You are on page 1of 15

Twelve-Tone Invariants as Compositional Determinants

Author(s): Milton Babbitt


Source: The Musical Quarterly, Vol. 46, No. 2, Special Issue: Problems of Modern Music. The
Princeton Seminar in Advanced Musical Studies (Apr., 1960), pp. 246-259
Published by: Oxford University Press
Stable URL: http://www.jstor.org/stable/740374 .
Accessed: 28/06/2013 15:39
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Oxford University Press is collaborating with JSTOR to digitize, preserve and extend access to The Musical
Quarterly.

http://www.jstor.org

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

TWELVE-TONE

INVARIANTS AS

COMPOSITIONAL

DETERMINANTS

By MILTON

AT

BABBITT

the present moment, when many of the jagged edges of abrup-

tion have been smoothedby time and practice, there are those
presumablyin the spiritof mediationand moderation- would
minimize,not so much Schoenberg's achievementas a composer, as
the degree to which the twelve-tonesystemis genuinely"revolutionary"
in its nature and implications,the degree to which it imposes new
demands of perceptionand conceptionupon the composerand listener,
and - therefore- the degreeto which it admitsof furtherand extensive explorationand discovery.
Such an attitude does a disservicenot only to Schoenberg,but to
the cause of understanding,particularlysince it so often involves the
invocationof the alleged historical-analogical
originsof the operations
of the system,along with conjecturesas to Schoenberg'smode of and
motivationfor arrivingat the system.However intriguingsuch conjecturesmay be, they are as irrelevantas they are futile; however pedasuggestivea quasi-geneticapproach
gogicallyconvenientand intuitively
may be, eventuallyit succeeds only in obscuringboth the character
of the systemand the profound differencesbetween the twelve-tone
systemand those musical systemsin which the "historicalforerunners"
of the twelve-toneoperations appear. The crucial point here is that
these "forerunners"are not independentand fundamentalstructural
determinants,but means of immediate procedure, neithernecessarily
presentnor, if present,of more than local significanceand influence.
who -

Thereforeit is appropriateto precede even so informala discussion


as the one to follow with the reminderthat the twelve-tonesystem,
formulable,
like any formalsystemwhose abstractmodel is satisfactorily
the
stipulated
can be characterizedcompletelyby statingits elements,
246

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

Twelve-Tone Invariants as Compositional Determinants

247

relationor relationsamong these elements,and the definedoperations


upon the so-relatedelements.Such a characterization,though explicitly
presentedin verbal format the earlieststage of the twelve-tonedevelopment,is likewiseeasily and explicitlyinferableas the maximum procedural intersection
among the "classical" twelve-toneworksof Schoenberg,
and
Webern,
Berg.
If the elementsof the twelve-tonepitch systemare, indeed, "traditional" ones,both insofaras theyare pitchclasseswithclass membership
definedby octave equivalence,and as thereare twelvesuch pitch classes
-corresponding to the chromaticallyequal-temperedquantization of
the frequencycontinuum--evenhere essentialdeviationsmust be noted.
In the twelve-tonesystemthereis a one-to-onecorrelationbetweenpitch
notationand presentedpitch,as opposed to the many-to-onecorrelation
of triadic-tonalmusic; therecan be no such distinctions
as thosebetween
or
and
functional
between
"dissonance,"
enharmonicallyidenexplicit
tical "consonance" and "dissonance." The independentassumptionof
octave equivalence has been a frequentpoint of attack upon the system,
particularlyby those who assert that the correspondingassumptionin
the tonal systemservesto defineclasses of equivalent function;it need
be answered only that, similarly,this assumption in the twelve-tone
systemservesto defineclasses of equivalent order position.
It is in the definitionof relations among the elements that the
fromsystemsof the past, for relationsare
systemdivergessignificantly
definedentirelyby the impositionof a total linear orderingupon the
pitch classes, thus defininga twelve-tone"set" (designated: S). The
orderingemployed,in any given work, is inferablefrom-at mostall of the compositionalpresentationsof the set (and its transformations), and not necessarilyfrom any one compositionalpresentation.
By introducingthis principle as the basis of relationship,Schoenberg
not only effecteda fusionof the general systematicconstraintwith the
contextuallydefinedproperty-for,although the principleof formation
is definedfor all sets,the specificpitch class relationsdefinedby a set
- but established
are uniquelyassociatedwithit and its transformations
the means of a permutationalmusical system,as opposed to the combinational systemsof the past. Given a collectionof available elements,
the choice of a sub-collectionof these as a referentialnorm providesa
norm that is distinguishableby contentalone; such a system,and the
traditionaltonal systemis such, is thereforecombinational.But if the
referentialnorm is the totalityof elements,thereis but one such norm
in termsof content,and deviationsfromthis norm cannot exist within

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

The Musical Quarterly

248

the system.But if an orderingis imposed upon this totality,and taken


in the case of twelvepitch class
as a norm,thisnormis so distinguished,
elements,from the 12!' - 1 other possible orderings,that is, other
possible permutations.
Any considerationof the operationsof the systemmustproceed from
an awarenessof theirpermutationalnature.As a simpleexample: transin a combinationalsystem
position,exceptingthe identitytransposition,
resultsin the adjoining of pitcheswhich are not presentin the original
collection,and thus establishesa new sub-collection;transpositionof a
set resultsonly in a permutationof the elements.Also, compositional
traditionally,
impliescontourpreservation,a consideration
transposition,
that is, literally,meaninglessin definingtranspositionas a twelve-tone
operation,since contouris a functionof the registralspecificationof the
elements,and registralchoice is as undefinedby the structureof a set
as is duration,intensity,
timbre,or any of the otherattributesnecessarily
associated with a compositionalrepresentationof a set; as a result,a
set cannot be statedin musical notationwithoutthe additional qualificationthateach pitchsignbe takento signifythe totalpitchclass a member
of which it denotes. Since such a qualificationonly too easily leads to
but another confusionof systematicprinciplewith compositionalpermissive ("a tone may be stated in any octave"), it is both safer and
-to representa twelve-toneset in numericalnotation,by
more efficient
an ordered number couple succession,the firstmember of the couple
order positionin S ("order number"), the second signifying
signifying
the "pitch number" of the pitch class. The initial pitch class of S is
denotedby the couple (0,0), and is taken as the originof the coordinate
systemforboth order and pitch numbers,both of which range over the
integers0 - 11 inclusive,each integerappearing once and only once
as an ordernumberand a pitch number.In the case of ordernumbers,
this representsthe fact that twelve and only twelve pitch classes are
involved: in the case of pitch numbers,thisis the arithmeticalanalogue
of octave equivalence (congruencemod. 12).2 In this notation,the set
of the Schoenberg Third String Quartet, with registralrepresentation
chosen arbitrarily,
Ex. I

(iier)

4 8

is represented:0,0; 1,9; 2,8; 3,2; 4,5; 5,10; 6,11; 7,4; 8,3; 9,6; 10,1;
12! = (12 X 11 X 10X 9 ... X 1).

2 Two numbers, a and b, are said to be "congruent mod. 12" if, and only if,
a-b = k.12 where k is an integer (including zero).

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

Twelve-Tone Invariants as Compositional Determinants

249

11,7. Thus, the succession of differencesderived by the subtraction


(mod. 12) of a set number from the followingset number is the
ordered interval succession determinedby a set, and each of these
intervalnumbersdenotes,accordingly,a class of intervals.
The operation of transpositionapplied to any set can be represented by adding (mod. 12) an integer,O - 11, to each pitch number
of the set. Thus, if (a,b) is the couple signifyingan element of S,
then the transpositionis representedby (a,b+t), with t termed the
"transpositionnumber." Thus, the transpositionaloperation (designated: T) is convenientlyregarded as an operationon, a permutation
of,pitchnumbers;forany specifiedset,it could equally well be regarded
as effectinga permutationof order numbers,but the previous characterizationcorrespondsmore appropriatelyto the general conception
of transposition.
The totalityof twelvetransposedsets associated with a given S constitutesa permutationgroup3of order 12; as such it is closed, disjunct
with regard to any othercollect;onof sets T derived froma set whose
intervallicsuccession differsfrom that of any member of this totality.
Thus, it constitutesa combinationalcollectionof sets within,not only
the totalityof all possiblesets, but the totalityof sets derivablefroma
given S by the operationsof the system.
By virtueof the group structurethat it generates,and the additional
propertiesthat derive from the commutativeand transitivenature of
this group, a multitudeof attributesnecessarilyassociated with T can
be formallydeduced. The musical relevance of these attributescan be
decided only empirically,of course, but it is my purpose here to
examine only a few of those operationalinvariants(propertiesof a set
A "group" is a systemwhose elements (denoted a, b, c . .), an operation
(denoted *), and an equivalence relation (denoted=) satisfythe followingproperties:
1. Closure: If a, b are elements of the system,then a*b is an element of the
system.
2. Associativity: If a, b, c are elementsof the system,then (a*b)*c = a*(b*c).
3. Existence of an identity: There is an element of the system,e, such that,
for each element of the system(say, d), d*e = e*d = d.
4. Existence of an inverse: For each element of the system(say, d), there exists
an element of the system,d-, such that d*d- = d-*d = e.
In interpretingthe twelve-tone system as a group, the elements of the group
are twelve-tone sets, represented as permutations of pitch or order numbers; the
operation is the multiplication of permutations. "S" is the identity element. The
"order" of a group is the number of elements of the group.
In addition, the groups presented here have the propertyof "commutativity":
if a, b are elements of the system,then a*b = b*a.

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

250

The Musical Quarterly

that are preservedunder the operation, as well as those relationships


between a set and the so-operationallytransformedset that inhere in
the operation) which may be termed "musical invariants,"requiring
for their aural recognitionmerely the ability to perceive pitch class
and intervalclass identityand non-identity.
identityand non-identity,
The familiarinvariantassociated with T is that of preservationof
the interval number succession. For all its obviousness,it appears a
powerfullycohesive propertyin the light of the total non-invariance
of pitch classes with regard to order; that is, no order, pitch number
couple remainsfixed under T. Since each t produces a total derangement of the set elements,and the identicalintervallicsuccession,neither
of thesepropertiescan serve as the bases of differentiation,
in the search
for possible criteria for the compositionalhierarchizationof transpositions. Similarly,every value of t definesa regular permutation,but
an invariant basis for differentiationappears with the recognition
that complementaryt's (numbers whose sum is 0, mod. 12), and only
such t's, produce inverse permutations,of equal order. Beyond the
immediate boundary conditions on the intervallicstructureof a set
inferrable from this fact, a consequential musical propertyfollows
from the furtherfact that inverse permutationsproduce the same
number of order inversions.This measure of the extentof order rearrangementof the pitch classes can be described most easily by, for
the moment,regardingT as an operation on the order numbersof S;
an order inversionis each relationamong pairs of order numbersthat
violates the normal ascending relation among order numbers in S.
For example, in the set of Ex. 1, the application of, let us say, t = 4
to the set produces the followingorder numbersuccession: 7, 10, 0, 9,
8) produces the order
1,3,8,2,11,5,4,6.The complementaryt (t
The
number of order innumber succession: 2,4,7,5,10,9,11,0,6,3,1,8.
versionsproduced by each is 32.
Complementaryt's produce the same number of pitch adjacencies
with regard to S, both ordered adjacencies and reversedadjacencies.
(This conditionof adjacency is imposed merely in the light of the
simplest compositional exploitation of this property,which is immediatelyextensibleto pairs of pitchclassesassociatedwithany distribution
of order numbers.) If a set possessessuccessivepitch classes represented
by pitch numbersa and b, and successivepitch classes representedby
pitch numbersc and d (c may or may not be equal to b, and similarly
ford and a), and ifb - a -- d - c, thenthereis at such thata + t = c,
and b + t = d, so that under t, a and b are associatedwith the original

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

Twelve-Tone Invariants as Compositional Determinants

251

ordernumbersof c and d, and it thenfollowsthat under 12 - t, c and d


are associated with the original order numbersof a and b. So, too, for
reversedadjacencies, representedin the set by complementaryintervals.
The intervallicstructureof S, then,determinesthenumberof adjacencies
preservedunder a particulart and its complement,since this numberis
a functionof the multiplicityof and relative pitch placement of the
identical and complementaryintervalnumbers in S. Consider the set
of Ex. 1: the interval succession determined by disjunct dyads is
9,6,5,5,3,6. The interval between the identical 5's is 6, between the
complementaryintervals9 and 3 is also 6, and interval 6 is its own
complement.So, under the application of t- 6:
Ex. 2

the pitch content of disjunct dyads is preserved,and Ex. 2 can thus


be regarded as a permutationof the dyads of Ex. 1; if the succession
of dyads in Ex. 1 be numbered 1 -6 inclusive,then those in Ex. 2
are in the order5,2,4,3,1,6.In the usual cyclicnotation,the perinutation
is (1 5) (3 4). This demonstratesan immediate means of extending
serial transformation
to compounds of serial elements; I shall returnto
this aspect of the example later, but it should not be overlookedthat
in this possibilityof holding a pair of pitch classes (as opposed to a
pitch class) fixedwith regardto order and pitch content,thereis immanent the extensionto the fixed contenttrichord,tetrachord,hexachord,
etc., or, in other words, to the combinatorialset.
One more propertyof complementarilytransposedsets should be
indicated. This involves any segmentof S (by segmentis meant any
number of successiveset elements,although the propertyholds equally
for any selection of elements,non-consecutiveas well as consecutive),
and the correspondingsegmentsof any transpositionof the set and the
complementarytransposition.For example, consider the firstseven
elements of Ex. 1, and the correspondingelements of transpositions
with t -- 2 and t- 10:

Ex.3

(-)..(a):-

Considered with regardto this segmentof Ex. 1, both segments3a and


3b have the same number of pitch classes in common with it: four.

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

252

The Musical Quarterly

But, in addition,the patternof intersectionin termsof order numbers


of 3a with regard to 1 is: 0,1,2,5; the patternof intersectionof 1 with
regard to 3b is similarly: 0,1,2,5. The actual pitch classes involved,
naturally,are not necessarilyidentical,and are not in the presentcase.
This operational invariant resultingfrom complementarytransposition
is not only of obvious rhythmicand functionalsignificancecompositionally, but of essentialsystematicconsequence in the theoryof general
combinatoriality,
aggregatestructure,and the resultantmeans of hierarchizationof set segments.
The importance of transpositionalcomplementationalone would
serve to suggestthe systematicoperation of inversion(designated: I),
which is definableas complementationmod. 12 of each pitch number
of the t applied to all pitchnumbers
of S, as opposed to complementation
of S. Given a set element (a,b), I transformsit into (a,12 - b), or,
more generally (a, (12 - b) + t), since T is applicable uniquely to
the invertedset; with relationto the complex of sets generatedby T,
the invertedset assumes the local role of S. I and T commuteonly to
within complementation;therefore,the order of operations must be
specified,and I shall assume throughoutthisdiscussionthat T is applied
afterI (IT).
At this point,it is appropriateto considercomparable definitionsof
the remainingoperationsof the system,which reveal that retrogression
of order
(designated: R) can be regardedas affectingcomplementation
into (11 - a, b); thereforeRI (or the
numbers: (a,b) is transformed
the
since
merelyinvolvesthe simultaneous
commute)
reverse,
operations
application of both complementationoperations: (a,b) is transformed
into (11 - a, 12 - b). As in the case of I, t's are applicable to the
set numbers.
of S definedby theseoperations,
The presentationof the permutations
mod. transposition(in other words, this representationis independent
of the operation of transposition),in the usual group multiplication
table, with multiplicationpermutationsas the group operation,and S
denotingthe identitypermutation:4
RI
R
I
S
R
S
RI
I
I
S
RI
R
S
I
R
RI
4 The group table is read by choosing an element in the firstcolumn and the
element with which it is to be multiplied in the firstrow; the result of this multiplication is found at the intersectionof the row of the firstand the column of the second.

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

Twelve-Tone Invariants as Compositional Determinants

253

reveals,firstof all, that this collectionof permutationsis an instanceof


a group, and as such possessesthe propertyof closure, thus assuring
another combinationalaspect of the systemin the large, since the collection of sets determinedby these permutationsis disjunctwith regard
to any so determinedcollectionof sets one of whose setsis not a member
of this collection,and is identical with a collectionone of whose members is.
The revealed symmetrypropertiesindicate the hazardous connotations of such termsas "basic" or "original" set to denote other than a
set formnorm decided upon purelyon the basis of contextualconsiderations (temporal priority,for example), since such termscannot designate any attributeof set structurein a general sense, either from a
standpoint of internal propertiesor of relation to other set forms.
Similarly,the fact that the period of each of the permutationsis 2
should, in itself,dispel once and forall thosefutileattemptsto "equate"
these operationswith tonal functions.
A vast literatureof group theorysuppliesnecessarypropertiesof such
a structure,
and it is not withoutextra-musicalinterestthatthisparticular
group of permutationsis an instanceof a familiargroup structure,the
so-called "four-group."
The twelve-tonesystem,as system,is indeed "simple." It is simple
in its principlesof formationand transformation,
but enormouslycomin the necessaryinferencesthat can
plex and deep in its ramifications,
be drawn fromthese principles,for it is of the formalmodel of which
it is an exemplification
that Hermann Weyl has said: "From theseinsignificantlooking assumptionssprings an abundance of profound relationships; and mathematicsoffersan astounding variety of different
of this simple axiom system."5
interpretations
Inversion, in the traditional sense, implies inversion of contour.
In the twelve-tonesysteminversion,like transposition,
cannot be characterizedin termsof registralconsiderations,but merelyas that permutation of pitch class numbers (or, for particular purposes, of order
numbers) which resultsfrom the substitutionof complementarypitch
numbersin S; it followsthat there is an accompanyingsubstitutionof
a successionof complementaryintervalsforthe intervalsuccessionof S.
Even more, perhaps,than in the case of T, I must derive its "justification"fromits associated musical invariants.It must be emphasized
5 Philosophy of Mathematics and Natural Science, Princeton, 1949, p. 28.

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

254

The Musical Quarterly

that, although invariantsare associated necessarilywith the operations


in question,the degree to which theyare projectedexplicitlyin compositionaltermsdependsupon the emphasistheyreceivefromothermusical
components: rhythm,dynamics,register,phrasing,timbre, etc. Conversely,the desire for the compositionalexploitationof these fundamental propertiesmay be regarded as, at least, a partial determinant
of the compositionalcharacteristicsimposed on these components.
Consider the simplestinversionalinvariant: if (a,b) of S is transformedby IT into the correspondingorderelement (a, (12 - b) + t),
where (12 - b) + t may or may not be equal to b, then,corresponding to (c,(12 - b) + t) of S is (c,b). This property,as may be observed by referenceto our examples, is possessedalso by transposition
when t = 6; it holds for all t's applied to the inversion.
Again, the cyclic representationof the pitch class permutationseffectedby IT shows that all even t's produce similarpermutationsof six
cyclesof two elementseach (thus, regularpermutations),while odd t's
produce similar permutationsof five two-elementcycles and two unit
cycles. So, although complementaryt's still produce similar permutations,thisis merelybecause theyare eitherboth even or both odd. Odd
values of t, then, determinesix dyadic pitch classes between elements
of the same order number in I related sets, and even values of t
classes. These
determinefivesuch dyadic classes,and two single-element
latterrepresentset elementswhose order number,pitch number couple
remains unchanged under IT. It is for this reason that a necessary
condition for hexachordal inversionalcombinatorialityis that the sum
of the set numbersof the same order number in the I related sets be
odd. The pitch number of an element so fixed is equal to one half
such fixed elementsassociated with a given even t
of t; thus,
the.two
are unique, and are tritonerelated (since 12/2 = 6).
In addition to this partitionof the elementsof inversionallyrelated
sets into pitch classes, the IT operation also effectsa categorization
into interval classes. Since the intervalsbetween pitch classes of the
same order numberin I related sets are eitherall even or all odd, and
thus- each intervaloccursexactlytwice,it can be shown that pitch
classes in S whose pitch numbersdifferby 6 are associated with the
same intervaldeterminedby the elementof the same ordernumberin the
I related set.
The largely "note against note" presentationof the canon in the
second movementof Webern's Variations for Piano places these char-

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

Twelve-Tone Invariants as Compositional Determinants

255

acteristicsin the foreground.The initiallystated, I related formsof the


set (Ex. 4a) :
Ex. 4

(t)
(c)/

since t = 2 (taking the lower of the two sets as the reference"prime"


set), hold (1,1) and (11,7) fixed. Since the firsthexachord of S contains no tritonerelated pitches, there is no repetitionof pitch dyads
formed by elements of the same order number, so that, if the first
succession of six dyads is numbered 1,2,3,4,5,6, then the following
successionis a permutationof these: 6,4,1,5,3,2. As in the Schoenberg
example, although in a differentmanner, complexes of pitch elements
become, themselves,subjected to serial permutation.
The continuation of this canon demonstratesthe compositional
use of anotherinvariant.By choosing,as the initial pitches (or, as any
pitchesof the same order number,since the sums of the pitch numbers
of such pairs are equal) of the T formsof the S and the I related set,
elementsthe sum of whose pitch numbers,with regard to the original
referencepoint "gg" = (0,0), is equal to the original t - 2, the
pitch dyads resultingare identical with those created in the firstinversional juxtaposition.The recurrenceof repeated "a's" is merely one
manifestationof this general property.See Ex. 4b. Therefore,for each
firstelementof a set,thereis one and onlyone choice of the firstelement
of the I related set which holds the pitch dyads so fixed with regard
to a pre-definednorm. (In the light of the previousdiscussion,it is of
interestto point out that an equivalent statementof this condition is
that the initial I related formsbe transposedby complementaryvalues
of t.) Webern chooses fourof these twelvepossibilitiesto determinethe
pitch levels of the successivesections of the movement; the fiist and
last pitches of the I related sets which provide the pitch content of
the thirdand fourthsectionsof the work are shown in Ex. 4c. The first

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

256

The Musical Quarterly

movementof Webern's Quartet,Op. 22, employsthe same procedurein


a more elaborate and extendedmanner.
The morecommonand traditionalprocedureof "totallytransposing"
such a section - in my terms,the applying of the same t to both
simultaneouslystated forms- preservesthe interval succession but
not (except the single case where t - 6) the fixed pitch dyads: the
procedureunder discussionhere permutesthe intervalsuccessionwhile
retainingthe pitch contentof the dyads.
If we numberthe dyads of the firstsection from 1 - 12 inclusive,
the second section yields a permutationcorrespondingexactly to the
permutationof order numbers that resultsfrom applying t =- 5 to
the upper set, or - equivalently- t = 7 to the lower set, and - of
course - correspondingly
with the thirdand fourthsections.Thus, all
the propertiesassociated with the application of T to S are translatable
into propertiesof permutationsof dyads between I related sets.
Webern's particular choice of transpositionsappears to be related
primarilyto concerns of compositionalduration and external design.
The transpositionalchoice for the second sectionmakes possible a final
dyad for this section which is pitch identical with the initial dyad of
the movement,but with pitch componentsreversedas to set membership. The repeat of these firsttwo sectionsis founded on this identity.
determinedby exact analogy with
The thirdsection is transpositionally
the second section, throughthe "double function" of the final dyad
of the precedingsection. But the continuationof this basis of choice,
since the intervalbetween the firstand last elementsof the set is prime
to 12, would carry the work throughall possible twelve jointlydetermined transpositionsbefore returningto the firstdyad in its original
disposition.So, the fourthsectionemploysthe principleof interchange,
already introducedinto the workby the repeat of the firsttwo sections.
The fourthsectionthuseffectsa returnto the firstdyad of the movement,
while the repeat of the third and fourthsections as a unit resultsin
another interchange,necessarilythe exact reverseof the interchange
resultingfromthe firstrepeat.
Closely related to these invariantsis the property: if (a,b) and
+
1,c) are two successiveelementsof S (the provisionof succession
(a
is, actually, unnecessary,but is introducedhere for purposes of simplicity), and (d,e) is an elementof an I related set (where d may or
may not be identical with a or a = 1, and e may or may not be
identicalwith b or c), then the intervallicsuccessionb - e, c - e is

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

Twelve-Tone Invariants as Compositional Determinants

257

identical with the succession g - f, h - f, defined by d,f of the


initial set, and a,g and a + 1, h of the I related set. In the Contrapunctus Secundus of the Quaderno Musicale di Annalibera, which is
in many structuralrespectsclosely similar to the Webern movement
just discussed,Dallapiccola uses this propertyin the second half of the
piece as a means of unfoldingthe same intervallicprogressionby the
two canonic parts, while reversingtheir relation of temporal priority.
This property,as well as the fixeddyad property,is particularlysignificant as a harmonic factor when extended to include more than two
simultaneouslystated,I relatedsets.
As in the case of T, conditionsforthe retentionof pitch adjacencies
under IT are statable easily and fully.However, I merelyshall return
to Ex. 1, and examine the resultof applyingIT witht = 3, and t = 9.
Ex. 5
[

ar

The pitch contentof disjunct dyads is preserved,and the permutations


of these dyads under I for t = 3 is (1 5) (2 6), and for t = 9 is
(2 6) (3 4). Taken togetherwith the identitypermutationand the
permutationunder t -= 6, this group of permutationsleaves each
dyad twice fixed with regard to order, once with its component elementsin the order definedby S, and once with the elementsreversed;
of the four occurrences of each dyad, two maintain the order of
elementsdefinedby S, and two reversethisorder. More generally,if we
denote by A the identitypermutationof the dyads, by B the permutation on the dyads effectedby t = 6, by C and D the permutations
effectedby IT with t equal respectivelyto 3 and 6, the multiplication
table for thisgroup of permutationsis:
D
A
B
C
B
D
A
C
C
D
A
B
D
C
B
A
This group is isomorphicwith that formedby the permutationsrepresentingthe identity,I, R, RI operationsof the system.
Finally, the familiar phrase: "the identificationof the horizontal
with the vertical," implies much more than a mere compositional
prescriptivewith regard to the spatial distributionof the elementsof a

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

258

The Musical Quarterly

set when it is realized that adjacent pitch elementsof a set become


elementsof the same ordernumberin I relatedsets,when thesesets are
so chosen that the sum of the pitch numbersof any two elementsof the
same order number is equal to the sum of the pitch numbersof the
originallyadjacent elements.
Space does not permita considerationof invariance under R and
RI. However, it must be pointed out that the traditionalconceptionof
retrogressionas effectingthe temporal reversal of pitches constitutes
neithera meaningfuldescriptionnor a "justification"for its position
in the twelve-tonesystem.For, even with registralconsiderationsdisregarded, this characteristicis associated with but one transpositional
level of the retrograde;but all transpositionallevels of the retrograde
present the intervallicsuccession of the inversionin reversed order,
formspresentthe interwhile - necessarily- the retrograde-inversion
vallic succession of the prime (S) in reversedorder. Thus, the RI
forms,often regarded as the aurally most unrealistictransformations,
because the operation is viewed as applied to pitch successionrather
than to intervalsuccession,require for the perceptionof their relation
to S merely the ability to identifyinterval classes. In this important
sense,the RI formscan be regardedas the mostcloselyrelatedto S, and
are so employedoftenby Schoenbergin his compositional,"thematic"
presentationof successiveset forms: see, for example, the Variations
for Orchestra,the third movementof the Fourth StringQuartet, and
the Piano Concerto. In the pitch class order,intervalclass orderduality
related sets reside many
between retrogradeand retrograde-inverted
of the most importantpropertiesof such transformations.
Even so incompleteand informala discussionof so small a number
of the invariantsattendingthe operationsof the systemindicates,I hope,
somethingof the essential importance of this subject, analyticallyin
the "rational reconstruction"of compositions,and compositionallyin
comprehendingand masteringthe materialsof the system.If I have led
the discussion more in the direction of those aspects which suggest
the "macrocosmically"combinationalfeaturesof this basically permutational system,I could have - with equal appropriatenessand the
same means - examined the "microcosmically"combinationalfeatures
(particularlywhat are termedin group theory"imprimitivesystems"),
set structure(particularlywith regard to redundancyproperties),combinatoriality,generalizedpartitioning,derivation,sequence theory,and
related questions.
Certainly,any conjecturesabout "generalized" serialismmust con-

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

Twelve-Tone Invariants as CompositionalDeterminants

259

frontthe problemas to whethersuch alleged generalizationsresultin a


maintenance,an increase,or a decrease, of the number and scope of
such invariants,and whetherthe apparent "freedom"of such "generalizations" does not, in a deeper sense, reduce structuralresourcesrather
than augmentthem.
Likewise, I would insistthat a necessaryconditionfor the application of the permutationaloperationsof the twelve-tonesystemto orderable non-pitchelementsis that the rules of correlationbe so arrivedat
that theseinvariants,which are necessaryconsequencesof the pitch class
nature of the system,be susceptibleto musicallymeaningfulinterpretationsin theseother,perhapssignificantly
dissimilar,domains.
In conclusion,I can state only,withouthoping to have done more
than intimate the bases for such a statement,that an "exhaustion"
of the resourcesof the twelve-tonesystemin the relevantfutureis not
only unforseeable,but unthinkable.I trustI have begun to document,
and will be given the opportunityin the futureto furtherdocument,
the statementthat, in its vastnessof structuralmeans, its flexibility,
and its precision,the twelve-tonesystemcedes nothingto any musical
systemof the past or presentthathas engaged the mind of musical man.

This content downloaded from 156.143.240.16 on Fri, 28 Jun 2013 15:39:35 PM


All use subject to JSTOR Terms and Conditions

You might also like