Originally posted: Sun, 27 Jul 2003 22:47:51
Mind, Culture, and Activity 3(6), 1996.
This special issue of the superlative journal Mind, Culture, and Activity isn't a book, of course. But it provides a unitary set of papers around one main issue, and as such illumnates Latour's thoughts in ways that I have found quite useful. In particular, it highlights the advantages of Latour's approach, its disadvantages, and its relations to other approaches such as activity theory.
I say "Latour's approach" here because at this point Latour is no longer referring to his work as actor-network theory. Like John Law (see previous reviews in this blog), he has made a conscious decision to transition to post-ANT thought. Whatever that means. Despite some changes -- for instance, focusing on mediation rather than translation -- what Latour describes here is certainly part of a continuous trajectory that started with his ANT and "pre-ANT" work rather than a radical break. His critics also examine his work in terms of ANT. So I'll refer to Latour's argument as "ANT" or "post-ANT" throughout with the caveat that there may be important differences that I just don't get at this juncture.
The special issue includes Latour's essay "On Interobjectivity" along with commentaries by Michael Lynch, Marc Berg, and Yrjo Engestrom, and concludes with a response by Latour. (The issue also includes an unrelated article and two book reviews; I won't concern myself with these.) Latour's essay essentially tackles the micro-macro or agency-society or actor-network distinction that has concerned him since almost the beginning of his work. He criticizes attempts to separate the two, particularly as those attempts involve divining a macro-level structure that affords a God's-eye view of social interactions. Macro-level "forces," he says, cannot control us any more than we control our mediators.
For instance, he says, consider puppeteers. "If you talk to a puppeteer, then you will find that he is perpetually surprised by his puppets. He makes the puppet do things that cannot be reduced to his action, and which he does not have the skill to do, even provisionally" (p.237; cf. Bakhtin's discussions of dialogism in Dostoyevsky). "Is this fetishism? No, it is simply a recognition that we are exceeded by what we create" (p.237). Similarly, Latour argues, no macro-level force can be said to completely control us or pull our strings. "One can only associate mediators, no one of which, ever, is exactly the cause or the consequence of its associates" (p.237).
This insight leads Latour to a couple of conclusions. One is that skill cannot be attributed to a particular actor -- a statement that, taken to the extremes that Latour takes it, is anathema to activity theory and any other theory of learning and development. (On the other hand, you can see why Latour admires Hutchins' version of distributed cognition.) Another is that it doesn't make sense to talk in terms of levels or social structures -- action, rather, is shared by dispersed actants with various ontologies. Even the action of "summing up" (for instance, creating macro-level diagrams of activity systems) is a set of localized interactions meant to shrink a macro level down to human scale through inscriptions.
The commentaries are all interesting in different ways, but I want to focus on Engestrom's because it gives a lot of insight into both scholars' views. Engestrom praises Latour's essay for its similarity to activity theory, then criticizes it for its differences. In particular, Engestrom criticizes Latour for abandoning the idea of levels, for abandoning "cognition, volition, and emotion" (in short, learning and development applied to individuals), and for not accounting for a systematic durability beyond objects. In particular, Engestrom makes the argument that he has made elsewhere, that in a capitalist market economy the primary contradiction is between use value and exchange value. In short, while praising Latour, Engestrom disagrees with nearly all of the main claims in Latour's essay.
Latour understands this immediately and highlights it in his response. "How am I supposed to deal with this matter-of-fact rendering of everything I reject in my paper? ... Surely, all the differences in scale, timing, agencies that I pointed out in my paper cannot be put on the Procrustean bed of Yrjo's 'embeddedness'!" (p.268). Latour says that he is explicitly shifting away from the Marxist notion of dead labor, the notion of "a human in command, or more exactly, laborers who are empowered again by what has been taken away from them by fetishism and naturalization" (p.267). In short, Latour disagrees with the Marxist (and more generally modernist) tack of backgrounding the world of artifacts so as to foreground human ingenuity. Again, Latour is trying to draw our attention to a symmetrical understanding of the world, one in which humans and nonhumans alike are actants. "I am not trying to naturalize or mechanize humans by turning them into what are held by objects, I try to modify as much the humans who are no longer in command as its associates who are no longer objects, nor means, nor tools" (p.267).
The dialogue is fascinating for what it tells us about ANT (or I guess post-ANT) and AT. Reading Engestrom's other work through this lens, I discover a persistent reference to social structures, to the essential contradiction between use and exchange value as sort of a foundational contradiction in Western society, and to dead labor (although not usually expressed as such). And hand in hand with those is the persistent valorization of individual humans -- something I have done myself and that I make no apologies for, since I believe it was a necessary counterweight to the prevailing managerial models of human-computer interaction against which I was arguing. Latour's argument makes me pause and want to reevaluate AT, which is so rooted in 19th-century German philosophy and so focused on individual and group development. On the other hand, it makes me suspicious of Latour's account that he uses the interconnectedness of networks as an excuse not to account for individual and group development. Surely the sorts of everyday innovations we witness are not simply to be dismissed because of their interconnectedness!
And this gets me, again, to one of the chief complaints I have with ANT. I admire the persistent critique of abstract structures and the undertaking of integrating levels of analysis. I admire Latour's willingness to praise ethnomethodology. But ethnomethodology studies second-by-second interactions (what we might call micro-level interactions) and attempts to connect them to larger trends (what we might call macro-level interactions). ANT appears to do the latter but not the former. Does ANT use the flattening of levels as an excuse not to examine the routine, moment-by-moment interactions that by its own account make up the fabric of actor-networks? If not, then where are these accounts? Until they appear, I don't see why frameworks that do account for these interactions should pull up stakes.
Blogged with Flock