Dialect Boundaries and the Question of Franco-Provencal [Reprint 2013 ed.] 9027924805, 9789027924803

Citation preview

JANUA LINGUARUM STUDIA MEMORIAE N I C O L A I VAN WIJK DEDICATA edenda curat C. H. V A N S C H O O N E V E L D Indiana University

Series Practica,

147

DIALECT BOUNDARIES AND THE QUESTION OF FRANCO-PROVENCAL 5

by

GEORGE JOCHNOWITZ Richmond College of the City University of New York

1973

MOUTON THE HAGUE · PARIS

(g) Copyright 1973 in The Netherlands. Mouton & Co. Ν.V., Publishers, The Hague. No part of this book may be translated or reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publishers.

LIBRARY OF CONGRESS CATALOG CARD NUMBER: 72-88183

Printed in Hungary

In memory of my father JEROME JOCHNO WITZ

ACKNOWLEDGEMENTS

First and foremost, I would like to thank Professor Uriel Weinreich for his creative guidance and generously given help during every stage of the preparation and writing of this study. His assistance can never be fully acknowledged, and his untimely death was a tragic loss. Those of us fortunate enough to have worked with him are richer for having had this opportunity. I am thankful to Professor Robert Austerlitz for his help with the final corrections of this work, and to Professors Lawton P. G. Peckham and Harvey Pitkin for their careful reading of the manuscript, and for their useful suggestions. Among the many others who have always been willing to be of assistance, I would particularly like to mention Professor Marvin I. Herzog for his technical advice, and Professor Mario Pei for his bibliographical suggestions. I am indebted to Miss Agnes McKeon for her skilled drawing of the maps, to Mr. Ronald Stark, for his excellent typing, and to Mr. Anthony P. Quintavalla for his editorial help. Publication costs were in part paid for by an institutional grant given to the City University of New York by the National Science Foundation, and by the Research Foundation of the City University of New York. I appreciate this assistance. Finally, I wish to thank my wife, mother, children, and friends for help in ways too numerous to list.

TABLE OF CONTENTS

Acknowledgements

7

List of maps

13

Key to terms and symbols

17

1. Introduction

19 PART I. HISTORY OF THE PROBLEM

Introduction to Part I

25

2. The Dialect Boundary Controversy

27

3. Three Investigations of Franco-Provençal Boundaries

34

Orientation Gardette Structural Classifications of Sound Changes Escoffier Lobeck

34 34 39 42 50

PART II. ISOGLOSSES

Introduction to Part II

59

4. The Three Major Isoglosses

61

4.0 Introduction 4.1 Stressed Free Vowels 4.11 Location of Isoglosses 4.12 Discussion 4.121 Western Romance 4.122 Haudricourt and Juilland 4.13 Chronology 4.131 Phonemic Length 4.132 Palatalization of C 4.133 Changes in the Long Vowels

61 61 62 69 69 70 73 73 74 75

10

TABLE OF CONTENTS

4.2 Intervocalic Stops

4.3 4.4 4.5 4.6 4.7

4.20 Introduction 4.21 Location of Isoglosses 4.22 Discussion 4.221 Weinrich 4.23 Chronology Other Vowels Selected Grammatical Traits Bundling at Isoglosses 1, 2, and 3 Lexicon Franco-Provençal Unity

5. Other Isoglosses 5.1 Gascony and the Southwest 5.11 Interrelated Changes 5.111 Discussion 5.12 Other Southwestern Isoglosses 5.121 Initial and Medial Consonants 5.1211 Discussion 5.122 Final Stops 5.1221 Discussion 5.123 Other Final Consonants 5.124 [ké] Before Finite Verbs 5.125 Bundling at Isoglosses 4, 5, and 6 5.126 Lexicon 5.2 Remaining Isoglosses 5.21 Isogloss 7 5.211 Phonology 5.212 Selected Grammatical Trails 5.213 Bundling at Isogloss 7 5.214 Lexicon 5.22 Isogloss 8 5.221 Discussion 5.23 Isogloss 9 5.24 Miscellaneous Isoglosses

76 76 76 80 81 82 83 93 96 105 105 107 107 107 107 112 112 112 115 115 118 118 118 118 125 125 125 129 129 129 135 135 135 138

PART III. HISTORICAL DIVISIONS

Introduction to Part III

151

6. The Substratum-Superstratum Controversy

152

TABLE OF CONTENTS

11

7. Non-Linguistic Boundaries Agricultural Legal Boundaries Houses and Roofs Bresse History and Boundaries The Franco-Provençal Area Gascony Minor Isogloss Bundles

160 160 162 164 167 169 170 172 172

8. Conclusions Digressions into German and Yiddish Criteria and Boundaries Linguistic Change and Dialect Boundaries The Linguistic Position of Franco-Provençal

174 174 175 176 179

Bibliography

181

LIST OF MAPS

3-1 3-2 3-3 I-4 3-5 3-6 3-7 3-8 3-9 3-10 3-11 3-12 II—1 4-1 4-2 4-3 4-4 4-5 4-6 4-7 4-8 4-9 4-10 4-11 4-12 4-13 4-14 4-15 4-16 4-17 4-18 4-19

Area Covered by Lobeck, Escoffler, Gardette Forez Forez-Major Isoglosses Roanne Franco-Provençal Words French Words Provençal Words Lobeck Map A Lobeck Map Β Lobeck Map C Lobeck Map D Lobeck Map E Isoglosses 1-9 Palatal + A Non-palatal + A Ë Ë Ö Ö Loss of -TLoss of -D-PUnstressed Ö Monophthongization of au ô> ϋ i > è, ä, ài è> ä Final Unstressed -A Final Unstressed -Ü Loss of Unstressed Final Vowels Franco-Provençal According to Hasselrot nostron According to Hasselrot

35 37 41 43 46 48 49 51 53 54 55 56 60 63 64 65 66 67 68 77 78 79 84 85 86 87 88 89 90 91 92 94

14 4-20 4-21 4-22 4-23 4-24 4-25 4-26 4-27 4-28 5-1 5-2 5-3 5-4 5-5 5-6 5-7 5-8 5-9 5-10 5-11 5-12 5-13 5-14 5-15 5-16 5-17 5-18 5-19 5-20 5-21 5-22 5-23 5-24 5-25 5-26 5-27 5-28 5-29 5-30 5-31 5-32 6-1

LIST OF MAPS

Imperfect of être Imperfect with b or ν deux Plural Articles Isogloss 1 Isogloss 2 Isogloss 3 lundi, mardi, mercedi Lexicon Loss of b-v Distinction F->h Loss of Voiced Aifricates Loss o f - N R->arr Loss of -p, -k Loss o f - t Final -n Totally Lost al > au ké before Finite Verbs Isogloss 4 Isogloss 5 Isogloss 6 Lexicon Lexicon CA Vowel + s + Consonant Automatic Use of Subject Pronouns Disjunctive Pronouns Isogloss 7 Lexicon J> y Isogloss 8 Checked O PL > py KL > ky Isogloss 9 si > ei ni > ni j > h or jh s> h Lexicon Germanic Settlement

95 97 98 99 100 101 102 103 104 108 109 Ill 113 114 116 117 119 120 121 122 123 124 126 127 128 130 131 132 133 134 136 137 139 140 141 142 143 144 145 146 147 155

LIST OF MAPS

7-1 Law Boundary 7-2 Roof Boundary 7-3 Bresse 7-4 France in 890 7-5 France in 950 7-6 Southeastern France in 1032 7-7 Gascony 8-1 The Dauphiné Fan

15

164 166 168 170 171 171 173 178

K E Y TO T E R M S A N D

SYMBOLS

The ALF signifies the Atlas linguistique de la France of J. Gilliéron and E. Edmont, 35 fascicules. 1 Symbols of the ALF will be enclosed within square brackets [ ]. Words written in French orthography will be italicized. Latin and Vulgar Latin etyma, written in Latin orthography, will appear in UPPER CASE. The terms 'Northern French' and, when there is no ambiguity, 'French', will be used to refer to the dialects of Northern France. The term 'Provençal' will be used for the dialects of Southern France. This term will never be used in this study to refer to the dialects of the province of Provence alone. Here is the Mode de transcription of the symbols of the ALF as found on page 19 of the Atlas linguistique de la France, Notice servant à l'intelligence des cartes:2 Les lettres [a, e, i, o, u-b, d, f, j, k, 1, m, n, p, r, t, v, z] ont la même valeur qu'en français. VOYELLES, -[ce] = eu fr.; [υ] = ou fr.; [è] = e du fr. je. Π = voy, ouverte; ['] = voy. fermée; = voy. longue; [ w ] = voy. brève; [~] = voy. nasale; [~] = voy. demi-nasale; [,] = voy. tonique. les voyelles sans signes de quantité ou de qualité sont des sons dont on n'a pu préciser la quantité ou la qualité. CONSONNES, -[e] = eh fr.; [c] = eh de l'ail. Bach; [ç] = de l'ail, ich; [g] = g dur fr.; [h] = aspirée; [J] = / mouillée; [η] = η mouillée; [r] = r voisine de [ê]; [r] = r grasseyée; [r] = r fortement roulée; [s] = s dure fr.; [ç] = th dur angl.; [w] = w angl.; [w] = cons. de naît; [y] = y fr. de .yeux; [z] = th doux angl.; [j, 1, r] = 1,1, r prononcées la langue entre les dents. Les lettres superposées représentent des sons intermédiaires entre les deux sons marquées. Petits caractères = sons incomplets. Thus, the common ALF symbols listed below have the following IPA values : ALF i é

1 2

IPA i e

(Paris, 1902-1910). (Paris, 1902). 17

18

KEY TO TERMS AND SYMBOLS

ALF

ΙΡΑ

è à á ò ó υ u œ de è

ε a α 0 ο U y Θ œ 3

e

J 3 j θ ö

j

y 5 ζ ί

Ή

In the keys to the maps in Chapters 4 and 5, a fraction rule indicates a boundary between sounds, the symbol above the line referring to the sound used north of the boundary and the symbol below the line referring to the sound used south of the boundary.

1. I N T R O D U C T I O N

For many years there has been dispute over the dialectal fragmentation of France. Some scholars feel that there are two or three major dialect areas ; others say there is simply a good deal of local variation. In the Middle Ages, what is now France was considered to be divided into two language areas: the Langue d'Oc in the south and the Langue d'Oïl in the north. There has traditionally been some antipathy between the Northerner, who is supposed to be cold and industrious, and the Southerner, who is pictured as warm and spontaneous. In the 19th century, a group of writers known as the Félibrige, the most famous of whom was Mistral, decided to write serious literature in Provençal, or Langue d'Oc, or Occitan, as Southern French was variously called. But Southern French was not quite so uniform as the writers of the Félibrige thought, and they soon were faced the problem of whether to write in their local dialects, or use standard Provençal. The former possibly restricted the number of readers ; the latter suppressed local forms in the same way that French did, and so defeated the purpose of writing in dialect. The problem was never solved. The development of comparative linguistics in the nineteenth century inevitably led to a historical linguistic consideration of the dialectal diversity of France. In 1874, Ascoli defined French and Provençal according to the development of Latin stressed, syllable-final A, and in doing so, defined a third entity : Franco-Provençal. 1 Ι·η French, stressed syllable-final A always becomes [e] ; in Provençal it never does ; while in Franco-Provençal the change takes place only after palatals. (Ascoli, like Romanists, includes the velar stops [k] and [g] in the category palatals. This usage is followed in this study.) In looking for a single phonetic criterion to define Provençal, Ascoli may well have had a pre-conceived idea of what the Provençal area should be. The change from A =• [e], however, did not take place except after palatals in the area of Lyons and Grenoble - an area that has not traditionally been considered part of the Midi. So the marché ~ [marka] line was chosen to define Provençal, since MERCATUM is an example of an etymon in which a palatal consonant precedes stressed syllablefinal A, and a third area was defined. This division has been accepted by most scholars ever since. 1

G. I. Ascoli, "Schizzi franco-provenzali," Archivio glottologico italiano, III (1878), 61-120.

20

INTRODUCTION

Nevertheless, we cannot ignore a minority viewpoint which holds that a threefold division of France is arbitrary, and that in fact there is nothing but an enormous transition zone that separates the native habitat of French from the rest of the Romance-speaking world. This latter position has most recently been expressed by Robert A. Hall, Jr. 2 This controversy does not stem from inadequate information or dispute about the data involved. It is in the interpretation of the data that disagreement arises. Different goals and different criteria are implied in the analyses of various scholars. Since these criteria are only rarely stated outright, it is no wonder that no consensus has been reached on evaluating information and deciding what indeed is the correct way to describe the dialectal diversity in France. We will see that any number of isoglosses, representing phonetic, phonemic, morphological, syntactic, or lexical divisions can be drawn on a map of France. Two facts will become apparent from the following chapters : firstly, there are unmistakably isogloss bundles separating French, Provençal, and Franco-Provençal from each other, and secondly, there are other major isogloss bundles in France, as well as any number of non-bundling isoglosses with independent paths. The former group of isogloss bundles proves that France can indeed be divided into three parts. The other isogloss bundles and independent isoglosses force us to ask whether these divisions are meaningful, and if so, are they the only meaningful or the most meaningful divisions that exist? One might decide to weight isoglosses according to linguistic criteria. Thus, a phonemic change could be considered more important than a phonetic change, since its effect on the structure of a dialect is more profound. (A layman, on the other hand, may consider a non-structural change more important. For example, New York City residents, upon hearing a Bostonian, immediately react to the subphonemic difference heard in the vowel in hard, but remain unaware of the collapse of the /a/ and /o/ phonemes in Boston speech.) A structured group of phonemic changes, such as a chain shift, can be considered more important than a single phonemic change. But we may find that a single phonological phenomenon is not represented by a single isogloss, but by many different isoglosses. For example, there is no single line that can be drawn to separate the area where stressed free E became [ye] from the area where it remained [e]. The classificatory efficiency of the criterion is then seriously impaired. This leads us to another question that will be discussed in this study : Do some types of innovations stick together? We shall see that in some instances, series of structurally related changes occur in areas that can be clearly delimited ; in other cases, isoglosses do not bundle; in a few, e.g. E > [ye], a single phonetic change is not reflected by a single isogloss bundle at all. An attempt will be made to see why we have one type of picture in some cases and not in others. Reference will be made to the 2

Robert A. Hall, Jr., "The Linguistic Position of Franco-Provençal," Language, XXV (1949), 1-14.

INTRODUCTION

21

controversy between the Neogrammarians, who held that sound laws take place without exception, and Gilliéron, who stated that every word has its own history. This in its turn will lead to a confrontation of the Stammbaum model with the wavemodel of linguistic change. The techniques used in this study were directly suggested by Marvin I. Herzog's excellent pioneering work on the Yiddish of Northern Poland. 3 Herzog examines synchronic and diachronic data according to both structural and non-structural criteria. In addition, he relates non-linguistic information to linguistic boundaries. Herzog's approach thus represents a synthesis of the historically sensitive and sophisticated study of Sprachlandschaften (which, in German dialectology, succeeds the mechanism and idealization of the neogrammarians4) with the structuralist program for dialectology as first envisaged at the level of phonology by Trubetskoy. 5 Programs for such a synthesis of approaches were drawn up by Weinreich,6 Ivic,7 and others, but Herzog's work seems to be its first implementation on a monographic scale. This book is organized as follows : Part I, "History of the Problem", is a review of the work of previous scholars on this subject. Wherever possible the assumptions implicit in their viewpoints will be considered. Part I is composed of Chapter 2, "The Dialect Boundary Controversy", and Chapter 3, "Three Investigations of Franco-Provençal Boundaries". Part II, "Isoglosses", deals not only with the location of isoglosses and isogloss bundles, but with both synchronic and diachronic structural interpretations of the nature of these bundles. It is composed of Chapter 4, "The Three Major Isoglosses", and Chapter 5, "Other Isoglosses". Part III, "Historical Divisions", shows the nature and location of the various nonlinguistic divisions that parallel isogloss bundles. It is composed of Chapter 6, "The Substratum-Superstratum Controversy", and Chapter 7, "Non-Linguistic Boundaries". Finally, Chapter 8, "Conclusions", relates the dialectal picture of France to similar linguistic situations elsewhere, and, considers the consequent implications for the field of dialectology. This monograph is limited to examining the major dialect split in France. Thus, there will be no consideration of boundaries with Italian, Catalan, or Spanish dialects; nor will there be any discussion of the French of Normandy, Picardy, and Wallonia. 3

Marvin I. Herzog, The Yiddish Language in Northern Poland: Its Geography and History (Bloomington, Ind. and The Hague, 1965). 4 cf. V. M. Schirmunsky (¿irmunskij), Deutsche Mundartkunde (Berlin, 1962), Part I. 5 N. S. Trubetskoy, "Phonologie et géographie linguistique", Travaux du Cercle linguistique de Prague, IV (1931), 222-34; rpt. in his Principes de phonologie (Paris, 1957), pp. 343-50. 6 Uriel Weinreich, "Is a Structural Dialectology Possible?" Word, X (1954), 388^100. 7 Pavle Ivic, "On the Structure of Dialect Differentiation", Word, XVIII (1962), 33-53.

PART I

H I S T O R Y OF THE P R O B L E M

I N T R O D U C T I O N TO P A R T I

It is most interesting to note that the terms Langue d'Oc and Langue d'Oïl, denoting the dialects of Southern and Northern France respectively, had been in use for centuries before anyone tried to define what these entities were or to state where the boundary separating them was located. Modern linguistics did not exist before the latter part of the eighteenth century, and the only speech forms whose existence merited acknowledgment were literary languages. Thus, the question of whether dialect boundaries exist was not asked until attempts were made to map these limits. Chapter 2 explains the history and nature of the dialect boundary controversy. In many controversies, people can disagree violently without ever contradicting each other. That is also the case here. Consequently, Chapter 2 includes not only the opinions of the scholars who debated this problem, but also an attempt to discover the criteria implied by their positions. Chapter 3 is devoted solely to a consideration of four works by three contemporary scholars. The titles of two of these works, La Rencontre de la langue d'Oïl, de la langue d'Oc et du francoprovençal entre Loire et Allier by Mme S. Escoffier1 and Die Französisch-frankoprovenzalische Dialektgrenze zwischen Jura und Saône by Κ. Lobeck,2 immediately tell us that Ascoli's three-way division of France is taken for granted. The three authors in question do not discuss the existence of dialects; they investigate boundaries. Each of the four works discussed in Chapter 3 is the result of a personal investigation on the part of the author. Mgr Gardette, who was in charge of the team working on the Atlas linguistique et ethnographique du Lyonnais, did the fieldwork for his Géographie phonétique du Forez3 in 1937-1939. It and a companion volume, Etudes de géographie morphologique sur les patois du Forez4 (not discussed in Chapter

1 S. Escoffier, "La rencontre de la langue d'Oïl, de la langue d'Oc et du francoprovençal entre Loire et Allier", Publications de l'Institut linguistique romane de Lyon, XI (1958), 1-269. 2 Konrad Lobeck, "Die französisch-frankoprovenzalische Dialektgrenze zwischen Jura und Saône", Romanica helvetica, XXIII (1945), i-xii, 1-318 and 6 maps. 3 Mgr Pierre Gardette, Géographie phonétique du Forez (Macon, 1941), pp. 1-288 ( = Publications de l'Institut linguistique de Lyon, IX). 4 Gardette, Etudes de géographie morphologique sur les patois du Forez (Paris, 1941), pp. 1-82 (= Publications de l'Institut linguistique de Lyon, V).

26

HISTORY OF THE PROBLEM

3 because it includes no composite maps), both appeared in 1941, when Mgr Gardette was Professor at the Facultés catholiques de Lyon. Mme Escoffier, a native of Forez, who worked with Gardette as one of the coauthors of the Lyonnese Atlas, was Maître de Conférences à la Faculté catholique des lettres de Lyon in 1958, the date of publication of her two works. The first 61 pages of Lobeck's study first appeared in 1944, under the same title, as a dissertation at the University of Zurich. The complete work was published the following year. The author states in his Einleitung5 that his work is a continuation of the border study begun by L. Gauchat in his article, "Gibt es Mundartgrenzen?" 5 Lobeck goes on to say that he included in his own work some information found in the Glossaire des patois de la Suisse romande by Gauchat, Jeanjaquet, and Tappolet, and that he was helped by Jud in the selection of points visited and in the drawing up of his questionnaire. The actual fieldwork was done in 1935 and 1936. C. A. Robson's review of Escoffier's works 7 praises these three authors for their thoroughness and skill. "The immense technical competence of the whole Lyonnese school [including Mgr Gardette, of course] is beyond question, and Mme Escofiìer fully shares it". 8 Speaking of Lobeck, Robson says: "Lobeck's impeccable monograph . . . utilises every possible linguistic criterion to delimit this extensive frontier zone between Besançon and Mâcon... ", 9 The works in question deal primarily with data rather than with theory. The authors wish to show us the nature and location of the isoglosses that form the borders they studied. They succeed in doing exactly what they set out to do.

6

Lobeck, p. 1. L. Gauchat, "Gibt es Mundartgrenzen ?" Archiv für das Studium der neuren Sprachen und Literaturen, CXI (1903), 365-403. ' C. A. Robson, "Franco-Provençal: Definitions and Boundaries", Romance Philology, XVIII (1964), 63-75. 8 Robson, p. 65. 9 Robson, p. 64. 6

2. THE D I A L E C T B O U N D A R Y C O N T R O V E R S Y

The terms Langue d'Oc for the dialects of Southern France and Langue d'Oïl for those of Northern France are at least as old as 1284, when the poet Bernart d'Auriac used them to describe the speech of France.1 At around the same time, Dante distinguished between la lingua di si (Italian) and the languages of oïl and oc.2 Neither poet states what he considers to be the defining criteria for these languages. It would no doubt be unfair to say that they felt a language could be characterized by a single lexical item, such as the word for 'yes'. But they fixed on this difference as a handy representative of all others. It is not until one gets to the eighteenth century that anyone tries to describe the areas where French and Provençal are spoken. The Abbé de Sauvages, in his dictionary, under the listing for the word franchiman (Northern French) says : Il est aisé d'assigner à peu près les limites des deux pays : ils aboutissent à une espèce de zone ou de bande qui se dirige de l'est à l'ouest de la France, et qui passe par le Dauphiné, le Lyonnais, l'Auvergne, le Limousin, le Périgord et la Saintonge. C'est à cette bande limitrophe, ou frontière, pour ainsi dire, du gascon [provençal] et du français, que ces deux langues viennent se confondre; et il résulte de leur mélange, dans le langage du peuple, un jargon informe et dur à l'oreille qui n'a rien de bien décidé, ni pour le français, ni pour le gascon: on ne peut les distinguer qu'en s'écartant de la bande et allant vers le nord, ou vers le midi; ils paraissent alors se démêler peu à peu: car le passage de l'une à l'autre langue n'est point brusque; il se fait par des nuances qu'un voyageur attentif peut apercevoir lorsqu'il va par ex. de Paris à Antibes, ou à Perpignan; il voit le français s'altérer de plus en plus à mesure qu'il s'éloigne de la capitale; les idiomes, ou patois des provinces françaises deviennent plus barbares en s'approchant des limites des deux langues ; c'est pourtant encore du français: passé à ce terme, le ton change, le français disparaît, le gascon se développe, il devient insensiblement plus pur; mais au-delà de ce dernier état, qui a quelque étendue, il dégénère et va se perdre également par nuances, d'un côté dans l'italien, et de l'autre dans la langue espagnole. Mais, au lieu de traverser dans ce sens le royaume, on va du levant au couchant, en côtoyant pour ainsi dire les limites des deux langues, on trouvera que les nuances du gascon vont par des bandes parallèles à ces limites: en sorte que le bas peuple, ou les habitans d'une même bande, qui traversent en ce sens le royaume, parlent tous à peu près le même 1 Général Plazanet, "Essai d'une carte des patois du Midi, chapitre premier: Langue d'Oc et langue d'Oïl", Revue de géographie commerciale de Bordeaux (1913), 2 ( = Lefèvre Collection, L). 2 Plazanet, pp. 2-3.

28

HISTORY OF THE PROBLEM

langage, ou sont du même dialecte et s'entendent mieux entre eux qu'avec ceux de la bande voisine, mais plus éloignée de la frontière.3

Thus we see that the Abbé de Sauvages makes three points : firstly, that the border between French and Provençal runs across France from Dauphiné to Saintonge; secondly, the boundaries between Provençal on the one hand and French, Italian, and Spanish on the other are gradual ; and thirdly, there are noticeable dialect changes when one travels from North to South, but very slight ones when one travels from East to West. It is clear that the Abbé de Sauvages is taking the terms français and gascon for granted; it is equally clear that despite this pre-conceived idea he states that Provençal (gascon) has no clear limits at all but blends imperceptibly with the three Romance languages at its borders. He reconciles this contradiction by dismissing the transition dialects as "un jargon informe et dur à l'oreille. . . . " The term franchiman (or gabaï in certain areas), which is the Provençal word for Northern French or Northern Frenchman, itself implies a discernible difference between "we" (Southern) and "they" (Northern). I do not know whether or not there is a corresponding Northern folk-designation for méridional. Does a Bourbonnais peasant think of his Southern neighbor simply as an Auvergnat, or does he identify him with Limousins, Languedociens, etc. ? The next attempt to delimit the areas of the Langue d'Oïl and the Langue d'Oc was made by E. Coquebert de Montbret, a member of the Société des Antiquaires de France. After receiving dialect translations of the Parable of the Prodigal Son from all over France, he òoncluded that the division between the Langues d'Oc and d'Oïl was quite evident. "Et d'abord se présente ici une première division presque aussi tranchée, géographiquement parlant, que celle qui sépare la langue française en général des langues qui lui sont hétérogènes".4 (Ironically, Coquebert de Montbret agrees with the Abbé de Sauvages that the dialect boundary between French and Provençal is as sharp as the boundaries between French and other languages. The Abbé de Sauvages feels that all these boundaries are equally vague.) Coquebert de Montbret then describes the location of this boundary as follows : La ligne de démarcation dont nous avons parlé plus haut, et qui sépare la langue romane des patois du nord de la France, commencerait au S.-O. au bord de la Gironde près de Blaye, où les patois saintongeais confine au dialecte gascon; elle se dirigerait, à partir de là, à travers les départemens de la Charente-Inférieure et de la Charente, vers la partie orientale de celui de la Vienne, et vers la partie septentrionale de ceux de la Haute-Vienne et de la Creuse; puis, entrant dans les départemens de l'Allier, à l'est de celui de Puy-de-Dôme, du nord de ceux de la Haute-Loire, de l'Ardèche et de l'Isère, elle finirait par embrasser la Savoie et la Suisse romande. 5 3

L'abbé de Sauvages, Dictionnaire languedocien-français, nouvelle edition (Grand'rue, 1820), p. 350. 4 E. Coquebert de Montbret, Mélanges sur les langues, dialectes, et patois; précédés d'un essai d'un travail sur la géographie de la langue française (Paris, 1831), p. 21. 5 Coquebert de Montbret, p. 23.

THE DIALECT BOUNDARY CONTROVERSY

29

Coquebert de Montbret's division, somewhat more precise than the one given by Sauvages, differs from the latter's in Eastern France. Coquebert de Montbret includes Savoy and French Switzerland within the Langue d'Oc area, whereas Sauvages says his intermediate zone " . . .passe par le Dauphiné," which implies that Savoy and Switzerland go with the Langue d'Oli. Although Coquebert de Montbret has no doubts about the classification of Savoy and Switzerland, he is nevertheless still unsure about part of the dialect boundary in Eastern France. He says : "Nous manquons de renseignemens suffisans pour continuer vers l'ouest la ligne de démarcation à partir de Genève, afin de la lier à la partie orientale de l'Auvergne. Il faudrait pouvoir déterminer exactement à laquelle des deux grandes divisions de la langue d'Oïl ou de la langue d'Oc doit se rattacher le langage des départements de l'Ain, du Rhône, et de la Loire". 6 This statement implies that the line can be drawn as soon as the renseignemens suffisans are available. The author apparently has accepted the two-way split of France as a fact, and is concerned only with the location of the boundary line. He draws no implications from the fact that he has been unable to classify the speech of three departments. The next few decades saw no further research on this question. But in 1854, a group of Southern French writers formed an organization known as the Félibrige (the origin of this name is obscure). The most famous félibre was Frédéric Mistral. The Félibrige made an attempt to standardize the orthography of Provençal, and to write serious Provençal literature. They published a journal called the Armana Prouvençau, or Provençal Almanac. The second half of the nineteenth century, which was the period of the Provençal Renaissance, was also characterized by strong regionalist and racist feeling in both Northern and Southern France. During the 1870's, at the height of this chauvinistic feeling, two scholars, Charles de Tourtoulon and Octavien Bringuier, were sent by the Société pour Γ étude des langues romanes to go to Southern France and personally explore the frontier between French and Provençal. The work began in 1873 but was halted by the death of Bringuier. Tourtoulon continued the work alone for a while, but was forced to stop because of ill health. Only the Western third of the boundary was explored. The report was dated November 1875 and appeared in print in 1876. It includes a large and detailed map of the area studied. Tourtoulon and Bringuier used different criteria in different localities to determine the French-Provençal boundary. They conclude that it is a single line between the Gironde and the community of Agris, but that there is a belt of mixed speech east of this point. 7 We will see in Chapter 4 that a compact bundle of isoglosses in Western France fans out to form a fairly broad transition zone in Central France. This is the belt of mixed speech to which Tourtoulon and Bringuier refer. Although the authors do not explicitly state their criteria, they probably began their study in the hope of finding 6

Coquebert de Montbret, p. 6. Charles de Tourtoulon and O. Bringuier, Etude sur la limite géographique de la langue d'Oc et de la langue d'Oïl (Paris, 1876). 7

30

HISTORY OF THE PROBLEM

a clear isogloss bundle separating French from Provençal. In Western France they did exactly that. In Central France, despite their original intentions, they discovered a transition zone which they named the Croissant, because of its shape. This term has been accepted ever since. It is to Tourtoulon and Bringuier's credit that they were not prevented by preconceptions from noticing this wide belt of isoglosses. Who knows what they would have concluded had they been able to continue their explorations into Eastern France? In 1874, G. I. Ascoli offered a solution to the disagreements about dialect boundaries in Eastern France by stating that a third dialect area existed in Eastern France that is neither French nor Provençal, although it contains some aspects of both. In an article that must have appeared in 1874, although it is dated 1878,8 he says: Chiamo franco-provenzale un tipo idiomatico, il quale insieme riunisce, con alcuni caratteri specifici, più altri caratteri, che parte son comuni al francese, parte lo sono al provenzale, e non proviene già da une confluenza di elementi diversi, ma bensì attesta sua propria indipendenza istorica, non guari dissimile da quella per cui fra di loro si distinguono gli altri principali tipi neo-latini. 9

Thus, despite the various characteristics that Franco-Provençal shares with French and/or Provençal, Ascoli, by emphasizing sua propria indipendenza istorica, favors a Stammbaum-model analysis of Franco-Provençal - it is a Romance language in its own right. Moreover, the fact that he defines Franco-Provençal in terms of the development of Latin stressed syllable-final A (see Chapter 1) links him to the IndoEuropean linguists at the height of the neogrammarian successes, discovering and formulating phonetic correspondences. Ascoli's article was reviewed most unfavorably by Paul Meyer. To begin with, Meyer denies the existence of dialect areas altogether. He says : "A mon sens, aucun groupe de dialectes, de quelque façon qu'il soit formé, ne saurait constituer une famille naturelle, par la raison que le dialecte (qui représente l'espèce) n'est lui-même qu'une conception assez arbitraire de notre esprit". 10 In addition, he feels that the criteria selected by Ascoli, besides being arbitrary, were not important. Le nouveau groupe proposé par M. Ascoli, groupe qui, on l'a vu plus haut, n'offre aucune unité géographique, échappe-t-il du moins l'inconvénient de réunir des dialectes fort dissemblables? Pas le moins du monde: il réunit des dialectes qui offrent (et encore est-ce toujours bien sûr?) un très petit nombre de faits que M. A. a choisis entre beaucoup, comme particulièrement spécifiques. Il est de toute évidence que le dauphinois ressemble plus au provençal qu'au franc-comtois et au lorrain, et pourtant le lorrain, le franc-comtois et le dauphinois sont englobés dans le nouveau groupe de M. Α., duquel est exclu le provençal. Ces incohérences sont inévitables, quoi qu'on fasse, et c'est pourquoi je suis convaincu que le meilleur moyen de faire apparaître sous son vrai jour la variété du roman consiste non pas à tracer des β

See Sever Pop, IM dialectologie, première partie: Dialectologie romane (Louvain, 1950), p. 176,

n . 1. 8 10

G. I. Ascoli, "Schizzi franco-provenzali", Archivio glottologico italiano, III (1878), 61. Paul Meyer, untitled review in Romania, IV (1875), 294.

THE DIALECT BOUNDARY CONTROVERSY

31

circonscriptions marquées par tel o u tel fait linguistique, mais à indiquer sur quel espace de terrain règne chaque fait. 1 1

Meyer's colleague, Gaston Paris, agreed with Meyer that dialects do not really exist in France. " . . .dans une masse linguistique de même origine, comme la nôtre, il n'y a réellement pas de dialects; il n'y a que des traits linguistiques qui entrent respectivement dans des combinaisons diverses, de telle sorte que le parler d'un endroit contiendra un certain nombre de traits qui lui sont communes, par exemple, avec le parler de chacun des quatre endroits les plus voisins, et un certain nombre de traits qui diffèrent du parler de ces quatre endroits". 12 Meyer and Paris, in the preceding passages, have introduced a question of crucial importance. There has never been a definition of 'dialect' that enabled us to determine whether two given idiolects belong to the same dialect, or for that matter, to the same language. The criteria that have been used to determine dialect classifications are always selected ad hoc. Why pay attention to categories that are arbitrary? One can only discuss the geographic extension of linguistic features. What Meyer and Paris say is both original and true, but not necessarily relevant. An arbitrary category can nevertheless be useful. Moreover, even in an area as linguistically heterogeneous as the southern half of France, there are sizeable pockets of relative homogeneity. A discussion of the extension of linguistic features in a given area would be incomplete if no attention were paid to the fact that many features all extend to approximately the same line. Perhaps Paris was reacting against the strong regional hostilities of his time when he added: Et comment, je le demande, s'expliquerait cette étrange frontière qui de l'Est à l'Ouest couperait la France e n deux e n passant par des points absolument fortuits ? Cette muraille imaginaire, la science, aujourd'hui mieux armée, la renverse, et nous apprend qu'il n'y a p a s deux Frances, qu'aucune limite réelle ne sépare les Français du N o r d de ceux du Midi, et que d'un bout à l'autre du sol national n o s parlers populaires étendent u n e vaste tapisserie dont les couleurs variées se fondent sur tous les points en nuances insensiblement dégradées. 1 3

But Tourtoulon, who had devoted his own time and energy to mapping out part of this boundary, disagreed with Paris, as was to be expected. N ' e n déplaise à M. G . Paris, il y a bien deux langues françaises séparées par une frontière n o n imaginaire. "Et comment, dit-il, s'expliquerait cette étrange frontière qui, de l'est à l'ouest, [sic\ couperait la France en deux, en passant par des points absolument fortuits?" Explicable o u non, un fait est o u n'est pas, et lorsqu'il s'agit seulement de constater s'il existe o u n'existe pas, c'est une déplorable disposition d'esprit pour un observateur que de se préoccuper de l'explication à donner. 1 4 11

Meyer, p. 295. Gaston Paris, talk printed in the Bulletin historique et philologique du Comité des travaux historiques et scientifiques (1888), p. 134. 13 Paris, p. 135. 14 Charles de Tourtoulon, Des dialectes (Paris, 1890), p. 49. 12

32

HISTORY OF THE PROBLEM

L. Gauchat also felt that Meyer's and Paris's definition of dialects was unrealistic. He first gives what he believes are their three criteria for dialects, then attacks them one at a time. Ein unrichtiges Verfahren ist es ferner, wenn man zuerst eine Definition aufstellt und erst nachher sucht, ob so ein Ding vorhanden sei. Das hat man aber tatsächlich mit den Dialekten getan. Man hat gesagt, ein Dialekt müsse charakteristische Merkmale enthalten, die sonst nirgends vorkommen, er müsse von den Nachbardialekten durch ein ganz bestimmten Orten durchgehendes Zusammenfallen mehrerer (wenigstens zweier) Lautgrenzen deutlich geschieden sein. Innerhalb des Dialektes müsse eine ungetrübte lautliche Einheit herrschen. Da dies nicht vorkomme, gebe es keine Dialekte. Betrachten wir jeden einzelnen Punkt dieser Definition. Ich verstehe nicht, wie man Ascoli das Recht absprechen kann, eine eigene Definition der Eigentümlichkeit eines Dialektes aufzustellen, die darin bestände, dass eine gewisse Verbindung charakteristischer Merkmale den Dialekt ausmache, ob diese anderwärts vorkommen oder nicht? Welches Ding dieser Welt hat denn keine Eigenschaften mit anderen Dingen gemein?... Könnte man mit derselbe Logik sagen, die Grenze zwischen alt und jung, reich und arm etc. sei auch nicht festzustellen, und es gebe folglich keine alten und jungen, keine reichen und armen Menschen?... Wie die Mitglieder einer Familie, je älter sie werden, desto verschiedener sich gebärden, so können innerhalb des Dialektes Differenzen entstehen. Bleiben sie untergeordneter Art, so wird dadurch das Familienband nicht zerissen. Aber es kann auch eine Loslösung und Aufhebung der Gemeinschaft entstehen. 15

Gauchat, in the same article, maps a segment of the boundary between French and Franco-Provençal in the Jura Mountains of Switzerland. It is strange that a man who worked at locating a boundary should bring up the analogy with wealth or age, which are areas where significant boundaries do not exist. What is more significant in Gauchat's thinking is his implication, never stated outright, that although it is impossible to formulate a watertight definition of a dialect, the examination of particular criteria will locate meaningful divisions. He does not, however, explain how criteria might be evaluated or weighted. In 1913 Général Plazanet constructed a map showing the division between French and Provençal based on the previous ones that had been made. He divided France into four areas: French, Provençal, and two mixed areas, one essentially French, the other essentially Provençal. In Western France there is no mixed region, but in Central France the two mixed areas appear and grow wider as one goes east.16 Plazanet, nevertheless, does not discuss the term "Franco-Provençal", although he mentions it once.17 For the rest of the twentieth century, there is almost no dissent from the threefold division of France made by Ascoli - the criticisms of Meyer and Paris are virtually forgotten. The scholar Jules Ronjat, a native speaker of Provençal, made a personal 15

L. Gauchat, "Gibt es Mundartgrenzen?" Archiv für das Studium der neuren Sprachen und Literaturen, CXI (1903), 396-398. 16 Plazanet. 17 Plazanet, p. 36.

THE DIALECT BOUNDARY CONTROVERSY

33

study of those dialects which he can understand and those that he cannot. He includes Franco-Provençal dialects among the ones he cannot understand. He describes in detail the place where he finds the boundary between Provençal and its neighbors.18 This boundary is one of those considered by Plazanet. Ronjat is using intelligibility as his criterion. In doing so, he leaves many questions unanswered. Just how easy to understand must two dialects be before they are judged mutually intelligible? And what if comprehension is not mutual, but one-sided? These problems were not explored in linguistics until much later.19 Ronjat leaves another problem unresolved. He says he speaks Mistralian Provençal, i.e. the dialect of the area near the mouth of the Rhone. This area is linguistically and geographically as far as possible from the traditional boundaries Provençal shares with French, Spanish, and Italian, and approximately equidistant from each of them. Would Ronjat not have found different groups of dialects intelligible had he been a Gascon, a Limousin, or a Niçois ? Ronjat states that he finds certain Catalan dialects mutually intelligible with Provençal. Yet he has not included them in his study because " . . .la condition sociale et le développement littéraire leur ont fait une situation très différente de celle des parlers provençaux".20 This quotation betrays the fact that mutual intelligibility is not Ronjat's only criterion. The formation of the Félibríge, and the existence of words such as franchiman, indicate there is Provençal self-consciousness if not chauvinism. It would be most interesting to see the results of a survey in which French people were asked whether they were Northerners or Southerners. R. A. Hall is the one contemporary linguist who dissents from what is now the traditional point of view, and agrees instead with Meyer and Paris. Using several maps from the Atlas Linguistique de la France, he concludes that there is no such thing as Franco-Provençal - merely an enormous transition zone between France and the rest of the Romance world.21 Other contemporary scholars, such as Haudricourt and Juilland, or Martinet, simply take Franco-Provençal for granted without questioning the validity of the grouping. Bengt Hasselrot has made a slight change in the definition of FrancoProvençal, still using a single criterion, as follows : " . . . le franco-provençal est l'ensemble des parlers où A final précédé de palatale devient i (a, e) mais se conserve dans tous les autres cas (ne serait-ce que dans les cas où un déplacement d'accent vers la finale l'a préservé)".22 The three scholars whose works are discussed in Chapter 3, like most contemporaries, accept the three-way division of France as given. 18

Jules Ronjat, Essai de syntaxe des parlers provençaux modernes (Macon, 1913), pp. 2-9. See, for example, Charles Hockett, A Course in Modem Linguistics (New York, 1958), pp. 323-329. 20 Ronjat, p. 12. 21 Robert A. Hall, Jr. "The Linguistic Position of Franco-Provençal", Language, XXV (1949), 1-14. 22 Bengt Hasselrot, "Sur les adjectifs possessifs nostron, vostron en franco-provençal", Studia linguistica, XI (1938), 80. 19

3. THREE INVESTIGATIONS OF F R A N C O - P R O V E N Ç A L BOUNDARIES

ORIENTATION

The areas investigated by the three scholars whose works are examined in this chapter are shown on Map 3-1. Gardette's work was the first of the three to be published, and will therefore be presented first. Escoffier's books are much more recent than Lobeck's study. Nevertheless, since the area she considered is immediately adjacent to Gardette's, her work will be presented second. And continuing to move clockwise, Lobeck's area will be presented third. The nature and scope of the works discussed in this chapter are found in the introduction to Part I.

GARDETTE

Forez is a region in France coextensive with the Montbrison arrondissement, which is in the middle of the Loire department. The Géographie phonétique du Forez1 is a thorough study of the phonological isoglosses in and around this area. These isoglosses are shown on Map 52 of Gardette's study2 which is Map 3-2 of this work. This map includes fifty-one isoglosses, presented in no particular order. Lines 1, 9, 12, 14, and 47 treat changes in Latin stressed syllable-final vowels. Lines 13, 17, 48, 49, and 50 are concerned with the development of diphthongs. Lines 15, 16, and 18-22 indicate differences in the development of Latin intervocalic stops. Lines 2, 24, 37, 38, and 39 treat the development of other intervocalic and postvocalic consonants. Lines 43-46 are concerned with shifts in stress. And lines 25-31 indicate the limits of areas where dental consonants were palatalized before front vowels. All the other lines are boundaries for isolated changes. Here then are descriptions of the areas separated by Gardette's fifty-one isoglosses. With each isogloss, the more innovating or the more complicated of the two areas it separates will be mentioned first. Syllable-final vowels, i.e. vowels in absolute final position or else followed by a 1 Mgr Pierre Gardette, Géographie phonétique du Forez (Macon, 1941), pp. 1-288 ( = Publications de l'Institut linguistique de Lyon, IX). 2 Gardette, p. 258.

THREE INVESTIGATIONS OF FRANCO-PROVENÇAL BOUNDARIES

¡97

Map 3-1

35

36

HISTORY OF THE PROBLEM

single consonant or muta cum liquida (stop+[l, r]) cluster will be called free vowels. All other vowels are checked. 1. East of this line, -ARE, -ATU become [o] ; west of it, non-final A remains [a] ; e.g. PRATU>[prò] to the east, [pra] to the west. 2. West of this line, [a] or some other vowel is always inserted between front vowels and L. Thus FILARE becomes [fyalá] to the west and remains [fila] in the east. 3. East of this line, A preceded by a palatal becomes [i] ; west of it A remains [a]; e.g. COLLOCARE>[kuei] in the east, [kueá] in the west. 4. West of this line, I+final Ν becomes [i] ; east of it we have [i] ; e.g. V I C I N U > [vézi] to the west, [vèzi] to the east. 5. West of this line, O+final Ν becomes [ο, υ, u]; east of it we have [ö]; e.g. MANSIONE^- [mézu] to the west, [mézô] to the east. 6. West of this line, A+final Ν becomes [a] or [o]; east of it we have [ä]; e.g. PANE>[po] to the west, [pä] to the east. 7. West and south of this line, intervocalic -TR- becomes [ir] ; east of it, Τ disappears; e.g. to the west PATRE =-[pwèr] or [pair]; to the east, [par], 8. East of this line, final -U when necessary as a voyelle d'appui, is [ο, υ] ; e.g. ASINU > [án o ] or [ánu], 9. East of this line, stressed free Ë becomes [ié] ; west of it we have [e] ; e.g. PËTRA =• [pyé] to the east, [pé] to the west. 10. East of this line, final-A preceded by a palatal consonant becomes [i]; west of it it remains [a] ; e.g. *PIPITA> [pépi] to the east, [pèpyò] to the west. 11. East of this line, intervocalic -P- becomes [v]; west of it we have [b]. Between lines 11 and 11 bis words with [p] and [b] both occur, such as SAPA>[zàbò] but SAPONE => [ S ^ v d ] , 12. East of this line stressed free Ö beomes [œ, wè, yœ] ; west of it it remains [o]; e.g. SORO=-[sœr] to the east, [sor] to the west. 13. East of this line Ö+[y]>[woi, wai, wé]; west of it we have [ο, υ, u]; e.g. CRÜCE=-[krwé] to the east, [kru] to the west. 14. East of this line Ë and Ë + [y] become [ei, ai]; west of it Ë=>[è] and Ë + [y] =- [i] ; e.g. FIDE>[fèi], DIGITU>[dei] to the east, [fé], [di] to the west. 15. Intervocalic -T- becomes [d] west of this line; it disappears to the east; e.g. VITA=»[vidò] to the west, [vyò] to the east. 16. West of this line intervocalic -C- followed by A becomes [j]; east of it it disappears; e.g. SPICA> [ipijò] to the west, [ipyó] to the east. 17. Stressed AU remains [au] in the south; it becomes [o] in the north; e.g. CALIDU > [ t s a o ] to the south, [tsò] to the north. 18. In verbs such as SECARE, NECARE, intervocalic -C- followed by A becomes [y] to the east, [j] to the west; e.g. [sèyè] to the east, [sèjè] to the west. 19. West of this line, intervocalic -D->[z] in ALAUDA; e.g. [lcézétò]. 20. West of this line, intervocalic -D->[z] in forms of SUDARE where -D- is preceded by stress; e.g. SUDO>[euz].

THREE INVESTIGATIONS OF FRANCO-PROVENÇAL BOUNDARIES JT 16 » » '3 11 „W.

3

FOREZ3

« Gardette, p. 258.

Map 3-2

37

38

HISTORY OF THE PROBLEM

21. West of this line, intervocalic -D->[z] in certain forms of NODARE where -D- is preceded by stress; e.g. NODO>[nòz], 22. West of this line, intervocalic -D->[z] in the infinitives of SUDARE and AUDIRE; e.g. [suza], 23. West and south of this line, intervocalic -D->[z] in the infinitive NODARE. East and north of lines 19-23 intervocalic -D- disappears; e.g. [nuzà] to the west, [nwè] to the east. 24. West of this line, [z][tsàmyò], 25. West of this line, [s] and [z] before [y, i, u] become [e] and [j]; east of it they remain [s] and [z]; e.g. *VECINU>[vèji] to the west, [vézi] to the east. 26. North of this line, [t] and [d] before [y, i, u] become [te] and [dj] ; south of it we have [t] and [d]; e.g. DËCEM>[djèi] to the north, [di] to the south. 27. East of this line, [k] and [g] are palatalized before [i, u, e, a] ; e.g. CULU >[ku], 28. South of this line, [k] and [g] are palatalized before [i] and [u]; e.g. CULU >[kyó]. 29 North of this line, [1] before [i] and [u] remains [1] ; south of it we have [J] ; e.g. LUNA>[Jumò], 30 West of this line, [l] + [u]>[J]; e.g. LUNA>[lunò]. 31. South of this line, [n] before [i] and [u] becomes [n]; e.g. FINIRE>[fini], 32. West of this line C before A > [te]. Between lines 32 and 33 we have [ts] ; e.g. CALENDAS > [tsàlâd]. 33. East of this line C + A becomes [e] ; e.g. CALENDAS > [eàlâd]. 34. North of this line, CL becomes [çj] ; to the south it remains [kl] ; e.g. CLAVE >[çlô] to the north, [klâ] to the south. 35. North of this line, FL becomes [çl] ; to the south it remains [fl]; e.g. FLAMMA >[çlamo] to the north, [flam o ] to the south. 36. Within the areas shown by these lines, final -IVU becomes [i] ; e.g. LIXIVU =»[lési]; elsewhere, we have a rounded vowel, e.g. LIXIVU > [lisy ó]. 37. South of this line, final [1] always becomes [υ] ; north of it final [1] remains [1] ; e.g. FILU>[fyu] to the south, [fi] or [fil] to the north. 38. West of this line Ë + L > [yau] ; east of it we have [yèl] or [yèr] ; e.g. MEL > [mya o ] to the west, [myè] to the east. 39. South of this line, intervocalic [-1-] becomes [r]; e.g. TELA> [tyàrò], 40. East of this line, in words containing [bl] or [pi] and beginning with a dental, an [r] is introduced after the dental, e.g. STABULU > [étrabl]. West of the line we have [établ], 41. Within the area enclosed by this line [y]>[r]; e.g. MEL>[mra], 42. East of this line, final A preceded by a palatal becomes -[i] ; west of it we have -[a]; e.g. VACCA>[vaei] to the east, [vaea] to the west. 43. North of this line, FRAXINU> [frêne]; south of it we have [frès].

THREE INVESTIGATIONS OF FRANCO-PROVENÇAL BOUNDARIES

39

44. North of this line, ASINU> [an] ; south of it we have [az], This is analogous to the change in 43. 45. North of this line, DIEDOMENICA> [démets, dimëe] ; south of it we have [dyamöj, dyamädzo]. 46. North of this line, *COSERE>[kx>dr], *TORSERE> [tordr] ; south of it we have [kuz] and [tors]. 47. East of this line, PEDE > [pi], PETRA=> [pira] ; west of it we have [pié] and [piéra]. 48. North of this line, [a] + [y] preceded by a labial becomes [we], e.g. PATRE [pwèr] ; in the south we have [per] or [payr]. (TR>[yr] in Provençal.) 49. East of this line, E+[y]>[i], e.g. SEX>[si]; west of it we have [éy] or [ay], 50. East of this line, O + [y] =~ [wi] or [wi] ; west of it we have [woi, wai, wé, ο, υ, u] ; e.g. CRUCE> [kywi] to the west, [kru] or [krwé] to the east. 51. West of this line, the stress shifts to the long final vowel of feminine plurals; e.g. STELLA > [ityàlâ], STELLAS > [ityàla].

STRUCTURAL CLASSIFICATIONS OF S O U N D CHANGES

Most of these changes are not isolated phenomena. They can easily be classified into groups. For example, lines 15 and 19-23 are all concerned with the treatment of intervocalic dentals. South and west of these lines, in the Provençal area, intervocalic -D- has become [z] and intervocalic -T- has become [d]. This is an example of a chain shift-either of these changes explains the other. And line 24 is part of the same phenomenon. As -D- was changing to [z], all previous [z]'s in Provençal, from -S- for example, disappeared. It is not surprising that line 24 is east of 15 and 19-23. The area where [z] was lost is greater than the one where -D- becomes [z]. -D- could only become [z] where other [z]'s had already been lost. The chain shift in question is a pull chain in the sense of Martinet. 4 North and east of these lines, in the French and Franco-Provençal areas, intervocalic -[z]- was preserved. But intervocalic -T- and -D- were lost. . Line 11 is clearly related to the ones we have just discussed. West of this line, -Pbecomes [b] - a change analogous to -T- becoming [d]. East of this line, -P- is further weakened and becomes [v], which is not quite the same as the loss of -T- and -Dbut nevertheless an instance of the same French and Franco-Provençal tendency to weaken intervocalic stops. Lines 16 and 18 are a further manifestation of this same phenomenon, although here there is the additional factor of the palatalization of -C- to [ts, te, e] before -A. Nevertheless, the rule is the same : west of the line there is simple voicing to [dz, dj, j] ; east of it there is further weakening to [y]. The fact that affricates and fricatives 4

André Martinet, Economie de changements phonétiques (Berne, 1955), pp. 59-62.

40

HISTORY OF THE PROBLEM

exhibit the same voicing as stops may show that the voicing of intervocalic -C- antedated the palatalization. But this question does not concern us at present. The palatalization of C before A leads us to a different problem. Franco-Provençal, as we remember, was defined by Ascoli as that area where A preceded by C or G is fronted. This separates it from Provençal, where the change does not take place, and from French, where stressed, free A is fronted no matter what the preceding consonant. Line 3, a bit east of the ones we have so far considered, marks this boundary. Line 10, which runs parallel to it, marks an example of a related change : A > [i] after [y] or any palatal consonant. Lines 1, 9, 12, 14, and 47 represent an extremely significant group of changes. According to Wartburg, 5 the essential difference between Northern and Southern French was the result of the lengthening and subsequent diphthongization of stressed, free vowels in the north. South and west of the lines, these vowels remain essentially unchanged. North of them, the development is not uniform, but we can make the general statement that Ë, Ë, Ö, and Ö diphthongized, and that A becomes [o] in much of the Franco-Provençal area and [e] in French. Some of these diphthongs later became monophthongized again (e.g. stressed free Ö and Ö are now both [œ] in standard French), but their development remained distinct from that of unstressed and/or checked vowels. In Provençal, south of line 17, the diphthong AU was preserved. North of this line, consistent with other changes in the vocalic system, AU became [o], filling a hole in the pattern. When, as a result of various developments, a stressed vowel came in contact with a [y], the result was generally a monophthongization in Provençal. The one exception was the sequence A+[y]=-[ay], In French and Franco-Provençal dialects, the result is exactly the opposite - A + [y] is monophthongized but the other vowels + [y] remain diphthongs. These differences are shown by lines 13, 48, 49, and 50. We have seen that Provençal has been more conservative than French and FrancoProvençal, and has preserved reflexes of intervocalic voiced stops. We also saw that intervocalic -S->[z]>0. Similarly, intervocalic -N- has disappeared from local Provençal, as shown by lines 4, 5, and 6. Intervocalic -L- has become [r]. And postvocalic -L has become [υ] except after [i] and [u], where a back vowel was introduced before the -L. These changes are shown by the lines 2, 37, 38, and 39. In some varieties of Provençal, the loss of intervocalic -S-, -N-, and -L- preserved the distinction between simple and geminate Vulgar Latin consonants. Standard French lost the distinction between -NN- and -N-, between -LL- and -L-, and between -RR- and -R-. The loss of -N- and -L- in French would no doubt have created too many homonyms since -T- and -D- had already been lost. Lines 43-46 represent a single innovation in Provençal, where a shift in stress prevented the loss of antepenults in proparoxytons. Lines 25-31 are the northern boundaries of the palatalization of dentals [s, z, t, d, 1, n] before front vowels in Provençal. 5

Walther von Wartburg, Die Ausgliederung der romanischen Sprachräume (Berne, 1950).

THREE INVESTIGATIONS OF FRANCO-PROVENÇAL BOUNDARIES

Map 3-3 6

Gardette, p. 262.

42

HISTORY OF THE PROBLEM

The other lines are isolated phenomena. One of them, line 8, preservation of final -U, deserves to be called to our attention since it has been considered by some to be a defining criterion of Franco-Provençal. Thus we see that many of the fifty-one lines drawn by Gardette are manifestations of related phenomena. It is not too surprising that there should be such significant isogloss bundling. It is still less surprising when we learn that the thickest bundle runs along the Monts du Forez, which in addition to being a natural boundary, have served to mark political and ecclesiastical boundaries as well: between the former provinces of Auvergne and Forez, and between the dioceses of Clermont and Lyons. 7 Map 3-3 (Gardette's map 53) treats only the first seventeen of his fifty-one isoglosses.8 He considered these seventeen to be the oldest boundaries in this area. The others are either recent, or lexical or morphological ; and lexical and morphological boundaries usually do not bundle so clearly as phonological ones, which, as we have seen, are often structurally related to each other. These seventeen old phonological boundaries bundle in a striking way. They clearly indicate, at least at this point, a definite boundary between Provençal and Franco-Provençal. Yet despite this massive bundle, there is considerable diversity between villages on both sides of the boundary. ESCOFFIER

S. Escoffier's study treats the area immediately to the north of Forez. 9 Forez is the middle third of the Loire department-Roanne is its northern third. Mme Escoffier also studied adjacent areas in the departments of Allier and Saône-et-Loire. In the area covered by Mme Escoffier's study, three dialect areas meet, rather than the two of Gardette's. The dialect boundary between French and Provençal is not nearly so sharp as the one between Provençal and Franco-Provençal. Instead, there is a large transition zone, where the isoglosses run more or less parallel to each other. Thus, there are four general areas on the following map, 10 Map 3-4 : French in the north, Franco-Provençal in the southeast, Provençal in the southwest, and the transition zone, or Croissant, in most of the west. As stated in the introduction to Part I, the three authors whose works are discussed in this chapter did boundary studies rather than dialect studies. Consequently, Escoffier, like her two colleagues, does not distinguish between French and Bourbonnais dialects, between Provençal and Auvergnat dialects, or between Franco-Provençal and Lyonnais and/or Forézien dialects. '

Gardette, p. 267. Gardette, p. 262. 9 S. Escoffier, "La rencontre de la langue d'Oïl, de la langue d'Oc et du francoprovençal entre Loire et Allier", Publications de Γ Institut linguistique romane de Lyon, XI (1958), 1-269. 10 Escoffier, "La rencontre...", p. 174. 8

44

HISTORY OF THE PROBLEM

The following lines on Mme Escoffier's map correspond to lines we have already seen on Gardette's map: Escoffier's line Escoffier's line Escoffier's line Escoffier's line Escoffier's line Escoffier's line Escoffier's line Escoffier's line Escoffier's line Escoffier's line Escoffier's line Escoffier's line Escoffier's line Escoffier's line

3 corresponds 4 corresponds 5 corresponds 8 corresponds 9 corresponds 10 corresponds 11 corresponds 12 corresponds 13 corresponds 14 corresponds 15 corresponds 16 corresponds 17 corresponds 18 corresponds

to to to to to to to to to to to to to to

Gardette's Gardette's Gardette's Gardette's Gardette's Gardette's Gardette's Gardette's Gardette's Gardette's Gardette's Gardette's Gardette's Gardette's

line 9. line 12. line 14. line 3. line 8. line 11. line 16. line 15. lines 19-23 line 24. line 2. line 38. line 37. lines 4-6.

Lines 8 and 9, which run north and south on Gardette's map, turn east here, and thus separate Franco-Provençal from both French and Provençal. Both these lines have been used as defining criteria for Franco-Provençal. The remaining lines, 3, 4, 5, 10, 11, 12, 13, 14, 15, 16, 17, and 18 turn west here, and separate Provençal from both French and Franco-Provençal. Two other lines, 19 and 20, also separate Provençal from French and Franco-Provençal. They separate the Provençal areas where [zr] and [lr] may occur from the French and Franco-Provençal areas to the north and east where a stop is inserted and they become [zdr] and [ldr] respectively. These two lines, although their paths are parallel, do not coincide. Gardette did not discuss this phenomenon. Lines 1, 2, 6, and 7 separate French on the north from Provençal and FrancoProvençal on the south. Lines 1, 2, and 7 all concern the treatment of A. South of lines 1 and 2, -ATU and -ATA remain [-a]; north of them we have [é]. And south of line 7, final -A followed by a consonant remains [-a] ; north of it it disappears. Line 6 concerns Ö. South of this line, we find [ό, υ, u]. North of it we find the standard French [ce]. The particular area we are now concerned with shows Franco-Provençal agreeing with Provençal on this phenomenon. In most of the Franco-Provençal area, however, stressed, free Ö developed differently. Mme Escoffier's map of phonological features shows that most of the isoglosses that form the thick bundle between Provençal and Franco-Provençal fan out somewhat and form a less clearly defined bundle separating Provençal from French. Thinner bundles of isoglosses separate French from Franco-Provençal and Provençal, and Franco-Provençal from both French and Provençal. In 1958, Mme Escoffier studied lexical isoglosses in the same area where she had

THREE INVESTIGATIONS OF FRANCO-PROVENÇAL BOUNDARIES

45

done her earlier work.12 She selected twenty French words, twenty Provençal words, and twenty Franco-Provençal words, and made maps of the isoglosses limiting the areas where each of these words is used. The twenty Franco-Provençal words, with French glosses given by Mme Escoffier, are the following : 1. le jomard 2. la lettya 3. les joucles 4. la vogue 5. l'ambrossoir 6. la larmuze 7. l'amboutée 8. la faye 9. la caille 10. la dimanche 11. la cadole 12. étarni 13. l'applayure 14. le creusier 15. la dinsse, le dansillon 16. la polaille 17. le brosson, brossuron 18. la pousse 19. le mollon 20. l'ambre

'l'animal hermaphrodite' 'le petit lait' 'les courroies du joug' 'la fête' 'l'entonnoir' 'le lézard gris' 'la jointée' 'la brebis' 'la truie' 'le dimanche' 'la cabane' 'faire la litière des bêtes' 'la cheville d'attelage' 'l'antique lampe à huile' 'l'agacement des dents' 'la poule' 'le bec (d'un seau, d'un pot)' 'le pis de la vache' 'la mie de pain' 'l'osier blanc' 13

Map 3-5, illustrating these twenty isoglosses14 shows remarkable bundling when we stop to consider that lexical items do not fit into the kind of system that phonological items do. It might be argued that these twenty items are not representative, since Mme Escoffier deliberately chose Franco-Provençal words here, i.e. words not found in French or Provençal. But the facts that twenty such items could be found at all, and that their boundaries coincide so well with each other and with phonological boundaries cannot be lightly dismissed. The fact that these lines are concentric or vaguely parallel to each other seems to show that they all radiate from the same center of innovation. This point is doubtless the nearby city of Lyons. Escoffier's twenty French words are not standard French forms, but rather local, or dialectal French words found in the Bourbonnais area which is the northern part of the area under consideration. The words are as follows: 12

Escoffier, "Remarques sur la lexique d'une zone marginale", Publications de l'Institut linguistique romane de Lyon, XII (1958), 1-214. 18 Escoffier, "Remarques... ", pp. 26^13. 14 Escoffier, "Remarques... ", p. 45.

THREE INVESTIGATIONS OF FRANCO-PROVENÇAL BOUNDARIES

1. hâliyer 2. le coudre, la caure 3. le pot 4. égruger (le sel) 5. roucher 6. la bassie 7. (se) châiller 8. une houssine 9. la chaise 10. la nappe 11. 12. la trace, la haille 13. ajouter 14. se lapper à 15. la croye, le croyer 16. gougneur, gougner 17. l'auberelle 18. le coussat 19. les greutes 20. l'abergeoir

47

'sécher' 'le noisetier' 'la marmite' 'piler le sel' 'jeter, lancer' 'l'évier' 'se couvrir (dans le lit)' 'une baguette souple (pour fouetter)' 'la chaire' 'la fanon de la vache' 'la haye' 'traire' 'commencer à faire (un travail)' 'la pomme (ou poire) sauvage' 'le pommier (ou poirier) sauvage' 'reboutteur, remettre en place' 'le peuplier blanc' 'le houx' 'les cerises' 'le perchoir' 16

Map 3-6, which plots these twenty isoglosses,17 unlike the one showing the FrancoProvençal items, shows little unity or bundling. This is no doubt due to the nature of the area chosen. The Franco-Provençal area, as has long been recognized, has some sort of linguistic unity. Bourbonnais, although a part of the northern French area, has never been singled out for having or lacking particular features. Thus, the twenty items are local terms belonging to a vague area with no particular center - the bundling of the Franco-Provençal items becomes much more meaningful when compared with the disorder of the Bourbonnais words. Escoffier's twenty Provençal words are the following : 1. l'étève 2. les chambailles 3. la noire 4. la chindre 5. trezer 6. gaucher 7. apiter 8. le terme 9. achaber 16 17

Escoffier, "Remarques... ", pp. 54-75. Escoffier, "Remarques...", p. 77.

'le manche de l'araire' 'les jarretières' 'la puce' 'la punaise' 'piler (le sel)' 'tasser' 'attendre' 'le talus, le tertre' 'finir'

48

HISTORY OF THE PROBLEM

THREE INVESTIGATIONS OF FRANLO-PROVENÇAL BOUNDARIES

49

50

HISTORY OF THE PROBLEM

10. aner 11.la yèze 12. mouzi, mouzre 13. la draille 14. l'ambanne 15. l'itombè 16. l'arête 17. l'étouille 18. les retouillons 19. étouré 20. les novye

'aller' 'l'église' 'traire' 'le couloir à bois' 'le fanon (de la vache)' 'l'aiguillon' 'la barbe de l'épi' 'l'éteule' 'les chaumes' 'égoutté, mi-sec' 'les fiancés, les jeunes mariés'20

Map 3-7, showing these twenty Provençal items, is very informative.21 At first glance it seems as disorganized as the one which showed the Bourbonnais words. In the south central part of this map, however, we notice significant bundling. This bundling is in exactly the same place where the phonological bundle separating Provençal from Franco-Provençal is found. And the lexical bundle, just like the phonological one, fans out as it turns west, giving us the same transition area between French and Provençal that we saw before. Thus, the fact that Mme Escoffier deliberately selected words found only in one of the three areas under consideration turns out not to be misleading. The outline of the Franco-Provençal area, and in particular the Franco-Provençal-Provençal boundary, is clearly reflected in the lexical isoglosses. The vague French-Provençal boundary is even vaguer when lexical items are considered. LOBECK

Konrad Lobeck studied the boundaries between the Franco-Provençal dialects and the dialects of Burgundy and Franche-Comté.22 His study ends with six maps. The following are the isoglosses which appear in his Synthetische Karte A, Map 3-8. Vulgar Latin forms are, as usual, written in upper case letters ; the symbols within brackets, however, are the symbols used in the Glossaire des patois de la Suisse Romande. 3. In the Franco-Provençal dialects south of this line, and in most Provençal and French dialects, the sequence Á followed by a nasal consonant, followed by E or U, gives [a]. In the French dialects north of this line, however, this sound is fronted to [ä, è, ë]. South of the broken line, the sequence A + a nasal consonant becomes [ä]; north of it we have [ä, è, ë]. This line runs noticeably to the south of the solid line. 20

Escoffier, "Remarques... ", pp. 81-100. Escoffier, "Remarques...", p. 101. 22 Konrad Lobeck, "Die französisch-frankoprovenzalische Dialektgrenze zwischen Jura und Saône", Romanica helvetica, XXIII (1945).

21

52

HISTORY OF THE PROBLEM

4. South of this line, the sequence 1+nasal+U>[e]; north of it we have [Í]. It is interesting to note that in this case the Franco-Provençal dialects agree with standard French, whereas the supposedly French dialects of Franche-Comté and Burgundy agree with Provençal. 5. petra:pede. In most of the Franco-Provençal area, because of a shift of stress in PEDE, PETRA and PEDE developed differently, giving [pyera] and [pi] respectively. North of the dotted line, however, [i] is found in both words. feru.fera. South of this line, FERU>[fye] or [fyö], and FERA>[fira]. North of it, we have [fir] and [fira] respectively. 6. die lunae. South of this line, U was not fronted to [ü] before N + E , U. Thus, U + N + E > [ 5 ] . North of the line we have [ö, ë]. pluma. South of this line, U + a nasal consonant+A> [ö] ; north of it we have [ö, ë], 7. South of this line, DUOS and DUAS are distinct; north of it they have merged. 8. South of this line, NOVU and NOVE have merged ; north of it they are distinct. This development is exactly the opposite of the one shown by line 7. 9. North of this line, a palatal consonant+-ATUM, -ATAM>[i]. South of it, a palatal consonant+-ATAM > [a], although -ATUM > [i]. 10. South of this line, SICCU and SICCA have different stressed vowels ; north of it they have the same stressed vowel. 11. South of this line, É + N > [ ë ] or [ä]; north of it we have [a]. Map A, Map 3-8 of this monograph, shows that these isoglosses form two bundles : a thicker one running north of a thinner one. In general, south of these bundles, most towns are marked a, b, indicating that the stressed vowels in CLARU, CLARA, -ATA, and -ATAS are [a] or [o]. North of the bundles, the towns are marked A, B, indicating that the vowels in these words have become [e]. In the transition area between the isoglosses, most towns are marked χ, β, indicating that the vowels in these words are [a] and [o] when masculine, [e] when feminine. It seems that border areas often preserve distinctions found in neither of the adjacent areas by using different forms from each area. Map B, which is Map 3-9 of this study, showing additional phonological isoglosses, shows the same bundling found on Map A. In both cases, these isoglosses divide areas that differ from each other in relatively minor ways. The systematic differences between the Forez areas studied by Gardette are more significant than those found here. Two series of changes deal with umlaut phenomena before final -A in FrancoProvençal, and with different developments of nasal vowels in the two areas. Interestingly, the Burgundy-Franche-Comté area seems no closer to general Northern French than does the Franco-Provençal area. Maps C and D, showing grammatical and lexical isoglosses, present the same picture as maps A and B. Map E, which selected representative isoglosses from the four preceding maps, again shows the same pattern of bundling. Lobeck's maps C, D, and E are respectively Maps 3-10, 3-11, and 3-12 of this book.

THREE INVESTIGATIONS OF FRANCO-PROVENÇAL BOUNDARIES

53

PART II

ISOGLOSSES

I N T R O D U C T I O N T O PART II

Map II-l shows nine isoglosses, each of which represents an isogloss bundle. Part II is devoted to an examination of the isoglosses that make up each of these bundles, as well as of other isoglosses, some of which form other bundles and others which do not bundle. Map II—1 shows that Isoglosses 1, 2, and 3 form a miniature system. Each of them, at any point along their respective lengths, is always accompanied by one, never both, of the other two. These three lines reflect the three-way division of France first formulated by Ascoli, and are generally assumed to coincide with the major isogloss bundles in France. The validity of this assumption will be discussed again in the concluding chapter of this dissertation. Be that as it may, Chapter 4 is essentially devoted to a discussion of these three bundles. Two sets of changes that affected French and Franco-Provençal have been of particular interest to historical linguists - the development of stressed free (syllablefinal) vowels and the weakening of intervocalic stops. Both of these phenomena took place north of Isogloss 1. Chapter 4 is consequently divided into the following seven sections. In the first section we consider isoglosses related to the different developments of the stressed free vowels, with special attention to structural explanations that have been offered for these developments, and with an attempt at determining their chronology. In the second section a similar treatment is accorded to consonants. The third section considers the development of some other vowels whose paths follow Isoglosses 1 and 2. Grammatical isoglosses and their locations are considered in the fourth section. The fifth section reviews the bundling found at Isoglosses 1, 2, and 3. In the sixth section some lexical isoglosses will be examined for possible agreement with the isogloss bundles established through the study of phonological and grammatical phenomena. The seventh section considers the problem of Franco-Provençal unity. Map II-l further shows that Isoglosses 4, 5, and 6 form a miniature system analogous to the one formed by Isoglosses 1, 2, and 3. Some consonant changes in this area show signs of structural relationship to each other. Chapter 5 is therefore divided into two parts: Gascony and the Southwest, and Remaining Isoglosses. In the first part of Chapter 5, the interrelated changes are presented first, followed by a consideration of other isoglosses found in Southwestern France. The second part of this chapter is subdivided into first, a consideration of other bundles, and finally, a section on miscellaneous isoglosses. Despite the heterogeneity of the data considered in Chapter 5, the plan of Chapter 4 is followed as closely as possible.

60

ISOGLOSSES

Map I I - l

4. T H E T H R E E M A J O R I S O G L O S S E S

4.0 INTRODUCTION

If the isoglosses which appear in the maps of the present chapter were all superimposed on each other, a complex web of lines would be sure to appear. Nevertheless the pattern is far from chaotic. On the contrary, as the number of features differentiating the speech of France is increased, the isoglosses seem to bundle along a number of prominent lines. In Part III we will explore the historical foundation for the location of these isogloss bundles ; the purpose of the present chapter is to show that the bundles exist, and that they recur again and again as the outer limits of certain phonological, grammatical, and lexical innovations which arose somewhere in the Galloromance area but failed to engulf it all. In choosing subregional developments of a phonological type, we are treading on ground which has been well explored by Romance historical linguistics (although with insufficient attention to geographic detail). Grammatical and lexical developments discussed below do not have the same status of standard examples ; they are phenomena which emerged from our own sifting of materials of the ALF. Since the three major isoglosses with which this chapter is concerned form a miniature system, and since changes affecting stressed free A and the nasal vowels involve both Isoglosses 1 and 2, these isoglosses will be treated as a group and will not be discussed one at a time. We begin the phonological differentiation of France, starting with the vowels (4.1) and going on to the consonantism. Not surprisingly, the stressed vowels of Vulgar Latin have shown more systematic patterns of subregional development than unstressed ones; we return in 4.3 to the less regular developments among unstressed vowels.

4.1 STRESSED FREE VOWELS

Romance scholars define a free vowel as one in syllable-final position. A wordmedial syllable is considered to end with a vowel if a single consonant, or a muta cum liquida cluster, intervenes before the next vowel: a-na, a-kra. Vowels which are not free are called checked.

62

ISOGLOSSES

Stressed free I and U, and the checked vowels, either did not change at all or developed in the same way over all of France. In either case, we can draw no isoglosses. We will discuss isoglosses that separate the descendants of each of the remaining Vulgar Latin stressed free vowels. We will see that the geographic distribution of the descendants differs greatly in systematicity as between the several vowels. Whereas the descendants of A, and to some extent of Ö, show impressive patterning, those of other vowels have far less salient distributions. 4.11 Location of Isoglosses Map 4-1 shows the [a]-[e] boundary of six words which contained free stressed A in Vulgar Latin. Five of these words are infinitives of verbs and one is a noun. In all the words, A was preceded by a palatal consonant. (Embrasser developed out of "IMBRACHI ARE; danser< Galloromance *DANTSARE apparently contained no palatal consonant but developed in the Franco-Provençal area as if it had one nevertheless.) Let us call Isogloss 1 the isogloss separating the southern area where essayer (essaie EXAGIUM), ALF map 483, ends in [a] or [o], from the northern area where it ends in [e] or [i]. Map 4-2 shows the [a]- [e] boundary of three infinitives, a noun, and a past participle. In these words, stressed free A was not preceded by a palatal consonant. Let us call Isogloss 2 the isogloss separating the southern area where emporté (porté < PORTATUS), ALF map 456, ends in [a] or [o], from the northern area where it ends in [e] or [i]. The difference between the development of A after palatals, of course, is the criterion used by Ascoli to define Franco-Provençal. Maps 4-1 and 4-2 show, however, that the isoglosses for the infinitives do not coincide with the isoglosses for the other parts of speech. Thus, the area of France where A became [e] cannot be described by a single line, despite the fact that the change A»[e] has been used to set French, Provençal, and Franco-Provençal apart from each other. Just west of the Franco-Provençal area, at ALF points 801, 802, 803, 804, and 806, infinitives end in [e] while verb forms, as well as other parts of speech, end in [a]. This is true whether or not these vowels were preceded by palatals. At these points, the factor that conditioned the change A > [e] was morphological, not phonological as in the Franco-Provençal area. This morphological utilization of an incomplete sound shift has already been remarked on by Escoffier and Lobeck, both of whom observed that in these border areas, dialect differences are used to preserve the infinitive-past participle distinction, which is lost as one travels away from the border in either direction. Thus, *ALLUMINARE and *ALLUMINATUM are both [alume] in the North and [aluma] in the South, but are distinct in the border area, where for example *ALLUMINARE becomes [alume] and *ALLUMINATUM becomes [aluma].

63

THE THREE MAJOR ISOGLOSSES

65 377 454 455 483 812 M a p 4-1

ottacher danser embrasser empoigner essayer marché

64

ISOGLOSSES

4 5 6 emporté

Map 4-2

65

THE THREE MAJOR ISOGLOSSES

-412 852 101 I 1012 1014 1169 M a p 4-3

dix miel pièce pied p i e r r e (ye everywhere) riviere

ISOGLOSSES

919

«..

ι \ 1512

... . —s'Vio*

ν/ :: - - r r T ^

S I „. «o3^//; 0 j\\,o·' 1 4

+

.

.7. " ' '. 9»·-*. " · e—> ,y 9,7

[Γ.·Κ 94a , x. / .-TV- .·'""..·-'·" X/ ¿ s o /

V

fW'lW

' = «J7 : 950 6.5 : —- — I... 7" ·» «14 ·».-··" \,.J ÎSI*«· I r C li 3 « 636 * " '·••.7V¡.. „0 · " :.¿y-ì?.t.-··' »36 βΐ 7 1 M 1 * 6 36Í 75 7 ' ··* β »" I ..Λ " 770 , ' «44 ·»» ·66 .77 f.77 ·" „, 77. ·>0 „, 664/ „,· «1 ,Λ ... / 656 \ 733 ί·" 1 .7« >·· / ¡.„η 6 / '48 651 «9«.5 . 737 /.65 Ο " 7.« '«· ·40 ·»« ..7 / " · ... V 1 7.3 ,„ ··> 746 47 750 „.·> ... .74 -J759 ββ6 6.3, 675 ... 757 753 75β / ° 755 757 76 Β ... .71 _ 6.3 //6·4 676 760 β.3 ..4 ... •6,0

.,/ ... 691 / ..5 / > « 694 6«S

· '·,"'

7.0 ·' 7 .->14 n>/ ··«·..-·' "s 41»

763 6,6 '

68® .„ 7 , 0

6 9 8