MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

My apologies: There were a couple of typos in the original question. Hope I got them all.

Let $\kappa$ be an uncountable cardinal of cofinality $\omega$ and $M$ a model of size $\kappa$. We equip $\kappa^\kappa$ with the product topology and let $Aut(M)\subset \kappa^\kappa$ denote the topological group of automorphisms of $M$. The closure of $Aut(M)$ under the product topology is denoted by $\overline{Aut(M)}$ and $\partial Aut(M)=\overline{Aut(M)}\setminus Aut(M)$ denotes the boundary of $Aut(M)$.

My question is: If we know that (EDIT)$|\partial Aut(M)|\ge\kappa^+$(EDIT), can we say anything about the cardinality of $Aut(M)$, other than (EDIT)$|Aut(M)|\le|\partial Aut(M)|$(EDIT)?

Side note: If $f$ is in $\partial Aut(M)$, then $f$ is 1-1, not onto, and for every formula $\phi$ and every finite $\vec{a}$, $M\models\phi[\vec{a}]$ iff $M\models\phi[f(\vec{a})]$. I.e. $f$ is an elementary embedding.

share|cite|improve this question
For the product topology on $\kappa^\kappa$, you place the discrete topology on each factor? – Joel David Hamkins Aug 2 '12 at 19:21
Well, we can say that $|Aut(M)| \le \kappa^+$ because, choosing $f \in \partial Aut(M)$, the map $g \mapsto f \circ g$ is an injection $Aut(M) \to \partial Aut(M)$. So if $|Aut(M)| \ge \kappa^+$ (this isn't immediately apparent to me) then $|Aut(M)| = \kappa^+$. – Trevor Wilson Aug 2 '12 at 19:23
I was assuming that you meant the product of the discrete topologies on $\kappa$, by the way. – Trevor Wilson Aug 2 '12 at 19:24
At both Joel and Trevor: Yes, the topology on $\kappa$ is the discrete topology. – Ioannis Souldatos Aug 2 '12 at 21:24
@ Trevor: Your argument establishes $|Aut(M)|\le|\partial Aut(M)|$. My apologies there are a couple of typos in the question. It seems it has been a long day! – Ioannis Souldatos Aug 2 '12 at 22:17

By Trevor's comment, $|Aut(M)|\leq|\partial Aut(M)|$. From purely topological considerations we get this:

Let $\lambda=Aut(M)$.
The closure of $Aut(M)$ cannot have more than $2^{2^\lambda}$ elements. If $\kappa^+\leq|\partial Aut(M)|$, then we must have $2^{2^\lambda}\geq\kappa^+$. So in particular, if GCH or something similar holds below $\kappa$, we have $|Aut(M)|\geq\kappa$.

share|cite|improve this answer
Stefan, this is a nice observation making use of the fact that $\kappa$ is a limit cardinal. My original motivation stemmed from the fact that for $\kappa$ countable, the mere existence of one element in $\partial Aut(M)$ implies the existence of $2^\omega$ many automorphisms. This follows from a theorem of Su Gao in "On automorphism groups of countable structures", J. Symb. Log. 63, No.3, 891-896 (1998). – Ioannis Souldatos Aug 8 '12 at 15:36

I was able to prove the following for $\kappa$ of cofinality $\omega$:

Assume $M$ is a model of size $\kappa$, $cf(\kappa)=\omega$ and for all $\alpha<\beta<\kappa^+$, there exist functions $j_{\beta,\alpha}$ in $\overline{Aut(M)}^{T}\setminus Aut(M)$, such that for $\alpha<\beta<\gamma<\kappa^+$,

$$(*) j_{\gamma,\beta}\circ j_{\beta,\alpha}=j_{\gamma,\alpha},$$ where $\overline{Aut(M)}^{T}$ is the closure of $Aut(M)$ under the product topology in $\kappa^\kappa$.

Then there are at least $\kappa^\omega$ automorphisms of $M$. The proof is using Infinitary Logic and the assumption that $cf(\kappa)=\omega$ is fundamental.

I do not know of any way to get the above result directly.

(link to arXiv:

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.