Accueil > Actualités > Le Cycle BCL > Workshop : The phonology-phonetics mapping and related issues in Substance-Free Phonology

Cycle BCL - Elan Dresher, Charles Reiss, Tobias Scheer & Bert Vaux

Workshop : The phonology-phonetics mapping and related issues in Substance-Free Phonology

Mercredi 14 décembre 2022 - 9h-12h - Salle 128, Bâtiment de la MSHS

Workshop : The phonology-phonetics mapping and related issues in Substance-Free Phonology

Workshop organisé à l’occasion de la soutenance de thèse de Paolo Danesi.

Programme :

9:00 am - Elan Dresher - Some issues raised by Paolo Danesi’s thesis : Trade-offs between representation, computation, and transduction
9:40 - Charles Reiss - The unnaturalness of natural classes suggests that they come from nature not nurture
10:20 - Coffee break
10:40 - Tobias Scheer - Emergent primes : prerequisites and consequences
11:20 - Bert Vaux - How many representational states for phonological primes ? The question that won’t go away
12:00 - End

Lien zoom :
https://univ-cotedazur.zoom.us/j/9226426304

Abstracts

9:00 am - Elan Dresher - Some issues raised by Paolo Danesi’s thesis : Trade-offs between representation, computation, and transduction

Danesi observes correctly (p. 484) that ’there is a trade-off between computation and representation : the labour not done on one side will appear on the other side.’ As he shows further, the choices one makes on one side affect what one is able to do on the other. For example, if we put certain limits on permitted computations, we might have to complicate the phonological representations, with implications for constraints on representation such as the Contrastivist Hypothesis. In my remarks I would like to look again at some of these trade-offs. Danesi also discusses the mapping, or transduction, between phonology and phonetics, arguing that it is not universally fixed, but that substance-free phonological primes have arbitrary phonetic correlates. I would like to consider whether there is a third option between ‘fixed’ and ‘arbitrary’. I would also like to propose that trade-offs are not only between computation and representation, but that transduction also needs to be considered : simplifications to representations and to computation can come at the expense of complicating the phonology- phonetics transduction, and vice versa.

9:40 - Charles Reiss - The unnaturalness of natural classes suggests that they come from nature not nurture

In this talk, I will attempt to provide further arguments that we are not phonological amoeboids (to paraphrase Chomsky), and that acquisition is intractable without a universal set of distinctive features. The discussion will be based on the phonetic unnaturalness of certain phonological categories, the phenomenon of near-merger, and the question of what constitutes repetition in language.

10:40 - Tobias Scheer - Emergent primes : prerequisites and consequences

In this talk I argue that sonority cannot be emergent since there is no crazy mapping between phonetic and phonological consonant- / vowelhood (while there are crazy mappings in place and laryngeal properties). This matches the same observation regarding computation : crazy rules are only ever melodically crazy, i.e. regarding place and laryngeal properties.The acquisition of emergent primes consists in establishing i) the existence of a prime and ii) its association to a phonetic correlate. A logical consequence is that this association depends on nothing else than the environment. Since the environment may vary in arbitrary ways that the learning mechanism knows nothing about, so will the associations. Therefore in case of emergent primes, any phonetic property may be associated to any phonological prime (arbitrary mapping).Given the preceding, a list of what the learner needs to know before being exposed to linguistic input is established.

11:20 - Bert Vaux - How many representational states for phonological primes ? The question that won’t go away

"Danesi (p. v) tantalisingly states that "this thesis endorses unary primes. It remains to be seen whether binary primes produce significantly different results".

In this note I make a first step in this direction by revisiting three classes of phenomena where the predictive differences between unary and binary primes are the most clear : ternary patterns, exchange processes, and symmetric dissimilation processes. Skeptics of ternary patterns (e.g. Raimy 2021, Dresher 2021) have argued that they can be derived via other (more complex) means. Critics of alpha notation and advocates of privativity have argued that neither of these sorts of process exist (e.g. Bale, Papillon and Reiss 2014 and Poser 1987 respectively), but without addressing the wide range of such phenomena treated in the literature. I consider whether the fixes proposed by Raimy and Dresher can be generalised to the fuller range of attested ternary patterns, and in a similar vein examine whether the range of exchange and dissimilation phenomena can be insightfully reanalysed without recourse to binarity.

publié par Morgane Ftaïta - mis à jour le