7
$\begingroup$

Let $K\subset L$ be a field extension and $A, B\subset L$ be $K$-subspaces of $L$ of finite positive dimensions. Assume further that for every $a, b \in L$ and every nontrivial proper finite intermediate subfield $K\subset M\subsetneq L$, we have $$ \dim_{K}(aM\cap A) + \dim_{K} (bM\cap B)\leq [M:K]+1. $$ Then does the following hold? $$ \dim_{K}\langle AB\rangle\geq\dim_{K} A+\dim_{K} B-1, $$ where $AB=\{ab\mid a\in A, b\in B\}$ and $\langle AB\rangle$ is the $K$-subspace of $L$ spanned by $AB$.

Intuition: For every proper finite intermediate field $K \subset M \subsetneq L$, and for any $a, b \in L$, the assumption $\dim_K(aM \cap A) + \dim_K(bM \cap B) \leq [M : K] + 1$ restricts the "concentration" of $A$ and $B$ inside proper subfields. Roughly speaking, it avoids the case where a large part of both $A$ and $B$ live inside a small field extension — a situation that could lead to a "small" product span $\langle AB \rangle$.

Remark 1: Note that a similar result in a group setting is true. (Replace $K\subset L$ with an abelian group $G$, subspaces with subsets and dimension with cardinality.) That is, we have the following result:

Theorem: Let $G$ be an abelian group and $A$ and $B$ be nonempty finite subsets of $G$. Assume that for every nontrivial finite subgroup $H$ of $G$ and all elements $a,b\in G$, we have $|aH\cap A|+|bH\cap B|\leq|H|+1$. Then we have $|AB|\geq|A|+|B|-1$.

Remark 2: In the linear setting we have the following theorem in this paper: Let $K\subset L$ be a field extension and $A\subset L$ be a $K$-subspace of $L$ of finite positive dimension. Then either for every finite-dimensional subspace $B$ of $L$, we have \begin{align*} \dim_K\langle AB\rangle\geq\dim_K A+\dim_K B-1, \end{align*}

or there exists an intermediate field $K\subset M\subset L$, such that for every finite-dimensional subspace $B$ of $L$ satisfying \begin{align*} \dim_K\langle AB\rangle < \dim_K A+\dim_K B -1, \end{align*} we have $\langle AB\rangle M=\langle AB\rangle$.

Remark 3: Another related theorem from this paper: Let $K\subset L$ be a field extension in which every algebraic element of $L$ is separable over $K$. Let $A,B\subset L$ be nonzero finite dimensional $K$-subspaces of $L$ and $M$ be the stabilizer of $\langle AB\rangle$, i.e., $M=\{x\in L| x\langle AB\rangle\subset \langle AB\rangle\}$. Then \begin{align*} \dim_K\langle AB\rangle\geq \dim_K A+\dim_K B-\dim_K M. \end{align*}

Remark 4: For every nonzero finite-dimensional $K$-subspace $X$ of $L$, we denote by \begin{align*} \partial_A(X)=\dim_K\langle XA\rangle-\dim_{K} X. \end{align*} Let $\psi$ be the set of nonzero finite-dimensional $K$-subspaces $X$ of $L$ such that $\langle XA\rangle\neq L$. If $\psi$ is nonempty, we define the 1st-connectivity of $A$ by \begin{align*} \kappa(A)=\min_{X\in\psi}\partial_A(X). \end{align*} If $\psi$ is empty, we set $\kappa(A)\neq-\infty$. We define a 1st-fragment of $A$ to be an $M\in\psi$ with $\partial_A(M)=\kappa(A)$. A 1st-fragment with minimum dimension is called a 1-atom.

We now have the following theorem from this paper:

Theorem: Let $K\subset L$ be a field extension. Let $S$ be a $K$-subspace of $L$ containing $1$. Let $A$ be the 1-atom of $S$. If there exists a $K$-subspace $T$ of $L$ for which $\dim_{K}\langle ST\rangle<\dim_{K} S+\dim_{K}T-1<\dim_{K} L$, then $A$ is a subfield of $L$ properly containing $K$.

Remark 5: A result on large subspaces from this paper: Let $K \subseteq L$ be a field extension, with $\dim_K L$ finite. Let $A, B$ be non-zero subspaces of $L$ satisfying $\dim_K A + \dim_K B \geq \dim_K L$. Then $\langle AB \rangle = L$.

Remark 6: Let $L \subset L$ be a field extension. Define $\mu_{K,L}(r,s) = \min \dim_K <AB>$, where the minimum is taken over all $K$-subspaces $A, B$ of $L$ satisfying $\dim_K A = r$, $\dim_K B = s$. Define $\kappa_{K,L}(r,s) = \min_{h}(\lceil\frac{r}{h}\rceil+\lceil\frac{s}{h}\rceil-1)h $, where $h = [H : K]$ runs over the set of $K$-dimensions of all finite-dimensional intermediate fields $K \subset H \subset L$. It is proved in this paper that if $K \subset L$ is a field extension in which every algebraic element of $L$ is separable over $K$, then, for all positive integers $r, s \leq \dim_K L$, we have $\mu_{K,L}(r,s) = \kappa_{K,L}(r,s)$.

$\endgroup$
9
  • 1
    $\begingroup$ This seems related: mathoverflow.net/questions/141378 $\endgroup$ Commented May 30, 2024 at 7:09
  • 1
    $\begingroup$ can we replace $A,B$ to $S,T$ respectively and then forget about extra subspaces and the condition with quantifiers? $\endgroup$ Commented Jun 2, 2024 at 6:52
  • $\begingroup$ @Fedor Petrov: Yes, I just edited my question. Thanks! $\endgroup$ Commented Jun 2, 2024 at 21:31
  • 1
    $\begingroup$ Well, some unkind user might wonder why it took so long to get the question right. Post a question, maybe make one or two tweaks, sure, but going for the 16th time might make someone wonder whether you had any idea what you wanted to ask. $\endgroup$ Commented Jun 11, 2024 at 7:08
  • 2
    $\begingroup$ @GerryMyerson Seemingly you have not paid enough attention to the post's contents. Many of these edits involve adding resources over time, which could be beneficial. Unkind users should familiarize themselves with the nature and policies of Mathematics Overflow. FYI, editing posts to evolve the direction of the problem is a normal and acceptable practice within MO's guidelines. $\endgroup$ Commented Jun 11, 2024 at 7:20

0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.