Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Let $k$ be a field of characteristic zero, $X=\mathbb{G}_{m, k}=\mathrm{Spec}\ k[t, t^{-1}]$ the multiplicative group over k and $E=\mathcal{O}_X$ the trivial line bundle.

Consider the connection $\nabla: E \to E \otimes_{\mathcal{O}_X} \Omega^1_X$ defined by $\nabla(1)=\alpha \frac{dt}{t}$ for some $\alpha \in \mathbb{C}$, that is:

$\nabla=d+\alpha \frac{dt}{t}\wedge$

My question is: has $(E, \nabla)$ regular singularities?

Thanks for you help

share|improve this question
    
p^4: I didn't downvote you and yes it does. –  Donu Arapura Feb 5 '13 at 16:19
    
Dear Donu, could you explain why ? It seemed pretty obvious to me but then I found the paper mi.fu-berlin.de/users/esnault/preprints/helene/… page 2 and I was confused about the statement on sections –  pppp Feb 5 '13 at 16:39
    
It's obvious if you can use the standard Fuchs criterion that the singularities of the connection are logarithmic (cf Deligne's book). I get $t^s$ as a basis of horizontal sections in the example in the linked preprint. I'm sure they mean something else, but I don't have the time to read it carefully right now. –  Donu Arapura Feb 5 '13 at 16:57
    
That's the point. I don't see what exp(-t) has to do with horizontal sections –  pppp Feb 5 '13 at 17:43
add comment

1 Answer

Dear $p^4$,

I had a chance to glance further at the 4 author paper linked in your comment. Their connection form is $-dt + sdt/t$. This has double pole at $\infty$, so it isn't regular*. However, the connection in your question has logarithmic singularities at $0$ and $\infty$, so it is regular, as I said earlier. There is no contradiction.

*(Afterthought) This is a bit sloppy, since the Fuchs criterion is only a sufficient condition for regularity. But to see the non regularity, observe that the solution they write down $exp(t)t^{-s}$, which is correct (!), has bad singularities at $\infty$.

share|improve this answer
add comment

Your Answer

 
discard

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.