Sunday, September 1, 2013

De Man, Fish, and Simulation

This is the third post from my tryout at The Valve. It appeared there in mid-December of 2005. The two footnotes, however, are new. That conduit metaphor I mention in the third paragraph has come to play a major role in my critique of memetics, such as this post from 2011 and this more recent one, in my reflections prompted by Dan Dennett.
Though de Man was at Hopkins when I was there, I never studied with him, nor even read him. However, prompted by this and that, I decided to pick up Blindness and Insight and read around. The first (though not quite the only) essay I read is “Form and Intent in the American New Criticism.” Consider this passage (p. 25 in Blindness and Insight, 2nd Edition, Revised):
“Intent” is seen, by analogy with a physical model, as a transfer of a psychic or mental content that exists in the mind of the poet to the mind of a reader, somewhat as one would pour wine from a jar into a glass. A certain content has to be transferred elsewhere, and the energy necessary to effect the transfer has to come from an outside source called intention.
This much is what cognitive linguists would later identify as the conduit metaphor. The key text is Michael Reddy's “The Conduit Metaphor: A Case of Frame Conflict in Our Language about Language” (in Andrew Ortony, ed. Metaphor and Thought, 2nd edition, pp. 164-201). Reddy's article is based on 53 examples sentences. Here are the first three (p. 166):
(1) Try to get your thoughts across better

(2) None of Mary's feelings came through to me with any clarity

(3) You still haven't given me any idea of what you mean
Reddy's argument is that many of our statements about communication seemed to be based on the notion of sending something (the thought, idea, feeling) through a conduit, hence he calls it the conduit metaphor. He knows that communication doesn't work that way, but that's not is central issue. His central concern is to detail the way we use the conduit metaphor to structure our thinking about communication.

Thus both de Man and Reddy are concerned about how we think about communication, but Reddy is considerably more interested in the linguistic variations and details of a certain conception. While I would like to continue with de Man, I think it would be better to turn to Stanley Fish. We can pick some more de Man later if that seems useful.

In his seminal essay on "Literature in the Reader" Stanley Fish made a general point that the pattern of expectations, some satisfied and some not, which is set up in the process of reading literary texts is essential to the meaning of those texts. Hence any adequate analytic method must describe that essentially temporal pattern. Of his own method, Fish asserts (Is There a Text in This Class?, p. 28):
Essentially what the method does is slow down the reading experience so that “events” one does not notice in normal time, but which do occur, are brought before our analytical attentions. It is as if a slow motion camera with an automatic stop action effect were recording our linguistic experiences and presenting them to us for viewing. Of course the value of such a procedure is predicated on the idea of meaning as an event, something that is happening between words and in the reader's mind . . .
A bit further on Fish asserts (p. 32) that "What is required, then, is a method, a machine if you will, which in its operation makes observable, or at least accessible, what goes on below the level of self-conscious response."

In a slightly later essay (1973), “What Is Stylistics and Why Are They Saying Such Terrible Things About It?” Fish takes on an article by the linguist Michael Halliday, remarking that Halliday has a considerable conceptual apparatus - an attribute of many modern linguistic theories, lots of categories and relationships, all tightly defined. After quoting a passage in which Halliday analysis a single sentence from Through the Looking Glass, Fish remarks (p. 80):
When a text is run through Halliday's machine, it's parts are first dissembled, then labeled, and finally recombined in their original form. The procedure is a complicated one, and it requires many operations, but the critic who performs them has finally done nothing at all.
Now, though I am familiar with some of Halliday's work, I've not yet read that particular essay. Still, Fish's characterization seems fair, and would apply to many similar and even not-so-similar models. Note, however, that he frames Halliday's essay as one of many lured on by “the promise of an automatic interpretive procedure” (p. 78). That may or may not be a fair characterization of Halliday, but . . . .

Let us take Fish's call for a machine literally and imagine what literary study might be like if we had such a machine. The machine I am imagining would take the form of a computer programmed to simulate the human mind in sufficient detail so as to be capable of reading literary texts. Over the last 30 years or so AI folks have been saying such a machine is just around the corner and they no doubt will continue in that hope for the foreseeable future. More to the point, however, is that at the time de Man and Fish were writing the essays I quoted above, new work was being done in the computational simulation of language processes. A variety of techniques were developed for representing semantic information in computational form and these were coupled to parsers so that computers could answer questions about and construct paraphrases of simple natural language texts. Thus, while the prospect of a full and rich simulation of the human mind is ever distant, it is no longer unthinkable. It, or something like, can be and has been done in the small.

But let us fantasize about doing it in the large. What could we have it do? Well, since it can simulate us, we can have it do anything we can do. We can even create a community of them and have them do anything a bunch of us can do. We can have them all read the same text and argue about its meaning, adopting whatever theoretical perspectives they wish.

This takes us right back to Fish's remark about Halliday's system, that it takes texts apart and reassembles them, but otherwise does nothing at all. And so it is with our community of artificial critics. They simply reproduce us, in electronic guise. They do not crank out interpretations any more automatically and reliably and objectively than we do.

So what would be the point of building such an artificial community? The point is in the knowledge we would gain in constructing it and through observing its behavior. We cannot observe what happens in human brains as they read texts, chat with others about them, and even write learned commentary on them. But we could observe the computers doing these things, and thereby learn something about how texts work in our minds.

As we cannot do this, however, what is the point of entertaining the fantasy? Because it opens up, points the way to, a conceptual space in which we can do something, now.* Back in the 1970s I thought we could move into that space by constructing the appropriate models, and did considerable work to that end in my 1978 dissertation, Cognitive Science and Literary Theory. Since then it has become obvious to me that theories and models aren't enough. We need methods of practical analysis that are, in some sense, in advance of any models or theories we might have.** We must learn to discern patterns in texts for which we do not have any explanation, which do not seem to have any obvious bearing on meaning, but which are, nonetheless, there in the texts.

Of course, we can already do this, especially with poetry, where we've got meter, and rhyme, and other stuff. We have ways of describing those things, and theories that even say they are important. But the study of that “stuff” is peripheral to literary studies, even for those whose theories lead them to regard it as central rather than merely ornamental. The problem is we don't know how, in any systematic way, to relate those phenomena to meaning.

Perhaps we should bracket meaning for awhile. Or, at best, regard it warily and from a distance.

To be continued . . . .

* * * * *

* At the time I wrote this post the "digital humanities" weren't so visible as they now are, though that work was well under way by that time. That work is certainly in this conceptual space, though its rather different from what I had in mind when I wrote this piece.

** That is to say, description, which has become something of a hobby horse.

No comments:

Post a Comment