Die Gründe, die C. I. Lewis [5], [6] bewogen haben, neben der gewöhnlichen Implikation eine strikte Implikation einzuführen, sind bekannt. In der vorliegenden Arbeit wird aus ähnlichen Gründen eine strenge Implikation eingeführt, die jedoch einen engeren Begriff darstellt als die strikte Implikation. Mit einer Arbeit von Arnold Schmidt [7] hat meine nur geringe Berührungspunkte, da der Verfasser sich mit der strikten Implikation beschäftigt. Für diese wird ein relativ einfaches Axiomensystem angegeben und gezeigt, wie man durch geeignete Definitionen von Notwendigkeit und Möglichkeit bei eventueller Hinzufügung von Zusatzaxiomen die bekannten Systeme des Modalitätenkalküls erhält. Den Hauptteil meiner Arbeit bildet die Einführung meines Implikationsbegriffs; im zweiten Teil der Arbeit komme ich dann auf dieser Basis auch zu einer Einführung der Modalitäten.

Um die Art der Einführung des Kalküls zu motivieren, wollen wir kurz auf das eingehen, das uns mehr oder weniger deutlich vorschwebt, wenn wir von der strengen Implikation sprechen. Die strenge Implikation, die wir durch A → B wiedergeben, soll ausdrücken, daß zwischen A und B ein logischer Zusammenhang besteht, daß der Inhalt von B ein Teil des Inhaltes von A ist, oder wie wir es sonst ausdrücken wollen. Das hat mit der Richtigkeit oder Falschheit von A und B nichts zu tun. So würde man die Allgemeingültigkeit einer Formel A → (B → A) ablehnen, da sie den Schluß von A auf B → A einschließt und da die Richtigkeit von A nichts damit zu tun hat, ob zwischen B und A ein logischer Zusammenhang besteht.

The following remarks apply to many functional calculi, each of which can be variously axiomatized, but for clarity of exposition we shall confine our attention to one particular system Σ. This system is to have the usual primitive symbols and formation rules of the pure first-order functional calculus, and the following formal axiom schemata and formal rules of inference.

Axiom schema 1. Any tautologous wff (well-formed formula).

Axiom schema 2. (a) A ⊃ B, where A is any wff, a and b are any individual variables, and B arises from A by replacing all free occurrences of a by free occurrences of b.

Axiom schema 3. (a)(A ⊃ B)⊃(A⊃ (a)B). where A and B are any wffs, and a is any individual variable not free in A.

Rule of Modus Ponens: applies to wffs A and A ⊃ B, and yields B.

Rule of Generalization: applies to a wff A and yields (a)A, where a is any individual variable.

A formal proof in Σ is a finite column of wffs each of whose lines is a formal axiom or arises from two preceding lines by the Rule of Modus Ponens or arises from a single preceding line by the Rule of Generalization. A formal theorem of Σ is a wff which occurs as the last line of some formal proof.

The main purpose of this note is to prove (theorem 11, § 5) that, in any interpretation of the formalisation of Aristotelian syllogistic given by Łukasiewicz [4], it is always possible to associate with each element a a non-null sub-class φ(a) of some ‘universal’ class V in such a way that ‘Aab’ (all a are b), ‘Iab’ (some a are b) are equivalent respectively to ‘φ(a) is contained in φ(b)’, ‘φ(a) has a non-null intersection with φ(b)’. Similarly (theorem 6, §4) we show that in Wedberg's system [14] with primitives ‘Aab’, ‘a’ (not a), it is possible to find a mapping a → φ(a) as above such that ‘Aab’ is equivalent to ‘φ(a) is contained in φ(b)’ and φ(a‘ is equal to φ(a)’, the complement of φ(a) with respect to V. Thus, if we make the preliminary step of identifying elements a, b such that Aab and Aba both hold (i.e. taking equivalence classes with respect to the relation Aab & Aba), we are left with essentially only one kind of interpretation for these systems, namely the ‘normal’ interpretation by classes. Slupecki [11], [12] has proved that Łukasiewicz's system is a complete and decidable theory of the relations of inclusion and intersection of non-null classes, and Wedberg [14] has proved that his system is a complete and decidable theory of the relation of inclusion and the operation of complementation for nonnull, non-universal classes. Using the above-mentioned embedding theorem, we are able to obtain (theorems 9, 6, §§ 5, 4) very simple proofs of these results.

Commonly, when we succeed in showing a formula of quantification theory to be consistent, we do so by producing a true interpretation. Sometimes we achieve the same effect without even exceeding the resources of quantification theory itself: we show a formula consistent by producing a valid formula from it by substitution. Example: ‘(∃x)Fx ⊃ (x)(∃y) (Gxy ▪ Fy)’ is seen consistent by noting its valid substitution case ‘(∃x)Fx ⊃ (x)(∃y)(Fx ∨ Fy ▪ Fy)’. How generally available is this latter method? I shall show that it is available if and only if the formula whose consistency is shown is satisfiable in a one-member universe.

The “only if” part is immediate. For, if ψ is a substitution case of ϕ, then ϕ is satisfiable wherever ψ is; and ψ, if valid, is satisfiable in a one-member universe.

Conversely, suppose a true interpretation of ϕ in a one-member universe. Each predicate letter of ϕ is thereby interpreted outright as true or false, in effect, since there is no varying the values of ‘x’, ‘y’, etc. Now form ψ from ϕ by substitution as follows: change each atomic formula ϕi(e.g. ‘Fx’, ‘Fy’, ‘Gxy’) to ⌜ϕi⊃ϕi⌝ if its predicate letter is one that was interpreted as true, or to ⌜ϕi ▪ − ⊃ϕi⌝ if its predicate letter is one that was interpreted as false. Clearly ψ under all interpretations even in large universes, will receive the truth value that ϕ received under the supposed interpretation in the one-member universe. So ψ is valid.

I will give precise versions of the foregoing verbal definitions, by formalizing them within a formal metalanguage.

In order to avoid the necessity of formulating a meta-language, I will make use of VI 112, (i.e., Gödel's The consistency of the continuum hypothesis). Actually, the system there developed is unnecessarily strong for the limited purpose of this illustration. It will suffice to choose for the system of VI 112 without axioms C4 and E (axioms of replacement and choice); this system will be the logical basis of , and to this I will add some set-constants and some meaning postulates. The logical basis is roughly as strong as Zermelo's system with an axiom of infinity.

Within I will set up analogues of the previous definitions, applicable to a wide variety of object-languages. For the sake of concreteness it will be convenient to think of as having Zermelo set theory, without the axiom of infinity, as its logical basis – but it must be kept in mind that this is not an essential feature of the forthcoming definitions.

Our is certainly strong enough to have all primitive recursive functions calculable within it. This supplies us with the needed tool for talking about , within , by introducing a number of primitive recursive functions, properties, and relations describing the structure of . The exact definitions of these will depend on the exact form of , but for the forthcoming definitions it will suffice to know that these functions, properties, and relations are representable in for the given . This makes the resulting definitions most useful, since they can be applied to any of the type described.

A set S of natural numbers is called recursively enumerable if there is a general recursive function F(x, y) such that

In other words, S is the projection of a two-dimensional general recursive set. Actually, it is no restriction on S to assume that F(x, y) is primitive recursive. If S is not empty, it is the range of the primitive recursive function

where a is a fixed element of S. Using pairing functions, we see that any non-empty recursively enumerable set is also the range of a primitive recursive function of one variable.

We use throughout the logical symbols ⋀ (and), ⋁ (or), → (if…then…), ↔ (if and only if), ∧ (for every), and ∨(there exists); negation does not occur explicitly. The variables range over the natural numbers, except as otherwise noted.

Martin Davis has shown that every recursively enumerable set S of natural numbers can be represented in the form

where P(y, b, w, x1 …, xλ) is a polynomial with integer coefficients. (Notice that this would not be correct if we replaced ≤ by <, since the right side of the equivalence would always be satisfied by b = 0.) Conversely, every set S represented by a formula of the above form is recursively enumerable. A basic unsolved problem is whether S can be defined using only existential quantifiers.