Date: Sat, 12 Jul 2008 12:29:36 +0100 From: "Humanist Discussion Group \(by way of Willard McCarty <willard.mccarty-AT-mccarty.org.uk>\)" <willard-AT-LISTS.VILLAGE.VIRGINIA.EDU> Subject: 22.114 understanding language (in the right way)? To: <humanist-AT-Princeton.EDU> Humanist Discussion Group, Vol. 22, No. 114. Centre for Computing in the Humanities, King's College London www.kcl.ac.uk/schools/humanities/cch/research/publications/humanist.html www.princeton.edu/humanist/ Submit to: humanist-AT-princeton.edu Date: Sat, 12 Jul 2008 12:23:00 +0100 From: Willard McCarty <willard.mccarty-AT-mccarty.org.uk> Subject: understanding language (in the right way) At the end of their letter to the Editor of the Times Literary Supplement for 4 May 1962, commenting on an article by Yehoshua Bar-Hillel, "The future of machine translation", Karen Spärk Jones and T. R. McKinnon Wood of the Cambridge Language Unit, write as follows about machine translation: >Digital computers with fixed programmes cannot, >[Bar-Hillel] asserts, simulate human procedures >based on an information structure which is >continually changing. The only hope might be to >construct machines which are specifically >adapted to handling self-organizing systems of >this kind. Technological research on such >machines would, however, we maintain, have to be >prefaced by theoretical investigations of >language as a self-organizing system. In most >fields of research theoretical developments have >outstripped technological ones: in this field, >in contrast, this is not the case; technological >backwardness cannot be taken as a reason, or >excuse, for the lack of a general theory of >language…. it might on the contrary be said that >the main reason for the excessive cost of the >machine translation research programmes is that >the use of machines is substituted for thought... > >The point is that it is not that machines can’t >do it but that we can’t understand it >sufficiently to programme machines to do it. > >Any work in this field must start from an >analysis and understanding of language use; and… >anyone starting work in this field almost >immediately comes up against the fact that we >know virtually nothing helpful about it. In >particular, it becomes clear that the results of >traditional linguistic studies, which are >essentially descriptive, are unsuited to our >purpose, and, moreover, that they are often >incomplete even on their own ground. For the >primary requirement, which the analysis of >language use must fulfil, is that it must be >interpretable in terms of mechanizable >operations. This means that it must be concerned >with the way language works and not with a set >of comments on the results of the working. It is >this attempt to obtain a deeper understanding of >the way language works which makes the field >interesting for those doing research in it. >Their approaches may be completely different; >they may be concerned with translation, >abstracting, or retrieval, with the construction >of simplified model languages, or pidgin >languages, or meta-languages, with the analysis >of given texts, or the production of random >texts, with language games, or with learning >programmes; but the understanding of the way in >which we use language is their common aim. My question is this: how well does their statement stand up to what has been accomplished in the last 46 years? Are we any closer to being able to model language as a self-organizing system? It does seem to me that the undimmed lesson is the poverty of our theoretical undestanding of language -- or, to put the matter more generously and accurately, that the main benefit for us in all this is the drive to further and better questioning. But, then, I'm an old-fashioned humanist. Yours, WM Willard McCarty | Professor of Humanities Computing | Centre for Computing in the Humanities | King's College London | http://staff.cch.kcl.ac.uk/~wmccarty/. Et sic in infinitum (Fludd 1617, p. 26).
Display software: ArchTracker © Malgosia Askanas, 2000-2005