This is the fifth part of a series developed through the Department of Philosophical Theology at the Institute of Lutheran Theology’s Christ School of Theology, exploring major results in modern logic and their significance for philosophical and systematic theology.
The Löwenheim–Skolem theorems disclosed something remarkable about first-order logic: a theory may constrain its models quite strongly while nevertheless failing to determine the cardinality of the structures in which its sentences are true. The same theory may possess models of very different infinite sizes, and this fact already suggests that the relation between a formal theory and the structures satisfying it is more complex than a simple one-to-one correspondence between sentences and an intended domain.
The Compactness Theorem reveals a second and closely related feature of first-order logic, but one concerning not the size of models so much as the relation between finite portions of a theory and the theory taken as a whole. Its basic claim is this: if every finite subset T₀ of a first-order theory T has a model, then T itself has a model. Every finite part may be satisfiable in a different structure, and no single finite fragment need display the character of the eventual model of the whole theory; nevertheless, first-order logic guarantees that some structure satisfies all the sentences together.
This result became one of the central instruments of model theory because it permits the existence of structures to be established indirectly. Rather than constructing an infinite or nonstandard model object by object, one may show that every finite collection of the relevant conditions can be satisfied and then invoke Compactness to obtain a model of the entire theory.
Finite Satisfiability and the Whole Theory
Suppose that T is an infinite set of first-order sentences, and let T₀ be any finite subset of T, so that T₀ ⊆ T. There may be indefinitely many such finite fragments, and each may have a model quite different from the models of the others; Compactness does not require a single structure that already works for all finite fragments taken separately.
What it requires is only that each finite fragment be satisfiable. If that condition is met, then the whole theory T is satisfiable, even though infinitely many sentences must now be made true in one and the same structure.
The theorem has an equivalent formulation in terms of logical consequence. If T ⊨ φ, then there is some finite T₀ ⊆ T such that T₀ ⊨ φ. Thus, if a sentence φ follows semantically from an infinite collection of premises, it already follows from some finite portion of that collection.
This is an important point because it means that no particular first-order consequence requires an actually infinite body of premises essentially. An infinite theory may contain infinitely much information, but whenever one sentence is a semantic consequence of the whole theory, finitely many premises already suffice to force that consequence.
Why Completeness Yields Compactness
The connection with Gödel’s completeness theorem is both elegant and instructive. Suppose that T has no model; in that case T is semantically inconsistent, and we may write T ⊨ ⊥, where ⊥ represents contradiction.
Gödel’s completeness theorem tells us that whatever follows semantically in first-order logic is also formally derivable. Hence, if T ⊨ ⊥, then T ⊢ ⊥.
But every formal proof is finite, even when the set of available premises is infinite. A derivation of contradiction from T can therefore employ only finitely many sentences from T, which means that there must be some finite T₀ ⊆ T such that T₀ ⊢ ⊥.
By soundness, T₀ ⊨ ⊥ as well. Consequently, if the whole theory is unsatisfiable, some finite part of it is already unsatisfiable; taking the contrapositive gives the Compactness Theorem.
What first appears to be a theorem about infinite structures thus depends upon a striking interaction between syntax and semantics. The semantic fact that an entire infinite theory has a model is secured through the syntactic fact that any formal proof of contradiction would have to be finite.
An Infinite Model from Finite Requirements
A standard example displays the force of the theorem with unusual clarity. Let T = {σ₁, σ₂, σ₃, …}, where σₙ says that there are at least n distinct objects.
Every finite subset of T has a finite model. If, for example, a particular fragment contains only σ₁ through σ₁₀₀, then a structure containing exactly one hundred objects satisfies every sentence in that fragment.
The same reasoning applies no matter how large the finite fragment becomes. For any finite set of the sentences σ₁, σ₂, σ₃, …, one can choose a sufficiently large finite domain and thereby satisfy all of them together.
Compactness now tells us that the entire theory T has a model. Such a model must satisfy σ₁, σ₂, σ₃, … without end, and hence must contain at least n objects for every natural number n; therefore it cannot be finite.
Nothing in the argument required us to construct that infinite model directly. We established only the satisfiability of every finite portion of the theory, while Compactness guaranteed the existence of a structure satisfying them all at once.
Why Finitude Is Not First-Order Definable
The same pattern of reasoning reveals an important expressive limitation of first-order logic. Suppose there were a first-order sentence F that was true exactly in the finite structures.
Now consider the theory T = {F, σ₁, σ₂, σ₃, …}. Every finite subset of this theory would have a model, because if the largest size requirement appearing in a particular fragment were σ₅₀₀, one could simply choose a finite structure containing exactly five hundred objects; such a structure would satisfy F and all the relevant σₙ.
By Compactness, the entire theory would therefore have a model. Yet any model of the whole theory would have to satisfy F and so be finite, while also satisfying every σₙ and so containing at least n objects for every natural number n.
That is impossible. Hence there can be no first-order sentence whose models are precisely the finite structures.
This does not mean that first-order logic cannot describe particular finite structures. It can do that perfectly well, but it cannot express the general property of finitude in such a way that all and only finite structures satisfy the resulting sentence.
Nonstandard Models of Arithmetic
Compactness also provides one of the simplest routes to nonstandard models of arithmetic. Begin with a first-order theory of the natural numbers, expand its language by adding a new constant symbol c, and then add the sentences 0 < c, 1 < c, 2 < c, 3 < c, … .
Every finite portion of this expanded theory can be satisfied in the ordinary natural numbers. If a given finite fragment extends only through 1000 < c, one may interpret c as 1001 and thereby satisfy all the relevant sentences.
Compactness therefore guarantees a model satisfying the entire expanded theory. In that model, c is greater than 0, greater than 1, greater than 2, and so forth for every standard numeral.
The resulting structure cannot simply be the standard natural numbers, because within the standard natural numbers there is no natural number greater than every standard natural number. The model supplied by Compactness must therefore contain nonstandard elements.
The philosophical importance of this result lies in the fact that a first-order theory may satisfy all the axioms we associate with arithmetic while still having models that differ from the intended structure. Compactness here reinforces the lesson already emerging from Löwenheim–Skolem: first-order theories may determine a great deal without determining everything we may wish to fix about their models.
Theological Consistency and Finite Cores
The theological significance of Compactness begins with consistency. Suppose a theologian formalizes a body of claims concerning God, creation, incarnation, justification, sacramental presence, divine action, or some other doctrinal locus, and suppose the resulting first-order theory T has no model.
Compactness tells us that the problem cannot depend essentially upon the whole infinite or indefinitely extensible collection of assertions. There must be some finite T₀ ⊆ T that is already unsatisfiable.
This matters methodologically because it gives logical analysis a way of localizing inconsistency. Rather than claiming vaguely that an entire theological system is incoherent, one can ask which finite group of assertions cannot all be true together and then examine whether the difficulty lies in the doctrine itself, in the formalization chosen, or in assumptions introduced in moving from ordinary theological discourse into a formal language.
The theorem also yields the converse result. If every finite portion of a first-order theological theory is satisfiable, then the whole theory has a model, and in that sense Compactness gives a strong formal result about consistency.
Yet one must immediately distinguish this result from a much stronger theological conclusion. The fact that a theory has a model does not by itself establish that the theory is true.
Having a Model and Describing Reality
A structure may satisfy every sentence of a formal theory while interpreting its predicates, relations, functions, and objects in ways quite different from those intended by the theologian. If, for example, a theory contains a predicate Gx intended to mean that x is God, then the existence of a model in which some object falls under G shows only that the formal conditions imposed upon G can be satisfied within that structure.
It does not follow from this alone that the object in question is God, that the formal predicate adequately captures what Christian theology means by deity, or that the structure corresponds to divine reality. Model-theoretic satisfaction is a relation between a language and a structure; theological truth requires the further claim that the language, under its intended interpretation, says what is actually the case.
Compactness therefore gives theology something important but limited. It can show that finite satisfiability suffices for satisfiability of the whole first-order theory, and it can help identify the finite core of an inconsistency when no model exists.
What it cannot do is certify that a satisfying model is the intended theological interpretation. That distinction between formal satisfiability and theological truth becomes increasingly important as one moves from proof theory into model theory.
Compactness and the Limits of First-Order Description
Compactness reveals something fundamental about the character of first-order description. An infinite collection of sentences may impose indefinitely many conditions upon a structure, but if every finite combination of those conditions is satisfiable, then first-order logic guarantees a model satisfying them all.
This makes first-order logic exceptionally powerful as an instrument for establishing existence. At the same time, the theorem shows why certain features cannot be forced by first-order description alone: finitude is one example, and standardness in arithmetic is another.
Taken together with Löwenheim–Skolem, Compactness thus exposes a characteristic feature of first-order theories. They may constrain their models with enormous precision and still admit structures significantly different from the one the theorist initially has in mind.
None of this entails skepticism about mathematics, theology, or reference. It entails only that syntax by itself does not determine intended interpretation and that formal satisfaction should not be confused with truth about the reality under discussion.
For theology this is an important discipline because formalization can clarify consequences, identify contradictions, display structural possibilities, and distinguish assumptions that ordinary prose may leave entangled. Yet the existence of a satisfying structure remains a logical result about a theory and a model, not by itself a theological account of what makes the theory true.
Compactness therefore belongs naturally beside Gödel completeness and Löwenheim–Skolem as one of the central results defining both the power and the limits of first-order logic. Gödel showed that first-order semantic consequence can be captured by formal proof; Löwenheim and Skolem showed that first-order theories with infinite models generally cannot control the cardinality of those models; Compactness now shows that the satisfiability of an entire infinite theory is determined by the satisfiability of its finite fragments.
Together these results disclose a remarkable logical situation. First-order logic is strong enough to sustain rigorous reasoning about indefinitely complex structures while remaining too weak to determine, through its sentences alone, every feature of the structures we may intend.
For theology, the conclusion is not that formal logic reaches too little to be useful, but that its usefulness depends upon knowing exactly what has been established. Logic can tell us what follows from our formulations and whether those formulations can be jointly satisfied; theology must still ask whether the formulations say truly what is the case.
Bibliographical Note
The Compactness Theorem is closely connected with Gödel’s completeness theorem and became one of the fundamental instruments of twentieth-century model theory. It is commonly presented either as a consequence of completeness or by model-theoretic methods in its own right, and its applications to nonstandard models, non-definability results, and the existence of structures satisfying infinitely many conditions became central to the subsequent development of the field.
Standard treatments include Herbert Enderton, A Mathematical Introduction to Logic; George Boolos, John Burgess, and Richard Jeffrey, Computability and Logic; Wilfrid Hodges, A Shorter Model Theory; and C. C. Chang and H. Jerome Keisler, Model Theory.
No comments:
Post a Comment