# Is Gödel's completeness theorem a representation theorem?

In general a representation theorem is — according to Wikipedia — a "theorem that states that every abstract structure with certain properties is isomorphic to a concrete structure". Wikipedia gives a list of canonical examples.

[Side remark: Interesting enough, the only genuinely set theoretical example (Mostowski's collapsing theorem) is filed under "category theory", and Dedekind's lemma which is fairly easy to grasp is missing in this list.]

Can Gödel's completeness theorem be considered a representation theorem?

1. Consistent logical formulae (or sets of, ie. theories) can be considered as (descriptions of) abstract structures
2. The (set theoretical) models of such formulae (resp. theories) are concrete structures.

One important difference between Gödel's incompleteness theorem and some of the representation theorems listed at Wikipedia is, that not only in it's proofs (known to me) the concrete structures don't "come free with" the abstract structures — as most easily seen in the Dedekind case of posets — but that the proofs are non-constructive at all.

So I'd like to flank my question:

Is the Wikipedia "definition" of the notion of a representation theorem adequate or would practicing representation theorists refine it, e.g. to exclude "non-constructive" theorems.

• Representation theory is not the "theory of representing concrete theories". I also fail to see any reason whatsoever for the incompleteness theorem to be a "representation theorem", if anything it is the complete opposite. It tells us that under some constraints there is no concrete model. – Asaf Karagila Feb 15 '12 at 10:22
• Please help me to fix the misunderstandings: a) I used the term "representation theory" only in the tag, else I talked about representation theorems. b) I asked about the completeness theorem (sic!), not about the in completeness theorem. – Hans-Peter Stricker Feb 15 '12 at 10:27
• As for (a) my point was exactly aimed to the tag. Much like [logic] does not apply for everything which requires "a logical thinking" and [set-theory] has nothing to do with just any questions with sets. As for (b)... I guess that is what happens when you're hungover. However the point remains, the completeness theorem just connects syntax and semantics - it does not imply anything is isomorphic to another. – Asaf Karagila Feb 15 '12 at 10:36
• As for (a): I see your point, but I want to leave the tag, since there is no tag for "rep. theorems", and there is - presumably - a connection between "rep. theory" and "rep. theorem". As for (b): Maybe you are right, but for me it's discussion-worthy whether the compl. theorem really just connects syntax and semantics and whether it really does not imply anything about one thing being isomorphic to another. – Hans-Peter Stricker Feb 15 '12 at 10:43
• There is no direct connection between representation theory and representation theorems; this is why I approved the edit deleting the tag from the question. – anon Feb 15 '12 at 11:13