12.0346 basic problems in humanities computing?

Humanist Discussion Group (humanist@kcl.ac.uk)
Tue, 19 Jan 1999 21:58:46 +0000 (BST)

Humanist Discussion Group, Vol. 12, No. 346.
Centre for Computing in the Humanities, King's College London

Date: Tue, 19 Jan 1999 21:51:57 +0000
From: Willard McCarty <willard.mccarty@kcl.ac.uk>
Subject: basic problems?

Dear Colleagues:

A close friend and colleague, visiting me recently here in London, suggested
indirectly that one good way of understanding what our field is, in
particular the quality of its intellectual integrity, would be to identify the
basic scholarly problems with which it is engaged or should engage.

A prior problem -- one we would have to deal with but not get forever
stuck on -- is what the nature of these problems would be. I suggest
tentatively that we proceed on the assumption that they are broadly
interdisciplinary and that we not get bothered about their nature until
we have considered a list of candidates.

Let me risk a beginning.

>From my training as a literary critic I would say that a fundamental problem
squarely in our court is to determine the mechanical primitives of textual
analysis -- so that, for example, better text-analysis software can be written
-- though that's only one reason for tackling the problem.

One approach is to ask, in the scholarly analysis of a literary text, what
happens and how much of what happens can be reduced to mechanical terms?
This is at its beginning, it would seem to me, essentially a sociological
approach, and I would guess that it would involve a well-crafted survey of
working scholars. (A number of these have already been done, some by people
who read Humanist; I'd hope they would comment on such an undertaking.)
The danger with the approach is, however, in the tendency to assume that
computing mimics what we already do, that it is purely utilitarian. The
falsity of this assumption would be proven if the sociological project were
thoughtlessly undertaken, software then written and put out into the field,
but it seems to me we can save much grief by prior thought about the
questions we'd want to ask. Wouldn't the trick be to conduct the survey
at a sufficient level of generality and to concentrate on
objectives rather than narrowly on procedures?

Another approach, complementary to the first, might be to imagine
primitives, beyond the ones we now have, that would help literary-critical
analysis along. This would begin with those existing primitives to ask where
the deficiencies lie; what do our colleagues want to do that they cannot now
do with the computer? Again the servey might be useful to deal with the
dissatisfactions, collect the wish-lists and actually examine them
analytically. Then, on the way to imagining the new, we might consider
bridging efforts to bring into the fold of the approachable those techniques
that in fact currently exist but are inaccessible for one reason or
another. I am thinking of statistical techniques, which to the untrained
seem to require a suspension of all disbelief, and work in computer
science, frequently alluring but so often not quite in any form that an
ordinary literary critic can use.

I have the persistent notion that the way to imagine these primitives is
to think about the problem while engaging in literary-critical analysis
-- to become self-aware (in the operational sense), watch oneself work
and project imaginatively beyond the currently possible. In other words,
the best people to do this work are computing humanists.

Are the primitives a closed set? I don't see any way of knowing, especially
because the tools we have affect how we think about our material, and so a
newly imagined and implemented primitive is likely to change the terms of
the question. This is, in other words, a real problem, requiring pure
research. (Oh yes, remember "pure research"? I shudder to think how many
will consider this a quaint notion.)

What, then, would an art historian consider to be a fundamental problem
in humanities computing? A philosopher? An historian? Do the disciplines
as such make any difference to the question, or is the difference made
by the kind of data and the basic questions being asked? By "basic
questions" am I begging the question?



Dr. Willard McCarty
Senior Lecturer, Centre for Computing in the Humanities
King's College London / Strand / London WC2R 2LS
+44 (0)171 873 2784 voice; 873 5081 fax
maui gratia

Humanist Discussion Group
Information at <http://www.kcl.ac.uk/humanities/cch/humanist/>