Atelier de phonologie




Bienvenue ! Welcome!
L'Atelier de phonologie est un séminaire de recherche informel consacré à la phonologie : représentation, computation, acquisition et modalités d'apprentissage, interaction avec les autres modules linguistiques, etc. Y sont présentés des travaux en cours, des préparations de conférences, des revues de la littérature, ainsi que des présentations plus accomplies. Le séminaire, hébergé par le laboratoire SFL (Structures Formelles du Langage), est ouvert à la communauté de phonologues locale et internationale. Il se réunit environ toutes les deux semaines, le mercredi matin de 10h à midi, et est suivi d'un déjeuner sur place. Les archives des réunions passées (depuis l'automne 2012) est disponible ici.


The Atelier de phonologie is an informal research seminar dedicated to phonology: its representations, its computations, its acquisition and learnability, its interaction with other linguistic modules, etcetera. The seminar features informal presentations of work in progress, practice talks, literature reviews, as well as more polished talks. It is open to the local community as well as to international visitors. It is hosted at the SFL lab and it meets approximately every two weeks, always on a Wednesday morning, followed by lunch. The archive of past meetings (since fall 2012) is available here.





Où et quand / When and where


Où - Where:
59/61, rue Pouchet, Paris (cartes et informations sur les transports : here you can find a map and information concerning public transportation). 
Quand - When: 

le mercredi, de 10h à 12h. On Wednesdays, from 10am to noon.






Calendrier / Schedule 2018-2019

15 mai - Florian Breit


5 juin - Radwa Fathi


19 juin



12 Sep 
Salle 124















19 Sep 
Salle 124










3 Octobre
Salle 108

















































24 oct
Salle 311

















7 nov
Salle 124
























21 nov
Salle 124
















05 déc
Salle 104












19 dec
Salle 124















09 jan
Salle 311























23 jan
Salle 108


















06 fév
Salle 124








20 fév
Salle 255

























06 mar
Salle 159



















13 mar
Salle 159



















20 mars
salle 108


















03 avr
Salle 124



























17 avr
Salle 108













15 mai
Salle 124
John Alderte (Simon Fraser University)

Phonological regularity, perceptual biases, and the role of phonotactics in speech error analysis

This talk investigates a set of phonological patterns in the SFU Speech Error Database (SFUSED), with the goal of understanding if and how phonological grammar is involved in online processes of phonological encoding. It assesses 2,076 sound errors in English of various types (phonological substitutions, deletions, additions, etc.) for their phonological regularity (do they violate English phonotactics?). The results show that sound errors are much less regular (violate phonotactics more often) than reported in prior work, and in some contexts to not deviate significantly from chance expectations. The higher degree of phonological irregularity is attributed to methodology, because the methods for collecting speech errors in SFUSED are demonstrably less prone to perceptual bias. While
phonological encoding may require linguistic representations of planning units (e.g., segments), these findings suggest that some of the tools of phonological grammars, like syllable-structure algorithms, may not be required in the retrieval of phonological segments in words.



Guillaume Enguehard (Université d'Orléans, CNRS LLL) et Luo Xiaoliang (Tours/lll, École Polytechnique)

Force des voyelles et Branchement

Dans cette présentation, nous abordons la notion de "force" (degré de soumission aux phénomènes de fortition/lénition) dans le domaine des voyelles. Notre objectif est de montrer que la force des consonnes et celle des voyelles peuvent être unifiées. Pour cela, nous proposons que le seul facteur de force est la longueur. Plus précisément, les segments branchants sont plus forts tandis que les segments partageant leurs positions sont plus faibles. Nous abordons plusieurs phénomènes vocaliques illustrant cette hiérarchie.



Bien Dobui

Implications of Nasalization Blocking Strategies in Xochistlahuaca Amuzgo

Nasalization, both progressive and regressive, is an important phenomenon in the Amuzgo spoken in Xochistlahuaca in Mexico. However, on a reduced syllabic form of CCV, identity is easily lost. This presentation looks at different blocking strategies employed to block nasalization at the lexical and morphological levels and their implications.

Paralleling Stephen Marlett’s observation on Mixtec in his 1992 work, *NV and *LVn do not occur in Amuzgo (N being a nasal sonorant, L a non-nasal sonorant, V a vowel and Vn a nasal vowel). This attests to both progressive and regressive nasalization. On the other hand, and as may be expected, LV and NVn do occur.

1.            

ná

[nán]

‘hot’

2.           

wa̰

[wa̰]

‘EXIST.INHUM’

3.               

wa̰n

[ma̰n]

‘EXIST.HUM’

 What is particular to Amuzgo, however, is the template NTV, better analyzed as NTV given that NT becomes N when V is nasalized, e.g. by the 3SGHUM marker, as in examples 4 to 6. The difference between the template NTV, which also exists, is seen when the latter is nasalized without loss of the stop, as in example 7.

4.               

[ɲdj]
‘mouth.3SGPOSS’

ɲó
[ɲón]
‘mouth.3SGHUM.POSS’

5.               

[ma-ndaʔ]
‘PROG.SG-receive a gift’

ma-naʔ=an
[ma-nanʔ=an]
‘PROG.SG-receive 
a

gift.3SGHUM=3SG’

6.               

[ma-hndɛ]
‘PROG.SG-sell’

ma-hnɛɛ́
[ma-hnɛɛ́n]
‘PROG.SG-sell.3SG’

7.               

[ma-ɲ̍tjɔ̤∙]
ma-ɲtjɔ̤∙
‘PROG.SG-frolic’

[ma-ɲ̍tjɔ̤n∙]
ma-ɲtjɔ̤n
PROG.SG-frolic.3SGHUM

On the morphological level, nasal blocking is also carried by allomorphy in plurals and future markers, though of a different form. In the former, the nasal morpheme /n/ denasalizes as [l] (examples 8 and 9) where in the future marker, also /n/, an NT type is recalled, though a velar stop rather than an alveolar, [ŋ̩́k] (examples 10 and 11).

8.               

ʦʔán
‘person’

[nʔán∙]
‘PL.person’

9.               

ska∙
‘candle’

[l̩ka∙]
‘PL.candle’

10.           

tɔ ∙

‘to vomit’

[ń̩-tɔ∙]

‘FUT-vomit’

11.              

  huʔ

‘throw, expluse’

[ŋ̩́k-huʔ]

‘FUT-throw, expluse’

These strategies seem to imply a balancing of concerns against preventing identity loss through nasalization by maintaining what may appear to be an array of nasal allophones and allomorphs, but that actually form a strictly dedicated set that works to preserve identity.


Francesc-Josep Torres Tamarit (Université Paris 8 CNRS SFL) (avec Eulália Bonet, UBA)

Verb-clitic structures in Eivissan Catalan: recursive prosodic words and allomorphy

This presentation deals with stress shift in verb-clitic structures in Eivissan Catalan, an understudied Romance variety. Within Balearic Catalan, this is the only subdialect in which stress shift is restricted to apply only in second conjugation infinitives followed by pronominal enclitics, those that, as opposed to other conjugations, have penultimate stress when they are pronounced in isolation. Stress in second conjugation infinitives in Eivissan Catalan shifts one syllable to the right, that is, to the final syllable of the verbal stem, when one or more pronominal enclitics follow. There is no stress shift in imperatives followed by pronominal enclitics. We claim that pronominal enclitics in Eivissan Catalan adjoin to a recursive, maximal prosodic word, and that the domain for stress assignment is the minimal, embedded prosodic word. We further analyze two cases of stress-conditioned allomorphy (i.e. allomorphy of the infinitive morph and allomorphy of the verbal root) that occur in infinitive-clitic structures and that we analyze as cases of phonologically-conditioned allomorphy triggered by stress shift.


 Alexis Michaud (CNRS-LACITO) 


L'articulation phonétique/phonologie au miroir de l'intelligence artificielle : analyse des résultats de la transcription automatique de la langue Na (famille sino-tibétaine) par le logiciel Persephone

 

Les systèmes de reconnaissance automatique de la parole permettent désormais d'entraîner un modèle acoustique sur la base de deux ou trois heures d'enregistrements transcrits (pour un système mono-locuteur), au lieu de dizaines d'heures pour les outils antérieurs. Au-delà de l'intérêt pratique que présentent ces avancées technologiques pour les tâches de documentation linguistique, se pose la question de la mise en correspondance entre les modèles phonologiques et les modèles statistiques construits par les logiciels d'intelligence artificielle. En effet, l'algorithme réalise son entraînement sur la base de transcriptions fournies en entrée par le linguiste, transcriptions qui reposent sur un ensemble – plus ou moins élaboré, et plus ou moins explicite – d'hypothèses phonologiques. Le modèle acoustique, décalqué (par des méthodes statistiques) des transcriptions du phonéticien/phonologue jointes au signal acoustique, peut-il être interrogé par le chercheur, en un jeu de miroir ? Que peut nous apprendre la confrontation ainsi renouvelée avec le signal acoustique ?

     L'exposé vise à ouvrir des pistes pour élaborer une perspective phonologique sur ces questions nées à l'intersection de plusieurs disciplines. L'exposé, qui ne présuppose pas de connaissances préalables en traitement automatique des langues, commencera par une présentation du logiciel Persephone et de son application à la langue na de Yongning (famille sino-tibétaine). Des exemples précis de divergences entre la notation du linguiste et celle du logiciel seront offerts à la sagacité des participants, pour un exercice partagé de phono-philologie (ou philo-phonologie) numérique.




Joanna Zaleska (Leipzig University)


Defining phonologically relevant substrings

In this talk, I examine the question of how to restrict the application of a phonological process to the relevant portion of the linear string. As a test case, I take a particularly intricate pattern of phonological process application reported in Standard Indonesian (Lapoliwa 1981, Cohn 1989, Sneddon et al. 2010). The main challenge posed by the data is that the subparts of the word affected by different phonological processes are not nested but rather partially overlap. Some of these processes affect the lexical root and the suffix, but not prefixes. For others, the situation is reversed: Roots and prefixes are involved, while suffixes seem invisible. The analysis I propose for this set of data includes two types of restrictions on phonological processes. I argue that the root and suffixes in Indonesian form a single prosodic domain; the processes that fail to apply to prefixes are restricted to this domain. Prefixes, however, are attached to roots earlier than suffixes. Processes that seem to ignore suffixal material apply at a point at which suffixes are not yet present in the structure. In the final part of the talk, I will speculate about a possible reanalysis of the data, whereby the phonologically relevant portions of the linear string are identified exclusively by means of procedural mechanisms, as postulated by Scheer (2011) and D'Alessandro & Scheer (2015).


Jelena Stojković (Leipzig University)

(Vowel) Dissimilation is an Illusion – Looking from Afar 

This talk concerns a very complex vowel dissimilation pattern in the Afar language (Cushitic; Bliese 1981, Didier 2012, Parker & Hayward 1985): with the particular/singulative and plural morphemes low vowels dissimilate. The two dissimilations differ in resulting vowels /u, i, o/, loci and conditions for application, but are also the only two environments where this alternation is present. Supported with data from outside the nominal domain in Afar, as well as data from a typological survey of 25 languages, I claim that all these cases are misinterpreted and the apparent dissimilation is in fact featural affixation, where the height and place features involved in the change are actually floating information the respective morphemes are specified with.



Guillaume Enguehard (Université d'Orléans, CNRS LLL)

À propos de la cénématique (et de son application aux langues et aux jeux)

Le terme "cénématique", introduit par l'école de Copenhague, désigne l'étude de la structure immanente à l'expression d'un système sémiotique (c'est à dire sans considération pour ses aspects physiques, psychiques et sociaux). L'objectif de cette présentation est de montrer que cette approche holiste de la langue n'a pas pour seule vocation de fournir la description d'un réseau d'oppositions. Elle permet également de dériver les classes d'unités réalisées par les segments, les tons, les accents et les longueurs. Premièrement, j'argumente que les approches historiques de la phonologie sont mises en difficulté lorsqu'il s'agit de proposer un appareil théorique permettant à la fois une description univoque de l'objet étudié et une prédiction univoque de ses possibilités. Deuxièmement, je propose un raisonnement strictement déductif aboutissant à une typologie des unités cénématiques et à la procédure permettant de les identifier. Troisièmement, j'applique cette procédure à des exemples concrets afin de montrer que les unités prédites sont attestées dans deux types de systèmes sémiotiques : les langues et les jeux combinatoires abstraits. Je conclus en soulignant que cette approche n'est pas en opposition avec les écoles distributionnaliste, fonctionnaliste et générativiste. Elle les complète en remplissant l'une des cases du schéma réunissant les aspects extra-, intra- et inter-individuels de la langue.




Alexandre Vaxman

How to uniformly capture accentual regularities and exceptions both within a given accent system and across different types of systems using a single accent-assigning mechanism? In this talk, I will focus on exceptional affixes in lexical accent systems (accented dominant suffixes in Central Selkup and Uzbek) and in mixed systems (Eastern Literary Mari, Tundra Nenets). Since morphemes, on a par with syllables, are able to attract/repel word accent, I propose that morphemes also have weight, albeit unpredictable (“diacritic weight”). Since “weight” is an ordinal variable (as evidenced by phonological weight scales) and since diacritic weight is a type of weight, novel types of weight scales are predicted, which order either diacritic weight alone (in lexical accent systems), or both syllable weight and diacritic weight (in mixed systems). In the latter type of system, the two types of weight may be independent (Mari), or dependent (Tundra Nenets). Reference to such weight scales enables the parametric component of the grammar to correctly and uniformly assign word accent in all three types of systems (phonological, lexical and mixed). An advantage of this approach is that it makes accent location in languages like Tundra Nenets partly predictable, thereby reducing the amount of lexical specification of weight. I will show that it also allows for a straightforward account of Moses Columbian Salish, traditionally analyzed as involving a highly complex “strength hierarchy”, dominance and morpheme extrametricality.



Conor Youngberg (Université de Nantes)


Japanese dialects, the moraic nasal and diglossia


In this talk, I discuss the representation of the Japanese moraic nasal, typically transcribed as <N>. I propose that this segment is not a (phonological) consonant in every dialect. N is traditionally analysed as a sparsely specified coda or moraic segment consisting only of the feature [+nasal] (McCawley 1968, Itō 1987, Vance 2008; Labrune 2012), realised as a dorso-uvular glide intervocalically and finally (cf. Vance 2008) e.g. <hoN> [hoɰ̃] ‘book’, or as a homorganic nasal when preceding a consonant, e.g. <hoNdana> [hondana] ‘bookshelf’. Alternatively, N is viewed as a syllabic nasal or nasal back vowel vowel in word-final & intervocalic contexts and a coda elsewhere (Yoshida Y. 1999, Yoshida S. 2003). I claim that alternative representations are necessary in order to capture facts of N beyond homorganicity. I propose that in Tōkyō Japanese, the orthographic vowel-N sequence is phonologically a nasal vowel, while in Osaka Japanese N is a syllabic consonant. Finally, I propose that in Kagoshima Japanese, N is a true coda consonant. I frame the alternative representations of N within Strict CV (Lowenstamm 1996) and the TBU status of N is directly connected to the status of the V position within the relevant CV unit. I consider the accommodation of diglossia through licensing parameters (Cyran 2010). I conclude with an outline of forthcoming instrumental investigations of nasality in the dialects discussed.  




Guillaume Enguehard & Xiaoliang Luo (Université d'Orléans)

A Strength is Length solution to Edge and Bipositionality

Cette présentation se fonde sur la comparaison entre deux hypothèses récentes : Edge-and-Bipositionality (Lahrouchi & Ulfsbjorninn 2016) et Strength-is-Length (Luo & Enguehard à paraître). Notre objectif est de mettre en évidence les points de convergences et de divergences dans la manière dont ces hypothèses traitent la distribution de la sonorité. Nous argumentons que les généralisations de Edge-and-Bipositionality peuvent être formulées par Strength-is-Length avec plus de détail et moins de concepts théoriques.


Noam Faust (Université Paris 8, CNRS SFL)

"You may emerge, but on my terms!" said UG to the epenthetic vowel

Much of the phonology of Modern Hebrew (MH) emerged out of the phonologies of the native languages of the dominant group of founding speakers, the Ashkenazi jews of eastern Europe. For this reason, it is especially interesting to examine aspects of MH phonology that do not replicate patterns in those languages. One such aspect is the quality of its epenthetic vowel, [e]. This quality is surprising for two, related reasons. First, epenthetic [e] seems to be rare among the world’s languages (Hall 2006, Lombardi 2002). Second, and more importantly for the present purpose, it is not the epenthetic vowel of Yiddish or Russian. I claim that [e] is motivated by the interaction between the universal principles that govern the choice of epenthetic vowels and the specific morphology of Modern Hebrew. Specifically, among the three lexical vowels that alternate with zero, /e/ is the one that does so in the most contexts. I will present the proposal in Faust & Smolensky (2017), according to which an alternating vowel is lexicalized with an activity level (=strength) that is inferior to that of a non-alternating vowel. Within gradient harmonic grammar (Smolensky & Goldrick 2016), syncopating an alternating vowel is thus less of a violation of Max than syncopating a non-alternating vowel like /i/. Given this manner of lexicalizing weakness, the choice of the quality /e/ for the epenthetic vowel follows from the OT notion of epenthesis as a violation of an output’s dependence on the input: /e/ is selected because it least violates Dep. To generalize from MH, given (i) a language with V-zero alternations in lexical vowels, (ii) the absence of a designated, non-lexical epenthetic quality, and (iii) the universal principles of phonology, the quality of epenthesis is predictable. I will therefore be making the claim that the quality of epenthesis is both universally determined and emergentIt emerges from the data according to universal principles. 



Radwa Fathi (Université de Nantes, LLING) & Jean Lowenstamm (Université Paris Diderot, LLF)

The gender assignment pattern of French nouns

Au cours des 20 dernières années, l'identité et le rôle des traits phi a occupé une place importante dans les préoccupations des morphologues 'minimalistes'. Mais curieusement, ce n'est que beaucoup plus récemment que le Genre a été sérieusement étudié (dans le cadre en question), et que la question de l'assignation du genre aux noms a été confrontée. Les critères selon lesquels le genre est assigné sont-ils formels, sémantiques, ou impliquent-ils une combinaison ? Nous démontrerons que le genre des noms du français leur est assigné sur la base de critères formels dans des conditions définies. Lorsque ces conditions ne sont pas réunies, le genre est assigné arbitrairement. Dans la mesure où notre proposition ne reconnaît pas de rôle à des critères biologiques comme le sexe, il nous appartiendra de dériver les régularités évidentes que l'on observe dans le cas des espèces sexuées: de fait, si un locuteur du français est informé du fait que rétroscopiste est une profession, il n'a aucun mal à savoir a) que LE rétroscopiste et LA rétroscopiste sont tous deux des noms possibles, b) lequel des deux est une femme. Nous montrerons pourquoi et comment. La présentation de nos critères formels nous donnera l'occasion de consacrer une partie importante de notre intervention au développement d'une proposition nouvelle pour la représentation des consonnes flottantes du français.


Nicola Lampitelli (Université de Tours, LLL) & Francesc-Josep Torres Tamarit (université Paris 8, LSF)

Vowel length in Friulian: extrasyllabicity of voiced final obstruents and mora affixation

In this presentation we reflect on the sources of vowel length in Friulian. We will show that vowel length in this language is sometimes phonologically predictable and sometimes an instance of mora affixation in first conjugation verbs. Underlyingly voiced obstruents in word-final position cannot be parsed in coda position due to restrictions on the type of allowed codas in the language. Underlyingly voiced obstruents become extrametrical and devoice. Foot-Binarity enforces vowel lengthening. A formal analysis of these data will be presented within Turbidity Theory (Goldrick 2001, van Oostendorp 2008), based on the phonetic findings of Baroni & Vanelli (2000), which suggests that obstruent final devoicing in Friulian is an instance of incomplete neutralization. We will present new data on verbal morphology in Friulian and show that vowel length is also the expression of the Theme morpheme in first conjugation verbs, which ahs distinct allomorphs. One of the sources of vowel length in Friulian is therefore mora affixation. This analysis of morphological length in Friulian shows that there is no need for an L-shpaed morphome analysis of the data (Maiden 2004). In our analysis, each morph, including length, spells out a morphosyntactic feature.  



Ora Matushansky (université Paris 8, LSF) parler de


Russian athematic verbs, without stress (and with)


As is well-known, each Russian morpheme is accentually prespecified as accented, post-accenting, pre-accenting or unaccented. In addition, the morphophonology of the Russian verb involves a number of vowel hiatus resolution strategies such as deletion or conflation and glide formation. So what happens to various types of accent when the vocalic morpheme that it is associated with disappears from the surface representation?

 

In this talk I will first establish the accentual properties of various relevant morphemes in athematic verbs and additional processes that might influence the position of stress in Russian verbs. I will then demonstrate that different vowel hiatus resolution strategies influence the eventual position of stress in different ways and discuss what this means for various theories of lexical accent.

 



 Owen D.E. Edwards (université de Leiden)

Metathesis and Unmetathesis in Amarasi

I provide a complete analysis of synchronic CV -> VC metathesis in Amarasi, an Austronesian language of western Timor. Metathesis and unmetathesis realise a paradigm of parallel forms, pairs of which occur to complement each other throughout the language.

Metathesis in Amarasi is superficially associated with a bewildering array of disparate phonological processes including: vowel deletion, consonant deletion, consonant insertion and multiple kinds of vowel assimilation, any of which can (and do) vary by lect in their realisation. By proposing that Amarasi has an obligatory CVCVC foot in which C-slots can be empty, all these phonological processes can be straightforwardly derived from a single rule of metathesis and two associated phonological rules.

Three kinds of metathesis can be identified in Amarasi. (i) Before vowel initial enclitics, roots must undergo metathesis, responding to the need to create a phonological boundary between prosodic words. (ii) Metathesis occurs within the syntax to signal attributive modification. A syntactically metathesised form cannot occur at the end of a phrase and thus requires the presence of an unmetathesised form to complete it syntactically. (iii) In the discourse an unmetathesised form marks an unresolved event or situation. Such an unmetathesised form cannot occur in isolation and requires a metathesised form to achieve resolution. Metathesis in Amarasi is the central linguistic process around which linguistic structures are organised.

Amarasi metatheses also reflect fundamental Timorese notions of societal and cosmic organisation. Alongside weaving and other performed activities, metathesis is an important linguistic marker of identity in a region obsessed with similarities and differences of identity between different groups. The complementarity of Amarasi metathesis and unmetathesis within the syntax and within discourse reflects the Timorese division of the world into a series of mutually dependent binary and complementary pairs. As well as being the key which unlocks the structure of the language, metathesis is also a reflection of the structure of Amarasi society and culture.



Adèle Jatteau (SFL) 

Définir les domaines du « mot » phonologique : le cas du grec ancien

 

Les processus phonologiques s’appliquent souvent dans des domaines plus petits ou plus larges que le « mot » morphologique. Comment définir ces domaines ? Dans cet exposé, je m’intéresse au cas du grec ancien. Je montre que, alors que les processus segmentaux et phonotactiques (syllabation, clusters, distribution de /h/, assimilation) traitent le préfixe comme un objet indépendant de la base, les règles de placement de l’accent obéissent à un schéma plus complexe : le domaine accentuel est le mot préfixé dans son ensemble, mais l’accent ne peut pas remonter plus loin que la dernière syllabe du préfixe. Par ailleurs, alors que d’autres processus segmentaux s’étendent au mot affixé mais aussi aux clitiques, suggérant que l’ensemble clitique + hôte forme à son tour un « mot » pour la phonologie (Golston 1995, Agbayani & Golston 2010, Kiparsky 2003), l’accent ne s’étend pas aux clitiques. Pour rendre compte de ces phénomènes, je compare plusieurs traitements théoriques, et propose une analyse combinant domaines cycliques et prosodiques.

 



Florian Breit (University College London)

Reconciling Mutation and Modularity in Welsh

Initial Consonant Mutation (ICM) is a phenomenon in which words undergo a phonologically regular change that is conditioned by a morphosyntactic environment rather than a phonological one. For instance, Welsh /tad/ ‘father’ is variously realised with a nasal following the 1s possessive ([və n̥ʰad] ‘my fatrher’), with a voiced stop following the 2s possessive ([də dad] ‘thy father’), with a fricative following the 3s feminine possessive ([i θad] ‘her father’) or with no change at all following the otherwise homophonous 3s masculine possessive ([i tad], ‘his father’).

ICM phenomena are challenging for theories of the Morphosyntax–Phonology Interface. They appear to require either that morphosyntactic information is conveyed to the phonology and can condition alternation there, or that somehow certain phonological alternations can take place at some stage of morphosyntactic computation. Both options are incompatible with standard assumptions about the modularity of the mind that underpin current models of linguistic architecture (cf. Fodor 1983, Segal 1996, et al.). While linguists have long denied the latter option (phonology inside syntax), the former is still commonly employed, usually in the form of a module-transcending ad-hoc feature. For mutation, the basic story goes something like this: the morphosyntax annotates a node with a diacritic feature (say, [M]) which–-unlike regular syntactico-semantic features–-perseveres across the interface, where it can then condition phonological alternations such as ICM.

Virtually all the available theories of Welsh ICM (e.g. Ball & Müller 1992, Kibre 1997, Green 2007, Hannahs 2013) are, in one form or another, different riffs on the modularity-violating diacritic solution above. In contrast, I argue that even though Welsh ICM might seem like the ultimate modularity-breaker, a different analysis is possible. Taking inspiration from Lieber’s (1983) autosegmental account, I propose that mutations are the surface effect of floating phonological features that form part of the spell-out of morphosyntactic terminals, either as part of the underlying form of a trigger such as the various possessives above, or as the sole exponent of a terminal. Crucially, the insertion of the floating features is phonologically conditioned, so that a different floating feature may be inserted next to a oral vs nasal stops, for example. This overcomes various paradoxes known to arise from Lieber’s previous account.

It will be shown that the new account is not only compatible with modularity, but also empirically superior in many regards: it accounts for the productivity of the process, makes better predictions about variation and exceptionality, and brings within reach a first satisfactory account of items that appear to completely resist mutation (e.g. /ge:m/ ‘game’ is never affected by mutation and we have [və ge:m], [də ge:m], etc.). Based on this I conclude that mutation is best accounted for precisely when modularity is taken into account and respected.












Archives

Les archives des réunions passées (depuis l'automne 2012) sont disponibles ici.

The archive of past meetings of the Atelier is available here.




Contacts


Pour venir présenter ou être ajouté à la liste de diffusion, merci de nous contacter à l'adresse :
If you would like to come present to the Atelier or to be added to the mailing list, please contact us at:
noam.faust@univ-paris8.fr.