Matthew Wilken’s generally very good “Canons, Close Reading, and the Evolution of Method” is an interesting reading, not the least because he (to my reading) demonstrates the difficulty of intersection between “critique” and “composition,” in a loosely Latourian way. Or, to put it differently, his contribution is much better than his argument indicates.
As an overall project, the sort of distant reading and large-scale data interpretation he suggests are very interesting and (I think) enlightening. In part, my taste for this article has something to do with my interest in American literature and (particularly) regionalisms, so his examination of place-clusters on 252-253 seem very cool, as does the preoccupation his analysis suggests with international and foreign travel – “pretty diversely outward looking in a way that hasn’t yet received much attention,” as he puts it on 25.
I would quibble with some of his analysis, I think: I’m not sure what he would consider to be “much” attention, in the above passage, given the last few decades of research on American travel/tourism and transnational and literatures of encounters that are at the fore in many areas he’s referring to, or (for that matter) his apparent suggestion that attention to areas outside the U.S. contradicts what he refers to as the “standard model of the period,” that “its fiction is strongly introspective at both the personal and national levels, concerned largely with American identity and belonging” (252-53) – a model that, even where still accepted, bases itself precisely in the amount of anxiety expressed about what it meant to be an American abroad or confronted with those outside areas.
Even with that, however, I think he’s right in looking for broader patterns in these kinds of texts, the kind that become available through data mining. If, after all, the point is to look for patterns in these texts, then I agree that being able to actually do so in a way that can look for increasingly complicated patterns in increasingly large bodies of texts makes sense.
It’s this justification, though, that worries me. I appreciate the candor of his opening: “canons exist, and we should do something about them. The digital humanities offer a potential solution to this problem, but only if we are willing to reconsider our priorities for digital work in ways that emphasize quantitative methods and the large corpora on which they depend” (249). This move, it seems to me, misrepresents what the article does and, moreover, does a disservice to the real work of the article.
The misrepresentation occurs because of conflation, as a look at the opening moves suggest: he begins with the generally acceptable view that, despite the canon wars of 30 years ago, canons still exist. Immediately, however, he qualifies his claim: these are not canons “in the Arnoldian-Bloomian sense of the canon, a single list of great books,” but canons “in the more pluralist sense of books one really needs to have read to take part in the discipline” and the “books many of us teach in common to our own students” (249). It is these canons he takes aim at, canons that are just “adding a Morrison here and subtracting a Dryden there” (249).
Here, I have to separate (at least for my own enjoyment of the article) his project from his goals, because the move he claims as the impetus strikes me as, well, wrong. The problem is that the logical move he makes – “The idea of a Western canon (in, as he says, the Arnoldian sense of moralizing best ideas) was ethically and intellectually problematic; therefore, having canons (in the sense of books that are usually considered necessary to read for a field) is ethically problematical as well” – is not as obvious or unequivocal as he takes it to be. The problem he extrapolates from the second sense – what he refers to as a problem of abundance – was not the problem of the first sense. He writes of this overabundance that “[w]hat little we do read is deeply nonrepresentative of the full field of literary and cultural production” (251). Moreover, he notes, most of the texts we do read are “actively promoted by the major publishing houses and in any case almost exclusively books that have been vetted through commercial publication” (251).
Bluntly, the problem with the Canon (big “C”) was not that it was not sufficiently representational. The problem was that it was both prescriptive and exclusionary in a deliberative way. Even if one accedes completely that taste, aesthetics, etc. are implicated in socially constructed models of dominance by a controlling interest, it is not the same to say “these are different exemplifications of what we study” and “these are the only culturally permissible forms of aesthetic judgment, containing a natural virtue that must be emulated.” The Canon was problematic, essentially, because it didn’t permit the idea of canons – it was totalizing, an absolute on which one could found entire worldviews of taste.
This relationship hits at the heart of where I see his article going two directions. What the above-cited passages suggest are the competing impulses of critique and composition. The most interesting parts of the essay occur when he is generating information. The least interesting is when he situates it. The idea of looking for the place names in texts with an eye to figuring out the mental geographical ecologies of people producing texts is deeply interesting to me, as are the types of incremental change he suggests such data mining would allow us to look to. (I’m currently working with Heather on a somewhat similar project involving western expansion.) That is composition. His argument, however, that the point of doing so is to eliminate canons, or to make the study of literature “truly” representational, are critique: however unintentional, the overarching reasoning here is “there is a unified, compelling, ‘really real’ thing that is literature studies; if we get rid of human subjectivity, we can access it.” It’s not even that I disagree with what he seems to suggest that vision is (historicist cultural studies) or that there can be a correspondence with positivist science/data types without things (the irrational, the unconscious, etc.) getting lost. It’s that this argument doesn’t *need* these kinds of justifications, in my mind. The data he analyzes is perfectly interesting and perhaps even more so without claiming it as the answer to a “still existing” problem that he reinvokes.