We took analytics. . .

Analytics
. . . Polite literature is in the last analysis emotions; and all its charms and message must be spiritually discerned.

The first requisite, therefore, in teaching literature, would seem to be some certain means of reaching and engaging the sensibilities more directly. So far as their exercise is called for in teaching science the method in general use assures it. By it each student acquires not only systematic knowledge concerning the aspects and nature of the phenomena considered,—which is of the understanding, but also experimental knowledge of each quality and process,—which is of the feelings and becomes part of himself.

(Sherman, viii)

Willa Cather wrote a scathing account of her University of Nebraska-Lincoln literature professor Lucius Sherman, who wanted to reduce literature to its irreducible: the word. Sherman wrote a text in 1893 called The Analytics of Literature, where he advocated a mathematical and statistical approach to the study of literature. Such a scientific process, where a student would break the text into its component parts, would be guided by a teacher in a laboratory setting, where the elemental sources of the poem, the words, would be the materials of experimentation. Such a study could only lead the student to understand the sublime emotions that literature inspires. Sherman proposed a few experiments, including analysis of word count and frequency to devise statistical curves to show “force” and “tone quality.”

When I read his preface, I was acutely reminded of a scene in The Princess Bride, where Count Rugen tortures the Dread Pirate Robert in the Pit of Despair. As Westley’s life is sucked away by the count’s Machine, he is directed to “be totally honest on how the Machine makes you feel. . .Remember this is for posterity, so, be honest.” Cather must have felt that her poetic life was sucked out of her by force curves and frequency lists. I wonder what Professor Sherman would do if he could use a concordance program like TXM or build his own program in R. Texts by Shakespeare, Chaucer, and Browning are left, like the Dread Pirate Robert, “mostly dead,” their elemental life sucked out of them.

When Leibniz took a crack at trying to devise a symbolic language for mathematics so that he could one day create a perfect computing device, I wonder if he wanted to reduce all concepts to a symbolic logic, or to a Boolean if/then or true/false construct of concepts. Already, language and literature are in themselves symbolic, pointing to a truth outside of the word or sentence or phrase or line of poetry itself.

If one could peacefully cohabitate with Leibniz, Boole, Gödel, Cather, and Sherman, it’s not going to a battle to the death, but to the pain, and the pain is going to come from building the code that will allow the frequency lists and concordances to reveal something about the beauty or sentiment or sensation of the text. The trick, then, to becoming a “computing humanist” is to embrace both the mathematics of language and poetry and the language of discrete and symbolic mathematics.

I remember crying helplessly as a second grader trying to understand how to borrow when subtracting, and I suspect that all of my math anxiety comes from that moment. I understood that the 1 needed to be borrowed, but every explanation for the construct that allowed for that was dissatisfying and mystifying. What part of number theory did I not understand as an eight year-old? As soon as I got into more conceptual mathematics in trigonometry and calculus, I was much much happier, and started to enjoy the symbolic language of math. Perhaps Cather, francophile that she was, would enjoy the symbolic language of R, the computer language that would allow her to strip a text down to its parts in a few minutes of processing, so that the grunt work of finding “the least common multiple of Hamlet and the greatest common divisor of Macbeth,” could be done by a computer, counting it all out at last, so that the untangling of the “yarn” would be easier and easier.

The essence, though, of the story is still there for the analyst to pick through. Without a sense of the symbolic, the tune of the meaningful, the taste of the poetic, or the perfume of the aesthetic, the results of such calculations and computations are meaningless, and the analyst must set to work. The program cannot interpret, it can only offer up what the programmer asks for. As the women at Bletchly Park had to, so we must, as computing humanists, search for meaning in the data.

Willa_Sibert_Cather_University_of_Nebraska_sophomore_ca_1893 copy
(Willa Cather Archive)

6 thoughts on “We took analytics. . .

  1. . . . They foolishly praised the fabric and the colors. The emperor was very happy, At last a child cried out, “The emperor is naked.” . . .

    • I know that it seems like an empty exercise, creating programs that compute the literary quality of literature, and that it must seem like an obfuscatory move, but I think that the human creates the program, just as the human interprets what the program spits out. And the program will spit out exactly what is asked for. So there is, indeed, a place for human error after all!

      • What constitutes literature? To measure quality requires an exact definition agreeable to everyone. Why is Dostoyevsky considered a giant to the academic community when those of us of the unwashed persuasion find him wanting? What does Twain, Remark and Melville bring to the table that makes them more palatable? What guidelines must be in place to officially declare work to be of literature quality. – By whose standards! The easiest way I know to flunk an English (lit) course is to declare Hemmingway a hack on the final. THAT teacher will tell you you don.t know what you’re talking about – but if you try to pin her down she’d have to depend on hearsay evidence since there is no standard.

  2. I was familiar with the work they do at Gallup with TV ads and such. They bring in a small audience and then have them watch and rate the video in small chunks, each several seconds long. They ask people to measure the affect, either like or disklike, and the indicate the relative intensity of their reaction to each chunk. And then Gallup quantifies the aggreagte, including the emtoional shape of the ad. I thought, wow, one could measure poetry that way, line by line. And I’m sure poems would render into emotional shapes, which could be sorted out into categories. I mentioned it to my graduate advisor as an idea for an independent, interdisciplinary project, but he was utterly silent on the matter. Theme and style would not be in this type of analysis, except as they impacted the emotional reaction created by the piece.

  3. Like you, I was also quite surprised by the similarities between the mathematical endeavors described in the book and our humanistic endeavors. Leibniz’ dream of a symbolic language is, as you have mentioned, perhaps somewhat moot considering the symbolism language already consists of (I’m thinking here of Plato’s three levels of truth). Yet, I was struck by all the mathematician’s apparent sense of aestheticism (yes, for them, in terms of numbers). Perhaps, our endeavors in the humanities are even less restrictive than these great people experienced; after all, if we come up with a theory that has holes, we get to just claim that it only applies to certain aspects of literature (meter only applies to poetry, for example, and even within poetry not to all). Personally, I really felt devastated for Frege having to apparently throw out his entire life’s work…

  4. Courtney, it seems as though we were both getting at roughly the same idea in our posts for this week, and I think you put it quite well (if I do say so myself) when you noted that “[w]ithout a sense of the symbolic, the tune of the meaningful, the taste of the poetic, or the perfume of the aesthetic, the results of such calculations and computations are meaningless.” And not only that, but we were both able to work in some (in my mind) very astute pop-culture references (Indiana Jones and the Princess Bride? What a pairing…)

    I did wonder though, when you say that “texts by Shakespeare, Chaucer, and Browning are left, like the Dread Pirate Robert, ‘mostly dead,’ their elemental life sucked out of them,” was this you simply channeling what Cather would have thought of text-mining? Because it seem that although data (words…that “elemental life” of a text) is “sucked out” in being run through some program, the text, the story itself, still remains untouched and intact, to then be reconciled with the data it has provided. It’s like a dissection that, somehow, allows the poor little frog to still be sitting there happily on the operating table afterwards, croaking away. And in that way, it’s a fairly miraculous thing.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s