[Lingtyp] discovery procedures vs. diagnostics

Thu Dec 24 18:30:49 UTC 2020

Dear Adam,

Yes, there is lots of textual support. If you want to see discovery
procedures carried to an extreme, then read anything by Zellig Harris, for
example:

Harris, Zellig S. (1955). 'From Phoneme to Morpheme', Language 31: 190-222.
Harris, Zellig S. (1946). 'From Morpheme to Utterance', Language 22: 161-83.

Harris's book wasn't (originally) entitled '*METHODS in Structural
Linguistics*' for nothing. The whole book is devoted to discovery
procedures.

For the most empiricist of the descriptivists, the goal was to arrive at a
grammar of a language by performing a set of operations on a corpus of
data, each successive operation being one step farther removed from the
corpus. It followed then that the levels of a grammatical description had
to be arrived at in the order: first, phonemics, then morphemics, then
syntax, then discourse: ‘There is no circularity; no grammatical fact of
any kind is used in making phonological analysis’ (Hockett 1942: 20).
Martin Joos's comments after many of the papers in his edited volume *Readings
in Linguistics* (1957) will give you a flavour of discovery procedures. He
approves of Bernard Bloch's 'Phonemic Overlapping' article, even though he
himself recognized that Bloch's analysis was counterintuitive. But Bloch
had to discover the phonemes directly from the phones without using any
other sort of information, so his final analysis was weird.

There were less extreme structuralists in the US in this time period. Those
with a Prague School background, like Jakobson, of course. And Kenneth Pike
argued -- to the scorn of many of his colleagues -- that you had to use
grammatical information in a phonemic analysis. But even Pike did not
really break from operationalism, as he proposed discovery procedures to
'find' the grammatical information.

I think that in *Chomsky's Current Issues in Linguistic Theor*y (1964), you
will find a number of references to post-Bloomfieldian discovery procedures.

Fritz

Frederick J. Newmeyer
Professor Emeritus, University of Washington
Adjunct Professor, U of British Columbia and Simon Fraser U

On Thu, 24 Dec 2020 at 09:50, TALLMAN Adam <Adam.TALLMAN at cnrs.fr> wrote:

> Dear Bill (if I may),
>
> Thanks for your response and directing us to the passage where this issue
> is addressed. In fact, I see Seuren's response as a way of evading the
> methodological problems you highlight by conflating discovery procedure
> with diagnostic or criterion.
>
> Dear Fritz (if I may),
>
> Is there any obvious textual support for Chomsky's interpretation? I have
> not found any, but perhaps I am misreading things with a hindsight vision.
> After reading Longacre's book, I feel that the tagmemicist notion of
> 'discovery procedure' is very similar to the way 'diagnostic' or
> 'criterion' is typically used today (e.g. affix diagnostic, word
> diagnostic) - not just in generative literature, but in Dixon's
> presentation of 'Basic Linguistic Theory', for instance. Perhaps the notion
> articulated by Chomsky was an idea 'in the air' so to speak and was not put
> on paper.
>
> best,
>
> Adam
>
>
>
>
>
> Adam James Ross Tallman (PhD, UT Austin)
> ELDP-SOAS -- Postdoctorant
> CNRS -- Dynamique Du Langage (UMR 5596)
> Bureau 207, 14 av. Berthelot, Lyon (07)
> Numero celular en bolivia: +59163116867
> ------------------------------
> *De :* William Croft [wcroft at unm.edu]
> *Envoyé :* jeudi 24 décembre 2020 18:04
> *À :* TALLMAN Adam
> *Cc :* lingtyp at listserv.linguistlist.org
> *Objet :* Re: discovery procedures vs. diagnostics
>
> Adam's quotation from Seuren's review of *Radical Construction Grammar *prompts
> me to cite a passage of the book (pp 10-11) which Seuren ignored in his
> review:
>
> "It has been suggested to me that the methodological question is of
> relatively
> minor importance. In particular, reference is made to Chomsky's argument
> that it is unreasonable to ask linguistic theory for a discovery procedure
> for identifying the right grammar for a particular language (Chomsky 1957:
> 50–3). However, the problem which I am referring to is more basic than
> that. It is what Chomsky calls the condition of generality (ibid., 50),
> necessary for any adequate theory of grammar: ‘we must characterize the
> form of grammars in a general and explicit way so that we can actually
> propose grammars of this form for particular languages’ (ibid., 53–4).
> That is, for a particular language we can argue for and thus justify the
> analysis of that language's structures as an instance of the structures
> found in Universal Grammar.
>
> It is the condition of generality that I believe current syntactic
> theories fail. That is, the methods that linguists use to argue for their
> syntactic theories carry hidden fallacies which are largely unremarked
> upon. When these fallacious assumptions are uncovered, their abandonment
> leads us to a very different approach to syntactic theory than that
> advocated by formalist theories and even the functionalist syntactic
> theories referred to above."
>
> I've written about those hidden assumptions and fallacies in numerous
> places besides the book, particularly in the references cited below.
>
> Happy Holidays,
> Bill
>
> Croft, William. 2009. “Methods for finding language universals in
> syntax.” *Universals of language today*, ed. Sergio Scalise, Elisabetta
> Magni and Antonietta Bisetto, 145-64. Berlin: Springer.
>
> Croft, William. 2010. “Ten unwarranted assumptions in syntactic
> argumentation.” *Language usage and language structure,* ed. Kasper Bøye
> and Elisabeth Engberg-Pedersen, 313-50. Berlin: Mouton de Gruyter.
>
> Croft, William. “Word classes in Radical Construction Grammar.” *The
> Oxford handbook of word classes, *ed. Eva Van Lier. Oxford: Oxford
> University Press. Draft at http://www.unm.edu/~Papers/WordClassesRCG.pdf
>
>
> ------------------------------
> *From:* Lingtyp <lingtyp-bounces at listserv.linguistlist.org> on behalf of
> TALLMAN Adam <Adam.TALLMAN at cnrs.fr>
> *Sent:* Thursday, December 24, 2020 3:50 AM
> *To:* lingtyp at listserv.linguistlist.org <lingtyp at listserv.linguistlist.org
> >
> *Subject:* Re: [Lingtyp] discovery procedures vs. diagnostics
>
>
> *  [EXTERNAL]*
> Oh there was one more source I meant to attach by Seuren, in his review of
> Croft's 'Radical Construction Grammar'. It is really the only source that
> directly brings up the relationship (that I know of it), but its very
> succint (and silly).
>
> "Croft then proceeds to show (pp. 29–47) that, on this premiss,
> distributionalanalysis leads to a quandary caused by the fact that no two
> elements have exactly the same distribution under any system of
> categorization in anylanguage, which makes it hard to establish syntactic
> categories. Moreover,once categories have been established for one
> language, as a result of‘methodological opportunism’, a further appeal
> must be made to this ‘opportunism’ to establish cross-linguistic
> categories, such as those of noun,adjective, adverb, or subject, direct
> object, etc., which all have non-identicaldistributions in different
> languages. Therefore, established linguistic methodology is at fault and
> RCG will put things right.
>
> The first thing a professional linguist will say, at this point, is that
> thisquandary has been well-known since at least the 1950s. It was
> precisely thereason why the inductive method of distributional analysis
> meant to yield so-called DISCOVERY PROCEDURESwas abandoned as a method
> for discoveringcorrect linguistic analyses and replaced with the
> deductive method of theformation and testing of descriptive hypotheses.
> That is, one just positscategories for each specific language and assigns
> them certain structuraland/or semantic properties, and one then sees to
> what extent the machinery(or module) that is based on them yields the
> right results, according to an agreed set of eliminative adequacy
> criteria."
>
> Any other sources that try to draw out the relationship would be
> appreciated (even if they are not particularly informative or accurate,
> like the one above).
>
> best,
>
> Adam
>
>
>
> Adam James Ross Tallman (PhD, UT Austin)
> ELDP-SOAS -- Postdoctorant
> CNRS -- Dynamique Du Langage (UMR 5596)
> Bureau 207, 14 av. Berthelot, Lyon (07)
> Numero celular en bolivia: +59163116867
> ------------------------------
> *De :* Lingtyp [lingtyp-bounces at listserv.linguistlist.org] de la part de
> TALLMAN Adam [Adam.TALLMAN at cnrs.fr]
> *Envoyé :* jeudi 24 décembre 2020 11:20
> *À :* lingtyp at listserv.linguistlist.org
> *Objet :* [Lingtyp] discovery procedures vs. diagnostics
>
> Hello all,
>
> I wonder if anyone has read any sources that explicitly discuss the
> difference or relation between (in theory and/or in practice) between
> 'diagnostics', which are used to link up theoretical models with new data,
> and 'discovery procedures', which are disparaged, but seem to be, in some
> ways, ancestors of the former notion.
>
> Attached are some relevant citations if you've never heard of the notion
> 'discovery procedure'.
>
> best & happy holidays,
>
> Adam
>
>
>
>
>
> Adam James Ross Tallman (PhD, UT Austin)
> ELDP-SOAS -- Postdoctorant
> CNRS -- Dynamique Du Langage (UMR 5596)
> Bureau 207, 14 av. Berthelot, Lyon (07)
> Numero celular en bolivia: +59163116867
> _______________________________________________
> Lingtyp mailing list
> Lingtyp at listserv.linguistlist.org
> http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20201224/e4862d9c/attachment.htm>