G odel Without Tears 1

Incompleteness the very idea

Peter Smith
October 16, 2009 Why these notes? After all, Ive written An Introduction to G odels Theorems (CUP, heavily corrected fourth printing 2009: henceforth IGT ). Surely thats more than enough to be going on with? Ah, but theres the snag. It is more than enough. In the writing, as is the way with these things, the book grew far beyond the scope of the lecture notes from which it started. And while I hope the result is still pretty accessible to someone prepared to put in the time and eort, theres a lot more in the book than is really needed by philosophers meeting the incompleteness theorems for the rst time. After all, you might want to get your heads around only those technical basics which are actually needed for understanding philosophical discussions about incompleteness. So you need a cut-down version of the book an introduction to the Introduction ! Well, isnt that what lectures are for? Indeed. But theres another snag. I havent got many lectures to play with. So either (A) I crack on at quite a fast pace (hard-core mathmo style), cover those basics, but perhaps leave too many people puzzled and alarmed. Or (B) I do relaxed talknchalk, highlighting the really Big Ideas, making sure everyone is grasping them as we go along, but inevitably omit important stu and leave quite a gap between what happens in the lectures and what happens in the book. What to do? Im going for plan (B). But then I still need to do something to ll that gap between lectures and book. Hence these notes. The idea, then, is to give relaxed lectures, highlighting Big Ideas, not worrying too much about depth or ne-detail (or even about getting through all of the days intended menu of topics). Then after the lecture, Ill write up notes that expand things just enough, and then give pointers to relevant chunks of IGT. The idea, however, is for the notes to be more or less stand-alone, and to tell a brief but coherent story read by themselves. So occasionally Ill copy a paragraph or two from the book, rather than just refer to them. Warning: just occasionally in these notes, Ill no doubt apply that good maxim Where it doesnt itch, dont scratch. In other words, sometimes Ill say things that are not utterly rigorous, but I hope in unworrying ways. If you are bright enough to spot the slight cheats or corner-cutting, you should be bright enough to spot how to repair what I say, at the cost of a bit of fuss and bother, so no harm done!

Kurt G odel (19061978)

The greatest logician of the twentieth century. Born in what is now Brno. Educated in Vienna. At 23, his doctoral dissertation established the completeness of the rst-order predicate calculus. Later he would do immensely important work on set theory, as well as make contributions to proof theory. Talk of G odels Theorems, however, typically refers to his two incompleteness theorems in an epoch-making 1931 paper. Left Austria for the USA in 1938, and spent rest of his life at the Institute of Advanced Studies at Princeton. Always a perfectionist, after the mid 1940s he more or less stopped publishing. For a brief overview of his life and work, see http://en.wikipedia.org/wiki/Kurt G odel, or better though youll need to skip http://plato.stanford.edu/entries/goedel. Theres a nice biography, John Dawson Logical Dilemmas (A. K. Peters, 1997), which will also give you some sense of the logical scene in the glory days of the 1930s.

On formally undecidable propositions of Principia Mathematica and related systems I

This is the title of the 1931 paper which proves the First Incompleteness Theorem and states the Second Theorem. (The I indicates that it is the rst part of what was going to be a two part paper, with Part II spelling out the proof of the Second Theorem. But that was never written. Ill explain later why G odel didnt need to bother.) Even the title gives us a number of things to explain. Whats a formally undecidable proposition ? Whats Principia Mathematica ? youve heard of it, no doubt, but whats the project of that triple-decker work? What counts as a related system ? In fact, what is meant by system here? Take the last question rst.

2.1

Systems i.e. formal axiomatized theories

Our concern is with formal axiomatized theories. T is such a theory if it has (i) a formalized interpreted language L, (ii) an eectively decidable set of axioms, (iii) a formalized proof-system in which we can deduce theorems from the axioms. To explain, we rst need a denition: Defn. 1. A property P dened over a domain D is eectively decidable i theres an algorithm for settling, for any o D, whether o has property P i.e. theres a step-by-step mechanical routine for settling the issue, a suitably programmed computer could do the trick. A set is eectively decidable if the property of being a member of that set is eectively decidable. (i) Well take the idea of a formal interpreted language L to be familiar from earlier logic courses. There will be a syntax which xes which strings of symbols form terms, which form ws, and in particular which strings of symbols form sentences, i.e. closed ws with no unbound variables dangling free. There will be a semantics which gives the intended interpretation of L, xing truth conditions for each L-sentence. Crucially, to emphasize, Defn. 2. For a formalized language L, the syntactic rules of L must be such that the properties of being a term, a w, a w with one free variable, a sentence, etc. are eectively decidable. For whats the point of having a formal syntax if we cant then decided whether a string of symbols is or is not a w, for example? (ii) A theory T built in language L will have a certain class of L-sentences picked out as axioms. Again it is to be eectively decidable whats an axiom. (After all, if we build a theory and then cant routinely tell whether a given sentence is one of its axioms, what would be the point of that?) (iii) Just laying down a bunch of axioms would normally be pretty idle if we cant deduce conclusions from them! So a formal axiomatized theory T comes equipped with a proof-system, a set of rules for deducing further theorems from our initial axioms. Again, describing a proofsystem such that we couldnt routinely tell whether its rules are being followed would have little point. Hence we naturally require that it is eectively decidable whether a given array of ws is indeed a proof according to the rules. But it doesnt matter whether the proof-system is a Frege/Hilbert axiomatic logic, a natural deduction system, a tree/tableau system so long as it is indeed eectively checkable that a candidate proof-array has the property of being properly constructed according to the rules. So, in summary Defn. 3. A formal axiomatized theory T has a formalized language L, a certain class of Lsentences picked out as axioms where it is decidable whats an axiom, and it has a proof-system such that it is eectively decidable whether a given array of ws is indeed a proof according to the rules. Careful!!! To say that, for a properly formalized theory T it must be eectively decidable whether a purported T -proof of is indeed a kosher proof according to T s deduction system is not, repeat not, to say that it must be eectively decidable whether has a proof. It is one thing 2

to be able to eectively check a proof once found, it is another thing to be able to eectively decide in advance whether there is exists a proof to be discovered. Finally, a couple of useful notational conventions should be mentioned for future reference. 1. Greek letters, as in the weve just used, are schematic variables in the metalanguage in which we talk about our formal systems. 2. Particular expressions from formal systems and abbreviations of them will be in sans serif type. For more explanations, see IGT, 2.2, 3.13.3, 4.1.

2.2

Formally undecidable propositions and negation incompleteness

Defn. 4. T says: there is a formal deduction in T s proof-system from T -axioms to the sentence as conclusion. If is a sentence and T , then is said to be a theorem of T . So NB, ocially signies provability in T , a formal syntactically denable relation, not semantic entailment. Defn. 5. If T is a formal theory, and is some sentence of the language of that theory, then T formally decides i either T or T . Hence, Defn. 6. A sentence is formally undecidable by T i T and T . Another bit of terminology: Defn. 7. A theory T is negation complete i it formally decides every closed w of its language i.e. for every sentence , T or T Trivially, then, there are formally undecidable propositions in T if and only if T isnt negation complete. Of course, it is very easy to construct negation-incomplete theories: just leave out some seemingly necessary basic assumptions about the matter in hand! But suppose we are trying to fully pin down some body of truths using a formal theory. We x on an interpreted formal language L apt for expressing such truths. And then wed ideally like to build a theory T in L, whose axioms are such that when (but only when) is true, T . So, making the classical assumption that either is true or is true, wed like T to be such that either T or T . Negation completeness, then, is a natural desideratum for theories. For more explanations, see IGT, 3.4.

2.3

Deductivism, logicism, and Principia

The basic arithmetic of successor (next number), addition, and multiplication is childs play (literally!). It is entirely plausible to suppose that, whether the answers are readily available to us or not, questions posed in the language of basic arithmetic the language of successor, addition, and multiplication plus familiar rst-order logical apparatus have entirely determinate answers. These answers are surely xed by (a) the basic zero-and-its-successors structure of the natural number series and (b) the nature of addition and multiplication as given by the school-room explanations. So it is surely plausible to suppose that we should be able lay down a bunch of axioms which characterize the number series, addition and multiplication (which codify what we teach the kids), and that these axioms should settle every truth of basic arithmetic, in the sense that every such truth of the language of successor, addition, and multiplication is logically provable from the axioms. For want of a standard label, call this view deductivism about basic arithmetic. What could be the status of the axioms? You might, for example, be a Kantian deductivist who holds that the axioms encapsulate intuitions in which we grasp the fundamental structure of

the numbers and the nature of addition and multiplication, where these intuitions are a special cognitive achievement in which we somehow represent to ourselves the arithmetical world. But talk of intuition can be very puzzling and problematic. So we might prefer Freges view that the axioms are analytic, truths of logic or rather of logic-plus-denitions. On this view, we dont need Kantian intuitions going beyond logic. The Fregean brand of deductivism is standardly dubbed logicism. Famously, Freges attempt to be a logicist deductivist about arithmetic hit the rocks, because Russell showed that his logical system is inconsistent in a pretty elementary way (it is beset by Russells Paradox). That devastated Frege, but Russell was undaunted, and still gripped by deductivist ambitions he wrote: All mathematics [yep! all mathematics] deals exclusively with concepts denable in terms of a very small number of logical concepts, and . . . all its propositions are deducible from a very small number of fundamental logical principles. Thats a big promisory note in Russells The Principles of Mathematics (1903). His attempt (with Whitehead) to make good on that promise is Principia Mathematica (three volumes, though unnished, 1910, 1912, 1913). The project is to set down some logical axioms and denitions and deduce the laws of arithmetic from them. Famously, they eventually get to prove that 1 + 1 = 2 at *110.643 (Volume II, page 86), accompanied by the wry comment, The above proposition is occasionally useful.

2.4

G odels bomb

There are technical complications which means that Principia s axioms are not all clearly logical (in particular theres an appeal to a brute-force Axiom of Innity which in eect states that there is an innite number of objects, and to a very dodgy Axiom of Reducibility : for more on this see http://plato.stanford.edu/entries/principia-mathematica/). But leave those worries aside they pale into insignicance compared with the bomb exploded by G odel. For his First Incompleteness Theorem shows that any form of deductivism about basic arithmetic (not just Principia s) is in trouble. Why? Well the proponent of deductivism about basic arithmetic (logicist or otherwise) wants to pin down rst-order arithmetical truths about successor/addition/multiplication, without leaving any out: so he wants to give a negation-complete theory. And there cant be such a theory. G odels First Theorem says at a very rough rst shot that nice theories containing enough arithmetic are always negation incomplete. So varieties of deductivism, and logicism in particular, must always fail a major (and perhaps very surprising) result! Hold on! Ive heard of neo-logicism which has its enthusiastic advocates. How can that be so if G odel showed that logicism is a dead duck? Well, we might still like the idea that some logical principles plus what are more-or-less denitions together semantically entail all arithmetical truths, while allowing that we cant capture the relevant entailment relation in a single properly axiomatized deductive system of logic. Then the resulting overall system of arithmetic wont count as a formalized theory of all arithmetical truth since its logic is not formalizable, and G odels theorems dont apply. But more about that in due course.

2.5

The First Incompleteness Theorem just a bit more carefully, in two versions

Three more denitions: Defn. 8. The formalized language L contains the language of basic arithmetic if L has at least the standard rst-order logical apparatus, has a term 0 which denotes zero and function symbols for the successor, addition and multiplication functions dened over numbers either built-in as primitives or introduced by denition and has a predicate whose extension is the natural numbers.

The point of that last clause is that if N is a predicate satised just by numbers, then the w x(Nx (x)) says that every number satises ; so L can make general claims specically about natural numbers. (If L is already dened to be a language whose quantiers run over the numbers, then you can just use x = x for N.) Defn. 9. A theory T is sound if its axioms are true (on the interpretation built in to T s language), and its logic is truth-preserving, so all its theorems are true. Defn. 10. A theory T is consistent if there is no such that T and T . In a classical setting, if T is inconsistent, then T for all . And of course, trivially, soundness implies consistency. G odel now proves (more accurately, gives us most of the materials to prove) the following: Theorem 1. If T is a sound formalized theory whose language contains the language of basic arithmetic, then there will be a true sentence GT of basic arithmetic such that T GT and T GT , so T must be negation incomplete. However that isnt what is usually referred to as the First Incompleteness Theorem. For note, Theorem 1 is about what follows from a semantic assumption, namely that T is sound. And soundness is dened in terms of truth. Now, post-Tarski, we arent scared of the notion of the truth! But G odel was writing at a time when, for various reasons, the very idea of truth-inmathematics was under some suspicion. So it was extremely important to him to show that you dont need to deploy any semantic notions to get (again roughly) the following result: Theorem 2. For any consistent formalized theory T which contains a certain modest amount of arithmetic (and has a certain additional desirable property that any sensible formalized arithmetic will share), there is a sentence of basic arithmetic GT such that T GT and T GT , so T must be negation incomplete. Of course, well need to be a lot more explicit in due course, but that gives the avour of the result. The contains a modest amount of arithmetic is what makes a theory suciently related to Principia s for the theorem to apply. Ill not pause in this lecture to spell out that just how much arithmetic that is, but well nd that it is stunningly little. (Nor will I pause now to explain that additional desirable property condition. Well meet it in due course, but also explain how by a cunning trick discovered by J. Barkley Rosser in 1936 how we can drop that condition.) For the present, however, lets concentrate on the semantic version of G odels theorem, i.e. Theorem 1.

2.6

Theorem 1 is better called an incompletability theorem

Suppose T is a sound theory which can express claims of basic arithmetic. Then we can nd a true GT such that T GT and T GT . Of course, that doesnt mean that GT is absolutely unprovable, whatever that could mean. It just means that GT -is-unprovable-in-T . Now, we might want to repair the gap in T by adding GT as a new axiom. So consider the theory U = T + GT (to use an obvious notation). Then (i) U is still sound (for the old T -axioms are true, the added new axiom is true, and the logic is still truth-preserving). (ii) U is still a properly formalized theory, since adding an specied axiom to T doesnt make it undecidable what is an axiom of the augmented theory. (iii) U still can express claims of basic arithmetic. So G odels First Incompleteness Theorem applies, and we can nd a sentence GU such that U GU and U GU . And since U is stronger than T , we have a fortiori, T GU and T GU . In other words, repairing the gap in T by adding GT as a new axiom leaves some other sentences that are undecidable in T still undecidable in the augmented theory. And so it goes. Keep chucking more and more additional true axioms at T and our theory still remains negation-incomplete, unless it stops being sound or stops being eectively axiomatizable. In a good sense, T is incompletable.

How did G odel prove the First Theorem (in the semantic version)?

Lets take a rst pass at outlining how G odel proved the semantic version of his incompleteness theorem. Obviously well be coming back to this in a lot more detail later, but we can give just a avour of whats going on. We kick o two natural denitions. Defn. 11. If L contains the language of basic arithmetic, so it contains a term 0 for zero and a function expression S for the successor function, then the terms 0, S0, SS0, SSS0, . . . , are Ls standard numerals, and well use n to abbreviate the standard numeral for n. Henceforth, well assume that the language of any theory we are interested in contains the language of basic arithmetic and hence has standard numerals denoting the numbers. Defn. 12. The formal w (x) of the interpreted language L expresses the numerical property P i (n) is true on interpretation just when n has property P . Similarly, the formal w (x, y) expresses the numerical relation R i (m, n) is true just when m has relation R to n. Then the proof in outline form goes as follows: 1. Set up a G odel numbering We are nowadays familiar with the idea that all kinds of data can be coded up using numbers. So suppose we set up a sensible (eective) way of coding ws and sequences of ws by natural numbers so-called G odel-numbering. Then, given a formalized theory T , we can dene e.g. the numerical properties W T , Sent , Prf and P rov , where W T (n) i n is the code number of a T -w. Sent T (n) i n is the code number of a T -sentence. Prf T (m, n) i m is the code number of a T -proof of the T -sentence with code number n. P rov T (n) i n is the code number of T -theorem. 2. Expressing such properties/relations inside T We next show that such properties/relations can be expressed inside T by ws of the formal theory belonging to the language of basic arithmetic [takes a bit of work!]. We show in particular how to build an arithmetic w well abbreviate ProvT (x) that expresses the property P rov T , so ProvT (n) is true exactly when P rov T (n), i.e. when n is the code number of T -theorem. 3. The construction: building a G odel sentence Next the really cunning bit, but surprisingly easy we show how to build a G odel sentence GT such that GT is in fact equivalent to ProvT (g), where the numeral g denotes the code-number for GT . In other words, GT is true if and only if GT isnt a theorem. 4. The argument Suppose T GT . Then GT would be provable, and hence GT would be false, so T would have a false theorem and hence not be sound, contrary to hypothesis. So T GT . So GT is true. So GT is false and T , being sound, cant prove it. Hence we also have T GT .

There are big gaps to ll, but thats the strategy. (The proof of Theorem 2 then shows that we can get the same result using the same construction of a G odel sentence by dropping the assumption that T is sound, so long as we require a bit more by way of what the theory T can prove, and require T to have that currently mysterious additional desirable property. More about this in due course) Of course, you might immediately think something a bit worrying about our sketch so far. For basically, Im saying we can construct an arithmetic sentence in T that, via the G odel number coding, says I am not provable in T . But shouldnt we be suspicious about that? After all, we know we get into paradox if we try to play with sentences that say I am not true. So why does the self-reference in the Liar sentence lead to paradox, while the self-reference in G odels proof give us a theorem ? A very good question. I hope that over the coming lectures, the answer to that good question will become clear! Now read IGT, 1.13.4. 6

G odel Without Tears 2

Incompleteness and undecidability

Peter Smith
October 17, 2009 In Episode 1, we introduced the very idea of a negation-incomplete formalized theory T . We noted that if we aim to construct a theory of basic arithmetic, well ideally like the theory to be able to prove all the truths expressible in the language of basic arithmetic, and hence to be negation complete. But G odels First Incompleteness Theorem says, very roughly, that a nice theory T containing enough arithmetic will always be negation incomplete. Now, the Theorem comes in two avours, depending on whether we cash out the idea of being nice enough in terms of (i) the semantic idea of T s being a sound theory, or (ii) the idea of T s being a consistent theory which proves enough arithmetic. And we noted that G odels own proofs, of either avour, go via the idea of numerically coding up inside arithmetic syntactic facts about what can be proved in T , and then constructing an arithmetical sentence that via the coding in eect says I am not provable in T . We ended by noting that, at least at the level of arm-waving description that of Episode 1, the G odelian construction might look a bit worrying. After all, we all know that self-reference is dangerous think Liar Paradox! So is G odels construction entirely legitimate? Later, well see it that it certainly is. But rst I think it might well go a little way towards calming anxieties that some illegitimate trick is being pulled, and is certainly of intrinsic interest, if we give a dierent sort of proof of incompleteness that doesnt go via any worryingly selfreferential construction. So now read on . . .

Negation completeness and decidability

Lets start with another denition (sections, denitions and theorems will be numbered consecutively through these notes, to make cross-reference easier): Defn. 13. A theory T is decidable i the property of being a theorem of T is an eectively decidable property i.e. i there is a mechanical procedure for determining, for any given sentence of T s language, whether T . Its then easy to show: Theorem 3. Any consistent, negation-complete, axiomatized formal theory T is decidable. Proof For convenience, well assume T s proof-system is a Frege/Hilbert axiomatic logic, where proofs are just linear sequences of ws (it will be obvious how to generalize the argument to other kinds of proofs systems, e.g. where proof arrays are trees). Recall, we stipulated that T s formal language L has a nite number of basic symbols (of course thats no real restriction, since given two symbols, e.g. p and , we can construct an innite supply of composite symbols: p, p , p , p , etc.). Now, we can evidently put those basic symbols in some kind of alphabetical order, and then start mechanically listing o all the possible strings of symbols in some kind of order e.g. the one-symbol strings, followed by the nite number of two-symbol strings in dictionary order, followed by the nite number of three-symbol strings in dictionary order, followed by the four-symbol strings, etc., etc. Now, as we go along, generating sequences of symbols, it will be a mechanical matter to decide whether a given string is in fact a sequence of ws. And if it is, it will be a mechanical matter to decide whether the sequence of ws is a T -proof, i.e. check whether each w is either an axiom or follows from earlier ws in the sequence by one of T s rules of inference. (Thats all eectively decidable by Defns 2, 3). If the sequence is a kosher, well-constructed, proof, then list its last w , i.e. the theorem proved. 1

So, we can in this way, start mechanically generating a list of all T -theorems (since any T theorem has a proof, and by churning through all possible strings of symbols, we churn through all possible proofs). And that enables us to decide, of an arbitrary sentence of our consistent, negation-complete T , whether it is indeed a T -theorem. Just start dumbly listing all the T -theorems. Since T is negation complete, eventually either or turns up (and then you can stop!). If turns up, declare it to be a theorem. If turns up, then since T is consistent, we know that is not a theorem. Hence, there is a dumbly mechanical wait and see procedure for deciding whether is a T -theorem. We are, of course, relying here on a relaxed notion of eective decidability-in-principle where we arent working under time constraints (eective doesnt mean practically ecacious or ecient !). We might have to twiddle our thumbs for an immense time before one of or turns up. Still, our wait and see method is guaranteed in this case to produce a result in nite time, in an entirely mechanical way so this counts as an eectively computable procedure in the ocial generous sense (explained more in IGT, 2.2).

Capturing numerical properties in a theory

Heres an equivalent way of rewriting (part of) the earlier Defn. 12: Defn. 14. A property P is expressed by the open w (x) with one free variable in an arithmetical language L i, for every n, i. if n has the property P , then (n) is true, ii. if n does not have the property P , then (n) is true.

Now we want a new companion denition:

Defn. 15. The theory T captures the property P by the open w (x) i, for any n, i. if n has the property P , then T (n), ii. if n does not have the property P , then T (n).

So: what a theory can express depends on the richness of its language; what a theory can capture (mnemonic: case-by-case prove) depends on the richness of its axioms and rules of inferences. Ideally, of course, well want any theory that aims to deal with arithmetic not just to express but to capture lots of arithmetical properties, i.e. to prove which particular numbers have or lack which properties. But what sort of properties do we want to capture? Well, suppose that P is some eectively decidable property of numbers, i.e. one for which there is a mechanical procedure for deciding, given a natural number n, whether n has property P or not (see Defn. 1). Now, when we construct a formal theory of the arithmetic of the natural numbers, we will surely want deductions inside our theory to be able to track, case by case, any mechanical calculation that we can already perform informally. We dont want going formal to diminish our ability to determine whether n has this property P . Formalization aims at regimenting what we can already do: it isnt supposed to hobble our eorts. So while we might have some passing interest in more limited theories, we might naturally aim for a formal theory T which at least (a) is able to frame some open w (x) which expresses the decidable property P , and (b) is such that if n has property P , T (n), and if n does not have property P , T (n). In short, we want T to capture P in the sense of our denition. The working suggestion therefore is that, if P is any eectively decidable property of numbers, we ideally want a competent theory of arithmetic T to be able to capture P . Which motivates the following denition: Defn. 16. A formal theory T including some arithmetic is suciently strong i it captures all decidable numerical properties. And it seems a reasonable and desirable condition on a formal theory of the arithmetic of the natural numbers that it be suciently strong. 2

Suciently strong theories are undecidable

We now prove a lovely theorem: Theorem 4. No consistent, suciently strong, axiomatized formal theory is decidable. Proof We suppose T is a consistent and suciently strong axiomatized theory yet also decidable, and derive a contradiction. If T is suciently strong, it must have a supply of open ws. And by Defn 2, it must in fact be decidable what strings of symbols are open T -ws with the free variable x. And we can use the dodge in the proof of Theorem 3 to start mechanically listing such ws 0 (x), 1 (x), 2 (x), 3 (x), . . . . For we can just churn out all the strings of symbols of T s language, and mechanically select out the ws with free variable x. Now we can introduce the following denition: n has the property D if and only if T n (n).

The supposition that T is a decidable theory entails that D is an eectively decidable property of numbers. Why? Well, given any number n, it will be a mechanical matter to start listing o the open ws until we get to the n-th one, n (x). Then it is a mechanical matter to form the numeral n, substitute it for the variable and prex a negation sign. Now we just apply the supposed mechanical procedure for deciding whether a sentence is a T -theorem to test whether the w n (n) is a theorem. So, on our current assumptions, there is an algorithm for deciding whether n has the property D. Since, by hypothesis, the theory T is suciently strong, it can capture all decidable numerical properties. So it follows, in particular, that D is capturable by some open w. This w must of course eventually occur somewhere in our list of the (x). Lets suppose the d-th w does the trick: that is to say, property D is captured by d (x). It is now entirely routine to get out a contradiction. For, just by denition, to say that d (x) captures D means that for any n, if n has the property D, T d (n), if n doesnt have the property D, T d (n).

So taking in particular the case n = d, we have

But note that our initial denition of the property D implies for the particular case n = d: iii. d has the property D if and only if T d (d).

i. if d has the property D, T d (d), ii. if d doesnt have the property D, T d (d).

From (ii) and (iii), it follows that whether d has property D or not, the w d (d) is a theorem either way. So by (iii) again, d does have property D, hence by (i) the w d (d) must be a theorem too. So a w and its negation are both theorems of T . Therefore T is inconsistent, contradicting our initial assumption that T is consistent. In sum, the supposition that T is a consistent and suciently strong axiomatized formal theory of arithmetic and decidable leads to contradiction. So, if T is properly formalized, consistent and can prove enough arithmetic, then there is no way of mechanically determining whats a T -theorem and what isnt. We could, I suppose, call this result a non-trivialization theorem. We cant trivialize an interesting area of mathematics which contains enough arithmetic by regimenting it into a theory T , and then passing T over to a computer to tells us whats a theorem and what isnt. Its worth remarking on the key construction here. We take a sequence of ws n (x) (for n = 0, 1, 2, . . .) and then considering the (negations of) the ws 0 (0), 1 (1), 2 (2), etc. This sort of thing is called a diagonalizing. Why? 3

Well just imagine the square array you get by writing 0 (0), 0 (1), 0 (2), etc. in the rst row, 1 (0), 1 (1), 1 (2), etc. in the next row, 2 (0), 2 (1), 2 (2) etc. in the next row, and so on. Then the ws of the form n (n) lie down the diagonal! As well see, it is diagonalization (harmless and non-paradoxical) and not any worrying kind of self-reference that is really at the heart of G odels incompleteness proof.

A corollary about the decidability of logic

Defn. 17. A formalized logic is decidable if the property of being a theorem of the logic i.e. a sentence deducible from no premisses is decidable. It is familiar that standard propositional logic is decidable (doing a truth-table test or a tree test decides whats a tautology, and the theorems are all and only the tautologies). It is familiar too that theres no obvious analogue for deciding of an arbitrary sentence whether it is theorem of standard rst-order logic (a.k.a. the predicate calculus). But is there some other decision procedure? Well, Theorem 4 now has an interesting corollary: Theorem 5. If there is a consistent theory with a rst-order logic which is suciently strong and has a nite number of axioms, then rst-order logic is undecidable. Proof Suppose Q is a consistent nitely axiomatized theory with a rst-order logic and which is suciently strong. Since it is nitely axiomatized, we can wrap all its axioms together into one long conjunction, Q. And then, trivially, Q if and only if Q ; i.e. we can prove inside Q if and only if a certain related conditional is logically provable from no assumptions. So if the logic were decidable, and (1) we could mechanically tell whether the conditional Q is a logical theorem, then (2) we could mechanically decide whether is a Q-theorem. But since Q is a consistent suciently strong formalized theory (2) is impossible. So (1) is impossible the logic must be undecidable. Much later, well nd that there is indeed a consistent, nitely axiomatized, weak arithmetic with a rst-order logic, which is suciently strong the so-called Robinson Arithmetic Q ts the bill. So that will settle it: rst-order logic really is undecidable.

Incompleteness again

Theorem 3 says: any consistent, negation-complete, axiomatized formal theory is decidable. Theorem 4 says: no consistent, suciently strong, axiomatized formal theory is decidable. It immediately follows that Theorem 6. A consistent, suciently strong, axiomatized formal theory cannot be negation complete. Wonderful! A seemingly remarkable theorem proved remarkably quickly. But what can we learn from it? Well, note that unlike G odels own result Theorem 6 doesnt actually yield a specic undecidable sentence for a given theory T . And more importantly, it doesnt tell us that T must have an undecidable arithmetic sentence. So suppose we start o with a consistent suciently strong theory T couched in some language which just talks about arithmetic matters: then this theory T is incomplete, and will have arithmetical formally undecidable sentences. But now imagine that we extend T s language (perhaps it now talks about sets of numbers as well as about numbers), and we give it richer axioms, to arrive at an expanded consistent theory U . Now, U will still be suciently strong if T is, and so Theorem 6 will still apply. Note, however, that as far as Theorem 6 is concerned, it could be that U repairs the gaps in T and proves every truth statable in T s language, while the incompleteness has now moved outwards, so to speak, to claims involving U s new vocabulary.

G odels result is a lot stronger: he shows that some incompleteness will always remain even in the theorys arithmetical core. Still, the current theorem is surprising enough. Set down a purely arithmetical theory. Either it wont be suciently strong (will fail to prove some things youd want a formalized arithmetic to prove) or it is incomplete (so still will fail to prove some arithmetic truths). Finally, though, we should stress that the interest of Theorem 6 really depends on the notion of a suciently strong theory dened in terms of the informal notion of a decidable property of numbers being in good order. Well, obviously, I wouldnt have written this Episode if the notion of sucient strength was intrinsically problematic. However, making good that claim by given a sharper account of the notion of decidability takes some doing! And it takes much doing than we need to prove incompleteness by G odels original method. So over the next Episodes, we are going to revert to exploring G odels route to the incompleteness theorems. At this point, you can usefully read Chs 4 and 6 of IGT. You might also skim Ch. 5 but proof details there are perhaps only for real enthusiasts: in fact the arguments are about as tricky as any in the book, so I dont want you to get fazed by them!