ACF universal is the theory of integral domains

290 Views Asked by At

When studying David Marker's "Model Theory: An Introduction" book trying to understand the proof of Lemma 3.2.1 which says: $ACF_{\forall}$ is the theory of integral domains, I couldn't understand the last line of the proof which comes as followa: (I copy the whole proof)

Proof. The axioms for integral domains are universal consequences of $ACF$. If $D$ is an integral domain, then the algebraic closure of the fraction field of $D$ is a model of $ACF$. Because every integral domain is a subring of an algebraically closed field, $ACF_{\forall}$ is the theory of integral domains.

David Marker wants to show $ACF_{\forall}$ is the theory of integral domains, he is just satisfied with proving, as in another book :Model theory notes by Kevin Buzzard(which you can find it in page 12-13 from the link below*), that these two theories have the same models. But I myself think we have to show $ACF_{\forall} = ID$. The author has already proved that $Mod(ACF_{\forall}) = Mod(ID)$ and it is easy to show $ID\subseteq ACF_{\forall}$ also ,but I can't show the converse inclusion. Do you think if $Mod(T_{1}) = Mod(T_{2})$ and $T_{1}\subseteq T_{2}$, then we have $T_{2}\subseteq T_{1}$ for two theories $T_{1}$ and $T_{2}$? If yes our problem is solved? or ...

I also asked this question in this manner which couldn't lead me to my very mean if you want to see:
"A question about two theories and their models "

Please say if my understanding of the original proof is true or not? and help to prove it, please.

*http://wwwf.imperial.ac.uk/~buzzard/maths/research/notes/model_theory_notes.pdf

1

There are 1 best solutions below

7
On BEST ANSWER

The use of the phrase "is the theory of integral domains" in the statement and proof of Marker's Lemma 3.2.1 is very sloppy. What Marker actually means is that $ACF_{\forall}$ axiomatizes the class of integral domains and that is what the proof proves (and is what is needed at the point of use of the lemma to show that $ACF$ has algebraically prime models).

It is clearly wrong to say that $ACF_{\forall}$ is the (full) theory of integral domains (which is the only notion of theory of a class of structures that Marker has defined), as there are sentences like $\forall x \exists y\cdot x + y = 0$ that hold in any integral domain but are not purely universal.

The closest you can get to a true statement along the lines of Marker's actual statement of the lemma is that $ACF_{\forall} = ID_{\forall}$, but that isn't exactly what is wanted where the lemma is used.