-4

At the moment I am taking a course in logic (which includes an extensive amount of model thoery). In the last lecture we learned about quantifier elimination. On the last homework assignment we were asked the following:

Let ${\cal L}=\left\{ E\right\}$ be a binary equivalence relation symbol. For each of the following theories either prove that they have quantifier elimination or give an example showing that they do not have quantifier elimination.
a) $E$ has infinitely many equivalence classes all of size 2.
b) $E$ has infinitely many equivalence classes all of which are infinite.
c) $E$ has infinitely many classes of size 2, infinitely many classes of size 3, and every class has size 2 or 3.
d) $E$ has one equivalence class of size $n$ for each $n<\omega$ .

I'm working on the problem and I found a similar question from five years ago (here). I've gone through all the comments and linked resources, but I'm still stuck on all four sub-questions. Even with hints (that one can easily find across the links in the comments as well as similar questions), I can't even figure out if I'm supposed to prove or disprove the claims for any of them. Could someone please explain what's happening here? Answering a couple of the sub-questions would really help me develop the right approach any way of thinking.

Definition - A theory $ T$ is said to have quantifier elimination if for every formula $\varphi$ there is a quantifier free formula $\psi$ such that $T\models\varphi\iff\psi$.

Shavit
  • 47
  • 5
  • @AlexKruckman Maybe vague is not the appropriate word, but I meant to emphasize that I really don't know understand it, and more importantly how to work with it. – Shavit Jun 30 '25 at 17:49
  • @AlexKruckman That's the other problem, not really. In the lecture all we have seen (that I can clearly say is related) is; we saw that DLO has QE, we mentioned the MRDP theorem, we saw some examples in $\left( \mathbb{Z} , 0 ,s \right)$ and in $\left( \mathbb{Z} , 0 , <,s \right)$ (where s is the successor). But no criterion of sufficient conditions of any sort. – Shavit Jun 30 '25 at 17:58
  • @AlexKruckman No problem, I more then happy to answer. In the proof that DLO has QE we used sign conditions, with the fact that this theory is $\aleph_{0}$-categorical and complete. – Shavit Jun 30 '25 at 18:55
  • @AlexKruckman It might sound ridicules but I'm dead serious, if you give me your email I can send you the full notes my lecturer wrote ;) – Shavit Jun 30 '25 at 18:58
  • @AlexKruckman last example: "The proof follows the previous scheme. However, now there are more atomic formulas:in addition to the formulas x = ti, we will have formulas x < ti. Therefore, it is impossible suppose that all values xthat are not equal to any t1,...,tk behave the same and our technique considering the case when all equalities are false no longer passes. For these values x1,...,xn, the numbers t1,...,tk represent the intervals between these xs, and to find out the truth of the formula φ we need to try (in addition to all ti) at least one number from each interval". Thanks again!!! – Shavit Jun 30 '25 at 19:03
  • @AlexKruckman For (ℤ,0,) the proof was awfully long. It was induction based (from what I recall we assumed without lost of generality that all formulas are quantifier-free (I didn't understood this part tbh) then argued that a formula given formula $\psi({1},...,{})$ is a boolean combination of atomic formulas. then somehow using terms and by repeating some process we obtained a quantifier free formulation that is equivalent to the initial one. Sorry if my explanation is not the best, I really didn't understand this part of the lecture :) – Shavit Jun 30 '25 at 20:28
  • One useful thing for the counterexamples is that QE implies (but is not equivalent to) any embedding between two models is elementary (why?). For instance, for d, can you find an embedding of one model into another that is not elementary? – spaceisdarkgreen Jun 30 '25 at 21:32

0 Answers0