[PDF] A Classic Thesis Style et prior syllaba in hac

View - Download A Classic Thesis Style

Previous PDF

Next PDF

en vue de l"obtention du

DOCTORAT DE L"UNIVERSITÉ DE TOULOUSE

délivré par l"université de toulouse 2-le mirail discipline:sciences du langage zur erlangung der doktorwürde

DER NEUPHILOLOGISCHEN FAKULTÄT

DER RUPRECHT-KARLS-UNIVERSITÄT HEIDELBERG

présentée et soutenue par vorgelegt von

DANIELA MÜLLER

DEVELOPMENTS OF THE LATERAL IN OCCITAN DIALECTS

AND THEIR ROMANCE AND CROSS-LINGUISTIC CONTEXT

JURY Francesc Xavier Lamuela (Catedràtic, Universitat de Girona) Jean-Léonard Léonard (Maître de conférences HDR, Paris3-ILPGA et ILF) Patrick Sauzet (Professeur, Université de Toulouse2-Le Mirail) (directeur)

école doctorale:clesco

unité de recherche:clle-erss directeurs de thèse:patrick sauzet et edgar radtke rapporteurs:daniel recasens(catedràtic,universitat autònoma de barcelona) le1er décembre2011 There is a theory which states that if ever anyone discovers exactly what the Universe is for and why it is here, it will instantly disappear and be replaced by something even more bizarre and inexplicable. Douglas Adams, The Restaurant at the End of the Universe,1980, p.7 ABSTRACTThis thesis analyses sound changes that affected the lateral approx- imant inherited from Latin in Occitan dialects, in the Romance lan- guages, and in a number of other languages from around the world. Chapter1gives a comprehensive overview of the research carried out on the lateral approximant; it discusses articulation and acoustics as well as abstract representations of the sound. Chapters2to5are devoted to specific sound changes which occurred in Occitan dialects at different points in time. These developments are systematically compared to similar phenomena in Romance and other languages. In chapter2, I discuss the vocalisation of the dark lateral in preconsonan- tal and word-final position as well as intervocalically. It is argued there that Occitan and more generally Romance followed an unexpected pathway towards vocalisation, which cannot be explained by phonetic factors alone. Chapter3deals with palatalisation of the lateral in onset clusters. Rather than in articulatory assimilation, I propose that the origin of this sound change is to be sought in the frication which accompanied the obstruent + lateral onset clusters. Rhoticisation of the lateral, and its opposite, lambdacisation of the rhotic, is the topic of chapter4. In this chapter, I discuss duration factors in these sound changes and present experimental evidence to substantiate the idea that duration plays an important role. Finally, chapter5looks at the developments of the Latin geminate lateral in Gascon and other Ro- mance dialects; according to common opinion, the Latin geminate lateral underwent a retroflexion process, and I discuss how this might have been possible from a phonetic point of view.

Key words

Sound change - laterals - Occitan - Romance linguistics - phonetics v RÉSUMÉCette thèse se propose d"analyser les changements phonétiques qui ont affecté l"approximante latérale héritée du latin dans les dialectes occitans, dans les langues romanes ainsi que dans un certain nom- bre d"autres langues du monde. Le premier chapitre donne une vue d"ensemble des recherches phonétiques sur la latérale. Les aspects articulatoires, acoustiques et les questions de représentation abstraite y sont notamment discutés. Les chapitres suivants,2à5, sont consacrés à certains changements phonétiques spécifiques survenus dans les di- alectes occitans à des moments divers de leur histoire. Ces évolutions seront comparées de façon systématique à des phénomènes semblables dans les autres langues romanes et non-romanes. Le chapitre2traite de la vocalisation de la latérale vélarisée en position préconsonantique et finale de mot aussi bien qu"à l"intervocalique. Je propose dans ce chapitre que l"évolution de la vocalisation de la latérale en occitan et dans les langues romanes en général ne peut être expliquée à par- tir des seuls faits phonétiques. Dans le chapitre3, je discute de la palatalisation de la latérale dans un groupe consonantique en attaque syllabique. J"y fais l"hypothèse que l"origine de ce changement phoné- tique est à rechercher plutôt dans la friction qui accompagne le groupe consonantique du type obstruante + latérale que dans une assimilation de type articulatoire. La rhotacisation de la latérale, et son contraire, la latéralisation d"une rhotique sont traitées dans le chapitre4. Dans ce chapitre, je discute des facteurs de durée dans ces changements phonétiques et je présente des résultats expérimentaux qui soutiennent l"idée que la durée y joue un rôle important. Finalement, le chapitre5 regarde de près l"évolution de la latérale géminée latine en gascon et dans d"autres dialectes romans; étant donné qu"il est communément admis que cette évolution est due à un processus de rétroflexion, je discute les chemins possibles de ce changement d"un point de vue phonétique.

Mots-clés

Changement phonétique - latérales - occitan - linguistique romane - phonétique vii PUBLICATIONSSome ideas and figures have appeared previously or will appear soon in the following publications: Müller, Daniela & Sidney Martín Mota (2009), "Acoustic cues of palatalisation in plosive + lateral onset clusters". In Proceedings of Interspeech2009, Brighton,6.-10. September2009, Rundle

Mall: Causal Productions;1695-1698.

Müller, Daniela (2010), "Phonetic factors influencing /l/-rhoti- cisation in Greek". In Botinis, Antonis (ed.), Proceedings of the third ISCA Tutorial and Research Workshop on Experimental

Linguistics, Athens,25-27August2010;117-120.

Müller, Daniela (in press), "Le sort de /l/ dans les dialectes occitans : vélarisation et palatalisation". In Rieger, Angelica (ed.), L"Occitanie invitée de l"Euregio. Liège1981- Aix-la-Chapelle

2008. Bilan et perspectives/ Occitània convidada d"Euregio. Lièja

1981- Aquisgran2008. Bilanç e amiras/ Okzitanien zu Gast

in der Euregio. Lüttich1981- Aachen2008. Bilanz und Per- spektiven. Actes du neuvième Congrès International de l"AIEO,

Aix-la-Chapelle,24-31août2008. Aachen: Shaker.

Müller, Daniela (in press), "Un estudi critic de l"evolucion /l:/ > /ã/ > /d,dZ, tS/ en gascon e en gardiòl". In Actes du colloque " Nouvelle recherche en domaine occitan : approches interdisci- plinaires » Albi,11-12juin2009. Müller, Daniela (to appear), "Retroflexion de la laterala geminada latina en Gasconha - apròchis a travèrs l"istòria de la lingüistica". In Actes du dizième Congrès International de l"AIEO, Béziers,

12-19juin2011.

ix ACKNOWLEDGMENTSA number of people have assisted this thesis in its transition into the real world, and I would like to thank them for their efforts. I"d like to thank Patric Sauzet for his support throughout the past ten years and for innumerable last-minute reports and recommendation letters. Thanks are also due to Edgar Radtke for accompanying my academic development since my very first year and for supervising this co-tutelle thesis. I am greatly indebted to Daniel Recasens for his close reading of preliminary versions of this thesis and his perspicacious and perspicu- ous comments and remarks. I hope that this final version will live up to them. Several of my colleagues have read draft versions of some chapters: Mark Jones has given me useful comments, suggestions, and research ideas on /l/-palatalisation in onset clusters (chapter3), and Chiara Celata and Thomas Field have provided important feedback on the chapter on retroflexion of Latin /l:/ in Gascon and in other Romance dialects (chapter5). Moreover, a large number of colleagues have responded to several of my queries on LinguistList with references and discussions of topics pertaining to /l/ in a wide variety of languages; thank you to all! My early research on /l/-palatalisation in onset clusters (chapter

3) was carried out with Sidney Martín Mota, and Ana Paula Huback

pointed me towards the idea of frequency effects. Finally, Athanassios Vergados and József Krupp made sure that the translations from Latin in Appendix A are correct. My office-mates Stéphanie Lopez and Aurélie Guerrero corrected the French short abstract, Aurélia Lassaque the extended French summary, and Bill Spruiell, Trey Jones, Mikael Thompson, and Thomas Reiss prevented me from doing any harm to the English language. Especially Tom Reiss spent long dull hours and interesting minutes on reading and re-reading (and arguing about spelling) every single word written in this thesis rather than on, say, mowing the lawn. I gratefully acknowledge financial support from the Landesgradu- xi

CONTENTS

1characteristics of the lateral approximant9

1.1Articulation of the lateral10

1.1.1Tongue tip and blade10

1.1.3Tongue body14

1.1.4

Origins of positional variants of the lateral19

1.2The acoustic structure of the lateral21

1.2.1The formant structure of the lateral21

1.2.2

The zero in the transfer function caused by lat-

eral side channels24

1.3Gestures of the lateral26

1.4Phonological views of the lateral32

1.5Conclusion and outlook34

2 vocalisation of dark/ë/35

2.1 /ë/-vocalisation in preconsonantal and word-final po- sition35 2.1.1 The geographic and historical expansion of /ë/- vocalisation in preconsonantal and word-final position35 2.1.2 Articulation and perception of /ë/-vocalisation as a sound change in progress60

2.1.3Explaining /ë/-vocalisation72

2.1.4Conclusion83

2.2Vocalisation of dark /ë/ in intervocalic position83

2.2.1 The geographic and historical expansion of /ë/- vocalisation in intervocalic position84 2.2.2 Explaining the developments of intervocalic /ë/92

2.2.3Conclusion95

3 palatalisation of/l/in onset clusters97

3.1 The geographic extent of /l/-palatalisation in onset clusters97 3.1.1 The palatalisation of /l/ in onset clusters in Oc- citan97 3.1.2 /l/-palatalisation in onset clusters in Romance and elsewhere99 3.1.3 /j/-lateralisation in labial + palatal onset clus- ters105 3.2 The phonetic characteristics of stop + lateral onset clus- ters105

3.3Evolution of /l/-palatalisation in onset clusters108

3.3.1Articulatory explanations108

xiii xivcontents

3.3.2/l/-palatalisation in labial + lateral clusters as

an effect of analogy111 3.3.3 Experimental phonetic approaches to /l/-palatalisation in onset clusters114

3.3.4The lateral fricative hypothesis116

3.4Conclusions124

4 rhoticisation of the lateral

127
4.1 Rhoticisation of /l/ in Occitan, in Romance, and else- where127

4.1.1Rhoticisation in Occitan dialects127

4.1.2

Rhoticisation in other Romance languages129

4.1.3 Rhoticisation (and lambdacisation) cross-linguistically134 4.2 Phonetic differences and similarities between /l/ and /R/136 4.2.1

Differences and similarities in articulation be-

tween /l/ and /R/136 4.2.2 Differences and similarities in the acoustic struc- ture of /l/ and /R/137 4.3 Explaining /l/-rhoticisation (and /R/-lambdacisation)140

4.4Conclusion146

5 retroflex developments of the latin geminate lat- eral in gascon 149
5.1 Retroflex developments of Latin /l:/ in Gascon and elsewhere149

5.1.1Developments of Latin /l:/ in Gascon149

5.1.2

Developments of Latin /l:/ in Upper Aragonese

and Western Asturian152 5.1.3

Developments of Latin /l:/ in Southern Italian

and Sardinian dialects154 5.1.4 Retroflexion processes affecting the lateral else- where157 5.2 A century of puzzlement - explaining the fate of Latin /l:/ in Romance158

5.2.1Substrates and colonisations159

5.2.2Structuralist approaches162

5.2.3

How Latin /l:/ evolved in Gascon: the retroflex-

ion hypothesis163 5.2.4

How Latin /l:/ evolved in Gascon: the palatali-

sation hypothesis167 5.2.5 The mixed hypotheses: palatalisatio-ex-retroflexione and retroflexio-ex-palatalisatione168 5.3 Phonetic views on Romance retroflexion of Latin /l:/169 5.3.1 Articulatory and acoustic characteristics of retroflex consonants169 contentsxv

5.3.2Phonetic approaches to the evolution of retroflex-

ion of Latin /l:/ in Gascon and the other Ro- mance languages and dialects172

5.4Conclusion178

Appendix185

a the lateral of latin 187
bibliography 197
INTRODUCTIONThe evolution of language has attracted a lot of attention over the centuries, and sound change perhaps even more so than anything else. Changes in pronunciation relate not only to one sound becoming another sound or many sounds merging into one sound or vice versa, they can also affect the language"s functioning as a whole, by being incorporated at a very early stage into the morphological domain, by being conditioned by syntactic circumstances, by being correlated with fine semantic and pragmatic distinctions and by being carriers of social information about the speaker, her background and the context of the utterance. This thesis will attempt to throw some light on a small part of the phenomenon, taking as a starting point a sound which has rarely been addressed in a monograph: the lateral approximant.1 These sound changes are discussed as their occur in the dialects of Occitan. Indeed, Occitan with its very diverse dialectal developments, which have been rather well-documented over the centuries, is an excellent illustration of the phenomena studied in this thesis. However, since phonetic explanations are only justified if the phenomenon at hand turns out to have a large cross-linguistic incidence, each sound change is also traced in the other Romance languages and beyond. In this sense, all the chapters will show that the diversity of sound changes affecting the lateral consonant in Occitan dialects is neither unparalleled nor unprecedented.

Intended readership

The readership addressed by the present study is rather diverse: pho- neticians will find chapter1useful, where I have compiled the state-of- the-art of the research on laterals; linguists interested in sound change will find the case studies in chapters2-5worth reading; Romance scholars and dialectologists, especially Occitanists, can find informa- tion of the place of the sound changes with respect to not only other Romance languages, but also to the larger cross-linguistic context.2

Occitan dialects

Occitan is a Gallo-Romance language spoken in the area depicted in Figure1on page2.Bec ( 1963) provides an excellent and concise introduction to its history and its present state. The figure also shows1 Earlier monographs on the lateral or on a specific sound change involving the lateral include de Kolo vrat 2 Propinquity to Romance studies is signalled by the use of small capitals for Latin words and other etyma of Romance. 1

2 contents

Figure1:Occitan dialects. Map created by the Unité mixte de service de l"Université de Toulouse2- Le Mirail, used by permission. The numbers refer to the isoglosses which are described in the text. the inter-dialectal limits which are described as follows (from Patric

Sauzet, p.c.):

Isogloss1separates the Northern Occitan dialects (Lemosin, Auvernhat, Provençal Alpin) from the Southern Occitan di- alects (Gascon, Lengadocian, Provençal): In the Northern di- alects, the Latin velar stops palatalised in front of a back vowel (/ka, ga/ > /Ùa,Ãa/ or /Sa,Za/), while they are preserved as such in the Southern Occitan dialects, e.g. the Late Latin word camm¯inus("path"), of Gaulish origin, has given forms such as /Sa"mi/ or /Ùa"mi/ "chamin" in the Northern Occitan dialects, while the Southern Occitan dialects generally pronounce /ka"mi/ "camin".3 Isogloss2separates Gascon from the remainder of the Occitan dialects: In Gascon, word-initial /h/ developed into /f/, e.g. Latinfacereis /"h?/ "hèr" or /"ha/ "har" in Gascon, but /"fajRe/ "faire" or /"fa/ "far" in other Occitan dialects. Isogloss3separates Lengadocian from Provençal: The definite plural article, which developed from Latinillos, is "los" in Lengadocian, but "leis" in Provençal, e.g. /lu"z4?ls/ "los uèlhs" ("the eyes") in Lengadocian, but /lej"z4e/ or /li"zjø/ "leis uèlhs" in Provençal. Isogloss4separates the Northern Occitan dialects Lemosin and Auvernhat: While in Lemosin, the form of the partitive article

is "de + def. article", in Auvernhat it is only "de" without the3The pronunciation variants given here are not intended to be exhaustive.

contents 3definite article following, e.g. "del bòsc" or "de la lenha" ("wood") in Lemosin, but "de bòsc", "de lenha" in Auvernhat. Isogloss5separates the Northern Occitan dialects Auvernhat and Provençal Alpin: In Auvernhat, the first person plural ending is /-e/ or /-i/, while in Provençal Alpin it is /-u/, e.g. /"Zydi/ in

Auvernhat, but /a"Ãyu/ in Provençal Alpin.

The Occitan dialectological data adduced in this thesis have been gathered primarily from the linguistic atlases covering the Occitan- speaking area, as follows: alp Jean-Claude Bouvier & Claude Martel (eds.) (1975-1986),Atlas linguistique et ethnographique de la Provence. Paris: CNRS. alce Pierrette Dubuisson & Gérard Taverdet (eds.) (1971-1982),Atlas linguistique et ethnographique du Centre. Paris: CNRS. ally Pierre Gardette (ed.) (1950-1976),Atlas linguistique et ethnographique du Lyonnais. Lyon: Institut des Langues Romanes des Facultés

Catholiques & Paris: CNRS.

almc Pierre Nauton (ed.) (1957-1963),Atlas linguistique et ethnographique du Massif Central. Paris: CNRS. alal Jean-Claude Potte (ed.) (1975-1992),Atlas linguistique et ethno- graphique de l"Auvergne et du Limousin. Paris: CNRS. alloc Xavier Ravier (ed.) (1978-1982),Atlas linguistique et ethnographique du Languedoc occidental. Paris: CNRS. alg Jean Séguy & Xavier Ravier (eds.) (1954-1973),Atlas linguistique de la Gascogne. Paris: CNRS. allor Jacques Boisgontier (ed.) (1981-1986),Atlas linguistique et ethno- graphique du Languedoc oriental. Paris: CNRS.

Sound change

This thesis is about sound change, although cases of synchronic varia- tion, taken as a pool for potential sound changes, are also discussed. Just like any academic work, the analyses of the case studies offered in chapters2-5are underpinned by theoretical ideas which I will now present in the following paragraphs. John Ohala has argued that sound change originates in perception (e.g. Ohala 1983a,54,Ohala 1989,179,186,Ohala 2005a,419, among many other publications). When the listener misperceives a signal, either because of lack of attention of some kind or because the signal itself is ambiguous, she is likely to store a new exemplar (for exemplar theory, see below) that diverges in some respect from a previously stored token of the same sound or word-type, which she may then

4 contentschoose to try and reproduce in her own production. This is whatOhala

(1997,47) calls a "mini sound change": a momentary change in the mental representation of some sound, sound sequence or word. Cues for one sound can be mistaken as cues for another sound or several other sounds. When phoneticians conduct perception experiments, they are generally trying to establish which sounds or sound sequences are likely to give rise to some "mini sound change" and how frequent a particular type of "mini sound change" is expected to be. Inevitable variation in articulation, due to variation in speech rate, to various physical constraints, to adaptation to social situations, or to performance errors (slips of the tongue) in production, complicates the picture. Moreover, the speaker may emphasise an articulatory com- ponent of the sound sequence she thought she heard, assuming it was intended by her interlocutor, even when it was not. All this variability in production will lead to more "noise" in the signal.4Listener-speakers, however, are generally aware of all this variability and normally take it into account when perceiving speech. Thus, despite synchronic variation in both speech perception and production being the pool from which sound change (or phonologised variation) arises ( Ohala

1989,175-176,Ohala 1992c,39), the language or dialect remains stable

enough to fulfill its purpose of serving as a medium of communica- tion, without becoming petrified nor changing so rapidly as to prevent efficient communication. This means that only a few very of the "mini sound changes", which abound in natural communication, will give rise to more stable sound changes of the kind discussed in the present thesis. It is evident from the foregoing paragraphs that production and perception are intimately intertwined in the process of sound change. It is a continuous give-and-take between these two aspects of language, and the accounts of the sound changes affecting /l/ in this thesis will try to show this as much as possible. Nevertheless, this thesis is concerned more with the initiation of a particular sound change than with its spread, although questions pertaining to the latter aspect will occasionally be discussed too. Even for ongoing sound changes, such as /ë/-vocalisation in Brazilian Portuguese (see chapter2.1), it is difficult to obtain a general picture answering the questions of which section of the speech community contributes in which measure to sound change, where does it originate (in which geographic location, in which phonetic context), how is it perceived by listeners (does it carry prestige or is it stigmatised), in which part of the vocabulary does it become entrenched first and with which frequencies. This problem is even greater for sound changes which occurred centuries, sometimes millennia ago. Where possible, some attempts have been made to take4 What is noise and what is not in the speech signal is not easy to determine. Given the importance of redundant cues to the identity of speech sounds and their place within the phonetic context of an utterance, as well as the still largely underestimated role of fine phonetic detail, it is perhaps inexact to talk about "noise" at all. contents 5into account the contribution of frequency of occurrence of sounds or sound sequences to a particular sound change (see chapters2.1.3,

3.3.2, and5.3.2).

Frequency effects and considerations are a basic tenet of the usage- based exemplar theory advocated by Joan Bybee (for a very com- prehensive overview see

Silv erman2006,42-46). It is thought that

exemplars, i.e. the detailed representation in memory of specific acous- tic (and perhaps also articulatory) events5, lead to the formation of categories in language: frequent very similar exemplars will form a category prototype, according to phonetic context (see

Bybee 2001,

37-38). For instance, the prototype of a lateral, formed over the most

frequently occurring exemplars, in a front high vowel context may be a very clear, perhaps even slightly palatalised lateral, while the proto- type of a lateral in a back or low vowel context may be a darkish or dark lateral. Given moreover that categories are known to have fuzzy rather than clear-cut boundaries, some exemplars will inevitably be ambiguous with respect to their categorisation with one or the other prototype (a useful model for categorisation of any given exemplar is the Perceptual Assimilation Model of Catherine Best and colleagues

Best et al.2001,777,785-786,789-791)).

In departing in this way from the traditional phonemic principle, which aimed at distinguishing sharply between phonemes and allo- phones, at least in theory, we face a notational problem, namely what do the slashes (/l/) and the brackets ([l]) refer to? In this thesis, I almost invariably use slashes because I consider sounds and sound sequences which are stored in memory and are thus what the listener intends to pronounce and perceives from the actual physical signal. It has to be kept in mind, however, that it is still the actual soundsquotesdbs_dbs45.pdfusesText_45

[PDF] la sociologie est elle une science pdf

[PDF] incipit theatre

[PDF] excipit de l'assommoir

[PDF] l'assommoir incipit commentaire composé

[PDF] scl-90-r questionnaire pdf

[PDF] scl 90 r cotation

[PDF] scl 90 r version française pdf

[PDF] exemple plan de masse coté dans les trois dimensions

[PDF] échelle scl-90 révisée

[PDF] scl-90 r pdf

[PDF] exemple de permis de construire d'un garage

[PDF] plateforme q global

[PDF] scl-90-r questionnaire français

[PDF] exemple notice permis de construire maison individuelle

[PDF] rayon emergent