iA


Communism, bone and polished stone: the poetry of corpus linguistics

by Toma Tasovac. Average Reading Time: almost 3 minutes.

Poetry, too, is in the eye of the beholder. A street sign reading “Universe closed, use rainbow”  is too brilliant to be written off as something belonging only to the domain of traffic control. Not that traffic control is not poetic in its own right. After all, traffic control has given us not only traffic jam (“a marmalade made for the road?”) and traffic jams (“busy knee-length shorts worn by surfers?”) but also lollipop ladies (“candy-obsessed women of a certain generation?”). Poetry is everywhere.

I came to electronic lexicography and digital humanities via literary theory. That means that I have been trained (or brain-washed, if you believe in compounds of ideologic and hygienic persuasion) to blur the lines, not of sense and sensibility, but sense and senselessness. It also means that I have been drilled to see every text as inherently opaque, to argue that the text doesn’t really mean what it says.

Nowadays, when I am not actively pursuing topics like “Tsvetaeva’s spaces of (non)encounter” or “Architectonic discourse in Russian ornamental prose,” I can look for Jakobson’s poetic function in supposedly less likely places like dictionaries and databases. They may be lacking in terms of authorial intention (Eco’s intentio auctoris), but nobody can deprive me of my own intentio lectoris, let alone cleanse language of its own intentio operis. 

So, could linguistic corpora — in a semi-automatic homage to both Oulipo and Katherine Hayles (whose mother apparently was a computer) — be used as poetry machines?

Certainly. Here is one way: query a corpus and observe possible / interesting / defamiliarizing patterns of text to the right of the keyword in each line returned. In other words: read the right context of the concordance vertically by connecting parts of sentences from different sources that have been brought into proximity by the random capriciousness of the corpus at your disposal.

To make things more difficult in my first experiment, I decided I couldn’t reappropriate  just any part of the line: I could drop words from the end of the line, but I could claim a line only if I used the words immediately following the keyword. I could not change the order of the lines, and I gave each concordance line I took a separate line in the poem. I did, however, allowe myself to play with punctuation.

The query term was “weapons of” filtering out “destruction” because it was too obvious (though world-destruction did sneak in). The corpus: BNC. The raw data looked like this.

The result is a poem which could be about many things: communism, American imperialism, the London riots (or is it OWS?), the hobbits and — maybe, just maybe — tantalizing, forbidden, love. Not exactly Nobel-Prize-worthy, but then again: what is?

Weapons of…

Communism,
bone and polished stone.

Her own Britain would be no more than a temporary American
last resort.

A thousand angry people:
invasion and occupation,
obliquity and understatement. Parents and children stretch
world-destruction, Marty McFly returns to discover the mid-Eighties,
“The Thing”,
her own,
their own. One day America may find itself facing the last
morality:
an Emasotsha warrior.

Repression, reprisal and terrorism: natives are divided and
attack at football matches.
Their superior knowledge and direct executive
revolution are obscenity, blasphemy and
terror (Lidice, Oradour-sur-Glane).

British climbing’s hypothetical assassins?
Power and devices of infinite cunning for
the Elves and the peaceful sorceries of
his Kingdom were not those normally used by
past armies for displays in a popular pursuit.

Peter the Great’s military machine: muskets and
sexual mortification that she knew he would never describe.

One comment on ‘Communism, bone and polished stone: the poetry of corpus linguistics’

Leave a Reply