Wednesday, August 26, 2015

The Poverty of Literary Cognitivism 3: The key to the treasure IS the (form of the) treasure

As I was planning this series of posts I thought I’d follow my personal history by returning to the standard-issue cognitivism in Richardson’s essay-review and say a bit more about it than simply that it fails to deal with computation, and therefore yadda yadda. Then I’d move on to say a bit more about how the analysis of literary form relates to computation without, however, requiring that the analyst spend a couple of years reading the psychological literature before digging in. I’ve decided to reverse that order.

In this post I give three abbreviated examples of descriptive work. I pick up “Kubla Khan” from the previous post and say a few words about how that analysis relates to computation. I follow with some remarks about Heart of Darkness, and offer President Obama’s eulogy for Clementa Pinckney as my third example. I chose that text – despite the fact that it’s not what we ordinarily think of as a literary text, much less being a canonical text – because it has strong formal similarities with the other two texts and because we have a videotape of it and so can examine the ebb and flow of audience response. Despite their different genres, a lyric poem, a narrative, and a sermon, these texts share profound formal similarities. Perhaps those similarities are a clue to nature of the underlying mental processes. I conclude with some general remarks about literary form.

“Kubla Khan” – Three in Three

While the poem has been published with two, three, and four stanzas (as in the appendix I’ve attached), a cursory examination reveals it to consist of only two large movements. The first consists of 36 lines; is set in an exotic Orientalist landscape and has only four personal pronouns. The second, which has only eighteen lines; has no clear physical locus, but consists of a sequence of mental acts (remembering, hypothecating, imagining) and has sixteen personal pronouns. Despite these differences, both movements have the same overall structure.

While discovering these structure trees took care and attention to detail, it wasn’t rocket science. Much of the structure falls out of line-end punctuation, where periods dominate a colon (where present), colons (where present) dominate semicolons, and semicolons dominate commas. That will get you most of the tree structure, especially in the first movement.

Here’s the first movement:

1 tree

Note first of all the red-colored tree branches. The movement consists of three components (which are sometimes printed as separate stanzas); the middle of those in turn consists of three components; and the middle of those, again, has three components. Kubla Khan is introduced into the poem in the first line and his decree dominants the first 12 lines of the poem. The fountain appears in the middle (ll. 17-24) of the middle of the middle of first movement; it provides the agency that dominates that section of the poem. Neither Kubla nor the fountain is directly present in the last six lines of the first movement, which ends with the emblematic line: “A sunny pleasure-dome with caves of ice!”

Now, a cautionary note. The great virtue of such diagrams is that you can grasp them in a single glance and, by scanning the lines and points, examine the structure as though it were outside of time. But that is not how language works. We encounter words one after another. The nested triples that are so obvious in diagrammatic view are invisible word after word, with no word proclaiming, “we have reached the middle now, time to go back.” The green arrow in the following diagram marks the course of actual reading:

1 tree>

As the second movement is only half as long as the first, the tree structure is not so richly developed. But the same nested triples are evident:

2 tree

This movement opens with an imaginary damsel, ends in Paradise and smack in the middle of the middle of the middle we have a single line: “That sunny dome! Those caves of ice!” That is a repetition of the line that ended the first movement, albeit in a different emotional register.

Even the most casual reader will notice that that emblem – the dome and the caves – is repeated from the end of the first movement into the middle of the second movement. What this structural analysis and description reveals is that it is not just more or less in the middle, but that it is in the exact structural middle. Thus we have the propositions about the formal “envelope” of the poem:
1). The first movement ends with an emblem conjoining the pleasure dome and the caves of ice; this emblem reappears in the second movement at its structural center.
2). Structurally, as the fountain is to the first movement, so the emblem is to the second.
3). Structurally, as the emblem is to the first movement, so “drunk the milk of Paradise” is to the second movement.
These are descriptive statements about the formal structure of “Kubla Khan”, but they by no means exhaust that formal structure. As a simple example, when we look at the rhyme scheme we can see that it follows boundaries that are consistent with the tree structures except in the middle sections of each movement. In those sections rhyme patterns cut across those boundaries. These features characterize something that is objectively true of the poem. As such, they must be explained. What is the nature of the mental mechanisms at work? [1]

Computation in the Mind

In some respect, though perhaps not entirely so, those mechanisms are computational [2]. Consider the following simple arithmetic expression:
6 + 9 / 3 = ?
What’s its value? You can’t tell, as the expression is ambiguous as I’ve written it. It could be either of these unambiguous expressions:
(6 + 9) / 3 = ?
6 + (9 / 3) = ?
The first one evaluates to 5 while the second evaluates to 9. By adding parentheses I created structure that told you the order in which to perform the addition and division operations.

What I did in the analysis of “Kubla Khan” in the previous section is, in effect, inserted parentheses into the string of words so as to group it into a nested structure of strings and substrings. It’s not rocket science, but if you pay attention to what happens within those nested strings interesting things emerge. Thus in the first movement of the poem, section 1.1 is visual and spatial in character while section 1.2 is aural and temporal. Each is governed by a presiding agent: Kubla Khan in 1.1; the fountain in 1.2. Section 1.3 has no presiding agent but is both visuo-spatial and auro-temporal in character. Thus we might think of section 1.3 as ‘computing’ the semantic relationships between 1.1 and 1.2. (This kind of analysis is spelled out in considerable detail in the papers in [2].)

Notice the scare quotes around “computing” in that last sentence. If I had a model comparable to the one I’d built for “The expense of spirit” I wouldn’t use scare quotes. Alas, there’s more unspecified work to be done on that score. If you wish, however, you can think of 1.1, 1.2, and 1.3 as being three different mental spaces nested within the larger space of section 1. Here I’m talking of mental spaces in the sense of Fauconnier and Turner’s conceptual blending theory [3].

Now let’s relate those spaces to cognitive networks. The next image depicts some simple cognitive network without all the detail in example from the previous post in this series. Each node indicates some concept and is associated with some word. I’ve traced a path through the network that corresponds to some language string.


It turns out that that string consists of three sub-strings, as follows:


That gives us a tree with three branches, one for each leg of the path through the network:

triple branch

Or a large mental space enclosing three subspaces:

mental spaces

So, while I don’t have an explicit cognitive model for “Kubla Khan”, I do have a way of thinking about the relationship between a certain descriptive analysis of the poem and cognitive models that have seen extensive use in cognitive science. I can relate the structural trees for “Kubla Khan” to both cognitive networks, as an underlying model of more or less permanent mental objects and structures, and to the mental spaces of cognitive linguistics and blending theory, which represent temporary states of activation in the underlying permanent mental objects.

The fact that we don’t know how to develop a model of the mind powerful enough to deal with “Kubla Khan” text is certainly not a good thing. But the deeper principle at work is the ability to distinguish between the mental model (in the form of a network) and the text. The three-segment path, the tree, the mental spaces, those are functions of the text..

But the text is something of an illusion. As ink splotches on a page the text has a more or less permanent existence. But as a mental object, it may not have such existence, not unless it has been memorized. Those language strings and substrings are temporal phenomena, as are those mental spaces. They unfold in time.

Furthermore, as a term of critical art, the meaning of “text” is rather flexible, somewhat unclear. Just as “reading”, in professional usage, elides the difference between mere reading, which everyone does, and explicit and written interpretation, which is generally done only by critics (and their students), so "text" elides the distinction between those ink splotches (or sound waves) and the various mental and social processes that allow them to be meaningful. There are professional occasions where “text” means simply the physical objects, the signifiers. More often, however, “text” means something else or more than that, but it’s never clear just what because we have no explicit way of representing those other things. Cognitive science has given us a way of beginning to sort out the confusion and ambiguity that is built into our ordinary terminology: the text.

The notion of literary form as computational form allows us to systematically explore the relationship between the physical signifiers on the one hand and the mental signifieds on the other. The computational structure depicted in tree diagrams shows how the stream of signifiers is segmented as it is linked to, associated with, rendered into meaningful signifieds.

Heart of Darkness

Now let’s look at a very different kind of text, Joseph Conrad’s Heart of Darkness (HoD). HoD was originally serialized in three installments and is roughly 38,000 words long. It tells of the voyage that a steamboat captain, Charles Marlow, takes up the Congo River to retrieve the agent of an ivory-trading company, Mr. Kurtz, and, more importantly, to retrieve the ivory. Conrad tells the story through a complex double narration in which, long after the event, Marlow tells the story of that voyage to four men on a yacht in the Themes. One of them, unnamed (as are all the characters in the story but for Marlow and Kurtz) tells the story to us.

Forget all that for a moment. Look at the following chart:

HD whole envelope
Paragraph length in Heart of Darkness

Each bar represents a single paragraph of the text and its length is proportional to the number of words in the paragraph. They’re arranged in textual order from left to right. While the distribution is quite spikey you can see it has a roughly pyramidal envelope. [5]

See that longest bar? I started examining that paragraph when I realized that it told of things that hadn’t yet happened in the narrative, which had been told in canonical temporal order up to that point. So I looked at that paragraph, which occurs a bit over half-way through the text, and realized that it is the first time we learn much of anything about Kurtz. Up to that point he had been just a name and an objective. That paragraph contains a capsule history of him.

It seemed awfully long to me. Perhaps it’s the longest paragraph in the text, I thought. Since I had a digital text I decided, on a whim, to use MSWord’s count function to count the number words in each paragraph. It was tedious and took a couple of hours but it was easy enough to do.

Once I’d done it I created that chart to visualize the results. Sure enough, that extraordinary paragraph was the longest one in the text, and by a considerable margin. It was just over 1500 words long while the next two longest paragraphs weren’t even 1200 words long. I certainly wasn’t expecting the shape of the distribution and of course have no good explanation for it.

Now things get really interesting. For this paragraph is framed in an extraordinary way. This happens at the point where the boat is being attacked. Just before Marlow speaks this paragraph he tells us that his helmsman had gotten speared and dropped bleeding to the deck. At that point he breaks off his narration, makes some remarks to the men listening to him on the deck of that yacht in the Themes, and then launches into that capsule history of Kurtz. When he’s finished with the capsule history Marlow returns to the main narrative thread and tells his audience about pushing the dead man off the deck and into the Congo. So Conrad inserts Kurtz’s story into the moments when the helmsman bleeds out on the deck.

Isn’t that an extraordinary thing to do?

Moreover, here’s how he works his way back to the helmsman bleeding out on the deck:
But then, you see, I can't choose. He [Kurtz] won't be forgotten. Whatever he was, he was not common. He had the power to charm or frighten rudimentary souls into an aggravated witch-dance in his honor; he could also fill the small souls of the pilgrims with bitter misgivings: he had one devoted friend at least, and he had conquered one soul in the world that was neither rudimentary nor tainted with self-seeking. No; I can't forget him, though I am not prepared to affirm the fellow was exactly worth the life we lost in getting to him. I missed my late helmsman awfully,—I missed him even while his body was still lying in the pilot-house. Perhaps you will think it passing strange this regret for a savage who was no more account than a grain of sand in a black Sahara. Well, don't you see, he had done something, he had steered; for months I had him at my back—a help—an instrument. It was a kind of partnership. He steered for me—I had to look after him, I worried about his deficiencies, and thus a subtle bond had been created, of which I only became aware when it was suddenly broken. And the intimate profundity of that look he gave me when he received his hurt remains to this day in my memory—like a claim of distant kinship affirmed in a supreme moment.
That is, he’s comparing his helmsman, an African, with Kurtz, who was a remarkably accomplished, if impecunious (we learn later on) European. He decides that Kurtz was not worth the life of that helmsman.

The terms in which he describes the helmsman, in this passage and more extensively elsewhere, are not at all flattering. One can see why Chinua Achebe thought HoD to be racist [6]. What seems to tip Marlow’s evaluation in favor of the helmsman is simply that he’d worked with the man and a bond had formed between them. Kurtz, in contrast, was just a remarkable abstraction.

But that – the nature and meaning of the comparison between Kurtz and the helmsman – though central to the Conrad’s story, is secondary to our discussion here, which is about form. The place in the text where that weighing of one man against another takes place, THAT’s a matter of form.

As is the shape of that distribution of paragraph lengths. This is the longest paragraph in the text and it’s a bit after mid-way as measured by word count. I find it hard to believe that that distribution is either an accident or the result of conscious deliberation (if the latter, no one knows but Conrad). What psychological process produced it? Perhaps more importantly, how does it affect the way we read the text?

For reasons I discuss in my major working paper, Heart of Darkness: Qualitative and Quantitative Analysis on Several Scales [5], I’m pretty sure that this paragraph is structurally central in this text as the eruption of the fountain is structurally central to the first movement of “Kubla Khan” and the appearance of “That sunny dome! Those caves of ice!” is to the second movement. In all three cases we have, when considered abstractly, the same formal pattern: a violent intrusion into an ongoing discourse that happens roughly mid-way in that discourse. And yet we’ve got two very different texts, a romantic lyric on the one hand, and realist but also impressionistic narrative on the other.

Let’s examine a third example of that same abstract pattern.

Eulogy for Clementa Pinckney

President Obama delivered a eulogy for Clementa Pinckney on June 26, 2015, eight days after Pinckney and eight of his parishoners were murdered at a Wednesday night Bible study. Though I heard about the eulogy soon after Obama had delivered it, I didn’t watch it until a few weeks later.

Once I'd watched the video I decided to work on the text and I downloaded a transcript. I quickly saw that “grace” occurred many times in the second half, but not the first, and I sensed that I might be dealing with another of those texts with an intrusive center point. The structural center occurs when Obama introduces “the killer” (he never mentions the man’s name) and this is the only section where he mentions him. That’s also when he first mentions God’s grace.

Here’s the first and last paragraphs of that central section (paragraphs 21-27 in my count [7]):
We do not know whether the killer of Reverend Pinckney and eight others knew all of this history. But he surely sensed the meaning of his violent act. It was an act that drew on a long history of bombs and arson and shots fired at churches, not random, but as a means of control, a way to terrorize and oppress. An act that he imagined would incite fear and recrimination; violence and suspicion. An act that he presumed would deepen divisions that trace back to our nation's original sin. […]

According to the Christian tradition, grace is not earned. Grace is not merited. It's not something we deserve. Rather, grace is the free and benevolent favor of God as manifested in the salvation of sinners and the bestowal of blessings. Grace.
The word “grace” occurs for the first time in paragraph 23 and it occurs ten more times in this central section and then 24 more times in the second half. It occurs nine times in paragraph 47 (actually nine single sentence paragraphs which I treat as one for convenience), the penultimate paragraph:
Clementa Pinckney found that grace.
Cynthia Hurd found that grace.
Susie Jackson found that grace.
Ethel Lance found that grace.
DePayne Middleton-Doctor found that grace.
Tywanza Sanders found that grace.
Daniel L. Simmons, Sr. found that grace.
Sharonda Coleman-Singleton found that grace.
Myra Thompson found that grace.
The eulogy as whole, which is really a sermon in the vernacular of the African-American church, looks roughly like this:
A – Prologue: paragraphs 1-5
B – Pinckney: 6-17
C – Nation: 18-20
Ω – Grace: 21-27
C’ – Nation: 28-39
D’ – Pinckney: 40-44
A’ – Close: 45-48
The whole text is thus symmetrical about that middle section (Ω), which I explain in Obama’s Eulogy for Clementa Pinckney: Technics of Power and Grace [7]. Just as the fountain intrudes into the first movement of “Kubla Khan” and “That sunny dome! Those caves of ice!” into the second movement, and the murder of the helmsman (enclosing a précis of Kurtz’s career) intrudes into the river journey in Heart of Darkness, so this middle section, with its murders but also with its grace, interrupts Obama’s sermon. Three very different texts, one formal structure.

Who’d have thought?

This text was written for oral delivery and we have a record of that performance. We know, in a general way, how that performance moved people, how it worked. That structural center is the first time Obama gets a really strong audience response, with the organist improvising some riffs as well. His strongest response, not surprisingly, comes at the very end, when people rise and the band improvises riffs as he names the victims, one after the other.

What if we were to have people watch the video clip of Obama’s sermon while undergoing a brain scan? Could we then correlate brain activity with the ebb and flow of the sermon itself, with audience response to the sermon, and with the structural turning points we’ve identified through formal analysis?

All Together Now

The kind of symmetry exhibited in Obama’s eulogy is known as ring-form or ring-composition and has been examined extensively by classicists and Biblical scholars, but not so much by other scholars. The most significant exception is the late Mary Douglas, who spent much of the last decade or so of her career working on ring-composition in the Old Testament, but also in more recent texts (Thinking in Circles 2007 [8]). While neither “Kubla Khan” nor Heart of Darkness exhibit the all features Douglas has identified as typifying them (Thinking, pp. 36-37), they both exhibit a number of them, specifically, the features associated with a definite structural midpoint. Moreover, in my own work on ring-composition I have re-interpreted her seven criteria in computational terms [9].

Beyond that, I have identified a number of other texts as exhibiting ring-composition or significant aspects of it: Metropolis (the 1949 manga by Osamu Tezuka), Gojira (the 1954 Japanese film, aka Godzilla), three episodes of Disney’s Fantasia (Nutcracker Suite, Sorcerer’s Apprentice, and Pastoral Symphony), Apocalypse Now and even an essay in PMLA by Alan Liu. Mary Douglas has argued that Tristram Shandy is a ring-composition (Thinking, pp. 85-100), though I’ve not had time to verify her work against the text itself. Ring-composition clearly is not a form that is confined to ancient texts. Moreover, I have been contacted by John Granger, a Harry Potter scholar, who informs me that each of the seven books is a ring and that the series as a whole forms a ring [10]. He has also found rings in a number of other popular texts. While I’ve not examined those texts myself, I have no reason, on the surface, to doubt his claims. The technique is clearly alive, though I have no idea how widely it is distributed.

But this post isn’t primarily about ring-composition. It is about literary form and ring-composition is an example of that form. I chose it because it is not well-known, though some of the texts that exhibit it are quite well-known indeed. How many other ring-form texts are there?

Above all, if we are indifferent to the computational properties of texts, how can we participate in the “cognitive revolution” in a fundamental way, as that revolution was driven by the idea of computation? I understand that many literary critics object to the idea that the mind has a computational aspect and therefore that literary texts have a computational aspect as well. But if a literary cognitivist objects to this, then that cognitivist does not have a very deep understanding of the cognitive sciences. That kind of literary cognitivism is stillborn.

Appendix: Text of “Kubla Khan”

In Xanadu did Kubla Khan
A stately pleasure-dome decree:
Where Alph, the sacred river, ran
Through caverns meaureless to man
     Down to a sunless sea. (5)
So twice five miles of fertile ground
With walls and towers were girdled round:
And there were gardens bright with sinuous rills,
Where blossomed many an incense-bearing tree;
And here were forests ancient as the hills, (10)
Enfolding sunny spots of greenery.

But oh! that deep romantic chasm which slanted
Down the green hill athwart a cedarn cover!
A savage place! as holy and enchanted
As e'er beneath a waning moon was haunted (15)
By woman wailing for her demon lover!
And from this chasm, with ceaseless turmoil seething
As if this earth in fast thick pants were breathing,
A mighty fountain momently was forced:
Amid whose swift half-intermitted burst (20)
Huge fragements vaulted like rebounding hail,
Of chaffy grain beneath the thresher's flail:
And 'mid these dancing rocks at once and ever
It flung up momently the sacred river.
Five miles meandering with a mazy motion (25)
Through wood and dale the sacred river ran,
Then reached the caverns endless to man,
And sank in tumult to a lifeless ocean:
And 'mid this tumult Kubla heard from far
Ancestral voices prophesying war! (30)

     The shadow of the dome of pleasure
     Floated midway on the waves;
     Where was heard the mingled measure
     From the fountain and the caves.
It was a miracle of rare device, (35)
A sunny pleasure-dome with caves of ice!

     A damsel with a dulcimer
     In a vision once I saw:
     It was an Abyssinian maid,
     And on her dulcimer she played, (40)
     Singing of Mount Abora.
     Could I revive within me
     Her symphony and song,
     To such a deep delight ‘twould win me
That with music loud and long, (45)
I would build that dome in air,
That sunny dome! those caves of ice!
And all who heard should see them there,
And all should cry, Beware! Beware!
His flashing eyes, his floating hair! (50)
Weave a circle round him thrice,
And close your eyes with holy dread,
For he on honey-dew hath fed,
And drunk the milk of Paradise. (54)


[1] The following working paper says more about “Kubla Khan”, compares it to a very different, but related, poem, “This Lime-Tree Bower My Prison”:

STC, Poetic Form, and a Glimpse of the Mind. Working Paper, November 2013, 45 pp. Download:

I have full-dress accounts of both poems as well:

“Kubla Khan” and the Embodied Mind, PsyArt: A Hyperlink Journal for the Psychological Study of the Arts, November 29, 2003. Download:

Talking with Nature in "This Lime-Tree Bower My Prison." PsyArt: A Hyperlink Journal for the Psychological Study of the Arts, November, 2004. Download:

[2] This post has some useful remarks and is relatively short: Three Notes on Literature, Form, and Computation. September 10, 2013.

For the full-dress treatment, Literary Morphology: Nine Propositions in a Naturalist Theory of Form. PsyArt: An Online Journal for the Psychological Study of the Arts, August 2005, Article 060608. Download:

[3] Fauconnier, G. and M. Turner (2002). The Way We Think. New York, Basic Books.

[4] On the Poverty of Literary Cognitivism 2: What I Learned When I Walked off the Cliff of Cognitivism, URL:

[5] I’ve done quite a bit of work on Heart of Darkness and written it up in three working papers:

Heart of Darkness: Qualitative and Quantitative Analysis on Several Scales (2011) 48 pp. Download:

Paragraph Length in Heart of Darkness: Some Basic Numbers and Charts (2011) 6 pp. Download:

Periodicity in Heart of Darkness: A Working Paper (2011) 5 pp. Download:

[6] Achebe, Chinua. “An Image of Africa: Racism in Conrad's 'Heart of Darkness'” Massachusetts Review. 18. 1977. Online:

[7] You can find my detailed work on this text in two working papers, one is expository while the other contains analytic tables:

Obama’s Eulogy for Clementa Pinckney: Technics of Power and Grace (2015) 42 pp. Download:

President Obama’s Eulogy for Clementa Pinckney, Analytic and Descriptive Tables (2015) 14 pp. Download:

[8] Mary Douglas, Thinking in Circles: An Essay on Ring Composition. Yale UP. 2007.

[9] Pages 39-41, Ring Composition: Some Notes on a Particular Literary Morphology (2014) 59 pp. Download:

[10] Granger has a website, The Hogwarts Professor. Here is a link to his entries on “ring-composition”:

No comments:

Post a Comment