I haven't read the Bell Curve so I cannot say if it is racist or not, but Wikipedia has two quotes suggesting that the relationship between IQ and race is unresolved and that they suggest both environmental and genetic factors have something to do with it. Is this what sparked the controversy?
My own view is that while, as I say, I'm not sure, I suspect that the differences can be explained by environmental differences.
If it comes down to environment there's an explanitory escape clause - improve the environment and race-linked IQ differences disappear. That falls in line with social justice concerns and is an acceptable way forward. If, on the other hand, it comes down to real, demonstrable genetic differences, there's a problem - that's taboo science.
The usual way to placate social justice concerns is to say that race itself is ill-defined, and, coupled with figuring out just what IQ actually measures, the whole subject should be thrown out as junk science.
Can it be repaired? I think it can, but have no idea what the results would show.
What's needed is a tightening of both definitions - race and IQ. Cognitive scientists have been working on the second and came up with "g" which they argue is both internally and externally consistent. So what about race? Can that be pinned down?
It can be, if you just decide our impression of someone's race is misleading and go with a genetic definition instead. That would mean throwing out "self identification" of ethnicity and the kind of social constructs we might want to keep for political purposes, but it has the advantage of a clear definition for purposes of deriving associations - the epidemiology we are interested in.
How to pin down race genetically? Well, look at genes of course!
Scientists have figured out that lighter skinned East Asians get their skin color mostly from a non-working version of kitlg. Northern European people with lighter skin often have a poorly working version of SLC24A5. A small number of pale northern Europeans get their skin color from a non-working MC1R gene. From:
http://genetics.thetech.org/ask/ask288
Granted, that's a pretty shallow definition of what race is all about, but it has the advantage of getting the ball rolling. If we find no correlation with skin color (at least one way we talk about race) then fine, no need to press the case any further - it was all a mirage anyhow.
If there is a correlation, then we are back to nature/nurture conversations and have to start separating out those with background A from background B when the genes are the same. This is normal fare in statistical epidemiology.
The gold standard would be linking specific genes to g scores with an underlying explanatory theory - in the way you might determine someone's risk for breast cancer knowing how the genes interact.
In the end, and even if it all plays out nicely (I'm looking at you, reproducible study!), will we find out anything worth knowing? I seriously doubt it. And all the while we risk feeding oxygen to racists. They are even worse than Creationists when it comes to cherry-picking.