Some empires go about razing everything as they expand, grinding the culture and bodies of the conquered into dust, gloating in the lamentations of their women, &c. On one hand, if you're the kind of jerk who's inclined to go start a war of conquest in the first place, this probably feels amazing. On the other hand, this is a bit self-defeating: you're burning up what could be your own new resources. Attempts to do this generally burn bright, burn quick, and smell awful. Far smarter to march in, assert your fundamental supremacy, but do so while co-opting local elites, incorporating local culture, and making clear that participation in your empire has plenty of advantages. Your empire will be a hybrid thing, and you won't be able to distribute all the land to your most trusted cohorts, but your empire will also be more likely to exist in twenty years - and to exist as a more interesting, prosperous place.
The first, critical half of Understanding Class is a project of theoretical imperialism, but of the nice, friendly sort that co-opts local elites, hires their poets, marries their children to those of its generals, and only brings out a few examples for grisly execution - just to show them who's boss. Wright marches through the fields of Weberians, Durkheimians, and other sundry principates to clearly announce that Marxist class theory is supreme, that it the most fundamental nature of power relations across human societies, but that it is also a benign, flexible empire happy to contain all other theories within itself. Or, to switch metaphors a bit (but only a bit), he is like one of those Christian missionaries who shows the pagans that they were engaging in before was actually only an incomplete and implicit Christianity all along.
Two broad strategies characterize this ecumenism, although of course their particular implementation depends on the details of the theory he's engaging with. Both start off by acknowledging the legitimacy of forms of "class analysis" and social phenomena more generally that are pursued by non-Marxists, and which seem to take categories other than the bourgeoisie and proletariat (or whatever) as their object: subjectively recognized status groups, professions, und so weter. The first is that the Marxist account of antagonistic class relations rooted in property ownership and the production process represent the background condition for the more immediate relations that the other sociologists want to pursue. So, for instance, Weber and his epigones want to anchor class in exchange relations, but these exchange relations obviously depend on productive and ownership relations. The second, rather critical realist, strategy is to show the necessity of latent causal nexuses to any sociological theory with greater explanatory strategy than "shit happens." This has as its target those theories whose notion of "realism" is that sociological objects that are not a mere creation of the analyst but have some emic reality. Not so, we need to explain why some particular subjectively felt groups are more likely to arise than others, and objective class relations allows us to do this.
Like Caesar, Wright's anabasis also produces some excellent descriptions of the conquered. I'm not familiar with everyone he details, but let it be said that I've read Max Weber and Charles Tilly quite extensively, that both are very lucid writers, and that I still feel that I learned a great deal about their theories from Wright's gloss. Like any conquerer's gaze, it may distort, but if so it's a useful distortion of two sociologists I already admire and have learned a lot from.
The second, more speculative, part is fascinating but weaker, both analytically and (if it's not inappropriate to raise this objection) politically. If I were tempted to extend the metaphor, I could say that this is the fruit of offering too many concessions to local elites (meaning, interlarding his Marxism with too many sundry additional theories,) but I don't actually think that this is the case - except in the sense that there's a methodological weakness that's shared with much glib neoclassical economics. That methodological weakness is that he constructs an admirably clear theoretical model, on empirical foundations that seem to me to be very thin (one or two studies that have an alternative explanation), and then uses it to draw conclusions on the possibilities of "positive class compromise," i.e. conditions that would benefit both workers and capitalists when socialism is impossible. As arguments for Bernsteinian revisionism go, these are extremely weak. However, understood not as conclusions but speculations, they are inherently interesting, and those wishing to give greater substantiation (or falsification) to them may produce something quite useful. Bong-rip ideas have an important place in social science, not everything should be judged as if it's a finished product.
I probably ought have read Wright's earlier books on class first, in which he lays out his theories explicitly rather than in contrast with others - indeed I leave with the desire to get around to them soon - but didn't feel as though I failed to understand anything on that account, and the book was stimulating throughout. Recommended to anyone who's interested in the broad subject and isn't too intimidated by a bit of disciplinary inside baseball.