2
$\begingroup$

Let $A = \{a,b,c,d\}$ be a set of totally ordered alphabets, a Lyndon word over $A$ is a word $w$ in $A^*$ such that if $w=uv$ is a factorization of $w$ into non-empty subwords, then $u<v$ in lexicographic order. This is equivalent to being the unique minimum word (in lexicographic order) among all its rotations.

Let $\mathcal{L}(A)$ be the free Lie algebra on the set $A$. It is well known that Lyndon words in $A$ indexes a basis for $\mathcal(L)(A)$.

Assume $a$ is the least alphabet and let $w=a_1a_2\cdots a_k \in A^*$ (free monoid over $A$) define $L(w) = [a_1,[a_2,[\cdots[a_{k-1},a_k]\cdots]$ to be the right associative Lie monomial associated to $w$ in $\mathcal{L}(A)$. By rotation, I always mean another word $w^{'}$ which is obtained from $w$ by rotating the alphabets cyclically and further we also want this word to start with $a$.

Let $w = abcad$ and let $w^{'}=adabc$ be a rotation of $w$. My question is $L(w)$ and $L(w^{'})$ are linearly dependent as Lyndon words indexes a basis and $w$ and $w^{'}$ are rotationally equivalent.

More generally, I want to understand the linear dependence relation between the Lie monomials $L(w),L(w_1),\dots,L(w_k)$ where $w,w_1,w_2,\dots,w_k$ are all possible rotations of $w$ in the above sense.

How to understand the linear dependence relation between general right-associative Lie monomials and what makes the Lyndon words the indexing set for the basis of free Lie algebras? How the rotation of words affects the Lyndon words.

Kindly share your thoughts.

Thank you.

$\endgroup$
6
  • 1
    $\begingroup$ I am not sure I understand what you mean. But $\left[a,\left[b,\left[c,d\right]\right]\right]$, $\left[b,\left[c,\left[d,a\right]\right]\right]$, $\left[c,\left[d,\left[a,b\right]\right]\right]$ and $\left[d,\left[a,\left[b,c\right]\right]\right]$ are (in general) linearly independent. $\endgroup$ Commented Jan 29, 2020 at 23:56
  • $\begingroup$ It's been too long since I've thought about this stuff, but I want to recommend the book "Free Lie Algebras" by Christophe Reutenauer. It was an absolute pleasure to read, and I'm certain that the book has a good discussion of why there is a basis in correspondence with Lyndon words. $\endgroup$
    – Aaron
    Commented Jan 30, 2020 at 1:01
  • $\begingroup$ @darijgrinberg Can you please explain to me, how those monomials are linearly independent? I have edited the question, kindly reconsider this. Thank you. $\endgroup$
    – GA316
    Commented Jan 30, 2020 at 5:19
  • $\begingroup$ Not much to explain -- just compute them in the free Lie algebra and check that there is no nontrivial relation. I did it with SageMath and github.com/darijgr/sage-subspace . $\endgroup$ Commented Jan 30, 2020 at 9:41
  • $\begingroup$ @darijgrinberg Thanks for the code. It is very nice. Everything has been worked out for free Lie algebras. Can you please tell me how to check the linear dependence of $\left[a,\left[b, \left[c,d\right]\right]\right]$,$\left[b,\left[c,\left[d,a\right]\right]\right]$,$\left[c,\left[d,\left[a,b\right]\right]\right]$ and $\left[d,\left[a,\left[b,c\right]\right]\right]$ if we impose some commutation relations among the generators $a,b,c$ and $d$ using this sage codes?. Thank you. $\endgroup$
    – GA316
    Commented Jan 31, 2020 at 5:24

0

You must log in to answer this question.

Browse other questions tagged .