The argument I often have with people and never win is the one about poetry
as communication. Poetry, someone says, is at bottom, primarily, mainly,
centrally, inevitably and fundamentally a means of communication. Sometimes
they say it's a way of communicating emotions; especially if they think that
poems are a good way of expressing inner feelings with maybe some beautiful
language and some very apt and clever metaphors and images being used to do
it. But that's not what bothers me. It's the "means of communication" bit
that bothers me, because it is a normative statement posing as a descriptive
one, and you can tell this because whenever you hear someone say that poetry
is at its very heart, essentially, without question and above all else a
means of communication, they're usually saying it as a way of telling
someone off for not communicating *properly*. *Obviously* poetry is a means
of communication, so if a poem doesn't communicate it is a *failure*, a
degenerate case, The Thing Which Should Not Be. A variant argument asserts
that even if a poem *appears* not to be very communicative, it still is
communicating really because all poetry is finally and irrevocably and
incontestably a means of communication. You can't not communicate. Even if
you want not to. Even if you think you've managed not to communicate, you
have because you must have because, you see, poetry is a means of
communication.

I don't especially desire a poetry that communicates nothing whatsoever, but
I do resent its communicativity being taken so much for granted. It
pre-empts everything the poem might do to disrupt communications, or to
bring the performativity of certain figures of rhetoric into question; it
means measuring the "success" of the poem in terms of its ability to harness
that performativity, like a fresh microprocessor plugged into the
motherboard of rhetoric. This one passes; this one's defective.

It isn't, looping back to science, that science is purely and wholly
exterior to the social matrix. It is that it *infects* society with an
otherness that sociology is precisely unable to grasp or recognise because
it is concerned only with social inter-actions (which it considers to be the
authoritative Final Context of Everything). Some of the actants in science
are not social beings; some of them are machines, apparatuses, algorithms -
artifacts of human design to be sure, but not everything about them returns
to the father. J. G. Ballard wrote somewhere that in the future our
computational devices will have refined themselves down to the point where
they are invisible to us, capable of entertaining purposes indifferent to
our own and of acting in humanly unimaginable contexts. That is not an
outcome that can be plotted on any graph of the social: it is, however, an
outcome that some human beings are more or less consciously working towards.

Dominic