Editor for this issue: <>
cs.umn.edu>Vicki Fromkin says (I think in response to my comment about Harris and 'discovery procedures'): 'The question is not whether the structuralists followed what they preached. In fact all the years I was taught in that framework it was clear that they didn't. Rather -- it is the question of one's particular view of science. Empiricism 'at its roots' starts with the assumption that the only sure basis for knowledge is observation and experiment, that the scientist collects a large body of statements about particular events in the world or the laboratory, that by indcution, makes limited generalizations about classes of events, and proceeds to more general statements if above are verified, and evidence consists to a great extent to the methods used to obtain the generalizations. As Bloomfield stated: "The only useful generalizations about language are inductive generalizations" or Bloch & Trager "The linguist is a scientist whose task is to analyze and classify the facts of speech..." and Hocket: "Linguistics is a classificatory science whose objectives are to find (1) the universif of discourse.. and (2) CRITERIA TO MAKE CLASSIFICATIONS." ' First of all, I don't think that what's at issue here is a hypocritical discrepancy between theory and practice -- at least in Harris's case. Let me draw what I think is an accurate analogy in essential respects. If you look at the way a mathematical logician defines 'proof', the actual working practice of mathematicians doesn't come close to conforming to this definition. At best, what mathematicians give is proof sketches with many details omitted. But you could, in principle, give complete proofs if you wanted to -- what's standardly omitted is stuff that's so routine that anyone reading the proof is going to automatically fill in the gaps. One way of looking at Harris's *Methods* is as an attempt to secure the foundations of linguistic analysis in something like the same way -- providing the canons for rigor that, if not adhered to in actual practice for purely practical reasons, would nonetheless be the final arbiter of what was and wasn't a defensible analysis. Even that may have been too much to ask for, but it doesn't seem to me inherently unreasonable to want to give it a try. A couple of more general comments: I think Vicki is right that the most influential structuralists had a certain view of science that we would now properly reject. Whether 'empiricist' is quite the right word for it I'm not sure. Bloomfield tended to use the term 'mechanist' to describe his own mature philosophical outlook (where by 'mature' I mean 'subsequent to whenever exactly it was that he abandoned Wundtian psychology'). But I think that he in some ways gets an undeservedly bad rap for views that he formed in specific response to then prevalent ideas about language that we from our contemporary perspective would find just as unpalatable as he did. I wonder also if, at least in *Language* he might not have been in some ways deliberately overstating his case -- partly to provoke thought and perhaps partly pour epater les bourgeois. (Cf. Bob King's recent note on Joos.) As long as I'm at it, let me add a note about classification. I certainly would not want to come down on the side that taxonomy is all there is to linguistics (or science generally). But I think that there's a tendency for some contemporary linguists to think that taxonomy itself is somehow necessarily trivial. I think that biologists would find that view very strange. For that matter, the first real triumphs of scientific linguistics -- by anyone's definition, I should think -- were in part of a taxonomic nature: the development of the notion (to which we all still subscribe) of a language family and of the secure methodological footing upon which this notion was placed in the nineteenth century seems as clear a case in point as one could imagine. [End Linguist List, Vol. 2, No. 133]Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue