2
$\begingroup$

For a bridgeland stability condition $(P,Z)$ on $\mathcal{C}$ and $a > b$ we know that $Hom^0(A,B)=0$ for $A,B \in P(a), P(b)$ respectively.

I would like to see the geometric incarnation of this under mirror symmetry, in the simplest case of an elliptic curve. Thus from now on we work with a torus\elliptic curve.

Namely, under many simplifications if I understand, we can think of a bridgeland stability condition as a choice of symplectic form $\omega$ on $E$ (compatible with a fixed complex structure $J$).

Now every curve in $E$ is a lagrangian, and it should be semistable (i.e from $P(A)$) sorta when it's a special lagrangian, which I found is correlative to being length minimizing wrt to the natural metric $g$ that $\omega, J$ induce in its homology class.

I am not sure how to geometrically see the slope, nor how to then see that if $a>b$, then the fukaya hom vanishes, and that is my question.

$\endgroup$

1 Answer 1

1
$\begingroup$

Until someone better comes along here is some info about mirror symmetry on elliptic curve.

The rational line of slope $m/n$ for coprime $m,n$ is the appropriate stable vector bundle.

Then if you have rational lines $L_1 ,L_2$ with $L_1$ having bigger slope than to show $Hom(L_1, L_2)=0$ we analyze $L_1 \cap L_2$. THere ie one intersection point $(0,0)$ and the bigger slope translates to an orientation statement about the intersection point.

Namely, recall there a $\mathbb{Z}$ grading on $Hom(L_1,L_2)$ coming from a grading on the intersection points but to get just the $\mathbb{Z}/2$ one you only care about orientation.

So we see that the orientation will be $1 \pmod{2}$ and so in particular there are no homs decreasing slope.

As a further fun point; rotating the elliptic curve by 90 degrees fukayawise is the auto equivalence of derived sheaf category of the poincare bundle.

What's curious about this argument is that it was a $\mathbb{Z}/2$ argument; the reason there are no homs is because the grading modulo 2 won't allow it. That surprises me and I wonder if in higher dimensions it's similiar or if it uses the whole $\mathbb{Z}$ grading.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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