Ode to a Keatsian Turn: Creating Meaningful and Poetic Instances of Rhetorical Forms - Computational Creativity Research: Towards Creative Machines - page 193

and the systematic manipulation of words can be an effective means of navigating the
corresponding conceptual spaces (recall Goethe's maxim that “ words are often most
useful precisely when ideas fail us ”). By searching for opportunities for chiasmus, the
Sphinx is actually employing a simple form of dialectical reasoning. Given a thesis,
he fabricates its structural antithesis, and then uses chiasmus to forge a synthesis of
the two. The Sphinx is no Hegel, and he is certainly no Kant, but we must assume that
he applies some aesthetic and semantic filters to his formulations. For he does not
invert everything, but chooses to selectively invert theses whose antitheses appear
structurally and semantically sound. A computer that modeled the generative abilities
of the Sphinx would almost certainly be accused of mere generation. Yet its creator
might validly reply, Sphinx-like, “ Not necessarily. ”
So what might distinguish a computer's best efforts at chiasmus from those of
the Sphinx? Well, it would certainly help if it could display an appreciation of the
different shades of meaning carried by related forms of the same word-concept.
Consider Mae West's chiastic innuendo “ It's not the men in your life that counts,
but the life in your men. ” Mae uses deliberate equivocation here, by employing the
word “life” in two different senses—“life” as in personal life, and “life” as in zest
and vigor. Between these two senses, Mae stakes out a third sense, her sex “life” .
Equivocation like this is a form of trickery that often produces humor. Consider
another humorous example of chiastic equivocation: “ Children in the back seats of
cars sometimes cause accidents. Accidents in the back seats of cars sometimes cause
children. ” This is more than syntactic manipulation for its own sake. The repeated use
of “accidents” in two different senses—car accidents and accidental pregnancies—
produces a pithy commentary on life's surprises, and gives the impression that the
speaker has peeked behind the curtain of everyday language to glimpse a universal
truth. Each of these examples relies on word play, but each also evokes an unspoken
meaning that chimes with our experience of the world.
A computer can easily be programmed to scour a large text corpus for reversible
chunks of language such as “ hardly working ” and “ working hard ”, so as to generate
countless examples of chiasmus in the egregious vein of “ working hard or hardly
working? ”. Yet this would surely be a poor investment of anyone's time. Even a
more semantics-savvy generator, one capable of producing the political aphorism
“ for society to prosper, prosperity should be socialized ” from the independent text
chunks “ society to prosper ”, “ prosperity should' ' and “ be socialized ”, is hardly
worth the effort if all it can do is generate one instance of chiasmus after another.
As Truman Capote once said of Jack Kerouac on hearing of the latter's frenetic
stream-of-consciousness writing method, “ that's not writing, that's typing .” Even if
such a system could generate instances of chiasmus of a quality deemed usable by a
professional comedian, no professional would ever craft a whole act around a single
rhetorical device. Chiasmus, like other conduits for linguistic creativity, should not
be viewed as a party-trick. It should not be generated in bulk, nor sold by the yard.
Party tricks are the province of chumps like the Sphinx, in whose hands they are glib
generators of fakes rather than vehicles of self-expression.
Computer scientists are trained to embrace modularity, so it's tempting to imag-
ine how a standalone chiasmus generator might later unite with generators of other