Terry Eagleton has a MacIntyre-inflected, postmodernist-bashing piece in the most recent LRB, in which he has a fair bit of fun with a quite earnest-sounding book on typecasting, in which he unashamedly defends a materialist account of stereotypes. He begins by pointing out that many stereotypes have a basis in fact
It is an open secret, for example, that Ulster Protestants are not by and large dandyish aesthetes notable for their extravagant wordplay and surreal sense of humour. The English middle classes are for the most part less physically and emotionally expressive than Neapolitan dockers. It is unusual to meet a working-class Liverpudlian who dresses for dinner, other than in the sense of putting on a shirt. Corporation executives tend not to be Dadaists.
and then goes on to explain this feature by pointing to the commonalities of experience which to some degree constitute these groups
If a group of people have shared roughly the same material conditions over long periods of time, it would be astonishing if they were not to manifest some cultural and psychological traits in common... This does not mean that such people will all be clones of one another; but habits of mind, patterns of behaviour and emotional dispositions are bound up with the way we live with others, rather than being purely personal affairs.
More than that, some cultural shorthands are, after all, useful, as Eagleton points out:
Without stereotyping of some kind, social life would grind to a halt. If the plumber turns up to fix the drains dressed in tights and a tutu, I would naturally be liberal-minded enough to invite him to perform a few pirouettes at the sink; but if the bank manager insists on discussing my loan in Latvian, I might take my business elsewhere. Human freedom is a question of life being reasonably predictable, not of being joyously liberated from rules. Unless we can calculate the effects of our actions, which includes the way others might typically respond to them, we will be incapable of realising our projects effectively.
I have quite a lot of sympathy, firstly with the idea that there are genuine, non-trivial national, regional and even ethnic characteristics, some of which are really picked up on by stereotypes, and secondly, with the idea that predictability in human life is valuable, although I'm not sure that a claim quite as general as the one Eagleton makes legitimates stereotypes. He could easily, I think, fill it out though: the predictability in human behaviour derived from the use of stereotypes is, in certain situations, useful. Not being able to tell whether it would better serve your projects to go for a night out in a working men's club, or indeed a formal dinner at an Oxford college, in a tutu would be seriously disruptive of human projects. This, though, leaves a problem about explaining exactly what is wrong with stereotypes, since they are a quasi-legitimate - because not wholly inaccurate - predictive shorthand and such shorthands are useful.
This is an issue which Eagleton seems relatively unconcerned by though. We don't want to say, I assume, that all stereotypes are, if reasonably accurate, legitimate, precisely because social categories are malleable enough, sufficiently reflexive, to create individuals who fulfill them. It strikes me, for example, that whilst the Irish may well have had reasons, given their material situation, to evolve a culture which did not, in comparison to, say, that of the English upper-working class, place a high value on hard work, it also may be be the case that being told the Irish are all lazy may, if you're Irish, cause you work less hard than you otherwise would. Further, stereotypes do not only pick out relevant general features of some group or other quasi-accurately. They also tend to legitimate particular forms of treatment against that group.
Eagleton writes that the authors of the book he is having his fun with endorse the Foucauldian claim that 'naming is a form of exercising power', and comments that this
implies that power is always objectionable. It is not a view that the powerless generally share.
The second of those two claims, as far as it goes, quite correct. Note though, that in many cases one of the things that makes the powerless powerless is precisely that they have lacked power over the categories into which they have been placed. One needs only think of the use of the slogan 'the personal is the political' by feminists to understand that, I think. The first claim though, whilst it may be true of some Foucauldian thought, is not automatically tied to the originally stated Foucauldian view, as the experience of the use of the feminist slogan clearly demonstrates. Undoubtedly one of the things the powerless would like to be able to do is some authoritative naming of their own.
Authoritative naming, then, is not of itself the problem with stereotypes, as suggested above. The problem is that forms of treatment get tied to that naming. There are certain things, for example, that you can't do with a table - driving them to work for example. That tables cannot be driven to work is simply part of the meaning of what it is for something to be a table. Likewise, those under the rubric of some malicious stereotype are liable to treatment as according to that stereotype: the Irish can't be trusted to do an honest day's work, or near a drink, for example. The simple fact of differential treatment according to some category or other, though, cannot be the whole of the problem, for we would hardly want to say that murderers or moral saints should not, if correctly identified, differentially treated.
I don't think we would even want to say that stereotypes can never legitimate differential treatment, or at least differential attitudes. It would be odd, to say the least, to think that an Amazonian tribesman should be treated the same as a Londoner in our beloved capital, and surely, our preconceptions about what Amazonian tribesmen are like might be at least of some use in deciding how we ought to treat them. Even more than that, denying that we ought to name things, and ascribe moral properties to them in doing so, would strip our agency down to a radically disembodied core: it would not only be opaque how could I aim to instantiate any virtue at all, but unclear why I should not, since the freedom doing so would deny me is itself a moral ideal.
This, insofar as I understand Foucault, was what his later work focused on: the problem that whilst norms can clearly be oppressive, their total absence would be destructive of both the moral standpoint and human agency at all. Some kind of happy medium needs to be struck, where naming, and stereotyping as a form of naming, is legitimate under certain circumstances but not all. This is the difference I was trying to gesture at here: what we need to find is a distinction between
the kinds of attitude which legitimates wishing, even in jest, people's burial under thousands of tons of volcanic ash is not the same as the kind of attitude which regards people as unsatisfactorily Mediterranean in temperament. Think of the apparently once-widespread practice of hissing like leaking gas at Spurs fans, and the difference between that and thinking that Scots are typically a bit tightfisted, or the English rather standoffish.
On top of a requirement of basic accuracy, I'd suggest some liberal account of the person, of an autonomous agent partly but not wholly constituted by their commitments, who simply cannot be treated in certain ways, and talk about the centrality to the idea of tolerance of not liking the conduct that is tolerated. But then I would say that, wouldn't I?