Reflection by Devin Griffiths
This is a beautiful and finely-tuned analysis of characterization in Victorian realism. The central premise is that the Victorian novel, in studying how characters sketched on a page might come to “seem human,” anticipated central problems of artificial intelligence, as developed by a cluster of Anglo-American theorists in the 1940s-60s, and, conversely, that theorizations of AI provide formal models that unpack the uncanny dimensions of realist characterization.
I want to say more about the central analogy between the C19 and C20, and how Ward draws out its implications for ongoing debates about historicism, presentism, and what they exclude. But first, I’d like to think about the wider conversations of the book. One of the virtues of Seeming Human is its tight focus, which oscillates between close readings and 20th century computational theory. By the same measure, this approach excludes historical contextualization or a wider ranging consideration of methodological reverberations. How, for example, would an enlarged context distinguish Victorian approaches to character from the extensive discussions of characterization featured in eighteenth-century and Romantic literatures and their criticism? And how does the Victorian novel’s treatment of automatic behaviors and simulated intelligence fit into more extended histories of machinery and computing? Finally, I found myself questioning how the behavioral focus of this study, whether in the novel or in early theorizations of AI, ties in to behaviorism. One of Seeming Human’s most engaging and forceful arguments is around the superficiality of character, the way character is rendered through behaviors and performances that manufacture interiority. It made me ask how theories of behaviorism emerged from the behavioral analyses of nineteenth-century fiction, and also how behaviorism, in mediating the genealogy of the novel, might have conditioned the computer theorists, like Turing, who profited from a behaviorist conception of mind.[1] Behaviorism is interesting, too, because of alternatively agnostic and strict interpretations of what it means for mind: either that mind is inaccessible or simply does not exist. This distinction helps in thinking about genres like historical fiction (which often includes real people and actual communications) or more modern genres like Truman Capote’s “nonfiction novel.” Whereas Seeming Human groups characters with machines, as over against the human, if we include behaviorism, and so also the relation between characters and persons, I wonder if we would still insist that characters do not have an interior, even if that interior can only be reconstructed through surface effects. This is another way of asking how brief references to the posthumanism might be filled out, looking to work by Donna Haraway and others which works to collapse the distinction between the human and nonhuman.
One of the most fruitful if understated conversations in the book is its relation to Caroline Levine’s work on the technicity of form. Like Levine, Ward’s technological formalism fashions a mode of reading form as something ready to hand: portable, palpably active, and (in a qualified way) transhistorical. Which returns me to the question of historicism. One of the most provocative claims of Seeming Human is that current debates about presentism and historicism, initiated by the V21 manifesto, overlook an “historical middle … seldom invoked as a relevant historical interlocutor” (7). Behind the question of which middle, and why, it raises a formative question for all presentist, or transhistorical, or even Anthropocene studies: why that beginning? Why the Victorian novel, rather than the Romantic, or the modernist? The tacit claim (extremely useful, if wrong) of period-based historicism on the nineteenth century was that periods are marked by their integrity: this justified the analogy between text and context, between thin and thick description, between the book and either its “one culture” (George Levine) or the “cultural integrity of [its] empire” (Edward Said). Transhistorical studies tacitly question periodization precisely because they emphasize resonance beyond the period, even as they suggest a higher integrity. Ward takes the expansive totalities of world literature as an inspiration here, citing the way Wai Chee Dimock studies the “resonance of … forms across time” (8).
Such unity has long been a central object of critique in neighboring disciplines, including comparative literature, postcolonial, and diasporic studies. Seeming Human can be read as a reminder that when we draw an analogy between different times, just as when we draw an analogy between discourses or objects within the same time, we have to explain what that analogy makes possible: if one side serves to model, or influence, or reconfigure another; or if their resonance discloses some new phenomenon, some higher unity, and what it means for the history we are telling. Both presentist and transhistorical scholarship need to confront how analogies, drawn across time, will continue to matter when this present becomes someone else’s past.
[1] In Joshua Gang’s account, behaviorism, through its influence on I. A. Richards, was central to the development of close reading and the new criticism; while Heather Love has detailed its influence, by way of Gilbert Ryle and Clifford Geertz, on thick description and new historicism.