I particularly don't understand this point. Don't scientists separate things into "classes" and taxonomies all the time? One thing's an electron, another's a positron? Sometimes these particles interact explosively and annihilate each other -- literally speaking.
And the "conversion of abstractions into entities" sounds like the definition of reification to me. Certainly, any software programmer, particularly one using object-oriented techniques, consciously makes taxonomies and reifies abstractions all the time. These technologies run useful parts of the world.
And regardless of what the "sciences" think, the problem is that societies like to stratify themselves into these classes. I'm sure class theorists would prefer this weren't the case. Or do bosses usually go around gifting all their sweatshop subordinates egalitarian decisionmaking power over their corporation, because rigid economic classes are so 19th century?
Tayssir