:
Von welcher Wichtigkeit ist denn diese Beschreibung des g e g e n - war t i g e n
Phanomens, die fUr uns gleichsam zur fixen Idee werden kann. Dass wir darunter
leiden, dass die Beschreibung nicht das beschreiben kann, was beim Lesen der
Beschreibung von sich geht.
55 Philosophical Investigations, # 140.
56 Cf. Philosophical Investigations, #431-2, on the idea that understanding animates
signs, and #435-6, on the idea that understanding is a hidden rapid process.
57 Philosophical Investigations, #156.
226 DAVID STERN
59 Philosophical Investigations, # 159. ,
60 Philosophical Investigations, #165, #177.
61 See M. Heidegger: 1962, Being and Time, translated by J. Macquarrie and E. Robinson
(New York and Evanston: Harper and Row), #15-16.
63 This is very clearly stated in Philosophical Investigations , # 170; also the last paragraph
of #175.
65 On Certainty, #204.
Dept. of Philosophy
University of Iowa
269 English-Philosophy Bldg.
Iowa City, IA 52242
U.S.A.
M. R. M. TER HARK
THE DEVELOPMENT OF WITTGENSTEIN'S VIEWS
ABOUT THE OTHER MINDS PROBLEM*
The aim of this article is to examine the development Wittgenstein
undergoes between 1929 (the year of his return to Cambridge) and
1951 (the year of his death) in his approach to the other minds problem.
Wittgenstein's way of dealing with this problem is in terms of an analysis
of the use of psychological concepts in the third person. However, in
contrast to his extensive treatment of psychological concepts in the first
person in Philosophical Investigations (PI) and other posthumously
published works, he remains rather reticent about the correct analysis
of third-person attributions of sensations, emotions and thoughts.
Therefore it comes as no surprise that the little that has been written
about this aspect of Wittgenstein's philosophy of mind is controversial.
For instance, some commentators deny that Wittgenstein was ever a
behaviourist and even try to explain away the apparent behaviourism
in Philosophical Remarks (PR). Others, instead, read Wittgenstein as
a behaviourist even in his later works. According to R. Fogelin, for
instance, Wittgenstein takes it for granted in Zettel par. 488 that the
third person employment of psychological concepts simply give infor-
mation that can be verified by observation and the position he ascribes
to Wittgenstein is "straightforwardly behaviouristic". 1 And to S. Kripke
Wittgenstein's 'famous slogan' in Philosophical Investigations about our
'attitude towards a soul' sounds "much too behaviouristic".
2
In support of their claims commentators often act more on a hunch
than by offering extensive exegetical evidence. Consequently, whether
or not these claims are correct can rarely be seen from the little evidence
accompanying their interpretations. My main reason, therefore, for
making a detailed study of Wittgenstein's views about the other minds
problem, i.e., his analysis of third-person attributions of mental states,
is the fact that there is an extensive array of still unpublished remarks
in the Nachlass that throw considerable light upon the precise meaning
of published remarks. On the basis of the original context of published
remarks in the Nachlass and many still unpublished remarks about
other minds I shall argue the thesis that the development of this topic
from Philosophical Remarks to Philosophical Investigations has to be
Synthese 87: 227-253, 1991.
22S M. R. M. TER HARK
viewed as a (rather slow) transformation of a behaviouristic into a non-
behaviouristic way of thinking. In particular, I shall try to show that
the apparent behaviourism of Philosophical Remarks is real and cannot
be dismissed so easily. Such is the intent of Section 1. Next I shall
argue that forcing Wittgenstein's remarks about the third person, in
Philosophical Investigations and later works, into the position of behav-
iourism is an interpretation for which there is np licence. The burden
of Section 2 is that the meaning of Wittgenstein's emphasis on our
"attitude towards a soul" can be grasped only in the context of his
repudiation of logical behaviourism in manuscripts from 1941 and
1949/50. I shall end, in Section 3, with a discussion of Wittgenstein's
last, unpublished writings about the concept of pretending, the upshot
of which will be that his way of dealing with that problem is quite at
odds with a behaviouristic ignorance of the problem.
Before I attempt these tasks, there are some preliminary points
about the chronology of Wittgenstein's writings about the other minds
problem that need to be made. The first time Wittgenstein deals, in his
published writings, with the other minds problem is in Philosophical
Remarks chapter VI. The typescript published as Philosophical Re-
marks was compiled in 1930 out of manuscripts 105, 106, 107 and the
first half of lOS. Unlike many other 'definitive' typescripts, most of
the fragments in Philosophical Remarks have come unaltered from
manuscripts 105, 106, 107 and lOS. For instance, of the thirty-four
fragments composing chapter VI, twenty-eight fragments have found
their way unaltered into Philosophical Remarks. Changes concern only
the order and sequence of fragments. In that respect the status of
Philosophical Remarks is unique for in all his later writings Wittgenstein
continually changes his formulations or thoughts. Maybe this is the
reason that G. E. Moore spoke of the 'confused,3 character of Philo-
sophical Remarks. That Wittgenstein was not very satisfied especially
with chapter VI is proved by the fact that in less than two years later,
in 1931, Wittgenstein takes up only half of the fragments (nineteen) in
a comparable chapter in Big Typescript, called "Schmerz haben". The
remaining bulk of fragments in this chapter of TS 213 were culled from
manuscripts of the years 1931 and 1932 and, as we shall see, testify of
rapid and fundamental changes in Wittgenstein's thinking about other
minds. The story of this first phase of Wittgenstein's development will
be told in Section 1.
The second phase of Wittgenstein's development in relation to the
other minds problem occurs as late as 1941, in manuscript 123. This
WITTGENSTEIN ON THE OTHER MINDS PROBLEM 229
late dating may seem surprising and consequently is in need of some
explanation. Apart from a few rather implicit references to the other
minds problem in Notes for Lectures on "Private Experience" and
"Sense Data" and in the Blue Books, the first time Wittgenstein occu-
pies himself with this problem again and especially with his own earlier
behaviouristic approach occurs in 1941. This is not to say that he wrote
nothing about the philosophy of mind before 1941. On the contrary,
many of the fragments about private language in Philosophical Investi-
gations, par. 243-421 were drafted already as early as 1937/38. But
in these paragraphs of Philosophical Investigations and the relevant
manuscripts, the emphasis is on the first person use of psychological
concepts. In MS 123 (1941), however, he devotes himself exclusively
to the third person and those remarks shed considerable light upon the
history and precise meaning of Wittgenstein's most condensed aphorism
about other minds in Philosophical Investigations part II: "My attitude
towards him is an attitude towards a soul. I am not of the opinion that
he has a soul". The most recent source of this remark is MS 169,
written in 1949. The story of this second phase of Wittgenstein's devel-
opment will be told in Section 2.
The third phase of Wittgenstein's development is, as the second, a
critique of his first phase and is especially concerned with the problem
of pretending. These remarks were written between 1949 and 1951, in
manuscripts 169, 171, 173, 174 and 176. The latter three are the same
manuscripts in which Wittgenstein writes On Certainty and Remarks on
Colour and it is striking to note how he finally applies his philosophy
of certainty and doubt to the other minds problem in general and the
problem of pretending in particular. The story of this last phase of
Wittgenstein's development will be told in Section 3.
These preliminary deliberations about the chronology of Wittgen-
stein's writings are of course no substitute for a concrete examination
of their content. Addressing this task will offer more genuine insight
into the development of Wittgenstein's treatment of the other minds
problem.
1. THE BEHAVIOURISM OF THE PHILOSOPHICAL REMARKS
In this section I shall give a strongly behaviouristic interpretation of
Wittgenstein's remarks about other minds in Philosophical Remarks
and the corresponding manuscripts. Commentators
4
have often tried to
230 M. R. M. TER HARK
explain away the apparent behaviourism in Philosophical Remarks as a
rhetorical way of endorsing some form of non-behaviourism. However,
remarks in unpublished manuscripts unambiguously show that Wittgen-
stein's endorsement of behaviourism was not just apparent, but real.
Furthermore, these remarks explain how a rather bold version of behav-
iourism was forced on Wittgenstein by assumptions concerning the Ego
and the nature of self-knowledge he shared with, James and Russell.
Wittgenstein's path to behaviourism has roughly the following form:
in the early period of his later philosophy Wittgenstein subscribes to
the rather traditional view that mental experiences are known directly
from the inside. Given this conception of thought, Wittgenstein feels a
difficulty in attributing mental states to others. But whereas he accepts,
like Russell, that knowledge of other minds is knowledge by descrip-
tion, unlike Russell, he rejects the argument by analogy and conse-
quently is forced into behaviourism.
In fact the first two remarks written by Wittgenstein about other
minds, in 1929, hint at the distinction between knowledge by acquaint-
ance and knowledge by description:
Why do I call toothache "my toothache"?
When I say of someone else, he has toothache I mean with "toothache" as it were an
abstract [AbstraktJ form what I normally call "my toothache". (MS 107, p. 199)
In what follows I shall interpret the term 'abstract' in the sense that
it has to be contrasted with the concrete and direct acquaintance we
have with our own pain. Then it will be argued that the transition from
concrete to abstract, i.e., the attribution of mental states to others,
forces Wittgenstein into behaviourism.
In an important sense the appeal to knowledge by acquaintance was
meant by James and Rusell to dismiss talk of a metaphysical Ego as
pointless. Nonetheless, RussellS continues to speak of a cognitive re-
lation between a subject and an object since acquaintance by introspec-
tion clearly implies the occurrence of a mental act distinct from the
mental event with which the act is acquainted. James is even more
explicit in his acceptance of acquaintance and his rejection of the Ego.
One passage in Principles of Psychology vol. 2 is of special importance
in this context, for there James describes how the Ego and its objects
may be in large part an artifact of the methods typically used for
studying them:
Take the pain called toothache for example. Again and again we feel it and greet it as
the same real item in the universe. We must therefore, it is supposed, have a distinct
pocket for it in our mind into which it and nothing else will fit .... Thereupon of course
comes up the paradox and mystery. If the knowledge of toothache be pent up in this
separate mental pocket, how can it be known cum alio or brought about into one view
with anything else? This pocket knows nothing else; no part of the mind knows toothache.
The knowing of toothache cum alia must be a miracle. And the miracle must have an
Agent. And the Agent must be a Subject or Ego 'out of time'. J a ~ e s , p. 5)
Toothache is also Wittgenstein's favourite example in Philosophical
Remarks. Apart from Wittgenstein's choice of this example, the use of
some of his metaphors reveals the philosophical heritage of James. In
Philosophical Remarks he warns of the misleading comparison between
a toothache and a purse (PR, par. 62) or a matchbox (PR, par. 65) -
things one typically possesses in one's pocket. The point made by
Wittgenstein is that the comparison between pain and an object easily
misleads one into thinking that the object is in possession of an Ego,
which in its turn leads to the solipsistic inability to make sense of real
feelings other than one's own. A purse can have an owner, because it
can also be possessed by none. But to speak of pain as something one
possesses is pointless precisely because it is pointless to speak of un-
owned pains:
What is essentially private, or seems, has no owner. (MS 110, p. 7; TS 213, p. 508)6
The word'!, does not refer to an owner in sentences about having
immediate experiences nor does it show that the word 'I', in this meta-
physical sense, is superfluous here. Wittgenstein therefore says that we
could have a language from which'!, is omitted from sentences describ-
ing immediate experiences. When we leave out the word'!, and instead
say 'There is toothache' we can still describe the phenomenon formerly
described by 'I am in pain'.
Now while this shows clearly that it is wrong, from Wittgenstein's
point of view, to speak of an Ego that possesses something, the pro-
posed elimination of the word 'I' is not exactly pellucid, for what is
expressed by 'There is toothache'? In Philosophical Remarks par. 58,
Wittgenstein says that 'There is pain' "could have anyone at all at its
centre ... but the one with me at its centre has a privileged status".
How are we to understand this privileged status? My answer is that this
privileged status is the same as the knowledge by acquaintance we have
of our immediate experiences. The proposition 'I have toothache' is
verified by direct acquaintance with the experience of pain, although
an identification of an owner is not required by this verification. In the
context of a discussion concerning the differences between propositions
which are conclusively verifiable and propositions which are not thus
verifiable (i.e., hypotheses), Wittgenstein gives as examples of the for-
mer:
"Do you see this indicator move; if it will have reached 10' you will feel headache". Is
such a proposition not verifiable?
Or: "The red circle you see now will gradually change into a rectangle". That also seems
indeed to be directly verifiable. (MS 107, p. 250)
'I have toothache' is a proposition conclusively verifiable by reference
to phenomenal experience. That this verification by reference to phe-
nomenal experience is a form of acquaintance by introspection is clear
from this remark:
The proposition 'A is in pain' relates undoubtedly to [bezieht sich zweifellos auf] my
experience of pain. (MS 107, p. 271)
Like Russell, Wittgenstein holds that the meaning of 'I am in pain'
consists of a relation and thus a relation between a subject and an
object. Equally like Russell, he holds that the relation between subject
and object is incorrigible ("zweifellos"). Together these two features
constitute the nature of acquaintance by introspection. Moreover, this
adherence to knowledge by acquaintance is also well conveyed by
Wittgenstein's own retrospective account of it in 1932:
One might (falsely) conceive of the matter thus: The question "How do you know that
you have toothache" is therefore not being posed, because one experiences this from the
toothache (itself) at first hand, whereas that a person is in the other room, is experienced
from second hand, for instance through a voice. The former I know via immediate
observation [unmittelbare Beobachtung], the latter 1 experience indirectly. Thus: "How
do you know that you have toothache" - "I know it, because I have it" - You learn it
from the fact, that you have it; but for that must you not already know that you have it.
(MS 113, p. 104)
In this manuscript Wittgenstein for the first time
7
points out that phe-
nomenological propositions are not verifiable at all, and consequently
he abandons his earlier position. Meanwhile, however, he was still
treating the self-ascription of bodily sensations as it were based on
acquaintance by introspection.
These considerations determine our next step. We must examine how
Wittgenstein's introspective and verificationistic approach to the first
person comes to be juxtaposed with a behaviouristic approach to the
third person. In this he distances himself from Russell but conforms to
the behaviouristic doctrine to which Carnap adheres.
The main thesis defended by Carnap, in 1928,8 is that the epistemo-
logical core ("Kern") of our knowledge of other minds consists of
observations of behaviour; the content of experience is only a "Neben-
tei! vom Physischen". Only the observation of physical behaviour is
necessary for the justification of sentences about other minds. Alleged
claims to knowledge of the content of experience of other persons can
all be reduced to what Carnap considers the epistemological core. In a
second article dating from 1932
9
Carnap explicates his behaviourism in
more detail by stressing that the meaning of sentences about one's own
mind and other minds can be interpreted physically as asserting "the
existence of a physical structure characterized by the disposition to
react in a specific manner to specific stimuli". The principle behind this
translation of psychological concepts into physical language is verific-
ationistic: "A sentence says no more than what is testable about it".
Consequently, the meaning of sentences about other minds is entirely
exhausted by the observation of physical behaviour.
An important assumption made by Carnap is that ascriptions of
mental predicates in the first person and in the third person would be
symmetrical in respect of verification. First-person as well as third-
person statements obtain their "rational support" from observation
of behaviour. Although Wittgenstein never followed Carnap in his
behaviouristic translation of the first person, the same symmetrical
assumption is apparent in his early writings, the difference being only
a difference in method of verification: first-person statements are
verified by reference to phenomenal experience, i.e., via acquaintance
by introspection, whereas in the third-person case, statements are
verified by observation of physical behaviour. When I ascribe to some-
one else toothache, Wittgenstein says, I mean by 'toothache' an 'ab-
stract' of what I normally call my toothache. The contrast between
'concrete' and 'abstract' alludes to the problem how communication
between me and you about our sufferings is possible. The problem is
that, as the meaning of pain in the first person is based on a "zweiJellose
Beziehung" to the experience of pain eo ipso, one is deprived of this
privileged access in the third person. The relation between the first
person and the third person is of course symmetrical: I am acquainted
with my pain and not with his as he is acquainted with his pain and not
with mine. At best it is only possible to make an inference by analogy,
i.e., on the basis of self-observation 'concluding that the behaviour of
other people is in many ways analogous to our own and then inferring
that it must have analogous causes. Wittgenstein, however, cuts off this
route to any such argument by analogy, and therefore I shall argue, is
forced into behaviourism. I
In Philosophical Remarks par. 62, Wittgenstein opposes the following
approach to the other minds problem:
Very simply, I know what it means, that I have toothache, and when I say, that he has
toothache, then I mean, that he has now, what I had then.
In this fragment Wittgenstein does not attack the argument by analogy
as such but its more broader assumption that 'He has toothache' can
be explained by reference to what one knows from one's own case. As
Russell, a defender of the argument by analogy, puts it: "What people
would say is what we should say if we had certain thoughts, and so we
infer that they probably have these thoughts".10 Wittgenstein offers
two reasons for denying that this appeal to the same makes sense. The
first reason is rather obvious from what has been said before about the
elimination of the 'Ego'. Traditionally the difference between 'I am in
pain' and 'He is in pain' has been explained by reference to the one
who possesses pain. 'Pain' in 'I am in pain' and 'He is in pain', according
to this view, both refers to the same sensation but this same sensation
is attributed to different Egos. By precluding identification of an owner,
Wittgenstein, of course, cuts off this route. Another possibility to in-
terpret 'He is in pain' as referring to the same sensation as 'I am in
pain' is that, although these sentences are not about an Ego, they are
about a body. The meaning of 'He is in pain' is then conceived of as
locating a pain in his body. Starting from one's own case one could
then simply extend the idea of what is immediately felt into other
people's bodies. This possibility is also excluded by Wittgenstein. Be-
lieving that one can simply extend the idea of what is immediately felt
into other people's bodies is not only based on a false analogy but also
will give you only an idea of having feelings in their bodies, not of their
having feelings. In a manuscript from 1932 this is expressed very clearly:
Instead of saying the other is in pain we say "His tooth does ouch" [macht au]. "But
you don't want to deny that the other can have what you have". Surely not. I want to
know only how I have to imagine this. I know, for instance, in the case of the grey hair
that you and I could have. Also in the case of bad teeth. But when I apply this concept
in the case of pain I will never come from my pain to his but from my pain in my tooth
to my pain in his tooth. (MS 156, p. 59)
In short, 'toothache' in 'I have toothache' and 'He has toothache' is
not univocal. 'Toothache' cannot refer to the same sensation in both
sentences. It is also obvious that 'toothache' cannot refer to different
sensations in both sentences, otherwise, one has to be; in a position to
distinguish between them, and clearly, no one is. But as we have seen
in the first person, the meaning of the term is given by acquaintance
with some sensation. The consequence seems to be that the meaning
of 'He is in pain' is not at all about a sensation, but exhausted by the
observation of physical behaviour.
A possible way out of the almost inevitable behaviouristic position
would be to maintain, as is done by proponents of the argument by
analogy, that in order to verify the sentence 'He has toothache' it is
necessary to use a representation of phenomenal experience or an inner
ostensive definition of pain. Wittgenstein's insistence on acquaintance
by introspection and his rejection of the argument by analogy cuts off
this route. In Philosophical Remarks he says:
If I pity someone else, because he is in pain, I do imagine the pain, but I imagine that
I am in pain. (PR, par. 65)
This is a slightly different formulation of the thought quoted above
(see p. 230) that knowledge of other person's pain is an abstract from
knowledge of our own pain. 'Pain' is essentially private ("wesentlich
privat") and can only be known from within, via acquaintance by
introspection. The implication seems to be that pain with which I am
not acquainted is something I cannot represent. In a remark preceding
the one just quoted Wittgenstein stresses exactly that point:
A matchbox that the other has I can imagine but not someone else's pain, that is pain I
do not notice [spure]. (MS 107, p. 287)
Consequently, while communicating about our sufferings, we are
forced into the same symmetrical position: I notice my pain, but he
doesn't and I do not notice his pain, but he does. But if I can only
imagine pain with which I can be acquainted and if in that sense I
cannot imagine the pain of someone else, isn't the only conclusion to
draw that the content of experience of other persons is logically redun-
dant and epistemologically a "Nebenteil". Although not expressed in
236 M. R. M. TERHARK
these technical terms this very same thought is contained in the follow-
ing remark:
When I say 'A is in pain', then I use the representation of pain in the same way as 1 use
the concept of flowing, when 1 speak of an electric current flowing. (PR, par. 64)
In MS 107 this is followed by:
,
When suddenly hearing from the room next to us in an unknown voice the utterance 'I
am in pain' we do not understand [verstehen] it. (MS 107, p. 285)
The point of the latter remark is that it would be tempting to use a
representation of the pain as it is felt in order to understand the sound
coming from the room next to us, but that as a matter of fact it doesn't
help at all. On the contrary, in verifying the sentence no representation
of phenomenal experience is used and the only way to understand the
utterance is to observe physical behaviour and probably the circum-
stances in which it occurs. The utterance does not inform us of anything
unless we understand it, that is, unless we can test it. And we can test
it iff we know what physical behaviour would verify it. Deprived of
that sort of observation we do not understand the utterance at all.
This constraint upon communication about our sufferings forces
Wittgenstein into behaviourism. And this is what we do in fact find:
If 1 make myself understood by him in language, then it has to be understanding rver
stehen] in the sense of behaviourism. That he has understood me is a hypothesis, as is,
that 1 have understood him. (MS 110, p. 8)
In order to specify in more detail what Wittgenstein means by the term
'behaviourism' more must be said about his use of the term 'hypothesis'.
What Wittgenstein means here with hypothesis has to be opposed
with what he calls 'genuine propositions'. The latter are directly and
conclusively verifiable by reference to phenomenal experience. Hypoth-
eses, instead, are not conclusively verifiable and they are not true or
false in the same sense in which genuine propositions are. They are
probable and have to await future confirmation. Thus conceived, the
hypothetical nature of our knowledge of other minds is indeed indistin-
guishable from Carnap's logical behaviourism, according to which the
meaning of sentences about other minds has to be explained in terms
of dispositions to behave in specific ways in specific circumstances.
However, the meaning of 'hypothesis' in relation to behaviourism is
not exhausted by 'a disposition to behave'. Wittgenstein uses the term
in this context with the extra implication that error or illusion is always
possible. One of his many characterisations of 'hypothesis' is:
Propositions for which it counts that one can always be mistaken I call hypotheses. (MS
109, p. 16)
Applied to other minds it is obvious what 'possibility of error' means:
the other can always be pretending to be in pain or, simply lying. In
the language of behaviourism, which restricts itself to the observation
of outer behaviour, one cannot account for the distinction between real
pain and simulated pain and the case of a perfect imitator cannot even
be formulated consistently.ll Wittgenstein's reference to the problem
of pretending has escaped the notice of many commentators. But in
the following two fragments his use of 'hypothesis' refers to the possibil-
ity of error identified as the possibility of pretending:
(A) The two hypotheses, that other people have toothache, and that they behave just
as I do but don't have toothache, possibly have identical senses. That is, if I had, for
example learned the second form of expression, I would talk in a pitying voice about
people who don't have toothache, but as behaving as I do, when I have it. (PR, par. 64)
(B) The two hypotheses, that others have pain, and that they don't, and merely behave
as I do when I have it, must have identical senses if every possible experience confirming
the one confirms the other as well. In other words, if a decision between them on the
basis of experience is inconceivable. (PR, par. 65)
These two fragments have been the topic of much debate in the
critical literature. 12 The upshot of it is very unsatisfying. Several authors
have tried to read into (A) and (B) not only a non-behaviouristic
approach but also a rejection of behaviourism. No other evidence is
cited in support of these interpretations than (A) and (B) themselves.
However, (A) and (B) can be interpreted equally well along behaviour-
istic lines. Consequently, a decision to choose between these conflicting
interpretations has to be based upon some other evidence. This is the
reason for examining the original context of (A) and (B).
Both fragments occur for the first time in MS 107, (A) at page 270
and (b) at page 287. The context of (A) is different from the context
in Philosophical Remarks, the context of (B) is identical, but its initial
formulation is slightly different. These facts will make clear that
Wittgenstein's use of two almost identical fragments in a 'definitive'
typescript, as Philosophical Remarks is, ought to have caused more
surprise than it in fact did.
The fragment after (A) in MS 107 is the following:
(C) A proposition so conceived, that it can be true or false without any possibility of
controlling this is completely detached from reality, and doesn't function I Iworkl I any
longer as a proposition. (MS 107, p. 270; PR, XXII, par. 225)
On the basis of (A) and (C) the following behaviouristic position could
be ascribed to Wittgenstein. Detached from reality is a proposition like
this: 'One can never know what goes on in his mind for behind the
f ~ d e of his outward behaviour he can think qr feel totally different
or just nothing'. In this case our knowledge corresponds only to the
observable behaviour, not to the supposed content of experience. Even
if we made a mistake here the mistake could never be discovered which
means that one cannot speak of a mistake. The inevitable conclusion
for Wittgenstein to draw is that the sentence 'He is in pain' and the
sentence 'He merely acts as if in pain' are identical in meaning, i.e., in
both cases the epistemological core of our knoweldge consists of the
observation of behaviour.
But maybe this conclusion is too quick. For from the fact that Witt-
genstein explicitly realises this consequence of his verificationistic be-
haviourism one could be tempted to conclude that, unlike Carnap, he
is not willing to identify the two sentences. More can be learnt from
the formulation of (B) in MS 107. It is split into two parts:
(D) Is there yet no difference between the hypotheses that the others are in pain and
that they are not and just behave as I do, when I am in pain?
(E) According to my principle both assumptions must be identical in meaning if all
possible experience confirming the one confirms the other. If that is to say no distinction
is conceivable on the basis of experience. (MS 107, p. 287)
The question-form in (D) obviously refers to (A) at page 270 of the
same manuscript. Moreover the question suggests even a correction of
(A), that is the two sentences can be distinguished after all. Fragment
(E) equally suggests a correction, for it says: if every possible experience
confirming the one confirms the other, then according to the principle
of verification the two sentences must be identical. The conditional
seems to imply that it is possible to find experiences that do differentiate
between the two.
David Pears remarks in this context that the phrasing of (B) indicates
Wittgenstein's reluctance to commit himself to the identity of the two
hypotheses, but "there is no hint of any good way of avoiding the
identification". 13 However, Pears, as so many others, forgets to involve
the fragment following (B) in the discussion about pretending:
To say that the others are not in pain, presupposes that it makes sense to say, that they
are in pain. (PR, par. 65; MS 107, p. 287)
Here is a hint of avoiding the identification: to be able to speak about
pretending presupposes logically to be able to speak about genuine
cases of pain. The statement 'He seems to be in pain, but is not' has
a sense iff the statement 'He is in pain' has a sense. This logical prior
status of attributions of genuine pain implies that it inust be possible
to distinguish between cases of pretending behaviour and cases of genu-
ine pain-behaviour. The conclusion of this reconstruction can be that
Wittgenstein, in Philosophical Remarks, alternates between a behav-
iouristic ignorance of the problem of pretending and a confrontation
with the problem without committing himself either to behaviourism
or introspectionism. In any case, the fact that Wittgenstein did publish
both (A) and (B) indicates that his views still had to be cut out, not to
say they;were in a 'confused state'.
Conclusive evidence for my interpretation of (B) comes from TS 213
in which Wittgenstein takes up (B) and omits (A) and writes another
remark in which he explicitly says that we do distinguish between cases
of real pain and cases of feigned pain:
Behaviourism. "It seems to me, I am sad, my head droops so much". Why does one not
pity, when a door is unoiled and weeps in opening and closing? Do we pity the other,
that behaves as we do, when we are in pain, - on philosophical grounds, that have led
to the result that he suffers, as we? Physicists might just as well make us scared, by
assuring us, that the earth is not compact after all, as it seems, but consists of loose
particles, that wander around erratically. "But we would not pity the other, if we knew,
that it is just a doll, or that he merely feigns his pain?" Naturally - but we do have quite
specific criteria for something being a doll, or for someone feigning to be in pain and
these criteria oppose those we call criteria for something not being a doll (but for instance
a human) and for someone not feigning his pain (but really being in pain). (TS 213, p.
509)
In this important fragment
14
many new lines of thinking about other
minds announce themselves but must await further elaboration until
1941 and even 1949/50. The first line of thought is stated in the last
two sentences. The sceptical thesis that others can always feign their
pain and that the only way to find out what the real meaning of their
behaviour is, is rebutted by Wittgenstein: we do not need to infer
hypothetically from the outer to the inner because there are in fact
behavioural criteria that do distinguish between cases of feigned pain
and cases of real pain. From the fact that Wittgenstein takes up at page
507 of this chapter fragment (B) from Philosophical Remarks it can
safely be concluded that he already meant to distinguish between these
two cases in MS 107 in 1930. So concerning pretending, Wittgenstein
does not commit himself to a behaviouristic ignorance of the problem
like Carnap. As Wittgenstein lets the problem rest until 1949, I, too,
shall let it rest and return to it in Section 3.
A second line of thought prepares us for SeGtion 2. It is e x p r ~ s s e
in the first part of the quote above. There Wittgenstein suggests that
our knowledge of other minds is at a fundamental level not a question
of hypotheses and in that sense he distances himself from an intellectual-
istic tradition in which both behaviourism and introspectionism share.
Wittgenstein's question, if we pity someone else when he is in pain on
account of philosophical considerations, expresses doubt with regards
to his own earlier solution that our knowledge of other minds is a
hypothesis. This can be seen from the analogy he draws with the
physicist's notion of nature. The point of the analogy is that our ordi-
nary non-hypothetical conception of space as something we experience
which is full of matter cannot come into conflict with the hypothetical
conception of space as empty, or is a mere aggregate of electrons. They
cannot come into conflict because they are incommensurable, i.e., the
fullness of space with matter does not correspond to the fullness of
space with electrons (and thus to emptiness) in the physical conception,
but to the frequency of electrons. IS Consequently, the ordinary concep-
tion of space is as adequate as the physcial one and is certainly not
improved upon by the use of physical hypotheses. By analogy, our
ordinary conception of other minds is non-hypothetical and is not im-
proved upon by the use of hypotheses concerning inner mental states,
or as in Carnap's proposal, the neurophysiological structure underlying
dispositions to behave. This insight is as yet not further elaborated by
Wittgenstein and is still not incompatible with a form of behaviourism.
However as regards the philosophy of mind, the problem of psychologi-
cal sentences in the first person will want the centre of his attention,
especially between 1932 and 1938. Not until 1941 does Wittgenstein
resume the thread of his analysis of the third-person case.
2. "EINSTELLUNG ZUR SEELE"
Having laid the groundwork for some of his later views about the
correct analysis of the third-person case in TS 213, Wittgenstein moved
on, roughly between 1932 and 1939, to work out the correct analysis
of the first-person case. The upshot of his extensive treatment of the
first person is that self-ascriptions of pain are primarily expressions
rather than statements and that in expressing one's pain no claim to
privileged access or knowledge is made. That the analogy with natural
expressions affords Wittgenstein a means of overcoming the deadlock
in the discussion between dualists and behaviourists l simply take for
granted here.
16
But this cannot be emphasised enough: the importance
Wittgenstein attaches to natural expressions of pain is a clear manifes-
tation of his relativisation of the mainly intellectualistic approach of
logical behaviourists and dualists. Realising this pre-linguistic foun-
dation of the concept of pain implies for Wittgenstein that in an impor-
tant sense it is pointless to speak of knowledge here. My first thesis in
this section will be that in 1941 he follows the same strategy with
regard to our knowledge of other minds. More precisely: at a certain
fundamental level neither the concepts of knowledge nor of certainty
and doubt have any bearing on the sentence 'He is in pain', and
consequently, the meaning of that sentence cannot be exhaustively
equated with a claim to knowledge supported or refuted by the obser-
vation of behaviour. My second thesis will be that this relativisation of
the intellectualistic starting-point for posing the other minds problem
has as its result that the behaviouristic translatibility of psychological
concepts into physical concepts is untenable.
An important part of Wittgenstein's strategy to show that the other
minds problem is mainly the product of a detached intellectualistic view
is to abandon the symmetrical construction of the first and the third
person and to present an alternate, asymmetrical construction of their
relation. Instead of saying that both the first and the third person are
about observation and knowledge, Wittgenstein will say that only the
attribution of mental states to others is based on observation and implies
a claim to knowledge. However, his analysis does not stop here and if
it did one indeed could accuse Wittgenstein of naive behaviourism.
The following two fragments show that Wittgenstein's asymmetrical
proposal amounts to more than simply observation of behaviour in the
third-person case:
The inner is hidden from us means, it is hidden from us in the sense that it is not hidden
from him. And it is not hidden from the owner [Besitzer) in the sense, that he expresses
it and that we give credit to the expressions under certain conditions and that no mistake
exists there [und es da den [rrtum nicht gibt]. And this asymmetry of the game is expressed
by the sentence, the inner is hidden from us /the other/. (MS 169, p. 56)17
Can one know what goes on in someone else, in the way he knows it? - But how he
knows it? He can for instance express his feeling /express his experience/. Doubt about
the fact whether he really has this experience - analogous to the doubt if he has such or
such illness - does not enter into the game, and therefore it is false to say he knows
what he experiences. The other, however, can very well doubt it whether he has that
experience. Doubt does enter into the game, but precisely because of that it is also
possible that certainty [Sicherheit] exists. (MS 176, p. 47)18 ;
The most salient feature of this asymmetrical construction is that even
in the third-person case, at a certain level, mistakes are excluded.
Normally the term observation does imply the possibility of error.
Consequently, the meaning of the sentence 'He is in pain' is not exhaus-
ted by observation of behaviour. I shall argue below that a significant
part of the meaning of 'He is in pain' is constituted by what Wittgenstein
calls' Einstellung zur Seele'.
If at a certain level it is pointless to speak of mistakes in our attri-
bution of mental states to others, it seems impossible for Wittgenstein
to accept the traditional epistemological view according to which we
have 'only indirect outer evidence' of other minds. In MS 173
19
he
rejects precisely this roundabout approach to other minds which readily
leads to the sceptical conclusion that we are always in the dark about
other minds:
The distinctive feature of the inner [vom Seelischen] seems to be, that it has to be guessed
at from the outer of the person and is known only from within. But when through
accurate consideration this conception vanishes into thin air, the inner indeed has not
become the outer, but for us there is no longer direct inner evidence and indirect outer
evidence for the inner. (MS 173, p. 33)
It is not as if I had direct evidence for my inner, he only indirect evidence. But he has
evidence for it, (but) I do not. (MS 173, p. 42)
The question to be answered here is what Wittgenstein means by
'evidence'. The use of the term is not epistemological, for he has
eliminated the traditional epistemological properties of 'evidence', re-
spectively, 'direct inner' and 'indirect outer'. To the traditional epis-
temological approach Wittgenstein opposes a logical one:
What I want to say is rather, that the inner distinguishes itself for its logic from the outer.
And that it is certainly the logic which explains the picture of the inner and the outer
/the expression 'the inner'/, makes it understandable. (MS 173, p. 34)
The inner is not only connected with the outer by experience, but also logically. (MS
173, p. 36)
From a logical perspective, it seems, the fact that evidence makes
someone else's feeling merely probable is irrelevant, for Wittgenstein
says:
And the evidence, to the extent that it is uncertain, it is not, because it is only outer
evidence. (MS 173, p. 42)
That is, Wittgenstein does not use the concept of the. inner to explain
the uncertainty and unpredictability of the outer, as is usually done in
epistemology. The reverse is true: the indeterminacy of human life
provides an explanation for the use of the concepts of soul and inner.
What matters from a logical point of view is the fact "that we construct
a statement on this involved sort of evidence, and hence that such
evidence has a special importance in our lives and is made prominent
by a concept" (Remarks on the Philosophy of Psychology, II, par. 709
[RPP]). The role of the notion of evidence in the case of uncertainty
I shall postpone until Section 3, here I shall confine myself to the role
of the notion of evidence in the case of certainty. My point will be that
Wittgenstein's famous remark that our attitude towards him, as an
attitude towards a soul, has to be understood in terms of his application
of the logical distinctions between knowledge and certainty to the other
minds problem. In the following fragment he makes clear that in a
sense we cannot speak of knowing a mental state of another person
because of the role of the notion of evidence:
If 'I know .. .' means: I can convince the other if he gives credit to my evidence, then
one can say: I may be as certain [sicher] about his mood, as about the truth of a
mathematical proposition, but it is still false to say: I know his mood. (MS 174, p. 13)20
If I thus know, that he is glad, I feel myself certain, not uncertain in my satisfaction, and
that, one could say, is not knowledge. (MS 169, p. 1).
Wittgenstein wishes to say that the evidence for knowledge is of a
different kind from the 'evidence' in the case of certainty. In the latter
case he sometimes speaks of 'sure evidence' [sichere Evidenz], meaning
that it is evidence we accept as sure and that we go by in acting surely
(cf. On Certainty, par. 196). It is precisely this emphasis on acceptance,
on certainty of reacting and absence of doubt or error that Wittgenstein
wishes to convey with his metaphor 'Einstellung zur Seele'. A very first
draft of that metaphor makes this clear:
Instead of saying 'Attitude towards a soul' one could also say: 'Attitude towards a human
[zum MenschenJ'. I could always say of a human, that it is an automaton (that I could
learn at school during lessons in physiology) and yet it would not influence my attitude
towards the other. I could even say it of myself.
But what is the difference between an attitude and an opinion? I might say: the attitude
comes before opinion.
An opinion can be mistaken. But how should a mistake look like here? (MS 169, pp.
60-61)
Wittgenstein's point is that an attitude is logically antecedent to an
opinion and thus to a supposition or an hypothesis. A hypothesis about
other minds has point only where there is room for doubt and error.
An attitude is precisely the absence of such room.
However, even this emphasis on attitude can seem to be a veiled
form of behaviourism, as Kripke thinks. Therefore, I shall trace the
very first starting point of Wittgenstein's famous dictum in his manu-
scripts for there is explicit evidence that Wittgenstein's repudiation of
logical behaviourism was a major theme in the emergence of his later
emphasis on attitudes.
According to logical behaviourism, sentences about other minds are
hypotheses to be checked by (future) experience. Supposing someone
else to be in pain means forming an hypothesis concerning a disposition
to behave in a specific manner in specific circumstances. In the following
fragment Wittgenstein wishes to stress that the behaviouristic doctrine
is a detached and purely intellectual construction which itself is only
possible on the basis of something more primitive:
How when I said: supposing someone in pain, means to suppose something that would
only be confirmed by that sort of behaviour. Such an attempt at a translation into a
behaviouristic way of expressing seems somehow foolish [kindisch]. Why? (The experi-
ence that the attempt is foolish must be taken seriously)
It is an undertaking to secure [sichern 1 something, that is secured after all. (MS 123, 23-
5-41 )21
What is secured after all comes out very clearly in a passage written
two weeks later:
If someone asks: What is the difference between the representing of a pain and pain-
behaviour, I would explain: in the former case you imagine something painful, a stabbing
pain, a feeling, let's say in the mouth !tooth/ - in the latter case a position or movement
of the body. Now it is peculiar, that when I really represent pain I do not, it is true,
represent the other in a painful position, but I pull myself a painful face. (MS 123, 3-6-
41)
The point that needs to be made about this passage is that Wittgenstein
carefully avoids both the argument by analogy as well as the behaviour-
istic translation of sentences about minds, without professing himself
to offer a third theory. What he has to offer is a careful description of
the way in which we set up and maintain pain-language in real life. In
real life we do not regard sentences about other minds as 'assumptions'
or 'hypotheses' to be checked against experience, neither as hypotheses
in the sense of inferences by analogy nor as hypotheses concerning
future behaviour. The way we act, or better, react" shows what is
secured after all when ascribing pain to others. Consequently, the
meaning of psychological concepts in the third person is not entirely
exhausted by their reference to observable behaviour. The following
remark is rather illuminating in this respect:
The objection against a behaviouristic way of expressing of propositions about the im-
mediate experience is not that his way of expressing would not be about experiences but
about something else. But that actually we playa different /somewhat different/ game
with the expressions of experiences in comparison with the descriptions of behaviour. -
Not that is an objection, that this way of expressing is about outward behaviour, for
what it is about shows itself not unconditionally in the propositions and their ostensive
definitions, but in the system of the use of the propositions. If someone says anxiously:
'He complains awfully', one can say, that he does not speak of behaviour. (MS 123, 3-
6-41)
What is needed to let the problem of other minds dissolve is a
careful description of the way our propositions about other minds are
interwoven with our daily practices. Such a description will allow us to
see beyond the distinction between the inner and the outer. Wittgen-
stein's objection is that behaviourism as a matter of fact makes the
same appeal to ostensive definitions as its opponent and his criticism
focuses on the supposed use of such ostensive ceremonies. What he
wishes to stress is that we do not describe behaviour when we pity
someone else in pain. The reason we don't intend to describe behaviour
is the fact that the one in pain is expressing his pain. An expression is
not a physical property of a thing, the body, of which a physical descrip-
tion can be given. For example, if someone is wounded in his hand it
is not the hand who says that he is in pain. He says it, the person, with
his mouth and his eyes. Observing him, we do not react to the body
conceived physically as a thing, but to his expressions, to him as a
person. For instance, we do not say 'That poor hand is in pain' and
we do not console the hand, but rather look the other in the eye (ct.
PI, par. 286). So the game we are playing while watching someone in
pain is not one of describing but one of reacting to expressions of pain,
that is, our attitude towards a soul is part and parcel of the game.
The emphasis on attitude is far from having behaviouristic impli-
cations. Our immediate reactions to other people's pain shows that
inner and outer are internally related. Insofar as one treats someone
else as a human being eo ipso one treats him as somebody with an
inner. So Wittgenstein is not a metaphysical behaviourist, for far from
denying the inner he connects it morejirmly, i.e., logically to the outer.
Neither is he a logical behaviourist for the meaning of psychological
sentences in the third person in an important sense is not given by
descriptions of behaviour.
3. PRETENDING
The precise meaning of 'Einstellung zur Seele' can be grasped only if
we finally involve the problem of pretending in our discussion, for one
could still be tempted to argue as follows: even if Wittgenstein's empha-
sis on attitudes were acceptable, there would still be scope for the other
mind's scepticism. There might, in principle, still be pain-behaviour but
no pain, that is, we might, in principle, observe people's pain-behaviour
and yet not take up the appropriate attitude towards them.
22
Conse-
quently, the reference to attitudes cannot bring out the difference
between 'He is in pain' and 'He merely acts as if in pain'.
We have seen above that Wittgenstein, already in 1932, criticises a
behaviouristic perspective within which the problem of pretending can-
not even be formulated consistently. He suggested that there are quite
special criteria that distinguish between real pain and simulated pain.
In 1941, in the same manuscript in which he criticises behaviourism,
he refers to the problem of pretending in his way:
"But even if all the particularities of behaviour occurred, I could always imagine that he
is not in pain". That is what one says, and there must be a ground for it. In that a main
feature of the grammar of the expression 'being in pain' has to reside. (MS 123, 23-5-
41)
What is striking about this passage is Wittgenstein's unwillingness to
formulate the problem of pretending in traditional epistemological
terms. He seems to suggest that the problem pertains to the nature of
different language-games and is in that sense grammatical and logical.
Epistemologically the problem is formulated in terms of the evidential
relation between pain and pain-behaviour, that is the relation between
inner and outer. The consequence of this emphasis on the evidential
relation is that in the case of pretending one is readily inclined to say
that the relation between the inner and the outer is disturbed. When
people pretend, an absolute gap seems to be created between obser-
vation of behaviour and the true inner facts of behaviour. Pretending
seems to devalue, to annul, all evidence.
Wittgenstein's way out of the problem will be to say that pretending
is a different kind of language-game from the game in which someone
expresses his pain sincerely and somebody else reacts to it sympatheti-
cally. The problem now becomes, in the first place, a problem concern-
ing the logical relations between different kinds of language-games. 23
Wittgenstein's opposition to the sceptical consequences attached to
pretending can be reconstructed in three stages: (i) hidden psychic
medium is irrelevant; (ii) there is evidence for pretending; and (iii)
pretending is a language-game which logically presupposes more primi-
tive games. I shall discuss these stages in that order.
Of course, uncertainty tends to be thought of in terms of situations
where one feels that one is being deceived or duped as in (i). In these
situations somebody deliberately conceals something and disguises his
behaviour in such a way as to make his inner unrecognisable. But
Wittgenstein gives a number of instances where uncertainty is created
by exactly the opposite:
Think, that we not only don't understand the other when he hides his feelings, but also
often not, not when he is hiding them, indeed when he is doing his utmost to make
himself understood. (MS 169, p. 43)
Uncertainty here has nothing to do with a 'hidden' inner which lies
beyond one's reach, but rather with an outer which remains closed,
like a handwriting that cannot be deciphered. It is also possible to
hide one's thoughts from somebody by expressing them in a language
unfamiliar to him (RPP II, par. 564) or by withholding one's diary (Last
Writings, par. 974). These cases do not involve anything metaphysically
hidden, so that it can never be found. On the contrary, there is rel-
evance in what is being hidden here. In all these cases uncertainty
cannot be attributed to the fact that one only possesses indirect outer
evidence and that what the evidence should prove is elusive. The ex-
amples show clearly enough that it is not elusive, so that Wittgenstein
remarks:
One might even say: The uncertainty about the inner is an uncertainty about something
outer. (MS 174, p. 13)
There is evidence, or signs, for pretending (ii):
Above all has pretending its characteristic outer signs. How else could one speak of
pretending at all. (MS 169, p. 68)
Wittgenstein wishes to say that both signs for d,eception and signs for
sincerity 'inside' can be recognised on the outside. So the issue is not
an epistemological contrast between inner and outer. There must be
signs for pretending, since it would otherwise be impossible to expose
somebody as an impostor: one can only expose somebody if one knows,
no matter how, that he wears a mask. Thus Wittgenstein has no reason
to abandon his idea that inner processes too have meaning only within
the rule-guided activities of language-games. He is explicit about this
with regard to pretending:
Also what goes on in him is a game, and pretending is not present [gegenwiirtig] in him
like a feeling, but like a game. (MS 169, p. 49)
In other words, to ask whether somebody is posing or not is to ask
about the kind of language-game that is being played, not about our
knowledge of the relation between his behaviour and something inside
him.
Pretending is a different kind of language-game from the language-
game of spontaneous expression of feelings of (iii). As the Hintikkas
put it: "Wittgenstein never admits that we can, for instance, drive a
wedge between pain and pain-behaviour in primary language-games.
What happens is that another (secondary) language-game is superim-
posed on the primary one". 24 The following two remarks seem to sup-
port this interpretation:
The expressions of my feelings can be unreal [unecht]. In particular they can be feigned.
That is a different language-game from the primitive language-game of genuine ex-
pressions. (MS 169, p. 63).
Thus I wish to say, that there is an original, genuine expression of pain; thus that the
expression of pain is not to the same extent [gleichermassen] connected with pain and
pretending. (MS In, p. 1)
That the language-game of pretending is a different game from that of
genuine expressions can be seen from the fact that a child has to learn
a great deal before it can dissemble, so that we only talk about pretend-
ing in a fairly complicated pattern of life. If the pattern were not that
complicated we should be able to imagine a newborn child dissembling.
That this cannot be imagined is also clear from the fact that language
is constitutive for pretending. For like children, animals do not have
the ability to pretend: a dog does not dissemble, nor is it sincere (PI
II, xi, p. 229).
The internal relation between pain and expression remains intact,
regardless of the possibility of pretence. This language-game does not
disturb the original pattern, but makes the concept pain more complex.
In 'I feel pain', the sincere expression, the relation between expression
and sensation, is internal; the expression is the sensation and cannot
be conceived of without the sensation, otherwise, the expression is
embedded in the language-game of pretending. The relation between
expression and feeling is not internal here; expression and sensation
can be separated. That the relation is not internal also appears from
the fact that there is no primary expression in this language-game. At
any rate 'I am merely pretending that I feel pain' is not regarded as an
expression of pretending, in the way that 'I feel pain' is regarded as an
expression of pain. There is a spontaneous expression of pain function-
ing on a more fundamental level than the language-game of pretending,
in the sense that the certainty in the sincere language-game is not
prejudiced by the possible uncertainty in the insincere language-game.
On the contrary, precisely the fact that there is room for doubt and
uncertainty in the insincere language-game implies that there is cer-
tainty, too, witness Wittgenstein's logic of 'knowing' and 'certainty'.
Just as it is meaningless to talk about doubt and therefore about know-
ing in the language-game of the first person, so it is meaningless not to
talk about certainty, if on a different level, where it is possible to talk
meaningfully about doubt and knowing: the language-game of the third
person.
However, it would be misleading to conclude from this, as the Hin-
tikkas seem to do, that the observation of hiding behaviour does not
necessarily change the evidential situation, in the sense that it is "not
harder to discover whether someone really is in pain". 25 This is mislead-
ing to say because it ignores the very special nature of the evidence for
judging whether someone really is in pain. Wittgenstein calls this sort
of evidence 'imponderable evidence' (PI II, pp. xi, 228) - such as
subtleties of glance, of tone, of gesture - and his question is: What
does imponderable evidence accomplish? To answer this question Witt-
genstein compares the evidential situation in the case of pretending
with the role of evidence in mathematical judgements and judgements
about colours. In the case of the latter two there is in general complete
agreement in judgements, whereas "There is in general no such agree-
ment over the question whether an expression of feeling is genuine or
not" (PI II, xi, p. 227). This lack of agreement does not imply that
one is never certain that someone else is pretending; what it does imply
is that some third person is not sure and that I cannot convince him
(PI II, xi, p. 227).
The remarks in Philosophical Investigations devoted to the role of
'imponderable evidence' are few in number, but in his manuscripts he
says more about it:
Why did I wish to say, '2 x 2 = 4' is objectively sure. 'This man is in pain' is subjectively
sure. (MS 169, pp. 35-36)
The meaning of 'objectively sure' is obvious. In mathematics or judge-
ments about colours disagreement may exist, however, total agreement
can be reached and is reached more often than not by giving proofs or
by checking. Why does this sort of objective certainty not exist in the
case of judgements about the genuineness of expressions of feelings?
It should be obvious by now what a wrong answer is: it would be
pointless to say that objective certainty does not exist here because we
cannot look into someone else's mind. That would be appealing to the
epistemological contrast between direct inner evidence and indirect
outer evidence, whereas Wittgenstein wants to do away with these. The
lack of objective certainty is a constitutive feature of the language-game
in question. Consequently, the presence of subjective certainty and of
uncertainty is equally constitutive. More precisely, it is a constitutive
rule for our language-games in which we judge about the genuineness
of expressions of feelings that they are based on 'fine shadings of
behaviour', like subtleties of glance. There is no uniformity or regularity
of behavior here and consequently our concepts do not have fixed
limits. Therefore, the uncertainty whether someone is dissembling his
feelings or not has not so much to do with an inner that conceals itself
behind an outer but with the connection between elastic concepts and
all but unspecifiable external circumstances. As Wittgenstein puts it:
We play with elastic, yes also flexible concepts. That means however not, that they can
be reformed as one pleases /without offering resistance/, that is that they are useless. For
if trust and mistrust were not to have a foundation in objective reality, then they would
be only of pathological interest. (MS 169, p. 37)
For instance, variability and irregularity are an essential part of the
human physiognomy, so that the concepts of emotion, which are mainly
based on facial expressions, lack focus and have a kind of elasticity.
Here evidence includes often 'imponderable evidence', that is, there is
no sharp borderline between sufficient evidence and insufficient evi-
dence. Thus there is an elastic margin in which the evidence for real
laughter is insufficient, but not so insufficient as to be evidence for the
opposite.
So what does imponderable evidence not accomplish? It does not
lead to knowledge in the sense that there are no (universally) valid
principles from which a proof can be derived:
"I am sure, he is in pain". What does that mean? How does one use it? What is the
expression of certainty in behaviour; what makes us sure? Not a proof. That is, what
makes me sure, does not make another sure. But there are limits to the discrepancy.
(MS 169, pp. 31-32)
This fragment also indicates what imponderable evidence does ac-
complish: it can make me sure. And the subjective certainty in this case
is a form of reacting, in short an attitude towards a soul:
Do not conceive of being certain as a state of mind, a kind of feeling, or something like
that. The important thing of certainty is the way of acting, not the expression of the
voice with which one speaks. (MS 169, p. 32)
CONCLUSION
My interpretation of Wittgenstein's remarks about other minds has
suggested the following view of his development from Philosophical
Remarks to Philosophical Investigations. In the former Wittgenstein
endorses a form of logical behaviourism that reminds strongly of Car-
nap's behaviourism. Together his analysis of the first-person case and
his rejection of the argument by analogy provide the main reason for
Wittgenstein's behaviouristic approach to other minds. Between 1933
and 1938 Wittgenstein devotes himself to a non-introspectionistic as
well as a non-behaviouristic analysis of the first-person case in terms of
'expressions' and consequently he attacks his own earlier analysis of
the first person in terms of acquaintance by introspection. In this period
he says almost nothing about the third person, in any case, nothing
that hints at a criticism of logical behaviourism. Probably he thought
that his criticism of acquaintance by introspection and his own construc-
tive account in terms of 'expressions' implied automatically a rejection
of behaviourism about other minds. However, not until 1941 does he
come to realise that he still has to offer arguments that demonstrate
the untenability of logical behaviourism. In the last two years of his
life he finally succeeds in giving a coherent account of our knowledge
of other minds that is compatible with his non-epistemological but
logical descriptions of language-games, forms qf life and the primacy
of human action. If these descriptions of language-games still make a
behaviouristic impression to some it is because of the character of
language-games: language-games are 'outer' phenomena. But they are
'outer' phenomena that constitute the meaning of concepts of 'inner
phenomena' .
NOTES
* I wish to thank the copyright owners of Wittgenstein's Nachlass, Prof. G. E. M.
Anscombe and Prof. G. H. von Wright for graciously permitting me to quote from
unpublished work by Wittgenstein.
For the designation of the unpublished manuscripts and typescripts I have made grateful
use of the system of reference devised and published by von Wright in his 'The Witt-
genstein Papers' in von Wright, G. H.: 1982, Wittgenstein, Oxford, Basil Blackwell.
1 See Fogelin, R.: 1976, Wittgenstein, London, Routledge and Kegan Paul, p. 176.
2 See Kripke, S.: 1981, 'Wittgenstein on Rules and Private Language', in Block, 1. (ed.),
Perspectives on the Philosophy of Wittgenstein, Oxford, Basil Blackwell, pp. 238-313;
esp. p. 303.
3 According to Moore, Wittgenstein himself used this term to refer to the remarks
published as Philosophical Remarks. See Moore's letter to Russell 13 March 1930 in
Russell, B.: 1968, The Autobiography of Bertrand Russell, 1914-1949, London, Allen &
Unwin, p. 436.
4 See, for instance, Hacker, P. M. S.: 1987, Insight and Illusion, Oxford, p. 224; Pears,
D.: 1988, The False Prison, vol. 2, Oxford, pp. 308, 309; and Bouveresse, J.: 1976, Le
My the de !'[nteriorite, Paris, p. 350.
5 See, Russell, B.: 1970, 'Knowledge by Acquaintance and Knowledge by Description',
in Mysticism and Logic, London, pp. 152-68.
6 This is a more accurate formulation than the quotation of Wittgenstein by Schlick
("immediate data have no owner"), quoted in its turn by Hacker, P. M. S.: 1975, Insight
and Illusion, 2d ed. Oxford, p. 195.
7 Usually one refers to Moore: 1955, 'Wittgenstein's Lectures in 1930-1933', in Mind 64,
p. 12, for Wittgenstein's first abandonment of verificationism. However, the remarks of
Moore give no datings and their content is only suggestive.
8 See Carnap, R.: 1928, Scheinprobleme in der Philosophie. Das Fremdpsychische und
der Realismusstreit, Berlin.
9 See Carnap, R.: 1959, 'Psychology in Physical Language', in Ayer, A. J. (ed.), Logical
Positivism, New York, pp. 165-98. Appeared originally in Erkenntnis, III (1932-33).
10 See Russell, B.: 1948, Human Knowledge. Its Scope and its Limits, New York, p. 483.
11 In his Scheinprobleme Carnap acknowledges that in the case of other minds the
possibility of error always exists. But to him this is only a problem if one makes an appeal
to the content of experience. If a case of pretending occurs, Carnap says, the core of our
knowledge corresponds to a state of affairs in reality, but not to our putative knowledge
about the content of experience. According to Carnap this proves that the content of
experience is only an eliminable "Nebenteil".
12 See Kenny, A.: 1973, Wittgenstein, London, chapter 7; and Pears, D., op. cit. p. 312.
13 See Pears, D., op. cit., p. 313.
14 Strangely enough, this fragment was not selected by R. Rhees for his compilation of
Philosophical Grammar, while it occurs in three sources: MS 114 (p. 26), TS 211 (pp.
752-53) and TS 213.
15 See Philosophical Remarks, par. 36: "The visual table is not composed out of elec-
trons".
16 For an interpretation of the first case largely based on unpublished material see ter
Hark, M. R. M.: 1990, Beyond the Inner and the Outer, Wittgenstein's Philosophy of
Psychology, Dordrecht, Kluwer Academic Publishers.
17 Von Wright underestimates the contribution of this manuscript to PI II, see von
Wright: 1982, Wittgenstein, Oxford, Basil Blackwell, p. 134. For instance, PI II, par. viii,
ix and x are here drafted for the first time. The main part of MS 169, however, is devoted
to an analysis of the concepts 'inner' and 'outer'.
18 In the manuscript of On Certainty, par. 426-637 are written. However, an important
discursive stretch of remarks concerning other minds is written between what has since
been published as On Certainty, par. 523 and 524.
19 Besides remarks on the inner and the outer this manuscript contains part III of
Remarks on Colour.
20 Besides On Certainty par. 66-193, this manuscript contains thirty pages devoted to the
inner and the outer.
21 This manuscript was written between 25 September 1940 and 6 June 1941. Between
16 May 1941 and 16 June 1941 Wittgenstein writes about 'seeing-as' and behaviourism.
22 See Locke, D.: 1968, Myself and Others, Oxford, pp. 81-84.
23 With this interpretation I follow in the main the Hintikkas' view in Hintikka, M. B.
and Hintikka, J.: 1986, Investigating Wittgenstein, Oxford, Basil Blackwell, pp. 279-84.
24 Ibid., p. 281.
25 Ibid., p. 283.
Faculty of Philosophy
University of Groningen
A-Weg 30
9718 CW Groningen
Holland
C. GRANT LUCKHARDT
PHILOSOPHY IN THE BIG TYPESCRIPT*
PHILOSOPHY AS TRIVIAL
One of the most distinctive features of Wittgenstein's 'later philosophy
is philosophy's concern with itself. What philosophy is, and what it
should be, are two questions that occupied much of Wittgenstein's
thought. It is very easy to read passages from his later works as suggest-
ing a profound cynicism about the answers that should be given to both
questions. Philosophy as traditionally conceived and practiced seems to
come in for much criticism, not to say abuse. And philosophy as it
ought to be can very easily be seen as both self-deprecating and self-
destructive. What is left for philosophy to do, once its limitations are
understood, is repair work - that of a handyman, not of an architect
or even a builder. And once those repairs are done, if they are done
well, the repairman can be dismissed. Philosophy, in short, can be
viewed as putting an end to itself.
Certainly part of traditional philosophy's appeal is its claim to univer-
sality and primordiality. Conceived as exploring and constructing foun-
dations on which most of the rest of human thought rests, it is easy
to see why philosophy's twin metaphors of itself as underlying and
overarching render it so deep, so exalted, and so important. Investiga-
ting 'being qua being' grabs the imagination, if not the soul. By contrast,
"assembling reminders for a particular purpose" must seem paltry fare,
hardly enough to make a single meal of, much less the diet of a life.
One response to such invidious comparisons between what Wittgen-
stein's philosophy has to offer, and what most of the rest of philosophy
claims to provide is to turn one's attention away from attractiveness,
toward the uncosmetic question of truth. If Wittgenstein's views as to
the nature of philosophy are true, then no matter whether anyone is
attracted to them. And no matter whether the philosopher operates at
the intellectual level of a plumber, whom one calls in to fix one's
problem, and then sees no more of. At least, following this metaphor,
philosophy would be an honest job, which someone has to do.
This is one response to the charge that Wittgenstein's view of philos-
ophy is self-deprecating and self-destructive. But it is, I believe, inap-
Synthese 87: 255-272, 1991.
256 C. GRANT LUCKHARDT
propriate. For I believe that another reading of his views is possible -
one that emphasizes the depth and importance of philosophy as it ought
to be practiced, in contrast to the false profundity promised by so many
other philosophers. According to this reading, those passages in his
writings that might be read as suggesting that philosophy is either, in
the words of recent commentators, 'superficial' or 'trivial,l or 'impo-
tent',2 and as suggesting that Wittgenstein is dismissive and cavalier
about its practice, need to be balanced against others that suggest
exactly the opposite.
Several passages in the early 100s of the Investigations can be, and
have been, read to imply the former sort of view. In the following
section of this paper, I turn to those passages in an attempt to make
the strongest case for this view. Many of these passages have the same
manuscript source, however, and so I propose in the second part of this
paper to look closely at that source, as well as some of the surrounding
passages in the Investigations, to get a fuller and different picture of
his views about philosophy. Indeed, the manuscript to which I shall
refer is entitled 'Philosophy'. It is one of some nineteen 'chapters' of
the 'Big Typescript', which Wittgenstein compiled in 1932-33 from
manuscripts he wrote between 1930 and 1932. That is to say, virtually
immediately upon his return to philosophy in 1929, Wittgenstein began
to write about the nature of philosophy itself - both as it is traditionally
conceived, and as he thought it ought to be practiced. Within four
years, this material was molded into a complete chapter of the massive
typescript that he produced. Although he was never satisfied enough
with this typescript to have it published, Wittgenstein continued to
return to it and to draw remarks from in his later writings. Indeed,
selected remarks from two large sections of it appear virtually verbatim
in the final post-war version of the Investigations. One of these sections
is the chapter entitled 'Philosophy', several remarks from which appear
in the early lOOs. It is the remainder of these remarks - those that
appear in the typescript but not in the Investigations - that I shall be
concerned with here, for many of them provide a significant context
for the remarks that made it into the Investigations. It is not that these
published remarks argue clearly for the view of philosophy as trivial
and impotent; surely many members of two generations of interpreters
have not read Wittgenstein this way. Rather, what I shall argue is that
placed back in their original context, and supplemented by other re-
marks there, it is simply impossible to construe Wittgenstein as dispar-
PHILOSOPHY IN THE BIG TYPESCRIPT 257
aging the importance either of traditional philosophy or of the new
'method' of philosophy he claimed to have found.
I should say that I am aware of a double risk in arguing this way.
First, Wittgenstein was patently dissatisfied with the 'Big Typescript',
to the extent that he did not want it published. Surely there is a risk
in using an early piece of a philosopher's work, which he worked on
to change into the later published work, to cast light on the later work,
for one is always faced with the question: If he had been satisfied with
a sentence or a remark the way it was, why did he change or eliminate
it? Second, it is arguable that as a general rule, Wittgenstein's changes
were almost always changes for the better. Anyone who has worked
with the manuscript and typescript sources cannot fail to notice that
when he went back to them, his changes almost always read better,
and usually one can appreciate exactly why he made them. (In the case
of typescripts, with which we are dealing here, the principle is even
more universal. As the typist was typing the original material, Witt-
genstein would enclose possible corrections next to it in sets of double
slashes. The words and phrases between slashes are usually so evidently
better that one wonders why he even bothered to have the original
versions typed.) So one who probes into the early versions of material
which was culled through, changed, and added to to produce the Investi-
gations faces the twin difficulties that Wittgenstein preferred the later
versions, and the changes in them are arguably better.
Nevertheless, I believe there is much of value in the 'Big Typescript',
and other material from the early 1930s. Some commentators have
made much of the differences between this material and the 'later'
philosophy, and so the term 'middle period' has been coined to desig-
nate not just a chronological, but a distinct philosophical period of
Wittgenstein's life. With a few exceptions, however, what strikes me
as noteworthy about the 'Big Typescript' is not its differences from the
later period, but its striking similarities and overlap. And in particular,
it is noteworthy that the conception of philosophy he had developed
by 1933 was not essentially changed, or added to, in the post-war
Investigations. Rather, the by now famous remarks about philosophy
in the early 100s are for the most part verbatim excerpts from the 1933
'Big Typescript'. It remains to be shown that the truncated version is
the one that allows for the interpretation that Wittgenstein is dismissive
toward past and future philosophy, and that the earlier more complete
treatment suggests something quite different.
PHILOSOPHY IN THE INVESTIGA TIONS
In considering Wittgenstein's views on philosophy, three questions need
to be distinguished: (1) What did he think of the status and importance
of the claims of traditional philosophy?; (2) What did he think of
the status and importance of Wittgensteinian critical philosophy (i.e.,
philosophy that is critical of (1))?; and (3) What, if anything, did he
think there was left for philosophy to do once itl has shown the claims
of traditional philosophy to be either false or confused or meaningless?
Certain passages in the Investigations can be read as (1) denigrating the
importance of the claims of traditional philosophy, (2) denigrating the
importance of the Wittgensteinian enterprise itself, and (3) implying
that effective criticism puts an end to the philosophical enterprise.
If one were to make the case that Wittgenstein is contemptuous of
the status and importance of traditional philosophy, there would be
perhaps no better place to begin than with Investigations 118. There
he asks whence his investigation (lit.: 'the observation') derives its
importance, since "it seems only to destroy everything interesting, that
is, all that is great and important? (As it were all the buildings, leaving
behind only bits of stone and rubble.)" The answer he gives does not
deny that his philosophy destroys; rather, it denies the importance of
the object of his destruction, claiming that it is much less significant
than one might have thought. "What we are destroying", he says, "is
nothing but buildings in the air" (Luftgebaude). The clear implication
is that whereas the results of traditional philosophy are taken by those
who formulate them to be far-reaching and important, they are in fact
insignificant - mere human constructions that are not well-grounded.
In destroying them we are not destroying anything, much less all, that
is great and important. The claims of traditional philosophy are not
important at all, according to this interpretation.
The very next remark in the Investigations seems to bear this out,
for it characterizes the objects of philosophical criticism as pieces of
'plain nonsense', and as 'bumps' on the head of understanding. The
metaphor hardly suggests something to be taken seriously - no serious
malady, which could threaten someone's life or well-being, but merely
a bump on the head, which might in time go away on its own, without
the help of a diagnostician or surgeon or therapist. Concern over these
bumps seems out of place. Further, what critical philosophy reveals is
not just nonsense, but plain (schlichten) nonsense - nonsense that again
one would be foolish to take seriously, and that one might well be
embarrassed for ever having been duped by. The philosopher who
uncovers it, like the lover of Forms who descends to the cave, must
find it tedious and boring, not to say unpleasant, having to spend his
or her time dealing with such foolishness. After all, it is with people
whose stupidity rivals that of flies that one is dealing (Philosophical
Investigations 309; hereafter, PI), and showing them the way out of
their difficulties can hardly be more intellectually than point-
ing out to the pitiful souls in the cave that they are only watching a
puppet-show. Thus it is that both the objects of philosophical criticism,
viz., the claims of traditional philosophy, as well as the advocates of
the claims, seem to be regarded by Wittgenstein not simply with scorn,
but with an attitude approaching contempt. And given this attitude, it
is easy to appreciate why the conceits of philosophy and philosophers
must be dashed utterly to the ground.
There would appear to be a problem with this view of philosophy.
For if traditional philosophy is totally valueless, then Wittgenstein's
original question remains - What does make our investigation impor-
tant, if all we are doing is destroying buildings in the air? There seems
here to be a double edge to Wittgenstein's sword: If what he is de-
stroying were important, then the question would be, why do you want
to destroy it? Isn't change, or reform, called for, rather than destruc-
tion? On the other hand, if the object of destruction is not important,
then why bother with it? Housewreckers who go after pasteboard shacks
can hardly be said to have earned a day's wages. At least those who
tear down real buildings earn their pay. That is to say, the activity of
Wittgenstein's own critical philosophy seems to be called into question.
'What makes our investigation important?' appears not so much as a
rhetorical question, but an unanswered one, or one whose answer is
- 'Nothing'. This self-denigration can be read as extending beyond the
point made above, because, on the one hand, Wittgensteinian critical
philosophy (hereafter, 'critical philosophy') appears to be a simple
matter requiring little effort, and on the other, one that is best done
quickly, and then abandoned. That is to say, critical philosophy appears
not only to criticize nothing worth criticizing, but also to require little
effort, and to have nothing left to do once it has completed its simple
but worthless task.
Professor Anscombe's translation of Luftgebaude seems to amplify
the impression that critical philosophy is easily done. Houses of cards
are virtually self-destructive, requiring only one more card, or a breeze,
or a flick of one's finger to bring them down. And even a more literal
translation suggests that critical philosophy is easy. It isn't clear how
one dismantles buildings in the air, but surely, it appears, doing so
can't require much effort. Russell's view that Wittgensteinian critical
philosophy has 'lazy consequences' and that it renders philosophy as
"at best, a slight help to lexicographers, and at worst, an idle tea-table
amusement,,3 might seem borne out as well by the last phrase of 118.
Not only do we destroy buildings in the air, but, we then "clear up the
ground of language on which they stand". At least in this passage,
Wittgenstein gives no reason for clearing the ground. Is the idea that
one is preparing it for future buildings, or does the need to do so arise
merely out of aesthetic fussiness? And although the next remark does
state that the bumps on the head of understanding "make us see the
value of the discovery", the fact that they are merely bumps surely
vitiates the value of the discovery. In 119, in fact, it is apparently not
even necessary to treat or remove the bumps. The philosopher merely
uncovers them. Presumably they'll go away on their own. PI 309, which
I have already mentioned, also seems to bear out the point about the
ease with which critical philosophy is done, but perhaps even to add
an additional flavor of intellectual superiority. The fly, stupid as it is,
or victim of the perspective it has, cannot realize the simplicity of the
solution to its difficulty. We, on the other hand, see clearly and easily
how simple it is to resolve its predicament. Just retrace your steps, and
you're free! The solution is obvious to one who sees it, and easy to
follow for one to whom it is shown.
Parts of PI 124 and 126 may very well seem to sustain this interpreta-
tion. Remark 124 says that philosophy cannot interfere with the use
of language, that it cannot ground it, and that it leaves everything as
it is. Remark 126 says: "Philosophy simply puts everything before us,
and neither explains nor deduces anything. -Since everything lies open
to view there is nothing to explain". It is easy to conclude from this
that there is little work for the critical philosopher to do. Nothing is to
be changed, or justified, or explained; the status quo is apparently to
be maintained, and one needs simply to reiterate it: "The work of the
philosopher consists in assembling reminders for a particular purpose"
(127). The proper work of the critical philosopher is to remind, not to
create, which is to say, nothing new will come out of his work. Indeed,
if something new came out of it, it wouldn't be philosophical, since
'philosophy' names "what is possible before all new discoveries" (126).
Parts of the Investigations seem to suggest that criticism is all there
is to Wittgensteinian philosophy. Since it is wholly a meta-discipline,
there would be nothing for it to criticize were it not for the errors and
confusions of traditional metaphysicians. And once it has done that, it
will vanish. Among other passages, Wittgenstein's famous remark that
the real discovery is "the one that makes me capable of stopping doing
philosophy when I want to" (PI 133) strongly suggests that the goal of
philosophy is to quit doing it. Similarly, once the philosopher has shown
the fly the way out of the bottle, it can fly on. But the philosopher's
work is done. Having achieved his goal (Ziel), he has nowhere else to
go, and nothing else to do. Indeed, Wittgenstein's use of words such
as 'goal', 'aim' and 'result' in the Investigations may also suggest that
the goal of philosophy is simply to put an end to it. The clarity we are
aiming at (anstreben), he says in 133, means that "philosophical prob-
lems should completely disappear", and it is the uncovering of plain
nonsense and the bumps on the head of understanding that are the
'results' (Ergebnisse) of philosophy.
In summary, according to the interpretation I have been outlining it
is difficult to imagine how any thinker who set out to reform his disci-
pline could set his sights any lower, or render his discipline any more
insignificant. As a potential job description, this prescription for philos-
ophy surely would be the last to be funded by any organization or
agency, for the limited scope of what it can do, the unimportance of
what it does, the wholly negative results it reaches, and the ease
with which it can be done seem to render it trivial. Following this
interpretation, David Pole's question - why not just have philosophers
take a drug that would cause them to lose interest in philosophy - is
well-taken.
PHILOSOPHY AS IMPORTANT
Parts of the Investigations can be read, as we have seen, as lending
some support to the idea that the problems of traditional philosophy
are trivial. On the other hand, there are passages, such as remark
111, that speak against this reading: "The problems arising through a
misinterpretation of our forms of language have the character of depth.
They are deep disquietudes; their roots are as deep in us as the impor-
tance of language." It is this latter kind of point - that the problems
of philosophy are anything but trivial - that is emphasized in the 'Big
Typescript' (hereafter 'BT'4) , and so it is to this typescript we must
turn if we are to find a way of reconciling remarks such as those about
Luftgebaude with those that describe philosophical problems as having
depth. The 'BT' contains, I believe, all important distinction that allows
us to see how, in one sense of the word, traditional philosophical
problems are not deep, but how, in another sense, they are. The
first half of the Luftgebaude remark - the part that asks where our
investigation gets its importance from, since it s,eems to destroy every-
thing that is great and important - occurs in the 'BT', but the second
half, which says that all we are destroying are buildings in the air, does
not. The remark that immediately precedes it, however, helps to explain
the sense in which what we are destroying is taken to be great. It reads:
(Questions of different kinds occupy us, for instance 'What is the specific weight of this
body', 'Will the weather stay nice today', 'Who will come through the door next', etc.
But among our questions there are those of a special kind. Here we have a different
experience. The questions seem to be more fundamental than the others. And now I
say: if we have this experience, then we have arrived at the limits of language.)5
I understand this to say ~ h t the sense ('experience') that there are
ordinary questions, on the one hand, and fundamental ones, on the
other, is itself an illusion. To attempt to make the distinction is to begin
to transcend the bounds of the sensible. Earlier in the 'BT' Wittgenstein
also rejects this distinction between the fundamental and the accidental.
This time it is with regard to philosophical problems. 'Roughly speak-
ing', he says,
according to the old conception - for instance that of the (great) western philosophers
- there have been two kinds of problems in fields of knowledge: essential, great, universal,
and inessential, quasi-accidental problems. And against this stands our conception, that
there is no such thing as a g rea t, essential problem in the sense of 'problem' in the
field of knowledge.
6
In denying the distinction between great and fundamental and impor-
tant problems versus the inessential and accidental ones, Wittgenstein
is denying the very distinction that would allow one to say that he
regards philosophical problems as trivial. 'Trivial' must stand in contrast
to 'non-trivial', or 'important', but Wittgenstein's point is that there
are simply problems and questions, about which no distinction in terms
of depth can be made. Hence to say that the philosophical problems
he is destroying are not important, etc., is not the same as saying that
they are unimportant or trivial. It is simply to say that they are on a
par with all others.
To deny the distinction between essential and inessential problems
is not to say that traditional philosophical problems may not be deep
in another sense, and indeed the 'BT' contains many references to
profundity and depth. What is being referred to, however, is not their
own depth, but the depth of the unease and discomfort they create.
"The effect of a false analogy taken up into language", he says, is
"a constant battle and uneasiness (as it were, a constant stimulus)".7
Elsewhere he refers to the 'irritating' character of grammatical unclar-
ity,
8
of problems as "intangibly weigh(ing) down our consciousness" ,9
of the 'uneasiness and confusion' they create,lO of solutions to the
problems "calm(ing) us after we have been so profoundly uneasy", 11
of that "particular peace of mind that occurs"12 when they are solved,
of the philosopher as "exaggerat(ing), shout(ing), as it were in his
helplessness",13 of a false picture lying "at the bottom of our feeling
of helplessness", 14 of the "conflict in which we constantly find our-
selves",15 and of the 'turbulence' of philosophical conjectures and ex-
planations.
16
All of these expressions give a very different picture, I
suggest, than that given by talk of bumps on the head. The problems
of philosophy, one might say, are deep ones of the soul, not superficial
ones of the body.
In presenting my devil's advocate interpretation of the Investigations,
I presented the problems of philosophy as occurring in others, but as
recognized and solved by the Wittgensteinian critical philosopher. But
according to much in the 'BT', this view is incorrect, in two different
ways. First, as we have just seen, the problems of philosophy are
pictured by Wittgenstein as being problematical for the very person who
falls into the traps that language sets. It is not that the Wittgensteinian
philosopher has to produce the unease, discomfort, and feeling of
helplessness for the traditional philosopher; rather he is portrayed as
already experiencing it, and crying out for help. The unease of tra-
ditional philosophy is the unease of the traditional philosopher, and
the Wittgensteinian philosopher is not there to make trouble, but to
confront it.
But even this way of putting things tends to distort, according to
much of the 'BT', for it tends to suggest that there is a difference of
person between the traditional philosopher and the critical philosopher.
But it is important to recognize that in the 'BT', the philosopher who
is bewitched and befuddled by grammatical errors is often, if not
usually, the same person as the philosopher who must discover his own
solution to his problem. It is, after all, our consciousness that is weighed
down, our feeling of helplessness, our conflict that we find ourselves
in, and we who are so profoundly uneasy. In a manuscript from the
early 1930s Wittgenstein had written: "Philosophy is an instrument
whose only use is against philosophers and against the philosophers in
US".17 And in the 'BT' he says: "Work on philosophy is - as work in
architecture frequently is - actually more of a kipd of work on oneself.
On one's own conception. On the way one sees things. (And what one
demands of them.),,18 The metaphor of the fly and fly-bottle in the
Investigations can make it appear otherwise - as if it were a matter of
us and them. We see the way out for those confused souls who have
been so stupid as to fall into the traps of language. But at least in the
early thirties, Wittgenstein saw himself and his readers as all falling
into those same traps. "Language contains the same traps for every-
one" ,19 he says in the 'BT', and it is clear that he thinks the author of
the Tractatus continues to be as vulnerable to them now as ever.
Since those who discover the way out of philosophical difficulties are
the same as those who got themselves in in the first place, it follows
that those who fall into them can be no more stupid than those who
discover the way out. In the 'BT' Wittgenstein emphasizes that philo-
sophical problems arise not from stupidity, or by mistake. His idea,
rather, is that the traps that language sets for us are among the easiest
of all for humans to fall into. In PI 109 Wittgenstein speaks of the 'urge
to misunderstand' the workings of our language, and in the 'BT' he
speaks of the 'inclination' (Neigung)2o people have to think exactly the
way they do, and thus to become "deeply imbedded in philosophical,
i.e., grammatical confusions". 21 Furthermore, he remarks how easily
all of us are 'seduced' by language. It is this very feature of language,
he says, that explains Kant's scandal:
One keeps hearing the remark that philosophy really makes no progress, that the same
philosophical problems that had occupied the Greeks are still occupying us. But those
who say that don't understand the reason it must be so. The reason is that our language
has remained the same and it seduces us into asking the same questions over and over.
As long as there is a verb 'to be' which seems to function like 'to eat' and 'to drink',
as long as there are adjectives like 'identical', 'true', 'false', 'possible', as long as one
talks about a flow of time and an expanse of space, etc., etc., humans will continue to
bump up against the same mysterious difficulties, and stare at something that no explana-
tion seems able to remove.
22
Elsewhere he asks why grammatical problems are so tough and seem-
ingly ineradicable. His answer is that this is because "they are connected
with the oldest thought habits, i.e., with the oldest images that are
engraved into our language itself (Lichtenberg)". 23 These old images
are retained and preserved in our current linguistic practices, such that
language contains an "immense network of well-kept false paths" .24
Not to tread where so many others have gone before is the difficult
thing, which is to say, what is to be expected is that we "following all
of (our) instincts (will) live within the herd that has created this
language as its proper expression". 25 As with Plato, there is in Witt-
genstein a strong linkage between the masses, custom, and error. And
because of this linkage, there can be no question of triviality. What
runs deep in us all, what is instinctive, what is contained in our history
and above all in our language, may bring us into error, but cannot be
taken lightly.
Again, as with Plato, what one faces when one reckons with philo-
sophical error is not simply an intellectual problem. According to Witt-
genstein, these errors are involved with the will. The difficulty of philos-
ophy, he says, is "not the intellectual difficulty of the sciences, but the
difficulty of a change of attitude. Resistances of the will must be
overcome".26 It is that they are so deeply felt that makes them so
difficult to solve, and so the task of the philosopher who wishes to solve
them, either in himself or in others, is by no means easy. The depth
of the commitment to errors is tracked by the deep difficulty in solving
them. Wittgenstein sees the philosopher who wants to stand against the
desires of the herd - as standing against all temptation and inclination
- as standing against what is most natural, both in himself as well as
others. In two passages at the beginning of the section of the 'BT'
labelled 'Philosophy', Wittgenstein compares the difficulty of correcting
philosophical error with that of correcting error in science:
As I have often said, philosophy does not lead me to any renunciation, since I do not
abstain from saying something, but rather abandon a certain combination of words as
senseless. In another sense, however, philosophy requires a resignation, but one of feeling
and not of intellect. And maybe that is what makes it so difficult for many. It can be
difficult not to use an expression, just as it is difficult to hold back tears, or an outburst
of anger.
(Tolstoy: the meaning (meaningfulness) of a subject lies in its being generally under-
standable. - That is true and false. What makes a subject difficult to understand - if it
is significant, important - is not that some special instruction about abstruse things is
necessary to understand it. Rather it is the contrast between the understanding of the
subject and what most people wan t to sec. Because of this the very things that are most
obvious can become the most difficult to understand. What has to be overcome is not a
difficulty of the intellect, but of the will. )27
This latter passage brings out another difference between the solutions
to philosophical and scientific problems. Not only does the critical
philosopher confront strong desires to see things certain ways, but he
faces the paradox that those things that need tp be seen are in some
sense the most obvious. "One of the greatest impediments for philos-
ophy", Wittgenstein says, "is the expectation of new, deep / / unheard
of / / elucidations". 28 The novel and deep are what is expected of any
good philosophical answer to a problem, and so when the familiar and
the simple are presented, it must seem that they can't be solutions.
And yet it is the familiar and the simple that Wittgenstein calls "the
philosophically most important aspects of things".z9 In a remark
strongly suggestive of Plato, Wittgenstein says, "Learning philosophy
is rea 11 y recollecting. We remember that we really used words in this
way".30 Unlike Plato, of course, what we are to recollect is grammar,
which we learned when we learned (in this life) to speak.
Wittgenstein gives an example of a familiar rule of grammar to
illustrate his claim that "what we do is to bring words back from their
metaphysical to the normal use in language": "The person who said
that one cannot step into the same river twice said something wrong;
one can step into the same river twice". 31 The point can of course be
extended beyond Heraclitus to all sorts of theories of identity, and
Wittgenstein himself extends it to all of philosophy: "this is what the
solution of all philosophical difficulties looks like. Their answers, if
they are correct, must be homemade and ordinary". To those who
require novel answers, answers such as these will clearly never satisfy,
and so it is difficult to convince anyone of them.
Another factor that Wittgenstein speaks of as making the task of
philosophical correction anything but easy stems from his belief that
philosophical problems involve feeling as much as intellect. Because of
this, the philosopher is required not just to point out their errors to
those (including himself) who are philosophically confused, but to re-
solve their uneasiness and deep disquietudes as well. It is not enough,
so to speak, to point out that the person has fallen into a trap, for the
trap must be removed and the discomfort alleviated. And this means
that the person in the trap must acknowledge the solution to his problem
as being the correct one. Without acknowledgment the philosopher
may have solved a problem of the intellect, but will not have solved
the one involving the will. "One of the most important tasks", Witt-
genstein says,
is to express all false thought processes so characteristically that the reader says, "Yes,
that's exactly the way I meant it." To make a tracing of the physiognomy of every error.
Indeed we can only convict someone else of a mistake if he acknowledges that this
really is the expression of his feeling.
For only if he acknowledges it as such, is it the correct expression. (Psychoanalysis.)
What the other person acknowledges is the analogy I am proposing to him as the
source of his thought. 32
The reference here to psychoanalysis brings to mind the talk in the
Investigations of philosophy as therapy, but the point made here is one
that is not made there - viz., that a doctor's cure is not a cure at all if the
patient doesn't respond. A cure that doesn't work, however satisfying it
may be to the person who administers it, is simply not a cure, and so
the philosopher must be sensitive whether the unease the philosophical
difficulty has brought about has been removed - to whether, in Wittgen-
stein's words, "complete satisfaction comes". 33 This only happens, he
says, when "no question remains".
Throughout the 'BT' Wittgenstein suggests several strategies for 'cur-
ing' philosophical problems. He offers no discrete list, but the text is
studded with suggestions. Often for example, pointing to an analogy
that has been followed, but that was either followed too far, or was
incorrect to begin with, or that was not recognized as an analogy will
bring about a cure:
If I correct a philosophical mistake and say that this is the way it has always been
conceived, but this is not the way it is, I always point to an analogy that was followed,
and show that this analogy is incorrect. 34
A philosophical cure can also consist, he says, in 'rejecting false
arguments,.35 At other times it will suffice to point out that a word has
more than one meaning:
The particular peace of mind that occurs when we can place other similar cases next to
a case that we thought was unique, occurs again and again in our investigations when we
show that a word doesn't have just 0 n e meaning (or just two), but is used in five or six
different ways (meanings).36
Other times philosophical therapy requires that we 'establish a rule'. 37
An example of this might occur
if a society came together without clearly written rules, but with a need for them; indeed
also with an instinct following which they observed certain rules at their meetings; but
this is made difficult by the fact that nothing is clearly expressed about this and no
arrangement is made which clarifies the rules. Thus they in fact view one of them as
president, but he doesn't sit at the head of the table and has no distinguishing marks,
and that makes doing business difficult. Therefore we come along and create a clear
order; we seat the president in a clearly identifiable spot, seat his secretary next to him
at a little table of his own, and seat the other full members in two rows on both sides of
the table, etc. etc.
38
'
The philosophical difficulty with this, Wittgenstein says,
lies only in understanding how establishing a rule helps us. Why it calms us after we have
been so profoundly uneasy. Obviously what calms us is that we see a system which
(systematically) excludes those structures that have always made us uneasy, those we
were unable to do anything with, and which we still thought we had to respect.
39
At other times what is necessary is for the philosopher to discover a
single word, what Wittgenstein calls 'the liberating word' (das erlosende
Wort).40 This is the kind of word that "finally permits us to grasp what
up until then has intangibly weighed down our consciousness". Once
found, the philosopher "delivers the word to us with which one / / I / /
can express the thing and render it harmless". 41 Once the liberating
word is found, Wittgenstein suggests that using it to dispel problems
is quite easy. It is finding it that is the hard part. Apropos such words,
he compares philosophical problems to locks on safes,
which can be openedby dialing a certain word or number, so that no force can open the
door until just this word has been hit upon, and once it is hit upon any child can open
it. 42
Other times what will work to effect a philosophical cure is the
familiar "assembling (of) reminders for a particular purpose". An
example of this is:
The conflict in which we constantly find ourselves when we undertake logical investi-
gations is like the conflict of two people who have concluded a contract with each other,
the last formulations of which are expressed in easily misunderstand able words, whereas
the explanations of these formulations explain everything unmistakably. Now one of the
two people has a short memory, constantly forgets the explanations, misinterprets the
conditions of the contract, and therefore continually runs into difficulties. The other
one constantly has to remind him of the explanations in the contract and remove the
difficulty.43
Last, but surely not least, Wittgenstein speaks of developing a per-
spicuous representation. A perspicuous representation "produces just
that comprehension / / understanding / / which consists in 'seeing con-
nections'. Hence the importance of finding intermediate cases. ,,44 Our
grammar, which is said to be "lacking in perspicuity", 45 is to be "laid
out completely clearly". 46 Such a synoptic view Wittgenstein describes
as being "of fundamental significance for us". 47 It 'solves' the philo-
sophical problem which is "an awareness of disorder in our concepts"
by "ordering them". 48
All of these techniques and devices of the philosopher are described
by Wittgenstein as being of service in the dissolution of philosophical
problems - dissolution, he says, "in the actual sense of the word -
like a lump of sugar in water" .49 As I have so far presented the case,
the problems need to be dissolved because of the unease and discomfort
they bring to the person involved in them, whether that is the philos-
opher himself or another person. But it is important to note that in the
'BT', the value of philosophy is not merely therapeutic. Its importance
derives from at least two other factors.
First, philosophy can serve as much a preventive as a critical and
corrective role:
Language contains the same traps for everyone; the immense network of well-kept false
paths. And thus we see one person after another walking the same paths and we know
already where he will make a turn, where he will keep on going straight ahead without
noticing the turn, etc. etc. Therefore wherever that false paths branch off I should put
up signs which help one get by the dangerous places.
5o
A perspicuous representation in particular would seem to contain such
signs. Taking steps to avoid, and allow others to avoid, the misery of
philosophical discomfiture in the first place might even be seen as
preferable to philosophical therapy. An ounce of prevention, after all
On the other hand, in the 'Philosophy' chapter of the 'BT' Witt-
genstein makes some statements that suggest a totally non-utilitarian
justification of the value of philosophy. One of the section titles for this
chapter reads: "The method of philosophy: the perspicuous representa-
tion of grammatical / / linguistic / / facts. The goal: the transparency of
arguments. Justice". 51 'Justice' is not a word that Wittgenstein uses
often (never in the Investigations), and its striking use in this context
suggests that he saw this goal (Ziel - the same word, incidentally, he
uses in connection with the fly-bottle) of philosophy as an important
one. Indeed, in the text he says "Our only task is to be just. That is,
we must only point out and resolve the injustices of philosophy, and
not posit new parties -and creeds".52 I take it that what Wittgenstein
regards philosophy as being unjust to,ward is language, and that doing
away with those injustices is an end that is desirable in itself. In a
manuscript written in late 1930, Wittgenstein says that in our civili-
zation, "even clarity is sought only as a means to (an) end, not as an
end in itself. For me on the contrary, clarity! perspicuity are ends
in themselves". 53 With the achievement of these ends will come the
disappearance of the old parties and creeds - determinism, idealism,
solipsism, behaviorism, the picture theory, etc., together with the exag-
gerations and distortions they contain. ("It is difficult", he says, "not
to exaggerate in philosophy". 54) And that this is desirable in itself
hardly requires arguing, at least for one who is committed to truth.
NOTES
* I wish to thank Professor Robert Arrington for his comments on an early version of
this article.
1 J. F. Findlay, in his recent Wittgenstein: A Critique (London: Routledge and Kegan
Paul, 1984), states that Wittgenstein recommends superficiality in philosophy (p. 210),
and 'prefers' 'trivialization' (p. 211).
2 O. P. Baker and P. M. S. Hacker, in An Analytical Commentary on Wittgenstein's
'Philosophical Investigations' (Oxford: Basil Blackwell, 1980), express the concern that
Section 124 of the Investigations 'might be taken to intimate the impotence of philosophy'
(p. 236).
3 Bertrand Russell, My Philosophical Development (New York: Simon and Schuster,
1959), p. 217.
4 All quotations from the 'Philosophy' chapter of the 'Big Typescript' are taken from
my translation in Synthese 87, 3-22.
5 'Big Typescript', p. 7.
6 Ibid., p. 5.
7 Ibid., p. 6.
8 Ibid.
9 Ibid.
10 Ibid., p. 9.
11 Ibid., p. 10.
12 Ibid.
13 Ibid., p. 13.
14 Ibid., p. 17.
15 Ibid., p. 15.
16 Ibid., p. 19.
17 L. Wittgenstein, TS. 219, p. 11, quoted with the permission of the executors of
Wittgenstein's literary estate.
PHILOSOPHY IN THE BIG TYPESCRIPT
19 Ibid., p. 15.
20 Ibid.
21 Ibid., p. 14.
22 Ibid., p. 15.
23 Ibid., p. 14.
24 Ibid., p. 15.
25 Ibid.
26 Ibid., p. 5.
27 Ibid.
28 Ibid., p. 12.
29 Ibid.
30 Ibid.
31 Ibid., p. 8.
32 Ibid., p. 7.
33 Ibid., p. 14.
34 Ibid., p. 6.
35 Ibid.
36 Ibid., pp. 1O-1I.
37 Ibid., p. 10.
38 Ibid.
39 Ibid.
40 Ibid., p. 6.
41 Ibid.
42 Ibid., p. 11.
43 Ibid., pp. 15-16.
44 Ibid., p. II.
45 Ibid.
46 Ibid.
47 Ibid.
48 Ibid., p. 13.
49 Ibid., p. 14.
50 Ibid., p. 15.
51 Ibid., p. 9.
52 Ibid., p. 13.
271
53 L. Wittgenstein, Culture and Value, ed. G. H. von Wright, tr. Peter Winch (Chicago:
University of Chicago Press, 1980), pp. 6-7.
REFERENCES
Baker, G. P. and Hacker, P. M. S.: 1980, An Analytical Commentary on Wittgenstein's
'Philosophical Investigations', Basil Blackwell, Oxford.
Findlay, J. F.: 1984, Wittgenstein: A Critique, Routledge and Kegan Paul, London.
Russell, Bertrand: 1959, My Philosophical Development, Simon and Schuster, New York.
Wittgenstein, Ludwig: (1932-33), TS. 219 of the von Wright catalogue.
Wittgenstein, Ludwig: 1980, Culture and Value, ed. G. H. von Wright, trans. Peter
Winch, University of Chicago Press, Chicago.
Wittgenstein, Ludwig: 1991, 'Philosophy: Sections 86-93 of the so-called "Big Type-
script"', ed. Heikki Nyman, trans. C. G. Luckhardt and M. A. E. Aue, Synthese 87,
3-22.
Dept. of Philosophy
Georgia State University
Atlanta, GA 30303-3083
U.S.A.
DAVID PEARS
WITTGENSTEIN'S ACCOUNT OF RULE-FOLLOWING
Wittgenstein's treatment of rule-following in Philosophical Investi-
gations is compelling but enigmatic. It gives people a strong sense of
the direction in which it is moving and yet there is no general agreement
either about its starting-point or about its destination. The theory on
which he turns his back is agreed by all to be Cartesianism, but there
are many versions of that picture of the mind, and anyway, much
depends on the question which it is taken to be answering. Is the
problem the threat of scepticism about constancy of meaning? Or, is it
the difficulty of understanding what gives meaning a stability which is
not in doubt? Or perhaps, what needs to be explained is the credence
generally put in a person's own account of what he means. The diver-
gence between the different views of Wittgenstein's destination is
equally striking. Some take his conclusion to be that meaning is fixed
solely by agreement in judgements. Others think that he neither sought
nor claimed to have found any single criterion of the correct use of
language, but treated it as a system with many different ways of main-
taining its stability, none of which would serve in sufficiently adverse
circumstances.
In this paper no attempt will be made to deal with all these issues.
It will contain notes on several aspects of Wittgenstein's treatment of
rule-following, but no comprehensive interpretation. It will, however,
suggest a general thesis about some of the interpretations of his later
philosophy of language which have been put forward by others. Al-
though he offers no theories and confines himself to judicious descrip-
tion, there are two ideas which dominate this part of his work. One is
the idea that what at first sight looks like a report is often itself part of
the very thing that would have been reported if it had been a case of
thing and report. This idea is one of the things that lead people to
attribute a particular theory to Wittgenstein. They feel that because he
is moving in a single direction, he must end with a single theory.
One aspect of this idea is familiar. If I have told someone to add 2,
and later say, "I already knew at the time when I gave the order that
he ought to write 1,002 after 1,000," I am not reporting what went
Synthese 87: 273-283, 1991.
274 DAVID PEARS
through my mind at the time; what I mean is something like, "If I had
been asked what number should be written after 1,000, I should have
replied '1,002,.1 But two more points need to be added to this. First,
this conditional is not my hypothesis about what my reply would have
been, but a manifestation of the very intention with which, if I am
right, I gave the original instruction. Second, some distance out on the
same line there are cases where what a person; says later might be a
further development rather than a manifestation of something that
already existed at the time. There is an intermediate example, halfway
to that point, near the end of Part I of Philosophical Investigations:
"You said, 'It'll stop soon.' - Were you thinking of the noise or of your pain?" If he
answers "I was thinking of the piano-tuning" - is he observing that the connexion existed,
or is he making it by means of these words? - Can't I say both? If what he said was true,
didn't the connexion exist - and is he not for all that making one that did not exist?2
The idea - that what looks at first sight like a report is often itself
part of the very thing that would have been reported if it had been a
case of thing and report - is obviously a strong antidote to Cartesianism.
What is not so obvious is that it is closely connected with a second
leading idea in Wittgenstein's later writings. That is the idea that what
looks like a description of a sensation is often (but not always) a form
of words substituted early in the speaker's life for the natural expression
of the sensation which would have been described if it had been a case
of thing and description. This is, of course, a different idea and it
produces a different result when it is applied to intentions: the natural
manifestation of an intention is appropriate action, and so, if at first
sight it is surprising that we have non-inferential knowledge through
our intentions of our own futures, it becomes less surprising when we
realise that the tendency to use the right words in advance is an offshoot
of the tendency to do the right thing later. But though this is a different
idea, it is connected with the first idea and reinforces it by showing
how, in certain cases, language is grafted onto pre-existing patterns of
behaviour. The two ideas operating together make a powerful combi-
nation, and they give Wittgenstein's later philosophy of language a
unity which is easily mistaken for the unity imposed by a single theory.
It is often instructive to go back to an earlier stage in the development
of his philosophy and to look for the first emergence of a problem, or
even earlier, for the ideas in which it was lying latent. In the Tractatus
the theory of names leaves many questions unanswered but it is explicit
WITTGENSTEIN'S ACCOUNT OF RULE-FOLLOWING 275
on two points: the naming of objects is a rule-governed activity, but
the rules can never be stated but only shown by the actual application
of the names. This is not presented as an unfortunate consequence of
the picture theory, an unsolved problem to be cleared up later, but as
the central point of the early semantics:
There is a general rule by means of which the musician can obtain the symphony from
the score, and which makes it possible to derive the symphony from the groove on the
gramophone record, and, using the first rule, to derive the score again. That is what
constitutes the inner similarity between these things which seem to be constructed in such
entirely different ways. And that rule is the law of projection which projects the symphony
into the language of musical notation. It is the rule for translating this language into the
language of gramophone records.
The possibility of all imagery, of all our pictorial modes of expression is contained in
the logic of depiction.
3
However, the inner similarity or shared form is not something which
can be represented in the same way:
Propositions can represent the whole of reality, but they cannot represent what they must
have in common with reality in order to be able to represent it.
In order to be able to represent logical form, we should have to be able to station
ourselves with propositions somewhere outside logic, that is to say outside the world.
4
This is not mystery-mongering. The point is that the best that anyone
could do if he were asked to state the rule linking a name to its object
would be to say" 'a' must be used as the name of a". But though this
gestures in the direction of a rule, it does not specify it. That would be
impossible. The only way to convey the rule to someone else would be
to show him what it is by applying the name to the object:
What signs fail to express, their application shows. What signs slur over, their application
says clearly.
The meanings of primitive signs can be explained by means of elucidations. Elucidations
are propositions that contain the primitive signs. So they can only be understood if the
meanings of those signs are already known.
s
This should be compared with 201 of Philosophical Investigations:
What this shows is that there is a way of grasping a rule which is not an interpretation,
but which is exhibited in what we call "obeying the rule" and "going against it" in actual
cases.
6
The point is put very clearly in the lectures that Wittgenstein gave in
276 DAVID PEARS
Cambridge in 1930-32. He is explaining the leading ideas of the
Tractatus and the example that he uses is a sentence expressing an
expectation:
"Whatever necessary conditions I lay down for the fulfilment of an expectation must be
added into the expression of the expectation ....
What expression and fulfilment have in common is shown by the use of the same
expression to describe both what we expect and its fulfilment .... This common element
in expectation and fulfilment cannot be described or expressed in any proposition ....
What is "in common" between thought and reality must already be expressed in the
expression of the thought. You cannot express it in a further proposition and it would
be misleading to try".
7
Here as in the Tractatus what is "in common" is the form which is
shared by thought and state of affairs given the rules for the use of the
words used in the expression of the thought.
In philosophy it often happens - perhaps more often than in other
disciplines - that a change in the point of view from which a phenome-
non is described makes a problem out of something which had pre-
viously seemed unproblematical. In his early theory of language, Wittg-
enstein took the standpoint of logical analysis and asked how words
derive their meanings from things. If complex words derived, or at
least, could derive their meanings through simple words, the essential
links with the world would be revealed most clearly in the case of
simple words which had no intermediaries. It did not seem to be a
problem that the rules for the use of simple words could not be stated,
or that the rules for the use of complex words relied on those unstatable
rules. On the contrary, there would have been a problem if analysis
had not terminated in this way.
Later, when Wittgenstein took the standpoint of the philosophy of
mind and asked what we are really doing when we apply words to
things, the very thing that had seemed unproblematical became the
central problem. If the content of a rule can only be shown by someone
following it, is there any real difference between a speaker following a
rule and a rule following a speaker? Why not say that the so-called
"rule" is receiving progressive, but always incomplete, specification
from what the speaker does? There is a description of an interesting
case halfway between these two points in the long discussion of rule-
following in Philosophical Investigations:
Imagine someone using a line as a rule in the following way: he holds a pair of compasses,
and carries one of its points along the line that is the 'rule', while the other one draws
the line that follows the rule. And while he moves along the ruling line he alters the
opening of the compasses apparently with great precision, looking at the rule the whole
time as if it determined what he did. And watching him we see no kind of regularity in
this opening and shutting of the compasses. We cannot learn his way of following the
line from it. Here perhaps one really would say: "The original seems to intimate to him
which way he is to go. But it is not a rule".8
The next case out on this line would be the case of u ~ improvisation.
But before the search for the positive content of Wittgenstein's dis-
cussion of rule-following can be started, something needs to be said
about its negative aspect. What exactly is the theory that he is rejecting?
The Blue Book gives a typical example of the rejected theory. A
philosopher claims that, when we hear the word "yellow", we under-
stand it because we get images of the colour. This primitive explanation
of understanding the meaning of a general word would work only if
each of us could always derive from his image all the correct applications
of the word. But the ability to do that is an instance of the very
thing that needed to be explained and, therefore, cannot provide its
explanation.
9
Here it is worth noting that what has to be explained is
our correct understanding of the word and not our knowledge of the
way in which we take it, correctly or incorrectly, although, of course,
the second of these two achievements is contained in the first one.
Similarly, the longer and more detailed treatment of rule-following
in Philosophical Investigations is an investigation of following a rule
correctly, both when the rule is one that governs the use of a descriptive
word and when it is one that governs the development of a mathematical
series. This is enough to show that these discussions are not powered
solely, or even mainly, by interest in the question, "Why do we allow
so much authority to the rule-follower's own account of what he is
doing?" They are focussed onto the correctness of his performance and
the criteria for its correctness.
If someone understands the meaning of the word "yellow", he will
be able to pick yellow objects from a variously coloured set. If he
understands an algebraic formula, he will be able to write down the
series that it generates. Both these performances involve a direction of
fit which is the opposite of that involved in the description of an object.
So they seem to offer no foothold for Wittgenstein's idea that what at
first sight looks like the description of an inner thing may itself be part
of the very thing that would have been described if it had been a case
278 DAVID PEARS
of thing and description. However, the idea can be stretched to cover
this kind of example: an outward performance, which looks at first
sight as if it must be derived from an inner model, is really a direct
manifestation of the very same ability that would have been required
for the production of the right inner model.
The emphasis on mental pictures both in the Blue Book and in
Philosophical Investigations
lO
may seem to take the discussion away
from the philosophy of language down a disused side-track. True,
Russell had tried to revamp the old idea that images are the real bearers
of meanings,l1 but what made images a prime target for Wittgenstein's
attack was that they seemed to approximate most closely to what his
opponents wanted - "a picture the intention of which cannot be ques-
tioned, that is, a picture which we don't interpret in order to understand
it, but which we understand without interpreting it". 12 Or, to put their
ideal in another way, "the meaning mustn't be capable of interpretation.
It is the last interpretation". 13
It is often assumed that this ideal had inspired Wittgenstein's own
earlier treatment of sentences as pictures, but that is less than a half-
truth. He never took images to be the real bearers of meanings. What
he did was to generalise the concept of a picture so that it included
sentences. However, it is true that he treated the sense of a sentence
as a perfect shadow of the state of affairs that would verify it, a shadow
which could actually be seen in the sentence.
14
Or putting it the other
way round, even if a sentence could not be given a further interpreta-
tion, it could always show its sense immediately to anyone who had
grasped the rules for its useY However, it was only later that he
realised that this put all the weight on the distinction between random
and rule-governed vocalisation.
His criticism of the attempt to base meaning on images is only part
of an argument with a much wider scope: neither images nor analytical
formulae, nor anything else that might occur in the mind of a person
who understood the meaning of a word could possibly determine, by
itself, the correct use of the word. The net is cast even more widely to
include other kinds of mental event. Someone who is trying to grasp
the meaning of an unfamiliar word may suddenly feel that he has got
it. Russell regarded this as the achievement of acquaintance with a
universal. 16 Wittgenstein, naturally, acknowledges the phenomenon but
points out that these flashes of understanding cannot anticipate all the
correct uses of the word. Nothing can completely prefigure practice.
17
The positive content of his account of rule-following is more difficult
to identify. If we do not, and could not, derive the correct uses of
words from images or from any other key that might occur in our
minds, what does ensure that we use them correctly? Presumably, each
person has to use the vocabulary in the same way throughout his
speaking life, and if language is to be a means of communication, in
the same way as everyone else. But what is this using; of words in the
same way? His early theory of language had tied identity of meaning
to the identity of the unstatable rules which were supposed to link
words to things. So in his later philosophy of language he faced the
daunting task of explaining the criteria of identity of linguistic rules.
The question "What ensures that we use our vocabulary correctly?"
means "What tells us for sure that we are in that happy position?"
It is worth observing that there is another interpretation of the ques-
tion, which is recessive in his later writings. It could mean, "What
brings; it about that we use our vocabulary correctly?" The question,
taken in this way, is recessive, because he habitually avoids inquiries
which would take him across the border from philosophy into science.
However, it is not entirely absent, because at several points in his later
work there are brief expressions of the idea that language is grafted
onto pre-existing patterns of behaviour. In Lectures on the Foundations
of Mathematics he points out that human beings find it natural to
continue certain segments of open-ended series in the same way, and
that this is a fact of the greatest importance for the philosophy of
language.
1s
These natural tendencies are pre-linguistic and they can
be observed in the most primitive sorting of objects.
Near the beginning of this paper it was suggested that commentators
often show a tendency to look for a single theory in Wittgenstein's later
philosophy of language when no such theory is there to be found.
Several explanations of this tendency were offered, and another one
can now be added. When people read his elaborate answer to the first
version of the question, "What is the criterion of sameness of rule?",
their understanding of it may be influenced by what he might have
allowed himself to say in a full answer to the second version of it. The
shadow of a possible unifying scientific theory is cast on the actual
investigation of the criterion of identity and perhaps makes them see
in it a unity which is not really there.
But is it true that the investigation does not lead to the discovery of
any single criterion of identity of rule? Perhaps two people who agree
280 DAVID PEARS
in their applications of a word must be following the same rule. True,
agreement in judgements for any period of time might be followed by
divergence later. But that would still allow us to attribute to Witt-
genstein the view that this is our only criterion, and that, even if it
never gives us a decisive positive answer, it progressively approximates
to that result by testing for divergence in more and more cases. How-
ever, he only says that if language is to be a me;ans of communication,
there must be agreement in judgements,19 and he avoids committing
himself to the theory that this is the only resource available to someone
using a descriptive vocabulary. It is, therefore, reasonable to infer that
he deliberately allowed for the possibility of a solitary language-user
who would treat the truth of his predictions about his environment as
a criterion of the regularity of his usage.
20
In fact, it would be plausible
to maintain that the fulfilment of this criterion is taken for granted
when the other one, agreement in judgements, is used.
If this pluralistic treatment of criteria of sameness of rule seems no
great advance on the earlier theory, that rules can only be shown in
practice, it is worth asking what kind of advance might reasonably have
been expected. Wittgenstein describes rule-following very fully, deals
with teaching, both successful and unsuccessful, contrasts the practice
with other related but different practices, like waiting for inspiration,
and examines the phenomenology in order to discover what it feels like
to follow a rule. If he fails to give a single decisive criterion of sameness
of rule, that maybe because the concept is more like an idea of reason,
to which the use of language approximates by eliminating divergences
without any finality. It would not be surprising if his move from seman-
tics to the philosophy of mind led to a holistic application of his orig-
inally atomistic concept of showing.
I shall end this paper with a note on immunity from mistake. As
already explained, grasping the meaning of a descriptive word does
not, in general, involve immunity from mistakes in the use of it. It is
only certain inward-looking performances which exhibit this immunity,
and the most important one for rule-following is expressing an inten-
tion. A learner may try, but fail to follow a rule correctly, but he cannot
try, but fail to express the intention which he mistakenly believed that
the rule required him to form.21 To put the point in another way, we
bow to the authority of his sincere expression of his intention in a case
like this, and take sincerity to entail truth. One aspect of this important
doctrine has already been mentioned. If we ask what makes it possible
for a person to express his intention without any possibility of error,
the answer is not that he turns his attention inwards, considers it
and finds the correct description of it obvious?2 It is, rather, that his
expression of it is a manifestation of the same tendency that will later
produce the intended compliance with the rule. That, of course, is an
answer to a question on the borderline between philosophy and science,
like the observation that the use of the word "pain" is grafted onto the
natural expression of pain. The special feature of the use of the word
"intention" is that it is grafted onto a tendency to do something later
and it refers to that later performance. But though it is a more com-
plicated case than "pain", it has the same general structure: language
is based on an established pattern of pre-linguistic events.
However, Wittgenstein, as already remarked, does not pursue this
line of inquiry. His concern is with the other question, "What are the
criteria for ascribing intentions?" This is connected with the question
of immunity from mistake in the following way: if it really were possible
for a rule-follower's sincere expression of his intention to be mistaken,
that could only be because what he said might turn out to be false by
some independent criterion. If no such independent criterion can be
found, sincerity will entail truth.
This is the principle that powers the Private Language Argument. If
a philosopher suggests that his pain may be unlike other people's pain
in spite of the fact that he and they have been taught the use of the
word "pain" in the usual way, so that their reports of their own pains
are equally authoritative, then he must be supposing that there is
another criterion for this type of sensation, independent of all the
ordinary criteria. So we can challenge him to produce it. The application
of the principle is very clear in this case: if a mistake of the kind
suggested really were possible, then there must be some new criterion,
independent of all the ordinary criteria, which would show when such
a mistake had been made. But no such criterion can be found, and
without it, the suggestion collapses.
But the application of the principle is much less clear when the
independent criterion required for the possibility of a mistake is not a
new criterion, independent of all the ordinary criteria, but an ordinary
criterion which is independent of the resources used by the person
himself. It is, for example, arguable that people can make mistakes
about their own intentions when they are more complex and more
emotionally charged than the intention to follow a rule in a certain
282 DAVID PEARS
way, but in a problematical case it is not easy to judge the independence
of the criterion that would be used to establish that such a mistake had
actually been made. If someone does not carry out a project about
which he might well have had inhibitions, when would this behavioural
criterion be sufficiently independent of resources available to him in
advance, so that his original expression of his intention could be re-
garded as mistaken rather than insincere?
There is no easy answer to this question, but the difficulty is under-
estimated in many discussions of immunity from error. There seem to
be two reasons for the underestimate. One is that in the limiting case
of immunity from error the problem does not arise. If I feel a pain, it
is mine, and even someone who believed that I might conceivably be
mistaken about the type of the sensation could hardly claim that I might
have mistaken the sufferer for myself.23 I do not turn my attention
inwards in order to find out whether the affected subject is my ego or
someone else's. I do not even choose the mouth which uses the word
"I". In fact, nothing would be lost if this word were omitted from my
report and I merely said "There is pain here". For this is a case where
the grafting of language onto the pre-established sequence of events
adds nothing significant to it. If I could choose the mouth that spoke,
I would have a new way of faking, but I still could not make a mistake
about my ownership, unless there were some more radical change in
the way in which my sensations came to me. Wittgenstein, as usual,
avoids the scientific question, "How do we come to speak a language
with this particular immunity from error?", and merely asks what I
would say in circumstances strange enough to put a strain on the
language, but not so far out that they would make it unusable. His
treatment of the self-ascription of sensations is an extreme application
of the idea that what looks at first sight like a report of a thing - in
this case, the relation of ownership - is really an exemplification of it.
However, this is the limiting case of immunity from error, and when
we take up the question of a sensation's type, we find a very different
situation. If I cannot be mistaken about certain types of sensation,
when I have them, it is certainly not for this reason. But the existence
of the limiting case may well have led people to underestimate the
complexity and difficulty of the problem in the central cases.
The other factor which has contributed to this underestimation is the
failure to appreciate the difference between claiming that a person may
make a mistake about his own intention and claiming that one person's
pain might be quite different from another's in spite of the fulfilment
of all the ordinary criteria of identity in both cases. The first of these
two claims, unlike the second, does not require a new criterion of
identity, independent of all the ordinary criteria. But this crucial differ-
ence is easily overlooked, especially if the disputants use the customary
Wittgensteinian shorthand and discuss the question whether sensations
can, or cannot, be treated like objects, and steer the disqussion towards
cases where they are expressed rather than described.
1 Philosophical Investigations, I, 187.
2 Ibid., I, 682.
NOTES
3 Tractatus Logico-Philosophicus, 4.0141-4.015.
4 Ibid., 4.12. Cf. letter to Russell 19 August 1919, para. 5, (Notebooks 1914-1916, p.
130). ,
5 Ibid., 3.262-63.
7 Wittgenstein's Lectures, Cambridge, 1930-32, pp. 35-36.
9 The Blue and Brown Books, pp. 11-12.
10 Philosophical Investigations, I, 139-41.
11 Russell: 1956, 'On Propositions', in Essays in Logic & Knowledge, ed. R. C. Marsh,
London: the article was written in 1919.
12 The Blue and Brown Books, p. 36.
13 Ibid., p. 34.
14 Ibid., pp. 35-36.
15 Tractatus Logico-Philosophicus, 4.023.
16 Russell: 1912, The Problems of Philosophy, WiIIiams and Worgate, London, Ch. V.
17 Philosophical Investigations, I, 195-97.
18 Wittgenstein: Lectures on the Foundations of Mathematics, pp. 182-84.
20 See D. F. Pears, The False Prison, Vol. II, Ch. 14.
21 This is not to say that a person can never make a mistake, which is not purely verbal
about his own intention.
22 This is not to say that introspection is never needed by a person whose intention is
questioned.
23 See the discussion of ownership of sensations in Philosophical Remarks, 57-66.
Christ Church
Oxford University
Oxford, OX1 1DP
England
MICHAEL LEE KELLY
WITTGENSTEIN AND MAD PAIN
1. LEWIS ON MAD PAIN
Despite Wittgenstein's most valiant efforts in Investigations, Zettet,I
and other scattered writings, philosophers continue to be philosophers
and speak utter nonsense, or to be charitable, to speak what Witt-
genstein would call utter nonsense. I have in mind here what has been
written on pain, though of course the list can be multiplied greatly. My
target will be David Lewis's "Mad Pain and Martian Pain",
2
but I shall
take him on in a rather narrow way. What I mean by this is that I shall
not attack the theoretical claims he makes in his article for his particular
"materialist theory of mind". Rather, I shall try to apply some Wittg-
ensteinian therapy on a claim for which he gives no argument, for
which he admits he has no argument, but also for which he thinks it is
unimportant whether an argument is supplied. Let me stop being so
mysterious and allow Lewis to speak for himself.
He writes:
There might be a strange man who sometimes feels pain, just as we do, but whose pain
differs greatly from ours in its causes and effects. Our pain is typically caused by cuts
and burns, pressure and the like; his is caused by moderate exercise on an empty
stomach. Our pain is generally distracting; his turns his mind to mathematics, facilitating
concentration on that but distracting him from anything else. Intense pain has no tendency
whatever to cause him to groan or writhe, but does cause him to cross his legs and snap
his fingers. He is not in the least motivated to prevent pain or get rid of it. In short, he
feels pain but his pain does not at all occupy the typical causal role of pain.
3
Lewis goes on to say that we would doubtless think of this man as a
madman, and so he dubs this madman's pain "mad pain". This whole
scenario strikes me at first blush as the problem of other minds in
reverse; I shall explain what I mean in a moment (again, I apologize
for being so mysterious). As I said, Lewis does not argue for his claim
that there might be such a thing as mad pain. Again I quote:
I said there might be such a madman. I don't know how to prove that something is
possible, but my opinion that this is a possible case seems pretty firm. If I want a credible
theory of mind, I need a theory that does not deny the possibility of mad pain. I needn't
Synthese 87: 285-294, 1991.
286 MICHAEL LEE KELLY
mind conceding that perhaps the madman is not in pain in quite the same sense that the
rest of us are, but there had better be some straightforward sense in which he and we
are both in pain. 4
Now, if a student of mine wrote in a paper, "I don't know how to
prove this, but my opinion that it's possible is pretty firm", I would
comment snidely on his paper and mark off severely. But Lewis is an
important philosopher and not a student of mind so I shall simply
comment here on his claim that, because he is of the firm opinion that
mad pain is possible, a credible theory of mind must not deny the
possibility of it. I take this as analogous to the claim that, because an
important philosopher was of the firm opinion that it is possible that
space is absolute, a credible theory of physics must not deny the possi-
bility of it. This analogy speaks for itself.
5
Now, what does all of this have to do with Wittgenstein? Well, clearly
it can be made to have a lot to do with Wittgenstein, for certainly he
had a lot to say about pain, and unless I, and many others, read him
incorrectly, he would want to say, if not what Lewis says about mad
pain is outright nonsense, at least there is something terribly wrong
with it. What, if anything, is terribly wrong with it is what I shall be
examining in this paper.
2. THREE INTERPRETATIONS OF MAD PAIN
First let me explain what I mean when I say that Lewis's story of mad
pain strikes me as the problem of other minds in reverse. Whereas the
problem of other minds deals with the possibility that others who act
like we do when we feel pain really do not feel pain, Lewis's story deals
with the possibility that others who act nothing like we do when we
feel pain nevertheless really do feel pain. So far, this fits in with some
of the traditional accounts of the problems of other minds. Here is how
it differs: the problem of other minds is usually at least partially an
epistemological problem, How do we know that he really feels pain?
Or, how do we know that when she acts as though she feels pleasure
she isn't really feeling what we feel when we're in pain? But Lewis, at
least at the beginning of his article, drops the epistemological problem.
Presumably we all know that his sufferer of mad pain is in pain, other-
wise I see no reason why we would think that he is mad, as Lewis says
we would. It might be a bit odd if someone turned his mind intently
WITTGENSTEIN AND MAD PAIN 287
to mathematics, crossed his legs and snapped his fingers after a bit of
physical exercise on an empty stomach, but we would not call him
"mad" unless we thought he did this while feeling pain. We would just
call him a loony mathematician. But how would we know that he is in
pain whenever he goes through this series of behavior? Setting aside
this question for a moment, let me posit three interpretations of Lewis's
scenario of mad pain, going in order from the strongest to the weakest
claim (I believe that Lewis accepts the first interpretation; the other
two are possible interpretations, thus they are worth looking at):
(1) Our madman (call him "Dave") tells us that he is in pain and
does feel pain, and even though he does not act in the least like he
is in pain (indeed, when he needs to solve a math problem he works
hard to bring on more pain) we believe him; however, we do think
he is insane, because only a madman would want to induce all this
pain.
(2) The situation is the same as in (1) except here we do not believe
Dave; instead we think he is crazy, because only a madman would
claim so sincerely that he is in pain when he so clearly (or at least
we think he so clearly) is not in pain.
(3) Dave does not tell us he is in pain when he feels pain, because
he does not know that he has a pain, because he has learned to use
the word 'pain' in roughly the way we have (with one major difference
that 1 shall mention in a moment), and so attributes pain to those
who exhibit pain behavior; furthermore, he wrongly believes that he
has never felt pain, because (and here is where his learning of the
word 'pain' differs from ours) pain never served the normal causal
role for him,6 so as a child no one ever said "Oh you hurt yourself",
or asked "Where does it hurt?" etc., when he felt pain, and so he
never came to associate the word 'pain' (or the concept 'pain') with
his feeling of pain.
These are three interpretations of scenarios of mad pain; 1 am suppos-
ing for the moment that they do not contain numerous absurdities,
though they do. 1 am not supposing that they exhaust all possible
interpretations; clearly there are others, and 1 shall add to the list later,
but these three will give me something to start with.
You may be tempted to object that 1 am confusing the issue greatly.
After all, 1 pointed out that for Lewis there is no epistemological
problem (or at least, that epistemology is not his concern when it comes
to mad pain); we could say perhaps that it is more of an ontological
question. But what distinguishes my three scenarios is the difference in
epistemological relationships and what consequences arise from them.
What differs between (1) and (2) is that in (1) we believe Dave when
he says he is in pain, and in (2) we do not. What differs between (2)
and (3) is that in (3) Dave does not know or even believe that he feels
pain. If epistemological relationships do not matter, then anyone of the
above scenarios will do.
To forestall an objection, Lewis writes:
I can hear it said that I have been strangely silent about the very center of my topic.
What is it like to be the madman ... ? What is the phenomenological character of his
state? If it feels to him like pain, then it is pain, whatever its causal role or physical
nature. If not, it isn't. It's that simple!7
As if that were simple. It is not clear whether the points that this
last quotation raises should be called epistemological, but in any case,
this seems to rule out scenario (3). It seems to, but does it? Even this
is not that simple, for Lewis is not necessarily saying that madman
Dave will know that this "thing" that feels like pain, this feeling of
pain, is pain, even though it is pain. But is any of this coherent? To
whom must it feel like pain? Well, to him. Yet if he does not think
that it is pain that he is feeling, then does it make sense to say that it
feels like pain to him? And what does pain feel like? Well, surely we
all know what pain feels like; at least I do. How? Do I know it from
my own case? Pardon my lack of subtlety in doing so, but I have now
set the stage for Wittgenstein's entrance.
3. WITTGENSTEINIAN THERAPY FOR MAD PAIN
Look at Section 149 in Remarks on the Philosophy of Psychology II.
Here Wittgenstein gets anachronistically close to a direct comment on
Lewis's article.
Perhaps someone will say: How can you characterize the concept 'pain' by referring to
the occasions on which pain occurs? Pain, after all, is what it is, whatever causes it! -
But ask: How does one identify pain?
The occasion determines the usefulness of the signs of pain.
In Zettel Sec. 532-34, he writes:
The concept of pain is characterized by its particular function in our life.
WITTGENSTEIN AND-MAD PAIN 289
Pain has this position in our life; has these connexions; (That is to say: we only call "pain"
what has this position, these connexions).
Only surrounded by certain normal manifestations of life, is there such a thing as the
expression of pain.
Clearly these conflict strongly with Lewis's story of mad pain.
But of course I do not want simply to pit the words of Wittgenstein
against the words of Lewis. I want to use them to suggest that Lewis's
sentences simply do not make sense, though on the surface they may
seem to. (Let me clarify this. I am not saying that what Lewis says is
false, except in some rather broad sense of the term, and I am not
claiming that mad pain is impossible, rather I am simply saying that it
is nonsensical to say that there is any such thing; that is, I do not think
that there is any way to know what it would be like for there to be
such a thing, as Lewis describes it.)
So, how am I to use Wittgenstein here? Well, let us reexamine Zettel
Sec. 532: "The concept of pain is characterized by its particular function
in our life". Well, what particular function does mad pain have in our
lives or would it have if there were such a thing? Lewis answers this in
his original description: it turns our attention to mathematics and away
from everything else; it is not particularly unpleasant, for we do not
want to get rid of it; it makes us cross our legs and snap our fingers,
but not - I want to say, not in a pained way, but this would bias the
case in my favor - what should I say? It makes us cross our legs but
not as if we were writhing, and we do not groan as we do so.
But if all of this is so, then why do we call it 'pain', mad or otherwise
Lewis's only answer seems to be, because it feels like pain! Now, I do
not know exactly what sense to make of the peculiar phrase "feels like
pain";8 the closest phrase to this that I'm familiar with is "feels pain",
for which we have a rather clear-cut function and the function of "mad
pain" does not come close to this. It does not even have a slight family-
resemblance to it, unless the family is a white Anglo-Saxon Protestant
one that has just adopted a Zulu baby.
Here, I think, is the sort of picture Lewis has that makes him think
his account of mad pain makes sense. All of us have a clear picture,
he thinks, of what pain feels like. Take a pin and jab it into your arm:
there, that is a pain. Now take this "thing" and in your imagination
dissociate it from its causes and effects. (One of the things that Lewis
would say is an effect of pain is our writhing and groaning; I would not,
and I think Wittgenstein would not, accept this terminology. Rather I
would say that it is an expression of our pain. But remember, I am
simply describing Lewis's picture here.) His picture is much like that
given in Hume's discussion of causation. We know what a billiard ball
looks like, and we can imagine its rolling apart from its usual causes
and effects. We can imagine the ball's rolling being caused by nothing
at all, or a vocal command, and its effect upon hitting a second ball
not being that of sending the second ball off in a straight-line direction,
but rather of the first ball rolling backward over its original path.
9
Now,
if all of this makes sense with billiard balls, I think Lewis would say,
why not with pain? They are equally imaginable. We can call those
mad billiard balls, this mad pain. Remember, however, Zettel Sec. 250:
"That one can 'imagine' something does not mean that it makes sense
to say it".
One response to Lewis is to say that the grammar of objects, the
grammar of looking at and seeing objects, even the grammar of feeling
objects, is very different from the grammar of feeling pain. As Lewis
said, if it feels like pain, it is pain.
10
This cannot be said about seeing
objects or feeling objects. If I reach into my pocket and pull out a coin
that feels like a dime, it may, nevertheless, be a penny that I extract. 11
Also part of the grammar of pain is the expression of pain behavior.
The tendency, and this is what I think Lewis has done, is to make a
picture of pain: there is an outer picture, or a picture of the outer
aspects of pain, and an inner one (as if it makes sense to say we have
an inner picture of pain).12 Lewis tosses away the outer picture and
replaces it with another. It is as if he gets rid of the outer picture of
pain behavior by the rule of logical simplification, leaving just the inner
picture, and then as the rule of logical addition allows, he hooks it up
with any picture he so chooses. That is how he comes up with 'mad
pain'.
Yet I do not see how this can get us past scenario (3), since we
clearly learn how to use the word 'pain' in roughly the way Wittgenstein
describes it:
A child has hurt himself and he cries; and then adults talk to him and teach him
exclamations and, later, sentences. They teach the child new pain-behavior. (Investi-
gations I, Sec. 244)
Our sufferer of mad pain, Dave, however, has never cried when he felt
his mad pain, so he has never learned to associate the word 'pain' with
this sensation that accompanies his desire to do math. Yet if even Dave
does not think that he feels a pain, I do not see what right we have to
say that pain enters here at all.
Let me suggest a fourth scenario that does on first glance look more
plausible than the three I have already given; it seems, at least, to have
some of the grammar of pain present. Imagine that Dave at one time
was perfectly ordinary. He once felt pain as the result of cuts and burns
and pressure. He once tried to avoid pain, and would; writhe and groan
when he did feel pain. And so, he learned to use the word 'pain' in
the same sort of way that the rest of us did. But then something
happened. As the result of a neurological disease, all of this changed.
Now when he feels pain - but wait, this time I am biasing the case in
Lewis's favor; let me say instead of "now when he feels pain" (after
all that is the issue we are trying to decide - does it make sense to say
that he feels pain in the case?) let me say, now when he says things
like, "Oh, I have a pain", instead of writhing he casually crosses his
legs ..and snaps his fingers and concentrates on math, and does nothing
to avoid all of this.13 Now, shall we say that he is in pain? Rather, I
think we would say not that the neurological damage made his pain
have different causes and effects, but that the neurological damage
affected his seman tical ability: he no longer knows what pain means
(at least, not in the first person case).
Perhaps Lewis would say, "Yes, but this is really like your second
scenario: Suppose he really does feel pain, and despite his testimony
we incorrectly deny that he does". Here finally we seem to have an
honest disagreement: two interpretations of a situation. An unfortunate
ordinary fellow suffers a neurological disease. As a result, he claims he
feels pain when none of the occasions, connections, or surroundings
are present. 14 My interpretation is that, because none of these are
present, we should simply say that he has lost some of his seman tical
competence. Lewis, on the other hand, I am supposing, would say,
"No, he really does feel pain - mad pain".
So, now, do we really have an honest disagreement? If so, I shall
have to concede Lewis's point. (I am assuming that to have an honest
disagreement both positions must make sense.) I submit that I do not
have to do so.
I have good reason for saying that Dave has simply lost some of his
semantical competence: he once used the word 'pain' in the right way,
when the right surroundings were present. Now he does not. His use
of the word 'pain' is now quite esoteric (which is not to say "mad,,).15
It may have some use - we now know that when he says he is in pain
that he is doing math, etc. But I suggest that this use of the word 'pain'
has no more in common with the ordinary use of the word 'pain' than
the use of the word 'bank', meaning a financial institution, has to the
use of the word 'bank', meaning the land next to a river. 16
On the other hand, Lewis has no good reason to say that Dave is
really feeling the same sensation as he was before. ,Recall Wittgenstein's
suggestion from Investigations II, p. 207:
Always get rid of the idea of the private object in this way: assume that it constantly
changes, but that you do not notice the change because your memory constantly deceives
you.
What difference would it make if this were so, that Dave's "private
object" (his pain) constantly changed, but he did not notice this because
his memory constantly deceived him. Well, if this had some "external"
consequences, then of course that would be the difference that it makes;
but then what does the private object matter? We would know that it
was the case by some external criteria, but if there is no external effect,
then there is no practical or detectable difference (for, ex hypothesi,
there is no practical or detectable "internal" criterion, since Dave's
memory changes along with his sensation). And so, what can Lewis
mean when he says that Dave feels pain? Wittgenstein says in Investi-
gations I, Sec. 353.
Asking whether and how a proposition can be verified is only a particular way of asking
"How d'you mean?" The answer is a contribution to the grammar of the proposition.
If Lewis can give us no practical or detectable difference that would be
made if all of this were to happpen, then there can be no way of
saying whether it is true that the "private object" that is felt after the
neurological damage (Dave's supposed pain) is the same as that felt
before the damage. And if we do not know what it would be like for
it to be true or false, then there is no answer to the question, "How
d'you mean?" In other words, despite appearances, the claim means
nothing, and all of Lewis's talk of "mad pain" is, as I have held all
along, utter nonsense.
4. CONCLUSION
I have explicitly criticized Lewis in this paper, but with only minor
redirection these arguments can be brought to bear on many other
recent discussions, for instance, Thomas Nagel's "What Is It Like to
Be a Bat?" comes immediately to mind. That Lewis's and Nagel's
theoretical notions are so disparate suggests how widespread and deeply
held are certain intuitions that, if ever held up to I the appropriate
therapeutic scrutiny, would be seen to be incoherent nonsense. This
only serves to show once again the importance of Wittgenstein's
thought, and the need to meet him on his own ground.
Now, if I may be allowed a closing flight of fancy, let me announce:
"The Fight of the Century. In this corner: Ludwig Wittgenstein. In that
corner: David Lewis. The outcome: Wittgenstein wins by a knockout;
Lewis, meanwhile, sits in his corner, crosses his legs, snaps his fingers,
does math, and suffers mad pain".
NOTES
1 The works of Wittgenstein to which I refer are: 1958, Philosophical Investigations, 3d.
edition, translated by O. E. M. Anscombe, Macmillan Publishing Co., New York; 1967,
Zettel, edited by O. E. M. Anscombe and O. H. von Wright, translated by O. E. M.
Anscombe, University of California Press, Berkeley; and 1980, Remarks on the Philos-
ophy of Psychology, Vol. II, edited by O. H. von Wright and Heikki Nyman, translated
by C. O. Luckhardt and M. A. E. Aue, Blackwell, London. I refer to each by section
number, with the exception of Part II of Investigation, to which I refer by page number.
2 David Lewis: 1980, 'Mad Pair:! and Martian Pain', in Readings in Philosophy of Psychol-
ogy, Vol. I, edited by Ned Block, Harvard University Press, Cambridge, pp. 216-22.
3 Ibid., p. 216. Lewis never tells us anything about what happens to this "strange man"
when he suffers cuts, burns, pressure and the like. Lewis's wording seems to suggest
(though it is far from clear) that this man does not feel pain in such cases. I would simply
like to point out here that if he does not react as we do when in these cases we feel pain,
he will not live long to do his math, cross his legs, and snap his fingers.
4 Ibid.
5 One may object that the type of possibility that Lewis talks about is logical possibility,
and indeed our theory of physics does not rule out the logical possibility of absolute
space. But physical theories have little to say about logical possibility, and they may just
as well rule out the logical possibility of absolute space, given the general irrelevance of
logical possibility in such a theory (other than the need for the theory itself to be
consistent). What is important here is simply this: what, at some time before theorizing,
seems to be possible need not be so afterwards.
6 I am not comfortable with talk of the "causal role" of pain; it strikes me as an awkward
and probably misleading way of speaking. It is, however, Lewis's way of speaking, so in
presenting this scenario I have followed him.
7 Lewis, op. cit., p. 221.
8 For the best sense that I can make of it, see note 10.
9 See Hume's Enquiry Concerning Human Understanding, 1st ed., 1748, Section IV,
Part I.
10 If this means anything, it must mean what Wittgenstein says in Investigations Sec. 246:
It can't be said of me at all (except perhaps as a joke) that I know I am in pain [for
this read "that my sensation feels like pain"]. What is it supposed to mean - except
perhaps that I am in pain? ...
The truth is: it makes sense to say about other people that they doubt whether I am
in pain; but not to say it about myself.
11 The question of our supposed incorrigibility about pain has been much discussed, and
our perplexity over it is likely caused by our confusion about the grammar of feeling
objects with the grammar of feeling pain, and by the attendant belief that something like
what Lewis says about mad pain makes sense.
12 Lewis seems to forget that, if it does make any sense to talk of an "inner picture",
part of the inner picture is the unpleasantness of pain and our desire to avoid it.
13 I have doubts that even this scenario is neurologically or physiologically plausible, but
I posit it as Lewis's best last hope. Note, however, that as Lewis's case demands, all the
causes and "effects" of pain for this man have changed - all, except for the "effect" of
his verbal utterances and conscious awareness concerning his supposed sensations of pain.
Lewis never notes the oddity of this, that is, that all the "effects" but these are different.
14 The exception is, of course, his saying, "I feel a pain".
15 Unless Dave's neurological damage is too severe, we could persuade him that he is
not in pain at all. We would simply remind him that pain is a distracting, unpleasant
sensation that can be debilitating, and so on.
16 And surely Lewis is claiming that pain and mad pain have more in common than that
or else he would not say that a credible theory of mind must account for mad pain (in
such a way as to treat it as pain). Certainly our theories of finance need not account for
river banks.
Dept. of Philosophy
Florida State University
203 Dodd Hall
Tallahassee, FL 32306
U.S.A.
KENT LINVILLE AND MERRILL RING
MOORE'S PARADOX REVISITED
Wittgenstein "once remarked that the only work of Moore's that greatly
impressed him was his discovery of the peculiar kind' of nonsense in-
volved in such a sentence as, e.g., 'It is raining but I don't believe it'''. 1
Present practice is to refer to the difficulties generated by sentences of
this form, as well as to sentences of the form "I believe that p but not
p", as "Moore's paradox".
2
Despite Wittgenstein's great reputation
and regard for the importance of Moore's "discovery", little interest
has been generated in the topic. And yet, central issues in epistemology
and the philosophy of language are involved in the resolution of this
paradox. Since this is not generally appreciated, we begin our discussion
by establishing what some of those important issues are, thereby credit-
ing Wittgenstein's assessment of the importance of the paradox. Then,
we develop an account of the aberrant nature of Moore's sentences
(hereafter labelled "MS") that is indebted to Wittgenstein, and which
challenges the assumptions that motivate the standard form of dis-
cussion of these sentences initiated by Moore.
Moore says "such a thing as 'I went to the pictures last Tuesday, but
I don't believe that I did' is a perfectly absurd thing to say [i.e., assert],
although what is asserted is something perfectly possible logically".
3
Moore makes two claims here (besides stating the basic fact that MS
would be absurd to assert), claims which have become canonical in
subsequent literature. First, the peculiarity or absurdity of MS cannot
reside in the sentence itself, since it is not self-contradictory, and sec-
ond, that its absurdity (therefore) arises only in speech (expressing, as
it apparently does, a possibility in thought). So what puzzles Moore
and his followers is why the assertion of MS should be absurd, when
"what it asserts" might be true.
Explanations offered of that absurdity, with the exception of Wittgen-
stein's,4 rest on one or another version of the doctrine that saying or
asserting implies believing.5 To say "not-p", according to this view, is
to imply that one doesn't believe that p; this implied "I don't believe
that p" conflicts with one saying "I believe that p", thereby generating
the absurdity of such an assertion. And because "I don't believe that
Synthese 87: 295-309, 1991.
296 KENT LINVILLE AND MERRILL RING
p" is only (non-deductively) implied and not entailed, the assumed
non-self-contradictory character of MS is maintained.
Wittgenstein, on the other hand, gives a quite different account of
the absurdity of MS. On his view, '''I believe that this is the case' is
used like the assertion 'This is the case"';6 that is, "I believe that p"
is a form of the assertion that p. Hence, MS conjoins a form of the
assertion that p and an assertion that not p and i,ts absurdity arises from
that conflict.
Now that we have described these two competing accounts of how
the clauses of MS come into conflict, it is obvious each results from a
different interpretation of the function of "I believe". In Moorean
accounts, that belief phrase is assumed to be self-referential - to say
something about, be descriptive of, the speaker. Thus they see the
paradox to be that of explaining how a statement such as "I believe it
is raining" (understood to be "about the speaker") can conflict with
the other clause of MS, since it is clearly about some other (logically-
independent) subject matter - the weather. Because their response to
MS is guided by that question, the explanations they produce involve
the prior assumption that "I believe" is self-referential; and that un-
examined thought gives rise, in turn, to the other two staples in
Moorean discussions: the ideas that MS is not self-contradictory and
(therefore) that its absurdity arises only in speech.
Our discussion, following Wittgenstein's lead, argues that the three
ideas we have identified as guiding Moorean solutions of the paradox
are muddled, so our study, if successful, "dissolves" Moore's paradox
by showing it to be the outgrowth of misunderstandings.
7
And since
our criticisms aim to show that the philosophical requirements which
prompt Moorean invocations of the doctrine that saying implies believ-
ing result from antecedent confusion, direct criticism of that doctrine
will be ancilliary to our overall strategy, which is designed to show that
"I believe it is raining but it's not", for example, is absurd because it
consists of two contradictory assertions about the weather.
We begin by locating the ideas motivating standard discussions of
Moore's paradox in a different, and undeniable important, philosoph-
ical setting.
Wittgenstein hints at the historical significance of the primitive idea
in Moorean discussions of MS (that "I believe" is self-referential) in
the following remark:
MOORE'S PARADOX REVISITED 297
The significance of such possibilities of transformation, for example of turning all state-
ments into sentences beginning 'I think' or 'I believe' (and thus, as it were, into descrip-
tions of my inner life) will become clearer in another place. (Solipsism)8
The paradigmatic statement of solipsism is of course found at the
conclusion of Descartes's "First Meditation", where he vows "resolute
attachment" to the thought that I, "having no hands, no eyes, no flesh,
no blood, nor any senses, ... [nonetheless] falsely bdieve myself to
possess all these things".9 And this is the genesis of the philosophical
heritage which informs Moorean sensibilities. For the assumed meaning
of "I believe", around which standard discussion of Moore's paradox
turns, is also at the heart of the Cartesian enterprise in the Meditations.
Indeed, both the skepticism of the "First Meditation" and the cogito
of the "Second Meditation" depend upon a self-referential construal of
those words. For consider, since Descartes's skeptical hypothesis is
intended to immunize him from error, providing an antidote to the
threat posed by the malin genie, the "doubt" expressed in the resolution
to "consider myself falsely believing" cannot involve full-bodied disbe-
lief; to disbelieve a proposition ("I have hands, eyes ... ," for example)
is to believe its contradictory ("I believe that it is not the case I have
hands, eyes ... "), and that would leave Descartes equally exposed to
error.
So what Descartes's maneuver comes to, what his "hypothesis" re-
quires, is detaching belief from truth. This non-epistemic rendering of
"I believe" is spelled out in the Principles, where, explaining the cogito,
he characterizes judgments involving perception as though they func-
tioned to describe inner-experience rather than to assert "representa-
tional content":
[IJf I say I see, or I walk, I therefore am, and if by seeing and walking I mean the action
of my eyes or my legs ... my conclusion is not absolutely certain; because it may be that,
as often happens in sleep, I think I see or I walk, although I never open my eyes or
move from my place .... But if I mean only to talk of my ... consciously seeming to see
or to walk, it becomes quite true because my assertion now refers only to my mind,
which alone is concerned with my feeling or thinking that 1 see and I walk. 10
Now we can tie these thoughts of Descartes's directly to the issue at
stake in the resolution of Moore's paradox, by imagining an "Eighth Set
of Objections" to the Meditations, opening with the following question:
Dear Descartes, in your Meditations you take 'I am seated here by a fire' or 'This is a
hand' to be dubitable; whereas statements such as 'I think (believe/doubt) that this is a
hand' you hold to be indubitable. Now you apparently understand this epistemological
difference to reflect a semantic contrast, a difference in what is asserted in the two cases.
The former assertions go beyond reports or' one's experience, claiming to report an
independently existing reality; while the latter assertions, when carefully considered, are
seen to be nothing but descriptions of one's own states of mind, and so they are, according
to the cogito, indubitable. But then, why can't we assert such a thing as 'I believe I have
hands, but I don't'? Surely it would be perfectly absurd to .assert such a thing, although
I can't see that it should be, if you are correct in your underStanding of what its conjuncts
actually say.
So Descartes, too, might have been called upon to explain the odd-
ness of MS. And with this historical perspective in view, we now return
to standard discussions of Moore's paradox, to consider the related
claims they make that MS is impeccable in thought, and that it becomes
marred only when embodied in speech. For unless the entire thrust of
Wittgenstein's later philosophy is fundamentally misguided, it will turn
out that precisely because Mooreans are correct in claiming MS is
absurd in speech (on permanent holiday, if you will), they are wrong
in maintaining its intelligibility in thought. For the claim that an ex-
pression's meaning is a function of its having a place, and the place it
has in speech and communication is at the center of Wittgenstein's later
philosophy. Among the numerous examples we might cite as reminders
of this, the following, though long, is especially apt, since it relates that
theme directly to Moore's paradox:
How would it be, if a soldier produced military communiques which were justified on
grounds of observation; but he adds that he believes they are incorrect. ...
The communique is a language-game with these words. It would produce confusion if
we were to say: the words of the communique - the proposition communicated - have
a definite sense, and the giving of it, the 'assertion' supplies something additional. As if
the sentence, spoken by a gramophone, belonged to pure logic; as if here it had the pure
logical sense; as if here we had before us the object logicians get hold of and consider -
while the sentence as asserted, communicated, is what it is in business. As one may say:
the botanist considers a rose as a plant, not as an ornament for a dress or room or as a
delicate attention. The sentence, I want to say, has no sense outside the language-game.
This hangs together with its not being a kind of name. As though one might say "'I
believe ... " - that's how it is' pointing (as it were inwardly) at what gives the sentence
its meaning.ll
So the Fregean distinction between assumption and assertion, and the
idea of a "pure logical sense", which formal logic is to catch in abstrac-
tion from the empirical settings of assertion, is a myth, or so Witt-
genstein maintains. Meaning is inextricably scenic: "Words have their
meaning only in the flow of life". 12
Bringing this theme to bear on the useless sentences here at issue, we
begin by pressing for a precise characterization of the phenomenology
implied by the Moorean claim that "what it [MS] asserts" is a perfectly
possible, thinkable state of affairs. Taking "thinkability" literally invites
the following question: Can we have the thought "I e ~ i e v e it's raining
but it's not" without absurdity? Surely not. If we try to think that MS
is so, that strikes us as being every bit as ludicrous as trying to imagine
its public declaration; passing it through our head produces the same
sense of absurdity as passing it through our lips.
The Moorean response to this will no doubt be that the intelligibility
being claimed for MS is not based on the grounds that we can think it,
but rather, that it is possible to imagine (or "entertain" the thought
of) oneself believing it is raining, say, and that it's not raining. And
this, our objector insists, shows that what MS asserts is perfectly pos-
sible logically.
Among traditional philosophers who assume meaning and under-
standing either are or essentially involve mental states or processes,
such appeals are common fare. As a reminder of that, we recall two
important examples, the first from Hume, the second from Wittgen-
stein's Tractatus:
Suppose a person present with me, who advances propositions to which I do not assent,
that Caesar dy'd in his bed, that silver is more fusible than lead, or mercury heavier than
gold; 'tis evident, that notwithstanding my incredulity, I clearly understand his meaning,
and form all the same ideas, which he forms. My imagination is endow'd with the same
powers as his.u
'A state of affairs is thinkable': what this means is that we can picture it to ourselves.
14
Of course the Investigations argues that whether or not a sentence
has sense and what sense it has is not determined by the products of
the imagination:
It is no more essential to the understanding of a proposition that one should imagine
anything in connection with it, than that one should make a sketch from iLlS
Nor are such items sufficient to secure understanding; pictures, whether
images or sketches, public or private, are themselves signs which can
be variously interpreted, so they require rather than produce under-
standing:
Imagine a picture representing a boxer in a particular stance. Now this picture can be
used to tell someone how he should stand, should hold himself; or how he should not
hold himself; or how a particular man did stand in such-and-such a place, and so on. 16
But this general and rather abstract discussion will most likely not do
to correct the sense that we do have a successful appeal to the imagi-
nation in connection with MS, that here "a picture is conjured up
which ... [does] fix the sense unambiguously; I a picture ... which
seems to make the sense of the expression unmistakable". 17 Exposing
this apparition of sense to be the result of confusion requires more
detailed discussion.
Our beliefs are where we stand cognitively, so to speak; indeed,
though this analogy is weak it is not overworked by noting that the
verisimilitude of any representation of oneself falsely believing will be
limited in a manner akin to a drawing picturing where one is standing:
A painter does not draw the spot where he is standing. But in looking at his picture I
can deduce his position by relation to the things drawn. On the other hand, if he puts
himself into his picture I know for certain that the place where he shows himself is not
the place where he is.
18
Analogously, the person pictured to exhibit what MS asserts is, per-
force, a surrogate for "I" in "I believe", and so is subject to its logic.
Therefore, if we imagine the person depicted to be believing falsely,
we can "deduce" one is not then thinking from within that picture,
i.e., from within the "logical space" of "I believe"; for when belief is
thus identified ("I believe ... "), one cannot (logically) proceed to indi-
cate that the belief is false - one's current beliefs are one's representa-
tion(s) of reality. When we envisage ourselves from afar, as Hume was
wont to say, we observe both belief and contrary reality, and so the
relevant belief words are observational; but of course "I believe" is not
observational. We do not say "I believe it's raining" because we have
taken note of ourselves grabbing a raincoat or our saying "I believe it's
raining"; even more absurd is the thought that we base those words on
what we imagine ourselves to be saying and doing.
This criticism of the Moorean appeal to the imagination - where they
assume we find or may produce an exhibit of what MS allegedly asserts -
may be thought to mistake epistemological for seman tical differences;
that is, a Moorean may object that, of course, "I believe" is not ascribed
on the basis of observation, but that's because such ("intentional")
states are "known directly", rather than by inference from behavioral
symptoms, as in the case of "You (He) believe(s)". So although it is
correct that non-first-person belief words are required to articulate the
Moorean thought experiment, still, that in no way shows the thought
is not cogent. For making the required correction (substituting the
second- or third-person form of words) does not alter the "propositional
content" of one's thought.
That something along these lines would be Moore's response is clear
in his following comment:
'I went to the pictures last Tuesday, but I don't believe that I did' is absurd to say,
although what is asserted is something which is perfectly possible logically: it is perfectly
possible that you did go to the picture and yet you do not believe that you did.
19
Of course Moore is not alone in his unselfconscious assumption that
changes in pronoun make mere grammatical rather than logical alter-
ations in what is said or asserted (given the "indexicals" are understood
to "identify the same referent"). For example, the still widely used
canonical "A believes that p" is also employed on the assumption it
has the same truth-value, expresses the same proposition, no matter
what we substitute for "A". And, the assimilation underlying this prac-
tice takes us to an important conceptual source of the thought we
earlier identified as basic or primitive in both Descartes's thinking and
Moorean accounts of MS, namely, that "I believe" is self-referential.
No one will deny that prefixing "You (or He) believe(s)" to a simple
assertion such as "It is raining" does change its subject matter - transfor-
ming a proposition about the weather into one about the person's
belief. And, the assimilation of the first-person to those other forms of
the verb leads one to think a similar transformation is effected there,
and thus, to suppose those words are self-referential.
In opposition to the assimilation guiding such thinking, Wittgenstein
enjoins:
Don't look at it as a matter of course, but as a most remarkable thing, that the verbs
'believe', 'wish', 'will' display all the inflexions possessed by 'cut', 'chew', 'run'. 20
But before we consider how "believes" actually functions in our lan-
guage, it is important to appreciate the general nature, and thus the
full significance, of the presuppositions shaping the "matter of course"
(a priori) thinking Wittgenstein warns against. As is made clear by
the following questions which he poses in opening the Investigations's
discussion of Moore's paradox, those presuppositions are manifes-
tations of the very picture of language his later philosophy labors to
show is at the bottom of philosophical perplexity generally:
How did we ever come to use such an as 'I believe ... '? Did we at some time
become aware of a phenomenon (of belief)?
Did we observe ourselves and other people a!1d so discover belief?21
The conception of language, of reality, and qf the relation between
them that comes with affirmative answers to Wittgenstein's questions
is familiar to readers of the Investigations, which of course opens with
a quotation from Augustine illustrating that philosophically bedeviling
picture of language. The offending "primitive idea of the way language
functions,,22 is, in outline, this: words are names, language functions
to report the existence of what is named; and since words only name
what is there to be named, the nature of things is independent of and
prior to the ways in which we use language to talk about them. Learning
the meaning of "believe( s)", then, is learning to identify the phenome-
non of human experience it names - belief. "I believe" is the form of
words we have for identifying the presence of the phenomenon in
oneself; "You" and "He believes" are words and phrases which identify
that phenomenon in others. Moreover, when gripped by this model we
are inclined to think of sentences as names, too (perhaps along Fregean
lines, as a kind of complex proper name). Therefore, since we assume
"I believe it is raining" and "It is raining", for example, must name
different processes, when one is affirmed and the other denied (as in
MS), we think that that can't be a contradiction.
Discussion of the philosophical conception of language and com-
panion metaphysics, we have pointed out in the background of Moorean
discussions, is beyond the purview of this brief study. However, there
is a direct and untoward consequence of those ideas for first-person
belief talk which is perhaps sufficiently jarring to loosen their grip. This
consequence is remarked by Wittgenstein in the following comment:
A: 'I believe it's raining'. - B: 'I don't believe so'. Now they are not contradicting each
other; each one is simply saying something about himself.
23
Standard discussions of Moore's paradox are focused on the phrases "I
believe" or "I don't believe" nested in a single sentence, and thus
in one mouth, and so such discussions fail to reveal the implausible
consequence which their construal of those words has for an under-
standing of their role in contexts of interpersonal disagreement in belief.
But as Wittgenstein's above remark calls to attention, the Moorean
construal of those phrases effectively elides that role, rendering such
remarks bits of autobiography: A: "I feel anxious" - B: "I don't".
However, if I say "I believe the Prime Minister will soon fall from
office", and an interlocutor responds "I don't believe that he will",
there is disagreement.
Once we turn from the monolithic view of meaning and reference
which arises from a "name-thing" (" 'Fido'-Fido") conception of lan-
guage, we see quite clearly "I believe" and "He believes" do not have
the same function. Wittgenstein expresses their salient difference this
way: "'I believe that this is the case' is used like the assertion 'This is
the case'; and yet the hypothesis that I believe this is the case is not
used like the hypothesis that this is the case". 24 That is, "I believe
that the Prime Minister will fall from power" takes a position on, makes
a truth-claim about, the Prime Minister; whereas "He believes that the
Prime Minister will fall from power" takes no such stand (as witnessed
by the fact "He believes the Prime Minister will fall from power, but
he won't" is an impeccable assertion).
Wittgenstein brings out the conceptual difference in those belief
phrases from yet another angle, by asking the following:
I say of someone else 'He seems to believe ... .' and other people say it of me. Now,
why do I never say it of myself, not even when others rightly say it of me? - Do I myself
not see and hear myself, then? ... 25
The short answer is "seeing and hearing myself" is irrelevant, because
"I believe that p" announces a verdict about the truth-value of p; not
the condition of the person rendering that verdict;26 and others can say
of me that I seem to believe because their words - "He seems to believe
that p" (said of me) - do form a hypothesis about me, not a verdict
aboutp.
In sum then, "I believe" and "He believes" are different instruments
which perform functions as distinct as that made familiar by John Austin
between "I promise" and "He promises". "I believe", rather than
altering the subject matter of assertions, attenuates or otherwise modi-
fies statements. One form this takes is brought out in Urmson's classic
article "Parenthetical Verbs", in which he compares "I believe that p"
to "It is probable that p".27 Though again, one sometimes offers up a
proposition on the platter of "I believe" as a matter of conversational
form, thereby acknowledging that p is controversial (that others may
doubt or deny it), though nonetheless, certain of it oneself. In neither
case, though, is one adverting to a state or condition of the speaker;
as'Wittgenstein cautions: "Don't regard a hesitant assertion as an as-
sertion of hesitancy". 28
We must now anticipate and defend against various objections to our
account of "I believe", objections motivated by the thought that in
saying "I believe" surely something gets said abput the speaker. Such
criticisms can be developed in numerous ways. We concentrate here
on central examples, leaving it to the reader to adapt the principle
guiding our responses to those further cases. The error in all such
objections is the same: recognizing that information about the speaker
is made available by his speaking, the attempt is made to locate that
information in his words "I believe", whereas properly it should be
assigned elsewhere.
Starting with a simple case, of course a person can be saying some-
thing about himself when using "I believe". "Can you run a mile in six
minutes?" - "Yes, I believe I can". But that this remark is about the
speaker is not a function of the words "I believe", but of the second
occurrence of the pronoun. Or again, we answer questions like "Do
you believe Gorbachev will succeed in reforming the Soviet econ-
omy?" by saying "Yes, I believe he will", or "No, I don't believe he
will". Don't such answers show we are saying something about our-
selves? The erroneous assumption here is that the question is about the
respondent. But surely the question is about Gorbachev, though of
course addressed to the respondent. Again: "If you say 'I believe he's
happy', inferences about your behavior are possible. Doesn't that show
something gets said about the assertor?" No, inferences about my
behavior can also be made on the basis of my saying "He's happy",
but it doesn't follow that that assertion is about the speaker. Of course
if you assert "I believe ... ", that is your assertion and inferences about
you are possible. But that you have spoken and thereby opened up the
possibility of inferences being made about yourself is not the same
thing as asserting something about yourself.
Similar kinds of confusion arise from projecting features of the con-
text of utterance onto the use of "I believe". Someone applying for
Conscientious Objector status, for example, may well say to officials,
"I believe violence is evil". But it is not that his words are about
himself, rather he is the object of inquiry, an inquiry in which the
authorities are interested in the person. Their pursuing the question of
whether violence is evil, even had the applicant said straight off "Vio-
lence is evil", would be out of place in such proceedings. The following
observation of Wittgenstein's helps clarify what is at issue here, which
is how words that are about some other subject matter can, nonetheless,
be spoken to provide information about oneself:
The language-game of reporting can be given such a turn that a report is not meant to
inform the hearer about its subject matter but about the person making the report.
It is so when, for instance, a teacher examines a pupil. (You can measure to test the
ruler.)29
In asking, "Who discovered America?" the teacher is seeking to learn
something about the pupil. Yet the student can only perform success-
fully by using words which are not about himself, by saying, for exam-
ple, "Columbus, I believe".
Finally, further occasion for the misinterpretation of the logic of
assertions arises when traditional talk of truth conditions is applied to
propositions modified by "epistemic verbs". This kind of difficulty is
illustrated by David Lewis, when he states:
If someone says "I declare that the Earth is flat" (sincerely, not play-acting, etc.) 1 claim
that he has spoken truly: he does indeed so declare. 1 claim this not only for the sake of
my theory but as a point of common sense. Yet one might be tempted to say that he has
spoken falsely, because the sentence embedded in his performative - the content of his
declaration, the belief he avows - is false. Hence 1 do not propose to take ordinary
declaratives as paraphrased performatives ... because that would get their truth con-
ditions wrong.
30
Applying Lewis's reasoning to Moore's sentences of course inclines one
to hold that the assignment of conflicting truth values to its clauses
poses no problem of logical consistency. But this inclination and Lewis's
claim, that it is supported by common sense, are clearly wrong. The
truth is, rather, to the contrary, as George Lakoff observes:
Note that in statements it is the propositional content, not the entire sentence, that will
be true or false. For example, if 1 say to you 'I state that 1 am innocent', and you reply
That's false', you are denying that 1 am innocent, not that 1 made the statement. That
is, in sentences where there is an overt performative verb of saying or stating or asserting,
the propositional content, which is true or false, is not given by the sentence as a whole.
3
!
Of course "believe" is not a performative verb, but Lakoff's point
holds there, too. Lewis is confusing sincerity conditions with truth
conditions, conflating questions of truth with questions of truthfulness;
these are two different language-games or modes of assessment, both
of which can be engaged in with respect to "I declare", "I state", or
"I believe" statements, unlike "I have pain" statements, for example.
Before we conclude with a final criticism of standard treatments of
Moore's paradox, we should make note of one implication of this study
for studies of belief. Our account of the uselessness of MS, if correct,
inverts orthodox treatments of belief. This reorientation requires inde-
pendent development, of course, but we can sket,ch how the developed
story would parallel Wittgenstein's well-known "reversals" of the tra-
ditionally held primacy of appearance language over physical object
language,32 or of doubt over belief.
33
On the account of believing
developed here, belief is not, as traditionally assumed, the source of
assertions (a mental reservoir, a neural network, or whatever from
which assertions flow); rather, belief is itself "a new joint in", an
elaboration or modification of, the concept of assertion.
To this point in our discussion, our criticism of the standard account
of the absurdity of Moore's sentences has attacked the ideas which
motivate it; we have said nothing about the Moorean solution itself,
about their saying-implies-believing doctrine. We close by showing why
that doctrine, though perhaps true,34 does not explain the infelicity of
MS.
Those who maintain that saying implies believing gloss "implies" in
terms of what might be called the associated rights of "givings" and
"takings" in communication. Moore writes in explanation, "If we hear
a man say ... , we should all take it that ... ".35 Toulmin claims that a
forecaster's "It will rain" is about the weather and only "implies, or
gives people to understand" what his beliefs are.
36
Again, Nowell-
Smith develops a notion of contextual implication, which he explains
in terms of justified takings: "A statement p contextually implies a
statement q if anyone ... would be entitled to infer q from p in the
context in which they occur". 37 In sum, the idea that there is an impli-
cation relation between saying and believing is explained as amounting
to the claim that there is an inference ticket from asserting to a state
of the assertor.
As a principle of inference, however, the doctrine saying-implies-
believing is not something that a person can employ about himself. We
do not infer what we ourselves believe from what we say; we cannot
say "I said 'p', therefore (ceteris paribus), 'I believe p'''. Saying does
not imply believing, it implies "He believes": "'He said p', so (ceteris
paribus) he believes p". To make the principle here relevant to MS,
one must stretch it to include our taking the same interest in our own
words as another can. But when one works out that possibility, the
data which the principle is being invoked to explain (viz., the absurdity
of asserting MS) disappears:
If I listened to the words of my mouth, I might say someone else was speaking out of
my mouth.
'Judging from what I say, this is what I believe.' Now, it is possible to think out
circumstances in which these words would make sense.
And then it would also be possible for someone to say 'It is raining and I don't believe
it', or 'It seems to me that my ego believes this, but it isn't true.' One would have to fill
out the picture with behaviour indicating that two people were speaking through my
mouth.
38
NOTES
1 Norman Malcolm: 1958, Ludwig Wittgenstein: A Memoir, London, p. 177.
2 This practice probably derives from Wittgenstein; cf. his discussion of these sentences
in Ludwig Wittgenstein: 1968, Philosophical Investigations, 3d ed., trans. G. E. M.
Anscombe, New York, Part II, pp. 190-92.
3 G. E. Moore: 1942, "A Reply to My Critics", The Philosophy ofG. E. Moore, ed. P.
A. Schilpp, Evanston, pp. 542-43.
4 In addition to Wittgenstein's extensive discussion of the paradox, both in the Investi-
gations and elsewhere (cited below, note 11), this discussion draws on Kent Linville and
Mervill Ring: 1972, "Moore's Paradox: Assertions and Implication", Behaviorism I, pp.
87-102.
5 For an extensive bibliography of the major writings on Moore's paradox, see Jaakko
Hintikka: 1962, Knowledge and Belief, Ithaca, p. 64, footnote. To that, add Hintikka's
own work in that book, pp. 64-76; J. L. Austin: 1962, How to do Things with Words,
Cambridge, pp. 48-49; Norman Malcolm: 1963, Knowledge and Certainty, Englewood
Cliffs, pp. 16-17; P. H. Nowell-Smith: 1954, Ethics, London, pp. 80-81; B. C. van
Fraassen: 1984, "Belief and the Will", Journal of Philosophy 81, pp. 235-56; and Bernard
Williams: 1973, The Problems of the Self, Cambridge, p. 137. Considerations of length
lead us to only assert the basic sameness of these accounts, leaving verification to the
reader.
6 Wittgenstein, op. cit., p. 190. Cp.: "The sentence 'I want some wine to drink' has
roughly the same sense as 'Wine over here'!" Ludwig Wittgenstein: 1980, Remarkson
the Philosophy of Psychology, trans. G. E. M. Anscombe, eds. G. E. M. Anscombe and
G. H. von Wright, Chicago, vol. I, section 469.
7 "Something surprising, a paradox, is a paradox only in a particular, as it were, defective,
surrounding. One needs to complete this surrounding in such a way that what looked
like a paradox no longer seems one". Ludwig Wittgenstein: 1978, Remarks on the
Foundations of Mathematics, 3d ed., Oxford, section VII, no. 43.
8 Wittgenstein, Investigations, Part I, p. 24.
9 Rene Descartes: 1967, The Philosophical Works of Descartes, trans. Haldane and Ross,
Cambridge, vol. I, p. 148.
10 Ibid., p. 222, emphasis added.
11 Wittgenstein, Remarks on the Philosophy of Psychology, vol. I, sec. 487-88.
12 Ludwig Wittgenstein: 1980, Remarks on the Philosophy of Psychology, trans. C. G.
Luckhardt and A. E. Aue, eds. G. H. von Wright and Neikki Nyman, Chicago, vol. II,
sec. 687.
13 David Hume: 1967, A Treatise of Human Nature, ed. Selby-Biggs, Oxford, p. 95.
14 Ludwig Wittgenstein: 1963, Tractatus Logico-Philosophicus, trans. D. Pears and B.
McGuinness, London, 3.001.
15 Wittgenstein, Investigations, Part I, p. 396.
16 Ibid., note p. 11.
17 Ibid., pp. 426, 352.
18 Simone Weil: 1970, First and Last Notebooks, trans. Richard Rees, London, p. 146.
19 Moore, op. cit., p. 542.
20 Wittgenstein, Investigations, p. 190. Cp.: "Must the verb 'I believe' have a past tense
form? Well, if instead of 'I believe he's coming' we always said 'He could be coming' (or
the like), but nevertheless said 'I believed ... ' - in this way the verb 'I believe' would
have no present. It is characteristic of the way in which we are apt to regard language,
that we believe that there must after all in the last instance be uniformity, symmetry:
instead of holding on the contrary that it doesn't have to exist". Wittgenstein, Remarks
on the Philosophy of Psychology, vol. I, sec. 907.
21 Wittgenstein, Investigations, p. 190.
22 Ibid., Part I, p. 2.
23 Wittgenstein, Remarks on the Philosophy of Psychology, vol. II, sec. 419.
25 Ibid., p. 191.
26 "That he believes such and such, we gather from observation of his person, but he
does not make the statement 'I believe .. .' on the grounds of observation of himself.
And that is why 'I believe p' may be equivalent to the assertion of p. And the question
'Is it so?' to 'I'd like to know if it is so?,". Wittgenstein, Remarks on the Philosophy of
Psychology, vol. I. sec. 504.
27 J. O. Urmson: 1952, "Parenthetical Verbs", Mind LXI, reprinted in Essays in Concep-
tual Analysis, ed. A. Flew (London, 1956).
29 Ibid., pp. 190-91.
30 From David Lewis: 1972, "General Semantics", in Harman and Davidson, eds.
Semantics for Natural Language, Dordrecht, p. 210, as quoted in, George Lakoff: 1975,
"Pragmatics in Natural Logic", in E. L. Keenan, ed., Formal Semantics of Natural
Language, Cambridge, p. 256.
31 Ibid., p. 257.
32 Ludwig Wittgenstein: 1967, Zettel, eds. G. E. M. Anscombe and G. H. von Wright,
trans. G. E. M. Anscombe, Oxford, sec. 413-26.
33 Ludwig Wittgenstein: 1969, On Certainty, eds. G. E. M. Anscombe and G. H. von
Wright, trans. Denis Paul and G. E. M. Anscombe, Oxford, sec. 115, 160.
34 We use "perhaps true" advisedly; ct. Henry A. Alexander Jr.: 1967, "Comments on
Saying and Believing", in Epistemology: New Essays in the Theory of Knowledge, ed.
Avrum Stroll, New York, pp. 159-78.
MOORE'S PARADOX REVISITED
35 Moore op. cit., p. 543.
36 Toulmin, op. cit., pp. 52, 85.
37 Nowell-Smith, loco cit.
Linville:
Dept. of Philosophy
Oxford College/Emory University
Oxford, GA 30267
U.S.A.
Ring:
Dept. of Philosophy
California State University
Fullerton, CA 92634
U.S.A.
309
THEODORE R. SCHATZKI
ELEMENTS OF A WITTGENSTEINIAN PHILOSOPHY
OF THE HUMAN SCIENCES
ABSTRACT. In this paper, a Wittgensteinian account of the human sciences is con-
structed around the notions of the surface of human life and of surface phenomena as
expressions. I begin by explaining Wittgenstein's idea that the goal of interpretive social
science is to make actions and practices seem natural. I then explicate his notions of the
surface of life and of surface phenomena as expressions by reviewing his analysis of
mental state language. Finally, I critically examine three ideas: (a) that the goal of
interpretive inquiry is realized through a descriptive, context-constructing method that
enables investigators to grasp the instincts, mental states, and experiences ("Geist")
expressed in surface phenomena; (b) that uncovering rules plays a minor role in this
enterprise; and (c) that surface phenomena not only can be made natural but also have
causes and are subject to causal explanation.
Only to a limited extent did Wittgenstein explicitly advocate a definite
conception of the Geisteswissenschaften. It is possible, however, to
extract from his later works a consistent, albeit partial, account of this
field of knowledge. For the notions of the surface of human life and of
surface phenomena as expressions provide a guiding link with which
his ideas on a number of topics can be interrelated and their significance
for the human sciences established. This paper aims to carry out this
synthesis. I shall first discuss Wittgenstein's naturalistic conception of
the goal of interpretive social investigation, introducing notions such
as instinct-behavior and commonalities in human existence which will
subsequently receive greater concretization. Following this, I shall de-
scribe his notion of the surface of human life, and then analyze the
expressivity of surface phenomena by reviewing his analysis of mental
state language. The remaining sections will examine, with reference to
surface phenomena and their expressivity, Wittgenstein's views on: (a)
how to realize the goal of interpretive inquiry; (b) the roles played by
rules in such inquiry; and (c) the proper focus of the causal investigation
and explanation of human life.
1. THE GOAL OF INTERPRETIVE SOCIAL INVESTIGATION
Most of Wittgenstein's remarks that pertain explicitly to human science
concern so-called "interpretive" inquiry. Interpretive inquiry comprises
Synthese 87: 311-329, 1991.
312 THEODORE R. SCHATZKI
those investigations that aim to render at first strange, unfamiliar, or
puzzling actions and practices understandable (verstiindlich). Such
efforts are found predominantly in history and social anthropology, but
also in sociology, political science, and psychoanalysis.
Wittgenstein never indicated exactly what sorts of questions and
perplexities rendering practices understandable resolves. He was clear,
however, that, in most cases, making a practic,e intelligible does not
mean seeing the rationality in it. For, in his eyes, rationality is but an
occasional feature of life that presupposes a nonrational foundation.
Most practices, accordingly, do not make sense in the sense of being
rational. Rendering a practice intelligible means, instead, making it
seem natural. For Wittgenstein, a practice seems natural when an inves-
tigator locates it within the range of the humanly familiar, usually by
comprehending it as an expression or development of something com-
mon to all mankind and occasionally by relating it to what he or she
has learned about the vicissitudes of human life. As this formulation
implies, Wittgenstein was not the relativist that many of his interpreters
portray him as being. He was a staunch exponent of an extensive
commonality in human situations and ways of acting and being which
not only grounds but also constitutes interpersonal and intercultural
comprehensibility. As he writes:
One sees how misleading Frazer's explanations are - I believe - by noting that one could
very easily invent primitive practices oneself, and it would be pure luck if they were not
actually found somewhere. That is, the principle according to which these practices are
arranged is a much more general one than in Frazer's explanations and it is present in
our own minds, so that we ourselves could think up all the possibilities. (Wittgenstein
1979, pp. 65-66; hereafter RFGB)
In fact, if the "meaning" (RFGB, p. 69) of a practice is what is grasped
when that practice is rendered intelligible, then in the majority of cases
the meaning of a practice consists in concretizations or elaborations of
elements of human commonality.
Wittgenstein's criticisms of Frazer exemplify this conception of the
task of interpretive social science. Frazer is convicted of two central
mistakes. First, he misunderstands the nature of ritual in general, that
is, he misidentifies the general element of commonality of which it is
an elaboration. Frazer interprets ritual as instrumental action, thus
viewing it as an elaboration of the purposiveness found in human
existence, whereas Wittgenstein believed that ritual should be viewed
ELEMENTS OF A WITTGENSTEINIAN PHILOSOPHY 313
as "instinct-behavior" (RFGB, p. 72) expressive of certain aspects of
the instinct component of our common humanity (see also Wittgenstein
1966, p. 53ff.; hereafter LC). Second, Frazer fails to make specific
practices, e.g., the Beltane Fire Festivals, intelligible. When attempting
to come to grips with what is going on when children ceremoniously
"mak[e] a show" of burning a human being, Frazer retreats into a
hypothesis about the historical origin of this practice. According to
Wittgenstein, a historical hypothesis might fail to resolve Frazer's per-
plexity. Such a hypothesis brings historically antecedent phenomena
into association with the festivals. In doing so, however, there is no
guarantee that it will reveal the elements of commonality concretized
or elaborated in the festivals, and this is what it must accomplish if it
is to succeed in making these practices intelligible. Moreover, and most
importantly, if the hypothesis does accomplish this, it is not the fact
that the festivals originate in certain phenomena, but the resemblance
of these antecedent phenomena to the festivals that is responsible for
this success (see Section 3). In other words, if a historical hypothesis
makes a practice intelligible, it does so not because it is a historical
hypothesis, but because it happens to construct a context in which the
practice can be seen as natural. Historical origins qua origins cannot
make practices intelligible.
Of course, a social scientist might be interested in historical origins
simply because he or she is interested in knowing where a practice or
institution came from. More generally, an investigator might be inter-
ested in what brings certain phenomena about, i.e., their causes. An
investigator who pursues these interests does not aim to make practices
intelligible, but instead, to explain them causally. I do not think that
Wittgenstein saw anything wrong with causal explanation and the search
for origins and causes. He often characterized psychology and physiol-
ogy, for instance, as investigations of the causes of activity. What
Wittgenstein clearly did oppose is searching for causes when engaged
in the interpetive enterprise.
2. SURFACE PHENOMENA AND EXPRESSIONS
The guiding element in my synthesis of Wittgenstein's idea is his notion
of the surface of human life. Wittgenstein used the term "phenomenon"
(usually Erscheinung, sometimes Phanomen) as a general expression
for anything encounterable in experience. He wrote, for example, of
phenomena of electricity, of language, and of personal experience,
where, in each case, the phenomena of X are the experientially encoun-
terable appearances of X. The "surface of human life" is comprised,
accordingly, by the phenomena of human life, the aspects of human
life that are encounterable as such in experience. It embraces speech
acts, behavior, gestures, and facial expressions, on the one hand, and
thought experiences (e.g., Wittgenstein 1969, p. 8, hereafter BB), pains
(e.g., Wittgenstein 1967, sec. 288; hereafter PI), and other conscious
mental processes (seelische Vorgiinge), on the other.
By acknowledging the latter entities as surface phenomena, Witt-
genstein differentiated himself from behavioralism, which restricts the
surface of human life to behavior, the spatial-temporal appearances of
human life in the external world. More like Kant, the surface of life
comprised for Wittgenstein the totality of inner and outer, i.e., mental
and behavioral appearances of life. Each individual, moreover, has sole
experimental access to those elements of the mental surface of his own
life that are not simultaneously expressed in behavioral phenomena. In
outline, therefore, Wittgenstein followed tradition in acknowledging a
fundamental duality among the appearances of life. He differed from
tradition in arguing, first, that access to the mental surface of one's
own life is an ability engendered through and certified by social interac-
tion, and second, that mental phenomena such as feelings and pains
must not be assimilated to external objects. It should be mentioned
that Wittgenstein neither critically analyzed the idea that certain aspects
of life are encountered as such in experience nor defended his views
on which aspects these are.
1
Moreover, his notion of the surface of life
does not obviously apply to the experience of acting.
An important feature of the phenomena of human life is that they
express (ausdriicken) mental states, instincts, and experiences. Mental
states such as beliefs, expectations, moods, and emotions are not, like
conscious mental processes, surface appearances, that is, they are never
encountered as such in experience: "The psychological verbs to see, to
believe, to think, to wish, do not signify phenomena" (Wittgenstein
1970, sec. 471; hereafter Z); "'But "joy" surely designates an inward
thing'. No. "Joy" designates nothing at all. Neither any inward or any
outward thing" (Z, sec. 487). Mental states are, instead, expressed by
surface phenomena (cf. Z, sec. 506). This means, first, that surface
phenomena are the experiential, temporal or spatial-temporal appear-
ances of these states and, second, that what there is in the world to
mental states is nothing beyond these phenomena.Wittgenstein de-
scribed this relation by saying that appearances, in the circumstances
in which they occur, are what such states "consist in". For example,
what John's expecting Paul consists in, what there is in the world to
his being in this state, might be his pacing to and fro in his living room,
feelings of excitement and tension, and glances thrown at the clock on
the wall. In general, mental state concepts reflect the interest that a
group of people takes in certain patterns (Muster,' PI, p. 174), or
looks, that reappear in the surface appearances of their lives. When a
particular pattern appears in a given life, members of this group say
that the person is in such and such a state. This state, however, is
nothing above and beyond the appearances involved. Mental state
concepts are simply linguistic expressions for how it is with people
whose lives display certain looks (cf. PI, sec. 571).2
Surface phenomena are also able to express other surface phenom-
ena. In such cases, the entities expressed are something above and
beyoild the appearances that express them, and hence do not merely
"consist in" the latter. The two most prominent examples of this second
class of expressed entities are instincts and states of consciousness. For
Wittgenstein, instincts are not "inner" states but primitive, pre-linguis-
tic reactions (cf. Z, sec. 545; Wittgenstein 1975, sec. 545 (hereafter
~ C ; and Z, sec. 540-541). As such, they lie on the surface, and
certain more complex nexes of action and speech are extensions and
developments ofthem (Z, sec. 545; Wittgenstein 1976, p. 420; hereafter
CE). These more complex nexes can be called "expressions" of these
instincts, as suggested by Wittgenstein's description of ritual as ,"in-
stinct-actions". Similarly, mental processes such as sensations of sound
and taste, conscious thoughts (cf. Z, sec. 493), and pains (cf. Z, sec.
485) are surface phenomena, though not, like instincts, external be-
havior. They, too, however, can be expressed in behavior (cf. PI, sec,
288), especially in linguistic behavior such as "Oh, how beautiful!" or
"Boy, that smarts".
A final important feature of Wittgenstein's picture of the surface of
human life is that the phenomena of X form a particular sort of unity.
In his words, they have a particular face (Gesicht), or physiognomy (on
this topic, see Finch 1977, Chapter 11). When, for instance, Witt-
genstein wrote that belief (Z, sec. 514) or doubt (CD, p. 419) has a
particular physiognomy, what he meant was that the appearances of
belief or doubt, in the circumstances in which they occur, have a certain
316 THEODORE R. SCHA TZKI
look to them (cf. PI, p. 174 on grief, and Wittgenstein 1983, vol. VII,
sec. 17 (hereafter RFM) on experimentation). This thesis is an instance,
of course, of the wider doctrine that instances of a given concept share
a certain look, or face (cf. Z, sec. 376), by virtue of a host of similarities
and resemblances between them.
Now, an important issue in the philosophy of the human sciences
concerns the propriety of using one's own ordinary language mental
state and action concepts when describing lives in other cultures. This
issue seems to arise in an acute fashion on Wittgenstein's account of
such concepts. For according to him, which faces strike (beruhren, Z,
sec. 376) the members of a linguistic community in the phenomena of
life (thus, coordinately, what mental state/action concepts they employ)
depends on their needs and interests. If, consequently, basic or wide-
spread needs and interests vary from community to community, the
physiognomies that the members of different communities see in the
appearances of life will similarly vary. But if so, then the faces that an
investigator encounters in the lives of the people she studies may differ
from the faces they encounter there. And if this is the case, then, since
the use of mental states and action language reflects and itself propa-
gates the physiognomies people see in appearances, the mental state
and action concepts in terms of which her subjects carry out their lives
will not correlate with her own such concepts. Describing their lives
with her own such concepts, consequently, would only seem to miscon-
strue what they're about.
The problem with this argument, theoretically speaking, is that it
fails to distinguish between describing people's lives and capturing what
it is like to be one of them (see Runciman 1983). I shall not, however,
elaborate this claim, because from Wittgenstein's point of view there
is a tough of unreality to this theoretical discussion. As mentioned, he
believed that human beings share an extensive array of situations and
of ways of acting and being. From his remarks, mostly on Frazer, we
can gather that he viewed this commonality as consisting in at least:
(1) common basic needs and emotions; (2) similar physical environs
(phases of the moon, seasons, and animal and plant life); (3) common
facts of life (birth, death, and sex); (4) common gestures and primitive
reactions; (5) common goals and interests; (6) environmental entities
having the same significance (e.g., lions and thunder as threatening);
and (7) common practices (e.g., teaching, disciplining, and celebrating).
Where such commonalities are at work in the lives an investigator
studies, it follows: first, that faces familiar to the investigator from her
own linguistic community often will appear in these lives; second, that
this occurs because the same sort of thing is going on in these lives as
goes on in lives in the investigator's community when the same face
confronts her there; third, that the faces that strike the investigator will
also most likely strike them; all meaning, fourth, that there is no
problem with using the investigator's mental state anQ action language
to describe their lives - doing so neither falsifies the appearances nor
distorts what it is like to be one of them. Making these points leaves
open, of course, the exact extent to which they are true in any particular
situation. Since Wittgenstein, however, believed in extensive common-
alities, he thought that cross-cultural description and understanding are
far less problematic than they are sometimes portrayed as being (cf. PI,
sec. 206). For him, phenomenal physiognomics does not relativistically
confine mental state/action language to the linguistic community within
which it has organically grown.
However, it is not clear whether Wittgenstein is entitled to assert the
existence of most of these commonalities. If apprehension of physiog-
nomies is relativized to a community's interests and needs, it is not
obvious how an investigator can know that the familiar look she appre-
hends in a foreign community is not merely the projection of her own
community's ways of understanding and seeing. Unless it is possible to
transcend this relativization, it is impossible to be certain that most of
the above commonalities exist. As we shall see in Section 3, Wittgenste-
in's account of understanding does not allow for such transcendence.
3. SPIRIT AND OVERVIEW
Making a practice intelligible requires uncovering the instincts, mental
states, and experiences that are expressed in the surface phenomena
constituting it. Wittgenstein formulated this idea by writing that what
must be grasped in order to make a practice intelligible is the spirit
(Geist) expressed in it (the "inner nature" of the practice, RFGB, p.
74). This spirit is not the significance the practice has for its participants
(Cioffi 1981, p. 215), but the central instincts, mental states, and experi-
ences it expresses. The reason why grasping spirit makes practices
intelligible is that the instincts, etc., comprising it are, by and large,
concretizations and elaborations of types of instincts, etc., shared by
the investigator. Usually, therefore, grasping spirit is a means of "redis-
318 THEODORE R. SCHA TZKI
covering the I in the Thou" and thus of locating the practice within the
range of the humanly familiar.
How did Wittgenstein think a social scientist should go about grasping
the spirit expressed in a practice? He suggested that she observe the
same procedure recommended to philosophers who seek to understand
concepts and essence: constructing overviews (ilbersichtliche Darstel-
lungen). The social scientist should collect toget)1er different practices
related to the one to be understood in such a surveyable fashion that
she sees the connections among them, that is, how they hang together:
"This overview brings about the understanding which consists precisely
in the fact that we see how [these practices] hang together (die Zusam-
menhdnge sehen). Hence the importance of finding connecting links"
(RFGB, p. 69; translation modified).3 Underlying this recommendation
is the idea that practices are "associated", that, e.g., the Fire Festivals
of Europe, or the practices found in a given society, form "a multiplicity
of faces with co, 'lmon features which continually emerge here and
there" (RFGB, p. 74; translation corrected). Associated practices dis-
play a common physiognomy, and one grasps this physiognomy by
gaining an overview over the practices. Grasping this physiognomy,
however, is the same as grasping spirit. Thus, the inner nature of
a practice is unveiled either by placing the practice in the context
(Zusammenhang) of the wider culture in which it occurs, or by juxtapos-
ing it alongside similar practices in different cultures (cf. RFGB, pp.
75-76). Notice that grasping the spirit of a practice does not require
penetrating or digging below the surface phenomena of life. It requires,
instead, gaining a proper overview of the surface. Notice, too, that,
unbeknownst to Wittgenstein, it will sometimes be difficult for an inves-
tigator to construct an overview. In order to construct one, she must
know which practices are "related" to the one under investigation; in
many cases, she will know which practices these are only by under-
standing the spirit expressed in her object; and she can grasp this spirit
only by constructing the overview.
This exposition reveals that Winch's analysis of cross-cultural under-
standing, in his article "Understanding a Primitive Society", does not
describe Wittgenstein's views on this topic. Winch begins from the
assumption that each society has its own concept of (or rules for) the
intelligibility of human proceedings. He then argues that cross-cultural
understanding requires that the social anthropologist bring his concept
of intelligibility into relation with that of the society under investigation
in such a way as to create "a new unity for the concept of intelligibility
having a certain relation to [his] old one" (Winch 1977, p. 176). In
other words, understanding foreign practices necessitates achieving a
wider view of human life which incorporates the subjects' and the
investigator's ways of understanding human activity into a new perspec-
tive that goes beyond while retaining features of both. Hence, on
Winch's account, commonalities serve merely as a ground upon which
more expansive ways of understanding are erected. do
not suffice to ensure understanding but only make adequate, nonethno-
centric understanding possible.
For the most part, this account is clearly not Wittgenstein's. As we
have seen, in that majority of cases where the spirit of a foreign practice
consists of elements or types of elements shared by the investigator,
this commonality guarantees that grasping spirit is possible and that
accomplishing this makes the practice intelligible. In this vein, Witt-
genstein wrote:
All these different practices show that it is not a question of the derivation of one from
the other, but of a common spirit. And one could invent (devise) all these ceremonies
oneself. And precisely that spirit from which one invented them would be their common
spirit. (RFGB, p. 80)
Further, when the spirit contains unshared elements or types of ele-
ments, then commonality, in conjunction with the investigator's knowl-
edge of the vicissitudes of human life, will usually suffice to ensure that
spirit can be grasped and that understanding can thereby be achieved.
Only when the spirit is so incongruous with any sort of thinking the
investigator knows of could there be room, in Wittgenstein's scheme,
for Winch's Hegelianesque transformation of the investigator's concept
(or principles or rules) of intelligibility. It is significant, however, that,
although Wittgenstein considered (imaginary) cases of extremely
opaque practices, he never suggested that we react to these practices
by carrying out this sort of transformation. He suggested, instead, that
we either renew our efforts to make sense of them using our old
concept of intelligibility or abandon the effort and admit that we cannot
understand them (see, e.g., his discussion of the wood trade at RFM,
vol. 1, sec. 143ff.). To state Wittgenstein's views baldly: there is either
sufficient commonality and hence understanding or insufficient com-
monality and, as a result, no understanding.
Wittgenstein's claim about insufficient commonality is obvious. Sub-
suming practices under universal features of human life, however, can-
not provide much in the way of understanding. Such features are ex-
tremely general, and many different practices instantiate anyone of
them. Acquiring more particularized and robust understandings of indi-
vidual practices, consequently, requires some other, perhaps dialogical,
process. This conclusion is even more compelling in cases where the
spirit to be grasped contains unshared elements. ~ o r e o v e r as intimated
at the end of Section 2, Wittgenstein's account of mental state concepts,
when combined with the absence of a transformational model of under-
standing, makes it impossible both in general and in practice to distin-
guish between the apprehension of commonality and the projection of
one's community's ways of understanding and seeing.
4. RULES
My discussion of Wittgenstein's views on rules focuses on the scope he
accorded rules in human life. My aim is to show that interpretive social
science, as he conceived it, is concerned with rules only to a limited
extent. Wittgenstein's ideas on rules have been the subject of intense
discussion; I shall make no effort, however, to respond to the literature.
I shall also not examine the topic addressed by most of his remarks on
rules, viz., what it is to follow an explicitly formulated rule, for instance,
a rule of calculating.
It has become' dogma among some interpreters that Wittgenstein,
although he never explicitly indicated this, thought that all language
use, indeed, all behavior, is rule-governed. This dogma is a mistake.
The idea that all behavior is rule-governed exemplifies a rationalist,
or intellectualist, view of life. Wittgenstein, however, believed that
underlying and running through the layers and realms of rational, intel-
lectually-governed human conduct are a bedrock and veins of non-
rational proceedings: ordinary life is incompletely governed by logoi.
This thesis is reflected in his constant reminders that action, not
thought, underlies language (e.g., GC, sec. 475; Z, sec. 391; and RFM,
vol. VI, sec. 39), and that action is what underlies thought, observation
and reason (e.g., Z, sec. 8; PI, sec. 456; GC, sec. 110, 204). Since
action underlies reason, it follows that reason is not the measure of
practices (CE, p. 417). This viewpoint also animates, of course,
Wittgenstein's account of what it is to follow a rule: following a rule
does not require an intellectual act (e.g., an intuition or interpretation)
concerning how to act, nor does it involve following a second rule that
governs the application of the first. Rather, it simply consists in going
on (reacting) in particular ways within certain "surroundings", i.e., in
the context of what others do and the sort of instruction and correction
that pertain to the behavior in question (see RFM, vol. VI, sec. 42; for
an interpretation of rule-following similar to mine which also develops
ideas resembling those discussed at the end of this section, see Hintikka,
1989). In any case, to view all behavior and/or language use as rule-
governed is to see a core of ratio lying below the surface of human life,
and thus, to contravene Wittgenstein's admonishments in this context
not to dig below the surface (RFM, vol. VI, sec. 31). Let us consider
more precisely what his remarks imply about the scope of rules.
They suggest that Wittgenstein would have restricted the expression
"following a rule" to situations in which either an explicitly formulated
rule is "involved in" (BB, p. 13) behavior - e.g., consulted in the course
of acting - or action results from training (instruction and correction) in
which explicit rules are cited (cf. RFM, vol. VI, sec. 45). After all, he
wrote that "following a rule is a particular language-game" (RFM, vol.
VII, sec. 52, my emphasis) and that following a rule has a particular
physiognomy (RFM, vol. VII, sec. 60). Wittgenstein recognized, how-
ever, that behavior and language use can be and are sometimes ex-
ecuted and learned without consultation with or citation of rules (e.g.,
DC, sec. 95, 140). Much of the time, accordingly, people are incorrectly
described as "following rules". Some of what they do on those oc-
casions, however, might still be, in Wittgenstein's words, "in accord-
ance with a rule" (BB, pp. 87-98; and PI, sec. 54). For the notion of
a rule is connected to the notion of a regularity (e.g., PI, sec. 208) and
in at least two places, Wittgenstein acknowledged that it is correct to
describe a practice as proceeding according to rules when "an observer
can read these rules off from the practices of the game - like a natural
law governing the play" (PI, sec. 54; and cf. BB, p. 98). In other
words, behavior proceeds according to rules when it is possible to codify
behavioral regularities in the form of rules.
4
Of course, a regularity-
codifying rule, as Wittgenstein painstakingly worked out, is not the
same as a description of a regularity. It is a proposition, merely with
the form of a description, that is treated as a specification of how
actions must turn out if they are to count as correctly performed.
The key point about regularity-codifying rules is that they often,
if not usually, are not formulable. Wittgenstein explicitly drew this
conclusion in the case of the use of ordinary language concepts (e. g. ,
PI, sec. 84; and Z, sec. 350) on the grounds that any rule, formulated
either by an actor or by an observer ,on the basis of a finite number of
uses of a concept, will be unable pre facto to cover all possible accept-
able uses.
5
Moreover, since any finite chunk of behavior displays an
indefinite number of regularities, it accords with an indefinite number
of "law-like" rules (see Rosen 1983). This meflns, however, that the
behavior is no more in accordance with anyone or other of these rules
(BB, p. 13). These arguments apply to all behavior regardless of
whether it results from or involves explicit rules or is interwoven with
correction and instruction that do not cite rules. Action cannot be
encompassed by rules, and not only, as Winch claims, because what it
is to follow a rule cannot be understood as grasping a rule for the
application of the rule (Winch 1958, p. 55).
Since observation does not usually yield rules governing language
use, and since actors are unable to formulate such rules (BB, p. 25),
Wittgenstein asked, "What meaning is the expression 'the rule by which
he proceeds' supposed to have left to it here?" (PI, sec. 82). In other
words, hasn't it become pointless to describe the actions concerned as
proceeding according to rules? Of course, one might feel compelled,
especially in the presence of correction and instruction, to insist that
there must be rules at work keeping us consistent with and intelligible
to one another. This temptation merely expresses, however, our dis-
satisfaction with the surface phenomena. Instead of being content sim-
ply to note the surface and to say "This is what we do", we insist on
penetrating the surface to uncover the logos that secures our perfor-
mances. For Wittgenstein, however, there is no such stratum of reason.
There is only the surface (and a physical substratum). "Our mistake is
to look for an explanation where we ought to look at what happens as
a 'primordial phenomenon' (Urphanomen). That is, where we ought
to have said: this language-game is played" (PI, sec. 654; translation
amended).
For Wittgenstein, therefore, neither linguistic nor nonlinguistic be-
havior is generally governed by rules. (It is possible to construct an
argument pertaining to nonlinguistic behavior that parallels the one just
presented concerning linguistic behavior.) People follow rules when
rules are an explicit feature of the surface; and people can be said to
act in accord with rules in the exceedingly rare cases where behavioral
regularities are subject to comprehensive codification in a set of unique
rules. Notice that it follows, I do not think paradoxically, that in most
of the cases where people follow rules, they do not act in accordance
with rules.
Incidentally, I think that the strongest argument for the ubiquity
thesis is grounded in the connections Wittgenstein drew between lan-
guage and regularities (e.g., PI, sec. 207), on the one hand, and be-
tween regularities and rules (e.g., PI, sec. 208), on tqe other. If any-
thing is to be recognizable by us as language it must display regularities;
and, as we've seen, one type of rule is regularity-codifying formulations.
It seems to follow, accordingly, that rules exist wherever there is lan-
guage (cf. DC, sec. 61-62; and also Z, sec. 612). This argument fails,
however, to confront the above problem: How does one formulate
unique sets of rules that cover all possible (acceptable) uses of a given
concept? It might also be mentioned that John McDowell has pointed
to the following statement (at RFM, vol. VI, sec. 28) as evidence for
the ubi4uity claim: "Following according to the rule is FUNDAMEN-
TAL to our language-game. It characterizes what we call description"
(McDowell 1984, p. 341; see also Kerr 1986, p. 110). However, the
opening paragraph of this section addresses the issue of how someone
knows that the color he sees is red; and, in the Blue Book, Wittgenstein
characterized an ostensive definition of the word "yellow" as a "rule
of the usage of the word" (BB, p. 12, cf. p. 90). Now, an ostensive
definition is an example of an explicit rule that is cited in the teaching
of language. Using color words, accordingly, is correctly described as
"following a rule". Hence, I do not think that Wittgenstein was making
a general point about all description whatsoever, merely one about the
description of colors in particular.
The consequences of this discussion of rules for interpretive social
science are straightforward. Explicitly formulated rules are surface phe-
nomena. As such, they help comprise the set of surface phenomena,
the spirit of which an investigator aims to grasp. I do not think that it
is correct, however, to construe either explicitly formulated rules or
regularity-codifying rules as part of the spirit of any practice. Not
only did Wittgenstein nowhere write or imply that surface phenomena
"express" rules of any sort, but regularity-codifying rules do not pertain
to surface phenomena until they are formulated. Thus, although rules
are part of interpretive social science's object of study, they are not
part of what the social scientist aims to grasp (spirit) in order to under-
stand this object. Making practices intelligible, an enterprise which, as
we've seen, is keyed to the surface, thus concerns itself with rules only
to the extent that rules are surface phenomena.
This conclusion runs counter to the most widely acknowledged
"Wittgensteinian" philosophy of social science, that of Winch's (see
also, e.g., Apel 1967). In Winch's account, understanding a practice
requires a grasp of the usually nonexplicit rules governing it. For, a
practice is defined as a regularity in action; regularities are "constant
recurrence[s] of the same"; what counts as the same is determined by
rules; and the identity-establishing rules definitive of any given practice
are the largely nonexplicit ones followed by the people carrying it out.
In Winch's view, therefore, understanding a given surface phenomena
(a practice) requires a grasp of something beneath the surface which
governs it (nonexplicit rules) (Winch 1958, pp. 83-87). Winch is wrong,
however, to maintain that, in Wittgenstein's account, nonexplicit rules
govern practices, more specifically, that what practices are is defined
by nonexplicit rules that determine what counts as "the same". Essen-
tially the only rules that pertain to practices are explicit ones. And even
where explicit rules are at work, it is action (what we do) that ultimately
determines what "the same" is (e.g., RFM, vol. VII, sec. 40). To state
all this in terms of meaning: since interpretive social science aims to
uncover meaning, it does not strive to uncover rules, for rules do not
determine the meanings of practices.
5. CONCLUSION: DESCRIPTION AND EXPLANATION
As we have seen, the comparative, context-constructing method Witt-
genstein advocated for interpretive social science is descriptive in nat-
ure. Executing it consists in arranging descriptions of phenomena that
are related to the practice under investigation in such a way that there
results a grasp of the spirit of the practice and the practice is thereby
rendered natural. This procedure does not require explanations or
hypotheses about origin, although as indicated, explanations and hy-
potheses can help construct the context (see Cioffi 1981, pp. 219-
20). Furthermore, since Wittgenstein tied theories to hypotheses and
explanations (see, e.g., PI, sec. 109; and BB, p. 18), it follows that
theories, too, are not required in interpretive social science.
This does not mean, however, that Wittgenstein opposed the use of
explanations, hypotheses, and theories in the human sciences. Both
overviews and theoretical explanations have secure places in the
Geisteswissenschaften if different components of this enterprise strive
to realize different cognitive achievements. And sure enough, Witt-
genstein acknowledged that physiology and psychology have legitimate
explanatory work to perform. What he opposed was the idea "that these
explanatory sciences provide the sort of comprehension sought after in
interpretive social science (or in aesthetics and philosophy). I want to
conclude by suggesting how the above account of the sprface of human
life identifies the focus of these explanatory efforts.
The surface phenomena of human life that Wittgenstein construed
as expressions of mental states and processes have causes, where by
"cause" Wittgenstein meant a connection, possibly not open to view
(BB, p. 6), that is established either by experiment and statistics or by
tracing a mechanism (LC, p. 13f.; and CE, passim.). He spoke, in this
vein, of the causes of behavior (BB, pp. 11-12, 15), the causes of word
association (PI, p. 216), the causes of pain (Wittgenstein 1968, sec.
235; hereafter NFL), and the causes of the experience of seeing an
aspect (PI, p. 193, ct. pp. 201, 205). All these effects are surface
phenomena. He also spoke of the causes of certainty (RFM, vol. VI,
sec. 47; DC, sec. 429; and PI, sec. 325), of being able to follow a rule
(PI, sec. 217), of belief (Z, sec. 437), of fear (PI, sec. 476), and of
sadness (Z, sec. 509). In these cases, I take him to have been speaking
of the causes of those surface phenomena that are expressions of these
mental states. Since attributions of causality are, for Wittgenstein, hy-
pothetical explanations (BB, pp. 15,88; and PI, sec. 475), by using the
notion of cause as he did, he indicated that the entire apparatus of
hypothesis, explanation, theory, experimentation, and statistics can be
brought to bear on surface phenomena. We thus find him referring in a
matter-of-fact and uncritical manner to psychological and physiological
experiments, explanations, and causation (e.g., PI, sec. 412, 493, pp.
203,212; BB, pp. 6,88; LC, p. 20; and NFL, p. 235).
As is well known, the major question Wittgenstein raised about these
sciences is whether they can become systematic. In any science, the
"craving for generality" expresses itself in the desire to reduce the
explanation of events to the smallest number of primitive laws (BB, p.
18). Against this, Wittgenstein stressed the variety of causes, e.g., of
dreams, play, and punishment (LC, pp. 47, 49, 50). There also exists
the familiar issue of whether there is any regularity to the causes of the
variety of phenomena that count as expressions of a given mental state;
that is, whether there exist type correlations between, say, physiological
processes and instances of people being in that state. Wittgenstein was
skeptical about the likelihood of such correlations. Our use of mental
state and action concepts is keyed to physiognomic features of the
surface, and these concepts are learned through observation, partici-
pation, training, and correction, all of which are likewise keyed to the
surface. It is true that some of the causality at work in human activity
also lies on the surface, e.g., training bringing <;lbout "the phenomena
of understanding, obeying etc." (BB, p. 12, see also p. 111). But to
the extent that this causality lies below the surface, as for instance, in
the case of physiological causality, we should not expect physiological
regularities to correlate with physiognomies and the patterns of phe-
nomena that are expressions of a given mental state.
Earlier I criticized Wittgenstein's accounts of what is involved in
and how social scientists should go about understanding actions and
practices. I also indicated possible problems with his notion of the
surface of life. I believe he is correct, however, that at least most of the
rules governing human life are explicit and that behavior and conscious
mental phenomena have causes.
APPENDIX: WITTGENSTEIN AND GERMAN PHILOSOPHY
The sketch of a philosophy of the human sciences presented in this
paper substantiates and further develops the idea that Wittgenstein was
a twentieth century descendent of what Karl Mannheim called (early
nineteenth century) "conservative thought" (Mannheim 1971, Chapter
5; see Nyiri 1976 and 1982; and Bloor 1983, Chapter 8). Of the features
attributed by Mannheim to this style of thought, those most prominently
displayed by Wittgenstein's views on human-social reality and human
science are emphases: on physiognomic as opposed to constructive
understanding; on concreteness as opposed to abstractness; on parti-
cularity as opposed to systems; on the nonrationality as opposed to the
rationality of human life; and on understanding norms (rules) via the
actual instead of vice versa.
Moreover, since Mannheim identified Lebensphilosophie as the cen-
tral late nineteenth and early twentieth century descendent of conserv-
ative thought, it is of particular interest to note the apparent influence
exerted upon Wittgenstein by this school of thought. Lebensphilosophie
was influential in the German-speaking world, both when Wittgenstein
grew up and when he lived in Austria in the teens and twenties. One
of its prominent doctrines was the neo-Hegelian idea that the products
and sensible appearances of life manifest, embody, or make present
mental/spiritual/intellectual contents and states. Wittgenstein's idea
that surface phenomena "express" mental states is a version of this
doctrine (though he did not follow most life-philosophers in according
these contents and states an "ideal" existence in contrast to the "real"
existence of their spatial and temporal bearers). Wittgenstein even
occasionally used one of this school's terms of art:
(e.g., Z, sec. 110,534).
It seems likely, furthermore, that the life-philosopher Oswald
Spengler, whose direct influence Wittgenstein acknowledged (Witt-
genstein 1980, p. 19), was the concrete link between Wittgenstein and
Lebensphilosophie. Wittgenstein appears to have derived (at least) two
important notions from Spengler: the idea that the appearances of X
have a particular physiognomy, or look, and the idea that compre-
hending a practice is accomplished by grasping, through the construc-
tion of an overview, the spirit expressed in it. The first idea strongly
resembles Spengler's notion of the "habitus" of X: "the type of external
appearance pertaining to X alone", the "character and style of X's
stepping-into" the spatial realm of what has become (Spengler 1918,
p. 158). The second idea resembles Spengler's notion of physiognomics
(cf. RFGB, p. 69), a procedure in which the meaning of an individual
entity, the element of the "inner soul" it expresses, is grasped by
incorporating the entity into a living, inwardly felt unity constructed by
comparative methods (RFGB, pp. 146-54).6
NOTES
1 Wittgenstein did not make any specific comments about the notion of a phenomenon.
He seems to have believed, however, that the appearances encountered in experience
are in some way independent of the linguistically-expressed conceptual resources of, and
hence the descriptions given by, the people who encounter them. Appearances, that is,
are entities with a being of their own. Moreover, people can apprehend what these
appearances in themselves are. One of the many remarks suggesting that this was Wittgen-
stein's view is the famous passage (at PI, sec. 90) where he equated the possibilities of
phenomena with the kind of statements we make about them. At first sight, this passage
appears to contravene my reading. Notice, however, that Wittgenstein wrote, "the kind
of statements we make about phenomena" (my emphasis). This formulation implies that
appearances are some way and that we make statements about them on the basis of this.
The kind of statements we make about phenomena do not define the possibilities of how
phenomena in themselves are, but instead, the possibilities of what for us these phenom-
ena are instances of, given how they in themselves are. How appearances in themselves
are also determines what aspects of life are encountered as such in experience.
2 It is worth adding that Wittgenstein maintained that it is possible to see in appearances
the mental states they express. It is usually not the case, consequently, that a person
learns that someone else is in a given state by first observing an appearance and then
making an inference to the presence of the state (though, of course, this process also at
times occurs) (See Z, sec. 220, 225).
3 Incidentally, similar ideas appear in Wittgenstein's remarks on psychoanalysis (see Le,
pp. 45-51).
4 I am unclear whether Wittgenstein thought that only those regularities that occur in
certain surroundings, e.g., amid correctional and instructional behavior, can be correctly
codified in rules.
5 This argument implies that at least most of the rules cited when teaching language do
not adequately represent the linguistic practice they are used to help inculcate. I write
'most' because passages, such as PI, sec. 558 and Z, sec. 350, suggest that Wittgenstein
thought that rules can sometimes be formulated.
6 An earlier version of this paper was presented in May 1988 at the Inter-University
Center of Postgraduate Studies in Dubrovnik as part of a course on Wittgenstein and the
Philosophy of Culture. I wish to thank the course participants for their comments and
the course organizers, Tore Nordenstam and Kjell Johannessen, for the invitation.
REFERENCES
Apel, Karl-Otto: 1967, Analytic Philosophy of Action and the Geisteswissenschaften, D.
Reidel, Dordrect.
Bloor, David: 1983, Wittgenstein: A Social Theory of Knowledge, Columbia University
Press, New York.
Cioffi, Frank: 1981, 'Wittgenstein and the Fire-Festivals', in Irving Block (ed.), Perspec-
tives on the Philosophy of Wittgenstein, MIT Press, Cambridge.
Finch, Henry Le Roy: 1977, Wittgenstein - The Later Philosophy, Humanities Press,
Atlantic Highlands, NJ.
Hintikka, Jaakko: 1989, 'Rules, Games, and Experiences: Wittgenstein's Discussion of
Rule-Following in the Light of his Development', Revue Internationale de Philosophie
43(2), 279-97.
Kerr, Fergus: 1986, Theology after Wittgenstein, Blackwell, Oxford.
Mannheim, Karl: 1971, 'Conservative Thought', in Kurt Wolf (ed.), From Karl
Mannheim, Oxford University Press, New York.
McDowell, John: 1984, 'Wittgenstein on Following a Rule', Synthese 58, 325-63.
Nyiri, J.: 1976, 'Wittgenstein's New Traditionalism', in Essays on Wittgenstein in Honour
of G. H. von Wright, Acta Philosophica Fennica 28 (1-3), 503-12.
Nyiri, J.: 1982, 'Wittgenstein's Later Work in Relation to Conservatism', in B. McGuin-
ness (ed.), Wittgenstein and his Times, Blackwell, Oxford, pp. 44-68.
Rosen, Michael: 1983, 'Critical Theory: Between Ideology and Philosophy', in Sollace
Mitchell and Michael Rosen (eds.), The Need for Interpretation, Athlone Press, Lon-
don.
Runciman, W. C.: 1983, A Treatise on Social Theory, vol. 1, Cambridge University Press,
Cambridge.
Spengler, Oswald: 1918, Der Untergang des Abendlandes, Wilhelm Braumuller, Wien.
Winch, Peter: 1958, The Idea of a Social Science and its Relation to Philosophy, Routledge
and Kegan Paul, London.
Winch, Peter: 1977, 'Understanding a Primitive Society', reprinted in Fred R. Dallmayr
and Thomas A. McCarthy (eds.), Understanding and Social Inquiry, University of
Notre Dame Press, Notre Dame.
Wittgenstein, Ludwig: 1966, in Cyril Barrett (ed.), Lectures and Conl'ersations on Aesthet-
ics, Psychology, and Religious Belief, Blackwell, Oxford.
Wittgenstein, Ludwig: 1967, Philosophical Investigations, G. E. M. Anscombe (trans.),
MacMillan, New York.
Wittgenstein, Ludwig: 1968, 'Wittgenstein's Notes for Lectures on 'Private Experience'
and 'Sense Data", Philosophical Review LXXVII, 271-320.
Wittgenstein, Ludwig: 1969, The Blue and the Brown Books, Blackwell, Oxford.
Wittgenstein, Ludwig: 1970, in G. E. M. Anscombe and G. H. von Wright (eds.), Zettel,
G. E. M. Anscombe (trans.), University of California Press, Berkeley.
Wittgenstein, Ludwig: 1975, in G. E. M. Anscombe and G. H. von Wright (eds.), On
Certainty, Denis Paul and G. E. M. Anscombe (trans.), Blackwell, Oxford.
Wittgenstein, Ludwig: 1976, 'Cause and Effect: Intuitive Awareness', Peter Winch
(trans.), Philosophia 6 (3-4), 409-25.
Wittgenstein, Ludwig: 1979, 'Remarks on Frazer's Golden Bough', in C. G. Luckhardt
(ed.), Wittgenste'in: Sources and Perspectives, J. Beversluis (trans.), The Harvester
Press, Brighton. Originally published in Synthese 17 (1967), 233-53.
Wittgenstein, Ludwig: 1980, in G. H. von Wright and Heikki Nyman (eds.), Culture and
Value, Peter Winch (trans.), University of Chicago Press, Chicago.
Wittgenstein, Ludwig: 1983, in G. H. von Wright, R. Rhees, and G. E. M. Anscombe
(eds.), Remarks on the Foundations of Mathematics, G. E. M. Anscombe (trans.),
MIT Press, Cambridge.
Department of Philosophy and Committee on Social Theory
University of Kentucky
Lexington, KY 40506-0027
U.S.A.

Ludwig Wittgenstein (Auth.), Jaakko Hintikka (Eds.)-Wittgenstein in Florida_ Proceedings of the Colloquium on the Philosophy of Ludwig Wittgenstein, Florida State University, 7–8 August 1989-Springer

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ludwig Wittgenstein (Auth.), Jaakko Hintikka (Eds.)-Wittgenstein in Florida_ Proceedings of the Colloquium on the Philosophy of Ludwig Wittgenstein, Florida State University, 7–8 August 1989-Springer

Uploaded by

Copyright:

Available Formats

WITTGENSTEIN

You might also like