Thursday, September 18, 2003
It's good to be poor

I see, having a few minutes to surf the blogs before bed, that there is an interesting multiblog argument going on over Jacob Levy's recent New Republic column. Levy promises to respond to several of his critics, and I'm hoping that it's not too late to join in the fun. The article is about the Wall Street Journal's recent complaints that poor people don't pay enough taxes.

Here is the core of Levy's argument:

The general form of these arguments ("lucky duckies" as well as the arguments from the left) is: If we subject everyone to the same rules, institutions, or conditions, then there will be political demand to make them fair or otherwise tolerable. If we only subject some people to them, then some may be unfairly singled out or burdened; there will be opportunities to divide the citizenry, play the interests of some against those of others, and to undermine the overall desirable outcome. The only detail that changes from argument to argument is the class to which one tries to yoke people--the class of taxpayers, the class of potential soldiers, the class of recipients of government checks, etc.

Now, Levy has a point, but he has skipped an essential element of nearly all discussions of political equality: what exactly is a political system supposed to equalise? Equalising net tax rates are, I think, a fairly pathetic goal for a political theory. The simplest justification for progressive taxation is to raise the time-honoured - and nowadays nearly cliché - difference between formal and substantial equality.

I've already claimed that political systems should not be about equality, they should be about optimality, but that's just me. I would make the argument that many social institutions should treat people differently, that their goal should be the best possible outcome (by whatever standard is appropriate) for each person, rather than identical outcomes. To me, that means a guy with a billion dollars can live with a much higher tax rate than a guy with a minimum wage job. The self-development based case would go something this: a guy with a billion dollars can buy just about as much self-development as money will buy. His opportunities are less diminished by taxing his money away than the guy on minimum wage. Alternatively, I can make the case on the grounds that there is no way a guy with a billion dollars works 100,000 times more than someone with ten grand. All wealth is socially created (screw you, Ayn Rand!), and the only way a person can have that kind of wealth is when they owe the rest of the world for it. That would be a bit more conventionally Marxist, but I don't want to make the usual labour-value-based argument about exploitation. I don't think the guy on minimum wage is any less tied to the rest of society for his earnings than Bill Gates, but I'm quite okay with saying that he deserves his money more.

But actually, I mostly agree with Levy. Binding people to institutions is sometimes necessary precisely because of its effect on social cohesion. Frankly, participation in social institutions is rarely voluntary at all and usually can't be genuinely voluntary under any realistic circumstances.

I just think that that's a crappy argument for taxing the poor, but then I'm not very sympathetic to the Wall Street Journal's incessant complaining about taxes anyway. I don't think there is any good reason for state policy to assist in creating pressure for tax cuts in the US. I pay 50% income taxes and 21% VAT on a lot less than a top 1% income, and I don't complain about it. I suspect the conservative rich in America are just whiners.

Nota Bene: This should have been pulbished last night, but Blogger was on the blink.

Methodological Individualism and Actor-Network Theory

Cosma Shalizi brings up two points in reference to my third post on language rights that I'd like to address: First, that I haven't made a case against methodological individualism, and second, that I haven't made a case for Actor-Network Theory.

First, methodological individualism. Methodological individualism, at least as it applies to the social sciences, usually means that our theories about human activity should only make reference to the actions, and often reasons, that individuals do things and to the side-effects, expected and unexpected, of those actions. I have indeed put forward a theory that is not methodologically individualist. So, let me make a case against methodological individualism. Credit where credit is due, I stole part of this argument from Dan Sperber.

I can't for the life of me figure out how anyone could do historical linguistics in a methodologically individualistic way. Historical linguistics is the study of how languages change in time. Naturally, languages are spoken by individual people, but when we talk about language change, we are almost exclusively talking about a group phenomenon. If some individual somewhere changes the way they speak, without any reference to other people, it's of no interest to historical linguistics. It is only when the languages of whole communities of people change, and change in a largely uniform way, that it becomes interesting. Historical linguistics depends on this tendency for the whole community to change together. It simply makes no sense to say that linguistic change is something that happens to individuals or as something that follows from individual actions.

An individualist theory of historical linguistics would have to explain language change in terms of psycholinguistics. I don't see how that is reasonable, and I see even less reason to reject presently collectivist theories of historical linguistics just because they aren't grounded in psycholinguistics.

I subscribe to what is, depending on how you want to look at it, either a very weak form of reductionism or a moderately reductionist form of holism. I think that it's a good thing when theories with different objects are compatible and I think that there is often knowledge to be gained in working out that they are compatible, but I see no special reason to order our theories on the basis of some sense of fundamentalness and then use that scale to privilege some theories over others. It would be nice if our theories of psycholinguistics were compatible with historical linguistics, and since almost all of the theories of psycholinguistics and language acquisition that I know of allow for people's language to change over time and over generations in response to all sorts of things, I don't think that this compatibility is essentially a problem. But, no theory of psycholinguistics that I know of actually explains any theories of historical linguistics. Historical linguistics has functioned for centuries without privileging individualist explanations and I don't see any likelihood of change soon.

I think this same logic applies to anything that smacks of cultural anthropology, and a theory of socially mediated tools definitely fits the bill. People are born into a world that already exists, and the tools available to them are already chosen by others. The cultural machinery that supports them is largely hidden from view. Just as language can only persist as a social phenomena, collectives like states and companies and cultural phenomena like the division of labour only persist when they are supported by groups.

The grand contradiction of language rights is that they are inherently collective in nature, while the traditional liberal rights are usually understood as inherently personal. Nations and corporations are no different.

And that leads me to Actor-Network Theory. I could have made the same point somewhat differently. I might, for instance, have invoked more traditionally socialist characters like Stafford Beer, who discusses a similar sort of collective thinking in The Brain of the Firm. However, Bruno Latour draws attention to something that I think is important to the kind of collectivism I'm advancing. Collectives - as I've described them - are never just groups of people. Collectives always have non-human elements which are essential to its ability to think and act as a whole. A collective always includes at the very least mediating tools like codes for communication and structures that divide and order labour and enable the transformation of cognition into action. They can include other essential elements as well, and Latour makes a particular point of the importance of texts as cognitive elements. I think this is quite illuminating in areas like legal theory, where a document like a national constitution can be a genuine actor within the collective.

I should also point out that the kind of collectivism that I'm advocating is quite different from folk collectivism. I agree with Margaret Thatcher that there is no such thing as society, but there is such a thing as General Motors. That distinction between things that are just a bunch of people and things that are collectives is very important to the case I'm trying to make.

Next: I think I'm going to delve into the comments here for my next post, then Seth Edenbaum, for why I prefer not to use intrinsic value in normative political theory.

Update: Cosma responds, first citing Elster, then Sperber, Boudon and a book I haven't read. The book appears, according to the reviews I've found, advance some kind of psycholinguistic explanation for language change. I'll have to see it to evaluate it. The rest - reevaluating what it means for languages to be in contact, and rethinking the notion that there is a motivated distinction between internal and external causes for language change is something I'm pretty sympathetic towards. Sperber's article - which I'm pretty sure is the one I stole the example from historical linguistics from - seems to largely support my case, offering up a very weak kind of individualism and an epidemology of representations in place of strong methodological individualism. I am not as keen on the idea of an epidemology of representations, prefering instead an anthropology of tools. However, there is a fair amount of overlap between the two, something I think somebody with more time on his hands ought to explore in conjunction with a contrast of Dennett and Vygotsky. The Boudon paper is quite interesting, but not so convincing as all that. To identify all social explanation exclusively with individualist theories is, I think, really pushing the envelope beyond reasonable limits. I think it is not only perfectly reasonable, but absolutely necessary, to have theories of institutions which evoke the actions of institutions and explanations which make sense from an institutional point of view. I don't really see how an economics which includes the division of labour or a political science which deals with modern states, can genuinely do without it.

I have my differences with Elster, and most of them revolve around the key issue of methodological individualism and my rejection of public choice theory as a firm foundation for the social sciences. I see even less how public choice theory can explain lingusitic change than psycholoinguistics. However, I agree with Elster that a Marxist theory of institutions is not in terribly good shape, and I think he has identified the kind of thinking that has helped put it there. Once you start seeing the capitalist class, or the proletariat, as a kind of collective actor, you are asking for trouble. Many of his complaints about an objective teleology are spot-on. "Because it benefits the capitalists" is not an explanation, and does not substitute for a social theory. But that is not the kind of explanation I'm putting forward. I am trying to explain how a CEO who considers himself socially conscious, in charge of a company whose stock holders might be predominantly liberal, can find his firm exploiting people and polluting the environment; or a better example, how a nation of people who are largely socially liberal and when presented with the basic principles of social democracy agree with them, can have and continue to have over many years a state which goes repeatedly against those same principles. I think a theory of collective identity, collective action and collective cognition is necessary to this. I claim that collective outcomes can arise from structural conditions which are just as cognitive as human choices, and I think these outcomes should be compatible with a set of microfoundations for institutions, but they do not require them in order to be useful.

I also want to reject ad hoc classes of collective. Only those entities structurally capable of cognition and action can be classed as collective in my ontology. I have not advanced a theory of classes in this framework because this kind of thinking lends itself poorly to a theory of classes. This really is a key point which distinguishes what I am saying - which really is not that ambitious as social science - from the type of sociological explanation that methodological individualism came into being to oppose.

Wednesday, September 17, 2003
I'm back

I've finally managed to return to normalcy after having the flu all weekend and then taking a Chinese placement test on Monday. I've been unable to find a Dutch class that's more than one night a week, except for classes that run during the day and one class that starts at 5pm three nights a week. I can't fit that into my work schedule, so I'm going to do an intensive taalbad towards the end of the year instead of taking Dutch this semester.

I feel a little guilty about not studying Dutch until December. I guess there's some bit of me that is trying to be a good immigrant. So, this term I'm studying Chinese and Russian at KU Leuven. I have taken both simultaneously once before back in Montreal - it didn't go well. But this time, I don't have any other classes, so I'm hoping it will work out okay. I've managed to test into second year Chinese - which was all that I aspired to do. Since I haven't used Chinese at all in two years and I only know fan2ti3zi4, I'm taking it as a victory. I've taken first year Chinese several times, and I really, really, really, don't want to have to start all over again.

Anyway, I'm going to write a post for A Fistful of Euros, and then I'm going to start responding to comments and posts on my series on language rights, starting with Cosma Shalizi, because he brings up a pair of important points.

I have a chunk of code to debug first - I'm using minimum description length theory to identify terminology from corpora. It's working a good deal better than I could reasonably have hoped for. I'm pretty sure I'm on to something here, and I need to spend a couple hours teasing a good formula out of it. After lunch though, I expect to be writing.