Respect for the significance of his work and his diligence in reading widely and deeply in literary theory must temper any impulse to launch a full-scale critique of Herbert Simon's essay. And yet what other response can one make to Simon's monolithic vision of cognitive science as a unified paradigm capable of bringing harmony to the warring schools of literary criticism? The problem here is not his criticism of criticism: there are certainly many equally normative accounts generated by literary critics. The difficulty lies, rather, with the problem of the science itself, with the now wholly indefensible claim that one version of cognitive science-the AI symbol processing model-is somehow the only possible account of human cognition. What may have once been true in the grand old days of AI, in say 1969 when Simon first published The Sciences of the Artificial, is no longer the case.
Indeed, if one looks closely at the debates in cognitive science and philosophy today, one would be hard pressed to distinguish them from the debates in critical theory. Fundamentally, they are different versions of the same debate, a debate whose object is not so much literature or cognition per se but rather the whole discourse of discourse. So perhaps the best way to stage a critique is to make a small and respectful deconstructive turn on Simon's own text and show exactly where it contradicts itself. In doing so, one would be demonstrating that accounts of cognitive science are themselves caught in the economy of textuality and hence subvert the very theory expounded in them.
Near the end of his essay, Simon remarks on the endless critical battles that go on in literary criticism and notes that
[E]ach school appears to describe some particular mode of evocation, hence of meaning, and then to claim it as the correct one. If claims of uniqueness and exclusive correctness were abandoned, as they surely must be, peaceful coexistence could be wholly restored. (24)
The Olympian tone of this pronouncement obscures its contradictory relation to itself: the claim can only be made by someone who, in fact, is operating from a stance of "exclusive correctness." In Simon's case the vehicle of exclusion is omission. Nowhere in Simon's piece will a literary critic untutored in cognitive theory find any indication that a great debate is presently taking place regarding the nature of information processing in the brain and the coherence of the notion of the symbol as the fundamental unit of mental representation. And nowhere, in a paper first delivered
in 1990 and then subjected to "extensive revision," does one find acknowledgement that at the very moment of delivery, there already existed a rival paradigm which directly challenged the orthodoxy of the AI symbol-processing approach. Significantly enough, that paradigm, most broadly called simply connectionism, stands almost exactly in relation to classical AI as deconstruction once did to structuralism.
What one finds instead in Simon is a kind of disembodied dogma cloaked in the voice of passive agency so characteristic of official science:
The confidence with which, in the previous sections of this paper, I have spoken of the structure of human memory as resembling an indexed encyclopedia, or of the process of evoking meanings from memory is based on the fact that just such processes have been embodied in computer simulations, and the behavior of the resulting programs in the face of varied tasks has been found to resemble closely human behavior in the same tasks. (8)
What epistemic status can a phrase like "found to resemble closely human behavior" possibly have? Who found? By what standards? One could parse that pronouncement, as Dreyfus long ago did, by simply saying that the AI community found that a program designed to reify all of the classical assumptions about symbol processing operates in a way perfectly consistent with those assumptons.
Yet if one searches diligently through Simon's text, one can find, buried in a parenthetical aside, one tiny trace, one small hint of dissent in the supposed utopia of cognitive philosophy. Just as Simon starts to suggest that another AI program, ISAAC, can somehow reason and understand, he pulls back and wants us to
pass over such philosophical issues as whether ISAAC understands what it is doing. (I think it does; John Searle thinks it doesn't. But no matter.) (23)
No matter? Of course not; it's "only" a matter of a small difference of opinion. One can ignore the fact that Searle has a different way of modeling propositional thought, that Searle has written extensive critiques of the AI project, and that fundamental parts of the knowledge-representation paradigm which underlie AI have long, complex histories of models and counter models and long-standing conflicts over the difference between a syntactic and a semantic theory of truth. Perhaps the parenthesis is a trace of the presence of Searle at the original lecture. If so, how would the essay have read if any of Simon's other peers had been present: Crick, Smolensky, Fodor, Kripke, or David Lewis? No matter; what appears now by way of "extensive revision" is intended only for literary critics. One cannot be bothered with those technical philosophical questions.
At this point, near the limit of a brief commentary, one is tempted to use this reference to Searle as an opening through which the whole text of Derrida's Limited Inc. could be inserted into Simon's essay. Then the reader, upon reaching that point where it doesn't matter anymore, could watch with amusement as Derrida slays the Searle monster, leaving Simon to face, alone, the reign of deconstructive terror. One would then recommend to Simon that he navigate back through the virtual realm of intertexuality and closely read Derrida's Speech and Phenomenon, considering particularly the argument that Husserl's distinction between indication and expression cannot be sustained. That demonstration does matter because it marks the beginning of the end of the hegemony of the symbol and the notion that consciousness can model itself by separating out propositions from attitudes. And it matters fundamentally, because that demonstration, in a text which "found" that no account of mental processing is consistent with itself, undercuts both Simon's cognitive theory and his literary musings. Perhaps then Simon might at least consider the possibility that the discourses of these two versions of culture are of a piece, each one an alibi for the other.