That one learns from failure is a well-worn cliché in the business world, but that shouldn’t blind us to the fact that it is true. Or can be. For learning isn’t automatic. You have to work for it.
And sometimes you have to work for a long time. That’s certainly been the case with the string of failures or quasi-failures that have bounded my intellectual life.
Structuralism Can’t Explain “Kubla Khan”
It starts with the spring semester of my senior year at Johns Hopkins. I was in a course on Romantic literature taught by Earl Wasserman and I became interested in Coleridge’s “Kubla Khan” . I wrote a course paper on the poem and then, over the next two-and-a-half years, a Master’s Thesis.
My basic intention was to do a structuralist analysis of the poem, most in the mode of Lévi-Strauss on myth, but also with a bit of Roman Jakobson. It didn’t work out that way. In fact things went so badly that at one point I considered scrapping the thesis, bailing on the Master’s degree, and becoming a musician. It was in the wake of that crisis that I came back to the work, reconceptualized it and wrote a thesis, but one that was very different from I’d imagined I would be doing.
I knew going in that I couldn’t take structuralism off the shelf, apply it to them poem, and the analysis would pop out. I’d have to do a bit of preliminary work of a theoretical kind. The idea was to create the theory that would allow structuralist analysis to yield up the secrets of “Kubla Khan”.
It was the theoretical effort that fell apart. The thesis I ended up writing was mostly descriptive in character: the poem has two parts, each of them has three parts, and so on, mostly the rhyme scheme matches up, but sometimes it doesn’t, and so on. It went on like that for over 70 pages. I liked what I had done. But it wasn’t what I’d set out to do.
It didn’t explain anything, much less the poem’s meaning. It just described a very complicated linguistic object. And the description just lay there.
But, at the same time, that description seemed inescapable. At the time literary texts seemed to be very ‘fuzzy’ objects, their meanings highly variable, mostly as a function of what critics wanted of them. The patterns I described weren’t like that. They were pretty much there, in the text, and didn’t require any interpretive theory. Just look at the language and describe what you see.
Where’d those patterns come from?
Cognitive Networks Can’t Explain “Kubla Khan”
So, thesis in hand, I want off to get a doctorate in English at SUNY Buffalo. I arrived there in the fall of 1973 and showed my thesis around. People were impressed, but couldn’t help me.
In the spring of ’74 Ralph Henry Reese, a fellow graduate student, introduced me to Prof. David Hays in the Linguistics Department. He was a social scientists and a computational linguist and was familiar with that cognitive science stuff I’d been reading just before I left Hopkins.
He too was impressed with the thesis, but didn’t have anything to offer me immediately. However, we liked each other and I became his student and joined his work group. The idea, at least from my side, was that I would learn his semantic theory and then use it to explain that pattern I’d found in “Kubla Khan.”
After about a year or so of work I used Hays’s semantic theory to explain (critical aspects of) the structure of Shakespeare’s Sonnet 129, “The Expense of Spirit”. I chose that sonnet because Roman Jakobson has worked on, figuring my semantic work would thus complement his syntactic, morphological, and phonological work. That resulted in an essay I published in the special Centennial issue of MLN (Modern Language Notes) in 1976 .
I figured that analysis would be preliminary to the analysis of “Kubla Khan”. It turns out that that’s all there was. I just couldn’t produce a plausible account of the semantic underpinning of that poem. Instead I aimed my dissertation at a more sophisticated treatment of the Shakespeare sonnet.
Now, when I set out after an explanation for the structure of “Kubla Khan”, I certainly didn’t have any clear idea of what that explanation would be like. But it WAS clear enough to me that I couldn’t account for it with Hays’s computational semantics. Perhaps when the theory was more fully developed, but not now.
Thus there is a very obvious way in which my doctoral work represented a personal failure: It didn’t achieve what I’d set out to do, which was to explain “Kubla Khan”. But I didn’t experience it as a failure because working with Hays and his other students had been a tremendous intellectual experience. Even if the theory couldn’t be made to do a certain thing I very much wanted to do, it was a very interesting set of ideas, ideas that I continued to explore and develop after I left Buffalo in 1978.
* * * * *
What do we have? In 1969 or so I set out to do a structuralist analysis of “Kuba Khan” and failed. Instead I ended up with a complex description of a kind I’d never seen before. In 1973 I set out to somehow explain that description and failed. Instead I ended up assimilating a wide range of ideas from the cognitive sciences and learning and contributing to a sophisticated computational semantic model.
I was now working in an intellectual universe I hadn’t dreamed of when I set out to college in the fall of 1965 and that pretty much didn’t even exist at the time.
Description Comes Home
Before I left Buffalo Hays and I had coauthored a review of the computational linguistics literature for a journal that was then called Computers and the Humanities . In it we sketched out a thought experiment we called Prospero. Prospero was a computational model of Shakespeare, or at least enough of a mind that it could ‘read’ Shakespeare plays in some ‘interesting’ way. The idea is that, while we can’t get inside a person’s head while they’re reading a text, or attending the theater, we can examine what steps a computer takes as it works its way through, say, Hamlet.
We didn’t specify a delivery date for Prospero. We simply located it in the indefinite future. But I did have a delivery date in mind. I figured that in 20 years or so I’d be working with Prospero.
Well, 20 years from the mid-1970s would have been the mid-1990s. Was I working with Prospero, or something like it, at the time? No. Not even close. In another two decades, actually a bit less, IBM’s Watson would beat humans at Jeopardy. But Watson didn’t have anything remotely like the capabilities required for Prospero. Nor is a Prospero-like system on the current horizon. It just doesn’t compute.
This failure wasn’t has “hard-edged” as the first two. But it was nonetheless a failure. The field – call it cognitive science – didn’t develop as I’d expected it to. For that matter, no one quite expected it to develop in the way that it did. We were all somewhat taken by surprise .
What was going on in the mid-1990s, though, is that literary critics had a long last discovered the cognitive science. But that discovery did not include the work that I’d done two decades ago. And the work they were doing was so very different from what I’d done that it didn’t make sense to think of us as embarked on the same intellectual enterprise despite the fact that we shared a title, “cognition”.
In thinking that over, attending some conferences with these critics, and corresponding with them, I decided that, if they were going to be cognitive critics, then I had to be something else. I’d be a naturalist critic.
But that’s just packaging, branding as they say. Cognitive Poetics and Cognitive Rhetoric are simply two brands of cognitive criticism. And then there is Literary Darwinism, which is the most prominent brand among the bio-cultural critics. So, I’ll be Naturalist.
And, I realized, that the central characteristic of naturalism as I was coming to understand it is an emphasis on description, which is something neither the cognitivists or the bio-culturalists do. Naturalist critics describe texts.
In thinking over the many ways my work differed from theirs I realized that, above and beyond theories and models, I’ve been motivated by descriptions of the formal features of texts. There is/was “Kubla Khan”. But that’s not the only text I’ve published on, let alone worked on. I’d also published on Sir Gawain and the Green Knight , and on Much Ado About Nothing, Othello, The Winter’s Tale, and The Tempest in one article , the Winnebago Trickster tales, Oedipus, Hamlet, and Pride and Prejudice, in another essay . In each case my starting point was descriptive.
What Took You So Long?
Why did it take me so long to realize that? Because I wasn’t trained to think that description of form was important. Yes, literary criticism has something called “formalism”, but it tends to function in opposition to “history”, which is all about context. In literary criticism formalism is a philosophical and methodological doctrine that justified examining texts as autonomous objects unbeholden to authors, readers, or society and history. The careful description of form was and is not within the remit of academic literary criticism.
When I set out to investigate “Kubla Khan” I wasn’t looking for its form. It was looking for the mechanisms of its meaning. That I found its form was, yes, pleasing. But it was also puzzling. I didn’t know what to do with it. First Failure. And so I went off to graduate school in search of semantics. It was in graduate school that I identified the formal features that I subsequently published in those articles [4, 5, and 6]. What I found in graduate school was a semantic model, but it couldn’t handle the task I’d set it. Second Failure. And then Prospero didn’t happen. Prospero would have been an extension of earlier efforts to explain and understand human cognition. Third Failure.
Description is what I had left in the wake of those failures. It’s also where I’d started. I’d created this complex description – of “Kubla Khan” – that I didn’t understand and so wanted to explain it. On the way toward an explanation I found other descriptive problems, each motivating me toward a different kind of explanatory effort. Even if it was explanation that interested and attracted me, it was description that drove me.
I didn’t turn to the newer psychologies out of a dissatisfaction with Theory. I turned to them, initially, to solve a single descriptive problem, one involving one of the most important and most investigated poems in the English language, “Kubla Khan”.
And so, in the last few years, I’ve been theorizing about description  and describing various texts , most of them films, in fact. It’s hard but interesting work. And it’s “hard-edged” in a way that interpretation is not. It makes failure possible and thus opens the way for subsequent learning.
 I’ve been through this personal intellectual history in other somewhat different documents. Touchstones is the oldest of these, and most informal. It get’s us to and through “Kubla Khan”: Touchstones • Strange Encounters • Strange Poems • the beginning of an intellectual life, URL: https://www.academia.edu/9814276/Touchstones_Strange_Encounters_Strange_Poems_the_beginning_of_an_intellectual_life
A series of posts on Lévi-Strauss is more directly concerned with ideas and concepts: Beyond Lévi-Strauss on Myth: Objectification, Computation, and Cognition, URL: https://www.academia.edu/10541585/Beyond_Lévi-Strauss_on_Myth_Objectification_Computation_and_Cognition
Here’s the most recent working paper: On the Poverty of Cognitive Criticism and the Importance of Computation and Form, URL: https://www.academia.edu/15395772/On_the_Poverty_of_Cognitive_Criticism_and_the_Importance_of_Computation_and_Form
If arrives the most sophisticated account of description.
 Cognitive Networks and Literary Semantics. MLN 91: 952-982, 1976, URL: https://www.academia.edu/235111/Cognitive_Networks_and_Literary_Semantics
 William Benzon and David G. Hays, Computational Linguistics and the Humanist. Computers and the Humanities 10: 265 - 274, 1976, URL: https://www.academia.edu/1334653/Computational_Linguistics_and_the_Humanist
 Sir Gawain and the Green Knight and the Semiotics of Ontology. Semiotica 21: 267 - 293, 1977, URL: https://www.academia.edu/238607/Sir_Gawain_the_the_Green_Knight_and_the_Semiotics_of_Ontology
 At the Edge of the Modern, or Why is Prosper Shakespeare's Greatest Creation? Journal of Social and Evolutionary Systems, 21(3), 259-279, 1998, URL: https://www.academia.edu/235334/At_the_Edge_of_the_Modern_or_Why_is_Prospero_Shakespeares_Greatest_Creation
 The Evolution of Narrative and the Self. Journal of Social and Evolutionary Systems, 16(2): 129-155, 1993, URL: https://www.academia.edu/235114/The_Evolution_of_Narrative_and_the_Self
 I’ve posted three working papers theorizing description, all of them available at Academia.edu: https://independent.academia.edu/BillBenzon/
Description as Intellectual Craft in the Study of Literature (2013) 33 pp.
Description 2: The Primacy of the Text (2013) 41 pp.
Description 3: The Primacy of Visualization (2015) 48 pp.
 Beyond the formal papers listed in notes 4, 5, and 6 I’ve three formal papers about Coleridge, two on “Kubla Khan” and one on “This Lime-Tree Bower My Prison” and a large handful of working papers on three literary texts (Heart of Darkness, “The Cat and the Moon”, and “The Road Not Taken”), two manga (Metropolis, Next World, both by Osamu Tezuka), and various films and cartoons, including (Apocalypse Now, Gojira, and Disney’s Fantasia), and President Obama’s eulogy for Clementa Pinckney. All are available at Academia.edu: https://independent.academia.edu/BillBenzon/