PKL Rev11

Where the Axioms Come From

Gödel and Günther: a meaning that is perceived, and a will between two interpretations

Between 1954 and 1959 Kurt Gödel and Gotthard Günther wrote to each other about logic. Gödel had shown that no formal system of arithmetic that proves nothing false proves everything true. Günther was looking for a logic with more than two values that was to grasp thinking about thinking. It turned into neither a dispute nor an agreement, but into a conversation at whose centre stands a question.

Gödel asks where the axioms come from, and answers: from the meaning of the primitive terms, which we perceive, not create; he does not call this proved. Günther asks how a decision is made between two interpretations of equal standing (ebenbürtig), and answers: not by a theory, but by an act of will.

Behind this stands a question that is older than both and undecided to this day: Is mathematics discovered or invented? Whoever holds it to be discovered must say where its objects are and how we know of them. Whoever holds it to be invented must explain why it does not bend as one wishes, and why it fits nature. Mathematicians and physicists of rank stand on both sides, and neither side has convinced the other. Gödel belongs to the discoverers, and he knows that he cannot prove it. Günther's question runs across both camps: it concerns not mathematics but the logic with which one does it, and the place at which it does so is what this is about.

The letters are read alongside Gödel's lecture of 1951 and three texts of Günther's, of 1957, 1978 and 1979. The line from the letters to the late texts Günther does not draw himself; we draw it, and where two passages stand side by side, the sentence says who puts them together. What Günther reads into the oldest propositions of logic is developed in Appendix A; whoever does not know him yet does best to begin there. The letters are written in German; the translations from them are ours.

A correspondence

The occasion was not a problem of logic but an application. In 1953 an American relief committee for émigré scholars wanted to support Günther's application to the Bollingen Foundation and asked Gödel; Günther wrote to him on 2 August 1953. So Charles Parsons reports it in his introduction (CW IV p. 458). On 17 September 1956 the two met in Princeton (Parsons, ibid.; the place is named by Gödel in letter 10). Gödel's last letter is dated 7 January 1959.

Reflection, in stages

Günther's first question concerns the law of the excluded middle. One of its two formulations he reads as "the formulation of the law of the excluded middle for an infinite hierarchy of successive possible points of view of determination" (our translation) (letter 1, 29.4.1954, CW IV p. 480). Gödel agrees conditionally, and in the fourth letter he says what he accepts and what he does not:

The reflection on the subject treated in idealist. phil. (i.e. your II theme of thinking), the distinction of stages of reflection etc. seems to me very interesting & important. I even consider it quite possible that this is "the" way to the right metaphysics. The rejection of the objective meaning of thinking connected with it (but in truth entirely independent of it), however, I cannot go along with.

(our translation; "d." written out as "of", "u." rendered "&") (letter 4, 30.6.1954, CW IV p. 502)

Gödel thus accepts the stages of reflection. What is disputed is what follows from them for the objectivity of what is thought. Abstract entities, he writes, "seem rather to form a second level of reality, which however confronts us just as objectively. & independently of our thinking as nature" (our translation) (p. 504; the full stop after "objectively" renders the one after "objektiv" in the print).

Both use the word metaphysics without shyness. Günther wants to use the results of mathematical logic, then called logistics, "for a new metaphysics" (our translation) (letter 3, 23.5.1954, CW IV p. 486). What he means by this, and why for him it is no special case of poetry, is in Appendix A.

Gödel 1951: The meaning of the primitive terms is perceived

Three years earlier Gödel had presented his own answer, in a lecture that remained unprinted in his lifetime (Gödel 1951; pages after the Collected Works, vol. III). Whoever axiomatizes mathematics finds no finite list for it. The axioms "are not arbitrary, but must be correct mathematical propositions, and moreover, evident without proof" (p. 305), and their sequence does not end: "nor can there ever be an end to this procedure of forming the axioms, because the very formulation of the axioms up to a certain stage gives rise to the next axiom" (p. 307).

From what does such an axiom follow? Not from identity. The axioms are "by no means tautologies (in the sense of being in any way reducible to a = a), but still do follow from the meaning of the primitive terms under consideration" (p. 321). And this meaning, according to Gödel, is not made:

What is wrong, however, is that the meaning of the terms (that is, the concepts they denote) is asserted to be something man-made and consisting merely in semantical conventions. The truth, I believe, is that these concepts form an objective reality of their own, which we cannot create or change, but only perceive and describe.

(p. 320)

That is Gödel's answer, and it carries its reservation in the sentence: "I believe".

Against the view that mathematics is our creation he has reasons, not proofs. If there were absolutely undecidable mathematical propositions, he writes, that "seems to disprove the view that mathematics is only our own creation; for the creator necessarily knows all properties of his creatures" (p. 311; the alternative from which he infers this is treated in Appendix B). And: "If anything like creation exists at all in mathematics, then what any theorem does is exactly to restrict the freedom of creation. That, however, which restricts it must evidently exist independently of the creation." (p. 314) Whether undecidable propositions could be decided "arbitrarily by a new act of creation", he means to treat in a footnote whose number is missing. The editors note: "No footnote in the manuscript deals with this question." (p. 314, note b) The reference leads nowhere.

In the end Gödel calls only one view refuted, the nominalistic one. According to it the mathematical propositions express solely syntactical conventions and are "void of content" (p. 316). His objection: to show that the axioms are tautologies in a suitable language, one must already presuppose them, "the truth of the axioms must first be assumed and then it can be shown that, in a suitably chosen language, they are tautologies" (p. 317). He claims no more:

Of course I do not claim that the foregoing considerations amount to a real proof of this view about the nature of mathematics. The most I could assert would be to have disproved the nominalistic view, which considers mathematics to consist solely in syntactical conventions and their consequences. Moreover, I have adduced some strong arguments against the more general view that mathematics is our own creation.

(pp. 321 f.)

Does thinking change its object?

In the letters of 1954 and 1955 the point of dispute lies open. Günther distinguishes two identities, an irreflexive one and one of reflection; there are "two in principle different identities of the object with itself" (our translation) (letter 5, 2.10.1954, CW IV p. 508; the print has the misprint "vorschiedene" for "verschiedene"). A year later he determines the second. Whoever reflects on reflecting itself finds no object at hand:

The same must therefore first be made an object. I.e., it must be "posited" as a logical object. By thinking reflection itself, I turn it from a subject-like process into an object-like state of affairs. Result: identity of reflection means that the logical object, by being thought, changes in its logical properties.

(our translation) (letter 6, 19.6.1955, CW IV p. 514; "the same" is the subjective act of thought, the process of reflection; "changes" underlined in the print, here italic)

There are thus two types of objects, one "which, by being thought, suffers no change", and one "in which such a change must be reckoned with" (our translations) (p. 516).

Gödel answers with the case of intuitionistic logic, which he calls a reflection on thinking. Its objects are

not independent of thinking, but nevertheless they are changed "by the act of thought that occupies itself with them" just as little as the objects of physics.

(our translation) (letter 7, 10.8.1955, CW IV p. 518)

Gödel separates what Günther joins: depending on thinking, and being changed by thinking. The first he grants, the second he denies. As a picture from software, and no more than a picture: for Gödel thinking is here a read access without side effect; Günther demands that the side effect be reckoned with here.

And Gödel names what would convince him. That a many-valued logic is the adequate means of expression for Günther's difference could only be proved "by an axiomatization of logic on this basis & the success of the consequences developed from the axioms" (our translation) (p. 518). Günther would have to start from an "analysis of the sense of your truth values"; he calls this the "cardinal point" (our translations) (p. 520).

Günther 1957: Two interpretations, one decision

In "Metaphysik, Logik und die Theorie der Reflexion" Günther first reports Fichte: it is left to consciousness how it wants to interpret its relation to the world. "That is the Fichtean 'deed' or primordial action of the I." (our translation) (p. 14) Then he applies this himself:

If we now apply this way of looking at things to Aristotelian logic, we must state that classical consciousness realizes a fundamental metaphysical decision in the design of a logic that posits the relation between the I and being as immediate identity. […] But this relation "is" not identity! Consciousness can, however, decide to interpret this relation as identity.

(our translation) (p. 14)

The other interpretation is equally open: to understand itself as opposed to being. Aristotle erred when he believed that the situation of the "I think something" was unambiguous, "which is why one system of axioms could definitively fix the lawful character of this relation" (our translation) (p. 14). And both interpretations weigh the same:

Both possibilities of interpretation are formally absolutely of equal standing and of equal rank, and logic has to take both into account axiomatically.

(our translation) (p. 15)

Günther calls the decision metaphysical, and the word has a narrow sense here: a presupposition about how I and being stand to each other. The old doctrine of the world as a whole, the soul and God is not meant. A logic makes such a presupposition, and it cannot prove it, for there is a second one that is formally of equal standing with it. So we summarize his argument.

Two interpretations require two axiomatics. Günther sets four further principles against the four principles of classical logic, under the heading "Axiomatics of total rationality" (our translation) (p. 32); Appendix A reproduces the table and the way to it.

The way leads through stages of reflection. Reflection can be repeated: "Since, then, every R-stage can in its turn become the object of a further reflection, this series of iterations is infinite and can lead to no conclusion within itself." On it "an axiomatics and a finite formalism […] are utterly impracticable" (our translations) (p. 20). Only the whole of these repetitions has a law: the self that reaches the third stage "is no longer iterable" (our translation) (p. 23).

Both, then, are dealing with a sequence without end. Gödel asks for the next axiom; the axioms up to a stage, he writes, "give rise to" the next. Günther asks for the concluding one; on the repetitions themselves an axiomatics is "impracticable". The juxtaposition is ours.

Gödel's expectation: with necessity

In 1957 Günther indicates that his stages stand in relation to Russell's types (p. 4, n. 1). Gödel takes this up. He remembers an older manuscript of Günther's, and in 1959 he writes the sentence about it that contains his expectation:

You identified at that time iterated reflection with the theory of types and total reflection with a typeless (i.e. combining all types into one) logic. One should think that from philosophical insights into the nature of reflection the right axioms of a typeless logic would have to result with necessity, which would be a tremendous advance over the procedure of "trial and error" applied today.

(our translation) (letter 14, 7.1.1959, CW IV p. 534; the print writes "muessten", "gegenueber", "waere" without umlauts)

Gödel expects that the axioms of a new logic would have to result from insight with necessity. Günther had said of the old one in 1957 that its basic relation "is" not identity; consciousness could decide for it. For Gödel the axioms come from insight, for Günther from a decision. That the two sentences stand side by side is our arrangement; Günther's sentence is the older one and not written as an answer.

Günther 1978: The filling is given a place in the calculus

"Idee, Zeit und Materie" takes up the two interpretations again and gives them their sharpest form. Classical negation is an exchange relation (Appendix A), and one can look at it in two ways. As "decided": then one starts from what is. Or as "undecided": then one starts from the relator, which "only indicates a state of suspension" (our translations) (print p. 237). Thinking has the freedom "to understand its theme, in the 'reflection-into-other', as an already decided state of affairs (Eleatic being) or, in the 'reflection-into-self', as a decision not yet made (Heraclitean process)" (our translation) (print p. 242). These are the two themes: being, about which it has been decided, and the happening in which the decision is still pending. Of the two starting situations Günther writes:

Both posits appear to us as of equal standing with each other, and we can discover no theoretical ground to prefer the one to the other. In every moment of our waking consciousness, however, we must decide for the one or the other starting situation.

(our translation; "posits" for "Setzungen") (print p. 237)

The word "ebenbürtig" ("of equal standing") already stood in 1957. Now Günther says who decides. Where to every concept "there adheres a primordial ambiguity that is theoretically irresolvable", "a will must decide that no longer asks for grounds and would also ask in vain – because there are none left" (our translations) (print p. 246). That thinking was first developed as thinking of objects, he calls, "seen from the will",

a free, but moreover intelligent choice between two semantic interpretations, of equal standing, of one and the same formal logic. We say: the choice was intelligent! For, seen from thinking, this was a practical-technical necessity.

(our translation) (print pp. 246 f.)

The choice has a consequence in history: with it "the symmetrical exchange relation of the two themes shifts into a historical relation of rank" (our translation) (print p. 247). And this consequence is to become visible in the calculus. The theory of the second theme must be changed in such a way "that in its calculus its historical fate comes to expression in a manner appropriate to a calculus" (our translation) (ibid.). For this Günther separates the logical value from the location at which it can occur. Whoever programs has a picture for this, and only a picture: a variable that exists before a value has been assigned to it.

Tafel III A B C D E 1 1 2 1 2 3 α β γ δ ε

We separate the logical value from the ontological location at which it may possibly occur – or not. If it then actually occurs at the said location, that is a pure factum brutum, which must be interpreted not as logical necessity but only as a "posit", i.e. as decision or action. (our translation)

Gotthard Günther, Idee, Zeit und Materie, in: Beiträge zur Grundlegung einer operationsfähigen Dialektik, vol. 3, Hamburg 1980: the sentence p. 247, the table p. 248.

Günther reads the table himself. In column B there is "only the simple alternative between filling and non-filling of the empty place" (our translation) (print p. 248). In column C the places could remain empty, be filled as shown, "and finally the values can be interchanged" (our translation). In column D, besides the inserted value sequence, "a permutation sequence of five further value-fillings (Wertbesetzungen) is still possible, which already permits six hermeneutic variants of the concept of being. Systematically, the mutual relation of these interpretations can only be analysed heterarchically (heterarchisch)." (our translation) (print p. 249) To every value-filling there corresponds, in Günther, an interpretation; for column D he says so explicitly. Consciousness faces the decision "whether at all and, if so, with which value a given empty place is to be filled" (our translation) (print p. 248).

Here his answer is not the old one. That consciousness determines its relation to the world by a deed, he himself cites in 1957 as Fichte's doctrine. What is new is not that he calls the choice an act of will, but that he gives the filling a place in the calculus, the whether and the value: he separates the value from the location, and that a value occurs at a location is then to be understood only as a "posit". In the calculus as Günther draws it here there stands the space of value-fillings, in which the values can be interchanged; the rank between the interpretations is what the changed calculus is first to bring to expression. So we read the table.

Günther does not thereby claim a beginning out of nothing. Of the "problem of creation (Schöpfung) out of nothing" he writes: "Its source is a mythological fiction. If an action is to arise, something must always already be there at which the process of action sets itself going." (our translations) (print p. 245) Gödel, in his own English word, had objected against "creation" that what restricts it must exist independently of it. In both, the doing presupposes something that it does not itself bring forth. Gödel concludes from this that the concepts are perceived, Günther that action starts from what is given. The juxtaposition is ours.

What of this can be computed

The edition Rev11 has a theorem about the space of fillings: a pattern can be filled with two values exactly when its signs do not go beyond one (proved [38]); «At four places there are fifteen distinct patterns; eight of them can be filled with two values — and both numbers are proved [39], [40] and not looked up.» (edition, chapter 1) We assign this space to the place that Günther gives to the filling; the edition itself does not do so.

Over every pattern of length four that can be filled with two values lie exactly two truth tables ([41]): «the first place chooses, everything else follows from the pattern.» The two tables «come apart by a global exchange of values — they agree at no place whatever.» (edition, chapter 1) Beside this we place, for patterns with two classes, Günther's inserted value sequence and the interchanged one. Which of the two occurs, the edition does not derive, nor that one occurs at all: «… solely the that of enactment is no structural feature.» (edition, chapter 7)

Gödel's cardinal point, an "analysis of the sense" of the truth values, it does not fulfil: it carries no semantics of the values (edition, C.4). Of the step from m to m + 1 values it says: «each step adds exactly m new elementary contextures when going from m to m+1 values [70], and the image of the old stage does not reach the new element [71]. What is added was not there before.» (edition, chapter 6)

Charles Parsons writes in his introduction that Günther seems never to have tried to write down axioms and rules of inference for his many-valued calculi (CW IV p. 473). The edition has rules, Kaehr's of 1981 ([119]), and it proves where they hold and where they do not ([164] to [166]; edition, O4).

Neither Gödel nor Carnap

Neither the choice of an interpretation nor the filling of an empty place makes a proposition true in Günther: the interpretations are of equal standing, and the calculus has the empty place before it is filled. Günther thus stands, as we read him, neither with Gödel nor with the view that Gödel calls refuted. For this view Gödel refers to Hans Hahn and Rudolf Carnap (p. 315, n. 23).

On choice, Günther stands close to Carnap. He quotes him himself: also "the interpretation of logical signs (e.g., by a determination of truth conditions) can be freely Chosen" (Idee und Grundriss, p. 34; Carnap in Günther's quotation, "Chosen" so in the edition). And in Günther in 1978:

Contemplative reflection cannot do otherwise than waver, undecided, back and forth between two world pictures that are theoretically of equal standing – and therefore interchangeable. Action, however, decides; and it does so quite pragmatically.

(our translation) (print p. 246)

For both, no theory decides. The boundary runs at two other places.

The first is the presupposition. Logical positivism is for Günther himself one interpretation of the calculus among several, and one that "nolens volens presupposes an unreflected (physicalist) world view. Reality is here equated with thinghood." (our translation) (p. 35) We read this so: whoever takes the interpretation for a convention has already chosen one and does not know it.

The second is the consequence. The freedom of convention of the calculi Günther places beside the world views that "condemn" a logic of the humanities "to a loose systematics of mutually interchangeable systems" (our translations) (p. 34). In his own case it does not stop at interchangeability: the decision turns the symmetry into a relation of rank that has a history, and he demands that the calculus show this history. The line against Carnap is drawn by us, not by Günther.

He does not hold positivism to be an error: "The positivists are not to be blamed. Their attitude is the necessary result of a philosophy gone bankrupt." (our translation) (p. 6) He sees in it a consistent step of the modern age: "In logical positivism truth becomes convention" (our translation) (p. 48). And of mathematics, which today is built up postulationally: "Truth is no longer received, it is produced by definition." (our translation)

Action or event

In the text of 1978 stands the question towards which everything runs. Marx demands that action should replace contemplation. This demand, Günther writes,

forces, in unexpected paradox, a second theoretical-contemplative attitude of consciousness, troubled by the question of what action, as opposed to mere event, is actually supposed to be.

(our translation) (print p. 243)

To whoever thinks from being, "the problem of time appears only as a disturbing factor", where the task is "to explore the 'eternal' laws of the real" (our translations) (print p. 244); this is the subject of the sheet on Physics.

Günther connects the question with the two interpretations. Two-valued logic permits, "with exactly the same structural means, semantically two interpretations […], of which only one is laid claim to in the concept of event. On this one-sided laying of claim rests our entire Western, philosophical tradition and metaphysics." If one turns to the second, "which tends not towards what is already decided (being) but towards the undecided, a concept of action results quite naturally" (our translations) (print p. 246). The event, then, is for Günther not how the world is, but how one of the two interpretations reads it.

He gives the concept: "We are now dealing, then, with an event steered by a will conscious of its goal. Not only does something happen, but something is done." (our translation) (print p. 246) And he says what the will can do: its intentions could, "since for the time being nothing other than a symmetrical exchange relation is there at which its activity can kindle, in principle only consist in making a relation of rank out of the symmetry of the world interpretations" (our translation) (ibid.).

Günther 1979: A criterion

In "Identität, Gegenidentität und Negativsprache" (1979) Günther says why no ground is to be had for action. He says it with Schelling:

Schelling does grant that thinking must have a ground, and that is being. But action may not be traced back to any ground, otherwise there would be no freedom. Instead it emerges from an un-ground (Ungrund), as we have set out further above.

(our translation) (p. 53)

With "as we have set out further above" Günther makes Schelling's sentence his own. Then he names a criterion, and it hangs on one word: exchange. That is what he calls the relation of two sides that can stand in for each other, like the two values under negation; Appendix A leads up to it.

Here we come upon the only criterion for the difference between event, namely that something happens, and volitional action, namely that something is done. If the primordially symmetrical exchange relation of two world-points that function as possible starting places of the perception of the world (centres of subjectivity) is disturbed in such a way that all properties of the exchange are lost, we would speak of pure, or absolute, objectivity. Conversely, if we encountered in the world, primordially considered, only completed exchange relations, we would have to speak of an absolute subject.

(our translation) (p. 56)

Both poles he calls in the next paragraph "only regulative principles of thinking and acting": "For the empirical world no constitutive significance belongs to them." (our translations) (p. 56) What there is lies, so we read it, between them: more or less disturbed exchange. That the difference between event and action is thereby a matter of degree and not an either-or, we read from the context; it does not stand in the sentence of the criterion.

The words are those of 1978: happens, done. That the criterion answers the question of 1978 is our assignment; Günther does not refer back. The measure is now a different one: in 1978 a will decided between two world interpretations, in 1979 the exchange relation "of two world-points" stands in the sentence, of two centres of subjectivity. The two concepts he takes from Kant; that the poles are only regulative is self-evident "for every thinker who has gone through the school of Kant" (our translation) (p. 56). Only the distinction of regulative and constitutive, he adds, makes it possible "to speak of action – and with it subjectivity – being technically constructible in progressive stages" (our translation) (p. 56). A page later he applies Kant's distinction to the circles of which the edition treats: "The Hamiltonian full cycles undoubtedly correspond to what in Kantian philosophy is called a regulative concept." (our translation) (p. 57; on this edition, C.8)

And the conclusion of 1957? Then the self of the third stage of reflection was "no longer iterable". Now Günther writes: "Negativity, once it has left the limits of the triadic behind, is in principle unfathomable." (our translation) (p. 57) That he thereby takes back the axiom of 1957, he does not say.

Back to the question

Gödel saw the concepts as a reality that we perceive and describe, and did not himself hold this to be proved. Günther saw two interpretations of equal standing and a will that decides between them and "no longer asks for grounds". In the calculus he gave the filling a place: whether and with which value an empty place is filled. The rank of the interpretations his calculus does not show; it is yet to show it.

What is computed of this is the space of fillings. Which filling occurs follows from no theorem of the edition. In Günther, so we read him, this is no defect: it is the place of the will.

Günther gave a concept of action and named a criterion. A construction he does not present; his word is "constructible". Open is what he asked: "what action, as opposed to mere event, is actually supposed to be".


On citation. The letters are cited after volume IV of Gödel's Collected Works, with number, date and page; they are written in German, and the translations are ours. Gödel's lecture in English, as it has come down to us. Günther is cited after the editions of vordenker.de, which have converted the spelling, "Idee, Zeit und Materie" after the print; his German is given in our translation. Emphases in quotations are those of the source. About the edition Rev11 (anchor 24fa57a, 31 October 2026) this sheet claims nothing that it does not say itself.


Sources

Correspondence: Kurt Gödel, Collected Works, vol. IV: Correspondence A–G (Oxford 2003), Gotthard Günther, pp. 476–535 (quoted: letters 1, 3, 4, 5, 6, 7, 10, 14; German); introduction by Charles Parsons, pp. 457–476.

Gödel 1951: Some basic theorems on the foundations of mathematics and their implications, Gibbs Lecture, delivered on 26 December 1951 in Providence; Collected Works, vol. III: Unpublished Essays and Lectures (Oxford University Press, New York 1995), pp. 304–323. Quoted with the page of this volume.

Günther: Metaphysik, Logik und die Theorie der Reflexion, Archiv für Philosophie 7 (1957), pp. 1–44; quoted after the vordenker.de edition (pages of that edition). · Idee, Zeit und Materie, Hegel-Jahrbuch 1976, pp. 168 ff.; reprinted in: Beiträge zur Grundlegung einer operationsfähigen Dialektik, vol. 3, Hamburg 1980, pp. 236–254; quoted after the print of the Beiträge, pp. 236–249. · Identität, Gegenidentität und Negativsprache, lecture, Belgrade 1979; Hegel-Jahrbuch 1979, pp. 22–88; quoted after the vordenker.de edition (pages of that edition). · Idee und Grundriss einer nicht-Aristotelischen Logik (1959); quoted after the vordenker.de edition (2004; pages of that edition).

Edition Rev11: chapters 1, 6, 7; C.4, C.8; O4.