Vaught’s conjecture

A chain of three positions, with the first before the second and the second before the third, stays the same arrangement if its positions are renamed A, B, C or 7, 8, 9. Renaming is not a new structure. This finite example explains how counting works; the conjecture concerns infinite structures.

Can one complete rulebook have more genuinely different countably infinite structures than can be listed one by one, yet fewer than there are real numbers? The conjecture says this intermediate number cannot occur. The rulebook uses a countable collection of symbols and ordinary first-order statements about individual elements.[1]

Can a complete rulebook describe an intermediate-sized family of genuinely different countably infinite structures: too many to list, but fewer than the real numbers? A three-position ordered chain retains its structure when its positions are consistently renamed. This finite illustration teaches the identification used in the count; it does not give different models of one complete theory.

A model is a structure satisfying a theory, including its specified relations and operations. An isomorphism is a bijection preserving all of those, so changing element names does not increase the count. A countable language has at most a list’s worth of symbols; first-order quantifiers range over individual elements. A consistent complete theory decides each sentence but can still have nonisomorphic models.

Let $N(T)$ count isomorphism types of countably infinite models of such a theory. The proposed alternatives are $N(T)\le\aleph_0$ or $N(T)=2^{\aleph_0}$. Here $\aleph_0$ is the size of the natural numbers and $2^{\aleph_0}$ is the size of all infinite binary sequences, also called the continuum. The claim excludes intermediate model counts without assuming a general exclusion of intermediate sizes for all sets.[1]

A known theorem leaves at most one possible intermediate size: the smallest size of an uncountable set. That size need not be smaller than the set of real numbers; if the sizes coincide, it is already the allowed large case. What remains is to exclude a rulebook with that intermediate number of models whenever it is smaller. The cited paper derives further restrictions in a stronger language, without proving this exclusion for every complete first-order rulebook with countably many symbols.[1]

The 2026 paper studies stronger infinitary settings without proving the original universal statement.[1]

A rulebook can settle every statement in its language without picking out just one structure. Counting structures after removing differences of names measures how much freedom the rules still leave. Vaught connects that freedom with the sizes of infinite sets: are the families allowed by logical rules more restricted than arbitrary collections?[1]

Completeness controls what can be said in first-order sentences, while isomorphism compares whole structures. The gap between these notions lets a single theory have a rich classification problem even after every sentence is decided. Vaught asks how that richness can be organized as an infinite family.

Scott analysis connects counting with how much descriptive complexity is needed to distinguish structures. The cited paper studies such complexity in a stronger infinitary setting, where countable combinations of formulas are permitted. That methodological connection explains why descriptive set theory and model classification meet here; the paper’s restricted results do not settle the original first-order claim.[1]