Just a couple minutes ago, I submitted a new paper to the arXiv, actually my first in last October. Oddly enough, I’m fairly sure I’ll be submitting a second within the next week. My projects have an odd habit of wrapping up in big bunches. Go figure.

Anyways, it’s not public on the arXiv yet, but you can read it on my website. By Monday evening, it should be available at this link.

It’s got the somewhat overly fancy name of “Wall-crossing morphisms in Khovanov-Rozansky homology,” which I’m sure will generate lots of attention, because everything with “Khovanov” in its name is hot these days (I’ve heard it suggested, only half-facetiously, that “Khovanism” is now its own branch of mathematics).

Before going further you may want to do preliminary reading (or you may want to ignore me): if you don’t know what a Vassiliev derivative is, Dror (as usual) explains it better than I ever could. If you don’t know what Khovanov-Rozansky homology is…well, that’s harder. You can try to read what I have to say on the subject, or Jake Rasmussen’s excellent paper. Hopefully, we’ll eventually have a page around here that explains this, but it’s a little early for that.

Since we’re blogging here, I’ll be honest; it took me a while to warm up to this paper (and as is, I don’t think it’s my best work). At the beginning, I was basically just taking on faith from Nadya (my co-author) that this stuff was worth writing up. However, I eventually came around to the view that Khovanov-Rozansky homology is so mysterious at the moment, a paper which says “Here’s this funny construction that lets you generalize a standard construction from knot theory to KR homology. Can any of the rest of you make heads or tails of it?”

Which, incidentally, really is the content of the paper. It does exactly one thing, which is construct a natural categorification of the Vasilliev derivative of the HOMFLY polynomial. Of course, the precise sense in which things are “natural” is a touchy subject, but I think we can justify it. We lay out something a schema for constructing Vassiliev derivatives of categorifications, which would require that a certain natural map exists between the KR homology of knots differing by a crossing (that’s a wall-crossing map). Then we look in the most obvious place, and–lo, and behold–there it is.

I’ll admit, I can’t prove any exciting theorems about this categorification, or (which would be more exciting) I can’t prove any exciting theorems *using* this categorification. But KR homology is exciting, and Vassiliev derivatives are exciting, so it should be even more exciting when they collide. Not to mention that the sheer miraculousness of the stars aligning correctly to produce an obvious wall-crossing map indicates *something* is going on here. My personal feeling is that these maps are pointing the way to the functoriality of HOMFLY homology (and can be thought of as a narrow, crippled version of such). Hopefully, someone else can provide the necessary context to put in its right place.

Is the reference “[Shi2] Nadya Shirokova, The finiteness result for Khovanov homology., preprint, 2006” available somewhere?

I hadn’t followed this during Nadya’s talk at Berkeley last October, and I’m curious to see what the statement “Khovanov homology restricted to the subcategories of knots of bounded crossing number have finite type” means.

Nadya says that it is not fit for human consumption yet, and I believe her.