<ambiguity type=“two apparently unconnected meanings are given simultaneously” n=“3”>
Most of the ambiguities I have considered here seem to me beautiful.
--William Empson, Seven Types of Ambiguity p. 235.
Coding requires choices. Computers need a level of explicitness and specificity that a human reader is trained to read past. Editing literary works for a digital space requires tagging various elements of a poem, novel, or play with identifying marks that can be read by a machine. Like so many CAPTCHA tests, discourse is difficult to automate and often requires a human interpreter. For me, what’s been most fun about learning to code in this “Year of Code” has been confronting the often hidden, embedded choices that underlie an act of reading. As a literary critic, I am trained to tolerate, even celebrate, certain forms of ambiguity. Ironies. Puns. Metaphors. The flourishing rhetoric of the demagogue. Each intrudes on the discursive space of history with an elegance that can seem, as William Empson wrote, beautiful.
To be beautiful, an ambiguity must be recognized unambiguously.
In this regard, I sometimes compare coding with New Criticism. I love both for the confidence with which they articulate ambiguity. Both attach themselves to literary works and translate their inner tensions for an outside world. (Paradox sits alongside metadata in the peanut gallery of discourse.) Both attempt to stand outside of their object and articulate its parts, its inner workings, its contradictions--contradictions that must be rendered commensurable so as to be reconstituted as the interlocking parts of a functioning mechanism. One judges code like one judges a pudding or a machine. One demands that it work.
Of particular interest to me lately has been the difficulty of incorporating real historical data and references into the metadata of literary works. XML elements like <person>, <place>, and <date> point out from the text into a world of time and space. Gerard Genette theorized these elements as “paratext”--the parts of a book that point out and situate each text within a world of writing. A paratext marks a book’s place in time and space, and therefore records its movements across time and space. Markup encodes these movements anew. Marking up means constantly shifting across the boundaries of the text and the openness of history which must--paradoxically--be inscribed, closed, defined, and controlled in order to be pointed to.
Is there something about a marked up text that is more beautiful than a text stripped of the knowledge of its own context?
The third kind of ambiguity, Empson argues, involves bringing together two seemingly disconnected meanings. He concludes with a reference to Proust’s In Search of Lost Time as an ultimate example. By being in one time, but constantly reaching back and attempting to reconstitute the past, Proust’s narrator tries to straddle time--we become, in Proust’s words, “giants immersed in Time.” Perhaps this is the condition of metadata? A text that steps outside itself?
Empson puts it this way:
Proust, at the end of that great novel, having convinced the reader with the full sophistication of his genius that he is going to produce an apocalypse, brings out with pathetic faith, as a fact of absolute value, that sometimes when you are living in one place you are reminded of living in another place, and this, since you are now apparently living in two places, means that you are outside time, in the only state of beatitude he can imagine. In any one place (atmosphere, mental climate) life is intolerable; in any two it is ecstasy. Is it the number two, one is forced to speculate, which is of this encouraging character? Is to live in n+1 places necessarily more valuable to live in n? (Seven Types of Ambiguity, p. 131)