[Lingtyp] comparative concept of "listeme"?

[Martin Haspelmath]

In a database (or dictionary), I think one could at most code elements 
that *must be memorized* (because they are idiosyncratic in some way).

The term "listeme" (coined by Di Sciullo & Williams 1987) refers to 
elements that *are memorized*, though perhaps these authors (and others 
who adopted the term, e.g. Harley's 2006 morphology textbook) do not 
always make a clear distinction between these two.

In recent work (mostly unpublished), I have made a distinction between 
the two concepts, as follows:

* /inventorial items/ are elements which are part of the /inventorium/, 
the set of forms and constructions that MUST be stored (= that are part 
of the language as a system of social conventions)

* /mental(ic) items/ are elements which a language user has stored (= 
that are memorized, as part of the language user's mental lexicon, or 
/mentalicon/)

(For the term /inventorium/, see my 2023 paper on the term 
/construction/ and "constructicon": 
https://constructions.journals.hhu.de/article/view/539/591; for 
/mentalicon/, see this 2022 conference handout: 
https://zenodo.org/record/6408756)

In practice, of course, it's not only impossible to know which elements 
have been stored by particular speakers/signers (this may differ from 
speaker to speaker), but it's also quite hard to know which elements 
*must* be stored. Many English compounds look regular and compositional, 
but they must still be stored because they are not predictable (e.g. 
/*soil container /vs. /flower pot/).

Best,

Martin

On 17.09.23 18:29, Adam James Ross Tallman wrote:
> Hello all,
>
> I was wondering whether anyone had proposed, written about or even 
> just thought of a comparative concept of the "listeme" (the term comes 
> from Pinker's "Words and Rules", I think, but I seemed to have just 
> picked it up in grad school without inquiring about where it came from).
>
> I'm trying to mark up multimorphemic lexical entries in my database 
> with whether they are listemes. Sometimes I don't know whether the 
> form should be a listeme or not, because "memorized chunk" does not 
> translate into a single criterion for identification. Does anyone else 
> have experience coding listemes in databases in a similar fashion? Has 
> anyone proposed a listeme comparative concept?
>
> best,
>
> Adam
>
>
>
> -- 
> Adam J.R. Tallman
> Post-doctoral Researcher
> Friedrich Schiller Universität
> Department of English Studies
>
>
-- 
Martin Haspelmath
Max Planck Institute for Evolutionary Anthropology
Deutscher Platz 6
D-04103 Leipzig
https://www.eva.mpg.de/linguistic-and-cultural-evolution/staff/martin-haspelmath/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20230917/d4dcf465/attachment.htm>