File /Humanist.vol22.txt, message 110

Date: Tue, 08 Jul 2008 13:36:36 +0100
From: "Humanist Discussion Group \(by way of Willard McCarty              <>\)" <willard-AT-LISTS.VILLAGE.VIRGINIA.EDU>
Subject: 22.108 a frustrated becoming?
To: <humanist-AT-Princeton.EDU>

               Humanist Discussion Group, Vol. 22, No. 108.
       Centre for Computing in the Humanities, King's College London
                     Submit to:

         Date: Tue, 08 Jul 2008 13:29:00 +0100
         From: Willard McCarty <>
         Subject: a frustrated becoming?

In the Times Literary Supplement for 28 October 1983, in a review of 
a book by Geoff Simons (then Chief Editor of the National Computing 
Centre, Manchester), Are Computers Alive? Evolution and New Life 
Forms (Harvester, 1982), H. C. Longuet-Higgins gives due attention to 
the authors' quasi-religious evangelism, e.g.

 >Are computers alive? Yes! And today they truly
 >represent an emerging family of living species
 >in the world -- *that* is the startling argument
 >of this landmark book. . . How will we relate to
 >living machines? *We need to find the answer
 >soon*.... Machines are evolving limbs, senses,
 >brains, cognitive faculties, emotion, free will,
 >and the capacity for reproduction. A machine
 >capable of self-reproduction, of sensing the
 >changing world and of taking appropriate
 >adaptive action -- *must surely be regarded as alive*." [authorial italics]

The studied ignorance of current fact and violent leap of faith 
required to believe in such a pronouncement are especially in these 
quieter times easier to dismiss as evidence of self-delusion if not 
madness, but let us not do that. Let's put these and other such 
ravings on hold for now.

In sober moments, Longuet-Higgins notes, Simons does manifest the 
"enhanced appreciation of nature's own information technology [that] 
is one of the first signs of recovery from computer mania. Indeed", 
the reviewer goes on to note, "it could be argued that IT's most 
enduring contribution to our culture will have been the way in which 
it has forced the psychologist to look for computational accounts of 
the way our own minds and senses work." (I object to the use of 
"computational" here as a projection from the machine onto human 
cognition and perception, but never mind that for now.) 
Longuet-Higgins cites the example of automatic speech recognition. As 
a result of attempts to build systems to do this,

 >we find that existing knowledge is quite
 >inadequate for the purpose, suggesting that we
 >have been failing to ask the most important
 >questions -- those having to do with the
 >computations which the human auditory system
 >must carry out in order to unscramble an audible
 >signal into a sequence of English words -- much
 >as the code-breakers of Blechley Park during the
 >war used computers for decoding enemy
 >messages.... In fact, the more one discovers
 >about the acoustic intractibility of real
 >speech, the more respect one acquires for the
 >effortless ability of human beings to understand
 >it. It is by no means safe to assume that it is
 >only a matter of time before man-made systems
 >put our own mental and perceptual faculties to
 >shame.... it was such a claim, made nearly twenty
 >years ago, that led to the most humiliating
 >failure yet associated with the enterprise of artificial intelligence.

The reference is, of course, to the great machine translation project 
on which the whistle was blown by the Automatic Language Processing 
Advisory Committee (ALPAC) report, "Language and Machines: Computers 
in Translation and Linguistics" (1966). As Yorick Wilkes has said, it 
had been obvious for some time that the model of language on which 
work had proceeded was  simple-minded and that a much more 
sophisticated one was required. It's significant that ALPAC 
recommended further funding be put into linguistics research rather 
than into the building of translation= systems.

Longuet-Higgins thus gives us a recognizable, quite comforting 
argument, and one that should be familiar to everyone by now: the 
main point of computing cultural artefacts, such as speech or written 
text, is to isolate what cannot be computed, revealing to us our own 
ignorance of something quite specific, and so pushing us to know more 
and better. Fine. But what about the evangelical rant of Manchester's 
National Computing Centre's Chief Editor? What do we make of that?

If we look at what is being done with computing in the humanities 
today, 25 years later, we find, apart from construction of scholarly 
resources of many different kinds, (a) analytic investigation of 
artefacts, such as literary texts by means of computational stylistic 
techniques, simple concording, textual markup etc.; and (b) synthetic 
reconstructions blurring into VR constructive imaginings of possible 
worlds. The former is achieved by keeping one's distance and 
respecting difference between computing and whatever the reality is 
taken to be -- i.e. Longuet-Higgins' argument. The second, however, 
comes about when one forgets about the difference and, as it were, 
looks with or from within the computational environment to what it 
makes possible to render. (This, I suspect, is very well known to 
game-players.) Could it be that Simons' anthropomorphic projections 
and related visions of the time, such Arthur C. Clarke's HAL, in 
2001: A Space Odyssey, had something to do with the batch-orientated 
form which computing then took? Could it be that a frustrated desire 
to identify was responsible?



Willard McCarty | Professor of Humanities
Computing | Centre for Computing in the
Humanities | King's College London | Et sic in infinitum (Fludd 1617, p. 26).


Humanist Main Page


Display software: ArchTracker © Malgosia Askanas, 2000-2005