University of Southern California Dissertations and Theses

A Descriptive Analysis Of The Gascon Dialect Spoken At Donzac (Tarn-Et-Garonne), France

(USC Thesis Other)

A Descriptive Analysis Of The Gascon Dialect Spoken At Donzac (Tarn-Et-Garonne), France

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content C o p y r ig h t by R e in e C a r d a illa c K e lly 1966 A DESCRIPTIVE ANALYSIS OF THE GASCON DIALECT SPOKEN AT DONZAC (TARN-ET-GARONNE), FRANCE by Reine Cardaillac Kelly A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Linguistics) June 1966 UNIVERSITY OF SOUTHERN CALIFORNIA T H E G R A D U A T E S C H O O L U N IV E R S IT Y PA R K L O S A N G E L E S . C A L IF O R N IA 9 0 0 0 7 This dissertation, written by . ..................................................... under the direction of hex....Dissertation Com mittee, and approved by all its members, has been presented to and accepted by the Graduate School, in partial fulfillment of requirements for the degree of D O C T O R OF P H I L O S O P H Y Dean DISSERTATION COMMITTEE i\j - s/ ~ r — Chairman TABLE OF CONTENTS Page 0. Introduction..................................... 1 0.1. Geographical Location ......................... 1 0.2. Linguistic Location ........................... 4 0.3. Gascon Dialectology ........................... 6 0.4. The Dialects of Tarn-et-Garonne ............. 8 0.5. The Vitality of Dzcs........................... 13 0.6. Sources of Information...................... 18 0.7. The Field Investigation...................... 21 0.8. Scope and M e t h o d ............................. 23 0.8.1. Phonology .................................. 24 0.8.2. Combining Morphs into Morphemes .......... 25 0.8.3. Primary Forms ............................. 29 0.8.4. The Survey of Derivation ................ 31 0.8.5. Additional Criteria ...................... 37 1. Phonology.......................................... 40 1.1. The Segmental Phonemes...................... 40 1.1.1. Non-Syllabics ............................. 40 ii Page 1.1.2. Syllables................................. 63 1.2. The Syllable................................. 69 1.3. The Suprasegmental Phonemes.................. 87 1.3.1. Stress ................................... 87 1.3.2. Internal Juncture and Pitch ............. 9 3 1.4. The Phonological Phrase...................... 94 1.4.1. Phrase-Stress ............................ 94 1.4.2. External Junctures ...................... 95 1.4.3. Illustrative Text........................ 97 1.4.4. Intonation Contours ...................... 105 1.4.5. Contrastive Accent ...................... 110 1.5. Free Alternation of Phonemes............... 113 2. Borrowing and Bilingualism...................... 124 2.1. The Regional French Phonological System . . . 127 2.2. The Assimilation of Loans.................... 131 3. Words and Minimal Phrases........................ 141 3.1. Words.......................................... 141 3.2. Compounds vs. Phrases........................ 144 3.3. Minimal Phrases............................... 146 iii Page Morphology....................................... 150 1. Basic Morphophonemics ........................ 150 2. Inflection................................... 157 4.2.1. Form-Classes ............................ 157 4.2.2. Nouns and Adjectives ................... 158 4.2.2.1. Gender .............................. 158 4.2.2 .2 . N u m b e r .............................. 161 4.2.2.3. Alternations in Stems............... 165 4.2.2.4. Sub-Classes of Adjectives ........... 172 4.2.2.5. Numerals............................ 176 4.2.3. Pronouns ................................. 180 4.2.3.1. Inflected Pronouns ................. 180 4.2.3.2. Personal Pronouns................... 187 4.2.3.3. Uninflected Pronouns............... 188 4.2.4. Verbs................... 189 4.2.4.1. Morphophonemic Alternations ......... 189 4.2.4.2. Verb F o r m s .......................... 195 4.2.4.3. Classification of Verbs ............. 206 4.2.5. Uninflected Word-Classes ............... 222 4 .2 . 5 .1. Adverbs ............................ 222 Page 4.2.5.2. Pronominal Adverbs .................. 225 4.2.5.3. Prepositions........................ 227 4.2.5.4. Conjunctions ........................ 230 4.2.5.5. Coordinators........................ 231 4.2.5.6. Attributive Universals............. 234 4.2.5.7. Particles ............................. 236 4.2.5.8. Interjections ........................ 237 4.3. Derivation................................... 239 4.3.1. Alternations in Stems ................... 240 4.3.2. Suffixation .............................. 243 4.3.2.1. Nouns ................................. 245 4.3.2.2. Verbs................................. 298 4.3.2.3. Adjectives.......................... 321 4.3.2.4. Adverbs ............................... 334 4.3.3. Prefixation .............................. 336 4.3.3.1. Verbs ................................. 336 4.3.3.2. Adverbs............................... 339 4.3.3.3. Adjectives.......................... 340 4.3.3.4. Nouns ................................. 340 4.4. Compounding................................... 341 v Page 4.4.1. Exocentric Compounds ................... 343 4.4.1.1. Nouns ................................. 343 4.4.1.2. Adverbs .............................. 351 4.4.1.3. Pronouns ............................ 352 4.4.2. Endocentric Compounds ................... 352 4.4.2.1. Nouns ................................. 352 4.4.2.2. Adverbs .............................. 356 4.4.2.3. Verbs ................................. 357 4.4.2.4. Pronouns ............................ 358 4.4.3. Problems of Classification ............. 358 4.5. Reduplication................................. 360 4.6. Homophones................................... 361 5. Phrases........................................... 363 5.1. External Sandhi.............................. 363 5.1.1. General Sandhi .......................... 365 5.1.2. Special Sandhi .......................... 371 5.2. Phrase-Structure ............................ 375 5.2.1. Endocentric Phrases ...................... 376 5.2.1.1. Nominal Phrases ...................... 376 5.2.1.2. Adjectival Phrases................. 383 vi Page 5.2.1.3. Pronominal Phrases .................. 387 5.2.1.4. Verbal Phrases ...................... 391 5.2.1.5. Adverbial Phrases .................... 408 5.2.1.6. Phrasal Prepositions ................ 409 5.2.1.7. Phrasal Conjunctions ................ 411 5.2.2. Exocentric Phrases ...................... 412 5.2.2.1. Prepositional Phrases ................ 412 5.2.2.2. Phrasal Sentence-Linkers ........... 415 5.2.2.3. Phrasal Attributive Universals . . . 416 5.2.2.4. Absolute Expressions ................ 416 Bibliography ......................................... 419 0. Introduction The language under study is a nearly extinct, unwritten patois of Southern France spoken at Donzac, denartement of Tarn-et-Garonne, province of Gascony. It will be called "Donzacais" (abbreviated Dzcs.) after the name of the in habitants of the village. I shall refer to it as a "dia lect" in order to avoid the derogatory connotation and the flavor of illiteracy suggested by "patois." The French parler would be an even better term because of the desired emphasis on speech rather than on reading or writing. Since this dissertation is an independent descriptive study of one dialect, it can only contribute to dialectology by furnishing new linguistic evidence. The relative place ment of dialects in the Donzac area, however, is a knotty problem and an interesting one. I shall therefore devote part of the introduction to this question. 0.1. Geographical Location Donzac, administratively part of the canton of Auvillar, 1 2 is located in the westernmost corner of Tarn-et-Garonne. Situated on a small elevation overlooking the valley, the village attracts vacationers because of its green country side and its historic church steeple, a square turret built in the 14th century by the English occupants. Easily acces sible are Valence-d1Agen, a nearby market—town of some im portance, and Montauban, the capital city of the departe- ment. approximately 31 miles east of Donzac (see map, p. 3). The closest cultural centers are Bordeaux (Gironde), 102 miles northwest of Donzac, and Toulouse (Haute-Garonne), 55 miles southeast. The distance from Donzac to Spain is about 88 miles due south. Donzac is situated on the eastern side of the province of Gascony, across the river from the Bas-Quercy, a hilly region which resisted annexation by the French kingdom until the end of the 15th century. The duchy of Gascony, bounded by the Atlantic, the Garonne River, and the Pyrenees, was united to the French province of Aquitaine as early as the 11th century, but was later disputed for four centuries by the English monarchy. Gascony extended over all or part of eight modern deoartements where the various Gascon dialects are spoken to this day. Donzac, one mile south of the Garonne River, is a rural to Bordeoux QUERQ.V0 #Puyrairol cLosp eyres Loyroc . # Vdlence-dfcgen ^ Donzac _SL-Nicolas-de-la-G ravt Castelsarrasin >£unes • y ' i Auvillor s '* • St.-Joan-du- Bouzet TARN-ET-GARONNE TARN e x ' ^ HAUTE-GARONNE Toulouse K ilom etres 4 community whose farmers grow grapes, fruit, cereals, and vegetables. A post-World War II transition from old- fashioned to mechanized farming methods is rapidly changing the pattern of living. Nearly all the inhabitants except the local merchants are engaged in agriculture. The 1962 census taken by the Secretariat d1Etat au commerce interieur gave Donzac a total population of 514, of whom 201 reside in the village and 313 in surrounding hamlets. 0.2. Linguistic Location The indigenous dialects of modern Gascony have been handed down by oral tradition from generation to generation since the Middle Ages. As a result of the centralizing policy of the monarchy and the growing prestige of the Francian dialect, all other dialects had by the 15th century dwindled to the status of "patois." Dzcs. is linguistically and geographically a dialect of Lanaue d1Oc.1 a group which in the Middle Ages covered over a third of the Gallo-Roman territory. The dividing line ■4?he chief criterion is the form of the affirmative particle equivalent to French oui. The dialects evolved a form such as Dzcs. [o] from the Vulgar Latin neuter demon strative hoc, in contrast to forms of the oil type, which W. D. Elcock derives from the compounded V.L. hoc il (The Romance Languages [New York, I960], p. 87). between Lanaue d'Oil and Lanaue d1Oc having since receded, the number of Occitanian speakers was estimated in 1926 at 10,000,000, that is, approximately one-fourth of the French 2 population. Two of the southern dialects have achieved the status of languages, namely Provengal in the east, most purely spoken in Arles (Bouches-du-Rhone), and Gascon in the west, 3 preserved in its oldest form in the High Pyrenees. Between those extremes lies a large, ill-defined area which may be — * labeled Parlers Lanauedociens after Charles Bruneau. De pending upon the criteria, Dzcs. could be placed either with the central dialects or with the Gascon dialects. Every thing considered, Dzcs. appears to be best classified as a ^This is the figure given for the year 1926 in Sever Pop's La dialectologies aoercu historioue et mdthodes d'en- guetes linquistiques (Louvain, 1950). It would be hazardous to suggest more recent figures. In Gascony, as elsewhere, the vitality of the dialects varies so much from village to village that one would have to count the speakers by direct investigation within a relatively short period of time. Such a task can hardly be expected to be added to the field work for the new atlases, the primary concern there being to rescue what is left of many nearly extinct dialects. ^There is no place here to argue the status of Gascon as a language. It is structurally so different from the surrounding Romance languages that I personally tend to side with Gerhard Rohlfs and Sever Pop, who are in favor of call ing it a language. 6 Gascon dialect, the evidence for which I shall present later. 0.3. Gascon Dialectology The first sample of Gascon as a language different from Provenqal appeared in a multilingual poem written by the ionaleur Raimbaut de Vaqueyras between 1190 and 1203 (El- cock, pp. 387, 390). But only traces of early written Gas con exist. Since the days when the troubadours who were Gascon by birth chose to write in Provenqal, Gascon has been generally thought of as a spoken tongue rather than as a vehicle for literature. This lack of a written standard must have been a prime factor in the formation of the many highly individualized dialects which are still spoken today. A. Dauzat remarks that of all the Gallo-Roman dialects, the Gascon group is the most diversified. 4 A glance at the maps of the ALG shows at once that by vocabulary differences alone a Gascon speaker will make himself conspicuous only a few miles away from his village. Phonetic differences— one of the key criteria for dialect- ologists— often pass unnoticed by the average speaker. Such 4Jean Seguy, Atlas linguisticrue et ethnocrraphiaue de la Gascogne (Paris, 1954- ). is the case with the variants of the nasal phoneme /n/, which may be actualized as [n] in one dialect and as [q], 5 in all positions, in another. Conversely, the presence of unknown units in the phonemic inventory of a neighboring dialect becomes an obstacle to understanding. For example, the native of Donzac who travels seven miles west to Layrac will hear his [fa] 1faire1 pronounced (cf. point 648 on map 596 of ALG) with a glottal fricative which is strange to him: [ha]. He will communicate comfortably in the new speech only after a period of adaptation during which he will learn that only some of his labio-dental fricatives are to be changed to glottal fricatives. Should the Donza- cais travel the same distance south (into Gers), he will hear his [fa] pronounced [hs]. This is one of many such examples. If one adds the far-reaching differences in morphologic structure, it becomes obvious that the speakers of these dialects barely understand one another. It is clear that with such diversity of dialects, any useful contribution to Gascon dialectology must define its ^True, one system then has a phoneme /n/ and the other /rj/. But since any one system has only two nasals in all, palatal /n/ and one other, speakers have been conditioned to make no distinction between non-palatal nasals. territory with the utmost precision. W. von Wartburg, H.-E. Keller, and S. Pop deplore the improper localization of forms in past studies. The common flaw of grammars and dictionaries^ is that the authors tried to cover too large an area. 0.4. The Dialects of Tarn-et- GaiQfrhe The most complex dialectal diversity in all Gascony is found in the left-bank portion of Tarn-et-Garonne where Donzac is located. As indicated by the blank spaces in 7 Jules Ronjat's comprehensive map, this critical area has received hardly any attention from dialectologists, though other dialects of the departement have been at least sketch ily studied since World War I. Those dialects which seem to be more or less related to Dzcs. are, in the north, varie ties of Languedocien and Quercinol (right bank of the Ga ronne), and in the south, various dialects of the Gascon Plain, particularly Lomaane. 6For example Gabriel Roques, Grammaire Gasconne (Dia- lecte de l'agenais) (Bordeaux, 1913), who tries to encompass Gascon and Languedocien. 7Grammaire istorique fsicl des Parlers Provengaux Mo- dernes (Montpellier, 1930-1941). The map appears at the end of Vol. IV. 9 8 In the ALG the points closest to Donzac are 649SO: Saint-Nicolas-de-la-Grave, 11 miles east, and 659NO: Saint- Jean-du-Bouzet, 10 miles southeast. Jean Seguy considers the western area to be only vaguely represented by these two points, since the dialect of 649SO is entirely Quercino1 and that of 659NO entirely Gascon. Referring to the dialectological position of left bank communities in Tarn-et-Garonne, Jean Seguy writes: "En effet, 1'appartenance des communes ... est des plus decon- certantes." The field-work should be "un travail precis et minutieux sur le terrain, commune par commune, et dans cer- 9 tains cas maison par maison." Since no such laborious in vestigation has been done, Dzcs. remains one of the unre corded dialects. Since my knowledge of neighboring dialects is skimpy, I cannot determine which features of Dzcs. are distinctive O °The numbering system of the ALG. a regional atlas for the Nouvel atlas linquistique de la France, is identical with that of Gillieron's ALF. Points which have been added carry the same number, followed by letters indicating direc tion. 9Quoted from a personal letter dated 26 January, 1962 from Professor Jean Seguy of the Institut d1Etudes Meri- dionales de la Faculte des Lettres at Toulouse, author of the ALG. 10 for the area and should therefore be set up as criteria for a synchronic classification.1* ^ I rely, therefore, upon an analysis of Dzcs. kindly communicated to me by Jean Seguy. His classification of Dzcs. rests primarily upon etymologi cal data.'*'1 Positive Gascon features are as follows: vocalization of Latin -L: CALET > Dzcs. kau 'il faut' (transcription of Dzcs. mine, broad phonetic). CAELU > seu 'ciel'; vocaliza tion of implosive L in kauke 'quelque',* loss of R: V0STRU> I bQste 'votre1, VENDERE > bende 'vendre1, but R stays in I VENTRE > bentre 'ventre'; palatalization of S after yod in parese 'paraitre'; preservation of -N as dental nasal: PANE I ^ > pan 'pain', VINU > bin 'vin' (a feature of Lomagne); per fects of the Latin second and third conjugations yield u in faskuri ' je fis' , biskut 'il vit'; BOVES > byus 'bceufs' just as OVOS > yus 'ceufs' . The negative Gascon features are: consistent preservation of initial F-: FACERE > fa 'faire'; ^Should competent dialectologists establish these criteria— perhaps with the additional help of this study— the field-work could be continued at a fairly rapid pace to cover the other dialects. ^1 am indebted to Professor Seguy for his expert assistance, especially for his procuring a tape recording of Dzcs. made in January, 1962, on which his analysis is based. 11 intervocalic -LL- > 1 in akelos 1ces', but final -LL > t: AUCELLOS > auzets 'oiseaux'; the preservation of -N- (versus Gascon lyo < LUNA) is not significant in this area. In summary, Dzcs. has predominantly Gascon features, but the sound correspondences between the Dzcs. forms and the Gascon forms are not always consistent, and there are a few features which are foreign to Gascon. Three questions remain unanswered. Are the non-Gascon traits encountered in other dialects, and if so in which? What would be the results of a classification of the criti cal dialects based on synchronic criteria? Why does this particular area present such an intricate network of dia lects? I cannot answer either of the first two questions, but I believe that the answer to the second would bring to light a revealing set of criteria, phonological as well as morpholog i ca1. The reasons for the baffling position of Tarn-et-Ga- ronne left bank dialects can be sought in the history of the area and in its geographical location. Certain historical events— in addition to the general considerations previously discussed— probably affected this small area more than the rest of Gascony. After an early Romanization (1st c. A.D.), there must have been frequent 12 shifts in population. One of the causes was that nearly every century from the 5th to the 13th brought new invaders. 12 According to Louis Canet, the passage ox the invaders is verified by archeological evidence plentiful in Tarn-et- Garonne. Also, during the feudal system, the viscounty of Auvillar (including Donzac) passed back and forth between Gascon nobles and Lomaane viscounts. This is a possible origin of Lomagne dialectal features among others more typically Gascon. When Valence-d1Agen, Dunes (1269), and Donzac (1271) were built into strongholds IBastides). they were populated by bringing in people from all directions— another source of dialect admixture. What were the influ ences on the southern dialects of Tarn-et-Garonne between the 13th and the 19th centuries is probably not known. Only in 1808 was this deoartement created out of bits of adjoin ing territories, some of which belonged to non-Gascon dia lectal groups. The left bank dialects of Tarn-et-Garonne cannot be said to have evolved in geographical isolation. To the con trary, the country is flat and easy of access, whereas in 1 ° Lectures d'histoire locale sur le Tarn-et-Garonne 'Montauban, 192 5). 13 the isolation of the High Pyrenees Gascon is staunchly preserved. At the entrance of a rich alluvial plain, the southern part of Tarn-et-Garonne was presumably more invit ing to newcomers than the less privileged northern section. In short, this area appears to have been a crossroads of sundry dialectal influences which have left their stamp on today 1s vernaculars. 0.5. The Vitality of Dzcs. Dzcs. was spoken in the 1930's by a majority of speak ers of French descent. Everyone learned to understand the local dialect by hearing it spoken in the fields, in the village post-office, in grocery stores, and also in nearby market-towns. Children, who spoke French with their fami lies, would pick up snatches of the vernacular from servants and neighbors. Those people who were fluent in Dzcs. were not necessarily true bilinguals: some had difficulty in expressing themselves in French; others lavishly interlarded their French with vernacular words. Well-to-do landowners, people with above-average formal education, and school children made use of French most of the time. The local dialect, forbidden on the premises of the school, was felt by all to be more appropriate for 14 discussions on rural topics and for conversations with older people. French was also used by all natives when dealing with outsiders who knew no Gascon or with those whose dia lect differed greatly from Dzcs. It is of interest to note that the locals would not condone many mistakes from anyone who attempted to speak their dialect. They would then switch to French. This concern for proper language, whatever the lan guage, stemmed from the pride that bilinguals took in both idioms. There were occasions for both. French was valued as the formal language, the standard taught in the schools. It appeared superior because there are rules by which to speak it, because it is written, and because there is a prestigious French literature. Parents wanted their chil dren to improve themselves by speaking French, the "nice" 13 language which goes beyond the limits of the community. The vernacular, on the other hand, had a sentimental sig nificance which somewhat made up for its apparent lack of standard. It reflected the love of patrimonial land; it . Pop reports on p. 282 that Proven9al women play a more important part than men in the propagation of French. Here both men and women are equally eager to discourage the use of the dialect by the young. 15 seemed most natural and best fitted for familiar topics. Natives, then, were attached to the old dialect, but they recognized that it was outmoded. That "notre patois est quand meme une belle langue" was the belief of some rare Felibres who wished to see it taught in the schools. On the opposite side stood those who shared the widespread opinion that all patois are vulgar tongues best identified with the lower classes and peasants. The prime factor in the decline of many a Gascon dia lect is that all parents in most communities ceased some time back to speak Gascon with their children. At Donzac this practice must have begun soon after the end of World 14 War I. Consequently, most of today's middle-aged natives no longer use the vernacular; they can follow a conversation in it but they can participate little if at all. The number of speakers of Dzcs. has decreased steadily for other reasons. The nationwide movement that drove countryfolk to the cities began to be severely felt at 1^This is indicated by comparing the ages of children with the language(s) they speak. The S. Canin family is typical. Both parents speak exclusively Dzcs. with each other and with their elder daughter, who was born in 1913, but all speak exclusively French with the younger daughter, born in 1926. 16 Donzac in the late 1930's. The subsequent shortage of rural workers was remedied by the importation of outsiders with extraneous linguistic background. Some could speak a dia lect more or less similar to Dzcs., but the majority were Italians and Spaniards who came as farm-hands and remained as share-croppers or farm-owners. Some of the new arrivals soon began to speak Dzcs. in their own distorted way; at best, they acquired a smattering of it. Finally, the clannish value of the dialect diminished as the villagers changed their way of life. With improved means of communication and with a mixed population, the vernacular was gradually felt to be passe. Sentiment and tradition which tended to keep it alive lost their force. By 1960, the number of speakers of authentic Dzcs. dwindled to a handful of natives past 65 years of age. As these people die out, the dialect will die with them. Another important aspect of the lack of vitality of Dzcs. is the change in usage which I found to be striking, particularly in the past ten years. Upon returning to Don zac after a six years' absence, I noticed that in the case of doublets the French word is now used more often than the native word. Also, certain lexical items have been aban doned and replaced with a minimum of adaptation by a French 17 word. At times I found myself using more native words than did the older speakers. As I was trying to elicit Dzcs. forms and failing to do so, I finally produced the words myself. My informants agreed that those were "good" words indeed, but rarely used today: "ces mots, tu les attrapes au vol." The penetration of French, which caused the changes in Dzcs. usage, must be distinguished from linguistic borrow ing. When a native form is ousted by a French form, the dialect can be said to lose vitality. But this is not the case when the dialect adds loan-words to its vocabulary. Especially in building up its peripheral vocabulary, Dzcs. has used French extensively as a lending language. The number of French loan-words which have been admitted to Dzcs. indicates that borrowing has been more frequent than original invention. This mode of enrichment can be inter preted as a sign of weakness in this language, pointing ultimately to absorption. Linguistic inertia, combined with the inferior position of the language because of the social pressure of French, led to the present stage, wherein the language has declined by losing its popularity as well as some of its vocabulary. One can surmise that with the social change of the 20th 18 century this language would have had a limited future even without the recent decrease in the number of speakers. 0.6. Sources of Information In order to obtain reliable information about Gascon dialects, one has to go out among the people. Written sources such as the works of the Acrenais poet Jacques Jasmin or an occasional compilation of poems written by a local blacksmith must be discounted, for written sources exhibit enough dialect admixture to forbid a proper localization of forms. The paucity of reliable informants at Donzac presents an investigator with a major problem. Once the non-natives have been deducted, the investigator is left with a small number of potential subjects. But some have resided else where, many others have married outsiders; hence the danger of mutual compromise between closely related dialects is present. Fortunately, I can test the reliability of a prospective informant because of my knowledge of Dzcs. and because I have not been exposed to any other dialect. I was born in Donzac and lived there until age seven. Until 1957 I lived a part of each year at our city residence in Bordeaux, the other part on our Donzac estate. I lost 19 contact with the dialect in 1957, when I came to the United States. At present my vocabulary is limited, but I can still carry on an everyday conversation in the vernacular and I understand it thoroughly. I began a preliminary collection of data in 1960 with Agnes Cardaillac (AC), my mother, who now resides in Cali fornia. AC, age 77, was born in Laspeyres (Lot-et-Garonne), 2.5 miles west of Donzac, but at the age of 18 she married a native of Donzac and began to speak Dzcs. Her residence there was uninterrupted for over 25 years. I found that some variants from the dialect of Laspeyres (a right bank community) had slipped into her predominantly Dzcs. idio lect; the data elicited from her had to be re-checked and complemented in the field. Informant AC has been particularly helpful in prepar ing a tentative questionnaire and in gathering a culturally pertinent vocabulary of some 2,000 entries. With the in tention of compiling a glossary, I later added new data in the field while filling the various gaps in the morphology. Early in my investigation, the selection of informants was done partly by correspondence and partly by means of tape recordings. Several of the natives solicited were very helpful in suggesting the best speakers, that is, those 20 who have not been subjected to dialectal influences. How ever, not all those consequently selected showed a willing ness to serve as informants, but three of them accepted an invitation to participate in tape recordings which were made in May, 1963. The material to be recorded was left largely at the discretion of the group, my purpose being to obtain as large a sample of their speech as possible. Aside from passages to be translated from the French and the recitation of a few proverbs, the bulk consisted of lively conversa tions on rural topics and on current events. All three informants— Gustave Delpech (GD), 70, Jeanne Esther Delpech (ED), nee Rajalot (wife of the former), 67, and Daniel Pascalet (DP), 66— are farmers who were born and raised in the community and have lived there continuously. The first two reside in the heart of the village, and DP resides a quarter of a mile north. Upon my arrival at Donzac, I was able to secure two additional informants. Both are literate farmers, born at Peninet, a hamlet half a mile south of Donzac. They are 15 Marcelle Canin (MC), nee Buisson, 71, and Marguerite • L5I soon found out that in the case of this mixed mar riage, MC1s genuine Dzcs. has remained untouched by her husband's different dialect. On occasion, it is he who 21 Videau (MV), nee Soubsol, 80. I selected GD and MC as primary informants because of their diction and their willingness to serve as informants. The other informants were used chiefly to check on doubtful forms and on discrepancies in the responses of the primary informants. In this way I searched for the possible exis tence of several different dialects in the community. 0.7. The Field Investigation During my two months' stay at Donzac in the summer of 1963, I tried to gather a maximum of reliable data by re checking the data I had already collected and by adding as much vocabulary as possible. The techniques I used to elicit data differed only in details from those in common practice. By listening to the conversations of natives in the home in which I lived, I had to ask fewer questions. I used the direct method and pointed at objects or placed linguistic forms in a context which the informant and I had in common. For example, I asked: "What did your father say when this happened?" Verbs and some verb tenses were elicited by my asking accommodates his speech to that of his wife. This is in accord with E. Haugen's observation of similar cases among American-Norwegians. 22 questions in the dialect, such as "What do you use this for?" or "What would you do if . . .?" In addition, my in formants volunteered to give much useful information, which at times they had gathered between sessions. For many points of morphology, though, translations from the French turned out to be the only way of avoiding roundabout answers and of not influencing the informant by my own errors. Paradigms were elicited partly by substitutions in a sentence frame and by expansions and partly by other means, because too much repetition soon tired the attention of the informants. Furthermore, the lexical elements in a frame had to be changed often enough to keep the informants un aware of the purpose of the exercise. If I failed to do so, they began to anticipate the next question and occasionally to respond with a new type of overcorrection, which consist ed in creating some formal difference between morphemes out of their reluctance for homophones. I sometimes had diffi culty in preventing them from becoming introspective about language. Much of the material that the informants furnished was unknown to me. Instead of relying on their proposed trans lations, I asked them to illustrate their meanings by em ploying full sentences and by providing some sort of 23 cultural context. This they did well, and, by their com ments, they often indicated an obvious "feel" for the structure of their language. 0.8. Scope and Method The purpose of this dissertation is to examine the phonology and the morphology of the Dzcs. dialect; the syn tax is therefore outside the scope of this study. However, since the forms of Dzcs. words vary considerably with their environment, it seems indispensable to include a treatment of sandhi phenomena and to examine the structure of phrases. Inasmuch as this dialect is little known, I shall in clude a good many examples to illustrate the descriptive statements and shall present evidence in support of my in terpretation of linguistic facts. All examples will be glossed in standard French. These glosses are often imper fect approximations of the meanings of Dzcs. words, but a Gallo-Romance language like French serves better than, say, English. Where no other source could be found, the French equivalents for certain Dzcs. terms of botany and for names of agricultural tools have been furnished to me by Sylvain Canin, husband of informant MC. It is necessary at this point to present some of the 24 descriptive techniques which will be used in this disserta tion. The choice of one method of analysis rather than another has been guided by the general principle that the method should be tailored to fit the language under study rather than the other way around. 0.8.1. Phonology Generally speaking, the phonetic transcription of the sounds of Dzcs. used in this study is adequate for the con struction of the phonemic alphabet, but admittedly broad. Only those phonetic features which are of consequence for the selection of a given phonemic interpretation are dis cussed in some detail. Whenever the phonemic analysis raises no problem, only the main allophones are described. Phonetically similar segments are subsumed under a single phoneme when their distribution is non-contrastive, that is, when such segments are in complementary distribu tion, in partial complementation, or in free variation. Free variants include idiolectal variants and sounds which are interchangeable for all speakers. The phonemic inven tory based upon these premises accommodates the speech of all informants, whatever the rate of delivery. Certain features which distinguish slow from fast speech will be 25 noted, but clear-cut differences between the two are con spicuously lacking. The choice of phonetic and phonemic symbols has been largely guided by common practice among American structural ists. In the interest of economy and visual neatness of typewritten text, however, the total number of symbols and diacritic signs used has been reduced to a minimum; hence, for example, the choice of the IPA symbol y (rather than u) for the high front rounded vowel. Whenever feasible, the same symbol is used to represent one of the allophones of a phoneme as well as that phoneme itself. The distinction between the two is simply made by placing phonetic symbols in square brackets and phonemic symbols between slant lines. In the morphophonemic notation, capital letters and dia critics are preferred to more artificial symbols. 0.8.2. Combining Morphs into Morphemes Morphs are classified as allomorphs of the same mor pheme provided three main requirements are met: non-con trastive distribution, close semantic relationship, and occurrence in parallel formations. 26 16 As proposed by Harris, morphs qualify as members of a single morpheme if their distribution is non-contrastive, that is, if they are mutually exclusive or in free varia tion. The optional forms of whole words such as /dambe/ ~ /kme/ 1avec1, which have no morphemic relevance, will be presented in a special section, "Free Alternation of Pho nemes ." Scholars have focused attention on the delicate problem of defining with some accuracy the concept of semantic simi larity to be used in the identification of morphemes. Whether two forms are "same" or "different" is often debat able. No satisfactory solution has been reached, and yet some semantic criterion must accompany the examination of form and distribution. Recognizing that observable differences in meaning should not be brushed aside in the analytical process, I propose to require that, if not semantically identical, two morphs must be very closely related in meaning to be com bined into a single morpheme. Also, homophonous forms which belong to different word-classes are here considered to show 16Methods in Structural Linguistics (Chicago, 1951), pp. 198-201. 27 a sufficient semantic difference to analyze them as separate morphemes. I am referring here to homophones comparable to English bite (noun) and bite (verb), which I prefer to treat as different morphemes. But a verb such as bite is regarded as one morpheme both in "he bites" and in "the saw bites," since the two morphs can be subsumed under the definition 'make a sharp incision' and the slight semantic difference can be attributed to the difference in environment. On the 17 other hand, I agree with Bolinger that homophones of the same word-class which have remotely related meanings must be analyzed as separate morphemes and that in such cases some environment can be found wherein they contrast. The criterion of semantic identity or near identity is, 18 like similarity, difficult to define, but I feel that in •^In "Complementation Should Complement," Studies in Linguistics. VIII (1950), Bolinger objects to Nida's treat ment of pill (medicine) and Pill (objectionable person) as one and the same morpheme occurring in different environ ments . Bolinger notes that there is contrast in "What a pillI" and that the two nouns could be shown to be related only by means of a metaphor or pun. A comparable example is that of Dzcs. /bestjo/, best regarded as two homophonous nouns but distinct morphemes, one meaning 'animal', the other 'simpleton'. 1 f t See, for example, D. L. Bolinger, "Identity, Simi larity, and Difference," Litera. I (1954), 5-16, and Sol Saporta, "Morpheme Alternants in Spanish," Structural Stud ies on Spanish Themes, ed. H. R. Kahane and A. Pietrangeli 28 its application there is less room for subjective interpre tations . It is true that working with the more stringent of the two criteria now in use results in enlarging the size of the morphemic inventory. But perhaps the principle of economy of inventory should be applied only insofar as it is compatible with a regard for the current meanings of the forms under study. A third criterion is necessary when the distribution of allomorphs is morphologically conditioned. To be as signed to a single morpheme, two or more morphs must occur 19 "in parallel formations." For example, the morphs /-a/ in Dzcs. /ma/ 'raa1 and /-o/ in /nosto/ 1notre1 are condi tioned by the specific base morphemes /m-/ and /nost-/ re spectively; these morphs are analyzed as two allomorphs of the morpheme feminine because they occur with that meaning, in complementary distribution, and immediately following possessive adjective bases. It will be noted that members of a single morpheme do not necessarily have to be phonemically similar; in other (Salamanca, 1959), pp. 27-28. 19H. A. Gleason, An Introduction to Descriptive Lin guistics (New York, 1961), p. 88. 29 words, allomorphs may exhibit partial or total suppletive alternation. This is permissible and useful in the descrip tion of inflected forms. Morphs which occupy given posi tions in an inflectional paradigm clearly belong to the system even if they have anomalous shapes. For example, Dzcs. /baw/ 1je vais1 is a suppletive verb form which oc curs— in lieu of an expected regular form */ani/— as first person present in the conjugational system of the verb /ana/ 'aller'. On the contrary, the usefulness of suppletion is limited when it comes to setting up derivational paradigms. Formal similarity as a criterion in the analysis of deriva tives will be discussed later on. 0.8.3. Primary Forms In analyzing complex words, Hockett has objected to setting up any one allomorph of a base or affix as primary, on the grounds that this is tantamount to granting to that form a historical priority over other allomorphs of that morpheme or other morphemes in the word. According to him, one should not say, for instance, that English baked is "formed" from bake by a "process of suffixation"; this is 30 20 the model of description called "item and process." Jerrold J. Katz has recently replied to Hockett that such 21 ordering is to be construed as "psychological reality," since, as he demonstrates, certain operations in the speak er's brain need to be performed before others. Although speakers can hardly be expected to be con scious of the mental processes by which they form one word from another, observation shows that on occasion they prob ably are. One informant clearly identified— in his own words— the Dzcs. suffix /-e/, meaning 'tree', out of a series of examples by starting consistently from the name of the fruit; thus, /pumo/ 'pomme' : /pume/ 'ponunier1; /pero/ 'poire' : /pere/ 'poirier'. For this speaker, who is un aware of the loss of /-o/ of, say, /pumo/ in /pume/, the word /pume/ contains the morpheme /pumo/, and the name of the tree is actually formed from the name of the fruit by a derivational process. This and other similar examples that native speakers volunteered to give point to the "item and process" model 20"two Models of Grammatical Description" Word , X (1954), 211. 21"Mentalism in Linguistics," Language. XL (1964), 133. 31 of description as a realistic one for a statement of present day word-formation. Accordingly, /pumo/ is here described as the primary form of the base; the final vowel /-o/ is automatically deleted to avoid hiatus when the suffix /-e/ is appended to form the derivative /pume/. The principal criteria for the establishment of the primary forms themselves, however, must be simplicity and descriptive convenience. These criteria are particularly useful in describing sets of inflected forms. So, for example, the allomorphs of adjective bases are most suc cinctly described by starting from the base as it appears with the feminine inflection rather than with the masculine, that is, from /balent-/ rather than /balen/ 'travailleur1. The relationship between the allomorphs of the base is then statable in terms of the behavior of the last consonant of the primary form. 0.8.4. The Survey of Derivation Whether a set of words should be regarded as members of a derivational paradigm or as monomorphemic lexical items is often difficult to determine objectively. On what grounds is the investigator to decide which prospective de rivatives actually belong in a descriptive account of 32 derivation? Certain structuralists, among them Saporta (p. 29) relinquish resorting to the insight of native speakers be cause of the danger of obtaining such preposterous responses as that of the speaker of English who construed sex and sextet as related. Should he use his own judgment alone, the investigator may be influenced by diachronic considera tions— a reproach frequently addressed to descriptivists. A. Martinet would rather not rely upon the "linguistic 22 feel" of either the informant or the analyst. He points out that an acceptable analysis of derivatives should be based on formal similarity but that an arbitrary decision would have to be made as to the amount of similarity re quired. His suggestion, then, is to limit the survey of derivation to the study of "linguistic behavior," that is to say, spontaneous innovations of the type of French Petain > petiniste, thereby taking account exclusively of active formatives. Unfortunately, the Dzcs dialect does not lend itself to this attractive approach. The number of coinages which can be safely identified as such is negli- About Structural Sketches," Word. V (1949), 31-32. 33 gible and productive affixes are relatively few. By way of I compensation, the dialect borrows terms like /petinisto/ 'petiniste' from French, so that the examination of most new derivatives offers no interest other than that of re vealing the existence of patterns of loan assimilation in the bilingual process. This will be the subject of a sep arate chapter. Nevertheless, a good many of the indigenous analogical formations of the past, those which have under gone little or no psychological transformation, are still recognized as derivatives by present-day speakers. The derivational processes involved can therefore be regarded as synchronieslly pertinent even though today their produc tivity appears to be potential rather than real. I propose a treatment of derivation which takes into account on the one hand the phonemic similarity of the derivational data and on the other the morphemic interpre tations that the informants have indicated by their own reactions to tentative derivatives. On the basis of many plausible responses from various informants, I have come to the tentative conclusion that in determining whether one word'was derived from another the decisive factor for the native speaker was usually the ob vious recurrence of a clear-cut modern meaning rather than 34 the close similarity of the forms themselves. Thus, it would be contrary to the views of the informants to analyze, for instance, Dzcs. /syrprenge/ 'surprendre1 as a derivative of /prenge/ 'prendre' or /anej/ 'aujourd'hui' as a deriva tive of /nej/ 'nuit1. The members of these pairs are not understood at present as having any semantic affinity, so that they are best regarded as distinct morphemic units, although they are historically related and contain phonem- ically identical elements. Similarly, the partial /-sebe/ in Dzcs. /resebe/ 'recevoir' which recurs in /apersebe/ 'apercevoir' (cf. -ceive in English receive, perceive) has no morphemic relevance since it is not recognized as carry ing independent meaning. On the other hand, forms of the type of the Dzcs. adverbs /arynan/ 'l'an dernier' and /arse/ 'hier soir' are treated as derivatives in accord with the native speaker's opinion that the partials /an/ and /se/ are semantically identical with the nouns /an/ 'an' and /se/ 'soir' respectively. If these adverbs are complex words, then, to be consistent, morphemic status also should be granted to residues like the /engw-/ of the adverb /engwan/ 'cette annee'; like cran- in English cranberry. these resi dues are non-recurring morphemes. The preceding examples suggest that recourse to the 35 "feel" of native speakers is advantageous because it dis cards interpretations not warranted by the present meanings 23 of the forms. However, since the native speaker's primary concern seems to be with meaning rather than form, the data cannot be considered as prospective derivatives unless they have been previously tested for formal similarity. This requirement was met in the examples cited in the preceding paragraph, since the elements of each pair tentatively re garded as one common stem were phonemically identical. How much formal similarity is here deemed to be suffi cient can only be stated in general terms and illustrated with examples. There is no avoiding a somewhat arbitrary decision. In the first place, this criterion need be ap plied only to stems; affixes may have phonemically disparate allomorphs in complementary distribution. Thus, a given stem may appear with a suffix /_o/ meaning 'female' or with a suffix meaning 'male' which has the allomorphs /-e/, /-at/, or zero, e.g. /droll-o/ 'fille' : /droll-e/ 'gargon'; ^The adverb /arynan/ 'l'an dernier' just cited is a case in point. It could be analyzed as three rather than two morphemes: /aro/ 'maintenant (il y a)' plus /yn/ 'un' plus /an/ 'an' with normal loss of /-o/ in /aro/. However, this analysis cannot here be admitted since it rests upon historical considerations. /awk-o/ 'oie' : /awk-at/ 'jars'; /pjot-o/ 'dinde' : /pjot/ 'dindon'. Considering the over-all structure of the dialect under study, stems may vary in shape and yet suggest that a derivational pattern is present. On the whole, the alter- nate forms of a stem must show little phonemic variation and must be of similar configuration. A stem may recur in a curtailed or in an augmented shape, but total suppletion is here ruled out as a possible type of alternation. Where a derivational relation is not otherwise well established in the system and where a morphophonemic replacement which is not phonologically determined is involved, the confirming criterion is the recurrence of that alternation elsewhere in the language. For instance, /pul-o/ 'poule' (female) and /put/ 'coq' (male) are said to have a common stem be cause the alternation /l ~ t/ is found at other points in this dialect. Here are a few pairs of words which, though comparable to those just cited, are not regarded as being derivationally related because of insufficient formal simi larity: /kabalo/ 'jument', /sibaw/ 'cheval'; /trezo/ 'truie', /tesun/ 'cochon'; /bako/ 'vache', /byw/ 'boeuf'; /fenno/ 'femme', /ome/ 'homme'; /sawmo/ 'anesse', /aze/ 37 0.8.5. Additional Criteria In this study the morphemic analysis is based on the principle of a one-to-one correspondence between morph and morpheme. This precludes making use of Hockett's device of the portmanteau morph, a morph which is assignable to sever- 24 al morphemes. The device consists m treating a form like French .gji, i.e. /o/, which carries simultaneously the mean ing of the preposition jk and that of the definite article I s . , as an allomorph of both of these morphemes. Comparable Dzcs. forms, which are less compact, are regarded as combi nations of two morphs so that, for instance, /aw/ 'au1 is analyzed as containing the preposition /a/ 'a' and a morpho logically defined variant /w/ of the definite article /lu/ 'le',25 The use of zero elements is limited to a minimum. A morpheme may have a zero allomorph if it is represented elsewhere by an overt shape, but no use is made of zero stems or zero morphemes. The presence of an overt change, including a minus feature, is given precedence over a zero 2^See Saporta, p. 31. 25«rh.is interpretation is similar to Hall's treatment of Italian. See Robert A.Hall, Jr., Descriptive Italian Grammar (Ithaca, New York, 1948), p. 37. 38 affix. Much simplicity is gained in the description of complex words by treating the arrangements of phonemic stresses which accompany these words, i.e. the stress patterns or superfixes, as separate morphemes. The immediate constitu ents of a verb form like /ajmabi/ 1j'aimais1 are a sequence of segmental phonemes (itself made up of stem plus two in flectional suffixes) and the stress pattern Since, except for compounds, there is only one strong stress per construction, all other stresses being phonemically weak, the significant feature in a given superfix is the position of that stress which will be defined in morphological terms. Thus, /ajmabi/ is one of the constructions wherein the stress strikes the initial vowel of the inflectional margin. As a result, two morphs such as the stems /ajm-/ in /ajmabi/ and /ajm-/ in /ajmi/ 'j'aime', which differ only in stress, are treated as identical rather than as two distinct allomorphs of the base. Although strictly speaking the stress pattern should be noted separately from the segmental span, a strong stress mark is here placed on each construction as a con venient device meant to symbolize the stress pattern of the whole unit. Of two equally serviceable interpretations, that one is 39 preferred which conforms with directly observable linguistic facts. For example, part of the general structure of the verb conjugational system is openly revealed by consistent correspondences in those sets of forms of one and the same verb which vary from speaker to speaker. This co-existence of optional variants in paradigms has been taken into ac count in determining the number of primary verb stems and in defining their distribution. The resulting interpreta tion of verb inflection turns out to be one of the most economical and it has the advantage of being justified by contemporary evidence even when the descriptive and the historical analyses coincide. 1. Phonology 1.1. The Segmental Phonemes 1.1.1. Non-Syllables Non-syllabics total 25 segmental phonemes: /p b t d k g ^ ^ c j f s z s z m n n l l r r j (q) w/. Of these 25 units, the first 22 are consonants; the last three have respectively the same points of articulation as the vowels /i y u/ but the non-syllabic character and the often frica tive mode of articulation of consonants. The chart on page 41 includes all non-syllabic phonemes as well as their main sub-phonemic variants. The latter are placed in square brackets. I. Formational statement of non-svllabic phonemes Phonemes Variants Description and Examples /p t k/1 [p t k] The voiceless bilabial, apico-dental, •^Each phoneme is understood to represent a class of phones; say, /p/ subsumes all voiceless bilabial stop phones. 40 Labial Dental-alveolar Palatal Velar Uvular Stops P t [t] k b d [d] g Affricates V c i V j Fricatives: Flat f [b] [d] j [g] Grooved [s] s A V s [z] z V z Hole U) w Nasals m n V n W Laterals 1 V 1 Vibrants: Flapped r Trilled r [R] H Phonemes Variants [p“ t“ k-] A/ [ k k *1 /b d g/ [b d g] [b d g] 42 Description and Examples and dorso-velar unaspirated stops 2 i i occur freely, e.g. [put] /put/ 'coq'; [teko] /teko/ 'cosse'; [kap] /kap/ 1tete1. All may be unreleased when breath- group final, e.g. [sap-] /sap/ 'il sait'; [defeat-] /debat/ 'dessous'; [enlQk-] /enlok/ 'nulle part'. In addition /k/ subsumes the vari ants: velar [k] before the low vowel [a], e.g. [kau] /kaw/ 'il faut1; palatal [k] before front vowels, e.g. [kin] /kin/ 'quel'; back velar [k] before back vowels, e.g. [kuro] /kuro/ 'quand1. The voiced correlates of /p t k/ have stop and weak-fricative allo- phones. Neither the stops nor the fricatives occur syllable-final. O Following E. Nida's procedure, "e.g." introduces some illustrative examples. A colon introduces a list that I attempted to make comprehensive. Phonemes /t d/ //* i/ 43 variants Description and Ejftampies The stop allophones occur in abso lute initial position and after nasals, e.g. [umbro] /umbro/ 'om bre'; [punde] /punde/ 'pondre'; [eqgwan] /engwan/ 1cette annee1. /*/ is also a stop after /l/, e.g. [taldepan] /taldepan/ 'reste du pain'. The fricative allophones occur elsewhere, e.g.[abe] /abe/ 'avoir'; [bydet] /bydet/ 'boyau'; [luzgats] /luzgats/ 'les chats'. [t d] In addition, [t d] are palatalized before [s z], e.g. [katsa] /kaca/ 'macher ' ? [bjadze] /bjaje/ 'voyage', [t] also occurs in the production [pitu] of the word [pitsu] 'petit'. This being the only occurrence of [t] as a single consonant, and the word being of imitative nature, [t] is treated as an allophone of /c/, i.e. /picu/ 'petit'. [ts dz] The voiceless and voiced apico- Phonemes /*/ /c j/ 44 Variants Description and Examples dental affricates are phonetically clusters of stops plus homorganic sibilants. Both occur intervocal ic , e.g. [kQtsek] /ko/ek/ 'tout de suite'; [tredze] /tre^e/ 'treize'. Only [ts] also occurs utterance- final, e.g. [alabets] /alabe// 'alor s' . [dz] The first element of the cluster may be a voiceless [d] approaching [t], e.g. [dudze] ~ [dudze] /du/e/ 'douze'. The latter production is particularly frequent in the speech of informant GD. _ v v v v _ Lts dzJ The voiceless and voiced fronto- palatal affricates are phonetically clusters of palatalized stops plus homorganic sibilants. Both occur intervocalic and medially after [n], e.g. [kutsa] /kuca/ 'coucher1; [pentse] /pence/ 'peigne'; [pendza] /penjk/ 'suspendre'. [ts] also PhfiAgmes /V /f/ /s z/ /& V 45 Variants Description and Examples occurs word-initial and word-final, e.g. [tsai] /caj/ 'chai'; [papats] /papac/ 'jabot'. [dz] As for /X/, /j/ is actualized in terchangeably as [dz] or [dz], e.g. [mainadze] ~ [mainadze] /rnajnaje/ 'enfant'. [f] The voiceless labio-dental flat fricative has no voiced correlate. There is no marked variant of the norm /f/. [f] occurs except syl lable-final, e.g. [afa] /afa/ 'af faire 1 . [s z] The voiceless and voiced apico- [s z] alveolar groove fricatives have dental variants [s z] after [t d]. The alveolar phones occur else where, e.g. [kasa] /kasa/ 'chas- ser'; [nuzau] /nuzaw/ 'nous'. [s z] The voiceless and voiced fronto- palatal groove fricatives have no variants. Neither [s] nor [z] 46 Phonemes Variants Description and Examples occur medially after [n]. [z] does not occur word-final, e.g. [saliteo] /salibo/ 'salive'; [esado] /esado/ 'houe'; [repes] /repes/ 'repas'r [zakasa] /zakasa/ 'bavarder'; [guteza] /guteza/ 'couler'. /j/ [j i] The voiced fronto-palatal flat fricative [j] occurs generally be fore a vowel, e.g. [jero] /jero/ 'il y etait'; [kyjg] /kyjo/ 'queue'. The semi-vowel [i] occurs after a vowel, e.g. [apei] /apej/ 'apres'. /■q/ [g] The voiced fronto-palatal hole fricative with bilabial co-articu lation has no variants. It occurs only in French loans and only after a consonant, e.g. [abitqelomen] /abitgelomen/ ’habituellement'. /w/ [w u o] The voiced dorso-velar hole frica tive with bilabial co-articulation is generally actualized as [w] before a vowel and as the semi- Phonemes Variants /m/ [m] /n/ [n m q] 47 Description and Examples vowel [u] after a vowel, e.g. [kwe] /kwe/ ’cuir'; [niu] /niw/ ’nid'. However, in the diphthong /aw/ the second element is interchangeably [u] or [9], e.g. [sibay] ~ [sibaq] 1cheval1. The voiced bilabial nasal has no variants and no distributional limitations, e.g. [mure] /mure/ 1museau1. The voiced dental nasal occurs ex cept before labials and velars, e.g. [ana] /ana/ 'aller'. [m] automatically replaces [n] before labials, e.g. [krumpa] /krumpa/ 'acheter*. The variant velar [q] I occurs before velars, e.g. [diqko] /dinko/ 1jusque'. ^Since the opposition /m n/ is neutralized in this position, one could posit an archiphoneme N. In this study /m/ will be written instead. Phonemes Variants /n/ [n] /I 1/ ' [1 1] /r/ [r R] Description and Examples The voiced palatal nasal has no variants and no distributional limitations, e.g. [brena] /brena/ 1vendanger1. The voiced dental and palatal lat erals have no variants and no dis tributional limitations, e.g. [dilys] /dilys/ 1lundi'; [malyk] /malyk/ 'gourdin1. The voiced single-tap vibrant [r] alternates freely with the French voiced uvular trill or fricative [r ]. The latter is prevalent in the speech of informant ED. As a rule it is more frequent in women's speech than in men's. However, all male informants use [r ] sporadical ly. [r] occurs medially intervo calic and after consonants other than [n]. [r] also occurs sylla ble-final, e.g. [laparaso] /laparaso/ 'bardane'; [broko ] Phonemes Variants Description and Examples /broko/ 'brindille'; [bigar] /bigar/ 'moustique'. /r/ [r R] The voiced multiple-trill vibrant [r ] alternates freely with French [r ], e.g. in the recording of the sample conversation (see section 1.4.3., line 13) informant GD pro nounced [enRetar] /enretar/ 1en re tard' . In the random occurrences of [R] replacing indiscriminately [r] and [r] the phonemic distinc tion /r ?/ is lost. [r] occurs word-initial, medially intervocal ic, and after [n],^ e.g. [res] ^In short, /r/ and /r/ are mutually exclusive except between vowels. To put it another way, the opposition /r r/ is relevant only between vowels, so that an archiphoneme could be set up for all other positions. The replacement of either of the Dzcs. vibrant pho nemes by French /r/ in all positions is a consequence of bilingualism. In the case of the male speakers who use [r ] occasionally, there is co-existence of the French system (which has only one vibrant phoneme) with the Dzcs. system (which has two vibrants). In the case of those female speakers who use French [R] almost exclusively, the replace ment of the Dzcs. system of vibrants by the French system is nearly complete. 50 Phonemes Variants Description and Examples /res/ 'rien'; [dariga] /dariga/ 'arracher'; [enrambulat] /enrambulat/ 'embrouille'. II. Phonemic evidence Non-syllabic phonemes are identifiable on the basis of the following evidence presented in this order: commutation with zero, minimal contrastive pairs, contrast in analogous environment. /p/ 'lui’, /pet/ 'pet'; /un/ 'ou', /pun/ 'pont'. /p b/ /pun/ 'pont', /bun/ 'bon'; /pet/ 'peau', /bet/ 'beau'; /palo/ 'pelle', /balo/ 'balle'. /p f/ /pek/ 'sot1, /fek/ 'feu'; /pawre/ 'pauvre', /fawre/ 'forgeron'; /pun/ 'pont', /fun/ 'fontaine'. /p m/ /pun/ 'pont', /mun/ 'mon'; /pew/ 'pou', /mew/ 'mien'; /pes/ 'poids', /mes/ 'mois'. /b/ /^w/ 'au', /baw/ 'il vaut’; /aket/ 'celui-la', /baket/ 'auge'. /b m/ /bet/ 'beau', /met/ 'il met'; /buno/ 'bonne', /muno/ 'moue'; /krabo/ 'chevre', /kramo/ 'il cal cine' . /b w/ A/ A/ /t // A d / A n / A/ A A A n/ A/ A z / 51 /bej/ 'il voit', /wej/ 'huit'; /bebyt/ 'bu', /bewymmos/ 'il boit un peu'. /ej/ ' j'ai', /fej/ 'fait'. /fej/ 'fait', /wej/ 'huit'. /ako/ 'ceci', /tako/ 'tache'. /PYt/ 'il sent mauvais1, /py// 'puits'; /birat/ 1tourne1, /bira// 1vous tournez1. /tuto/ 1toute1, /tudo/ 'buse'; /tus/ 'tes1, /dus/ 'doux'; /tala/ 'tailler1, /dala/ 'faucher1. /tja/ 1tuer1, /nja/ 'il y en a'; /nat/ 1aucun1, /nan/ 'ils en ont'. /en/ 'en1, /den/ 'dent'. /pude/ 'pouvoir', /du/e/ 'douze'; /kudeno/ 'cou- enne', /du/eno/ 'douzaine'. /diw/ 'Dieu', /nxw/ 'nid'r /didaw/ 'de', /naw/ 'haut'. /un/ ' ou', /sun/ 'son'. /kawsos/ 'pantalons', /kawzos/ 'choses' ; /kawsa/ 'chausser', /kawza/ 'causer'; /busin/ 'morceau', /s s/ /s // /z / /z y /z X/ /// // c/ // // /// // V /buzigo/ 'friche'. /pes/ 'poids', /pes/ 'poisson1; /bwes/ 'voix', /engres/ 'engrais'; /sibado/ 'avoine', /sibaw/ 'cheval'. /pas/ 'pas', /pa// 'paix'? /petas/ 'chiffon', /adiSa// 'au revoir'. /fajoj/ 'je ferais', /fazjoj/ 'je faisais'; /eloezu/ 'elle et moi', /elozezu/ 'elles et moi /peza/ 'peser', /deza/ 'deja'r /njezo/ 'niaise' /mjezo/ 'demie'. /seze/ 'pois', /se/e/ 'seize1; /bezi/ 'je vois' /e/is/ 1 eux' . /tryka/ 'frapper', /tryka// 'vous frappez'. /atrapa// 'vous attrapez', /papac/ 'jabot1, /ko/ek/ ’tout de suite', /kru/a/ 'croiser' /debua/ 'devouer', /kru/a/ 'croiser'. /kru/a/ 'croiser', /buja/ 'bouger'. 5Contrastive pairs are very difficult to find, the functional yield of the opposition // // being extremely low. A/ /s z/ /s c/ / V A 3 / A/ A g/ /g/ /m/ A n/ 53 /bua/ 1vouer1, /busa/ 'essuyer1. /desa/ 'laisser', /deza/ 'deja'; /sibaw/ 'cheval', /zilet/ 1veston'; /gratysa/ 'etriller', /estyza/ 'cacher1. /samow/ 'chameau1, /cakot/ 'flaque boueuse'; /proso/ 'proche', /poco/ 'poche'. /obei/ 'obeir', /lezi/ 'lire1. /aze/ 'hier', /ajen/ 'ayons'; /ensaza/ 'essayer', /partaja/ 'partager'. /yw/ 1 oeuf1 , /kyw/ ' cul'y /laba/ ' laver' , /klaba/ 'fermer a clef'. /kryn/ 'nuage', /gryn/ 'grain'; /zuke/ 'perchoir', /zuget/ 'il joua'. /aro/ 'maintenant', /agro/ 'aigre'; /ajre/ 'air1, /gajre/ 'guere'. /ej/ 'j'ai', /mej/ 'plus'; /krabo/ 'chevre', /krambo/ 'chambre'. /krama/ 'calciner', /krana/ 'faire 1'arrogant'; /mej/ 'plus', /nej/ 'nuit'. /n/ /n n/ /V /n nj/ /I/ /I 1/ /V A j/ /i ij/ /j/ /j V 54 /u/ 'ou', /nu/ 'non'? /un/ 'ou', /nun/ 'nom*. /anet/ 'anneau1, /anet/ 1agneau'; /tino/ 'cuve1, /tino/ 1teigne1. /oli/ 'huile', /nolo/ 'eau-de-vie'; /seka/ 'se- cher', /kan/ 'chien'. /brena/ 'vendanger'3 /nja/ 'il y en a'. /yno/ 'une', /lyno/ 'lune'; /o/ 'oui', /lo/ 'la' (postpositive d.o. pronoun). /lej/ 'lait', /lej/ 'lit'; /sulet/ 'tout seul', /sulet/ 'seuil'; /palo/ 'pelle', /palo/ 'paille'. /ej/ 'j'ai', /lej/ 'lit'. /bjelo/ 'vieille', /saminejo/ 'cheminee'; /yla/ 'ouiller', /embuja/ 'envoyer'. /lej/ 'lit', /ljejbis/ 'je l'y ai vu'. /sule/ 'grenier', /sulje/ 'soulier'; /mej/ 'plus', /mjej/ 'a demi'. /arpji/ 'j'enfonce mes griffes', /pqisen/ 'puis- s ant' . 55 /q/ /abita/ 'habiter', /abitqa/ 'habiteur1. /q w/ /pqisen/ 'puissant', /fwina/ 'fureter'; /sitqado/ 'situee', /zwato/ 'joug'. /w/ /ej/ 'j'ai', /wej/ 'huit'. /r/ /o/ 1 oui' , /or/ 'or'; /bun/ 'bon', /brun/ ' noeud dans le bois’. /r 1/ /tar/ 'tard', /tal/ 'reste'; /aro/ 'maintenant', /alo/ 'aile'. /r/ /os/ 'il est', /res/ 'rien'; /un/ 'ou', /run/ 'rond1. /r 1/ /baro/ 'il ferine' , /balo/ 'balle'; /run/ 'rond', /lun/ 'long'. /r r/ /purl/ 'pourrir', /muri/ 'mourir'; /bara/ 'ter mer ' , /fara/ 'il fera'; /kere/ 'chercher', /lire/ 'lis'; /tyro/ 'motte', the suffix /-yro/ as in /sewdyro/ 1soudure'. Ill. Gemination By gemination is meant the utterance of a consonant with retarded release. Although physiologically there is 56 only one closure rather than two successive productions of that consonant, phonemically the consonant will be doubled. The consonants involved are the nasals /m n/, the lateral /l/, and the stop /t/. If gemination as a border phenomenon is disregarded (see 5.1.1.II), consonants are found gemina ted in some 40 words, /ll/ occurs in approximately half of those words. The phonemic evidence is presented for the four geminates in decreasing order of frequency. /I 11/ /palo/ 'pelle', /pallo/ 'pale'; /bula/ 'voler1, /rulla/ 'rouler'; /pulo/ 'poule', /grullo/ 1sava- te1 . /n nn/ /anado/ 'allee', /annado/ 1annee'; /ana/ 'aller', /danna/ 'damner'; /sina/ 'taquiner', /sinna/ 'signer'. /t tt/ /asete/ 'asseoir', /disatte/ 'samedi'; /petas/ 'chiffon', /bjettaze/ 'aubergine'. /m mm/ /sameno/ 'il seme', /semmano/ 'semaine'? /pume/ 'pommier', /emmello/ 'amande'. IV. The phonemic interpretation Only the affricates and the non-syllabics /j q w/ need be discussed. 57 A. The affricates Are the four affricates [ts, dz or dz, ts, dz or gz ] to be treated as single units or as clusters of two pho nemes? The following considerations support the two-unit interpretation: (1) The sequences [dz, dz] are not always voiced through out; that is, native speakers may pronounce [dg] or [$z], [dz] or [^z] and occasionally write "dutze" 'douze'. (2) There is also fluctuation in the total or partial voicing of word-final [ts] in connected speech; for in stance, the [ts] of [bets] 'beaux' becomes [tz] or [dz] in [debetzomes] ~ [debedzomes] 'de beaux hommes'. (3) Native speakers are inconsistent in the syllabic divi sion of these complex sequences in intervocalic position, e.g. [kot-sek] or [kq-tsek] 'tout de suite'; [kat-sa] or [ka-tsa] 'macher1. (4) There are clusters analogous to [ts], such as [ks] and [tr] in a single syllable. (5) [ts] is more often than not interrupted by a morpheme boundary; e.g. [pets] 'peaux' is the plural of [pet] 'peau'. The evidence suggesting a single-unit interpretation seems to carry more weight. (6) In contradiction to point (2) above, there is generally 58 complete assimilation— by gemination— of [-ts ] to the ini tial consonant of the following word, [ts > p] in [deppulos] 'dix poules'; [ts > b] in [fazebbeze] 1faites voir'. (7) In contrast with point 3 above, native speakers con sistently divide such words as [pun-tsyk] 'pointu' with the sequence [ts] in the same syllable. (8) Other clusters— like [ks]— are not homorganic, whereas in [ts] the [s], alveolar elsewhere, is dental after [t] and in [ts] the [t], dental elsewhere, is palatalized before [s]. (9) There are no clusters *[ks ps ] parallel to [ts ]. (10) The suspicious sequence [ts] is not paralleled by a reverse sequence of the same segments in a single syllable; that is, *[st] is not permissible in word-final position; neither is *[st], (11) The phonological system includes the simple sibilants /s z s z/. (12) It must be added to point 5 above that there is no morpheme boundary interrupting [ts], that there are singular words like [pyts] 'puits' and [pretg] 'prix' ending in [-tg], and that [tg] patterns as a single consonant in verb inflections, e.g. [fazen] 'nous faisons', [faz£ts] 'vous faites'. 59 Since there is no clear structural pressure imposing any one interpretation and since the fluctuation in the speech of the natives may well be ascribed to the influence of French, points 1, 2, and 3 can be considered as phonemic latitudes. Point 5, on the other hand, cannot be dismissed. Before a solution for all affricates can be reached, another problem must be taken into account, namely the pho nemic status of the voiced affricates versus the voiceless. The opposition /c 3/ is fairly well established— though operative in few native words— because the occurrences of [j] are increased as the dialect borrows new words by adapt ing Regional French /-aza/ as /-a3e/, e.g. /tirade/ 'ti- rage' . The case of // // is less clear-cut. It must be borne in mind that the opposition // // is operative only in a few words and only in intervocalic position. Furthermore, since /// may be actualized as [dz], the margin of security for // // is considerably narrowed. The expected consequence is the confusion of the two phonemes. For informant MC /kru/a/ and /kru/a/ 'croiser' alternate freely. Although MC uses consistently [ts] in numerals such as [dutse] /du/e/ 'douze', she can hear no difference between [dutse] and [dudge], For MC, then, [ts] and [dz] are simply allophones 60 of one and the same phoneme, say ///. The other informants, however, still "prefer" [tg] to [dz] [dz] in certain words. Does this mean that they are aware of an actual phonemic difference or that one pronun ciation "sounds somewhat better" to them than the other? I believe that the opposition // // is breaking down and that the substitution of one phoneme for the other would not hinder communication. Such scanty evidence hardly warrants setting up a separate phoneme ///, unless for reasons of symmetry and structural pressure. The presence of /// in the system supplies /// with a voiced correlate and /j/ with a dental counterpart: ^ / v 'f c J I am of the opinion that the interpretation of both dental affricates as one phoneme would be equally serviceable and somewhat more accurate. The solution adopted here is to treat all four affri cates // / c ]/ as single units— with the compromise, that for the sake of simplicity in morphemic analysis the sus picious sequences will be written as two units when they belong to two different morphemes, i.e. /py// 'puits', but /bets/ 'beaux' with pluralizer /-s/. 61 B. The segments [j q w] and [i u] Are these separate phonemes or allophones of the vowels A y V? Three different environments are distinguished. [j] and [w] are fricatives, hence discrete consonant phonemes, before a vowel in syllable-initial position, that is, inter vocalic and either phrase-initial or absolute word-initial, [j] is common in intervocalic position, e.g. /pajen/ 'paien', /kyj°/ 'queue'; but [-w-] is rare, e.g. /mawoneste/ 'malhonnete'. [j] is phrase-initial in /jes/ 'il y est', /ja/ 'il y a', and [w] word-initial in /wej/ 'huit'. Between a consonant and a vowel the non-syllabic phones which are acoustically similar to [i u] are often frica tives, but sometimes occur as non-fricative varieties of [i u], phonetically semi-consonants. This varies from speaker to speaker, so that there seems to be a choice of two phonemicizations, say, /bjel/ or /biel/ 'vieux'. Actu ally, the only possible interpretation is to separate pho- nemically the non-syllabics from the vowels /i u/ and retain /bjel/. This is done on the basis of existing sequences of [j] plus homorganic [i]. The data include: /repapji/ 'je divague' and the theoretically possible /irpji/ 1j1enfonce mes griffes' and /rumji/ 1je rumine', the infinitives of 62 these verbs being respectively /repapja/ 'divaguer', /arpja/ 'enfoncer ses griffes’, and /rumja/ 'ruminer1. In addition, the sequence [j i] and the reversed sequence [ij] occur in /emboji/ 'j'envoie' and in /bolijana/ 'je veux y aller', the latter to be distinguished from /boljana/ 'il veut y aller1. There is minimal contrast in one pair, /nia/ 'nier', /nja/ 'il y en a'. The phoneme /j/, then, is actualized consis tently as a fricative in the positions stated in the preced ing paragraph and it subsumes a semi-consonant positional variant which may occur after a consonant. In this way the potential difference between the /j/ of /emboji/ 'j'envoie' and the /j/ of /repapji/ 'je divague' is treated as an allo- phonic difference. The interpretation adopted for /j/ applies also to /w/ and /q/, the distributions being more or less parallel. /w/ contrasts directly with /u/ in such rare cases as /buat/ 'voue', /kwat/ 'couve’, but no sequences of /w/ plus homor ganic /u/ could be found. It follows that the non-syllabic segment [q], which seems to occur only in loan-words after a consonant, is also independent of the vowel /y/, e.g. /pqisen/ 'puissant'. Even in the absence of contrastive pairs, a unit /q/ is easily integrated in the sound-system. The post-vocalic position remains to be examined. In 63 this position the non-syllabic phones [i u] are phonetically semi-vowels, e.g. [kuire] 'coude'; [adiu] 'adieu1. Since [i u] are in complementary distribution with the vowels [i u], a tentative phonemicization /kuire/, /adiu/ is sug gested. The marking of phonemic stress takes care of the distinction between the complex vowel nucleus and hiatus in /pais/ 'peres', /pais/ 'pays'? /faus/ 'faux', /aus/ 'aout'. But this notation is inadequate for transcribing weakly stressed syllables where the syllabic vowels would be un marked; for instance, /ei/ would stand indiscriminately for the diphthong in [zeina] 'gener' and for the hiatus in [obeisen] 'obeissant'. The solution adopted here is to treat the semi-vowels as sub-members of /j w/ because of complementary distribution between pre-vocalic and post vocalic positions and in view of further evidence to be presented in section 1.2.3 in connection with syllabic structure. The data above will therefore be re-written thus; /kujre/ 'coude', /adiw/ 'adieu', /pajs/ 'peres', /pais/ 'pays', /faws/ 'faux', /aus/ 'aout', /zejna/ 'gener', /obeisen/ 'obeissant'. 1.1.2. Syllables There are seven syllabic phonemes /i y e e a o u/. 64 They are placed on the vowel triangle with the main sub- phonemic variants of /o/. The remaining norms exhibit no appreciable variants other than nasalized allophones. Front Front Back unround round round High Higher-mid Mid Lower-mid Low a The definition of each class of vowel phones is accord ing to the chart. When loan-words are taken into considera tion, there are no limitations in distribution. This matter is discussed in section 1.2.1. The distribution of the allophones of /o/ is as fol lows. [o], a higher-mid close back rounded vowel, or [o-1], a mid centralized back unrounded vowel, occur unstressed word-final. This variation takes place in the speech of any one speaker, e.g. [fenno] ~ [fenno-1] 'femme'. A higher version of [9] approaching [u] is also heard in this posi tion . [9], a lower-mid open back rounded vowel, occurs in all closed syllables, e.g. [mgs] 'morceau'. 65 A glide [oo] is occasionally audible in word-initial stress- ed syllables and in monosyllables, e.g. [aketo^me] 1cet homme'; [09] 'oui'. [o], a mid back rounded vowel, or [9] occur elsewhere inter changeably . The distribution, a combination of free variation and complementation, is non-contrastive. The phoneme /o/ is arbitrarily chosen as the norm which subsumes all variants. All vowels have nasalized allophones in syllables checked by [m n n q], phonemically /m n n/, e.g. [len] /len/ 'loin'; [kSn] /kan/ 'chien'. The nasal consonants are al ways sounded. I. Phonemic evidence The vowel phonemes are established on the basis of these contrasts: /i e/ /ordi/ 'orge', /orde/ 'ordre'y /kabilo/ ' che- ville', /kabelo/ 'ecimage de mais1; /bibi/ 'je vis (vivre)', /bebi/ 'je bois1. /i y/ /akiw/ 'la', /kyw/ 'cul'; /ardit/ 'en avant1, /bendyt/ 'vendu1; /akit/ 'acquit', /resyt/ 'regu'. /y V /yn/ 'un', /un/ 'ou'; /kalyt/ 'fallu', /kalut/ 'il fallut'; /ynglo/ 'ongle', /unkle/ 'oncle'; /lynta/ 'huiler', /luntens/ 'longtemps'. /e e/ /sen/ 'cent', /sen/ 'nous soiraaes1; /mew/ 'mien', /mew/ 'miel'; /set/ 1soif', /set/ 'sept'; /parel/ 'paire', /parel/ 'pareil'; /pew/ 'pou', /empew/ 'ajout'; /pebre/ 'poivre', /lebre/ 'lievre'. /e a/ /el/ 'ceil', /al/ 'ail'; /new/ 'neige', /naw/ 'neuf'; /ajmet/ 'il aima', /ajmat/ 'aime'. /a o/ /kal/ 'il faut', /kol/ 'col'; /paw/ 'peu', /pow/ ’peur'; /espja/ 'guetter', /espjo/ 'regarde*. /on/ /pot/ 'il peut', /put/ 'coq'; /on/ 'on', /un/ 'ou'; /mos/ 'morceau', /mus/ 'mout'. II. The phonological system The inventory of the segmental phonemes can now be re capitulated and examined. Counting the borrowed phoneme /njp the total number of segmental units is 32. All are classified on the next two tables, the semi-vowels being best tabulated above the high vowels. 67 Labial Labio-dental Dental Alveolar Palatal Velar Stops P b t d k g Affricates / / v 'C C D Fricatives f s z V V s z Nasals m n n Laterals 1 V 1 Vibrants r r Front unrounded Front rounded Central neutral Back rounded Semi-vowels j u> w High i y u Higher-mid e o Lower-mid e Low a 68 The structure of the sub-systems calls for some com ments. The dialect utilizes the correlation voice/no-voice only in the obstruents— the stops, the affricates, and the fricatives. There is a hole in the pattern: /f/ has no voiced correlate. The sonorants— the nasals, the laterals, and the vibrabts— are all voiced and form discrete series. The vocalic sub-system is highly unsymmetrical. The three front unrounded vowels are paralleled by only two back rounded correlates. In other words, whether [o] or [o] is taken as the norm, there is only one mid back rounded vowel as against two mid front unrounded vowels. Now, it must be emphasized that the opposition /e e/ has too high a functional yield to allow the merger of /e/ and /e/. Symmetry, therefore— if we assume that phonologi cal systems keep tending toward symmetrical patterns— would be expected to develop through the emergence of a back rounded phoneme to fill the vacant slot. This may take place provided the diphthong [ao]— alternating at present with /aw/ (cf. 1.1.1.I)— becomes monophthongized to [o] and that [o] is eventually phonologized. Should the first of these conditions be fulfilled, the second would ensue. In the light of the present distribution of /o/, it is pre dictable that a new opposition */o o/ would have to develop 69 6 i in stressed open syllables, e.g. /boto/ 'fete', /gawto/ ’joue'. Another point of dissymmetry is the presence of a single front rounded vowel /y/. There is no indication that other front rounded phonemes should evolve, since phoneti cally similar phones are lacking. I believe that placing /y/ in a central column, thus aligning it with neutral /a/, is to impute unwarranted symmetry into the system. Such a procedure would conceal two facts: first, that /y/ is unique as a front rounded vowel, and second, that there is a total absence of back unrounded vocalic phonemes. 1.2 . The Syllable The syllable is a sequence of one to five phonemes of l which at least one is a vowel. No consonants are syllabic except for some fricatives and the vibrant /x / in a few interjections: /st/ (to hush someone); /pst/ (to attract attention); /ft/ (to chase cats); /sw/ (to chase birds); /trt/ (to chase pigs). Rules for the syllabification of native words can be formulated in this manner: ^In native words at least, /o/ and /aw/ are mutually exclusive in all other positions. 70 In word-medial position a consonant belongs with the following vowel, e.g. /ti-no/ 1cuve'; /pa-de-no/ 'poele'. The syllabic boundary falls between two successive con- 7 sonants unless they form an indivisible cluster below, e.g. /kar-ga/ 'charger'; /des-tim-bur-lat/ 'ecervele'; /paj-sel/ 'echalas'? /traw-la/ 'pietiner'. Clusters of two indivisible consonants are of two types: clusters consisting of a stop or /f/, plus /r/ or /l/, e.g. /a-by-kle/ 'aveugle'; /su-flet/ 'gifle' and clus ters made up of any consonant plus /j/, /g/, or /w/ except that /w/ plus /j/— like /x/ plus /l/— belong to two differ ent syllables, e.g. /fry-tje/ 'qui aime les fruits'; /glo-rjo/ 'orgueil'; but /bew-joj/ 'je boirais'. Sequences of three consonants are split after the first consonant, e.g. /ba-nas-tro/ 'hotte'; /ar-pjo/ 'doigt de poule'; /en-gwan/ 'cette annee'. When words are subject to compulsory or optional sandhi within breath-groups, the last consonant of a word belongs with the initial vowel of the next word, e.g. the production of /s ez afanat anej/ is /se-za-fa-na-ta-nej/ 'il s'est ^Consequently, two identical consonants belong to two different syllables, e.g. /en-naw-ta/ 'hausser'; /y-tll-le/ 'utile'. 71 depeche aujourd'hui1. The various types of syllables will be listed and ex emplified in decreasing order of frequency. Canonical forms are given with V standing for any vowel and C standing for any non-syllabic:^ CV /be-no/ 'viens' CVC /bol/ 1il veut' V /e/ 'et1 VC /in/ 'ou' ccvc /tres/ 'trois' ccv /pru/ 'assez' cvcc /tens/ 1 temps’ ccvcc /bjels/ 1vieux' vcc /yws/ 1 oeufs ' cccv /ka-drjo/ 'il faudrait ®This basis for the count is justified by the treatment of /j q w/. It may appear that an entirely different picture of the syllable would be obtained by re-counting the pho nemes on the more common basis: V = any vowel or complex vowel nucleus and C = any consonant. Such a count would diverge from mine in that the non-syllabics /j q w/ would not be added in with the consonants, hence the canonical forms would differ somewhat both in shape and in hierarchy. Actually, the results would still show the predominance of the pattern CV, with CVC ranking second. This is so because in any portion of the corpus /j q w/ occur seldom in com parison with the other consonants. 72 A count of canonical forms in isolated words shows that CV is by far the most common type and that CVC occurs more frequently than all the remaining types together. The pro portion of closed syllables versus open syllables is approx imately 35 per cent versus 65 per cent. The latter figure is considerably increased in connected speech, since the final consonant of a word is transferred to the initial vowel of the next word; for instance, /pas/ 'pas' is of type CVC, but it becomes CV in a phrase like /pa-zet/ 'pas lui1. The dialect exhibits, therefore, a marked preference for open syllables. I. The distribution of vowels Examination of the phonemes constituting the syllable leads to the following statements with regard to distribu- 9 tion. All vowels occur under stress in word-initial, medial, and final positions. The vowels /e a y u/ do not occur un stressed word-finally. All vowels except /e/ occur un stressed word-initially, although in this position /o-/ is 9There is no advantage in presenting the data in chart form when the examples are plenteous and the exceptions few. However, the chart form is quite appropriate for demonstrat ing the complex vowel nuclei. 73 found only in learned words like /opylenso/ 'opulence'. All vowels occur unstressed word-medially, but in this position the occurrence of /o/ and /e/ calls for a discussion. In indigenous words /o/ is uncommon in any unstressed syllable preceding the stressed syllable. Only in the se quence /omen/ of nouns and adverbs, for instance in the ad verb /paromen/ 'probablement', is /o/ found unstressed in the pretonic syllable. Moreover, according to a widespread native gradation pattern, /o/ is raised to /u/ when the stress shifts away from /o/, e.g. /mole/ 'moudre', /mulyt/ 'moulu'. There is, as a consequence, a strong tendency to replace any unstressed /o/ before the stressed syllable. Accordingly, some words like /mutyr/ 'moteur1 have been borrowed with /u/ for French /o/, while others currently exhibit fluctuation from speaker to speaker? some cases in point are /mwasgnyzo/ 'moissonneuse1, /pgpylasjun/ 'popula tion', /ipgteko/ 'hypotheque', /pr°pjetari/ 1proprietaire'. On the other hand, examples such as /sofyr/ 'chauffeur', /oto/ 'auto' wherein /o/ is consistently preserved indicate that the tendency to replace /o/ by /u/ is breaking down. Pretonic /e/ is also uncommon in native words. It does occur in pretonic position in closed syllables, e.g. /bjettaze/ 'aubergine', /perdyt/ 'perdu'; but /lejtygo/ 74 'laitue' and /wejteno/ 'huitaine' are the only two indige nous words in which /e/ was found unstressed in an open syllable (for these syllables re-defined see 1.2.III). Pre tonic /&/ is avoided, as in the case of /o/, by raising it one step, i.e. by substituting /e/ for /e/. This replace ment, which takes place in open syllables— with the two exceptions just noted— is very frequent in derivation, e.g. /sirezo/ 'cerise'; /sireze/ 'cerisier1. The loan-word /bwet^ro/ 'voiture' presents an interest ing case of incomplete assimilation. There are three alter nate pronunciations: /bwetyro/, /bwatyro/, and /bwetyro/. The first of these is the fully assimilated form, with /e/ in pretonic position as in native words. /bwatyro/, pre serving intact the French sequence /wa/, is the least as similated of the three forms; but /wa/ is too unfamiliar a sequence in the dialect to become fixed as such; hence the other two forms, /bwetyro/ lends itself to two interpreta tions. It may rank with /lejtygo/ 'laitue' as another ex ception in the making. Preferably, /e/ may be the compro mise vowel, its point of articulation being mid-way between the /e/ and the /a/ of the concurrent forms. 75 II. Diphthongs and triphthongs The center of the syllable may be occupied by a combi nation of a vowel with one or two of the non-syllabics /j w/. A composite of vowel plus semi-vowel will be called a falling diphthong, a composite of semi-consonant plus vowel, a rising diphthong. A composite of semi-consonant plus vowel plus semi-vowel is a triphthong. The various possible combinations are charted and ex emplified on pp. 76-77. A glance at the charts shows that the dialect makes ample use of diphthongs. There are 23 in all, that is, over half the possible number. In the major ity of cases many examples can be found in addition to the one given, but some diphthongs are uncommon, and, as indi cated by the blank spaces, many combinations are not util ized. Triphthongs are comparatively rare. Some comments are in order. The rarity of combinations of /y/ with /w/ is expected because of the narrow margin of security between those phonetically similar sounds. The falling diphthong /yw/ does occur in three items: /byw/ 'boeuf', /kyw/ ' cul', and /yw/ 'ceuf' . The dialect even ex ploits such possibilities as /j/ plus homorganic vowel /i/, e.g. /emboji/ 1j'envoie1, but no combinations of /w/ and /u/ have been found. jv Vj Vj wV wV Vw Vw i repapji ie divaaue fwina fureter liwse eclair miwrano qr^nad? e sjegut 11. suiY.il be jre verre krejrej ie croirai kwet meche kwensa coincer pew POU empewta alloncrer e . 1 V m^ezo demie djerbyt ouvert 1 , me j plus lejtygo laitue lwe loi pwensun poincon new neuf a t ja lUfiF pjala crier ajgo eau maj la me ler zwato jouq swasanto soixante kawzo chose gawdyflo toupie o t jo tienne bom jo enioleuse bo j bois mawoneste malhonnete owlo brebis -J 0> jv LJ. < Vj Vj w^ wV Vw Vw u glurjus oraueilleux bjulun violon 1 ku jre C<?ilde y byw boeuf jVj jVj jVw jVW wVj wVj 1 wVw wVw e mje j milieu jejdit ie lui ai dit kwe j cuit a jawbali ■ie le lui donne o angijoj i'irais 78 There is a marked paucity of examples of /wa/ and /wi/. The rare native words with /wa/ are /kwat/ 1couve' and /zwato/ 'joug'; the other items are learned loans like /swasanto/ 'soixante'y /swana/ 'soigner'; /pwatrino/ 'poi- trine'y /swajus/ 'soyeux' (versus the native noun /sedo/ 'soie'). That the remaining blank spaces on the chart are not filled appears to be structurally accidental. The search for examples has proved fruitless so far. The rising diphthongs with the high front rounded semi consonant /q/ are not included in the charts because of a consistent correspondence in the dialect to those diphthongs in French. Phonologically, the examples appear to be assi milated loans. That these words are not native is also demonstrated by their derivation. For instance, the ab stract noun /kuntinqasjun/ 'continuation' is derived not from the native verb /kuntyna/ 1 continuer1, but from the French verb. The various rising diphthongs with /q/ are: /qi/, /qi/ /prudqi/ 'produire'y /kundqi/ 1conduire'y /epqizomen/ 1epi xsement1y /pqisenso/ 'puis sance ' . /qa/, /qa/ /abitqa/ 'habituer'y /sitqa/ 'situer'y 79 /sitqasjun/ 'situation'7 /kuntinqasjun/ 'continuation'. /qe/ /abitqelomen/ 'habituellement'? /kuntinqelomen/ 'continuellement'. A falling diphthong /yj/ seems to be in the making. This is observable only in the masculine plural alternate forms of the definite article /lus/ 'les' before a voiced consonant (see 1.5.Ill.B.[l]). The four triphthongs occur seldom except for /joj/, which is the first person singular inflection of the condi tional and often of the imperfect. There is another occur rence of /wej/, the word /wej/ 'huit', which is virtually the only word with initial /w/. Ill. Semi-vowels and non-svllabics In the preceding section /j q w/ were handled according to the conventional method, which consists in treating such non-syllabic elements, that is, on-glides and off-glides, as components of diphthongs and triphthongs. This section is intended to suggest a different interpretation, which would dispense with all complex vowel nuclei and would yield a more satisfactory description of the structure of the syl lable . 80 In section l.l.l.IV.B /j q w/ were described as con sonants in pre-vocalic position. Rising diphthongs, i.e. sequences of type Cj, Cq, and Cw, as in /pjel/ 1 peau1 ; /kwe/ 'cuir'; /mjo/ 1mienne', are better interpreted as consonant clusters; some of these are of the "muta cum liquida1 1 type in which /j q w/ rank with the liquids /r l/, e.g. /tjo/ 'tienne'; /trop/ 'trop1; others can be set up as an addi tional type of cluster, e.g. /sjek/ 'il suit'. This will be done in section I.2.V. In falling diphthongs, the second elements are phoneti cally semi-vowels by position, that is, non-fricative phones which from the physiological standpoint are related to vow els. These phones, however, have this in common with con sonants, that they are non-syllabic. Phonologically, the interpretation of the semi-vowels as consonantal units is attractive, since, occasionally at least, they can be shown to function as consonants in the framework of the syllable. In connection with the distribution of vowels (see 1.2.1 above), it was observed that /e/ replaces /e/ in the pretonic open syllables of indigenous words, e.g. /new/ 'neige', but /neba/ 'neiger'. In both of the exceptions noted, namely /lejtygo/ 'laitue' and /wejteno/ 'huitaine1, it is remarkable that /e/ is followed by /j/. I consider 81 that the initial syllables of these two words are not open but closed, because /e/ is checked by the consonant /j/. These words, then, are no longer exceptions; they follow the native pattern which tolerates pretonic /e/ only in closed syllables. [-i] or [-j] also appear to check vowels when both alternate with implosive [-2] at word boundaries; for ex ample, the pluralizer /-z/ exhibits free variation in /lazgatos/ ~ /lajgatos/ 'les chattes1. This sound-change, observable in the speech of all informants, will be de scribed in section 1.5.III.A as a vocalization or a pala talization of [-z]. Whatever the phonetic shape of the alternants of [-z], the alternation can be interpreted as an exchange of consonantal sounds in closed syllables. The treatment of the semi-vowel [u] as /w/ is recom mended because it is also found to pattern as a consonant in a few instances. The alternation of [u] with a conso nant, either [b] or [l], is the prevalent derivational pattern, e.g. [flu] 'fil', [fila] 'filer'. Consequently, the consonant [-w-]— which does not occur intervocalic in base-words— is infrequent but not unknown in word-medial intervocalic position. The items under discussion are derivatives formed by the addition of the prefix [mau-] to to some base, e.g. [mawone^te] 1malhonnete'; [mawadret] 'maladroit'; [mawapres] 'malappris'. The semi-vowel [u] of the prefix, which is homophonous with the adverb [mau] 'mal becomes a consonant [w] in those derivatives. It is note worthy that [w] shifts to the next syllable, yielding the division /ma-wo-nes-te/. [w], then, disrupts the morpheme boundary and joins the following vowel. This behavior is typical of consonants according to the rules of syllabifi cation stated in section 1.2. Furthermore, the division of words into syllables re veals a similar patterning for the liquids and the non- syllabics. Both constitute the second elements of indivis ible clusters, e.g. /y-bro/ 'oeuvre'; /no-bjo/ 'mariee'. Also, the sequence /wj/, like /rl/, is broken by the syl labic boundary, e.g. /aw-jo/ 'il aurait'; /byr-la/ 'bruler' There is nothing, then, to preclude the interpretation of [y] as /w/ in, say, /piw-ze/ 'puce'. Several conclusions can be drawn at this pointy Only the consonantal interpretation is appropriate for /j q w/ in pre-vocalic position. The same interpretation for the semi-vowels is indicated because it takes into ac count those instances in which these phones pattern as con sonants . 83 The syllables whose peaks are falling diphthongs can be considered to have the phonemic structure of closed syl lables . The articulatory difference between semi-vowel and con sonant is non-significant, since a semi-vowel like [_u] can be shifted to the consonantal, intervocalic position [-w-], e.g. /maw/ 'mal'; /mawadret/ 'maladroit' (syllabified /ma-wa-dret/). Hence [i] and [j] are also represented by only one symbol /j/, whatever the environment. Inasmuch as /j t [ w/ are always non-syllabic, but other wise akin to vowels as well as to consonants, the terms "non-syllabics" and "syllables" are preferred to "conso nants" and "vowels" for the segmental sub-systems of this dialect. IV. Vowels in hiatus Hiatus is rare internally in native words. The vowels /ai/ form hiatus in the words: /ai/ 'hair', /pai/ 'par- bleu', and /pais/ 'pays'. In unstressed position the mono syllabic sequence /aj/ replaces hiatus in the derivatives of /pais/ 'pays': /pajzan/ 'paysan'; /despajza/ 'd^payser'. Interjections provide other cases of hiatus: /ai/ (to order a horse forward); /beisi/ (to order a horse to turn left); 84 /tea/ (to call pigs), /ay/ is found in the imitative verb /sengayta/ 1s1engouer1. Hiatus is rather common at word boundaries. As a gen eral rule, a word-final stressed vowel forms hiatus with any vowel following immediately, e.g. /2 setyiies-*/ 'si tu y es1. Hiatus is optional, but relatively frequent, when the word- final vowel is unstressed, even if the two contiguous vowels are identical, e.g. /2awportoanej1/ 'il le porte aujourd'- hui 1 ; /2baakaloal/ 'il va achever'. There is hiatus when function-words like /a/ 'a' are made up of a single vowel, e.g. /2awp6rtoalafjeroi/ 'il le porte a la foire'. Hiatus is sporadically avoided by aphaeresis. This appears as an irregular phenomenon opposed to the regular and widespread elision, or loss of a word-final vowel.^ The examples seem to be hackneyed phrases or highly recur- 11 rent expressions; /2ez gurifiando-*^umo (y)no ^ejnol/ 'elle est gourmande comme une fouine1, /2j a (y)m paw ke 1 atendjonl/ 'il y a belle •*"®No conditioning has been found for such unexpected cases of aphaeresis? the reason may be that the data are meager; I could elicit only random occurrences. In the examples, some of which“are found in the il lustrative text (see 1.4.3), the silent vowel is placed in parentheses. 85 lurette qu'on 1'attendait', /2gejto (a)li6i/ 'regarde ga', /2nus kadra (e)nsazal/ 'il nous faudra essayer'. In contrast to the scarcity of internal hiatuses in native words is the readiness with which the dialect intro duces loan-words with hiatus, e.g. /obei/ 'obeir' (doublet of /eskuta/), /kreatyr/ 'createur1, /poezio/ 'poesie', /poemo/ 'poeme', /eblui/ 'eblouir', /buat/ 'voue',1^ /lyeto/ 'luette', /ryelo/ 'ruelie'. The sequences */yo/ and */ko/ that would result from the integration of feminine loan words with native /-o/ would be exceptions. Such sequences are not permitted; hiatus is here avoided by epenthesis of the transitional glide /j/, e.g. /lungobyjo/ 'longue-vue', /mulyjo/ 'morue', /pansejo/ 'pensee (fleur)' (cf. native /pensado/ 'pensee, meditation'). V. The distribution of consonants All consonants (abbreviated C) appear singly in syl lable-initial position. Examples are limited to cases of rare occurrence, particularly word-initial, e.g. /cyka/ 'sucer', /lej/ 'lit1, /nolo/ 'eau-de-vie', /zero/ 'zero', /wej/ 'huit'. ■*-2in Regional French, the pronunciation with hiatus is favored; i.e. /vue/ 1voue1 is dissyllabic. 86 All C appear singly in syllable-final position with the exception of the voiced stops and // 3 f z z r/. The following C can be syllable-final only in word-final posi tion: /n 1 p s / c/. /“Hi/ occurs word-final in only one word, /agram/ 'chiendent'. The types of syllables with clusters of two C remain to be examined. In syllable-initial clusters of two C, the distribution is best stated with reference to the second C: (1) if the second C is /r/, the first C may be /f/ or any stop; (2) if the second C is /l/, the first C may be /f/ or any stop other than the dental stops /t d/; (3) if the second C is /j/ or /w/, the first C may be any stop, any nasal except /n/, one of the fricatives /f s z/, or the dental lateral /l/, e.g. CCV /fje-ro/ 1foire'; CCVC /lwes/ 'loi'; CCVC /bun-swer/ 'bonsoir'; CCVC /mjot/ 'muid'; CCVC /zwen/ 'jeune'. In syllable-final clusters the second C is generally /s/, but it is /k/ in rare cases like /bosk/ 'bois'. The first C may be /t k n 1 r/ or the non-syllabics (semi vowels) /j w/, e.g. of type CVCC /bets/ 'beaux'; /keks/ 'begues'; /mens/ 'moins'; /suls/ 'seuls'; /zurs/ 'jours'; 87 /rejs/ 'rois'; /faws/ 'faux'. Within the syllable, clusters of three C— two C plus a semi-C— are rare and were found only syllable-initial in verb forms, e.g. /ka-drjo/ 'il faudrait'. This word was often heard with reduction of th6 uncommon cluster /-drj-/ by loss of /-r-/, i.e. /kadjo/. 1.3. The Supraseamental Phonemes 1.3.1. Stress There are two phonemic degrees of stress at the word level: strong or primary stress /'/ and weak stress /°/. Unless pertinent to the discussion, the latter will be left unmarked. In citation forms the syllable bearing strong stress is slightly longer than the others. Those two stresses are in contrast in the pairs, /lapis/ 'celeris', /lapis/ 1lapins1 /bako/ 'vache', /ako/ 'ceci' /pero/ 'poire', /zero/ 'zero* /duso/ 'douce', /dusu/ 'douceur' /myson/ 'ils montrent', /mysan/ 'nous montrons1 /ajmi/ 1j'aime', /ajma/ 'aimer' /debenge/ 'devenir', /debengyt/ 'devenu'. 88 Simple words have one and only one primary stress. This stress is not completely free; it always falls either on the ultimate or on the penultimate syllable of a word. Therefore, the antepenult syllable of a three-syllable simple word automatically receives weak stress. I. Stress patterns of simple words The various stress patterns are listed and exemplified below. -- /shr/ 'couleuvre'; /bren/ 'son de ble' -- ” /pjoto/ 1dinde1; /negre/ 'noir' -~- -- /tesun/ 'pore'; /azk/ 1hier1 " /agaso/ 'pie'; /bulygo/ 'etincelle' -“ - “ /perdigat/ 'perdreau'; /eskanti/ 'eteindre' . Simple words of four or more syllables may have a secondary stress ['] two syllables before the primary stress. The occurrence of ['] being predictable even if some speaker makes use of it as a rule, it is sub-phonemic and will not be marked in phonemic writing. The longer patterns are: -w - -“ - -'- -“ - [maleruzo-*] /maleruzo/ 'malheureuse' -“ - ~ ” -'- [entamena] /entamena/ 1 entamer' 89 ” “ “ “ [enr ambul&do- 1 ] /enrambulado/ 'embrouillee' ” — — “ -- [eskarabilat] /eskarabilat/ 'degourdi' “ -- — [eskarabilado-* ] /eskarabilado/ 'degourdie' -w - “ " — — [endependentomen] /endependentomen/ 'independemment1. The distribution of stresses on simple words can be summarized thus: one primary stress followed or not by one weak stress and preceded or not by one to at most five pho- nemically weak stresses. Applied to syllables or to words in constructions, the term "stressed" will be used to mean "receiving a primary stress" as opposed to "unstressed," meaning "receiving pho- nemically weak stress." II. Stress patterns of compounds Compounds have at least one and at most two primary stresses. The units are made up of two full words (here termed main components) with or without one function word. The various arrangements of phonemic stresses which 13 occur in the data available are listed below with exam- 13 JThe data include native formations as well as units which show Fr. over-all patterns and which may be loan- translations. All the data available, barely 100 items, fit the description here presented, but due to the paucity of 90 pies. The presentation of a single example rather than two is meant to indicate that the pattern has limited occurrence. The breakdown of units into smaller words is given in paren theses . -~- /kawkyn/ (kawke yn) 1 quelqu'un1 /grampaj/ (gran paj) 'grand-pere' -'- “ -'- /tastosaw/ (tasto saw) ' trempette1 /pkndawzet/ (pan de awzet) 'cenelle1 -- -~- /penydo/ (pe nydo) 'pauvresse' -“ - -'- /tastasaw/ (tasta saw) 'gouter1 -- /abanze/ (aban aze) 1 avant-hier1 -x- -'— -w - /gastosawso/ (gasto sawso) 'trouble-fete 1 /damoblankc/ (damo blanko) 'hibou' ” /azedela/ (aze dela) 'avant-hier' /eskawfolej/ (eskawfo lej) 'chauffe-lit' -'- — ” -'- /portofenan/ (porto fenan) 'siege du charretier1 -- -'- -- -- /busunnegre/ (busun negre) 'prunellier' -- -- -- -'- -- /fringolastutos/ (fringo las tutos) ' cher- cheur de femmes' native compounds of certain types this description should be considered tentative. 91 " “ -- /merdodekukyt/ (merdo de kukyt) 1 gomrae du cerisier' In all compounds the second main component is stressed. The first main component may be stressed or unstressed. The stress patterns of compounds can be described as follows: (1) dissyllabic units bear a single primary stress on the final main component; (2) three-syllable units bear a single primary stress on the final main component provided this component is dissyllabic and stressed on the penult, e.g. /penydo/ 'pauvresse'. In other three-syllable units the initial main component may or may not be stressed; (3) units of four or more syllables bear a primary stress on both main components. More often than not the stress patterns of compounds do not differ from those of similarly constituted phrases, so that the foregoing stress patterns are of limited useful ness as criteria for the identification of compounds. Where the stress pattern of a compound is distinctive, there is an alternation of strong stresses with weak stresses, an alter nation which, as can be seen in the listing above, is pres 92 14 ent in the most widespread patterns. Thus, a simple word which normally loses its primary stress in phrases may pre serve it when used as the first main component of a com pound and vice-versa. There follow some examples of com pounds with stress patterns showing the typical alternation just mentioned. The first main component is unstressed, e.g. in dissyllabic compounds such as /sembun/ (sen bun) 'parfum' (vs. the clause /sem bun/ 'il sent bon'), in three-syllable compounds with final component stressed on the penult such as /binagre/ (bin Agre) 'vinaigre' (vs. the phrase /de bin agre/ 'du vin aigre'). The first main component is stressed, e.g. in other three-syllable compounds such as /awteskots/ (awtes kots) 'autrefois' (vs. the phrase /yn awte kop/ 'une autre fois'), . | ■ v . | , v \ in larger compounds such as /belofilo/ (belo filo) 'belle- i v 15 fille' (vs. the phrase /yno bslo filo/ 'une belle fille'). ^In support of the significance of this alternation of stresses are examples such as /tortikoli/ (torti koli) 'tor- ticolis1, a learned loan-word integrated without phonemic substitution but with a stress pattern which clashes with that of the end-stressed Fr. original. have occasionally heard the first main component 93 1.3.2. Internal Juncture and Pitch There is no internal open juncture; that is, within the word the transition from one phoneme to another is normal. This is true of simple words and compounds. Neither is there open juncture at word border-points within close-knit phrases. Furthermore, the constituents of such phrases are joined by sandhi phenomena (see 5.1 below). So, for exam ple, the absence of open transition and the subsequent re placement of /i/ by /j/ cause the phonemic stretch /boljadyza./ to be ambiguous. Depending upon the context, this utterance can be interpreted as /boli adyza./ 'je veux aider' or as /bol i adyza./ 'il veut l1aider'. Pitch variations are irrelevant at the level of the syllable or the word. Relative pitch levels are significant as constituent elements of the intonation contours which pertain to breath-groups. These formalized patterns will be with semi-strong rather than strong stress, e.g. [mau] in /mawentendyt/ 'malentendu', [apei] in /apejmjedzur/ 'apres- midi'. The evidence is far too scanty to grant phonemic status to secondary stress or to establish a possible rela tionship between the stress and the grammatical nature of the components involved; hence the notation /mawentendyt/ on the analogy of native compounds of the types /gastos&wso/ (vb. + N) 'trouble-fete', /kamboligo/ (N + N) 'jarretiere' wherein the two primary stresses are well established and are not dependent upon the nature of the components. 94 described in section 1.4.4. 1.4. The Phonological Phrase A phonological phrase or breath-group is a span of 16 one or several words characterized by the following proso dic features: a single phrase-stress, a terminal (external) juncture, and a given arrangement of pitch phonemes. An utterance may consist of one or more phonological phrases. These in turn may comprise one or more syntactic constructions of which the minimal types, or small close- knit phrases, will be exemplified in section 3.3. The boundary between two successive close-knit phrases is marked by the presence of potential pause and the optional charac ter of general sandhi modifications (see 5.1.1). 1.4.1. Phrase-Stress Phrase-stress is a group stress usually superimposed upon word primary stress and comparatively more prominent. There is only one of these per phrase. It usually strikes the last word of a phrase, including an enclitic function l^A meaningful utterance made up of a single word is to be included here. Such a condensed utterance takes on structural meaning precisely because it is accompanied by the same prosodic pattern as that of an equivalent longer phrase. For examples of these, see 3.3. 95 word which elsewhere bears weak stress. In rapid speech the prominence of phrase-stress is increased and the inten sity of preceding loud stresses is considerably reduced. In conjunction with an external juncture, phrase-stress serves as an initial criterion in the division of utter ances into phonological phrases. Once the external junc tures have been recorded and the intonation patterns estab lished, phonological phrases are identified. Henceforth, phrase-stress need not be marked by a special symbol; it is understood to be represented by the stress mark preceding an external juncture. 1.4.2. External Junctures An external juncture is a combination of features which terminate those portions of utterances called phonological phrases. In this dialect, the following features suggest the presence of an external juncture: (1) a potential or actual pause of non-significant duration — its actualization depending upon the speed of delivery and/or the emotional disposition of the speaker; (2) the non-occurrence of phrase-internal sandhi phenom 96 17 ena, (3) the occurrence of group-stress marking the end of a phrase; (4) the coincidence of this stress with prolongation of the stressed syllable when no unstressed syllable follows; (5) a comparatively more intense production of one syllable or word indicating the starting point of a new intonation pattern. One or more additional features lead to the distinction of three types of junctures; rising/t/, sustained /-*/, and falling /*/. These features are: /t/ an abrupt rise in pitch slightly above the last pitch level of the phrase, dwindling volume, and a brief interrup tion of voice; /-»/ some reduction of volume and a split-second pause, but no change in pitch and no end-rise after the last pitch level; /\/ a quick drop into silence. •^•^These two criteria are of limited value by them selves, since they may also indicate the break between two successive close-knit phrases. 18 1.4.3. Illustrative Text The following is an unrehearsed conversation between three villagers: informants DP, GD, and ED. Typically, there are shifts in sentence construction, repetitions, hesitations, chuckles, and sudden changes of topic— from corn-planting, to weather conditions, to weather predictions in connection with the yearly celebration at a nearby vil lage in honor of its patron Saint, hence to the celebration of a similar event at Donzac. The interlinear French translation is literal. It will be noted that this translation does not differ appreciably from colloquial French wording and usage, except for an occasional imperfect subjunctive and a few prepositions which would not be used in everyday French conversation. 1 DP 2e a prelaw -* fu mil -* ]furo lu samenes fy1 I Et a propos, le mais, quand le semes-(tu), toi? 2 GD 2e few -* l&w -* Eh, bientot; bientot 3 3pasque aben ajyt yn drolle &e tens -*3e t parce que (nous) avons eu un drole de temps, hein? 1®I am grateful to Dr. R. Zrimc for helping with the audition of the recording of this text. 4 5 6 7 8 9 10 11 12 13 14 98 2 i i ■ e o 4 Eh oui; 2 1 2. 1 , 2 3, e ke ame yn iber kumo lu ke abem pasat t et qu'avec un hiver comme celui que (nous) avons passe, 2 am aket fred de fupt It a fAj t avec ce froid de loup, qu'(il) a fait ... 2 , , 2 3| 2, saj pas se 1 as sentit -» ty -* (je ne) sais pas si (tu) l'as senti, toi, 1 V I 2 I V 1 ! zu l'ej plci sentit tuzur 4 moi, (je) l'ai rudement senti toujours. 2 2, 2 , V 2 apej -* krezjon de trabala -* Apres, (ils) croyaient de travailler, S | 31 v 3 ez estal la piezo -♦ (ga) a ete la pluiei 21, 2 , l v , 2 3,2 e aro ke pujon trabala ame lu printens - ♦ Et maintenant qu'(ils) pourraient travailler avec le printemps, 2 , vi 1 , v 1 2 11 ez estat tuzur la piezo -* e lu fret -* (ga) a ete toujours la pluie, et le froid! 2 , 3 _ 1 , 1 e fa k es tut en retar t Et (ga) fait que (c‘) est tout en retard! 3 3A 1 i 1 1 1 s i ba kale mete -* aro 4 (II) va falloir s'y mettre maintenant. 99 15 DP 2nuzaw ban kumenia - * 3 duinin 1 Nous, (nous) alions commencer demain. 16 GD 2 duman t Demain? 17 DP 2ban kumenia -* (Nous) alions commencer; 2 . 1 , ,11 - 2 3 I 3 18 ez en tren de munta lu semwer -* (il, i.e. mon fils) est en train de monter le semoir; 19 2ban kumensa de lu samena cfuinan I (nous) allons commencer de le semer demain. 20 2e ap&j -* kuntyneren -*2dyz u trej zurs -♦ Et apres, (nous) continuerons duex ou trois jours, 21 2e apej sederen lu semwer a yn 1awte 1 et apres, nous cederons le semoir a un autre. 22 GD 2ej la sazun -»2aro2 -* (C')est la saison, maintenant; o * 3 i 1 X 1 23 e fa bet tens -* aro et (il) fait beau temps maintenant; 24 2 j a m paw ke 1 atendjon t (il) y a un bout de temps qu'(ils) 1' attendaient' . 25 DP 2m& bel&w tapla puj6 fa awraj^ -* Mais peut-etre aussi bien (il) pourrait faire orage: 26 27 28 29 30 31 32 33 34 35 36 100 2 2 1. 2 , ' ane] -» sabes ke -♦ aujourd'hui, (tu) sais que ... DP, 2ez la boto a kastelsarazin -* GD (c1)est la fete a Castelsarrasin DP 2lu sent alpinjen I Saint Alpinien. 2 I | l 2 . v 2 e d abitydo fa awraje -* Et d'habitude (il) fait orage ... 2 2, .2 2 U. , . GD awraje suBen i orage, souvent. 2 - i i n i 2 2 % 1 DP e be digo -* se plew -* Eh bien, dis done. s'(il) pleut, 2faren kum a awbifa -* (nous) ferons comme a Auvillar, 2 la kiteren tumlaa -* (all laugh) (nous) la laisserons tomber1 Ah I eh bien, cette fete? 1 2 i 2 ED e se preparo -* Eh, (elle) se prepare ... T - N T - . 2 3 I 3 +2 , 2 DP se preparo t krezes ke -* (elle) se prepare? (Tu) crois que ... 37 38 39 40 41 42 43 44 45 46 101 2j awra de munde t (il) y aura du monde? 2e bel!ew i Eh peut-etre. 2se ty i ses -*2parleren dez awtes kots -» Si toi, (tu) y es, (nous) parlerons des autres fois: 2sabes ke2 - * (tu) sais bien que ... 2abem birat kawko ifalso de tenz en tens -» (nous) avons tourne quelque valse de temps en temps . . . 2t en rapelijos 2de la bira 2enlcero t (Tu) te le rappellerais, de la tourner, encore? 2 — 3 2 ^ i e m en rapelijoj - » sylomem belew -* Eh, (je) me le rappellerais, seulement peut-etre, (tu) sais bien, la tete me tournerait peut-etre, hein 1 1 ^ ■ 1 j3|2 2 2 kwake - ♦ j a de mumens -* ataw Quoique ... (il) y a des moments ainsi, P 3 2 2 I belew ke nu -» peut-etre que non; 102 47 ED 8nus kadra nsaza 4 (il) nous faudra essayer. 48 DP Snus kadra nsaza t (II) nous faudra essayer1 . . p -_1. 2n 2 I 2 49 me kadjo pas kam memo Mais (il) ne faudrait pas quand meme 50 2ke se futusen 2de 3nuzaw -* 1trop 4 qu'(on) se moquat de nous, tropl 2 2 t 51 kaw pasa - ♦ (II) faut passer ... 52 2per desys tut a nost aje aro 4 par-dessus tout, a notre age, maintenant. 53 GD 2o co 4 Oui, bien sur. __ 2, • » 2 l 2 2 . .31 2 54 DP ke bos fa - ♦ dxgo -* sabes ke -* Que veux-(tu) faire, dis done, (tu) sais bien que . . 55 2sun de kawzoz aro k on ne parlo I>e - * * (ce) sont des choses maintenant dont on parle, certes -r I3' 3 2 13 _ 3, . 3 31, 3 |2 56 la lengo -* pax -* raj -♦ om parlo prampu - * La langue, parbleu, ga va; on parle bien assez, 3 , 57 ED 2me om bem bjel - * mais on devient vieux, 103 sme de 3forsos - 2zero - * ■ 2n j a plys -» (all laugh) mais des forces, zero, (il n')y en a plus 1 2 I 3,1 buz amyzerep pla -♦ (Vous) vous amuserez bien, 3e zu buz atendrej1! et moi, (je) vous attendrai. 3e lDe ->1alor 2ez 3entendyt 1ataw 1 Eh bien, alors, (c1)est entendu ainsi: 3nu n angeren a la boto tud ctys 4 (nous) nous en irons a la fete tous deux. following is a free English translation of the illustrative text: DP By the way, when are vou planting the corn? GD Well, soon. Soon because that funny weather we had 's got everything upset. DP It sure did. GD And with a winter like the one we had, with that down right cold weather we had'. . . . I don't know about you, but I almost froze. Then, we thought we'd get to work, the rain came, and now that it's spring and we could work, another cold rainy spellL So everything's late. We better get to it now. 58 DP 59 GD 60 61 ED 62 The 104 DP We're starting tomorrow. GD Tomorrow? DP Sure, we're going to start. My boy is getting the planter fixed up; we're starting tomorrow. And then we'll keep going a couple of days, then we'll pass the planter on to the next man. GD The season's right and the weather's finally cleared up I DP But we may get a thunderstorm, yet. Today is celebra- tion-day at Castelsarrasin. Saint Alpinien's anniver sary, and usually it's a stormy day. CD Quite often stormy. DP Well, what can we do? If it rains, we'll do like the folks at Auvillar, we'll just watch it fall. (All laugh) Say, how about our celebration? ED We're getting ready. DP We are? Do you think we'll have much of a crowd? ED Maybe. DP If you're going, we'll talk about old times. We've danced many a waltz, don't you know? ED Do you still remember how to waltz? DP I'd remember all right, but maybe my head would spin, 105 heyl Though, there are times when it might not. ED We'll have to try. DP We'll have to try, but I wouldn't want people to make fun of us too muchl At our age we must sacrifice a lot of things. GD That's right. DP What's a man going to do? These are things that we can talk about, you know. Our tongues work fine; we talk a lot . . . ED But we're getting old. DP But strength, no morel (All laugh) GD Have a lot of fun, and I'll be waiting for you. ED All right, then, it's settled: the two of us will go and celebr ate 1 1.4.4. Intonation Contours An intonation contour is a morpheme whose constituent elements are pitch phonemes and an external juncture. Pitch levels are symbolized by numbers: /l/> /2/, /3/, /4/, pitch /4/ being the highest and the least common. Each contour is a frame of three pitch levels, the first number representing the pitch at the onset of the phrase, the second the pitch of the syllable bearing phrase- 106 stress, the third the pitch on which the last segment of the phrase is spoken. Intervening pitch glides, if any, are noted by interpolated numbers. So as not to clutter the notation, the last number need not appear when it is the 19 same as the preceding one; thus, /23t/ stands for /233t/. This presentation of intonation contours is limited to the main types which are given below, disregarding momentar ily features of contrastive accent to be dealt with in sec tion 1.4.5. As many examples as possible are extracted from the illustrative text. Reference to this text is made by quotation of line numbers. Only those intonations which do not occur in the text are illustrated from other material. I. Questions {233t} indicates that the phrase is a question to be answered by "yes" or "no," e.g. line 16 /2dumant/; also cf. lines 3 (last phrase), 36, 37, 42. This intonation mor pheme alone distinguishes yes-no questions from statements identically worded. {2211} signals a question requesting an informative 19 The last pitch level is not manifest when there is no unstressed syllable or only voiceless segments after phrase- stress . In such cases the contour has in actuality only two pitch levels, e.g. /3es trawkati/ 'il est troue'. 107 answer, e.g. line 1 /2kuro lu samenes ty1!/; also cf. line 34. {244t} is a variant of contour {233t} with added sur- prise or indignation, e.g. /21 az embalado4 t/ 'tu I'as ava- lee' . II. Exclamations {233-*} turns a statement into a humorous message or conveys such emotions as joy or anger, e.g. line 10 /3ez estal la piezo -*/; also cf. lines 32, 33, 57, 58, 59. {211-*} concurrently with an introductory exclamative element signals an exclamation, e.g. /2ke m ennyzes1 -*/ 'que tu m'ennuies'. Exclamations may also appear with the same contours as statements. III. Statements {211i} denotes a completed assertion, e.g. line 30 /2sul3en1 l/; also cf. lines 28, 38, 60. /21211i/ (or /2 (12) lli/) is a conditioned variant of the pattern above. The occurrence of the intervening /12/ is probably dependent upon the length of the span preceding phrase-stress. Thus, if there is a stretch of more than some six syllables before phrase-stress, the speaker drops from /2/ to /l/ and rises back to /2/ before he completes 108 the contour on pitch /l/. Compare the short statements just cited and the phrase on line 52 /2per desys tut a nost aje1 aro 1/ with the longer statement on line 19 /2ban kumensa de lu samena duman \/. The absence of this expected pitch undulation over a long statement yields a monotone melody and may denote low spirits or indifference, e.g. line 21 /2e apej sederen lu semwer a yn 1awte1\/. {222-*} in absolute initial or in utterance-medial posi tion signals a short unfinished statement, e.g. line 9 /2ajfej -♦ Icrezjon de trabala -*/; also cf. lines 15, 20, 26, 30, 31, 51. {232-*} adds to the pattern above the indication that the speaker intends to elaborate upon the topic, e.g. line 54 /ssabes -*/• /2122-*/ (or /2 (1)22-+/) is a variant of {222-*}; it be longs to longer unfinished statements to be completed in a few words, e.g. line 29, the phrase which has seven syl lables before phrase-stress /2e d abitydo f^1 awraje2-*/. {21232-*} announces that a long incomplete statement will be followed by further exposition, e.g. the statement at line 25 to be completed at line 29; cf. also line 18. {2121232-+} is a complex pattern which shows a combina tion of the variants already illustrated for unfinished 109 statements. An example is found at line 11 with windup at line 14. There are two successive glides from /2/ to /l/ in the course of the extra-long initial stretch. IV. Miscellaneous {211-*} is used to set off an introductory phrase, e.g. line 1 /2e a prepaw-*/; also cf. lines 39, 44, 61. Appositive and afterthoughts repeat the contour of the phrase immediately preceding, e.g. line 1 /2lu mil-*/ repeats the contour just illustrated. Similar contours are also 20 found on lines 5 and 6, 12, 45 and 46. A final brief rise above the last pitch level means that the speaker wishes to elicit some sort of approval or comment from the listener. This is a common form of ex pression, equivalent to ending a phrase with /3et/ 'hein?'. This end-rise is symbolized by /t/ replacing /\/ or /-*/ in statement and exclamation contours, e.g. line 13 /se fa k es tut en^etar t/y also cf. line 24. ^There are a few words spoken in low voice— phrase- final adverbs akin to hesitators— whose status is not clear. The information they carry is somewhat redundant; they are placed inr the background so that the stress level is not perceptible and the pitch is even or lower than the preced ing level. For lack of a better solution, these bits of the spoken chain have been set up as discrete phrases, e.g. line 14 /1aro1i/; also cf. lines 23, 45, 50. 110 The basic intonation contours will be symbolized by the following punctuation marks: /./ for statements and commands, i.e. {2111}, e.g. /suben./; /:/ for unfinished statements, i.e. {222-*}; /?/ for information questions, i.e. {2211}; /'./ for exclamations, i.e. {211-*}. 1.4.5. Contrastive. Accent This term refers to the various means that speakers employ to add emotional overtones to a message. Gascon speakers favor such forms of expression and use an array of gestures of their own. Only some of the features of con trastive accent which are conventional will be illustrated here. The intonation patterns will now be found altered by the introduction of additional prosodic and phonetic fea tures . I. Emphatic stress Emphatic stress, symbolized by ", involves the expul sion of a larger volume of air than is normal for word pri mary stress. The use of this stress adds insistence— seri ous or amused— to the semantic load of a particular word, thereby setting it off from the rest of the phrase. This stress always occupies the same place as word primary stress; it may precede phrase-stress or coincide with it, Ill in which case the syllable thus stressed sounds extra-loud, e.g. line 34 /2e be akelo iDOto i/; also cf. lines 44, 49, 52, 53, 55. A jump to a higher pitch level occasionally goes' along with ", e.g. line 13 /2e fa k es t^it en retar t/; also cf. line 62. II. Displaced stress Displaced stress, symbolized by ", differs from " by its special placement, which enhances the contrastive ef fect. This stress excludes word primary stress and falls on a syllable which would normally have weak stress. In addi tion to the irregular stress pattern of the word which re ceives ", there is some displacement of the intonation con tour, with possible absence of phrase-stress. Displaced stress, then, affects a whole phrase. It is used essen tially as a rhetorical device. Examples are on lines 14, 15, 50, 58, 61. The irregular stress patterns of the words involved are here compared with the normal accentuation: kale, kale; duman, duman; nuzaw, nuzaw; zero, zero? entendyt, entendyt. " is found at line 3 on the function word /de/ 'de' with a simultaneous jump to pitch /4/, which 112 21 further emphasizes the contrastive effect. h i . other feature? Ironical emphasis on a specific word is rendered by a sudden rise to pitch /3/ on a syllable which is not neces sarily stressed. An example of each type is found at line 56 /2 la lengo2 -♦/, /3om parlo prampu2 -*/; also cf. lines 7, 39, 43, 44, 47, 55. A pitch depression at level /l/ expresses mild emphasis or determination, e.g. line 8 /1zu/; also cf. lines 35, 49, 61, 62. The clipping of a phrase-final syllable contrasting with the normal prolongation of that syllable denotes bom bastic style or willful exaggeration. This feature— left unmarked— is observable at the very end of the phrases on lines 5, 6, 38, 56. Failure to make all compulsory sandhi modifications at word border-points is also a device which creates a bombastic effect; cf. lines 3, 5, 6, where such words as /ame/ 'avec', /aket/ 'ce' appear in unaltered shape. 21-The stress shift to the function word /ke/ 'que* on lines 26, 45, 54 is not a contrastive accent but an acci dental displacement of word primary stress due to prolonged hesitation. 113 1.5. Free Alternation of Phonemes This section is meant to account for the fluctuation in the phonemic shape of Dzcs. words. The words cited here exhibit alternate forms involving the exchange and/or the addition of segmental phonemes. This type of alternation, which concerns mainly nouns, verbs, and adverbs of everyday use, has no morphemic relevance. The occurrence of the variants may be free or re stricted. These variants are said to be in free variation when all informants use them interchangeably. This is the case in over half of the data quoted. In other instances, informants are at variance as to the form they use. Some of these discrepancies are probably of dialectal origin, while others can also be accounted for in other ways. Various attempts at finding some consistency in the distribution of the variants among the few informants available have failed. Disagreement in usage between any two informants, including those who reside close together, prohibits the grouping of idiolects into actual dialects. The lack of dialectal uni formity observable inside the community must be described by taking the peculiarities of each of the idiolects into account. 114 The variants which are presumably due to no other cause than dialectal diversity or idiosyncrasy will be set aside under the label "sub-dialectal variation." The other data have been arranged under the headings: "bilingualism," "analogy," and "latitudes in pronunciation habits." This classification is, in a sense, one of convenience. Each heading furnishes the most likely cause for the emergence of a set of variants. There is no guarantee, however, that the cause given is valid in all instances. No one cause pre- 22 eludes the possibility of dxalect admixture. The distinction between alternate forms in free varia tion and forms occurring in specific idiolects will be ap parent from the system of notation. An informant who dif fers from the norm by using his own version of a word will be identified after the listing of that variant. Alternate forms separated simply by the symbol ~ are in free varia tion. When each member of an alternation is identified with an informant, the implication is that other speakers use 22lf some speaker picked up a form which is similar to the Dzcs. equivalent but characteristic of a related dia lect, living or extinct, it is gratuitous to correlate the two and treat the pair, say, as evidence of some type of phonetic change within the Dzcs. dialect. The small number of informants available makes it difficult to establish which form represents the Dzcs. norm. 115 both indiscriminately. I. Bilingualism Given a Dzcs. word and a Fr. word of identical meaning and of similar configuration, bilingual speakers will employ certain words either in their native form or in a variant form which shows contamination by the Fr. form. In the first set of examples, the re-shaping of syllables involves the substitution or the insertion of phonemes which are common to both sound systems. The native form is given first. Some examples are: /raska ~ raskla/ 'racier', /perdyt ~ perdyt/ 'perdu', /duplida ~ dublida/ 'oublier', /ajsaple ~ ajsable/ 'desagreable', /pusiple ~ pusible/ 'pos sible', /bo/o ~ b6so/ 'bosse', /trabesa ~ trabersa/ (GD) 'traverser', /tretamme ~ tretemme/ 'traitez-moi', /pr^kyrabbus ~ pr^kyrebbus/ 'procurez-vous', /din ~ dan/ 'dans', /feni/ (MC) ~ /fini/ (GD) 'finir', /defesille/ (MC) ~ /difisille/ (GD) 'difficile'. In the examples below the Fr. sounds have been modi fied in accordance with the patterns of phonological assimi lation to be examined in section 3; thus, /pyrgatori ~ pyrgatwero/ 'purgatoire', /portofelo ~ portofylo/ 'porte- feuille', /sepawza ~ serepawza/ 'se reposer'. 116 The Fr. uvular phoneme /r/ is often adopted intact, e.g. in /puri ~ puRi/ 'pourrir1, /muri ~ muRi/ 'mourir', as a substitute for either one of the Dzcs. vibrants. It will be noted that this substitution nullifies the native dis tinction between /r/ and /r/. This is facilitated by the low functional yield of this opposition, which is operative only in intervocalic position. The contrast in analogous environment is clear, but the words do not often belong to the same word-class and no minimal pairs have turned up. il. Analogy The following are instances of external grammatical analogy: /faj ~ fa/ 'il fait', /bij ~ bk/ 'il va'; the short forms may be due to analogy with the form /a/ 'il a' of the common verb /abe/ 'avoir'. In /sjegyt ~ sjegit/ 'suivi', the first form is the regular past part, of class III verbs, of which /sjege/ 'suivre' is a member; the second form follows the pattern of class lib verbs like /feni/ 'finir', /fenit/ 'fini'. The alternation /saburi ~ sajuri/ 'je sus' in the perfect of /sabe/ 'savoir' parallels the alternate forms of the same tense of /abe/ 'avoir' (see below). There is internal grammatical analogy combined with 117 Fr. influence in the present subjunctive /feniski/ (MC) ~ /finisi/ (GD) 'je finisse'. The absence of /k/ throughout this tense is due to analogy with the present indicative /finisi/ 'je finis'. The merging of these tenses is facili tated by the fact that most regular verbs have homophonous forms for the first two persons of the present subjunctive or indicative. Ill. Latitudes in pronunciation habits The following are trends in alternate pronunciations of words which may be indicative of sound changes in progress or which may continue to co-exist as phonemic latitudes. A. Palatalization of implosive J-z]S3 [-z] becomes [-j] or [—±] at word boundaries; for in stance, /laz bulygos/ ~ /laj bulygos/ ~ /lai bulygos/ 'les etincelles'. The form /laz/ is a widespread sandhi variant of /las/ 'les' resulting from the assimilatory voicing of word-final /-s/ in anticipation of a vowel or voiced con- ^This phenomenon concerns only [-z], not [—s ]. There is no trace at Donzac of a phonetic change in progress in the dialects of adjoining communities where implosive [-s] word-medial is at the aspiration stage, e.g. [pa^kado] 'ome lette ' . 118 sonant, e.g. /laz gitos/ 'les canes' (see sandhi 5.1.1.I). The form /laj/ alternates with /laz/ only before a voiced consonant. This alternation seems at present to be limited to monosyllables of high incidence, specifically forms of the verbs /esta/ 'etre', /abe/ 'avoir', the negative adverb /pas/ 'pas', and certain nominal markers; thus, /ej/ for /ez/ in /ej minar/ 'il est douillet', /aj/ for /az/ in /aj bebyt/ 'tu as bu', /paj/ for /paz/ in /paj nat/ 'aucun', /muj/ for /muz/ in /muj debasis/ 'mes bas', /dyj/ for /dyz/ in /dyj dits/ 'deux doigts'; but /estez/ is not affected, /* 1 / e.g. /estez busis/ 'ces morceaux', 24 This sound change tends to eliminate the sequence /z/ + consonant, which is not found word-medial, and to re place it by the familiar sequence /j/ + consonant. The voiced implosive sibilant is either palatalized to [-j] with audible friction, or vocalized to [”i] particularly before non-fricative consonants, e.g. [pai mei] 'pas davantage'. This change has become a speech habit with informants MC and MV, who preserve [-z] only sporadically. The reverse is 24it is of interest to note in passing that similar sound-changes have already taken place in the course of the history of this dialect, e.g. V.L. [nqkte] > modern [nei] 'nuit'; V.L. [patre] > modern [pai] 'pere'. 119 true of other informants. This is exemplified at line 20 of the illustrative text where DP used /trej/ for /trez/ in /trej zurs/ 'trois jours'. The next step may be either to retain the alternation /-z ~ -j/ as a generalized phonemic latitude in the dialect or to eliminate all implosive /z/'s. In the latter event, the palatalization of /-z/ would have to be described as an automatic morphophonemic change; for instance, {pas} would have three alternants: /pas/ before pause or before an un voiced consonant, /paz/ before a vowel, and /paj/ before a voiced consonant. B. The influence of palatal [j] (1) Fronting of [u] This alternation is subsequent to the palatalization of implosive [-z] and is limited to occurrence in the masc. pi. definite article /lus/ 'les'. Of the varying pronunciations noted as /luz ~ luj/ in, say, /luz balats/ ~ /luj balats/ 'les fosses', the second one shifts [u] in the direction of [y]: [lui] [luHi] [lyi]. The allophonic variant [uH] is a high centralized vowel. In the change to [y] the point of articulation of [u] has been modified by assimilation to 120 (1), but lip-rounding is preserved. A sequence [yi], i.e. /yj/, is thus emerging as a correlate of the existing se quence /yw/. (2) Palatalization of a consonant preceding [j] The following consonants are affected: /sj/ > /s/, e.g. /fenisjon ~ fenison/ (MC) 'nous finissions' (imperfect of /feni/ 'finir'), /sjoski ~ soski/ 'je sois1 (present subjunctive of /esta/ 'etre') /zj/ > /z/, e.g. /krezjoj ~ krezoj/ ’je croyais1 (imperfect of /kreze/ 1croire') /lj/ > /l/, e.g. /eskuljs ~ eskule/ 'ecolier', /miljun ~ milun/ 'million' /nj/ > /n/, e.g. /nja ~ na/ 'il y en a'. (3) Absorption of a consonant by hi There is loss of [b] in the imperfect of /abe/ 'avoir', e.g. /abjo ~ ajo/ 'il avait'. There is loss of [r] in the few conditionals that have alternate forms with /r/ in stem- final position, e.g. /awrjo ~ awjo/ 'il aurait', /kadrjo ~ kadjo/ 'il faudrait', existing alongside normal conditional forms like /ajmijo/ 'il aimerait'. 121 C. Syncope of a consonant An uncommon consonant cluster is simplified to conform with prevalent sequences, e.g. /e(n)stryit/ 'instruit', /e(n)stryksjun/ 'instruction', /aust/ (GD) /aus/ (MC) 1aout1 . D. Other alternations There is dissimilation in /lewze/ (MC) ~ /lawze/ (GD) 'leger' and distant metathesis of /r/ in /dintra/ ~ /drinta/ (MV) 'entrer'. The alternation /e ~ i/ is due to the closing influence of the nasal in /den ~ din/ 'dans'. The fluctuation affects only the preposition, which occurs always with weak stress in connected speech. The adverb /dedin/ 'dedans' which bears strong stress is fixed with /i/. Because of the low functional yield of the opposition ///, ///, the merger of these two phonemes is in progress, e.g. /kru/a ~ kru/a/ 'croiser', /du/e/ (MC) ~ /du/e/ 'douze'. IV. Sub-dialectal variation /esta/ (MC) ~ /este/ (GD) 'etre' are alternate forms of the infinitive of this verb. Other variants in the conju gation of this verb as a present participle, in the perfect, 122 and in the imperfect subjunctive will be discussed below in 4.2.4.3.111.A.(3). For variants of the verb /aje ~ abe/ 'avoir' in the infinitive, as present and past participle, in the perfect, and in the imperfect subjunctive, see 4.2. 4.3.111.D.(6). Other examples of common dialectal variants which have been noted are: /esternyda/ (GD) ~ /esturnyda/ (ED) ~ /estyrni/ (MC) 'eter- nuer1? the imitative sound of this verb is also a likely cause for the formal fluctuation; i 25 /ok ~ ot/ ’le' (enclitic neuter d.o. pronoun); /esklot/ ~ /esklop/ (MC) 'sabot* /empezo/ ~ /empes/ (MV) 'amidon1 /ten ~ tens/ 'temps' /aw men ~ aw mens/ 'au moins' /dambe/ ~ /ame/ (GD) 'avec' 2^The form /ok/ occurs in other Gascon dialects as well. Whether or not it originated in Dzcs., the variant /ot/ is likely to have emerged as a result of analogy to the common word-final sequence /-ot/ of nouns and adjectives. The al ternation is further facilitated by the fact that breath- group-final voiceless stops are generally unreleased, in which case the audible difference between /k/ and /t/ is extremely slight. Since /ok ~ ot/ is an enclitic pronoun, it normally occurs in breath-group-final position, e.g. [digozQk- ~ digozqt"] 'dis-le'. /apraki ~ aprakiw/ 'par la' /abanze/ (GD) ~ /azedela/ (MC) ’avant-hier' /tabe/ (DP) ~ /taben/ (MC) ~ /tabes/ (GD) 'aussi' /arse/ ~ /arses/ (GD) 'hier soir1 /aso/ ~ /asos/ (GD) 1 ceci ' /ako/ ~ /akos/ (GD) 'cela' Since no analogy has been found to explain this final /s/ in the last four examples, a dialectal origin, or perhaps an idiosyncrasy in GD's speech, must be assumed. The forms with /-s/ are rejected by the other informants. 2. Borrowing and Bilingualism Linguistic borrowing is widespread in Dzcs. The limi tations of the old indigenous word-stock are offset by the heavy importation of French words rather than by original invention. Inasmuch as all speakers of Dzcs. are also na tive speakers of French, borrowing is here a bilingual proc ess, the spontaneous transfer from the speaker's French vocabulary of any item which is wanting in the rural dia lect . The French words are seldom admitted into Dzcs. without change. The borrowing language requires partial or total accommodation to the native phonological and morphological patterns.^ In the loan-word type, the assimilation consists ^The Fr. phonemic shape of a word need not always be altered to fit the Dzcs. phonological patterns. One and the same form of a word may happen to meet the structural re quirements of both languages. Despite their common Gallo- Romance origin, the two languages are structurally quite different, so that the incorporation into Dzcs. of a loan word in its original shape is the exception rather than the rule. 124 125 in substituting native phonemes and permissible sequences for French ones. In the loan-blend type, one of the con stituent morphemes of the lending language is replaced by a native one; that is, the French model has been partly ana lyzed into its grammatical components. The meaning of the unit is preserved intact. Loans are identifiable by means of several criteria— formal, semantic, and grammatical, or by a combination of these. Words which represent new concepts, admitted into Dzcs. in the last few decades, exhibit a conspicuous simi larity to French, and this similarity contrasts with the indigenous form of the everyday words. The collation of the modern words with their French equivalents permits one to uncover patterns of adjustment of the French models to the native structure. These patterns— generally distinct from early historical sound correspondences— apply also to 2 most of the learned and semi-learned words. Besides the over-all shape of the word unit, the inci dence of unfamiliar sounds such as /q/> unusual diphthongs, vowels in hiatus, and complex consonant clusters betrays ^1 call "semi-learned" such words as Dzcs. /ryjo/ 'rue' which are used infrequently in comparison with the familiar terms /kamln/ 'chemin' and /ruto/ ’route'. 126 the French origin of certain words. Others vacillate in form, indicating that their integration is in progress. In the case of doublets, the comparatively learned semantic content and the form of one of the items both serve to dis tinguish the loan-word from the native word. Various dis crepancies in word-formation are due to the borrowing of one of the constituent morphemes, stem or affix. Loan-blends will be exemplified and contrasted with native formations in the section which treats of derivation (4.3). Provided words are recognizably recent additions to the dialect or are learned words, the preceding criteria iden tify a sufficient number of loans to establish the patterns of assimilation. These phonemic and morphemic correspond ences are also manifest in the many instances in which speakers transfer Dzcs. words into their French vocabulary. For example, the word /ankare/ with the meaning 'mettre de face' is the manufactured French equivalent of the Dzcs. verb /enkara/. The lending language is a regional variety of French (abbreviated RFr.) which differs from standard French as to sound-system and allophonic distribution. Before turning to the assimilation of loans, I shall present a brief analy sis of the essential features of the RFr. phonological 127 system. 2.1. The RFr. Phonological System The RFr. phonemes are shown in the chart on p. 128. The vowel phonemes contrast in /ni/ nid. /ne/ na, /ne/ nait. /ny/ mi, /n// noeud. /na/ na, /anana/ ananas. /nu/ ngu s., /no/ nos. Except in word-final position in a few words, /e/ and /e/ are mutually exclusive, [e] occurring in closed sylla bles and [e] in open syllables. Unlike standard Fr., RFr. has only one /a/ and no pho nemic distinction between [o] and [o] or between [/] and the lower-mid front rounded vowel [ee]; that is, patte is homophonous with pate. paume with pomme. and ~ieune with -ieune. [q] and [ce ] are respectively allophones of /o/ and ///, occurring only when these vowels are checked or nasal ized, e.g. [mqRdoRe] /moRdoRe/ mordore. The vowel /a/, which occurs only with weak stress, is shorter than all others when word-final and keeps the preceding syllable closdd, e.g. [scele] /s/le/ seule with [ce] as in [soel] /s/l/ seul. [e ce a q] are here not discrete phonemes but nasal ized allophones of the respective vowels /e / a o/ when checked by nasal consonants. The symbol [n] is here used 128 RFr Dzcs w w Y u u b t d k g P b t d k g / / V V £ c j V V V V V s z s z f s z s z V V m n n m n n 1 1 V 1 R r r 129 to subsume the released nasal as well as the implosive nasal. The articulation of the latter is always audible but limited to its onset before pause, e.g. [bo] /bo/ beau. [bSn] /bon/ bon, [bqna/ bonne. /j/ is a fricative in /jeR/ hier and /peje/ paver? it is distinct from the vowel /i/ in, say, /baj/ bail vs. /ebai/ ebahi. /vj and /w/ may be set up as separate pho nemes from /y/ and /u/ respectively, although the evidence is very scanty, e.g. /ekqels/ ecuelle vs. /Ryela/ ruelle. /lwi/ louis vs. /eblui/ ebloui. The RFr. consonants /p btdkgfvszszmnnl R/ are identifiable from the following series of verbs and adjectives: /py/ m., /by/ ill, Ay/ ill, /dy/ /veky/ vi.cu, /egy/ aiqu, /fy/ iiit, Ay/ yu, /sy/ su, /azyre/ .az.uii, /esy/ echu. /zyre/ jure. /my/ mi, /ny/ m, /kan// /ly/ Jji, /Ryd/ jrudg. The RFr. phoneme /R/ represents the uvular fricative of standard Fr. Under this phoneme are also subsumed the api cal vibrants [r] and [r] that a few speakers still use sporadically according to the distribution of these phonemes in the Dzcs. system (see 1.1.1.I above). Most elderly 130 r v , speakers still pronounce a palatal lateral LU in many words v in place of modern [j]. Those who employ [l] consistently in accordance with Fr. spelling have an additional phoneme /I/ which contrasts with /j/ in such pairs as /vels/ veille vs. /pejs/ nave.3 Word-stress does not have phonemic status in RFr. The occurrence of strong stress is definable in these terms: the last vowel of a word is stressed unless it is /a/, in which case the penult is stressed. In loan-words the place ment of this stress is usually adopted into Dzcs. without change. The phonetic feature of vowel length automatically conditioned by various following consonants is disregarded here because it is of no consequence in borrowing. Only in the case of /a/, as noted earlier, is vowel length signifi cant . ^Some of the differences between RFr. and standard Modern Fr. are survivals of earlier pronunciation habits. For example, the trilled [r] and [r], which disappeared from standard Fr. in the 17th c., the distinction between [j ] and [l], and the preservation of nasal consonants after nasal ized vowels are perpetuated in the speech of bilingual peo ple probably because these features are still present in the sound-systems of various Gascon dialects. 131 2.2. The Assimilation of Loans The RFr. and the Dzcs. sound-systems are placed side by side for comparison on p. 128. Only the patterns of assimilation of the vowels need be examined in some detail. The adjustments required in passing from one system to the other are best set forth by considering each of the RFr. vowels separately along with its Dzcs. phonemic equivalent. The RFr. models are transcribed in broad phonetics. Pat terns of phonological assimilation are presented first and the most widespread patterns of morphological adaptation are noted next as divergences from the former. RFr. [i] > Dzcs. /i/, e.g. RFr. [taksi] taxi > Dzcs. /taksi/ [ibrids] hvbride > /ibride/, [penisilina] penicilline > /penisilino/. RFr. [e] (stressed) > Dzcs. /e/, e.g. [asje] acier > /asje/, [kafe] cafe > /kafc/, [volqntje] volontiers > /buluntje/, /oRloze/ horloger > /orloze/ (cf. the nat. base /reloje/ 'horloge1). RFr. verb forms ending in [e], i.e. ~ee. are borrowed in the Dzcs. shapes /_a, -at, -ado/ according to the forms of class I verbs, e.g. the infinitive [falsifje] falsifier > /falsifja/ (cf. the nat. stem /faws/ 'faux'), 132 the past participle [falsifje] falsifie. -ee > masc. /falsifjat/, fem. /falsifjado/y also [filme] > /filma/ 'filmer', /filmat/ 'filme', /filmado/ 'filmee'. By analogy to several native derivational suffixes which have the form /-at/, the RFr. suffix [ite] -ite is adopted as /-itat/ in loan-words like /elektrisitat/ 'elec- tricite'. The adaptation of an occasional monomorphemic word such as [komite] comite as /kumitat/ is probably due to the fact that the ending [te ] was misanalyzed as a suf fix . RFr. [e] (unstressed) > Dzcs. /e/, e.g. [telefqne] 1tele- phone > /telefono/, [televize] televiser > /telebiza/, [setadiRa] c'est-a-dire > /setadire/, [betiza] betise > /betizo/ (cf. the nat. stem /bestjo/ 'bete'). RFr. [e] > Dzcs. /e/, e.g. [kanseR] cancer > /kanser/, [poema ] poeme > /poemo/, [peRseptoeR] per cepteur > /perseptyr/, but pretonic [e] > /e/ (see 1.2.1 above). I RFr. [en] > Dzcs. /en/, e.g. [boten] Bottin > /boten/, [ekRiven] ecrivain > /eskriben/ (with nat. base as in the verb /eskriwre/ 'ecrire'), [faRmasjSn] pharmacien > /farmasjen/ (replacing nat. /aputikajre/)y but [vaksen] vaccin > /baksin/ presumably after the verb [vaksine] vacciner > /baksina/. 133 RFr. [y] > Dzcs. /y/, e.g. [otobys] autobus > /otobys/, [Rygbx] rugby > /rygbi/, the noun [Rasy] repu > /resyt/ with /-t/ by analogy with native past participles (cf. the past participle /resebyt/ 1regu'). RFr. [/] > Dzcs. /y/, e.g. [bl^] bleu > /bly/, [/R/] heu- reux > /yrus/ with nat. suffix /-us/ '-eux'. RFr. [ce] > Dzcs. /y/, e.g. [tRaktceR] tracteur > /traktyr/, [poRtafceja] portefeuille > /portofylo/ (a hybrid compound with second element borrowed), [soeR] soeur. i.e. reliaieuse > /syr/ (a doublet of nat. /so/ 'soeur, i.e. parente1 ), [kceR] coeur > /kyr/ (a term of anatomy, doublet of the nat. sentimental term /ko/ ' coeur 1 ) .^ RFr. [a] (pre-tonic) > Dzcs. /e/ word-medial, e.g. [Rapoze] reposer > /repawza/ (alongside nat. /pawza/), [RatR^ta] retraite > /retreto/, [kRatQna] cretonne > /kretono/, [samansa] semence > /semenso/ (cf. the nat. verb /samena/ 1semer'); but the result is /o/ in [depaRtaman] departement > /departomen/, [apaRtaman] appartement > /apartomen/ be cause the sequence [aman] -ement is reminiscent of the na tive noun-forming suffix /-omen/. 4RFr. [oen] is expected to yield Dzcs. /yn/, but I found no examples of modern loans. 134 RFr. [a] (post-tonic) > Dzcs. /o/, /e/, or /i/ in word- final position depending to some extent on gender adapta tion . The result is /o/ for all feminine words, e.g. [batoeza] batteuse > /batyzo/. Since most Dzcs. feminine words end in /o/, a final /o/ is added when the RFr. original ends in the so-called "mute e," i.e. when -g. is present only in the orthography of words like rue pronounced [Ry]. The hiatus which would result from the addition of /~o/ after the vowels /y/, /e/ is automatically resolved by inserting the glide /j/, e.g. [Ry] m a > /ryjo/, [b.l^] kJLsu.e. > /blyjo/, [fyze] fusee > /fyzejo/. The result is /e/ for most masculine words, e.g. [sufRe] soufre > /sufre/. Other examples are: /ibride/ 'hybride', /ministre/ 'ministre', /sjerze/ 'cierge'. There are, however, masculine loans adapted with Dzcs. /o/ or /i/ for RFr. /a/, e.g. [meRs] maire > /mero/, [telefqna] tele phone > /telefono/, [Rsmeda] remede > /remedi/ (replacing the now obsolete /putingo/). Loans like /artisto/ ~ /artiste/ 'artiste', /posto/ ~ /poste/ 'poste' are masculine nouns with alternate forms ending in /o/ or /e/. The RFr. masculine noun [sylfats] sulphate > /sylfato/, but the end ing /o/ being reminiscent of feminine words, the Dzcs. word 135 has fluctuating gender. RFr. [a] > Dzcs. /a/, e.g. [sinema] cinema > /sinema/, [opeRa] opera > /opera/, [baladmzs] baladeuse > /baladyzo/. RFr. [an] > Dzcs. /an/, e.g. [banka] banque > /banko/, [RestoRan] restaurant > /restoran/, [dantifrisa] dentifrice > /dantifrisi/ (cf. the nat. base /den/ 'dent'). RFr. [u] > Dzcs. /u/, e.g. [sufRa] soufre > /sufre/, [sufRe] soufrer > /sufra/, [futbql] football > /fubol/. RFr. [o] (stressed) > Dzcs. /o/, e.g. [oto] auto > /oto/, [nymeRo] numero > /nymero/, [kilo] kilo > /kilo/. RFr. [o] (pre-tonic) > Dzcs. /u/, but /o/ is preserved in certain loans; /o/ and /u/ alternate in others, /u/ is fixed in such words as /mutyr/ 'moteur1, /trutwer/ 'trot- toir', /kula/ 'coller1, but /o/ remains in /atomiko/ 1ato- mique', /kolonio/ 'colonie', /moto/ 1moto1. There is fluc tuation in /p^litiko/ 'politique1, /ip°teko/ 'hypotheque', /pr^kyra/ 'procurer'. In pairs of loans like /telefono/ 'telephone' vs. /telefuna/ 'telephoner', /bote/ 'vote' vs. /buta/ 'voter' the widespread Dzcs. alternation /o/ ~ /u/ is manifest. RFr. [q] (stressed) > Dzcs. /o/, e.g. [qR] aJL (noun) > /or/, [kqla] colie > /kolo/, [aloR] alors > /alor/ (besides the nat. /alabe//). 136 RFr. [q ] (pre-tonic) > Dzcs. /u/ or /o/ as [o] above, e.g. [kalifQRni] Californie > /kalifurnio/, [lqRnete] lorgnette > /lurneto/; but /o/ remains in /transporta/ 'transporter' (cf. nat. /purta/ 'porter'). RFr. [qn] > Dzcs. /un/, e.g. [avjqn] avion > /abjun/, [vagqn] wagon > /bagun/, [bombarde] bombarder > /bumbarda/, [bulqn] bouillon > /bulun/ (cf. nat. verb /byli/ 'bouil- lir' ) . The RFr. sequence [wa] (stressed) yields Dzcs. /we/, e.g. [fulwaR] fouloir > /fulwer/, [muswaR] mouchoir > /muswer/ (replacing the nat. /mukadu/), [bonswaR] bonsoir > /bunswer/ (cf. the nat. noun /se/ 'soir'). The sequence [wa] is bor rowed intact, however, in very recent loans like the RFr. compound noun [pjenwaR] pied noir > /pjenwar/? the loan- translation would be */pe negre/. The RFr. sequence [wa] (unstressed) is preserved in Dzcs. as /wa/ in a few words like [kwafoeR] coiffeur > /kwafyr/ (now replacing nat. /perykje/). In fully assimilated loans the final result is /we/ after an intermediate stage /we/, thus [vwatyRe ] voiture > /bwatyro/ > /bwetyro/ > /bwetyro/. 5This is a nickname given to French colonists recently returned to the metropolis from North Africa. 137 Conclusions The patterns of assimilation of the vowels will now be summarized and commented upon. The vowels /i e y a u/, which are common to both systems, are borrowed into Dzcs. without change, except for pretonic /e/, which is replaced by /e/- Although the vowels /e o/ also pertain to both systems, phonemic substitution is often required to conform to native distributional patterns. Thus, word-final stressed /-e/, being rare except in verbs, is replaced by the more familiar mid-front vowel /-e/. Also, pretonic /o/, which subsumes [o] and [q], is normally replaced by high back /u/ in accordance with the Dzcs. tendency to avoid pretonic /o/, hence the vacillation /o ~ u/ in a good many loan-words. But under the strong influence of Fr. pronunciation and spelling, pretonic /o/ is at present admitted in certain loan-words. On the other hand, the replacement of RFr. [qn] by /un/ is quasi-automatic, since the sequence /on/ is g generally unpermitted in this dialect. The three RFr. vowels /s/, ///, and [ce ]— the latter an allophone of ///— are unknown in Dzcs. The mid-front 6The rarely used Dzcs. /on/ for RFr. [qn] on must be retained as such to avoid confusion with Dzcs. /un/ 'ou1. 138 rounded vowels [jzf] and [ce] are both assimilated to high /y/, the only front rounded vowel in the Dzcs. system. Since /y/ is chosen rather than /e/ or /e/, the influential fea ture here is lip-rounding rather than tongue-height. Unless RFr. [a] is equated with some familiar gender termination— /-o/ for fern, and /-e/ or /-i/ for masc.— its adaptation is governed by position in relation to word- stress. In post-tonic word-final position mid-central neu tral [a] is integrated as a sub-member of mid-back rounded /o/, phonetically the most similar native phoneme. The pho nemic area covered by Dzcs. /o/ was seen to extend toward the center and to include a centralized unrounded positional variant [o-*] approaching [a] (see 1.1.2). Conversely, in pre-tonic position higher-mid front unrounded /e/ is adopted instead of /o/ as the equivalent of RFr. / a / . This is so, partly because /o/ tends to be rejected from that position, partly because the [o]-type phones that can occur before stress are all rounded. The assimilation of consonants presents few problems. With the exception of /v/, all RFr. consonants are present in the Dzcs. phonemic system; therefore, only minor adjust ments are necessary to comply with native sequences. 139 RFr. [v] is automatically replaced by Dzcs. /b/, e.g. [avjqfn] avion > /abjun/, [vagqn] wagon > /bagun/, [televize] televiser > /telebiza/. The RFr. voiced labio-dental frica tive /v/ is assimilated to the native voiced bilabial stop /b/. Since the substitute is /b/ rather than the voiced labio-dental /f/, the widespread opposition voice vs. voice less takes precedence over the difference in point of arti culation. The choice of a stop as a substitute for a frica tive is also facilitated by the fact that the Dzcs. phoneme /b/ subsumes a widely distributed fricative allophone [b]. RFr. [z] is adopted unchanged in unanalyzed loan-words, e.g. [zyska] iusqu'a > /zyska/ (competing with nat. /dinka/), [obeRzins] aubergine > /oberzino/ (ousting the nat. /bjettaze/). In loan-blends like /utilage/ 'outillage' (cf. the nat. noun base /ytis/ 'outil') the RFr. original [utilazs] was analyzed as containing the suffix [azs] / I v . equated with Dzcs. /-aje/. The RFr. phoneme /vj is easily integrated into the Dzcs. sound system as a non-syllabic counterpart of the high front rounded vowel /y/, e.g. [pqize] puiser > /pqiza/ (cf. the nat. noun base /py^/ 'puits'). RFr. words which have alternate pronunciations with either [j] or [i] in intervocalic position are borrowed 140 into Dzcs. with /!/, e.g. [mitRajoeze] ~ [mitRaldezs] mi trailleuse > /mitralyzo/. All such RFr. words are spelled with the letters ill. representing currently the sound [j], so that the resulting Dzcs. /!/ is consistent with the spelling. RFr. consonant clusters are often reduced to common native combinations; for example, /r/ drops out in [pRopRjeteRa] proorietaire > /propjetari/. RFr. [gz] is equated to Dzcs. [dz], i.e. ///, hence [egzekyte] executer > /e/ekyta/. Certain sequences of three consonants in ad jacent syllables are preserved unchanged, e.g. [qbskyrite] obscurite > /obskyritat/ (cf. the nat. base adj. /esky/ 1obscur'); others are simplified by eliminating the foreign element; for example, /k/ drops out in [eksteRjceR] exterieur > /esterjyr/, [sanktifje] sanctifier > /santifja/ (cf. the nat. noun /sen/ 'saint'). The Fr. clusters made up of /s/ plus consonant were avoided in the past when word-initial by a prosthetic [e-], hence Dzcs. /espektatyr/ 'spectateur' (15th c. Fr. word), /eskeleto/ 'squelette' (16th c.). In more recent loans like /spesjalisto/ 1specialiste1 (19th c.), /standardiza/ 'stan dardised (20th c.), such initial consonant clusters have become permissible sequences. 3. Words and Minimal Phrases 3.1. Words Words are readily identified by native speakers, how ever altered their form may be in phrasal combinations. The evidence is twofold. When asked to repeat an utterance at the speed of dictation, informants would separate certain fractions by short pauses and refer to these fractions as "words." Furthermore, the two informants (AC and GD) who attempted to write the dialect in an orthography of their own devising marked the same word border-points by spaces or apostrophes. In order to reflect the native speaker's feeling that the word is a pertinent structural unit, a phonemic stretch which can be quoted alone with meaning, the word will be taken as a unit of analysis and broken down into its mor phemic components. This section will examine the criteria used to identify words. We distinguish full words from function words. Though 141 142 it is not possible in this dialect to draw a neat line of distinction, it is generally true that function words are phrasally-bound units which occur with weak stress when em bodied in a phrase,"'’ whereas full words are free forms which for the most part occur stressed in phrases. I. Phonological criteria A combination of the following criteria can serve to identify words: A. the occurrence of at least one vowel marking a syllable peak except in a few interjections and in monosyllabic func tion words which are subject to compulsory elision of the final vowel before another vowel, B. the arrangements of stresses, including the occurrence of sub-phonemic secondary stress as described earlier in 1.3.1.I, C. word transition features: (1) the external sandhi phenomena to be described below in 5.1 or the occurrence of a real or potential pause, (2) with very few exceptions the occurrence of hiatus, ^There is one main exception: enclitic pronouns which occur in phrase-final position and automatically receive phrase-stress, e.g. /la byti./ 'je la vide' but /byto lo./ 'vide-la 1. 143 (3) the absence at the beginning of a word of the consonants /r / / j/, at the end of a word: (4) the absence of the consonant clusters /pi kl bl gl fl pr tr kr br dr gr fr/ and of the clusters composed of a consonant plus the non-syllabics /w j q/, (5) the absence of the consonants /b d g / j f z z r/ except as sandhi alternants, (6) the absence of the unstressed vowels /e a y u/. One of the following criteria is sufficient to mark the end of a word: (7) the occurrence of the consonant clusters /-sk -ks -ws -js -ns -Is -rs/ checking a syllable, e.g. /bosk/ 'bois', /p£ks/ 'idiots', /byws/ 'boeufs', /pjels/ 1cheveux'; these clusters do not occur word-initially and they are separated by a syllable boundary in word-medial position, e.g. /kles-ke/ 'noyau'. (8) the occurrence of the consonants /-n -1 -p -s -/ -c/ checking a syllable, e.g. /kan/ 'chien', /putral/ 'poi- trail', /pa// 'paix', /papac/ 'jabot'. (9) the presence of many consonant clusters at word 144 transition which are found neither in word-initial position nor at word-medial syllable transition. Without an exhaus tive investigation of permissible sequences, a complete listing of these unpermitted clusters cannot be given. Some examples are: /zb zd zg zz ss/ as in /dyz balats/ 'deux fosses', /paz de// 'pas dix', /luz zwens/ 'les jeunes', /las saminejos/ 'les cheminees'. II. Grammatical criteria Phonological criteria may be supplemented by morpho logical and syntactic criteria, such as: A. the occurrence of bound morphemes, including prefixes, suffixes, and bound bases, B. the occurrence of word-final inflectional suffixes show ing syntactic congruence with other words. 3.2. Compounds vs. Phrases With few exceptions the constituent elements of com pounds are subject to the same sandhi modifications as are the components of phrases. In the identification of com pounds, grammatical criteria— morphological as well as syn tactic— play an important role, while phonological criteria in general play a secondary role. 145 The following criteria, enumerated in decreasing order of importance, distinguish compounds from phrases: (l) grammatical criteria a. the indivisibility of the unit (i.e. the construction is closed and no further elements can be inserted), e.g. the adv. /en lok/ (prep. + N) 'milie part' . b. the absence of function-words joining the constituents, e.g. /kyro nrw/ 'dernier ne' (vb. + N) with no marker before / • / 2 the N /niw/. c. a form-class which differs from that of the main element, e.g. the N /biro 1 el/ 'louchon' (vb. + def. art. + N) . d. the inherent gender of a construction, e.g. the N /porto munedo/, m. 1porte-monnaie' (vb. + fern. N) or its indiffer ence to gender, e/g. the N /krebo de fame/ 'meurt-de-faim'. e. the plural formation of the unit. In those compounds which have an overt plural marker, generally only the final component is pluralized, e.g. with {-s} /busum blans/ 'aube- pines' (N + adj.). 2The specific function word implied is occasionally revealed in spontaneous French translations of indigenous words; for example, those speakers who do not have the standard French noun saumure in their vocabulary translate Dzcs. /ajgo saw/ as eau de sel. thus inserting the preposi tion .de. (Dzcs. de) between the two nouns. f. the absence of gender concord, e.g. a masc. adj. modify ing a fern. N in /gram maj/ 'grand'mere'. g. the syntactic patterning of a compound, for instance its ability to occur in phrases in the same slot as a simple word of the same form-class? e.g. the adj. /bum markat/ 'bon marche' can occur preceded by the modifying adverb /pla/ 'trfes' and so can simple adjectives. h. a reversal of the word-order of phrases, e.g. the adj. /ber/ 'vert' preceding rather than following the N in /ber zus/ 'verjus'. i. the occurrence of unique or obscure constituents, e.g. the initial element in the N /li/o krabo/ 'chevre-feuille'. (2) Phonological criteria a. certain stress-patterns (see 1.3.1.II), e.g. the com pound N /bjen fej/ 'bienfait' versus the phrase /bjen fej/ 'bien fait'. b. special combining forms resulting from phonemic modifi cations at word-boundaries other than those which are due to phrasal sandhi, e.g. /bew/ for /bet/ 'beau' in /bew paj/ 'beau-pere' (adj. + N). 3.3. Minimal Phrases A word is a minimum free form in the sense that it is 147 the smallest meaningful unit that the unsophisticated native speaker can easily single out and use as a citation form in linguistic discussions. In actual speech situations, how ever, there are limitations to the freedom of occurrence of certain types of words. Function words can seldom constitute a meaningful one- word utterance, although elliptical questions like /dympej?/ 'depuis ... ?' are not uncommon. On the other hand, a large number of full words frequently occur alone. For example, the adjective /troso./ 'menue' is a possible answer to the question /kumo 1 as trubado?/ 'comment l'as-tu trouvee?' In this dialect, all such full words belong to one of the following classes or sub-classes: adjectives, adverbs, verbs, numerals, pronouns other than oblique and relative, and interjections. Conversely, common nouns and certain verbs and verb tenses are bounded forms; that is, they always occur in conjunction with other words, the smallest combinations being those with one function word. Such short utterances, which may be called minimal phrases, are of the following commonest types. 148 I. Minimal phrases with nouns The smallest phrases with nouns as heads also contain a nominal marker or a preposition, e.g. /de kar./ 'de la viande', /dambe plaze./ ' avec plaisir'. II. Minimal phrases with verbs A. Types of verbs (1) Exclusively transitive verbs occur in clauses with at least an accusative pronoun, e.g. /aw f&w./ 1je le fais'. (2) Exclusively pronominal verbs do not occur without an oblique pronoun, e.g. /akato te./ 'baisse-toi'. (3) Some intransitive verbs such as /ana/ 'aller' require at least the prepositional phrase substitute /i/ 'y', e.g. /i baw./ 'j'y vais1. (4) The verb /esta/ 'etre1 is always accompanied by some function word like the pronominal adverb /ne/, e.g. /n es./ 'il l'est'. B . Verb tenses (1) Tenses of the subjunctive mood require the conjunction /ke/ 'que' preceding the verb form, e.g. /ke bele./ 'qu'il veuille'. (2) The present participle occurs only after the prepos tion /en/ 'en1, e.g. /en triman./ 'en se donnant de la peine'. 4. Morphology 4.1. Basic Morphophonemics Morphophonemic alternations in Dzcs. are unevenly dis tributed and of many different types. In this section are presented and exemplified only those alternations which are manifest in stems of more than one word-class or in stems as well as in affixes. Further exemplification of these and description of other alterna tions confined to a single word-class will be found in various sections of the morphology. I. Phonologically conditioned alternations A. Consonants (1) Reduction of consonant cluster? Clusters of two consonants which remain intact whatever their position in a word are marked by underscoring the second consonant in the stem formula; thus fres&A represents 150 151 the phonemic shape /fresk/ both in /fresko/ 'fraiche' and /fresk/ 'frais'. Clusters of two consonants which are un- permissible word-finally (see 1.2.V above) are automatically reduced in that position by deleting the second consonant. No symbol; e.g. prunt > /prunto/ 'prompte' vs. /prun/ 'prompt';^" bende > /bende/ 'vendre' vs. /ben/ 'il vend'. Word-final clusters of three consonants that would result from the addition of /-s/ are automatically reduced to per missible clusters of a single C + /-s/ by the same process, e.g. prunt > /pruns/ 'prompts'. (2) Replacement of ,/b d g z/ Voiced /b d g z/, which cannot occur word-finally or before /-s/, are replaced in these two positions. The 2 regular substitute for any one of these four consonants is •* -To be read: the stem prunt yields phonemic /prunt/ as in /prunto/ 'prompte' versus phonemic /prun/ in /prun/ 'prompt'. ^As noted by Hockett in A Course in Modern Linguistics (New York, 1958), p. 280, "The term 'automatic' refers to the fact that the primary form is replaced, but not to the particular replacement which is made." Therefore, several types of replacement are possible, as is the case here. The fact that these voiced consonants are not permitted in cer tain positions is a phonological requirement of the lan guage. But there is more than one type of replacement: one regular (most frequent), others irregular (less common). The latter will be listed in full later on, e.g. /kreze/ 152 its unvoiced counterpart, a partial assimilation to pause. I | I Symbol 0} e.g. eberid > /eberido/ 'eveillee' vs. /eberit/ 'eveille', /eberits/ 'eveilles'; yrug* > /yruzo/ 'heureuse', /yruzis/ 'heureux1 vs. /yrus/ 'heureux1. (3) Replacement of /l n/ Palatal /l/ and /n/ are fronted to dental /l/ and /n/ respectively in word-final position and before /-s/. Sym- v v | v . | v . . , bols L N; e.g. pareL > /parelo/ 'pareille' vs. /parel/ 'pareil', /parels/ 'pareils'; plaNe > /plane/ 'plaindre' vs. /plan/ 'il plaint'. This alternation is due to the non occurrence or the word-final clusters *ls, *ns, and it is automatic only to the extent that /l/ and /n/ are very rare in that position, e.g. in /putral/ 'poitrail', /kan/ 'chien1. (4) /y..~..b/ This is the replacement of labio-velar /w/ by bilabial /b/ in intervocalic position. Written phonemically, this alternation appears as a substitution of a stop for a frica tive; but only the fricative allophone [b] of /b/ can occur in this environment, so that the alternation is in fact an 'croire', /krej/ 'il croit'; /kryzo/ 'crue1, /kry/ 'cru'. 153 exchange of fricatives. It is automatic insofar as inter vocalic /w/ occurs only in a few derived items (see 1.2.Ill above). Symbol W? e.g. beWre > /bewre/ 'boire' vs. /bebyt/ 'bu'; klaW > /klaw/ 'cle' vs. /klaba/ 'fermer a cle'. (5) Vocalization of /i/ After a consonant the stem-final palatal /j/ is vocal ized to /i/ word-finally and before /-s/. This alternation is automatic and due to the phonological limitations in word-final C clusters. Symbol J; e.g. nobJ > /nobjo/ 'ma- riee' vs. /nobi/ 'marie'? ordin&rJ > /ordinarjo/ 'ordinaire' vs. /ordinari/ 'ordinaire'. B. Vowels (1) Loss of a vowel A base-final or a suffix-final unstressed vowel is lost when a vowel-initial suffix follows. This is the normal 3 process by which hiatus is avoided. Hiatus is invariably avoided at morpheme boundaries in the body of the word. This alternation is automatic to the extent that in the ^Two monosyllabic verb stems are an exception in that the stem-vowel is preserved rather than the inflection- initial vowel, e.g. /tja-/ + /ijoj/ > /tjajoj/ 'je tuerais' (see 4.2.4.3.III.A.[1-2]). 154 native vocabulary hiatus occurs only in a few words which are essentially onomatopoetic (see 1.2.IV above). No sym bol? e.g. /pigaso/ + /-un/ > /pigasun/ ’hachette'? /pud-/ + /e/ + /i/ > /pudi/ 1 je taille' . (2) Gradation /e ~ e/ Mid-open /e/ is raised to mid-closed /e/ in open syl lables when the stress shifts onto the following syllable. This automatic vowel change is due to the non-occurrence of unstressed /e/ in a pre-tonic open syllable (see 1.2.1 above). Symbol C? e.g. resSgo > /resego/ 1scie' vs. /resega/ 1scier'; m£te > /mete/ 'mettre1 vs. /metyt/ 'mis'. II. Morphologically conditioned alternations A. Consonants (1) Loss of /r/ This is the replacement of /~r/ by zero. This occurs in certain stems in absolute word-final position and before the plural suffix /-s/. Symbol [r]; e.g. the adj. pryme[r] > /prymero/ 'premiere' vs. /pryme/ 'premier'? the suffix -e[r] as in the N /burdilero/ 'metayere1 vs. /burdile/ 'metayer', /burdiles/ 'metayers'. But cf. ser > /sero/ 155 1chere', /ser/ 'cher’, /sers/ 1chers1. (2) /l ~ t/ This is the replacement of the voiced dental lateral /l/ by the voiceless dental stop /t/. It occurs under the conditions just stated for the loss of /r/. Symbol L; e.g. akeL > /akelo/ 1celle-la' vs. /aket/ 'celui-la', /akets/ 'ceux-la'. But cf. sul > /sulo/ 'seule', /sul/ 1seul', /suls/ 'seuls'. (3) Assimilatorv voicing In certain stems a single final voiceless consonant becomes voiced before a vowel-initial suffix. Symbol e.g. peg'' 'poids1 > /peza/ 'peser1 , /pezis/ 'poids'. But cf. pasA 'pas' > /pasis/ 'pas'. (4) /K ~ t/ This is the replacement of the stem-final voiceless palatal stop /k/ by /t/ before plural /-s/; that is, a re gressive assimilation, the occurrence before the dental sibilant /s/ of the homorganic stop /t/. Symbol K; e.g. enemiK > /enemik/ 1ennemi' vs. /enemits/ 'ennemis'. But cf. fek > /fek/ 'feu1, /fcks/ 'feux'. 156 (5) Loss of /n/ This is the replacement of stem-final /n/ by zero be fore plural /-s/. Symbol (n); e.g. razl(n) > /razin/ 'rai sin' vs. /razis/ 'raisins'. But cf. bulan > /bulan/ 'fau- cille courbe1, /bulans/ 'faucilles courbes'. (6) Extended stems Stems may occur in extended form before a suffix. Cer tain stems are augmented with V + C, others with a single C. Symbol -; e.g. gus-t 'gout' > /gustis/ 'gouts', /gustus/ 1 savoureux' vs. /gus/ 'gout'. B. Vowels (1) ,/e ~ e/ In certain stems mid-open /e/ is raised to mid-closed /e/ in closed syllables when the stress shifts onto the following syllable. Symbol £? e.g. finCstro > /finestro/ 'fenetre' vs. /finestrun/ 'petite fenetre'; Cste > /este/ 'etre1 vs. /estut/ 'il fut'. But cf. emmello > /emmello/ 'amande', /emmelle/ 'amandier'; perde > /perde/ 'perdre', /perdyt/ 'perdu'. (2) /e ~ e/ In certain stems mid-closed /e/ is lowered to mid-open 157 /e/ under stress. Symbol E; e.g. mEstr- > /mestreso/ 'mai- tresse' vs. /mestre/ 'maitre'; lEba > /leba/ 'lever' vs. /lebo/ 'il leve'. But cf. pela > /pela/ 'peler', /pelo/ 'il pele'. (3) /q_~_u/ /o/ is raised to /u/ when the stress shifts away from 4 /o/. This alternation is regular in all stem morphemes. Symbol O; e.g. tOrt 'boiteux' > /torto/ 'boiteuse' vs. /turteza/ 'boiter'y tOse > /tose/ 'tordre' vs. /tusyt/ 'tordu'. 4.2. Inflection 4.2.1. Form-Classes Stems are assigned to one of the following classes on the basis of inflectional characteristics and formal varia tion: (1) Nouns (N), which show number inflection (plural). In addition, nouns inherently belong to one of two large gender classes (masculine or feminine). ^But it is not automatic, since /o/ can occur un stressed in words like /kargomen/ 'chargement'. 158 (2) Adjectives (adj.), which are inflected for gender (masc. or fem.) and number (pi.). The choice of inflection is determined by concord with a N or N substitute. (3) Pronouns (pron.), which are N substitutes and which show one or more of the following grammatical categories: person and case by formal variation (see 4.2.3.2 below), gender (masc., fem., or neuter) and number (pi.) by inflection. The choice of number and gender inflection is determined by cross-reference to a N or a contextual feature. (4) Verbs (vbs.), which are inflected for actor-number and/ or for tense-mode-aspect (see 4.2.4.2). In addition to this classification there is a large number of uninflected words classed by syntactic criteria alone into adverbs, pronominal adverbs, prepositions, con junctions, coordinators, attributive universals, particles, and interjections. 4.2.2. Nouns and Adjectives 4.2.2.1. Gender I. Nouns The singular (sg.) form of a N constitutes its stem, which may be a base or a derived stem. Most N belong either 159 to the masc. or to the fem. gender-class. Assignments are usually arbitrary, so that the gender must be given for each N; thus fabo, f. 'feve', kujre, m. 'coude'. In N which have animate referents, the gender normally corresponds to the sex, e.g. ome, m. 'homme', male, m. 'male1, fenno, f. 'femme1, fymelo, f. 'femelle'; but N des ignating undomesticated animals for which no sex differen tiation is usually made may be of either gender, e.g. lebre, N V I 1 f. 'lievre', samow, m. 'chameau', semiK, f. 'punaise1, I piwze, f. 'puce'. The gender of a N is not consistently shown by its phonemic shape. A large number of N ending in a consonant or /e/ are masc., but exceptions are not uncommon, as can be seen in the examples above and in sun, m. 'son, bruit', sun, f. 'sommeil'; klaw, m. 'clou', klaw, f. 'cle'; set, f. 'soif'? talen, f. 'faim'; fame, f. 'manque de nourriture'. Except for rare N like maro, m. 'belier', manto, m. 'man- teau', native N ending in unstressed /o/ are fem.; but cf. learned or borrowed N like mero, m. 'maire', kSmisero, m. 'commissaire', kasko, m. 'casque', ipodromo, m. 'hippo drome', aristokrato, m. 'aristocrate' alongside ministre, I m. 'ministre', artiste, m. 'artiste'. Nouns ending in other vowels or in stressed /o/ are of either gender, e.g. 160 so, m. 'sou', so, f. * sosur' ; bapu, f. 1 vapeur' ; kutelu, m. 'petit couteau'; nymero, m. 1numero'. A few nouns are indifferent to gender, i.e. belong to both gender-classes, e.g. zitano 1gitane1, ibrono 1ivrogne1 II. Adjectives An adj. stem, i.e. base or derived stem, is defined as that phonemic stretch which precedes the fem. inflection /-o/. This primary allomorph of the stem always occurs un changed before the masc. inflection /-e/. The gender morphemes masc. and fem. are mutually ex clusive suffixes which follow the stem. These morphemes have the regular allomorphs {fem.} /-o/, {masc.} /$ ~ ~e/* One large group of adj. stems take masc. zero, fem. /-o/. This is the prevailing type of gender formation. An adj. stem of this group is represented by a formula such as sek which subsumes a fem. form /seko/ 'seche' and a masc form /sek/ 'sec'. Some other examples: estret 'etroit', kek 'begue', petit 'petit', kadyk 'sot', bjok 'depareille'. A second, somewhat smaller, group of adj. stems takes masc. /-e/, fem. /-o/. These are designated by the symbol e, thus prope > fem. /propo/ 'propre', masc. /prope/ 'pro- I ( pre'. Some other examples: abille 'habile', magre 161 'maigre', bewze 'veuf1, agre 'aigre', sale 'sale1. There is a small group of adj. stems which do not show gender; that is, these stems have a single form for masc. and fem. These are identified as adj. rather than N because they function as noun-modifiers. The singular form consti tutes the stem. These stems are marked with *. They are: braj* 'vrai1, bestjo* 'bete', forso* 'bien des ' , kado* 'chaque', memo* 'meme', the ordinal adj. trezjemo* 'troi- sieme', and higher ordinals. A few adj. are inflectionally defective. The fem. forms /njezo/ 'niaise', /bomjo/ 'enjoleuse', /mylo/ 'ste rile' have no matching masc. forms. The masc. form /fringyr/ 'amoureux1 has no matching fem. form. For the distribution of irregular allomorphs, see 4.2.2.3.II.C, 4.2.2.4, and 4.2.3.2 below. In summary, the allomorphs of the gender morphemes which occur with adj. stems are: fem. /-o -a -jo/, masc. /$ -e -u -un/. 4.2.2.2. Number Plural (pi.) number is indicated by an inflectional suffix. The pi. suffix is word-closing; in N it follows the stem and in adj., the gender inflection. Singular number is signaled by the absence of pi. inflection; that is, there 162 is no singular morpheme. With both N and adj. stems the morpheme pi. has three allomorphs: /-s -is $/, /-s/ being by far the most widespread. PI. /-s/ occurs with all stem allomorphs ending in a vowel, e.g. the N esplingo, f. 'epingle' > pi. /esplingos/ 'epingles'; the N musy, m. 'monsieur' > pi. /rausys/ 'mes sieurs' . Consequently, all adj. stems inflected for fem. form their pi. regularly with /-s/, as do all adj. stems inflected for masc. with /-e/, e.g. the adj. rette 'raide' > fem. pi. /rettos/ 'raides', masc. pi. /rettes/ 'raides'. All three allomorphs of {pi.} occur with stem allo morphs ending in a consonant, /-s/ being again the most common; e.g. to the masc. sg. adj. forms /dret/ 'droit', /dus/ 'doux', /gris/ 'gris' correspond the masc. pi. forms /drets/ 'droits', /dusis/ 'doux', /gris/ 'gris'. In N and adj., {pi.} /-is/ and /0/ occur only after stem-final sibi- lants, affricates, or consonant clusters; /-is/ and /// are in complementation relative to individual stem mor phemes . ^But in pronouns (see 4.2.3.2 below), {pi.} /-is/ oc curs once after the stop /t/ in /tutis/ 'tous' and once after the affricate /// replacing /t/ in /e/is/ 'eux', the pi. form of /et/ 'lui'. 163 One group of stems, the smallest, takes pi. zero. I _ These stems are marked with ; thus the N bwes-, f. subsumes a pi. form /bwes/ 'voix' overtly identical with the sg. form. These stems are: I _ 1 _ the adj. stems faw^.- 'faux' and gnz- 'gris', 0 the adj. derived by means of the suffix ~eg, e.g. frans4z= 'frangais1, the N stems: bi/=, f. 'cep de vigne' faws=, f. 'faux' bwes=, f. 'voix' kru/=, f. 'croix' dilys=, m. 'lundi' ly^=? f- 1lumiere' dimars=, m. 'mardi' pr£/=, m. 'prix' dimekres^, m. 'mercredi' printens=, m. 'printemps' dijaws=, m. 'jeudi' Py^=, m. 'puits' eskis=, m. 1dechirure1 setis=, m. 'siege' estyc=, m. 'etui' A second small group of stems takes pi. /-is/. These stems are marked with thus the N osA, m. ‘os' subsumes a pi. form /osis/ 'os', the adj. gras*, 'gras', a masc. Pi. form /grasis/ 'gras'. This group is composed of all N and adj. stems ending in sibilants, affricates, or consonant clusters other than those listed as pertaining to the pre ceding group. Some other examples: adj. stems raw]^ 164 I ^ I I 'rauque1, rusA 'roux', danjeruz* 'dangereux'; N stems bosk*, O I /v 1 m. 'bois, foret', debas~, m. 'bas (vetement)', pais*, m. ■pays', pes', m. 'poisson', repos'', m. 'repas1. All N which have this type of gender formation belong to the masc. gender class. The third and largest class of stems forms the pi. regularly with /-s/. This group includes all remaining N and adj. stems. These are left unmarked, e.g. the N fawtyl, I y m. 'fauteuil1, estiw, m. 'ete'. For semantic reasons, certain N always appear with pi. inflection, while others cannot be inflected for pi. The latter are identified as N rather than adj. because they belong to a gender class by definition. With N used only in the pi. the stems noted are under stood to be hypothetical sg. forms; thus, *fresA, m. 'frais1 inferred from the pi. form /frezis/ 'frais'. Among these are, for instance, N which designate pairs of objects, e.g. I I I v *kawso, f. < /kawsos/ 'pantalons', *estanalo, f. < , | v /estanalos/ 'tenailles', or a substance made up of individ ual units, e.g. *sendre, f. < /s^ndres/ ' cendres' , *fens'‘, m. < /fensis/ 'fumier'. Used only in the sg. are N such as: the names of the twelve months, e.g. zambje, m. 'janvier', 165 fewre, m. 'fevrier'; mass Nj e.g. bren, m. 'son de ble', tor, m. 'glace', zys, m. 'jus'; I V I collective N, e.g. munde, m. 'gens', gresilun, m. 'cresson'; abstract N, e.g. pa/, f. 'paix', perdun, m. 'pardon'; N referring to objects that are unique or are so regarded, e.g. s&w, m. 'ciel', bjerzes, f. 'Vierge', karmal, m. 'ere- v maillere', krinun, m. 'pointe du pain long', putral, m. 'poitrail'. 4.2.2.3. Alternations in Stems Only stems ending in a consonant are subject to allo- morphic variation. There are alternations in the stem- final consonant due to the presence of masc. zero and/or of the pi. morpheme. Unless specified otherwise, the following alternations occur; in adj. stems before masc. zero, followed or not by pi. /-s/ or pi. zero; in N stems before pi. /-s/. 166 I. Phonologically conditioned alternations6 (1) Reduction of C clusters This is the loss of the final C of an adj. stem ending in a cluster which is not permitted word-finally. All un marked stem-final C clusters are thus reduced. Adj. type nigawd > fem. /nigawdo/ 1nigaude1, masc. /nigaw/ 'nigaud', masc. pi. /nigaws/ 1nigauds'; e.g. fund 'rond1, minard 'douillet', entenerk 'dur d'oreille', blank ’blanc’, I v | I l lung, 'long1, brilent 'brillant', kurt 'court1, malawz 1malade1. Permissible C clusters are preserved word-finally in the adj.: rawk* 'rauque1, fresk* 'frais', faw_s= 'faux', and in the N bosk*, m. 'foret, bois'. (2) Replacement of /b d g z/ a. Regularly by unvoicing of /d g z/ Adj. type fad > fem. /fado/ 'folle', masc. /fat/ 'fou', masc. pi. /fats/ 'fous'; e.g. testyd 'entete', myd 'muet', flasid 'fletri', eskansarid 'epuise', peg 'idiot', pegu^* 6Those alternations which have been described as basic in section 4.1 are here listed first under each type of conditioning. 167 'poisseux', famyz'' 1 fameux' . b. Irregularly by loss of /-z/ only in kry[z] 'cru'. c. Irregularly by substituting /-w/ for /-b/ in : neb ~ i 7 1 1 new 'neuf1 and bib ~ biw 'vif'. (3) Replacement of ,/l n/ a. Regularly by fronting of /~l/ v v ( Adj. type bjeL > fern, /bj£lo/ 'vieille', masc. /bjel/ ■vieux, vieil1, masc. pi. /bjels/ 'vieux'. Also of this . V type is pareL 'pareil'. b. Irregularly by loss of /-*n/ only in the N ka[n], m. 'chien1, i.e. pi. /kas/ 1chiens1. (4) ,/ p ~ t/ Stem-final /-p/ regularly yields /~t/ before word- final /-s/. This is an automatic replacement due to the non-occurrence of the C cluster */ps/. Symbol P. N type taP, m. > sg. /tap/ 'coteau', pi. /tats/ 'coteaux'; i A Q e.g. luP, m. 'loup1, koP, m. 'coup', kaP, m. 'tete'. ^Where two forms of the stem are listed, that which occurs before fern. /~o/, here neb, is given first. ^That only N stems are subject to this alternation is fortuitous. 168 (5) Vocalization of ,/i/ Adj. type nesesarJ > fem. /nesesarjo/ 'necessaire', masc. /nesesari/ 'necessaire'; e.g. kuntrarJ 1contraire'. (6) ,/s ~ , i ,/ Not permitted word-finally and before /s/, stem-final /-z/ is automatically replaced by /j/ in the adj. mjez ~ mjej 'demi', i.e. fem. /mjezo/, masc. /mjej/. II. Morphologically conditioned alternations (1) Loss of , /r/ Adj. type esky[r] > fem. /eskyro/ 'obscure', masc. /esky/ 'obscur'. Of this type are the adj. stems: pryme[r ] 'pre mier', kla[r] 'clair', mady[r] 'mur', dare[r] 'dernier', lawze[r] 'leger', regylje[r] 'regulier', and all adj. stems derived by means of -e[r] and -je[r]; e.g. gawsefr] 'gau- cher'. (2) /I ~ t/ Adj. type saduL > fem. /sadulo/ 1repue1, masc. /sadut/ 'repu', masc. pi. /saduts/ 'repus'. Of this type are the adj. stems: akeL 'ce ... -la', beL 'beau', and nubeL 'nou veau ' . 169 (3) Assimilatorv voicing This alternation occurs in N stems pluralized with /-is/. N type mesA, m. > sg. /mes/ 'mois', pi. /mezis/ 'mois1. Of this type are: pesA, m. 'poids' and hypothetical fresA, m. 'frais'. (4) /K_~_fc/ This alternation occurs only before pi. /-s/. N type floK, m. > sg. /flok/ 'bouquet', pi. /flots/ 'bou- I I I quets1 ; e.g. abriK, m. 'abri' , limaK, m. 'escargot , amiK, m. 'ami', preseK, m. 'peche male', furmiK, f. 'fourmi'. This alternation is regular in N stems with only three ex ceptions: bek, m. 'bee', fek, m. 'feu1, and respek, m. 'respect'. Conversely, nearly all adj. stems ending in /k/ remain unaltered; thus sek 'sec' > masc. pi. /seks/ 'secs'. The alternation occurs once in the adj. fredelyK 'frileux' > masc. pi. /fredelyts/ 'frileux'. (5) bogs of /p . / a. Stem-final /n/ is deleted only before pi. /-s/. Adj. type bu(n) > masc. sg. /bun/ 'bon', masc. pi. /bus/ 'bons'. Of this type is also the adj. gasku(n) 'Gascon'. The loss of /n/ is the norm in N stems. The data 170 comprise some 100 indigenous N with stem-final /n/. Some two-thirds of these belong to the (n)-group, i.e. drop /n/ before /s/, while others belong to the n-group, i.e. have a single stem allomorph. Members of the n-group include: i. all N stems ending in /-en/, /-en/, e.g. ben, m. 'vent', mazen, m. 'pampre1, sen, m. 'saint', ii. all N stems ending in /-an/ except ma(n), f. 'main', e.g. an, m. 'an', arkulan, m. 'arc-en-ciel', emban, m. 'hangar', iii. all N diminutives derived with /-un/, e.g. esklapun, m . 'copeau', iv. the following N stems (lexically defined): kryn, m. 'nuage' plafun, m. 'plafond' pyn, m. 'poing' sun, m. 'son, bruit' brun, m. 'noeud dans le bois' trun, m. ' tronc' fun, f. 'fontaine' Members of the (n)-group include all other N stems / J 1 1 with /-n/, e.g. bezi(n), m. 'voisin', bi(n), m. 'vin', busi(n), m. 'morceau', medesi(n), m. 'medecin', awrinu(n), m. 'brugnon', besu(n), m. 'jumeau', ky(n), m. 'gros 171 i 9 morceau1, PY(n), m. 'point'. Loan-words such as sitnasju(n) f. 'situation', abju(n), m. 'avion' tend to join this group; in some of these the loss of /n/ is option al, e.g. /baguns/ or /bagus/ 'wagons' b. Stem-final /n/ is deleted before masc. zero with or without pi. /-s/ in the adj. stems picu[n] 'petiot' and milu[n] 'meilleur', i.e. fem. /miluno/ 'meilleure', masc. /milu/ 'meilleur'. (6) Extended stems Three N stems have alternants augmented with /-t/ be fore pi. /-is/. Symbol -t; thus gus-t*', m. > sg. /gus/ ’gout', pi. /gustis/ 'gouts'; also ares-t~, m. 'arret', pres-t^, m. 'pret'. Note; Due to a special sandhi phenomenon (see 5.1.2.III.A. 1) the two adj. masant ~ masan[t] 'mechant' and grand ~ gran[t] 'grand' atypically have two stem allomorphs in the sg. before masc. zero: one regularly constituted, e.g. ^Occasionally, as is the case here, the loss rather than the retention of /-n/ in the pi. serves to distinguish separate morphemes which have homophonous sg. stems; thus pynx, m. 'poing' vs. py(n)2, m. 'point'. l°In contradistinction, there is no such fluctuation in common words whose configuration is clearly indigenous. 172 /gran/, another with added /-t/, e.g. /grant/. 4.2.2.4. Sub-Classes of Adjectives Adj. stems are subdivided into: I. stems which form adverbs by means of the suffix /-omen/ and II. stems which do not. I. Adj. stems of the first sub-class are of three types: A. base adj. which constitute a large group and are usually descriptive in meaning, e.g. prunt 'prompt' (adv. pruntomen 'promptement1), dusA 'doux' (adv. dusomen 'doucement'); B. adj. stems derived by means of certain suffixes (see 4. 3.2.4), e.g. untuz* 1honteux1 (adv. untuzomen 1honteuse- ment'); C. ordinal adj. to be described with the numerals (4.2.2. 5), e.g. trezjemo* 'troisieme' (adv. trezjemomen 1troisieme- ment1). II. Adj. stems of the second sub-class are also of three types: A. derived adj. stems other than those of types B and C V j | above, e.g. degalant 'gaspilleur1, blagyr 'bavard'? B. adj. whose bases are used as first elements of compound 173 pronouns. These adj. are function-words of the nominal marker group (see 5.2.1.1.A). They are: (1) the distributive adj. kado* 1chaque1 not inflected for number, (2) the quantitative adj. kjkwk® 'quelque' fully inflected: C. adj. whose bases are homophonous with pronoun bases, but are separate morphemes. These adj. are nominal markers. The inflectional allomorphs with the pron. bases are either the same as those which occur with the adj. bases or differ ent allomorphs. These bases are ten in number: (1) the indefinite article yn (unstressed) is an adj. in flected for gender but not for number: /yn/ 'un', /yno/ 'une'. yn (stressed) is also a pron.; it can be pluralized with /-s/ as a bound constituent of pronominal phrases (see 5.2.1.3.(l)b) and as an element of compounds (see 4.4.2.4). (2) the definite article 1^ (unstressed) is built on the a base /l-/. The irregular gender allomorphs are masc. /-u/, fem. /-a/, yielding /lu/ 1 le’, /la/ 'la'. The pi. allomorph is /-s/. These forms are also accusative pron. and bound demonstrative pron.y the latter occur stressed as heads of 174 pronominal phrases (see 5.2.1.3.(8)a.i and b): /in/ 'celui', /la/ 'celle'. (3) the demonstratives akeste and akeL are adj. and free pron. Their inflectional allomorphs are of the regular types indicated. In relation to the speaker, akeste signi fies nearer position, e.g. /akesto/ 'cette ... -ci, celle- ci1, /akestes/ 1ces ... -ci, ceux-ci1; akeL signifies far ther position, e.g. /aket/ 'ce ... -1&, celui-la'. I (4) the interrogative and exclamative kin 'quel, lequel1 fully inflected is an adj. and a free pron. (5) the negative nad is an adj. and a free pron. inflected for gender but not for number: /nat/ 1aucun', /nado/ ' au- cune'. (6) the "indefinites" awte 'autre' and memo* 'meme' are adj. and stressed constituents of pronominal phrases (see 5.2.1.3.(l)c). (7) the cardinals {1} and {2} are inflected like adj. (see below). (8) the inclusive form /tut/ is an adj. and a free pron. The adj. tutx uses regular inflectional allomorphs, e.g. 175 masc. pi. /tuts/ 'tous', fem. pi. /tutos/ 1toutes1. The pron. tut^ differs from the adj. only in that it uses the irregular pi. allomorph /-is/ after the base, i.e. masc. pi. /tutis/ 'tous'. (9). the adverbial adj. tuts 'tout' is identical in meaning with the adv. tut4 'tout', but it is inflected like the in clusive adj. tutx. The adverbial adj. functions as an adj.- modifier (see phrases 5.2.1.2.[1]). (10) the possessive adj. bases are homophonous with the possessive pron. bases and some of them use different in flectional allomorphs. Possessive adj. are like pron. in that they have distinct forms for reference to various per sons and on this account both will be treated below in sec tion 4 .2.3.2 . D. In a sub-class of its own is the quantitative adj. forso* 'bien des, bon nombre de'. This base is non-recur ring and is not consistently inflected for number.^ H-This adj. is pluralized before a N, e.g. /forsoz awkos/ 'bien des oies'; but when no N follows, usage varies as to the presence or absence of pi. inflection, e.g. /n abjo forsos./ ~ /n abjo forso./ 'il en avait bon nombre'. In the latter instance this adj. occurs in a position often filled by quantitative adverbs; cf. /n abjo pla./ 'il en 176 4.2.2.5. Numerals Numerals include base forms, derivates, and compounds, all of which are described in this section for pure con venience . i. Cardinal? Of the cardinal numerals, few are inflected but all belong to an intermediate class of words which can fill the syntactic positions of both adj. and pron. They are stressed words. Numeral {1} is inflected only for gender: masc. /yn/ 'un', fem. /yno/ 1une1; these forms are homophonous with those of the indefinite article. Numeral {2} is fully in flected: masc. /dys/ 'deux', fem. /djos/ 'deux'; /-jo/ is the irregular fem. allomorph which occurs also with posses sive pron. Numeral {100} may be pluralized with /-s/ under certain conditions and so can numeral {1,000,000}. Other cardinals are uninflected. Bases and derivates for teens and tens appear in pho nemic notation in the table on p. 177. The division of forms into morphs is indicated by periods. avait beaucoup'. 1 2 3 4 5 6 7 8 9 10 20 177 Bases Teens Tens yn unz .e dy .s d.jo.s du^.e tres tre^.e tr.ento katre katorz.e kr.anto 1 sink 1 kinz.e sink.anto s j es se/.e swas.into set we j naw d ej6 1 bin sen 1 milo 178 Base morphemes having more than one allomorph include: (1) /yn/ ~ /unz-/ {2} /dy-/ ~ /d-/ ~ /du/-/ {3} /tres/ ~ /tre/-/ ~ /tr-/ {4} /katre/ ~ /katorz-/ ~ /kr-/ {5} /sink/ ~ /kinz-/ {6} /sjes/ ~ /se/-/ ~ /swas-/ The allomorphs of the bases are distributed as shown by the table. There are two suffixes: (teens) /-e/ and (tens) /-ento/ ~ /-anto/. Derivates with /-e/ are stressed on the last vowel of the base. Derivates for tens are stressed on the initial vowel of the suffix. Higher numerals are compounds identified by the pres ence of bound allomorphs like /diz-/ 'dix' which do not occur elsewhere. Compounds are stressed on each numeral component. Numerals 17, 18, and 19 are respectively /diz e set/, /diz e wej/, /diz e naw/, that is, composites of two bases joined by the coordinator j* 'et1 with initial base {10} appearing as /diz-/ for /de//. Compounds from 21 to 69 have the same structure. As first element in compounds before a vowel {20} has the sandhi-form /bint/ for /bin/ (see sandhi 5.1.2.III.A.[2]), e.g. /bint e yn/ 'vingt et un' /bint e 179 dys/ 'vingt-deux' (fem. /bint e djos/). Other tens occur ring as first elements drop /-o/, e.g. /trent e yn/ 'trente et un'. The vigesimal system is used from 61 to 99, so that numerals 70, 80, and 90 are not derivates but compounds, respectively /swasant e dej6/, /katre bin/, /katre bin de//. To the bases /swasanto/ and /katre bin/ are added numerals 1 to 19. sl 'et' is present after the first element of com pounds up to and including 79, absent in that position in compounds above 79, e.g. /swasant e diz e naw/ 1soixante- dix-neuf1, but /katre bin tres/ 'qautre-vingt trois', /sen yn/ 'cent un', /milo de// 'mille dix'. In multiples of 100, the numeral /sen/ 'cent' is pluralized only when it occurs as last element of compounds, e.g. /katre sens/ 1guatre cents', but /katre sen naw/ 'quatre cent neuf'. As initial element of compounds before a consonant, {5} has the sandhi-form /sin/ for /sink/ (see sandhi 5.1.2. II), e.g. /sin sens/ 'cinq cents'. milu(n) 'million' is a masc. noun which occurs preceded by other numerals or by other nominal markers, e.g. /sjez miius/ 'six millions'. 180 II. Ordinals Only widely used ordinals, those corresponding to the first five cardinals, will be examined here. As indicated, for example, by the form /sizjemo/ 'sixieme1 with base allo- morph /siz-/ for native /sjes/, higher ordinals appear to be loan-words adapted from French. Ordinals prym£[r] 'premier' and segund 'second' are fully inflected base adjectives. The next three ordinals show no gender; they are derived from cardinal bases by means of the suffix /-jemo/: trezjemo* 'troisieme' with base allomorph /trez-/ for /tres/, katrjemo* 'quatrieme' with base allomorph /katr-/ for /katre/, and sinkjemo* 'cinqui- eme' . 4.2.3. Pronouns 4.2.3.1. Inflected Pronouns The stem of a pron. is defined as that phonemic stretch which occurs before the fem. inflection or before the neuter inflection. Alternations in the form of the stem are few and will be dealt with as they occur. (1) Pronouns whose bases are homophonous with adjective bases have been presented in 4.2.2.4.II.C above. 181 (2) The free demonstratives aso 'ceci' and ako 'cela', and the bound demonstrative ,s<2. ' ce' (see 5.2.1.3. [8]) show neu ter inflection /-o/. These forms have no plural. The bound base /as-/ normally signifies nearness to the speaker, while /ak-/ signifies remoteness. The base /s-/ is merely demon strative in meaning. (3) The compound distributive kadvn is inflected for gender but not for number: /kadyn/ 'chacun', /kadyno/ 'chacune'. (4) The compounds masc. /kawkezyns/ 1quelques-uns' and fem. /kawkozynos/ 'quelques-unes' have indefinite numerative meaning. They are fully inflected but have no sg. forms. 4.2.3.2. Personal Pronouns Pronouns and adjectives, which have distinct forms for person reference, constitute the paradigm shown on p.182. Not all pron. are inflected. Only gender inflections appear in the table. Stress is marked on all forms except those that never occur stressed. The nominative-prepositional forms are stressed free 12 forms. The possessive pron. are stressed forms either 12These occur after prepositions as axes, e.g. /per zu/ 'pour moi' and either before or after verbs as emphatic Nominative- prepositional Oblique Possessive Pronouns Possessive Adjectives masc fem. masc. 1 sg. 1 pi. 2 sg. fam. 2 sg. def. 2 pi. v | zu masc. fem. 3 et el.o I nuzaw ty bus buz aw me nus te bus Reflexive se Accusative masc. fem. neut 1. u 1. a aw l.o ot Dative m.4w m.jo nist.e nost.o .1 , . ■ t.un t.jo b6st.e b6st.o s.un s.jo m.un nost.e t .un bost.e s .un I i fem. m. a nost.o t .a bost.o s. a oo io 183 13 free or bound. All other forms are phrasally bound. The oblique case forms of the first and the second persons are under-differentiated forms, while the third person shows differentiated forms for the reflexive, accusative, and dative. Of the bound forms, possessive adj. are never stressed; the remaining forms may be stressed or unstressed . . 14 according to syntactic position. The meanings of the nom.-prep. forms are here defined, /zu/ 'moi': the speaker /nuzaw/ 'nous': several persons including the speaker /ty/'toi': one person familiarly addressed /bus/ 'vous': one person deferentially addressed subjects, e.g. /et debalo./ 'lui, il descend1, /marunes ty./ 'tu ronchonnes, toi‘. l^Token examples: /lu sun/ 'le sien', /so de boste/ 'ce qui vous appartient', /ez mew./ 'il est a moi'. The distribution of these morphemes will be stated in 5.2.1.3. (1)a, (7)a.ii and 5.2.1.4.II.(3-4). ■^When enclitic, these forms receive phrase-stress; when proclitic, they are unstressed. The forms /lo/ 'la' and /ot/ 'le', being special enclitic forms, are always stressed. The parallel proclitic forms /la/ 'la' and /aw/ 'le', and the reflexive /se/ are always unstressed. The other forms (oblique forms and the special forms in the da tive and the accusative masc.) occur both in proclitic and enclitic positions and are stressed accordingly. The order and positional variants of pronouns when coupled with other clitic complements will be described in 5.2.1.4.I.A.(2). 184 /buzaw/ 'vous': several addressees without distinction of form of address /et/ 1lui': one male other than speaker or addressee /elo/ 'elle': one female other than speaker or addressee /e/is/ 'eux': several males or a mixed group /elos/ ’elles1: several females Personal pronouns consist of a base followed by a gen der suffix or, in the case of those pron. which are neuter or indifferent to gender, only the base. The base indicates both person and case; thus, a base like /me/ 'me, moi' sig nifies both first person sg. and oblique case. In the case of /bus/ 'vous', which doubles as nom.-prep. and oblique, case is signaled by syntactic position and subsequent stress difference.^ (l) Baig.e.s The distinct base morphemes for the various cases have ^Another type of analysis would recognize case suf fixes as independent units from initial person morphemes, for instance oblique /-e/ in /me/, /te/, and person mor phemes like 1st sg. /zu ~ m-/ and 2nd sg. fam. /t-/. Such an analysis would emphasize the role played by position, but it would demand an extensive use of zero elements and either the establishment of a zero possessive morpheme or the arbitrary division of, say, /nost-/ into a base /nos-/ and an overt possessive allomorph /-t/. 185 the allomorphs: nom.-prep.: 1 sg. /zu/ 1 pi. /nuzaw/ 2 sg. fam. /ty/ 2 sg. def. /bus/ 2 pi. /buzaw/ 3 /el- ~ et ~ e/-/ (see below) oblique: 1 sg. /me/ 1 pi. /nus/ 2 sg. fam. /te/ 2 /bus/ reflexive: 3 /se/ accusative: 3 /l-/ accusative neuter: /aw ~ ot/^^ dative: 3 /i/ (i.e. i^. 'lui') possessive: 1 sg. /m-/ 1 pi. /nost-/ 2 sg. fam. /t-/ 2 /bost-/ l6/ot/ 1le1 is the enclitic form of /aw/ 'le1 (see 5.2. 1.4.A.[2]). 186 3 A -/17 The possessive adj. bases are homophonous with the posses sive pron. bases just listed. (2) Gender inflections The gender morphemes with the pron. bases have allo morphs distributed as follows: {masc.} /-e ~ / ~ -un ~ -ew ~ -u/. The regular allomorphs occur: /~e/ with /nost-/, /bost-/, zero with eL, i.e. with shape /et/. The irregular allomorphs occur: /-un/ with /t-/ and /s-/, /-ew/ with /m-/, and /-u/ with /l-/. {fem.} /-o ~ -jo ~ ~a/. The regular allomorph /-o/ occurs 18 with /nost-/, /bost-/, /el-/, /l-/. The irregular allo morphs occur: /“jo/ with /m-/, /t-/, /s-/ and /-a/ with /I"/.19 The masc. and fem. morphemes have the same shapes and the same distribution when they follow the poss. adj. bases l^It is worthy of note that this dialect does not dis tinguish between sg. and pi. 3rd pers. possessors, hence /sun/ 'sien, leur1, /sjo/ 'sienne, leur'. 1®/lo/ 'la1 is the enclitic form of /la/ 'la'; cf. n. 16 above. ■L^The accusative forms /lu/ ' le' and /la/ 'la' are identical in shape with the def. art. and with bound demon strative pronouns. 187 as they have when they follow the poss. pron. bases, except for {masc.} /-un/, which occurs after the poss. adj. base /m-/ (as well as after /t-/, /s-/), and {fem.} /-a/, which 20 occurs after the poss. adj. bases /m-/, /t-/, /s-/. (3) Number inflection All forms which are inflected for masc. or fem. gender can also be inflected for pi. number. The allomorphs of the suffix {pi.} are: /-s/ and /-is/. The latter occurs only in nom.-prep. 3 after masc. zero, i.e. after the masc. sg. allomorph /et/ 'lui' whose pi. form is /e/is/ 'eux1; there is an irregular alternation in the base, the replace ment of the voiceless dental stop /t/ by the voiced dental affricate ///. The regular allomorph /-s/ occurs elsewhere In both poss. pron. and poss. adj. the presence of /~s/ en tails the loss of stem-final /-n/ as is frequent in N; thus sg. /tun/ 'ton, tien', pi. /tus/ 'tes, tiens'. I Dative ij. 'lui', indifferent to gender, can also be pluralized, but in irregular fashion. Its pi. shape is 20()nly after the possessive adj. bases /m-/, /t-/, /s-/ there occurs as an irregular sandhi alternant a fem. variant /-un/ replacing /-a/ before a vowel-initial fem. word, e.g. /ta pulo/ 'ta poule', but /tun awko/ 'ton oie'. Fem. /-un/ is homophonous with the masc. allomorph /-un/. 188 /luzi/ 'leur', atypically formed by prefixation of /luz-/, an element which is to be counted as one of the irregular allomorphs of {pi.}. The prefix /luz-/ is homophonous with the pre-vocalic form of accusative masc. pi. /lus/ 'les". 4.2.3.3. Uninflected Pronouns These forms are classed as pronouns by the criteria that they substitute for and fill the positions of personal pronouns or nominal phrases (see 5.2.1.1). Except for the relative pronoun, these pronouns are independent, i.e. have no antecedent. They are: (1) the relative Jig. ~ &i. which has the variant forms /ki/ 'qui, lequel' for reference to persons and /ke/ 'que, quoi' for reference to things. These variants occur only in the following positions: after prepositions, after the bound demonstrative 1^ and before the relative clause /ke sjoske/ ’que ce soit’. The single shape /ke/ 'qui, que, lequel' occurs elsewhere and refers to persons as well as things (see 5.2.I.3.[8]b). I I / 1 (2) the interrogative ] £ £ . ~ which has the forms /ki/ 'qui' for persons and /ke/ 'que, quoi' for things. (3) the negative res 'rien' for things and the compound 189 negative diayn 'personne' for persons. (4) the indefinite kawkumet 'quelque chose' for things and the compound indefinite kawkvn 'quelqu’un' for persons. (5) the indefinite personal joel 'on', which functions only as subject and substitutes for any nominative personal pro noun or nominal phrase referring to persons. The 3rd pers. reflexive pron. /se/ occurs with /on/ in pronominal verb phrases (see 5.2.1.4.I.A.[l]). 4.2.4. Verbs 4.2.4.1. Morphophonemic Alter nations In this section are presented basic alternations in inflected verb forms and alternations in verb stems which are comparable to basic alternations. Irregular alterna tions which are infrequent and for which no symbol is pro vided will be listed exhaustively in the inventory of irreg ular verbs (4.2.4.3.Ill). I. Phonologically conditioned alternations 190 A. Consonants (1) Reduction of C clusters Deletion of the last of two C takes place in verb-final position when the stem is followed by covert inflection, i.e. in 3 pres, and in 2 imper. Type djerbe > /djerbe/ 1ouvrir1 vs. /djer/ 'il ouvre', /djer/ 'ouvre'; e.g. parti 'partir', respunde 'repondre', I mOrde 'mordre'. (2) Replacement of /b d g z/ a. Regularly by unvoicing Type sabe > /sabe/ 'savoir' vs. /sap/ 'il sait'; e.g. resGlje 1 1 ^ 1 'recevoir' , PY^i 1puer', past part, bazyd 'eclos' (of baze + \ 1 1 'eclore'), sjEge 'suxvre', impf. stem diz- (of dire 'dire'). o 0 b. Irregularly by palatalization of /z/ to /j/. Symbol Z. Type beZe > /beze/ 'voir' vs. /bej/ 'il voit'. Also of this type are: kreZe 'croire', kuZe 'coudre', and plaZe 'plaire'. (3) Replacement of /I n/ Type byLi > /byli/ 'bouillir' vs. /byl/ 'il bout'; e.g. 191 plaNe 'plaindre'. (4) /w ~ b/ Type eskriwre > /eskriwre/ 'ecrire', /eskriw/ 'il ecrit' vs. /eskribes/ ' tu ecris', /eskribyt/ 'ecrit'; e.g. plGWre ■ I 'pleuvoir', diWre 'devoir', klaWre 'contenir'. (5) /m ~ n/ The stem-final bilabial nasal /m/ is dentalized to /n/ word-finally. Symbol M; e.g. drUMi > /drumi/ ' dormir' vs. /dron/ 'il dort'. This alternation is automatic insofar as /-m/ occurs word-finally only in the N actram. m. 'chiendent'. (6) /r ~ r/ The multiple trill /r/ which cannot occur word-finally is automatically replaced by the single trill /r/ in the verb kure > /kure/ 'courir' vs. /kur/ 'il court'. B. Vowels (1) Loss of a vowel The first of two adjacent vowels which is canceled may be inflection-internal, e.g. /kant-/ + /e/ + /_i/ > /kanti/ ' je chante', or stem-final, e.g. /prenge-/ + /-ijoj/ > /prengijoj/ 'je prendrais'. 192 (2) /e ~ e/ Type nGse > /nese/ 'naitre’ vs. /nesyt/ 'ne'; e.g. asCte 'asseoir', plSWre 'pleuvoir', raGte 'mettre', pise 'paitre'. (3) Palatalization of /i/ After a stressed vowel /i/ is automatically palatalized to /j/. No symbol; e.g. /embalere-/ + /i/ > /embalerej/ 'j1avalerai'; /remydijo-/ + /i/ > /remydijoj/ 'je remuerais'. II. Morphologically conditioned alternations A. Consonants These alternations are few and sporadic (see 4.2.4.3. Ill) . B. VoweIs (1) , /e ~ e/ This alternation occurs in the verb Gste > /este/ 'etre' vs. /estat/ 1ete'. (2) /o ~ u/ Type mOle > /mole/ 1moudre' vs. /mulyt/ 'moulu', /mulut/ ■ • *il moulut'; e.g. tOse 'tordre', mOrde 'mordre'. 193 (3) /e ~ e/ Type abutEla > /abutela/ 'botteler1 vs. /abutelo/ 1il bot- telle'. Also of that type are the following verbs and their compounds: agrymEla 'amonceler1 anEla 'mettre bas (brebis)' apEla 'appeler1 arEsta 'arreter' atEla 1atteler' betEla 'veler' brEga 'recurer' brEsa 'bercer1 bulEga 'remuer' defEla 'effeuiller' derebEla 'reveiller1 dEsa 'laisser' drEsa 'dresser' embEstja 'embeter' I ^ embwEta 'emboiter' empEwta 'ajouter' endEwta 'endetter' enkjEta 'inquieter' entEra 'enterrer' espEra 'esperer' grElla 'greler1 grEsa 'graisser' kapEla 'couvrir' krEba 'crever' lEba 'lever’ lEka 'lecher' myzEla 'museler' nEba 'neiger' nEga 1noyer' pEza 'peser' plEga 'plier' prEga 'prier' prEsa 'presser' prEsta 'preter' prusEda 'proceder' rastEla 'rateler' resEga 'scier' sEga 'faucher' 194 sErka 'chercher' trabEsa 'traverser' zEjna 1gener' zEli 'geler' zEta 'jeter1 (4) /u ~ o/ In certain stems /u/ is lowered to /o/ under stress. Symbol U. Type adUra > /adura/ 'adorer' vs. /adoro/ 'il adore1; but cf. pusa > /pusa/ 'pousser', /puso/ 'il pousse'. Also of that type are loan-verbs like pilUta 'piloter' and the following verbs and their compounds: akrUsa 'accrocher' apatUka 'mettre en tas' aprUba 'approuver' aprUsa 'approcher' bartUla 'parler a tort et a travers’ birUla 1tournoyer' bUla 'voler' bUle Irr. 1vouloir' bUta 'voter' damUra 'rester' deskUrna 'ecorner' I drUMi 'dormir' ebapUra 'evaporer' s embrUka 'se planter une epine' embU j a 'envoyer' embUmja 'enj oler' emp1U j a 'employer' empUna 'empoigner'^ erapUza 'imposer' estUfa 'etouffer' fagUta 'mal habiller' galUpa 'galoper' gaspUta 'grappillonner' gUlsa 'ailler' 195 kalUta 'gifler' kanUta 'mettre bas (chienne)' kUfa 'coiffer' kUla 'coller' kumpUza 1 composer' macUrla 'mal macher' I mUr i Irr . ' raour ir ' penjUrla 'ne pas suspendre d'aplomb1 prUba 'prouver' pUde Irr. 'pouvoir' pUrta 'porter' rebUlta 'revoIter' rekUlta 'recolter' reprUsa 'reprocher' rUlla 'rouler' rUta 'roter' I sUrti 'sortir' trambUla 'trembler sur ses jambes' trantUla 'trainer' tridUla 'frissonner' trUba 1trouver' tUka 'toucher 1 tUrna 1revenir' zUga 'jouer' 4.2.4.2. Verb Forms A verb form consists of a stem plus inflectional end ing. Paradigms and analysis of the sample regular verbs I J ! I ajma 'aimer', parti 'partir', menti 'mentir', and perde 'perdre' are shown in phonemic notation on pages 196-199. The constitution of the inflectional margin distin guishes finite from non-finite verbs. In the finite system the margin is a coalescent suffix, i.e. an inflectional se quence made up of two elements: a medial and a final, neither one of which normally occurs without the other. 196 Class I ajma 'aimer' Pres. part, ajma.n Past part, ajma.t Future Conditional 1 ajme.re.j 1 ajm.ijo.j 2 ajme.ra.s 2 ajm.ijo.s 3 ajme.ra. 3 ajm.ijo. 4 ajme.re.n 4 ajm.ijo.n 5 a jme .re ./ 5 ajm.ijo./ 6 ajme.ra.n 6 ajm.ijo.n Present Present S. 1 ajm. .i 1 ajm. .i 2 ajm.e.s 2 ajm.e.s 3 ajm.o. 3 ajm.e. 4 ajm.a.n 4 ajm.e.n 5 ajm.a./ 5 ajm.e./ 6 ajm.o.n 6 ajm.e.n Imperfect Imperative 1 ajm.ab .i 2 ajm.o. 2 ajm.abe.s 4 ajm.a.n 3 ajm.abo. 5 ajm.a./ 4 ajm.abe.n 5 ajm.abe ./ 6 ajm.abo.n Perfect Imperfect S. 1 ajm.er .i 1 ajm.es ,i 2 ajm.ere.s 2 ajm.ese.s 3 ajm.e .t 3 ajm.fese. 4 ajm.fere.n 4 ajm.fese.n 5 ajm.ere./ 5 ajm.fese./ 6 ajm.fero.n 6 ajm.ese.n 197 Class Ila parti 'partir1 I I Pres. part, parti.n Past part, parti.t Future Conditional 1 parti.re.j 1 part.ijo.j 2 parti.ra.s 2 part.ijo.s 3 parti.ra. 3 part.ijo. 4 parti.re.n 4 part.ijo.n 5 parti .re ./ 5 part.ijo./ 6 parti.ra.n 6 part.ijo.n Present Present S. 1 part. .i part.e.s 1 part. .i 2 2 pirt.e.s 3 par . 3 part.e. 4 part.e.n 4 part.e.n 5 part.e./ 5 pirt.e./ 6 p4rt.o.n 6 part.e.n Imperfect Imperative 1 part.jo.j 2 par . 2 part.jo.s 4 part.e,n 3 part.jo. 5 part.e./ 4 part.jo.n 5 part.j&./ 6 part.jo.n Perfect Imper feet S. 1 partisk.ur .i 1 partisk.us .i 2 partisk.ure.s 2 partisk.use.s 3 partisk.u .t 3 partisk.use. 4 partisk.ure .n 4 partisk.use.n 5 partisk.ure./ 5 partisk.use./ 6 partisk.uro.n 6 partisk.use.n 198 Class lib menti 'mentir' Pres. part, menti.n Past part, menti.t 1 2 3 4 5 6 Future menti.re.j menti. ra.s menti.ra. menti.re.n menti .re ./ menti.ra.n Conditional 1 ment.ijo.j 2 ment.ijo.s 3 ment.ijo. 4 ment.ijo.n 5 ment.ijo./ 6 ment.ijo.n Present Imperative 1 mentis, .i 1 2 mentis. 2 mentis.e.s 4 mentis.e.n 3 mentis. 5 mentis.e./ 4 mentis.e.n 5 mentis.e ./ mentis.o.n Imperfect 6 1 mentis.jo.j 2 mentis.jo.s mentis.jo. mentis.jo.n mentis. jo./ mentis.jo.n 3 4 5 6 Perfect Present S Imperfect S 1 mentisk.ur .i 2 mentisk.ure.s 3 mentisk.u .t 4 mentisk.ure.n 5 mentisk . ure ./ 6 mentisk.uro.n 1 2 3 4 5 6 mentxsk mentisk. mentisk mentisk, mentisk, mentisk, , i , s ,n n 1 mentisk.us .i 2 mentisk.use.s 3 mentisk.use. 4 mentisk.use.n 5 mentisk.use./ 6 mentisk.use.n 199 Class III perde 'perdre' Pres. part, perde.n Past part, perdy.t Future Conditional 1 perde.re.j 1 perd.ijo.j 2 perde.ra.s 2 perd.ijo.s 3 perde.ra. 3 perd.ijo. 4 perde.re.n 4 perd.ijo.n 5 perde.re./ 5 perd.ijA./ 6 perde.ra.n 6 perd.ijo.n Present Present S. 1 perd. .i 1 perd. .i 2 perd.e.s 2 perd.e.s 3 per . 3 perd.e. 4 perd. e .n 4 p&rd.e.n 5 perd.e ./ 5 perd.e./ 6 perd.o.n 6 pferd.e.n Imperfect Imperative 1 perd.jo.j 2 per . 2 perd.jo.s 4 perd.e.n 3 perd.jo. 5 perd.e./ 4 perd.jo.n 5 perd. jo./ 6 perd.jo.n Perfeet Imperfect S. 1 perd.ur .i 1 perd.us .i 2 perd.ure.s 2 perd.use.s 3 perd.u .t 3 perd.use. 4 perd.ure.n 4 perd.use.n 5 perd.ure./ 5 perd.use./ 6 perd.uro.n 6 perd.use.n 200 The medial suffix, which follows the stem, is a tense-mode- aspect indicator, referred to as "tense-marker." There are ten morphemes in this category listed below with their traditional names. The final suffix is an actor-number indicator, referred to as "person-marker." There are six morphemes in this category listed below with numbers 1-6. Finite tenses have all six forms with the exception of the imperative, which has only 2, 4, and 5. In the non-finite system person-raarkers are lacking. The presence of tense- markers for the present participle and past participle sig nals the contrast between these tenses and the infinitive. I. Stress There are four distinct stress positions defined as fo1lows: The stress pattern of the present participle is iden tical with that of the infinitive; that is, the stress strikes the penultimate vowel of the stem in class III verbs, the final stem-vowel elsewhere. The stress strikes the final vowel of the stem in the past participle, present subjunctive, 1, 2, 3, 6 present, and 2 imperative. The stress strikes the second vowel of the inflectional 201 margin in the conditional, but the initial vowel of the in flectional margin in the remaining verb forms. II. Stems There are three basic stem alternants: the infinitive (inf.) stem, the imperfect (impf.) stem, and the perfect (perf.) stem. In regular verbs the last two are derivable from the inf., which is taken as the primary allomorph. Tenses are grouped in the tables according to the stems in- volved. A. The inf. stem The inf. normally serves as stem for present partici ples, conditionals, futures of Cl. II, III, past participles of Cl. I, II, with replacement of the final vowel by /e/ for futures of Cl. I, and with replacement of the final vowel by /y/ for past participles of Cl. III. B. The impf. stem This is the phonemic stretch which appears throughout the impf. before the sequence /-ab-/ of verbs of Cl. I or before the sequence /-jo-/ of verbs of Cl. II, III. In regular verbs of Cl. lib this stem is equal to the inf. augmented with /-s/. In other regular verbs it is equal 202 to the inf. minus the last vowel. This stem normally recurs in presents, imperatives, and, save for Cl. lib, in present subjunctives. C . The perf. stem This is the phonemic stretch which appears throughout the perf. before the sequence /-er-/ of verbs of Cl. I or before the sequence /-ur-/ of verbs of Cl. II, III. In regular verbs of Cl. I, III this stem is equal to the inf. minus the last vowel. In regular verbs of Cl. II it is equal to the inf. augmented with /-sk/. On this stem are normally formed impf. subjunctives and Cl. lib present subjunctives. III. Tense-markers21 Tense-markers carry distinctions of time, mode, and aspect only loosely suggested by their traditional labels. The meanings of the tense-markers differ in no vital sense from their counterparts in standard French. Allomorphs of the tense-markers are listed below in the order of presentation on the first table. Most tense p 1 All are widely used, including the perfect indicative and the imperfect subjunctive which have fallen into disuse in spoken French. 203 morphemes have alternanats in complementation relative to verb classes and/or person-markers. The conditional mor pheme alone has phonologically determined allomorphs. {pres, part.} /-n/ {past part.} -£-. The combination stem + -d- yields an adj.-like verb form which may be inflected further for gen der and number? thus, masc. /ajmat/ 1aime‘, fern, /ajmado/ 1aimee'; both forms may take the pluralizer /-s/, e.g. /ajmats/ 1aimes'. {fut.} 1 /-re/? 2, 3, 6 /-ra/; 4, 5 /-re/ {cond.} /-ijo/ conditioned to /-jo/ after stem-final /-r/ or /-w/ {pres.} in Cl. I: 1, 2 /-e/; 3, 6 /-o/? 4, 5 /-a/; in other classes: 1, 2 /-e/; 6 /-o/; 4, 5 /_e/; 3 / {impf.} in Cl. I: 3, 6 /-abo/; in other persons /-abe/; in other classes: /“jo/ {pres. S.} /-e/ {imper.} in Cl. I: 2 /-o/; 4, 5 /-a/; in other classes: 2 0; 4, 5 /-e/ {perf.} in Cl. I: 3 /-e/; 6 /-ero/; in other persons /-ere/; in other classes initial /e-/ is replaced by /u-/ {impf. S.] in Cl. I: /-ese/; in other classes initial /e-/ is replaced by /u-/. 2 04 IV. Person-markers The meanings of person-markers correlate with those of nominative (emphatic) personal pronouns (see 4.2.3.2 above). Each person-raarker, symbolized by a number, is accompanied below by its traditional name and the corresponding nomina tive pronoun(s). 1: 1st pers . sg.: /zu/ 1moi1 2: 2nd pers. sg.: /ty/ ' toi* 3: 3rd pers . sg.: /et/ 1lui1, /elo/ 'elle1 4: 1st pers. pi.: /nuzaw/ 'nous' 5: 2nd pers . pi.: /bus/ 1vous1, /buzaw/ 1vous1 6: 3rd pers . pi.: /e/is/ 'eux', /elos/ 'elles' Allomorphs of the person-markers are: {1} /“i/ conditioned to /-j/ after stressed vowel {2} /~s/ ~ / (in imper. only) {3} $ ~ /-t/ (in perf. only) 14} /-n/ is} /-// {6} /-n/ V. Verb forms in prohibitions In contradistinction to affirmative imper. forms, the phrasally bound verb forms which are used in negative commands, i.e. followed by negative adverbs, are for all verbs structurally identical with the present S. Allomorphs of the person-markers for the three existing forms are therefore: 2 /-s/, 4 /-n/, 5 /-//. Also, the tense-marker /-e/ which appears before 2 /-s/ is formally the same as that of the pres. S., e.g. /eskutes pas./ 'n1ecoute pas'. The tense-marker which appears before 4 /-n/ and 5 /-//, however, may be optionally /-e/ or /-e/, whatever the verb class, e.g. /eskutem pas./ ~ /eskutem pas./ ’n'ecoutons , 2 2 pas . Two different interpretations present themselves, de pending upon which of the alternate shapes of the tense- marker in 4, 5 is taken as the norm. If /-e/ is chosen, then the three verb forms at issue are formally distinct from the corresponding forms of the pres. S. This would lead to recognizing the existence of an additional tense mode, call it imper. B., alongside imper. A. The tense- marker for imper. B. would have the allomorphs: 2 /-e/; 4, 5 /-e/, with /-e/ as a non-significant variant of /-e/. 22>rhe form /e/ is probably by analogy to the tense- marker /e/ in persons 4, 5 affirmative imperative of verbs of all classes except I. Although class I is the largest, most commonly used verbs belong to class III. 206 The preferred alternative is to take /-e/ as basic and to note its fluctuation with /-e/ in 4, 5, so that persons 2, 4, 5 pres. S. recur in combinations with negative adverbs as bound constituents of prohibitive phrases (see 5.2.1.4. IV.[l]b). A few illustrations will show the formal contrast be tween verb forms in positive commands and verb forms in negative commands: /tryko./ 'frappe', /trykes pis./ 'ne frappe pas'; /pagan./ 'payons', /pag|m plys./ 'ne payons plus'; /fenis./ 'finis', /feniskes pas./ 'ne finis pas'; /tja/./ 'tuez', /tjoskf/ zamej./ 'ne tuez jamais'; /bajzi./ 'vas-y', /j anges pas./ 'n'y vas pas'. 4.2.4.3. Classification of Verbs Verb classes are established on the basis of the final vowel of the inf. This criterion yields three classes of i i ! ! verbs: Cl. I -a, type ajma, Cl. II -i, types parti and i i 1 menti, Cl. Ill -e, type perde. -i verbs are assigned to I one of two sub-classes, namely Ila, type parti and lib, I type menti. Distinctive features of the latter are, as previously stated, the form with /-s/ of the impf. stem and the recurrence of the perf. stem in pres. S. Cl. I comprises most of the verbs of this dialect. It 207 is a productive class which absorbs Fr. loans in -er. i.e. Fr. /-e/ yields Dzcs. /-a/. Cl. lib is smaller, but is ex panding by integration of Fr. verbs in -ir. Cl. Ill is a closed class which is comparable in size to Cl. lib. Cl. Ila, the smallest, is also a closed class containing the verbs: v | * byLi 'bouillir1 PY^1 1puer' I l drUMi 'dormir1 senti 'sentir1 > I endrUMi 1endormir1 serbi 'servir1 l l mUri Irr. 'mourir' sUrti 1sortir1 I parti 'partir' I. Regular and semi-regular verbs Only those verbs are regular which follow the patterns of Cl. I, Ila, lib, or III and exhibit only phonologically conditioned alternations. The conjugation system of a regular verb can be represented simply by listing the inf. with appropriate morphophonemic symbols. The final vowel identifies the verb class; thus, akusa 'poursuivre1 belongs I I to Cl. I, m£te 'mettre' to Cl. III. For -i verbs, sub-class lib can be distinguished from Ila by noting the character istic /-s/ of the impf. stem; thus fyzis 1fuir'. Verbs whose stems exhibit the morphologically 208 conditioned alternations stated above in 4.2.4.1 are con sidered semi-regular; such is the case of, say, bulEga 'remuer1. Verbs of this sort will re-appear in the inven tory of irregular verbs only if they present some other peculiarity. All other verbs (marked "Irr.") are irregular, a total of 30, some of which have similarly inflected compounds. II. Types of irregularities There are four types of irregularities. Most wide spread is the irregular form of various stems. In second place are irregularities in inflection, which occur primar ily in past part, (see below), sporadically in other tenses. The next type is the anomalous distribution of the perf. stem; in addition to its normal recurrence in impf. S., this stem may also occur atypically in pres. S. and more rarely in past part, or imper. Least common are irregulari ties of stress patterns. Tenses showing partly or totally suppletive stems are enumerated here in decreasing order of frequency: fut. and cond., pres. S., past part., impf., perf., pres., imper., pres. part. All impf. S. have regularly constituted stems. The past part, inflection d has the irregular alter- 209 nants: z, t, (t). z subsumes the phonemic shape /z/ before O o fern, /-o/ and unvoiced /-s/ in the absence of /-o/, e.g. prez > fem. /prezo/ 'prise', masc. /pres/ 'pris1. t stands 0 for non-alternating /t/, e.g. dit > fem. /dito/ 'dite1, masc. /dit/ 'dit'. (t) indicates the loss of /-t/ when an unpermitted consonant cluster would result in word-final 1 / * / / 1 position, e.g. bis(t) > fem. /bisto/ 1vue1, masc. /bis/ 1vu'; fej(t) > fem. /fejto/ 'faite', masc. /fej/ 'fait1. Ill. Irregular verbs In this section are presented an analysis and an ex haustive listing of the irregular verb forms with illustra tions. The verbs are arranged by classes. Within each class are included the groups of verbs which share the same irregular features. Additional individual peculiarities of given members of each group are also described. All forms not listed follow the regular patterns of the respective classes. Unless specified otherwise, the listing of an anomalous fut. stem implies an identical cond. stem. A. Irregular verbs Cl. I (1) Short stem verbs ana 'aller' and Xa. 'faire' have in 1, 2, 3, 6 pres. 210 the irregular forms: 1 /baw/ 'je vais' /faw/ 'je fais1 2 /bas/ 'tu vas1 /fas/ 1tu fais' 3 /baj/ 'il va' /faj/ 'il fait' 6 /ban/ 'ils vont'^ /fan/ ' ils font' That is, these verbs have respectively the anomalous stems ka-, .fa- in 1, 2, 6 pres, and bai. fai in 3 pres., the lat ter two recurring in 2 imper. In 1, 2, 6 pres., person- markers are added directly to the stems, so that in the ab sence of tense-markers these stems carry the additional meaning "present." The person-marker /-w/ in 1 pres, is a suppletive variant of /-i/. In addition: the fut. stem of ana is anae-. e.g. /angerej/ 'j'irai', /angijo/ 'il irait1. To the perf. stem ana- are added inflectional suffixes of Cl. II, III, e.g. /angures/ 'tu alias', /angusi/ 'j'allasse'. The latter stem recurs as pres. S. stem, e.g. /ange/ 'il aille'. Atypically is inflected as if it were a member of Cl. II or III in all tenses except as noted above in pres, and in the regular pres. part, /fan/ 'faisant', e.g. /fazen/ ^3A11 informants except DP have the regular form /anan/ 'nous allons' for 4 pres. Instead, DP uses /ban/ by analogy to the irregular form for 6 pres. 211 'nous faisons', /fazjoj/ 'je faisais', /faskut/ 'il fit'. The inf. .£& serves as fut. stem, e.g. /faren/ 'nous ferons’ The cond. stem is also fa-. but this stem being monovocalic hiatus is avoided irregularly by deletion of the initial vowel /i/ of the tense-marker, e.g. /fajos/ 'tu ferais1. The impf. stem is faz- and the perf. stem fask-. the latter recurring as pres. S. stem, e.g. /faskuron/ 'ils firent', /faskusi/ 'je fisse', /fasken/ 'nous fassions'. The past part, fej(t) is wholly irregular. (2) tia 'tuer' Due to the absence of a stem vowel in the impf. stem, the stress automatically strikes the tense-marker in all forms built on this stem, e.g. /tjo/ 'il tue, tue', /tjon/ ’ils tuent'. In 1, 2 pres, the tense-marker /-e/ has the unique alternant /-o/, thus /tjoj/ 'je tue’, /tjos/ 'tu tues'. The pres. S. stem is tiosk-. e.g. /tjoske/ 'il tue' The cond. stem is tia-. but this stem being monovocalic, hiatus is avoided irregularly (cf. fa 'faire' above) by deletion of the initial vowel /i/ of the tense-marker, e.g. /tjajoj/ 'je tuerais'. (3) esta ~ Sste 'etre' Paradigms and analysis of this highly irregular verb 212 are shown on p. 213. Forms separated by the symbol ~ are optional variants representing alternate conjugation sys tems; that is, speakers who make use of inf. esta have regular Cl. I inflections in pres, part., perf., and impf. I S.; conversely, users of inf. Sste have regular Cl. Ill in flections in these three tenses. The past part, stem esta- is irregular only for users I of inf. Sste. The fut. stem is se-. There is a special non-recurring impf. stem er- to which are added the tense suffixes minus the initial part /ab-/. Irregularly this stem is stressed. In the pres. the stem has the suppletive allomorphs es- for 3, su- for 6, se- elsewhere. Irregularly the stress strikes the stem vowel in 4, 5 pres. Throughout, the pres. person-markers are added directly to the stem, so that in the absence of tense-markers the stem carries the additional 24 meaning "present." There is an irregular pres. S. stem siosk-. which re curs in the imper. The inflection in 2 imper. is atypically ^An alternate interpretation would regard the vowel in /sen/, /se// as a tense-marker of Cl. III. The form /es/ would then show the regular covert tense inflection of 3 pres. The remaining three forms, though, would still lack tense inflection. 213 I 1 esta ~ £ste 'etre' Pres, part.: esti.n ~ este.n Past part.: esta.t Future Conditional 1 se.re . j 1 se.jo.j 2 se.ra.s 2 se.jo.s 3 se.ra. 3 se.jo. 4 se.re.n 4 se.jo.n 5 se.re 5 se.jo 6 se.ra.n 6 se.jo .n Present Present S. 1 sc . j 1 sjosk. .i 2 se.s 2 sjosk.e.s 3 es . 3 sj&sk.e. 4 se.n 4 sjosk.e.n 5 sb 5 sjosk.e./ 6 su. n 6 sjosk.e,n Imperfect Imperative 1 er . . i 2 sjosk.e ,s 2 er.e.s 4 sjosk. e .n 3 tr .o. 1 5 sjosk. e ./ 4 er . e ,n 5 fr.e.ff 6 er.o ,n Perfect Imperfect S 1 est.er .i ~ est.ur .i 1 est.es .i ~ 2 est.ere.s 2 est.ese.s 3 est. e .t 3 est.ese. 4 est.ere.n 4 est.ese.n 5 est.ere 5 est.ese./ 6 est.ero.n 6 est.ese.n 214 equal to that of 2 pres. S. The tense-marker in 4, 5 imper. has the regular allomorph of Cl. Ill verbs. b . itTegulay veyfrs qj, iia i mUri 'mourir1 has irregular inflection in the past part, mor(t) 'mort'. C. Irregular verbs Cl. lib sufris 1souffrir' has the wholly irregular past part. sufer(t) 'souffert'. D. Irregular verbs Cl. Ill (1) -e verbs These verbs have the regular inflections of Cl. Ill -e verbs, but the last vowel of the inf. being stressed, the stress remains on that vowel in the pres, part., e.g. /bule/ 'vouloir', /bulen/ 1voulant1. Members of the group are: bUle 'vouloir', pUde 'pouvoir1, bale 'valoir', kale 'fal- O loir 1, and sabe 'savoir' . O Special stems in fut. and pres. S. are respectively: . py<3S,,. S. examples bud- bel- /budran/ 'ils voudront', /b£li/ 1je veuille' pud- pesk- /pudras/ 1tu pourras', /peske/ ’il puisse' bad- bal- /badra/ 'il vaudra', /balen/ 'ils vaillent' 215 fut. pres. S. examples kad- kal- /kadra/ 'il faudra', /kale/ 'il faille’ saw- saj- /sawrej/ 1je saurai', /sajes/ 'tu saches' In addition: pUde may have the cond. stems pudr- or pu-; that is, forms like /pudrjoj/ ~ /pujoj/ 'je pourrais' are free variants. The perf. stem and the past part, stem are both pusk-. e.g. /puskut/ 'il put', /puskyt/ 'pu'. kale is impersonal, i.e. takes no person-marker other than 3. The cond. stem is kadr- with unstable /r/, hence /kadrjo/ ~ /kadjo/ 'il faudrait'. bale and kale have stem- final /w/ for /l/ in 3 pres., i.e. /baw/ 'il vaut', /kaw/ 'il faut’. sabe has two perf. stems in free variation, namely the 2__ regular sab- and the irregular sai-. e.g. /saburi/ ~ /sa^uri/ 'je sus'. The latter recurs in the impf. S., e.g. /sajusi/ 'je susse'. 25 The verbs of this group lack the imper. except for sabe whose imper.: /sajes/ 'sache', /saj&n/ 'sachons', /saje// ’sachez' is modelled on that of abe 'avoir', which see below. 25xnformants declined to furnish forms for this tense on the ground that such forms are not used. 216 (2) -re verbs, type dire 'dire' I __ | dire 'dire' and rire 'rire' have respectively: impf. I I stems diz-, riz-, e.g. /dizes/ 'tu dis', /dis/ 'il dit', /rizjo/ 'il riait' perf. and pres. S. stems dig.-, rig-. e.g. /diguron/ 'ils dirent', /rigusi/ 'je risse', /digen/ 'nous disions' fut. stems reduced to di- ri-. e.g. /diran/ 'ils diront', /rijo/ 'il rirait'. I In addition: the perf. stem of dire recurs throughout the imper., e.g. /digen/ 'disons'. 2 imper. /digo/ 'dis' atypically takes the Cl. I tense-marker /-o/. The imper. of rire is regular, e.g. /ris/ 'ris', /rize^/ 'riez'. The | I past part, dit 'dit' of dire is wholly irregular. The past | part, of rire is built on the perf. stem and has only the masc. form /rigyt/ 'ri'. (3) -Wre verbs r tvoe beWre 'boire' The /-r-/ of the inf. beWre is missing in all forms other than the regular pres. part, /bewren/ 'buvant'. The fut. stem is reduced to bew-. e.g. /bewren/ 'nous boirons', /bewjo/ 'il boirait'. The regular impf. tense-marker /-jo/ is added to a stem beb-, the stem-final /b/ replacing the expected /-w/. This irregular alternation alone distin- 217 guishes the impf. from the cond., e.g. /bebjon/ 'ils bu- vaient', /bewjon/ 'ils boiraient'. Other tenses built on the impf. stem exhibit the regular alternation /w ~ b/ sub sumed under W, e.g. /bebi/ 'je bois', /bew/ 'il boit’. Of this type are: eskriWre 'ecrire1, klaWre 'conte- I | nir', the impersonal plGWre 'pleuvoir1, diWrex 'devoir' (i.e. English 'owe') and diWre2 'devoir' (i.e. English » 'must'). The latter differs from diWrei only in having the irregular pres. part, /diben/ 'devant' built on the impf. I stem. diWreo lacks the imper. I The verb biWre 1vivre' is also of this type, but it has a special stem bisk- for perf. and past part., hence /biskut/ 'il vecut', /biskyt/ 'vecu'. (4) -ge verbs, type tenae 'tenir' Irregularities are limited to the form ten- (instead of the expected teng-) of the impf. stem, e.g. /tenon/ 'ils tiennent', /ten/ 'tiens', and the form teng- of the pres. S stem which is equal to the perf. stem, e.g. /tenguren/ 'nous tinmes', /tengusi/ 'je tinsse', /tenge/ 'il tienne'. The inf. regularly yields /tengen/ 'tenant', /tengyt/ 1tenu', /tengerej/ 'je tiendrai', /tengijoj/ 1je tiendrais1 Of this type are the compounds of tenge and the verbs 218 benae 'venir', prenae 'prendre*, and their compounds. In addition: benae has the irregular 2 imper. /beno/ 'viens' with substitute stem vowel /e/ for /e/ and the Cl. I tense-marker /-o/. Compounds of benae have regular 2 im per., e.g. /suben te/ 1souviens-toi*. The past part, preg 'pris' of prenae is wholly irregular; this irregularity also applies to the compounds: aprenge 'apprendre', kumprenge * comprendre', syrprenge * surprendre1 . (5) -Ze verbs, type kreZe. 1 c r . i O . i f f e ' kreZe *croire' and beZe 'voir* have respectively fut. stems krei-. bei-, e.g. /krejras/ *tu croiras1, and cond. stems krer-. ber- with unstable /-r/, e.g. /berjoj/ or /bejoj/ 'je verrais'. In addition beZe has: perf. stem and pres. S. stem bisk-. e.g. /biskut/ *il vit', /biskusi/ *je visse', /bisken/ 'nous voyions', and a highly irregular past part. bis(t) * vu' . (6) ale ~ abe 'avoir' Paradigms and analysis of this highly irregular verb are shown on p. 219. Forms separated by the symbol ~ are optional variants representing alternate conjugation sys- , v terns; that is, speakers who make use of inf. aie preserve 219 aje ~ abe 'avoir' I V I v I I Pres, part.: aje.n ~ abe.n Past part.: ajy.t ~ aby.t Future Conditional 1 aw. r & . j 1 aw.jo.j 2 aw.ra.s 2 aw. j o. s 3 aw.ra. 3 aw.jo. 4 aw.re.n 4 aw.jo.n 5 aw.re ./ 5 aw.jo./ 6 aw.ra .n 6 aw.jo.n Present Present S. 1 e . j 1 1 ^ eg. .i 2 a. s 2 a j . e . s 3 1 a . 3 1 X a;j .e. 4 ab. e . n 4 1 V a^.e .n 5 ab.e./ 5 ag .e./ 6 a .n 6 aj.e .n Imperfect Imperative 1 ab. j o. j 2 1 X a;j .e .s 2 ab. j o. s 4 an.£ .n 3 ab.jo. 5 an . s ./ 4 ab.jo,n 5 ab.jo./ 6 ab. j o. n Perfect Imperfect S. 1 an.ur .i ~ ab.ur . i 1 aj .us . i ~ ab 2 x i aj.ure.s 2 X 1 an.use.s 3 M 1 , a] .u .t 3 an.use. 4 an.ure.n 4 an.use,n 5 X * 1 an.ure.p 5 X 1 / aj.use./ 6 an.dro.n 6 an.use .n 220 initial /aj-/ in pres, part., past part., perf., and impf. S.; other speakers use inf. abe and have initial /ab-/ in these four tenses. The form ab- of the perf. stem and of the impf. stem, which recurs in 4, 5 pres., is regular Cl. Ill for users of inf. abe, irregular for users of inf. aie. Irregular stems I include: fut. stem aw-. a suppletive stem in 1 pres., a reduced stem a- in 2, 3, 6 pres., and a stem an- in all other tenses. Inflections are regular as in Cl. Ill with two excep tions. Tense-markers are lacking in 1, 2, 6 pres., so that an anomalous stem like e_- also means "present." 2 imper. is atypically equal to 2 pres. S. as to stem and inflection. (7) Isolated irregularities I While the past part, of m£te 'mettre' is regularly , I metyd 'mis', the compound prum£te 'promettre' has the irreg ular past part, prumez 'promis'. The perf. stem d~ierbisk- of dierbe 'ouvrir' is consti tuted as in Cl. II, e.g. /djerbiskuri/ 'j'ouvris'. Besides the regular past part, djerbyd 1ouvert1 which has active meaning, there is an irregular form dvber(t) 'ouvert' which 221 • 26 has passive meaning. I s~i£qe 1 suivre1 has two past part, m free variation: 0 the regular sjegyd and, as if in Cl. II, the irregular I sjegid 1suivi1. koze 'cuire' is defective in that it lacks finite forms. The irregular past part, kwej(t) 'cuit' shows re placement of /o/ by /wej/. kunese 1connaitre' has stem-final /j/ for /s/ in 3 pres., hence /kunej/ 'il connait'. IV. Defective verbs In addition to those mentioned in the preceding sec tions, the following verbs have incomplete paradigms: baze 'eclore' has no pres. *espja 'regarder, epier' has only 2 imper. /espjo/ 'regarde' with stress on the inflection and 5 imper. /espja^/ 're- gardez'. *gejta 'regarder, guetter' has only 2 and 5 imper., e.g. /gejto/ 'regarde'. kere 'chercher, querir' has only the inf. 2^The former occurs only in phrases with the auxiliary abe 'avoir', e.g. /a djerbyt./ 'il a ouvert'; the latter occurs in phrases with linking verbs, e.g. /ez dyberto./ 'elle est ouverte'. 222 kwa 'couver1 has only inf. and past part. nCse 'naxtre' has no imper. ♦penjUrla 'ne pas pendre d'aplomb' has only 3 and 6 pres., e.g. /penjorlo/ 'il ne pend pas d'aplomb'. plaZe 'plaire' has no imper. The form bakx 'voila, voici' is uninflected and there fore not assignable to any one verb class. It is neverthe less identified as a verb because it functions as head in verbal phrases, for instance, when preceded by accusative J 1 / * * clitics, e.g. /lu bakx./ 'le voila1. 4.2.5. Uninflected Word-classes 4.2.5.1. Adverbs Adverbs (adv.) are uninflected full words identified by their ability to occur in utterance-final position im mediately following a full verb in the frame /ako se faj ... ./ 'cela se fait ... 1 In this frame the tense of the full verb fa. Irr. can be adjusted to future or past to accommodate temporal adv. Certain locative adv. are lexically best fitted after the I verb m£ te 'mettre' in the frame /ak6 se met ... ./ 'cela se met ... ' 223 In such a position an adv. functions as verb-modifier. This criterion identifies a large majority of the adv. of this dialect. There follows a list of common simple adv aban 'avant' abarezo 'pele-mele1 alabe/ 'a ce moment-la' alyrs 'ailleurs' anej 'aujourd'hui' apej 'apres' I ardit 'vivement, avec ardeur aro 'maintenant' ataw 'ainsi' awtur 'autour' aze 'hier1 bas 1en bas1 be 1 bien' belew 'peut-etre* betkop 'beaucoup1 biste 'vite' bjen 'bien' daban 1devant' dar e 'derriere' deza 'deja' dret ’debout' duman 1demain' I dympe j 1depuis' enkero 'encore' espres 'expres' gajre 'ne ... guere' guto 'ne ... goutte' ko/ek 1sur-le-champ' len 1 loin' lew 'tot, bientot' maw 'ma1' mens 'moins' mej 'davantage' milu 'meilleur' naw 'en haut’ parel 'pareil' pas 1ne ... pas' paw 'peu' 224 pla 'tres, beaucoup' plys 'ne ... plus' prampu 'bien assez' prose 'pres, a cote' pru 'assez' pryme 'd'abord' suben 'souvent' taben 'aussi' tan 'tellement' t£ir 'tard' trop 'trop' tuzur 'toujours' zamej 'ne ... jamais' In addition to functioning as verb-modifiers, a number of common adv. also function as modifiers of adj. and of 27 adv. The following uninflected forms, which do not fit in the frames in the position just defined, are identified as adv. by these criteria. (1) The interrogative adv. kumo 'comment', kuro 'quand', perke 'pourquoi', un 'ou' function as verb-modifiers and occur preposed to verbs, thus /ako : kumo se faj?/ 'comment cela se fait-il'. (2) The interrogative adv. kumbien 'combien' functions like 27in sentence-modifying function, adv. may occur in sentence-initial or in sentence-final positions before or after an external juncture, usually /“*/, or a potential pause indicated by the absence of sandhi phenomena. Words of other classes such as sentence-linkers may also occur in these positions. 225 other quantitative adv. as preposed modifier in nominal phrases (see 5.2.1.1.A.[l2]b.i). (3) The adv. m~ie-i 'a demi' and kawke 'environ' function as adj.-modifiers. (4) The adv. tut,. 'tout' functions as adv.-modifier. (5) The adv. mei 'd'autre'— distinct from mei 'davantage'— functions as postposed modifier of indefinite pronouns (see 5.2.1.3.[6]). This word is identified as an adv. by the negative criterion that it does not qualify as a member of any other word-class and because, like adv., it is an unin flected full word occurring with strong stress in phrases. 4.2.5.2. Pronominal Adverbs The pronominal adverbs (pron. adv.) ne 'en, le, de lui, d'eux, etc.' and ijg 'y' are phrase-bound function words. Their respective regular forms are /ne/ and /i/. The posi tional variants of us. will be described in 5.2.1.4.1.A.(2). The pron. adv. are substitutes for single words and also for groups of words including adv. They occur singly in the same syntactic positions as accusative or dative pron. Their functions and their domains as substitutes are enumerated at this point with token examples. 226 The pron. adv. ne. substitutes for certain types of: (1) direct object complements (see 5.2.1.4.II.[l]b), e.g. /a de bin./ 'il a du vin' : /n a./ 'il en a' (2) indirect object complements (see 5.2.1.4.II.[2]b), e.g. /me subeni de sa maj./ 'je me souviens de sa mere' : /men subeni./ 1je me souviens d'elle'. (3) subjective complements (see 5.2.1.4.II.[3]), e.g. /sun rises./ 'ils sont riches' : /ne sun./ 'ils le sont', /es kantunje./ 'il est cantonnier' : /n es./ 'il l'est' (4) attributes to verbals (see 5.2.1.4.IV.[2]), e.g. /ben de dela./ 'il vient de la-bas1 : /ne ben./ 'il en vient' (5) attributes to adjectivals (see 5.2.1.2.[7]a,b), e.g. /ez zaluzo d aket ome./ 'elle est jalouse de cet homme' : /n ez zaluzo./ 'elle est jalouse de lui', /sej syr ke fara masan tens./ 'je suis sur qu'il fera mauvais temps' : /ne sej syr./ 'j'en suis sur'. The pron. adv. i^ substitutes for certain types of: (1) indirect object complements (see 5.2.1.4.II.[2 ]b), e.g. /zogon a 1 estyzet./ 'ils jouent a cache-cache' : /i zogon./ 'ils y jouent' (2) attributes to verbals (see 5.2.1.4.IV[2]), e.g. /eren pu kamin./ 'nous etions sur la route' : /j eren./ 'nous y etions1. 4.2.5.3. Prepositions Prepositions (prep.) are identified as function words which appear unstressed followed by an axis in exocentric phrases (see 5.2.2.1). There follows a list of simple prepositions. ^ aban 'avant' apej 'apres' daban 'devant1 dambe 1avec' dar4 'derriere de 'de' debat 1 sous' din 'dans' I dympej 'depuis' en 'en' ende 'avec 1'intention de' enta 'chez' entre 'entre' 28some of the forms listed here, especially .a and de. occur with many different meanings and seem to be analyzable as homophonous but separate morphemes. The semantic differ ence, however, is partly due to the type of axis that fol lows the prep, so that the meaning of the prep, is often definable only in the framework of the phrase. For some illustrative examples of the various meanings of a and de in prepositional phrases, see 5.2.1.1.F.(l)a,b and 5.2.1.3. (7)a.i,ii. At least three markedly different meanings of /per/ may be distinguished at this point: perx denotes location, per3 purpose, and per3 the agent of the action. 228 kumo 'comme' per3 'par' kuntro 'contre1 selun 'selon' penden 'pendant' sin 'sans' perx 'a travers, sur, par' syr 'sur' pera 'pour' The forms aban, apej. daban, dare. debat 'dessous' are also adv. The form kumo is also a conjunction. The common prepositions given in the table on p. 229 occur combined with certain forms of the definite article. The variants of the morphemes combined are morphologically defined. As indicated in the table, in the combinations /aw/, /entaw/, /du/ /pu/, and /su/, the variants of the morphemes involved are further conditioned phonologically by an initial consonant in any word immediately following. The various combinations are listed in the first column on the table. The breakdown into components is given in the second column. In group 1 the prep, peri and svr have the respective variants /pel/, /syl/. In group 2 the forms /lu/ and /lus/ of the def. art. have respectively the variants /w/ and /s/. In the combinations of group 3 both components have special forms. The variants of the prep, are: /d/ for ^e, /p/ and /pe/ for peri . /s/ for syr. The variants Combinations Group 1 pellu 'a travers le' pella 'a travers la' pellas 'a travers les1 syllu 'sur le' sylla 'sur la' syllas 'sur les' Group 2 aw 'au' entaw ’chez le' as 'aux' entas 'chez les' des 'des' Group 3 du 'du' pu 'a travers le' 29 su 'sur le' pes 'a travers les' sus 'sur les' 229 Standing for non-occurrina per lu per la per las syr lu syr la syr las a lu (before a C) enta lu (before a C) a lus enta lus de lus de lu (before a C) per lu (before a C) syr lu (before a C) per lus syr lus A Q ^The occasional occurrence of /syl/ for /su/ seems to be random and is presumably by analogy to /syllu/. 230 of the def. art. are /u/ for /lu/, /s/ and /us/ for /lus/. 4.2.5.4. Conjunctions Conjunctions (conj.) are identified as introductory function words followed by a finite verbal in subordinate clauses. Simple conjunctions include: kan 'quand, meme si' paske ~ parsoke 'parce que' ke 'que, au point que' pqiske 'puisque' kumo 'comme* se 'si' kwake 'quoique' tanke 'tant que, aussi long- temps que' Examples: /kam penjaben lu tabat/ 'quand nous suspendions le tabac', /kan sejoj mjej mor/ 'meme si j'etais (lit. je serais) a demi mort', /k anabo dala/ 'qu'il allait faucher1, /kum eron estran]4s/ 'comme ils etaient etrangers', /s abja/ talen/ 'si vous aviez faim*. Where no main clause is present, the conjunction Jig. signals a command. The verbal is then in 3 or 6 pres. S., and an entire intonation contour, identical with that of a statement, accompanies the construction, e.g. /ke sen angem pas./ 'qu'ils ne s'en aillent pas', /k aw leziske./ 'qu'il le lise'. 231 4.2.5.5. Coordinators These are function words which connect functionally equivalent constituents. Coordinators are not semantically related to the construction they precede, so that they function as connectives rather than attributes. They may be subdivided into two types. 30 (1) Coordinating Universals These connect all categories of words and construc tions, including utterances by the same or different speak ers. They are: e 1et' sjoske ... sjoske 1soit ... soit' I I ni . . . ni 1 ni ... ni ’ u 'ou' me mais ' u . . . u ' ou ... ou1 sinu 'sinon' Examples: /de kawlets e kawkes pores/ 'des choux et quel- ques poireaux', /la bjerzes e lu bun drw/ 'la Vierge et le bon Dieu1, /ke sen ange e ke torne pas./ 'qu'il s1en aille et qu'il ne revienne pas', /ni ty ni zu/ 'ni toi ni moi', 30The terms "coordinating universals" and "attributive universals" and basically their definitions are borrowed from F. A. Agard's Structural Sketch of Rumanian (Baltimore, 1958), pp. 116-118. 232 /ni em pagan ni sim paga/ 1ni en payant, ni sans payer', /ni ke minjese ni ke bebu.se/ ' ni qu'il mangeat, ni qu'il but', /maridado me maleruzo/ 'mariee mais malheureuse1, /trabalom pla : me dromon gajre./ 'ils travaillent beaucoup, mais ils ne dorment guere', /dedin sinu deforo/ 'dedans sinon dehors', /beno : sinu damori asiw./ 'viens, sinon je reste ici', /sjoske et sjoske sum paj/ 'soit lui soit son pere', /sjoske a la bino sjoske din lu caj/ 'soit a la vigne, soit dans le chai', /dimars u dimekres/ 'mardi ou mercredi', /dambe u sanz elo/ 'avec ou sans elle', /u aribes u te baw kere./ 'ou bien tu arrives, ou bien je vais te chercher'. 31 (2) Sentence-Linkers These function words are limited to connecting sentenc es or utterances. They occur after one of the junctures and they are disjoined from the construction they introduce, i.e. set off by potential pause and absence of sandhi phe nomena. They are frequently but not necessarily preceded by one of the commonest coordinating universals £., jae., ji. 3lThis is the term used for this type of function words by W. Nelson Francis in The Structure of American English (New York, 1958), p. 417. 233 They include: alabe/ 1 alors 1 anfin 'enfin' apej 1ensuite' ataw 1ainsi1 awtromen 'sans cela' dun 1 done' perkunseken 'par consequent' purtan 'pourtant' sakela 'quand meme' sependen 'cependent' tapla 'tout aussi bien' The forms alabej6. apej. ataw. awtromen are also ad- , 32 verbs. Examples: /(e) alabe/ biskuri yno lebre./ '(et) alors, je vis un lievre', /(mez) anfin perke crides tan?/ '(mais) . I , enfin, pourquoi cries-tu tant', /(e) anfin la lebre m eskapet./ '(et) enfin, le lievre m'echappa', /(e) ataw me panet mun arzen./ '(et) ainsi, il me vola mon argent', /dun la meturi a la porto./ 'done, je la mis a la porte', /(me) 32Actually, these words fulfill a double function, that of sentence-linkers and that of sentence-modifiers, Of the many adv. that function in these ways, those four are in cluded here on two counts: they occur most frequently in this double function and their meanings as sentence-linkers are more or less distinct from their meanings as verb- modif iers. This semantic difference is sufficiently marked in the case of /awtromen/ 'autrement' to regard this form as two homophonous but discrete morphemes, cf. /ako se fajo awtromen./ (adv.) '9a se ferait d'une autre fagon' and /awtromen ako se fajo./ (coordinator) 'sans cela, ga se ferait'. 234 sakela se pob be tajza./ '(mais) quand meme, elle peut bien se taire’, /tapla m akabijoj akeb busin./ 'tout aussi bien, je finirais ce morceau'. 4.2.5.6. Attributive Universals Attributive universals are uninflected full words which function as attributes (or modifiers) of many if not all categories of words and constructions, including utterances. As illustrated below, they occur in initial or in final position. They are: enkero 'encore' sylomen 'seulement' memo 'meme' taben 'aussi' preske 'presque' tapaw 'non plus' raj 'peu importe' zyste 'a peine, juste' sadi/ 'dit-il, il dit, etc.' The words enkero and taben are also adverbs. The word tapaw occurs only in correlation with a preceding negative adverb such as pas 'pas'. Examples: /enkero kawkyn/ 'encore quelqu'un', /enk£ro a pe/ 'encore a pied', /enkero sin k aw beli/ 'encore sans que je le veuille', /memo zu/ 'meme moi', /memo sun ome/ 'meme son mari', /memo se leba/ 'meme se lever', /memo per bin sos/ 'meme pour vingt sous', /memo se buljo/ 'meme s'il voulait', 235 /memo preske trop espes/ 'meme presque trop epais', /preske digyn/ 'presque personne', /preske adret/ 'presque adroit', /preske kumo kaw/ 'presque comme il faut', /ako raj/ 'pa, pa n'a pas d'importance', /duman raj/ 'demain, peu importe', /debal lu krymet raj/ 'sous la mue, pa va', /ke kride -♦ ^sj./ 'peu importe qu'il crie', /sylomen akiw/ 'la seule- ment', /sylomen lu dimese/ 'le dimanche seulement', /sylomen a unz uros/ 'seulement a onze heures', /sylomen abjo duplidat de se prenge la klaw./ 'seulement il avait oublie de prendre la cle', /et taben/ 'lui aussi', /mus^r taben/ 'monsieur aussi', /taben paske plebjo./ 'aussi parce qu'il pleuvait', /zu tapaw/ 'moi non plus', /suz majnajes tapaw/ 'ses enfants non plus', /din la krambo tapaw/ 'dans le chambre non plus', /as pas pla kambjat tapaw./ 'tu n'as pas beaucoup change non plus', /zyste la familo/ 'la famille seulement', /zyste trez u katre kintaws/ 'a peine trois ou quatre quintaux', /zyste so ke kaljo/ 'juste ce qu'il fal- lait'. I The word sadi/. accompanied by juncture, introduces or follows a quotation, e.g. /o -* sadi/./ 'oui, dit-il', /sadi/ -* paz anej./ 'elle a dit: pas aujourd'hui', /ne turnan - * sadi/./ 'nous en revenons, dirent-ils'. 236 4.2.5.7. Particles There is a residue of uninflected function words which occur as bound elements in phrases. They are: (1) the partitive article iig. 'du, de la, des' . This article fills the same position in phrases as other nominal markers. It is distinct from the preposition j&g. 'de' in that phrases introduced by the article are endocentric whereas phrases introduced by the preposition are exocentric. Like other nominal phrases, the phrases introduced by the article have among their functions that of axis of a preposition (see 5.2.2.1.[3]), e.g. /dambe de kalaws/ 'avec des cailloux'. (2) the intensifier pas 'pas' which functions as preposed modifier of negative pronouns (see 5.2.1.3.[7]) and negative adverbs (see 5.2.1.4.IV[1]b). (3) the particles afIn 'afin', dinko 'jusque', and pramo 'a cause' which occur exclusively as bound constituents of phrasal prepositions (see 5.2.1.6) and phrasal conjunctions (see 5.2.1.7). (4) the particles 'de' and 'a' which function as con nectives a. in adjectival phrases with infinitive verbals as 237 attributes (see 5.2.1.2.[7]a), b. in pronominal phrases with adjectives as attributes (see 5.2.I.3.[7]a.iv), c. in verbal phrases with infinitive verbals as complements (see 5.2.1.4.II.[5]), d. in prepositional phrases with infinitive verbals as axes (see 5.2.2.1.[8]). (5) the emphatic particle ]££. 'que' which occasionally in- i 3 3 troduces a statement, e.g. /ke nu./ 'mais non'. (6) the sandhi connective /z/ which joins various types of constituents of verbal phrases (see 5.1.2.IV). 4.2.5.8. Interjections This term applies to a class of stressed words which occur primarily as independent utterances rather than as constituents of syntactic constructions. Here are some of the most common interjections. (1) Affirmative particles a. o 1oui' answers a positive question affirmatively. ^^This particle, which in other Gascon dialects serves to introduce any statement, is used sparingly and with em phatic meaning by Dzcs. speakers. 238 b. co 'mais oui' has emphatic affirmative meaning and may occur either alone or after q. 'oui'. I c. si 'si' answers a negative question affirmatively. I I d. sifet 'mais si' is the emphatic equivalent of si 'si'. (2) Negative particles a. nu 'non' is the negative response used in familiar ad dress . b. nani 'non' is the negative response used in deferential address. (3) Courtesy terms 1 a. adiw 'bonjour, au revoir, adieu' is used in familiar address as a greeting and for leave-taking. I b. adisa/ 'bonjour, au revoir' differs from adiw only in that it is used in deferential address. c. pleti 's'il vous plait, s'il te plait', spoken with falling intonation, is a courtesy formula which accompanies 34 . an order. Spoken with rising intonation, pleti 'comment' is used to request a repetition of an utterance. 34a 1so used are the stereotyped phrase /se te plaj/ (familiar) 's'il te plait' and the borrowed phrase /si bu plet/ (deferential) 's'il vous plait'. 239 (4) Commands to animals These interjections, some of which are onomatopoetic, are used with the meanings given in parentheses, e.g. ow (to order a horse to halt), bo (to order cattle to halt), ojo (to order a horse to turn right), a (to order cattle forward), are (to order animals backward), ulu (to chase dogs), aty (to urge dogs to attack). The following are calls to animals, e.g. te (dogs), mini (cats), tiru (ducks), lili (young ducks), beru (geese), tiku (turkeys); for other examples, see 1.2 and 1.2.4 above. (5) Exclamations Among these are the following: e 'eh', anen 'allons', te 'tiens', desigy 'bien sur', pardi 'parbleu' and its con- t tracted form /pai/. 4.3. Derivation In this dialect derivation is the strongest process of word formation. There are three types of derivation: suf- fixation, prefixation, and parasynthetic formation, that is, the simultaneous addition of a prefix and a suffix to a primitive. The term "primitive" refers to the basis from which a complex word is derived. Affixes which are closely related semantically and are 240 represented by different phonemic shapes are here treated as allomorphs of one morpheme provided they occur in com plementation relative to individual primitives. The allo- morph with the widest distribution is taken as the norm. 4.3.1. Alternations in Stems Basic alternations are exemplified under the headings of the various affixes. They are represented in the stem formulae by the symbols which have been defined earlier in section 4.1. Thus, the derivation of the vb. /resega/ 'scier' from the N /resego/ 'scie1 involves— among other changes— the basic alternation /e ~ e/; this is indicated l _ I by writing the N stem with £, thus resCgo, f. On the other hand, the use of the formula resEga for the derived vb., with the symbol E which stands for /e ~ e/, shows that the replacement of /e/ by /e/ is also manifest in the inflected forms of the vb., e.g. /resegat/ 'scie' but /resegi/ 1je scie1. The following are alternations which are applicable only in derivation and which occur more than once. No sym bols are assigned to these alternations. 241 I. Phonologically conditioned alternations A. Consonants /w ~ 1/ This is the replacement of the word-final fricative /w/ by the lateral /l/ before a vowel-initial suffix, e.g. /saw/ 'sel', /sala/ 1saler1. Stem-final /w/ is invariably replaced in stem morphemes before a vowel, but the alternant is either /l/ or, more rarely in derivation, the basic al ternant /b/, e.g. riw > /riw/ ’ruisseau' vs. /ribero/ 'ri viere'. This alternation is automatic only to the extent that prefix-final /_w/ is preserved in intervocalic position in a few derived adj. B. Vowels Palatalization of , /i/ /i/ is palatalized to /j/ in unstressed position, e.g. /pais/ 'pays', /despajza/ 'depayser'. This alternation is due to the fact that in native words, other than onomato- poetic words, when two vowels are in hiatus, one of them must bear the word-stress. 242 II. Morphologically conditioned alternations These alternations occur when an affix is added to certain stems. A. Consonants (1) /t - 1/ The voiceless dental stop /t/ is replaced by the voiced dental lateral /l/ in, e.g. /kutet/ 1couteau', /kutelo/ 'coutelas1. (2) ,/t - p/ The voiceless dental stop /t/ is replaced by the voice less bilabial stop /p/ in, e.g. /esklat/ 'eclat', /esklapun/ 'copeau'. (3) Palatalization The lateral /l/ and the nasal /n/ are palatalized to y ^ V /l/ and /n/ respectively, e.g. /farul/ 'verrou', /farula/ 'verrouiller'; /pyn/ 'poing', /pynat/ 'poignee'. (4) , /n ~ m/ The dental nasal /n/ is replaced by the bilabial nasal /m/ in, e.g. /kryn/ 'nuage', /krymus/ 'nuageux'. 243 (5) /r - r/ The single trill /r/ is reinforced into the multiple trill /?/ in, e.g. /fer/ 'fer' , /feraso/ 'pelle a feu*. B. Loss of Phonemes When followed by a suffix, a stem may occur in reduced form. The one or two phonemes deleted may be stem-final, e.g. /-a/ in /plura/ 'pleurer', /plurnisa/ 'pleurnicher', /-re/ in plewre/ 'pleuvoir', /plebina/ 'pleuvoir legere- ment'; occasionally the deleted phoneme is stem-internal, e.g. /j/ in/?elajre/ 'chiffonier', /pelarot/ 'chiffonier'. The alternations presented above, as well as unique alternations, are noted in parentheses as they occur in derivational relations. The distinction between the two types of conditioning is made by assigning to the symbol 0 0 the meaning "morphologically conditioned," thus ly/, f. 'lumiere' (// 0 0 k/) : alyka 'allumer', and to the symbol : the meaning "phonologically conditioned," thus semaw, f. 'comporte' (/w : l/) : semalot, m. 'cuvette'. 4.3.2. Suffixation Derivational suffixes are bound morphemes which follow the stem (or nucleus) and which may themselves be followed by inflectional suffixes. Derivates may contain one or two 244 derivational suffixes. A stem may be free or bound, and it may be a base morpheme or a derived stem. In this chapter the term "stem" is used in the sense "derived stem." With some exceptions to be noted later, derivates are 35 stressed on the word-final suffix. In most instances the stress pattern of a derivate therefore differs from that of the primitive. A stress mark on a given suffix is used as a simple device to represent the significant position of the stress on derivates formed by means of that suffix, thus {-ado}. The notation "with AA" preceding a series of examples refers to allomorphic alternations in the primitives. For reasons of economy, I do not list the most widespread of these alternations, namely the regular loss of the final vowel of the primitive (stressed or unstressed) before a vowel-initial suffix, for example, the loss of /-e/ in /tewle/ 'tuile' before the suffix /-ado/ in /tewlado/ 'toi- ture'. Suffixes are arranged according to the word-classes of the derivates. The order of presentation of the various 35Since the strong stress is limited to either the ul timate syllable of a word or to the penult, all dissyllabic suffixes automatically receive word stress. 245 suffixes is that of decreasing frequency of occurrence in the data. For practical reasons, however, suffixes which are homophonous but discrete morphemes are presented in sequence. The parasynthetic formation of verbs is best fitted under suffixation. 4.3.2.1. Noun? (1) C-eti) 'diminutive' is a productive suffix which denotes small size or a small part of a whole item. This suffix also yields pet names and designates the young of animals. It forms masc. N usually, but not consistently, on masc. primitives. In derivates referring to living beings, /-et/ is followed by allomorph zero of the morpheme 'male'. This suffix is added to: N bases with no AA bastu(n), m. 'baton' : bastunet, m. 'petit baton' faws=, f. 'faux' : fawset, m.'serpe' kuzo, f. 'citrouille' : kuzet, m. 'courge (article de couture)' suK, m. 'base du tronc d'arbre' : suket, m. 'billot' taP, m. 'coteau' : tapet, m. 'talus' tesu(n), m. 'cochon' : tesunet, m. 'petit du cochon' 246 N bases with AA al, m. 'ail' (/l 0 0 l/) : alet, m. 'ail jeune' kOt, m. 'cou* (/t 0 0 p/) : kupet, m. ' nuque' pyn, m. 1poing' (/n <» £/) : pynet, m. 'poignet' N stems with no AA pulo, f. 'poule' : pulet, m. 'poulet' with AA t ( drOlle, m . ' gargon' : drullet, m. 'petit gargon' Ext. N bases •4 .1 . _ ^ 36 with no AA tapyrl- (cf. taP, m. 'coteau') : tapyrlet, m. 'rai- dillon' with AA purtan- (cf. pOrto, f. 'porte') : purtanet, m. 'por- tillon' The following are pet names given to work animals and based on: Vb. bases with’ no AA lawra 'labourer' : lawret 'boeuf attele a droite' 3^That is, with no alternation other than the extension of the base. 247 Bound bases with no AA berm- : bermet 'vache attelee a gauche' brak- : braket 'vache attelee a droite' kawb- : kawbet 'bceuf attele a gauche' ^ The allomorph /-eto/ forms fem. N usually, but not con sistently, on fem. primitives. In derivates referring to living beings, the sequence /-eto/ consists of two suffixes: /-et-/ 'diminutive' and /-o/ 'female'. The primitives are: N bases with no AA kadeno, f. 'chaine1 : kadeneto, f. 'chaxnette' kyrgo, f. 'cruche1 : kyrgeto, f. 1cruchon' majzu(n), f. 'maison' : majzuneto, f. 'maisonnette' V v munje, m. 'gros haricot' : munjeto, f. 'haricot' I I pxnsos, f. 'pinces' : pinsetos, f. 'pincettes' pupo, f. 'mamelle, sein' : pupeto, f. 'tetine' sawklo, f. 'sarcloir a deux bees' : sawkleto, f. 'sarclette' tawlo, f. 'table' : tawleto, f. 'tablette' •^^These data are obsolescent. While all informants re call these names, only a few older men could provide precise glosses. 248 N bagjgs with AA furko, f. ' fourche1 (/k 0 0 s/) : furseto, f. 1 four- chette' N stems with no AA pulo, f. 'poule* : puleto, f. 'jeune poule' with AA drOllo, f. 'fille' : drulleto, f. 'fillette' Adi. bases with no AA segund 'second' : segundeto, f. 'cadette1 The allomorph /-ot/, 'diminutive' with occasional de- preciative or humoristic connotation, forms masc. N primar ily on fem. N. Sex differentiation is not made in derivates referring to living beings. The primitives are: N bases with no AA ■ v vi 38 balezo, f. 'balai' : balezot, m. 'petit balai' bariko, f. 'barrique' : barikot, m. 'baril, petit fut' baro, f. 'barre' : barot, m. 'barreau' 38There is no morphemic contrast between this and the rival form balezun. m. 249 — 1 a — I buriko, f. 'anesse' : burikot, m. 'anon' gaspo, f. 'grappe' : gaspot, m. 'grappillon' krabo, f. 'chevre' : krabot, m. 'chevreau' krambo, f. 1chambre1 : krambot, m. 'chambrette obscure' sako, f. 'sac de jute' : sakot, m. 'petit sac’ with AA semaw, f. 'comporte1 (/w : l/) : semalot, m. 'cuvette portee sur la tete' N stems with no AA pepin, m. 'pepere' : pepinot, m. 'petit vieux' (ironical) kano, f. 'chienne' : kanot, m. 'chiot' with AA pelajre, m. ' chif fonnier' (/j 0 0 0/) : pelarot, m. 'chiffonnier' (ironical) Bound stems with no AA marm- (cf. marmalo, f. 'marmaille') : marmot, m. 'marmot' Christian, names formincr nicknames with no AA zan, m. 'Jean' : zanot, m. with AA ■ , margento, f. 'Marguerite' : margot, f. I — — t pjCre, m. 'Pierre' : pjerot, m. Ext. N bases with no AA azik- (cf. aze, m. 'ane') : azikot, m. 'anon' Allomorph /-oto/ forms fem. N on: N bases with AA paw, m. 'barre de ridelle' (/w : l/) : paloto, f. 'plaquette de bois pour hausser un cuvier' Vb. bases with no AA pika 'piquer' : pikoto, f. 'petite verole' Allomorph /-un/ forms masc. N on: N bases with no AA balezo, f. 'balai' : balezun, m. 'petit balai' gat, m. 'chat' : gatun, m. 'chaton' pigaso, f. 'hache' : pigasun, m. 'hachette' with AA finCstro, f. 'fenetre' : finestrun, m. 'lucarne' N stems with no AA kareto, f. 'charrette' : karetun, m. 'charreton' 251 N stems with AA esklat, m. 'eclat' (/t 0 0 p/) : esklapun, m. 'copeau' Ext. N bases with no AA lard- (cf. lar, m. 'lard') : lardun, m. 'lardon' This suffix also occurs alone or followed by /-et/, /-eto/ with cumulative diminutive significance in nicknames derived from Christian names: with no AA franso, f. 'France' : fransun, f. zano, f. 'Jeanne' : zanuneto, f. with AA franswa, m. 'Frangois' : fransunet, m. Allomorph /-u/ forms a few masc. N on: N bases with AA kutCt, m. 1 couteau' (/t 0 0 l/) : kutelu, m. 'petit couteau' Past part, with no AA entamenad : entamenadu 'premier morceau du pain coupe' In blacro. f. 'verve' : blagasu 'moulin a paroles' this 252 suffix is preceded by the intensive suffix /-as-/. t Allomorphs of limited occurrence include: /-il/, /-in/, /-ol/ which form masc. N, and /-iso/ which forms fem. N on: N bases with no AA grano, f. 'graine' : graniso, f. 'gresil' kami(n), m. 'chemin' : kaminol, m. 1sentier' , t musko, f. 1mouche1 : muskil, m. 'moucheron, moustique1 W stems with no AA 1 I ^ N pepe, m. 'grand-papa' : pepin, m. 'pepere' (endearing) Ext, adi. bases with AA trusk- (cf. trOs'*, 'chetif') : truskil 'petit bouchon' (endearing term to address a child) (2) {-et2} 'activity or thing associated with' forms masc. N essentially on primitive vbs.: Vb. bases with no AA estyza 'cacher' : estyzet, m. 'cache-cache' kapysa 'faire une cabriole' : kapyset m. 'cabriole' siskla 'pousser des cris' : sisklet, m. 'cri pergant' 253 tira 'tirer' : tiret, m. ’tiroir’ yb. sjLems with no AA birUla 'tournoyer' : birulet 'tourniquet' Bound vb. stems with no AA pisurl- (cf. pisurleza 'pissoter sans cesse') : pisurlet, m. 'verge' Allomorph /-eto/ forms fem. N on: N bases with no AA ' . 39 eskino, m. 'dos' : eskineto, f. with AA kOsto, f. 'cote' : kusteto, f. 1cotelette* I , sClo, f. 'selle' : seleto, f. 'harnachement' Bound bases with no AA kar- (cf. kareza 'charroyer1) : kareto, f. 'charrette' Ext, bases with no AA sank- (cf. san, f. 'sang') : sanketo, f. 'sang cuit' •^^Cannot be glossed independently because of unique occurrence in the set phrase f& eskineto 'faire la courte echelle'. 254 Adi. bases with no AA blank 'blanc' : blanketo, f. ’ ragout a la sauce blanche1 dusA 'doux1 : duseto, f. 'inache, doucette1 Allomorph /-un/ forms masc. N on: Vb. bases with no AA bu/a 'boucher1 : bu/u(n), m. 'bouchon' kawsa 'chausser' : kawsu(n), m. 'chausson' turtila 'tortiller' : turtilu(n), m. 'gateau en tortillon' N bases with no AA bus, m. 'buis' : busu(n), m. 'buisson' Ext. N bases with no AA kambaz- (cf. kambo, f. 'jambe') : kambazu(n), m. 'jambon' The form /-is/ occurs with the vb base s£te 'asseoir' in the N setis=. m. 'siege' stressed on the base. (3) {-as} 'augmentative' denotes large size of an object or superlative degree of a quality with frequent connotations of depreciation or contempt. This suffix forms masc. N 255 usually, but not consistently, on masc. primitives. In derivates referring to living beings, /-as/ is followed by zero meaning 'male'. This is a productive suffix which appears with: N bases with no AA tesu(n), m. 'cochon1 : tesunas", m. 1gros cochon' fango, f. 'boue' : fangasA, m. 'bourbier' with AA ibrOno, m. 'ivrogne1 : ibrunas*, m. 'ivrogne invetere1 N stems with no AA ka[n], m. 'chien* : kanasA, m. 'gros chien' with AA dr6lle, m. 'gargon' : drullasA, m. 'gros gargon' Allomorph /-aso/ forms denominal fem. N usually, but not consistently, on fem. primitives. In derivates refer ring to living beings, the sequence /-kso/ consists of two suffixes: /-as/ 'augmentative' and /-o/ 'female': N bases with no AA I % bi(n), m. 'vin' : binaso, f. 'mauvais vin' fenno, f. 'femme' : fennaso, f. 'femme demesuree' 256 kambo, f. 'jambe' : kambaso, f. 'grosse jambe' kano, f. 'paresse' : kanaso, f. 'paresse extreme' ligo, f. 'lie' : ligaso, f. 'lien de gerbe usage' majzu(n), f. 'maison' :majzunaso, f. 'enorme maison' ma(n), f. 'main' : manaso, f. 'grosse main' pato, f. 'patte' : pataso, f. 'grosse patte' pumo, f. 'pomme' : pumaso, f. 'grosse pomme' with AA ibrOno, f. 'ivrogne' : ibrunaso, f. 'ivrogne inveteree' kwSso^ f. 'cuisse' : kwesaso, f. 'grosse cuisse' trCzo3 f. 'femme malpropre' : trezaso, f. 'femme tres sale' Ext. N bases with no AA putun- (cf. putu, m.^ 'baiser') : putunaso, f. 'gros baiser' Allomorph /-o/, denoting the comparatively larger size of an object, forms fem. N on masc. N. These derivates are stressed on the nuclei. N bases with no AA pajrol, m. 'chaudron' : pajrolo, f. 'grand chaudron' ^Identical in meaning with the rival form putet. m. 257 tupi(n), m. 'cafetiere en terre' : tupino, f. 1marmite en terre' with AA karjot, m. 'brouette' (/t 0 0 l/) : karjolo, f. 1voiture a cheval1 kutet, m. 'couteau' (/t 0 0 l/) : kutelo, f. 'coutelas' rastet, m. 'rateau' (/t 0 0 l/) : rastelo, f. 1 rateau en bois ' ridew, m. 'rideau' (/w : l/) : ridelo, f. 'ridelle' Allomorphs of limited occurrence include: /-ado/, /-elo/, /“6lo/, /-ero/ which form fem. N on: N bases with no AA I v v , agylo, f. 'aiguille' : agylado, f. 'aiguillon' 1 , kamizo, m. 'chemise' : kamizolo, f. 'camisole' with AA bekat, m. 'hoyau' : bekadelo, f. 'trident, fourche courbe' riw, m. 'ruisseau' : ribero, f. 'riviere* (4) {-omeni 0 0 -isomen} in sense 'fact or act of forms de verbal masc. N. This suffix remains productive through the agency of French. The form /-isomen/ occurs with is-vbs, the form /-omen/ with vbs. of other classes. Among the many N derived with /-omen/ are the following: 258 Vb. bases with no AA amyza 'amuser' : amyzomen, m. 'amusement' bate 'battre' : batomen, m. 'battement' bira 'tourner' : biromen, m. 1tournement1 kambja 'changer' : kambjomen, m. 1changement1 kumensa 1commencer' : kumensomen, m. 'commencement' manka 'manquer' : mankomen, m. 'manquement' trambla 'trembler' : tramblomen, m. 'tremblement' Vb. stems with no AA brutuna 'bourgeonner' : brutunomen, m. 'bourgeonnement' empuzuna 'empoisonner' : erapuzunomen, m. 1empoisonne- ment' karga 'charger' : kargomen, m. 'chargernent' plasa 'placer' : plasomen, m. 'placement' razuna 'raisonner' : razunomen, m. 'raisonnement' This suffix is often added to Fr. primitives; for example, aranzomen, m. 'arrangement' (cf. nat. vb. arenga 'arran ger'), enkurazomen, m. 'encouragement' (cf. nat. vb. enkuraja 'encourager'), zyzomen, m. 'jugement' (cf. nat. vb. v v . . . zyja juger ). There follow some examples of derivates in /-isomen/: 259 feag.es with no AA bumis 'vomir' : bumisomen, m. 1vomissement1 establis 'etablir* : establisomen, m. 'etablissement' Vb. stems with no AA I agrandis 'agrandir1 : agrandisomen, m. 'agrandissement1 degrusis 'degrossir' : degrusisomen, m. 'degrossisse- ment' refredis 'refroidir' : refredisomen, m. 'refroidisse- raent' radusis 'radoucir1 : radusisomen, m. 1radoucissement' The primitives are Fr. vbs. in, e.g. aplanisomen, m. 'apla- nissement' (cf. nat. vb. aplana 'aplanir, niveler1), elarzisomen, m. 'elargissement' (cf. nat. vb. alarzis 'elar- gir ' ) . Allomorph /-asjun/ occurs in complementation with /-omen/ relative to individual primitives. This suffix is also an active formative. The derivates are deverbal fem. N: Vb. bases with no AA awmenta 1augmenter' : awmentasju(n), f. 'augmentation' ezita 'hesiter' : ezitasju(n), f. 'hesitation' 260 inunda 'inonder1 : inundasju(n), f. 'inondation' iriga 'irriguer' : irigasju(n), f. 'irrigation' kumplika 'compliquer' : kumplikasju(n), f. 'complica tion ' kundenna 'condamner' : kundennasju(n), f. 'condamna- tion' repara 'reparer' : reparasju(n), f. 'reparation' separa 'separer' : separasju(n), f. 'separation' Vb. stems with no AA aprUba 'approuver' : aprubasju(n), f. 'approbation' , , 41 kulura 'colorer' : kulurasju(n), f. 'coloration' The primitives are Fr. vbs. in, e.g. kuntinqasju(n), f. 'continuation' (cf. nat. vb. kuntyna 'continuer'), sittias ju(n), f. 'situation'. The variant forms /-enso 0 0 -isenso/ occur with a limited number of vb. primitives. The derivational pattern 41It is not possible from the few data available to determine the status of the forms /-ysjun/ and /-isjun/ in relation to the suffix /-asjun/. The primitives with which these forms occur are either homophonous with or phonemical- ly adapted from Fr. vbs. The indigenous character of the derivational patterns is therefore dubious. Pending con clusive data, items such as the following appear to be learned loans: diminysjun, f. 'diminution' (vs. nat. vb. deminga 'diminuer'), kunstitysjun, f. 'constitution', puzisjun, f. 'position', prupuzisjun, f. 'proposition', pynisjun, f. 'punition'. 261 appears to call for /-enso/ after e-vbs. and /-isenso/ after is-vbs. The derivates are fem. N: Vb. bases plus /-enso/ with no AA krese 'croitre' : kresenso, f. 1croissance' kunese Irr. 'connaitre1 : kunesenso, f. 'connaissance' with AA nCse 'naitre1 : nesenso, f. 'naissance' Vb. bases plus /-isenso/ with no AA — v ! g - f , rezwi reji zwis 'jouir' : zwisenso, f. 'jouissance' The N sufrenso, f. 'souffrance' is an example of a loan-word which does not follow the pattern just described. The primitive is the Fr. vb. /sufrir/ souffrir. not Dzcs. sufrIs 'souffrir' which belongs to a different vb. class. (5) (-omen2] 'state of being' forms deverbal jnasc. N: Vb. bases with no AA estuna 'etonner' : estunomen, m. 1etonnement1 Vb. stems with no AA rezwi1 3 're jouir' : rezwisenso, f. ' re jouissance' 262 abate 'abattre' : abatomen, m. 'abattement1 embarasa 1embarrasser1 : embarasomen, m. 'embarras' eskawfa 'echauffer' : eskawfomen, m. 1echauffement' i i 42 kuntenta 'contenter1 : kuntentomen, m. 1contentement1 ept[izomen, m. 'epuisement1, built on a Fr. base, is one of many loan-words with more or less abstract meaning. (6) {-adoi} 'outcome or result of' forms fem. N predominant ly on primitive vbs.: Vb. bases with no AA engyla 'enfiler' : engylado, f. 'enfilee' fyma 1fumer' : fymado, f. 'fumee1 kaga 'aller a la selle* : kagado, f. 'bouse, fiente' pUrta 'porter' : purtado, f. 'portee' zEla 'geler' : zelado, f. 'gelee' Vb. stems with no AA kujuna 'plaisanter' : kujunado, f. 'plaisanterie, bevue' ^■^With this primitive /-omen/ also occurs with the meaning "act and result of." Additional data may reveal the existence of other meaning groups not definable at present on the basis of single examples. 263 nisa 'nicher' : nisado, f. 'nichee' pisa 'uriner' : pisado, f. 'pissee' sanna 'saigner' : sannado, f. 'saignee' N bases with no AA kluko, f. 'poule couveuse' : klukado, f. 1couvee' In the form -Vt. i.e. /-at/ with a-vbs. and /-it/ with I i-vbs., this suffix forms masc. N on: Vb. bases with no AA I I bumis 'vomir1 : bumit, m. 'vomissure' eskypis 'cracher1 : eskypit, m. 'crachat' I I muzis 'moisir' : muzit, m. 'moisi' peka 'pecher' : pekat, m. 'peche' (7) {-ado2} 'capacity of', including 'contents and duration of' forms mostly denominal fem. N: N bases with no AA brasA, m. 'bras' : brasado, f. 'brassee' buko, f. 'bouche' : bukado, f. 'bouchee' mati(n), m. 'matin' : matinado, f. 'matinee' palo, f. 'pelle' : palado, f. 'pelletee' 264 v v | selOj f. 1seau1 : selado, f. 'valeur d'un seau, plein seau' with AA I asjCto, f. 'assiette' : asjetado, f. 'assiettee' Ext. N bases with no AA ann- (cf. an, m. 'an') : annado, f. 'annee' ser- (cf. se, m. 1soir') : serado, f. 'soiree' zurn- (cf. zur, m. 'jour') : zurnado, f. 1journee' N stems with no AA kareto, f. 'charrette' : karetado, f. 'charretee' Vb. bases with no AA bela 'veiller' : belado, f. 'veillee1 The allomorph /-at/, which occasionally adds pejorative connotation, forms denominal masc. N: N bases with no AA bejre, m. 'verre’ : bejrat, m. 'grand plein verre' bentre, m. 'ventre' : bentrat, m. 'plein ventre' with AA pyn, m. 'poing' (/n ® n/) : pynat, m. 'poignee' 265 (8) {-ados} 'stir or spurt', forms fem. N on: Vb. bases with no AA bufa 'souffler' : bufado, f. 'souffle' flamba 'flamber' : flambado, f. 'flambee' zigla 'rejaillir' : ziglado, f. 'rejaillissement1 Vb. stems with no AA esklajra 'eclairer' : esklajrado, f. 'eclaircie' N bases with no AA aleno, f. 'haleine' : alenado, f. 'bouffee' v v awraje, m. 'orage' : awrajado, f. 'averse' Ext. N bases with no AA marsenk- (cf. mars, m. 'mars') : marsenkado, f. 'giboulee de mars' This suffix is also found in the N /pelado/ in the fixed idiomatic phrase /a pits e a pelados/ 'a diverses reprises, tant bien que mal'. These N have no other occurrence. The primitive has not been identified. (9) {-ad04} 'series of strokes, blows, shots' forms fem. N on: 266 N bases with no AA bastu(n), m. ’baton' : bastunado, f. 'bastonnade' kanu(n), m. 'canon' : kanunado, f. 'canonnade' Ext. N bases with no AA petar- (cf. pet, m. 'pet, detonation') : petarado, 'suite de pets, de detonations' Vb. bases with no AA brEga, 'rosser' : bregado, f. 'rossee' raska 'battre, rosser' : raskado, f. 1volee de coups' (10) [-u] forms abstract fem. N denoting a quality, condi tion, or state of being from primitive adj.: Adi. bases with no AA berd 'vert' : berdu, f. 1verdeur' dusA 'doux’ : dusu, f. 'douceur' espesA 'epais' : espesu, f. 'epaisseur' fred 'froid' : fredu, f. 'froideur' 0 9 i ^ i 43 fresk 'frais' : fresku, f. 'fraicheur' 43Rivai form freskvro. f. with semantically identical suffix /-yro/. 267 44 grand 'grand' : grandu, f. 'grandeur' larze 'large' : larzu, f. 'largeur' lung 'long' : lungu, f. 1 longueur' magre 'maigre' : magru, f. 'maigreur1 niiwt ' haut' : nawtu, f. ' hauteur ' with AA /kaw/ (masc.) 'chaud' (/w : l/) : kalu, f. 'chaleur lC^L 'laid' : ledu, f. ' laideur' Allomorph /-enso/ also forms abstract fem. N. Each of these N is matched by an adj. ending in (-en(t)} (i.e. masc. /-en/, fem. /-ento/), but the reverse is not true. The primitives are either bound or free forms. Examples of both N and adj. are given below. Primary adi. bjulent 'violent' diferent 'different' entellizent 'intelligent' insulent 'insolent' inusent 'simple d'esprit' N bjulenso, f. 'violence' diferenso, f. 'difference' entellizenso, f. 'intelligence insulenso, f. 'insolence' inusenso, f. 'simplicite d'esprit' pasjent 'patient* pasjenso, f. 'patience* 44]3etter derived from the masc. form /kaw/ than from the adj. stem formula kawd. 268 Secondary adi. N abundent 'abondant1 abundenso, f. 'abondance1 V | v ( brilent 'bnllant' brilenso, f. 'brillance' rekunesent 'reconnaissant1 rekunesenso, f. 'reconnais sance ' The phonemic configuration of the base betrays the French pattern in at least some instances, e.g. the primary adj. puisent 'puissant' and N puisenso. f. 'puissance', the secondary adj. obeisent 'obeissant' (cf. the nat. vb. eskuta 'obeir') and N obeisenso. f. 'obeissance'. Allomorph /-eso/ forms de-adjectival fem. N: Adi. bases with no AA rise 'riche' : riseso, f. 'richesse' saze 'sage' : sazeso, f. 'sagesse' zwen 'jeune' : zweneso, f. 'jeunesse' with AA I V v J bjCL 'vieux' : bjeleso, f. 'vieillesse' mOL 'mou' : muleso, f. 'molesse' The productive form of this suffix is /-eso/, the adaptation of Fr. -esse. hence derivates such as adreso. f. 'adresse', (cf. nat. adj.: adret 'adroit'), biteso. f. 'vitesse' (cf. I , nat. adv. biste 'vite'), tristeso. f. 'tristesse' (cf. nat. 269 I adj. triste 'triste'). Certain abstract fem. N are derived from primitives by means of allomorphs /-tat/, /-itat/ or /-etat/. The forms /-itat/ or /-etat/ tend to occur where the phonological limitations of the language do not permit the addition of the consonant-initial from /-tat/. Examples with /-tat/: Adi. bases with no AA bu(n) 'bon' : buntat, f. 'bonte' fjer 'fier' : fjertat, f. ’fierte' kla[r] 'clair' : klartat, f. 1clarte1 ser 'cher' : sertat, f. 'cherte1 with AA b£L 'beau' (/l : w/) : bewtat, f. 'beaute'45 Bound bases with AA mi- (cf. mjez ~ mjej 1 demi' , with /j 0 0 i/) : mitat, f. 'moitie1 Examples with /-itat/: 45This suffix clearly recurs in the form /-tat/ in the N bertat. f. 'verite'. By the criteria used here, however, there is insufficient phonemic similarity between the par tial /ber-/ and the adj. /braj/ 'vrai' to class bertat as a derivate. 270 Adi. bases with no AA egal 'egal' : egalitat, f. 'egalite' nyd 'nu' : nyditat, f. 'nudite' sulide 'solide' : suliditat, f. 'solidite' trankille 'tranquille' : trankillitat, f. 'tranquil- lite ' Bgjyjid bases with no AA neses- (cf. nesesarJ 'necessaire') : nesesitat, f. 1necessite' This suffix is frequently appended to Fr. adj., e.g. matyritat, f. 'maturite' (cf. nat. adj. madvTr1 'mur'), obskyritat, f. 1obscurite' (cf. nat. adj. eskvTr1 'obscur1). Examples with /-etat/: Adi. bases with no AA pawre 'pauvre' : pawretat, f. 'pauvrete' prope ’propre' : propetat, f. 'proprete' sale 'sale' : saletat, f. 'salete' Allomorph /-izo/ occurs with a small group of adj. to form fem. N: 271 Adj.. kas.s,s with no AA , I balent ’travailleur1 : balentizo, f. 1ardeur au tra vail’ tenant 'faineant' : fenantizo, f. 'faineantise' frans 'franc' : fransizo, f. 'franchise' I 1 gurmand 'gourmand' : gurmandizo, f. 'gourmandise' rainard 'douillet' : minardizo, f. 'sensibilite excessive' A variant form /-ado/ occurs appended to the adj. sek 'sec' in the N sekado. f. 'secheresse'. (11) {_ei3 'tree or plant bearing a fruit1 forms denominal masc. N: N bases with no AA abelano, f. 'noisette' : abelane, m. 'noisetier' awrinu(n), m. 'brugnon' : awrinune, m. 'brugnonier' emmello, f. 1amande' : emmelle, m. 'amandier' < , figo, f. 'figue' : fige, m. 'figuier' , v v , iranje, m. 'orange' : iranje, m. 'oranger' kastano, f. 'chataigne' : kastane, m. 'chataignier' pero, f. 'poire' : pere, m. 'poirier' pryno, f. 'prune' : pryne, m. 1prunier1 sitru(n), m. 'citron' : sitrune, m. 'citronnier' with AA ■ 1 awrxkOt, m. 'abricot' : awrikute, m. 'abricotier' v | v , eskxlOt, m. 'noix' : eskilute, m. 'noyer' mCsplo, f. 'nefle' : mesple, m. 'neflier' nuze, f. 'noix' (/z 0 0 g/) : nuge, m. 'noyer' p£so, f. 'peche' : pese, m. 'pecher' — * — I rOzo, f. 'rose' : ruze, m. 'rosier' sirCzOj f. 'cerise' : sireze, m. 'cerisier' The form /-je/ of this suffix occurs only with Fr. bases, I e.g. the partially assimilated loan-words olibo. f. 'olive' . * 1 | olxbic. m. 'olivier'; frSzo. f. 'fraise' : frezie. m. 'fraisier'. This suffix appears in the fem. form /-ero/ with AA in kudun. m. 'coing' (/n 0 0 n/) : kudunero. f. 'cognassier'. (12) £ — es} 'place or dwelling for', forms denominal masc. N: N bases with no AA ajgo, f. 'eau' : ajge, m. 'saignee' grano, f. 'graine' : grane, m. 'grenier' kyrgo, f. 'cruche' : kyrge, m. 'evier' 273 palo, f. 'paille' : pale, m. 'pailler' pizu(n), f. 'pigeon' : pizune, m. 'pigeonnier' Allomorph /-ero/ forms fem. N on: N To&$gs with no AA kusi(n), m. 'coussin' : kusinero, f. 'taie d'oreiller' I | ^ ^ lapi(n), m. 'lapin' : lapinero, f. 'etable a lapins' with AA 1 / v / v 46 vi furmiK, f. 'fourmi' (/k 0 0 1/) : furmilero, f. 1four- miliere (habitation)' Ext. N bases with no AA tawpin- (cf. tawpo, f. 'taupe') : tawpinero, f. 'tau- piniere' (13) l“e3} 'device intended for' forms denominal masc. N: N bases with no AA eskalo, f. 'echelle' : eskale, m. 'escalier' rat, m. 'rat' : rate, m. 'piege a rats' ^This uncommon consonant alternation is due to the Fr. equivalent form. Most of the informants now discard the form /furmigero/ which shows voiced /g/ for /k/ in the base, an alternation not uncommon in native derivation. 274 telo, f. 'toile' : tele, m. 'metier a tisser1 with AA kandClo, f. 'chandelle' : kandele, m. 'chandelier' JSq m k L ba&es with no AA zuk- (cf. azuka 'rentrer au poulailler') : zuke, m. 'perchoir' In this sense the suffix /-e/ appears with a Fr. base in v I l v orele. m. 'oreiller' (cf. nat. awrelo. f. 'oreille'). Allomorph /-ero/ forms fem. N on: N bases with no AA f* 'eau' : ajgero, f. 'egout' 1 ^ I kuzino, f. 'cuisine (preparation)' : kuzinero, f. 'cuisiniere (appareil)' (14) {-o 0 0 -ado 0 0 -eso) 'female' forms fem. N matching the masc. N derived by means of {zero 0 0 -e 0 0 -at 0 0 -aje) 'male' . The following examples are arranged according to the shape of the morpheme 'male'. Allomorphic alternations affect the stems only when followed by a suffix meaning 'male'. Deri- vates with /-o/, /-e/, and zero are stressed on the same syllable as the primitives. 275 Primitive Male Female Bound bases /-$/ /-o/ with no AA 1 bezin- 1 bezi(n), m. 'voisin' 1 bezino, f. 'voisine' gat- gat, m. 1 chat1 gato, f. 'chatte1 git- 1 git, m. 'canard' 1 gito, f. 'cane' kan- ka[n], m. 'chien' kano, f. 'chienne' 1 kuzin- kuzi(n), m. 'cousin' 1 kuzino, f. 'cousine' nenet- nenet, m. 'bebe' neneto, f. 'bebe' pjot- pjot, m. 'dindon' pjoto, f. 'dinde' l pulin- puli(n), m. 'poulin' 1 pulino, f. 'pouliche with AA v 1 aneL- anet, m. 'agneau' anelo, f. 'agnelle' bastard- bastar, m. 'batard' bastardo, f. 'ba- tarde' beteL- betet, m. 'veau' betelo, f. 'femelie du veau' f iL- fil, m. ' fils ' filo, f. 'fille' marsand- marsan, m. 'marchand' marsando, f . 'mar ch ande nebud- 0 nebut, m. 'neveu' nebudo, f. 'niece' nobJ- nobi, m. 'marie' nobjo, f. 'mariee' puL- put, m. 'coq' pulo, f. 'poule' Primitive Male 276 Female Bound stems with no AA pulet- with AA burdile (r ) ■ pajzant- rezent- Bound bases with no AA droll- I embesill- estr- pastr- with AA mEstr- pulet, m. 'poulet' burdile, m. 'metayer' pajzan, m. 'paysan' rezen, m. 'institu- teur ' /-e/ drolle, m. 'gar^on' embesille, m. 'imbe cile ' estre, m. 'machin, hoirane dont on a oublie le nom' pastre, m. 'berger' mestre, m. 'maitre' puleto, f. 'jeune poule' burdilero, f. 'meta- yere' pajzanto, f. 'pay- sanne1 rezento, f. 'institu- trice' /-o/, /-eso/ drollo, f. 'fille' I / erabesillo, f. 'imbe cile ' estro, f. 'machine, femme dont on a oublie le nom1 pastro, f.'bergere' l mestreso, f. 'mai- tresse' Primitive Male Female Bound stems with no AA kantajr- kantajre, m. 'chan- teur ' kantajro Bound bases /-at/, /-aje/ /-o/ with no AA a jn- ajnat, m. 'aine' ajnado, awk- awkat, m. 'jars' awko, f. guj- gujat, m. 'servi- teur, jeune homme' gujo, f. ma jn- 1 v majnaje, m. 'enfant' majnado, teuse' 47 (15) [-e[r]} 'holder of a job or performer of a task* in cluding such meanings as 'maker of, person in charge of, working with or on the premises of'. This suffix appears in the shapes: /-er-/ when followed by /-o/ 'female', /~e/ when followed by zero 'male'. The derivates are denominal masc. or fem. N according to the sex of the referent. The primitives are: 47prom a descriptive standpoint it is also possible to regard ajnad- as the base morpheme, in which case the suf fix 'female' is /-o/ and the suffix 'male' zero with auto matic unvoicing /-d-/ : /-t/. Historically, the stem ajnad- is a compound, but it is no longer apprehended as such. 278 N bases with no AA krabo, f. 'chevre1 : krabe, m. -ero, f. 'eleveur, -euse de moutons1 1 / I kuzino, f. 'cuisine (preparation)' : kuzinero, f. cuisiniere, cordon-bleu' prizu(n), f. 'prison' : prizune, m. 'prisonnier' v 1 v | zardi(n)j m. 'jardin' : zardine, m. 'jardinier' with AA byw, m. 'bceuf' (/y ® 0/) : bwe, m. 'bouvier' esklOt, m. 'sabot' (/t ° ° p/) : esklupe, m. 'sabotier' Ext. N bases with AA 1 1 I burdil- (cf. bOrdo, f. 'metairie') : burdile, m. -ero, f. 'metayer, -ere1 . N stems with no AA kareto, f. 'charrette' : karete, m. 'charretier' I , muli(n), m. 'moulin' : muline, m. 'meunier' This suffix corresponds to Fr. -er. -ere and appears in loan-words like buse. m. 'boucher', bulanie. m. 'boulanger'. The allomorph /-je(r)/ yielding masc. /-je/, fem. /-jero/ is found with: 279 N bases with no AA kantu(n), m. 'cote, bord' : kantunje, m. 1cantonnier' peryko, f. 'chevelure' : perykje, m. 'coiffeur' with AA eskOlo, f. 'ecole' : eskulje, m., -cro, f. 'ecolier, -iere' This suffix corresponds to Fr. -ier.-iere in loan-words like ubr i e. m. 'ouvrier' (cf. the nat. base vbro. f. 'oeuvre'). The allomorph /-ajr-/, yielding masc. /-ajre/, fem. /-ajro/, occurs with: N bases with no AA kampano, f. 'cloche' : kampanajre, m., -ajro, f. 'sonneur, -euse de cloches' v v p^lo, f. 'vieux chiffons' : pelajre, m. 'chiffonnier' with AA ■ i karjCro, f. 'chaise' : karjerajre, m. 'racommodeur de chaises' (16) {-ajr-} 'agentive' is a word-medial affix which occurs before /-e/ 'male' or /-o/ 'female'. The bi-morphemic suf fixes /-ajre/, /-ajro/ form deverbal N of matching genders on: 280 Vb- bases with no AA kanta 'chanter' : kantajre, m. 'chanteur' laba 'laver’ : labajro, f. 1laveuse1 lawra 'labourer' : lawrajre, m. 'laboureur' marida 'marier' : maridajre, m. 'intermediaire du mariage' sEga 'faucher' : segajre, m. 'faucheur' Vb. stems with no AA brena 'vendanger' : brenajro, f. 'vendangeuse' kasa 'chasser' : kasajre, m. 'chasseur' peska 'pecher' : peskajre, m. 'pecheur' resEga 'scier' : resegajre, m. 'scieur' zUga 'jouer' : zugajre, m. 'joueur' The allomorphs /-yr 0 0 -yz-/ occur: /-yz-/ in the presence of /~o/ 'female', /-yr/ before zero 'male'. The primitives are: Vb. bases with no AA prumena 'promener' : prumenyr, m. 1promeneur' pUrta 'porter' : purtyr, m., -yzo, f. 'porteur, -euse' ramuna 'ramoner1 : ramunyr, m. 'ramoneur' 281 sawba 'sauver' : sawbyr, m. 1sauveur1 Vb. stems with no AA sawta 'sauter1 : sawtyr, m., -yzo, f. 'sauteur, -euse' v I v , trabala 'travailler' : trabalyr, m. 1travailleur' The derivational pattern with /-yr/, /-yzo/, which is paral lel to the Fr. pattern with -eur. -euse. is gaining ground. The loan-word kwafvr. m., -vzo. f. 'coiffeur, -euse' based on the Fr. vb. coiffer (cf. Dzcs. kUfa) is a case in point. The allomorph /-en(t)/, yielding masc. /-en/, fem. /-ento/, occurs with Vb. bases with no AA plaNe 'plaindre' : planen, m., -ento, f. 'plaignant, -ante' rezis 'regir' : rezen, m., -ento, f. 'instituteur, -trice' (17) {-ade} means 'makeshift utensil, aid'. The various allomorphs of this morpheme yield deverbal N. /-ade/ forms masc. N on: Vb. bases with no AA debana 1devider' : debanade, m. 'devidoir' 282 kapEla 'couvrir' : kapelade, m. 'sorte de couvercle1 laba 'laver1 : labade, m. 'lavoir1 raaka 'meurtrir, battre' : makade, m. 'battoir' tira 1tirer' : tirade, m. 'timon1 /-adu/ forms masc. N on: Vb. bases with no AA estaka 1attacher' : estakadu, m. 'objet quelconque pour attacher' kaga 1aller a la selle' : kagadu, m. 'cabinets' i 48 kapEla 'couvrir' : kapeladu, m. 'sorte de couvercle' muka 'moucher' : mukadu, m. 'mouchoir1 Allomorphs of limited occurrence include: /-adero/, /-adiso/, which yield fem. N, and /-An/, /-ew/, /-in/ which yield masc. N from: Vb. bases with no AA — I — 1 bara 'fermer' : baradiso, f. 'cloture' liga 'lier' : ligan, m. 'lien de paille tresse* rulla 'depiquer au rouleau' : rullew, m. 'rouleau' tUka 'toucher, piquer' : tukadero, f. 'aiguillon' Q I ^°The optional variant of kapelade. m. 283 with AA I I mOle 'moudre' : muli(n), m. 'moulin' ■Vfe.— .stems with no AA , I sala 'saler' : sali(n), m. 'saloir' Ext, vb. bases with no AA tumbar- (cf. tumba 'tomber') : tuinbarew, m. ' tombereau' (18) {-et} meaning 'implement' is distinct from the above forms in that this suffix denotes more elaborate and ade quate tools. The derivates are deverbal masc. N; Vb. bases with no AA bufa 'souffler' : bufet, m. 'soufflet' cyka 'sucer' : cyket, m. 'sucette' siwla 'siffler' : siwlet, m. 'sifflet' This morpheme appears as zero, replacing /-a/ with the vb. base estaka 'attacher' : estaK, m. 'chaine d'etable' and the vb. stem embvka 'gaver' : embyK, m. 'gorgeoir'. These derivates have recessive stress, i.e. the nuclei are stressed. The preceding suffixes have given way to the productive 284 form /-wer/, the adaptation of Fr. -oir. which is added to primitive vbs. in their Fr. form. Some examples are: Fr. /muse/ 'moucher' (cf. nat. muka) : muswer,^9 m. 'mouchoir' Fr. /prese/ 'presser1 (> prEsa) : preswer, m. 'pres- soir 1 Fr. /rakle/ 'racier' (cf. nat. raska) : raklwer, m. 'racloir' Fr. /seme/ 'semer' (cf. nat. samena) : semwer, m. 1semoir' The names of modern machines are loan-words made up of the suffix /-yzo/, adapted from Fr. -use. appended to Fr. primi tive vbs., e.g. batyzo, f. 'batteuse', mwas°nyzo, f. 'mois- sonneuse', liyzo, f. 'lieuse' (cf. nat. vb. liga 'lier'). (19) A minus feature, the replacement of the vb.-final vowel /-a/ by zero (with relevant adjustments of the last consonant) signifies 'product or concrete result of a proc ess or completed action'. The derivates are masc. N with recessive stress: Vb. bases with no AA I / I / eskisa 'dechirer' : eskis, m. 'dechirure' 4^This form is ousting the Dzcs. word mukadu. m. 285 esklata 'eclater' : esklat, m. 'eclat' tryka 'frapper1 : tryK, m. 'coup' with AA empEwta 'allonger, ajouter' : empew, m. 'ajout' krida 'crier' : krit, m. 'cri' 0 7 plE^a 'plier' : plek, m. 'pli' resEga 'scier1 : resek, m. 'sciure de bois' O trawka 'trouer' : traw, m. 'trou' Vb. stems with no AA eskapita ’eteter' : eskapit, m. 'cime coupee du mais' This morpheme has the overt form /-o/ replacing /-a/ and forming deverbal fem. N which are stressed on the nuclei: Vb. bases with no AA muna 'faire la moue1 : muno, f. 'moue' rantra 'repriser' : rantro, f. 'reprise' raja 'rayer' : rajo, f. 'raie' trawla 'pietiner' : trawlo, f. 'trace de pieds' The allomorphs /-adyro 0 0 -yro/ yield deverbal fem. N: 286 Vb. bases with no AA raska 'racier' : raskadyro, f. 'raclure' sewda 'souder1 : sewdyro, f. 1soudure (reparation)' with AA I , mOle 'moudre' : mulyro, f. 'moulure' Vb. stems with no AA kUfa 'coiffer' : kufadyro, f. 'coiffure' v (20) {-iklo} 'collective' forms fem. N with frequent pejora tive connotation on: N bases with no AA baro, f. 'barre' : baralo, f. 'barriere, palissade’ with AA fZr, m. ' fer' (/r ° ° r/) : feralo, f. 'ferraille' N stems with no AA I I v pulo, f. 'poule' : pulalo, f. 'volaille' Bound bases with no AA v marm- (cf. marmot, m. 'marmot') : marmalo, f. 'mar- maille' 287 Adi. bases with no AA v menyd ’menu' : menydalo, f. 'menuaille' Allomorph /-aje/ forms masc. N without emotional con notation on: N stems with no AA t | v bezi(n), m. 'voisin' : bezinaje, m. 1voisinage' N bases with no AA plymOj f. 'plume' : plymaje, m. 'plumage' Vb. bases with no AA atEla 'atteler' : atelaje, m. 'attelage' utilaie. m. 'outillage' is built on the Fr. base outil (cf. nat. N vtis P m.). Allomorph /-adis/ forms masc. N expressing contempt on Vb. bases with no AA V | V | farfula 'tourner sens dessus dessous' : farfuladxs, m. 'desordre confus' 288 Vb. stems with no AA 1 abareza 'meler1 : abarezadis, m. 'melange' The other variants are found in a few derivates: /-a/, 1 I I 1 /-is/, /-omen/ which yield masc. N and /-ado/, /-ino/, /-o/ (with stress on the base) which form fem. N. The resultant forms are collective nouns. The primitives may be: N bases with no AA I | berme, m. 'ver' : bermino, f. 'vermine' fryt, m. 'fruit' : fryto, f. 'fruits' tewle, m. 'tuile' : tewlado, f. 'toiture' with AA bCstjo, f. 'bete' : bestja, m. 'betail' Vb. bases with no AA V | v ( A abila 'habiller' : abilomen, m. 'vetements' arnesa 'harnacher' : arnesis, m. 'harnais' (21) {-aje} denotes an action as well as its result. This suffix remains productive through the agency of Fr. -age. The derivates are deverbal masc. N: 289 Vb. bases with no AA afyta 'aiguiser1 : afytaje, m. 'aiguisage' , v atEla 'atteler' : atelaje, m. 'attelage' I I v drEsa 'dresser' : dresaje, m. 'dressage' I I v raarida 'marier' : maridaje, m. 'mariage' I I v munta 'monter1 : muntaje, m. 'montage' 1 i i 1 ^ i i pasa 'passer' : pasaje, m. 'passage' ^ _ I v repasa 'repasser' : repasaje, m. 'repassage' I I v tira 'tirer' : tiraje, m. 'tirage' yza 'user' : yzaje, m. 'usage' Vb. stems with no AA v i . v i v defEla 'effeuiller' : defelaje, m. 'effeuillage' degryna 'egrener' : degrynaje, m. 'egrenage' sawkla 'sarcler' : sawklaje, m. 'sarclage' I I V sylfata 'sulfater' : sylfataje, m. 'sulfatage' zardina 'jardiner' : zardinaje, m. 'jardinage' (22) {-Vdo} 'act of forms deverbal fem. N. The vowel V is determined by the vb. class of the primitive: /-ado/, /-ido/ occur after a-vbs. and i-vbs. respectively. The primitives are: 290 Vb. bases with no AA brama 'beugler' : bramado, f. 'beuglement' ' ~ i , brandi 'secouer' : brandido, f. 'secouee' entra 1entrer1 : entrado, f. 'entree (de la nuit)1 eskapa 'echapper' : eskapado, f. 'escapade' lEka 'lecher' : lekado, f. 'lechee' ' C 5 1 mentis 'mentir' : mentido, f. 'raensonge1 raunta 'monter' : muntado, f. 'montee' nEga 'noyer' : negado, f. 'noyade' reginna 'caracoler' : reginnado, f. 'caracole' sEga 'faucher' : segado, f. 'fauchaison' tumba 'tomber' : tumbado, f. 'chute, tombee' Vb. stems with no AA sabuna 'savonner' : sabunado, f. 'petite lessive' In the loan-blend turnado. f. 'tournee d'un commer^ant', this suffix appears with the base /turna/, adapted from Fr. tourner and semantically distinct from Dzcs. tUrna 'reve- nir' . (23) {-Vt} 'undergoer of an action (person or specific product)' forms deverbal masc. N homophonous with the masc. 291 past part, of the primitive vbs. The forms /-at/, /-it/ I I occur with a-vbs. and i-vbs. respectively: Vb. bases with no AA akyza 'accuser' : akyzat, m. 'accuse' awmenta 'augmenter' : awmentat, m. 'augmentation (terme de tricot)' V| V | byLi 'bouillir' : bylit, m. 'viande du pot-au-feu' deminga 'diminuer' : demingat, m. 'diminution (terme de tricot)' embita 'inviter' : embitat, m. 'invite' estUfa 'etouffer' : estufat, m. 'sauce de bceuf au vin' nEga 'noyer' : negat, m . 'noye' — 1 G ^ — 1 A rustis 'rotir' : rustit, m. 'roti' Vb. stems with no AA sala 'saler' : salat, m. 'pore sale' (24) {-adyro} expresses an injury or wound as well as the inflicting blow, stroke or thrust. This suffix forms de verbal fem. N on: Vb. bases with no AA byrla 'bruler' : byrladyro, f. 'brulure' 292 kupa 'couper1 : kupadyro, f. 'coupure' maka 'meurtrir, ecraser1 : makadyro, f. 'meurtrissure' Allomorph /-ado/ also forms deverbal fern. N on: Vb. bases with no AA gafa 'mordre' : gafado, f. 'morsure' grawpina 'egratigner' : grawpinado, f. 1egratignure' pesyga 'pincer' : pesygado, f. 'pincee' Allomorph /-at/ forms masc. N. The primitives desig nate either the instrument which deals the blow or the localization of the blow: N bases with no AA pato, f. 'patte' : patat, m. 'coupj de la main ou de la patte' with AA 1 I /v kOfo, f. 'coiffe' : kufat, m. 'coup sur la tete' Bound bases with no AA burm- (cf. burmelo, f. 'morve') : burmat, m. 'coup sur le nez, sur le visage' The form /-yro/ found in pika 'piquer' : pikvro. f. 'piqure d'insecte, traitement medical' corresponds to Fr. -ure and 293 yields borrowed derivates in place of native formations, e.g. Fr. /blese/ 'blesser1 (cf. nat. blasa) : blesyro, f. 'blessure' Fr. /kase/ 'casser1 (cf. nat. kupa 1couper, casser1) : kasyro, f. 'cassure' (25) {-aw} 'device or location serving a specific purpose1 forms masc. N chiefly on primitive N: N bases with no AA kambo, f. 'jambe' : kambaw, m. 'barre pour ecarter les pattes d'un cochon1 with AA dit. m. 'doigt' : didaw, m. 1de a coudre1 v ^ fjCro, f. 1foire' : fjeraw, m. 'foirail1 I I pOrto, f. 'porte' : purtaw, m. 'portail' _ _ * _ | v rOdo, f. 'roue' : rudaw, m. 'orniere' N stems with no AA kareto, f. 'charrette' : karetaw, m. 'chemin de tra verse Vfr. with AA kaca 'macher' {'/c 0 0 s/) : kasaw, m. 'dent molaire' 294 (26) {-at}, a formative of masc. nouns occasionally with specialized meaning, forms masc. N on fern. N: N bases with no AA ajgo, f. 'eau' : ajgat, m. 'pluie d'abat' ■ i kuzino, f. 'cuisine' : kuzinat, m. 'ensemble de legumes pour une soupe' V y palo, f. 'paille' : palat, m. 'litiere' with AA kOsto, f. 'cote' : kustat, m. 'cote (partie du corps)' Allomorph /-ero/ forms fern. N on fern. N: N bases with no AA garbo, f. 'gerbe' : garbero, f. 'gerbier1 palo, f. 'paille' : palero, f. 'meule de paille' with AA 1 V V » furmiK, f. ' fourmi' (/k 0 0 1/) : furmilero, f. ' four- miliere, ensemble de fourmis' (27) {-eno} 'collective numeral' forms fern. N denoting the approximate number of items in a collection. This suffix is added to certain cardinal numerals and does not appear with compound cardinals: 295 Numeral bases with no AA du^e 'douze' : du/eno, f. 1douzaine1 • 1 t kinze 1quinze1 : kinzeno, f. 1quinzaine' trento 1trente1 : trenteno3 f. 1trentaine' kranto 1quarante' : kranteno, f. 1quarantaine' Ext, numeral bases with no AA wejt- (cf. wej 'huit') : wejteno, f. 'huitaine' I , bint- (cf. bin 'vingt') : binteno, f. 1vingtaine' This morpheme appears in shape zero with the N sen, m. i 50 'centaine’, homophonous with the cardinal base sen 'cent1 . i 51 (28) {-aso} 'container made of' forms denominal fern. N: N bases with no AA burOj f. 'bourre' : buraso, f. 'maillot pour nouveau- ne ' 50Unlike the numeral, the masc. N with collective meaning occurs preceded by the def. art. vn 'un'. SlThe suffix {-(t)jero} in sense 'container for . . .' is appended to Dzcs. N bases like sawso. f. : sawsjero, f. 'sauciere'? tabat. m. 'tabac' : tabatjero, f. 'tabatiere'. However, derivates such as kafetiero. f. 'cafetiere' from kafe, m. 'cafe', wherein the suffix occurs in shape /-tjero/ after stressed vowel, suggest that both the form of the suffix and the derivational ratio are borrowed from Fr. 296 I v v j palo, f. 'paille' : palaso, f. 'paillasse' with AA f&r, f. ' fer' (/r 0 0 ?/) : feraso, f. 'pelle a feu' Allomorph /-ino/ appears with: N bases with AA I _ __ I tCro, f. 'terre1 : terino, f. 'terrine' Ext. N bases with no AA sedas- (cf. sedo, f. 'soie') : sedasino, f. 'tamis' (29) {-yro} 'merchandise for* forms deverbal fern. N: Vb. bases with no AA dawra 'dorer' : dawryro, f. 'dorure1 fyma 'fumer' : fymyro, f. 'fumure' lEba 'lever' : lebyro, f. 'levure' pintra 'peindre' : pintryro, f. 'peinture' sewda 'souder' : sewdyro, f. 'soudure' (30) Unclassified suffixes The following is a list of residual forms which, ac cording to this analysis, are isolatable suffixes, or pos 297 sibly historical components. They have rare occurrence, so that pending additional derivational data their meanings are not definable. Suffixes < Primitives ; Derived F -ado N pan, m. 'pain' panado, f. 'panure' -ado Ext. N base fengl- (cf. ren, renglado, f. 'rangee' m . 'rang 1) -an(t) N paisA, m. 'pays' pajzan, m., -anto, f. (/i 0 0 j/) 'paysan, -anne' -at N abeske, m. abeskat, m. 'eveche' I / A | eveque -e Vb. plaZe, 'plaire' plaze, m. 'plaisir' -ego N ma(n), f. 'main' manego, f. 1manche d'araire’ -el N kaP, m. 'tete' kabel, m. 'epi de ble' -el Bound base par- (cf. aparja parel, m. 'paire' 'accoupler') « y ^ V -elo N kaP, m. 'tete' kabelo, f. 'cime' -en Ext. N base ser- (cf. serado, seren, m. 'rosee du soir' f . 'soiree') -e Vb. luga 'louer' luge, m. 'loyer' -elo Bound base burm- (cf. burmat, m. 'coup sur le burmelo, f. 'morve' nez' ) 298 Suffixes Primitives Derived N -ero Vb. kaga 'aller a la selle1 kagero, f. 'colique' -erto Bound base kyb- kyberto, f. 1couverture' -estre Bound base kyb- kybestre, m. 'couvercle' -et Vb. nuza 'nouer' nuzet, m. 'noeud' — e t N ka£, m. 'tete' kapet, m. 'chapeau' 1 -ino Vb. 1 g langi 1languir' 1 langino, f. 'langueur' -isi Vb. 1 serbi 'servir1 1 serbisi, m. 'service' i o- H< 0 N kaP, m. 'tete' v * kabolo, f. 'epi de mais' -6no N kar, f. 'viande' (r 0 0 r) karono, f. 'charogne' ■ -u Vb. senti 'sentir1 sentu, f. 'odeur, sen- teur ' ■ -u Vb. syza 1 suer 1 syzu, f. 'sueur' minus feature Vb. 1 sjCge 1suivre1 O sj ek 'affilee' 4.3 . 2 . 2 . Verbs Verb formation by means of various suffixes represented i 5 2 by the phonemic shape /-a/ is particularly widespread. ^Verb endings such as this in other Romance languages have been interpreted in a variety of ways. The problem is complicated by disagreement as to what kinds of forms con stitute words and, too, by the obvious absence of clear-cut distinction between inflectional and derivational suffixes. 299 The derivational patterns involved are in general produc tive, facilitated by the existence of equivalent deriva tional relations in French. Dzcs. /-a/ is equated with the active French formative -er. Not considered here are loan- derivates which exhibit morphophonemic changes foreign to the dialect; for instance, the vb. utila 'outiller' is built on the Fr. N base /uti/ 'outil' augmented by /!/, not on the I Dzcs. N base vtis~. m. 'outil', which is unproductive. The lists of examples are not intended to be exhaustive. To summarize, the final vowel of an infinitive is analyzed as an integral part of the vb. stem by some, as all or part of the inflection by others, the principal criterion being descriptive convenience. The latter analysts either do not count vb. suffixes comparable to Dzcs. /-a/ among deriva tional morphemes, or else they regard pairs like Latin corona (N) : coronare (vb.) as instances of zero-derivation. Working with Dzcs., I am of the opinion that the inventory of the derivational suffixes of this dialect would be in complete if the vb. formative /-a/, the most prevalent of all suffixes, were to be left out. Not all languages use this type of vb. formation or use it so extensively. I re gard Dzcs. sek (adj.) 'sec' : seka (vb.) 'secher' (cf. Eng lish bright : brighten) as two free forms belonging to dis tinct form-classes; between the two words there is a formal as well as a semantic difference: the vb. derived from the adj. sek 'dry' means 'become dry'. It should be recognized that vb.-final /-a/ has a dual function. It is often a derivational suffix, so that a vb. stem, like any other stem, may be a base or a derived stem. Whether it is treated as an infinitive inflection or as part of the stem — this is the interpretation preferred in this study— the vowel /-a/ also serves as a convenient indicator of a par ticular conjugational system, i.e. a marker of a sub-class of vbs. 300 (1) [-a].} meaning 'use as instrument or implement' forms denominal derivates: N bases with no AA balezo, f. 'balai' : baleza 'balayer' v v ^ dalo, f. 'faux' : dala 'faucher' diske, m. 'disque' : diska 'passer le disque' esplingo, f. 'epingle' : esplinga 1epingler' sawklo, f. 'sarcloir' : sawkla 'sarcler' , I sisklet, m. 'loquet' : siskleta 'fermer au loquet' with AA brEs^, m. 'berceau' : brEsa 'bercer' V y farul, m. ' verrou' (/l 0 0 l/) : farula ' verrouiller' klaW, f. ' cle' : klaba 'fermer a cle' 1 s I mEstre, m. 'metre' : mestra 'metrer' 1 myzCt, m. 'museau' (/t » l/) : myzEla 'museler' pesA, m. 'poids' : pEza 'peser' telefOno, m. 'telephone' : telefuna 'telephoner' Ext. N bases with no AA arpent- (cf. arpen, m. 'arpent’) : arpenta 1arpenter' pencen- (cf. pence, m. 'peigne') : pencena 'peigner' 301 N stems with no AA ligaso, f. 'lien de gerbe usage1 : ligasa 1attacher grossierement' This morpheme appears as /-ela/ with the N base klaW. m. 'clou' (with AA) : klabela 1 clouer' and as /-eza/. palo. f. 'pelle' : paleza 1pelieter 1 . (2) {-a2} 'perform an action or task, partake in . . forms denominal derivates: N bases wi th no AA I , ba§0lo, f. 'vaisselle' : baSela 'faire la vaisselle’ kaso, f. 'chasse' : kasa 'chasser' kisto, f. 'quete' : kista 'queter' I ✓ | kuzino, f. 'cuisine (preparation)' : kuzina 'faire la cuisine' nado, f. 'nage' : nada 'nager' with AA I I galOP, m. 'galop' : galUpa 'galoper' kalOto, f. 'gifle' : kalUta 'gifler' salyt, m. 'salut' : salyda 'saluer' I . v» - v | trabal, m. 'travail' (/l “> 1/) : trabala 'travailler' zOK, m. 'jeu' : zUga 'jouer' Ext. N bases with no AA sawt- (cf. saw, m. 'saut1) : sawta 'sauter' (3) {_a3} means 'produce, bring forth'. The derivates refer to natural phenomena such as body functions, parturition, atmospheric precipitations. The primitives are N: N bases with no AA t I liwse, m. 'Eclair' : liwsa 'faire des Eclairs' pis, m. 'urine' : pisa 'uriner' salibo, f. 'salive' : saliba 'saliver' with AA gr£llo, f. 'grele' : grElla 'greler' * , nCW, f. 'neige' : nEba 1neiger' rOt, m. 'rot' : rUta 'roter' I f tOr, m. 'glace' (/r » r/) : tufa 'glacer' N stems with no AA gatun, m. 'chaton' : gatuna 'mettre bas (chatte)' graniso, f. 'gresil' : granisa 'gresiller' I , puli(n), m. 'poulin' : pulina 'poulicher' 303 with AA kanOt, m. 'chiot' : kahuta 1raettre bas (chienne)' Bound N bases with AA anEL- (cf. anelo, f. 'agnelle') : anEla, 'mettre bas (brebis)' I , , betEL- (cf. betelo, f. 'femelie du veau') : betEla 'veler' Ext. N bases with no AA bent- (cf. ben, m. 'vent') : benta 'venter' (4) {-a4} 'apply, spread' forms denominal vbs.: N bases with no AA pebre, m. 'poivre' : pebra 'poivrer' piastre, m. 'platre' : plastra 'platrer' pyrgo, f. 'purge' : pyrga 'purger' sufre, m. 'soufre' : sufra 'soufrer' with AA al, m. 'ail' (/l < » l/) : ala 'ailler' engrEs, m. 'engrais' : engrEsa 'engraisser' I , gOlso, f. 'gousse' : gulsa 'ailler' grEso, f. 'graisse' : grEsa 'graisser' 304 saw, f. 'sel' (/w : l/) : sala 'saler' I / V / v I surel, m. 'soleil' (/I 0 0 1/) : surela 'exposer au soleil' (5) {-a5} 'affix, make use of' forms denominal vbs: N bases with no AA arpjo, f. 'doigt de poule' : arpja 'faire des em- preintes de griffes' petasA, m. 'chiffon, piece' : petasa 'rapiecer' sewkle, m. 'cerceau' : sewkla 'cercler' with AA f£r, m. 'fer' (/r < » r/) : fera 'ferrer' fiw, m. 'fil' (/w : l/) : fila 'filer' meko, f. 'meche' (/k « s/) : mesa 'mecher' Ext. N bases with no AA lard- (cf. lar, m. 'lard') : larda 'larder' plant- (cf. plan, m. 'plant') : planta 'planter' with AA plumb- (cf. plun, m. Aplomb') (/n : m/) : plumba 'plomber' N stems with no AA bu/u(n), m. 'bouchon' : bu/una 'boucher une bouteille' 305 (6) £-ae} 'manifest, display' forms denominal vbs.: N bases with no AA embezo, f. 'envie' : embeza 1envier1 kaket, m. 'verbiage' : kaketa 'caqueter' kankan, m. 'medisance' : kankana 'medire' perdun, m. 'pardon' : perduna 'pardonner' I , prufit, m. 'profit' : prufita 'profiter' zyn, m. 'jeune' : zyna 'jeuner' with AA 1 v I reprOse, m. 'reproche' : reprUsa 'reprocher' (7) {-a7} 'gather . . . harvest . . .' forms denominal vbs. N bases with no AA breno, f. 'vendange' : brena 'vendanger' bryK, m. 'champignon' : bryka 'aller chercher des champignons' grano, f. 'gerbe' : grana 'glaner' N stems with AA I I gaspOt, m. 'grappillon' : gaspUta 'grappillonner' The form /-eza/ occurs with the N base fen, m. 'foin' : feneza 'faner'. 306 (8) {-a8} 'act as . . forms denominal vbs.: N bases with no AA zyje, m. 'juge' : zyja 1juger' with AA pilOto, m. 'pilote' : pilUta 'piloter1 Bound N bases with no AA trimar(d)- (cf. trimardo, f. mendiante) : trimarda 'mendier' i 5 3 (9) {-ae} 'perform an action associated with' forms vbs. mostly on N primitives: N bases with no AA estumak, m. 'estomac1 : estumaka 'bouleverser' kargo, f. 'charge' : karga 'charger' 53The difficulties encountered in defining the meanings of vbs. derived from N in various languages are well known, hence the necessity of at least one broadly worded gloss such as this to encompass what is here a relatively small number of left-over items. The semantic classification of suffixes yielding denominal vbs. has often been abandoned by students of languages such as English and French, one of the main reasons being that many such vbs. have multiple meanings. While this is true of literary languages, it is noteworthy that in this dialect the number of vbs. which belong to several sense-groups appears to be limited. plaso, f. 'place' : plasa 'placer' supOj f. 1soupe' : supa 'souper' with AA l I fagOt, m. 'fagot' : fagUta 'mal habiller* kOfo, f. 'coiffe, coiffure' : kUfa 'coiffer' ky(n), m. 'coin' (/n 0 0 n/) : kyna 'enfoncer, tasser • , trab£s, m. 'travers1 : trabEsa 'traverser' Ext. N bases with no AA gus-t'', m. ’gout’ : gusta 'gouter' Ext, prep, bases with no AA I , dintr- (cf. din 'dans') : dintra '(r)entrer' I (10) (a- . . . -i) in sense 'impart a characteristic, i. cause to become . . ., make . . .' forms transitive vbs. essentially on adj.: Adi. bases with no AA bastard 'batard' : abastardis 'abatardir' dus'' 1 doux1 : adusis ’ adoucir ’ I 1 q grand 'grand' : agrandi 'agrandir' larze 'large' : alarzis 'elargir' 308 — I — ' c rund 'rond' : arundi 'arrondir' Alternate forms of this affix which occur in complementation with individual primitives are: /a- . . . -a/, /en- . . .-a/ /en- . . . -a/, /ra . . . -i/,5^ /re . . . -i/, /es- . . . , /es- . . . -i/. Before bilabials the prefix /en-/ appears as /em-/. The primitives may be: Add. bases with no AA I _ 1 o fred 'froid' : refredi5 'refroidir1 0 I ^ * S A fresk 'frais' : rafreski 'rafraichir1 kla[r] 'clair' : esklaris 'eclaircir1 lung 'long' : alunga 'allonger1 nawt 'haut1 : ennawta 'hausser' 1 , petit 'petit' : apetita 'rapetisser' 5^It does not seem useful to take into account the ex tended variant /-si/ of the suffix /-i/ which occurs, for example, in /rakursi/ 'raccourcir'. Such vbs. are assumed to be loan-words on two counts. In the form /-si/, equiva lent to Fr. -cir. this suffix was found to occur only with bases like kurt 'court' which are nearly or entirely homo- phonous with their Fr. counterparts; cf. dvr 'dur' : dvrsis 'durcir'. On the other hand, the extended form -cir follows a typically Fr. pattern, that of noir : noircir. This ob servation is applicable to other extended forms of deriva tional suffixes which do not seem to belong to this dialect. Generally speaking, the Dzcs. suffixes, unlike certain Fr. suffixes (cf. the t- in Fr. biiou-terie). do not have vari ants augmented by the addition of a consonant in suffix- initial position. 309 rise 'riche' : enrisis 'enrichir' saduL 'repu' : asadula 'rassasier' surd 'sourd' : ensurda 'assourdir' with AA b£L 'beau' : embelis 'embellir' prOse 'pres, proche' : aprUsa 'approcher' Ext, adi. bases with no AA klajr- (cf. kla[r] 'clair') : esklajra 'eclairer' Adv. bases with AA maw 'mal' (/w : l/) : emmalis 'aggraver' Occasionally appended to N bases, this morpheme has the meaning 'impart the characteristic of . . .', e.g. piano, f, 'plaine' : aplana 'aplanir'; pynto, f. 'pointe' : apynta 'appointer'. The variants /-a/, /~i/ without prefix occur with: Adi. bases with no AA espes* 'epais' : espesis 'epaissir' kuntent 'content' : kuntenta 'contenter' las'" 'las' : lasa ' f atiguer' plen 'plein' : plena 'remplir' rusA 'roux' : rusis 'roussir' sale 'sale1 : salis 'salir' trempe 'mouille' : trempa 'tremper' zawne 'jaune' : zawnis 'jaunir' (culinary term) with AA abykle 'aveugle' (/k ° ° g/) : abygla 'aveugler1 1 I ^ enkj£t 'inquiet' : enkjEta 'inquieter' kawd 'chaud' (/d » f/) : kawfa 'chauffer' Adv. bases with no AA abarezo 'pele-mele' : abareza 'meler, melanger' I (11) {-i} 'show a quality, become . . turn . . .' intransitive vbs. on adj. primitives: Ad.j bases with no AA blank 'blanc' : blankis 'blanchir' , • s grand 'grand' : grandi ‘grandir1 magre 'maigre' : magris 'maigrir' negre 'noir1 : negris 'noircir' zawne 'jaune' : zawnis 'jaunir' with AA I I grOs 'gros' : grusis 'grossir' 310 forms 311 The variants /-a/, /a- . . . -a/, /ra . . . -i/, /re . . . -i/ occur with: Adj. bags? with no AA I I Q fred 'froid' : refredi 'refroidir1 0 , _ I /v fresk'' 1frais' : rafreskis 'rafraichir' lung 'long' : alunga 1allonger' mady[r] 'mur' : madyra 'murir' ! . .i petit 'petit' : apetita 'rapetisser' sek 'sec 1 : seka 1s^cher' (12) (de(s)- . . . -a} is a privative morpheme meaning 'strip . . . off, remove . . rarely 'remove from . . .' The alternate forms /de-/, /des-/, /dez-/ of this and other prefixes with vowel /e/ are distributed thus: /dez-/ before vowels, /des-/ before voiceless stops, /de-/ elsewhere. The primitives are usually N: N bases with no AA bentre, m. 'ventre' : debentra 'eventrer' I , buzigo, f. 'friche' : debuziga 'defricher' I v V | kabelo, f. 'cime' : deskabela 'couper la cime' I v v* , / kuraje, m. 'courage' : deskuraja 'decourager' 312 teko, f. 'cosse' : desteka 'dcosser' * I / tripo, f. 'tripe' : destripa 'etriper' with AA fflo, f. 'feuille' : defela 'effeuiller' I , ^ kOrno, f. 'corne1 : deskUrna 1ecorner' paisA, m. 'pays' (/i : j/) : despajza 'depayser' (in sense 'remove from . . .') Ext. N bases with no AA dent- (cf. den, f. 'dent') : dedenta 'edenter' gus-t*, m. 'gotit' : degusta 'degouter' with AA _ v> | A v rebel- (cf. rCbe, m. 'reve') : derebEla 'reveiller' N stems with AA parel, m. 'paire' (/l “ l/) : desparela 'depareiller' The form /de(s)- . . . -i/ occurs with: Adi. bases with no AA fresk'' ' frais' : defreskis 'defraichir' with AA I A 1 ^ grOs 'gros' : degrusis 'degrossir' The form /es- . . . -a/ appears with the extended N stem 313 kapit- (cf. kaP 'tdte') in eskapita 'eteter'. The form /-a/ occurs with: N bases with AA ■ , pCT, f. 'peau' : pela 'peler' Ext. N bases with no AA mazenk- (cf. mazen^ m. 'pampre') : mazenka 1epamprer' (13) {en . . . denotes movement into a given object or substance, the meaning being generally 'put into . . rarely 'put . . . into'(i.e. the primitive is acted upon). The prefix becomes /em-/ before bilabials. This morpheme forms denominal vbs.: N bases with no AA fango, f. 'boue' : enfanga 'embourber' gabjo, f. 'cage' : engabja 'mettre en cage1 kadeno, f. 1chaine' : enkadena 'enchainer' kambo, f. 'jambe' : enkamba 1enjamber' palo, f. 'paille' : empala 'empailler' (in sense 'put . . . into') puno, f. 'poigne' : empUna 'empoigner' with AA buko, f. 'bouche' (/u » y/) : embyka 'gaver' I A | A bwCto, f. 'boite' : embwEta 'emboiter' I __ , tEro, f. 'terre' : entEra 1enterrer' N stgmg with no AA buraso, f. 'maillot' : emburasa 'emmailloter' sako, f. 'sac' : ensaka 'ensacher' (14) {en- . . . -aa} 'expose to . . cover with . . . provide with . . .' forms mostly denominal vbs.: N bases with no AA 1 I fanno, f. 'farine' : enfarina 1 enfarmer' furko, f. 'fourche' : enfurka 'enfourcher' I V v I kuraje, m. 'courage' : enkuraja 1encourager' puzu(n), m. 'poison' : empuzuna 'empoisonner' with AA I . , benin, m. ' venin1 (/n < » m/) : embenima ' envenimer' brOK, m. 'epine' : s embrUka 'se planter une epine dGwte, m. 'dette1 : endEwta 'endetter' kryn, m. ' nuage' (/n «> m/) : enkryma ' ennuager' 315 Adi. bases with AA bOmjo, 'enjoleuse' : embUmja 'enjoler' The variant forms /a- . . . -a/, /es- . . . -i/, /-a/ com bine with: N bases with no AA embarasA, m. 1embarras' : embarasa 'embarrasser' fame, f. 'manque de nourriture' : afama 'affamer1 with AA ly/“, f. 'lumiere' (// “ k/) : alyka 1allumer1 Ext. N bases with AA pawr- (cf. adj. pawryg 'peureux1) : espawris 1apeurer' 0 (15) {a- . . . -a^/ 'assemble into . . .' forms primarily denominal vbs.: N bases with no AA ! , pilo, f. 'pile' : apila 'entasser1 with AA • , fagOt, m. 'fagot' : afagUta 'mettre en fagots' ■ , patOko, f. 'meule de foin' : apatUka 'mettre en tas' 316 Ext, bound bases with no AA parj- (cf. parel, m. 'paire') : aparja 'accoupler' / | / I The form /a- . . . -ela/ occurs with the N bOto. f. 'botte' (with AA) : abutEla 1botteler'. (16) {a- . . . -as} denotes a specific movement associated with the primitive N: N bases with no AA mure, m. 'museau' : amura 1tomber la face a terre1 with AA zinul, m. 'genou' (/l ® l/) : azinula 'agenouiller' Bound bases with no AA zuk- (cf. zuke, m. 'perchoir') : azuka 'rentrer au poulailler' (17) {-ez^} is an iterative suffix with an occasional pejorative tinge which forms vbs. on: N bases with no AA fango, f. 'boue' : fangeza 'patauger dans la boue' gluP, m. 'goutte qui coule' : glupeza 'couler goutte a goutte' 317 guto, f. 'goutte' : guteza 1couler1 pasA, m. 'pas' : paseza 'faire les cent pas' with AA esklOt, m. 'sabot' (/t 0 0 p/) : esklupeza 'marquer les pas avec des sabots' Ext. N bases with no AA putun- (cf. putu, m. 'baiser') : putuneza 'embrasser sans cesse' Bound bases with no AA kar- (cf. karetOj f. 'charrette1) : kareza 'charroyer' This suffix occurs preceded by the intensive suffix /-as-/ with: N bases with no AA brymo, f. 'brume' : brymaseza 'brumer' kyjo, f. 'queue' : kyjaseza 'battre de la queue' Bound bases with AA kud- (bound form of N kuire. m. 'coude') : kudaseza 'coudoyer' This suffix also follows the suffix -urla (see below). The extended variant /-eteza/ combines with the N base alo. f. 318 'aile' : aleteza 'battre des ailes, se debattre'. (18) {-ezas} 'behave in a given manner' with depreciative connotation forms vbs. on: Adi. bases with no AA feftant 'faineant' : fenanteza 'rester oisif' ryde 'rude' : rudeza 'rudoyer' with AA I y | ^ tOrt 'boiteux' : turteza 'boiter' Bound N bases with AA mar sand- 'marchand' (/s ° ° k/) : markandeza 'mar- chander'55 This suffix appears in the extended form /-teza/ with the pron. tv 'toi' : tyteza 'tutoyer' Allomorph /-a/ occurs with the adj. base krane ’ele gant, arrogant' : krana 'faire 1'arrogant' and the adj. stem kuntrarJ 'contraire* : kuntrarja 'contrarier'. (19) {-eza3} denoting change of state and approximation S^The alternation is probably due to contamination by the N markat, m. 'marche' unless the base is borrowed from Fr . 319 forms intransitive vbs. on: Adi. bases with no AA berd 'vert' : berdeza 1verdoyer1 negre 'noir' : negreza 1tirer sur le noir' fusA, 'roux' : ruseza 'devenir roux, murir (ble)' (20) Suffixes of low incidence The forms /-ala/, /-ula/, /-awna/ are pejorative and possibly frequentative in meaning. They combine with a few vbs. and N to denote an action performed carelessly and repeatedly: Vb. bases with no AA bada 'ouvrir la bouche1 : badala 'bailler' bira 1tourner' : birUla 'tournoyer' I v I tira 'tirer1 : tirala 'tirailler' with AA trambla 'trembler' (/l : <fi/) : trambUla 'trembler sur ses j ambes'56 C £ I 3°This suffix is also found in vbs. like tridUla 'fris- sonner', but the base is non-recurring so that its meaning is difficult to define. 320 with no AA pato, f. 'patte' : patawfia 'tripoter' The forms /-ina/j /-ina/j /-ni.sk/, /-nika/, /-yka/ have diminutive and frequentative value. The primitives are vbs .: Vb. bases with no AA maca 'macher1 : macyka 1machonner1 I ( tysis 'tousser1 : tysina 'toussoter1 with AA plCWre Irr. ’pleuvoir' (/-re 0 0 <//) : plebina 'pleuvoir legerement' plura 'pleurer' (/-a 0 0 //) : plurnisa 1 pleurnicher' Ext. N base with AA I , pen- (cf. p£, m. 'pied') : pennika 'agiter ses pieds' {-urla] has purely a pejorative sense. This suffix is added to primitive vbs. and occurs either alone or in con junction with the iterative suffix /-eza/: Vb. bases with no AA maca 'macher' : macUrla 'mal macher' } 321 "V | V | penja '(sus)pendre1 : penjUrla 'ne pas pendre d'aplomb pinta 'boire avec exc&s' : pinturleza 'boire sans cesse1 pisa 'uriner1 : pisurleza 'pissoter sans cesse' The suffix /-ana/ occurs once with incremental value with the vb. bate 'battre' : batana 'rosser1. 4.3.2.3. Adjectives (1) {-uz], i.e. masc. /-us/, fem. /-uzo/, in sense 'having the character of . . ., fullof . . .' forms denominal adj. N bases with no AA awra]e, m. 'orage' : awrajuz* 'orageux' 0 1 I * bilo, f. 'bile' : biluz 'qui se fait de la bile' fango. f. 'boue' : fanguz* ‘boueux1 0 firlango, f. 'fil qui pend' : firlanguz* 'dont les fils sont visibles' gino, f. 'malchance' : ginuz* 'malchanceux' I v v j gratylo, f. 'chatouille' : gratyluzA 'chatouilleux' lagano, f. 'suppuration des yeux' : laganuz^ 'dont les yeux suppurent' pusko, f. 'poussiere' : puskuz* 'poussiereux' 0 unto, f. 'honte' : untuzA 'honteux' 9 0 322 with AA benin, m. 1 venin' (/n 0 0 m/) : benimuz^ 1 veniraeux1 bisi, m. 'vice' (/i : j/) : bisjuz* 1vicieux' 1 I * glOrjo, f. 'orgueil' : glurjuz ’orgueilleux' kryn, m. 'nuage' (/n ® m/) : krymuz* 1nuageux1 pjetat, f. 'pitie1 : pjetaduzA 'pitoyable' rOno, f. 'crasse' : runuz* 'crasseux' 7 0 Ext. N bases with no AA gus-tA, m. 'gout' : gustuzA 'savoureux' nerb- (cf. ner, m. 'nerf1) : nerbuzA 'nerveux' sann- (cf. san, f. 'sang') : sannuzA 'tache de -sang' 0 sangin- (cf. san, f. 'sang') : sanginuz" 'sanguin' Bound N bases with no AA yr- (cf. bunyr, m. 1bonheur1) : yruz* 'heureux' N stems with AA burmelo, f. 'morve1 : burraeluz* 'morveux' 7 O Compound N with AA malyr, m. 'malheur' (/y e/) : maleruz* 'malheureux' 323 In sense 'full of', the suffix /-nek/ occurs with the N base berme. ra. 'ver' : bermenek 1vereux' i 57 (2) [-ajre} expresses a habitual tendency and carries such meanings as 'accustomed to, fond of'. The derivates are predominantly deverbal adj.: Vb. bases with no AA bada 'ouvrir la bouche' : badajre 'badaud' brama 'braire, pleurer' : bramajre 'braillard' duna 1donner1 : dunajre 'genereux' fringa 'courtiser' : fringajre 'leger, amoureux'58 kambja 'changer' : kambjajre 'qui aime changer' mjawla 'miauler' : mjawlajre 'miauleur' runka 'ronfler' : runkajre 'ronfleur' runa 'grogner' : runajre 'grognon' sina 'taquiner' : sinajre 'taquin' zapa 'aboyer' : zapajre 'aboyeur' “ ^'Distinguished from the homophonous N-forming suffix /-ajre/ by the criterion that, unlike the N derivates, the adj.-derivates may be preceded by such adj.-modifiers as the adv. me~i 'plus' or ola 'tres' . 58The optional variant of frinavr. an adj. used only in the masc. form. 324 Vb. stems with no AA gyla 1gueuler' : gylajre 'gueulard' putuneza 'embrasser' : putunezajre 'qui aime embrasser' N bases with no AA supo, f. 1soupe' : supajre 'qui aime la soupe’ This suffix occasionally appears with a Fr. base, e.g. in ruspetaire 'grincheux' derived from ruspeta. adapted from the colloquial Fr. vb. rouspeter (a 19th c. word); but this suffix does not seem to be productive at present. The variant form -j e[r ]. i.e. masc. /-je/, fern. /-jero/, forms denominal adj.: N bases with no AA despenso, f. 'depense' : despensje[r] 'depensier' fryt, m. 'fruit' : frytje[r] 'qui aime les fruits' kankan, m. 'medisance' : kankanje[r] 'medisant' Ext. N bases with no AA petard- (cf. petar, m. 'grand bruit') : petardje[r] 'bruyant' In fenno. f. 'femme' : fennasje 'coureur de femmes', this 325 suffix occurs preceded by the intensive suffix /-as-/. The variant form -e fr 1 also forms denominal adj.: N bases with no AA dreto, f. 'droite' : dretefr] 'droitier'59 gawso, f. 'gauche' : gawse[r] 'gaucher' Allomorph /-yr/, i.e. masc. /-yr/, fern, /-yro/, forms a few deverbal adj.: Vb. bases with no AA bula 'voler' : bulyr 'voleur' I , menti 1mentir' : mentyr 'menteur' Vb. stems with no AA blaga 'bavarder' : blagyr 'bavard, hableur' Of rare occurrence are the alternate forms -airol and -ande fr1 which yield adj. from reflexive vb. primitives, e.g. se banta 'se vanter' : bantajrol 'vantard'; se tryfa 'rire par moquerie' : tryfande[r] 'plaisantin'. (3) {-en(t)x3, i.e. masc. /-en/, fern, /-ento/, means 'having ^^Also occurs as dretj e[r]. 326 the quality or property of’. This suffix appears with bound basesj e.g. bjulent . 'violent1 (see 4.3.2.1.[10] for further examples) and with the imperfect stem of vbs. such as: with no AA amyza 'amuser1 : amyzent 1amusant' embarasa 1embarrasser' : embarasent 1embarrassant1 empUza 1imposer' : empuzent 'imposant' estuna 1etonner1 : estunent 1etonnant1 langis 'languir' : langisent 'languissant' plaZe 'plaire' : plazent 'plaisant' ragusta 'regaler' : ragustent 'allechant' sabe Irr. 'savoir1 : sabent 'savant' O (4) {-en(t)3} means 'in a condition of' and is also usually added to the imperfect stem of vbs.: with no AA • , mUri Irr. 'mourir' : murent 'raourant' pydi 'puer' : pydent 'puant' rire Irr. 'rire' : rizent 'riant' syza 'suer' : syzent 'suant' In lvzent 'luisant' this suffix is appended to the infini- I tive form of the vb. lvzis 'luire'. (5) {-yd}, i.e. masc. /~yt/, fem. /-ydo/, has the meaning 327 'characterized by' and frequently carries a derogatory connotation of excess or overabundance. This suffix forms denominal adj.: N bases with no AA awraje, m. 1orage' : awra^yd 'tres orageux' bentre, m. 'ventre' : bentryd 'ventru1 buro, f. 'bourre, poil' : bufyd 'velu' erbo, f. ' herbe' : erby<| 'herbu' furko, f. 1 fourche 1 : furky^i ' fourchu' pato, f. 'patte' : patyd 'qui a du poil aux pattes, lourdaud' with AA bO/o, f. 'bosse' : bu^yd 'bossu' I A , Os , m. 'os' : usyd 'osseux' O Ext. N bases with no AA brunk- (cf. brun, m. 'noeud dans le bois' ) : brunkyd ' qui a beaucoup de noeuds' Allomorph /-al/, without derogatory connotation, also forms denominal adj.: N bases with no AA fin, f. 'fin' : final 'final' 328 I mati(n), m. ’matin' : matinal 'matinal' (6) {-et} 'diminutive' with flattering or endearing conno tations forms adj. derived from adj.: Adi . frases. with no AA fresk* 'frais' : fresket 'plutot frais' grand 'grand' : grandet 'grandet' I I petit 'petit' : petitet 'tout petit' picu[n] 'petiot' : picunet 'tout petiot' saduL 'repu' : sadulet 'qui a bien mange' sul 'seul' : sulet 'seulet' The variants /-estin/, /-6t/, /-ulet/ are also added to a few adj. bases: with no AA grasA 'gras' : grasulet 'grassouillet' magre 'maigre' : magrestin 'maigrelet' pawre 'pauvre' : pawrot 'pauvret' (7) {-as1} 'augmentative' denotes a superlative or exces sive degree usually with strong pejorative connotation, though occasionally this suffix may also express compassion. The derivates are adj. derived from adj.: 329 Adi. bases with no AA bu(n) 'bon' : bunasA 1d'une bonte naive1 (sympathetic) fad 'fou' : fadas~ 'tout a fait fou' 0 grand 'grand' : grandas* 'trop grand' lurd 1lourd1 : lurdasA 'tres lourd' pawre 'pauvre' : pawrasA 'pauvre1 (sympathetic) rise 'riche' : risasA 'tres riche' Adi. stems with no AA patyd 'lourdaud' : patydasA 'tres lourdaud' 0 This affix has intensive value in word-medial position, e.g. in the N blagasu, m. 'moulin k paroles', in the adj. fennasje 1coureur de femmes'. (8) {-asa} denotes approximation to a quality and conveys a derogatory shade of meaning. This suffix forms adj. on adj .: Adi. bases with no AA blank 'blanc' : blankas* 'blanchatre' blund 'blond' : blundas*' ' d' un vilain blond' zawne 'jaune' : zawnas* 'jaunatre' 330 Alternants of this morpheme are: /-uz/, /-juz/ also added to adj. bases: with no AA berd 'vert' : berdjuz* 'verdatre' * — I 4% griz- 'gris' : grizuz 'grisatre' 0 negre 'noir' : negruzA 'noiratre' 0 (9) {-able} denotes possibility, including such values as 'apt to be, fit for being' where the primitive is a vb. (passive meaning) and 'capable of showing' where the primi tive is a N. This suffix appears with the imperfect stem of vbs. such as: with no AA ajma 'aimer' : ajmable 'aimable' kundenna 'condamner' : kundennable 'condamnable' minja 'manger' : minjable 'mangeable' repara 1reparer' : reparable 'reparable' with AA ais 'hair' (/i : j/) : ajsable 'haissable' N bases with no AA I I a prufit, m. 'profit' : prufitabl^ 'profitable' razu(n), f. 'raison' : razunable 'raisonnable' 331 The variant -ade fr 1. i.e. masc. /-ade/, fem. /-adero/ also occurs with primitive vbs.: with no AA marida 'marier' : maridade[r] 'en age de se marier1 plEga 'plier' : plegade[r] ' pret a plier' (10) {-an(t)}j i.e. masc. /-an/, fem. /-anto/, denotes an inherent quality or disposition. This suffix forms adj. on the imperfect stem of vbs.: with no AA ana Irr. 'aller1 : anant 'commode, facile h manier' benge Irr. 'pousser1 : benant 'plantureux, vigoureux' degala 'gaspiller' : degalant 'gaspilleur' This morpheme appears as /-adis/ in the deverbal adj. kambjadis 'de nature changeante' from kambja 'changer', The variants -iluz, i.e. masc. /-ilus/, fem. /-iluzo/, and -inuz add the connotation of an undesirable quality. The adj. are formed on: Vbs. (infinitive) with no AA I I drUmi 'dormir' : drumiluzA 'dormeur1 0 Vbs. (past part.) with AA mOr(t) 'mort' : murtiluzA 'malingre' Adi, bases with no AA blank 'blanc' : blankinuz* 'pale, qui semble blanc' 0 (11) [-ezj, i.e. masc. /-es/, fem. /-ezo/, denotes nation- 0 ality or appurtenance to a country, region, or city. This suffix forms adj. on proper names: with no AA azen 1Agen' : azenez= ’agenais' 0 dunzak 1Donzac' : dunzakez= 'donzacais' 0 franso, f. 'France' : fransez- 'francais' 0 4 Allomorph /-ol/ also forms adj. from proper names or N with no AA espano, f. 'Espagne' : espanol 'espagnol' I v v i 6 0 muntano, f. 'montagne' : muntanol 'montagnard' (12) {-arj), i.e. masc. /-ari/, fem. /-arjo/, may be re garded as an adj. formative with no clearly definable mean ing. The evidence in support of granting morphemic status to this suffix is scanty. The primitives are: 6°0ther adj. of nationality in common use appear to be loans from Fr., e.g. /aleman/ 'allemand', /alemando/ 'alle- mand'; /italjen/ 'italien', /italjeno/ 'italienne'. 333 Bound bases with no AA bulunt- (cf. buluntat3 f. 'volonte') : buluntarJ 'volontaire' neses- (cf. nesesitat, f. necessite') : nesesarJ 1necessaire1 Prep. with no AA kuntro 'contre1 : kuntrarJ 1contraire' N bases with no AA v y milu(n) 'million' : milunarJ 'millionnaire1 (13) Unclassified Suffixes The following are suffixes which function as formatives of adj. They have unique occurrence in the data and there fore no accurate definition of their meanings can be given. All denote a quality suggested by the primitives. The primitives include vbs. and N. -ajre forms the adj. kagajre 'lache, poltron' from the vb. kaqa 'aller & la selle'. -ard, i.e. masc. /-ar/, fem. /-ardo/, forms the adj. flemard I 'flemmard' (colloq. Fr.) with AA from the N fISmo. f. 'flemme' (colloq. Fr.). 3 34 -elyK, i.e. masc. /-elyk/, fem. /-elyko/, forms the adj. fredelvK 'frileux' with AA from the N fret. m. 'froid'. -yg, i.e. masc. /-yk/, fem. /-ygo/, forms the adj. pawryg 0 2. 1peureux' with AA /o 0 0 a/ from the ext. N base pawr- (cf. pow, f. 1peur'). 4.3.2.4. Adverbs {-omen} forms adv. of manner from primitive adj.: Ad-i. bases with no AA dare[r] 'dernier' : dareromen 1dernierement’ I /v | ^ fresk ’frais1 : freskomen 1fraichement1 larze 'large1 : larzomen 'largement' masant 'mechant' : masantomen 'mechamment' prunt 'prompt' : pruntomen 'promptement' pryme[r] 'premier' : prymeromen 'premierement1 I , pulid 'joli' : pulidomen 'joliment' 0 regyljefr] 'regulier' : regyljeromen 'regulierement' tal 'tel' : talomen 'tellement' zyste 'juste' : zystomen 'justement' Adj.. with no AA ajmable 'aimable' : ajmablomen 'aimablement' 335 diferent ’different* : diferentomen 'differeminent' ( i 61 endependent 'independant1 : endependentomen ’independamment' final ’final’ : finalomen 'finalement, enfin' nerbuz" 'nerveux' : nerbuzoraen 'nerveusement' 0 untuz* 1honteux' : untuzomen 1honteusement' 0 yruzA 'heureux' : yruzomen 'heureusement’ Ext, adi . bases with no AA awtr- (cf. awte 'autre') : awtromen 'autrement1 This suffix is also added to adj. in their Fr. form, e.g. Fr. /ytil/ : ytilomen 'utilement' (cf. nat. adj. vtille 'utile'). Also denoting manner, the suffix /-os/ combines with the vb. base tasta 'tater' to form the adv. tastos^^ stressed on the base. The suffix /-iw/ forms locative adv. from the bound 61-This adv. and the preceding are now being replaced by the loan adv. endeoendamen and diferamen under the in fluence of the Fr. irregular forms. 62This a<3v • occurs only in the prepositional phrase /a tastos/ 'a tatons'. The suffix /-os/ recurs in the comparable phrase /a gratipawtos/ 'a quatre pattes'; the primitive, however, is an obscure bound element which has no occurrence outside of this phrase. . 336 bases ak- (cf. ako 'cela') : akiw ’la* and .as- (cf. aso I 'ceci') : asiw 1ici'. 4.3.3. Prefixation All prefixes are unstressed elements. The stress pat tern of a primitive is not altered by the addition of a prefix. The process of prefixation seldom yields derivates of a different form-class from that of the primitive. There are allomorphic alternations only in the prefixes. 4.3.3.1. Verbs (1) (de(s)-} reverses the result of the action. For the distribution of the allomorphs /dez-/, /des-/, /de-/, see 4.3.2.2.(12) above. This prefix combines with: V)?. bases atEla 1atteler' : dezatEla 'deteler' ennyza 'ennuyer' : dezennyza 1desennuyer' fa Irr. 'faire' : defa Irr. 1defaire' kawsa 1chausser' : deskawsa 1dechausser1 kuZe 'coudre' : deskuZe 1decoudre' majla 1meler1 : demajla 1demeler' plEga 'plier' : desplEga 1deplier1 zynis 1 lier les bceuf s' : dezynis 'delier' 3 37 Vb. stems klabela 'clouer* : desklabela 'declouer' kUfa 'coiffer' : deskUfa 'decoiffer' kUla 'coller' : deskUla 'decoller' liga 'lier' : deliga 1delier' pencena 'peigner1 : despencena 'depeigner' pupa 1teter' : despupa 'sevrer' 6 3 (2) {r(e)-} in sense 'back, back again' denotes restora tion to a former state or position, or return to a former place. Before vowel-initial primitives the vowel of the prefix is deleted. The primitives are: Vb. bases apEla 'appeler' : rapEla 'rappeler' lEba 'lever' : relEba 'relever' • — I mCte ’mettre' : remCte 'remettre* Vb. stems I _ » _ adusis 'adoucir' : radusi 'radoucir' amena 'amener' : ramena 'ramener' aprUsa 'approcher' : raprUsa 'rapprocher' apUrta 'apporter' : rapUrta 'rapporter' ^^The mere repetition of an action is not signaled by a prefix but by the auxiliary verb tUrna 'a nouveau' (see phrases 5.2.1.4.I.B.[3]). 338 drEsa 'dresser' : redrEsa 'redresser' (3) {e(s)-} has intensive meaning and denotes the further ance of an action. The primitives are: Vb. bases kyra 'curer' : eskyra 'recurer' lEba 'lever' : elEba 'elever' 1 e * _ , taris 'tarir' : estaris 'epuiser' tira 'tirer' : estira 'etirer' . Vb, , -a-t.ems kawfa 'chauffer' : eskawfa 'echauffer' This prefix and the incremental suffix /-asa/ occur con comitantly in espudasa 'tailler tres long' derived from the vb. base puda 'tailler le vigne'. This prefix occurs in shape /ra-/ with the vb. stem gusta 'gouter' : ragusta 'regaler'. (4) Prefixes of low incidence The prefix /en-/ : /em-/ denotes movement away from a place, e.g. lEba 'lever' : enlEba 'enlever'; mena 'mener' : emmena 'emmener'; pUrta 'porter' : empUrta 'emporter'. The prefix /en-/ : /em-/ means 'inside' in, e.g. bara 'fermer' : embara 'enfermer'; beWre Irr. 'boire' : embeWre Irr. 'imbiber'. 339 The prefix /a-/ denotes either movement in the direc tion of the speaker or movement along with the doer, e.g. mena 'mener' : amena 1araener'; pUrta 'porter' : apUrta 'apporter'. /a-/ denotes movement downward in abate 'abattre' from bate 'battre'. /de(s)-/ has negative meaning with bases such as aorUba 1approuver' : desaprUba 'desapprouver'; espera 'esperer' : dezespera 'desesperer'. /de(s)-/ indicates the division into parts, e.g. kambja 'changer' : deskambja 'faire de la monnaie'y kupa 'couper' : deskupa 'decouper'. /de(s)-/ means 'away' in deskuca 'decoucher' from kuca 'coucher'. 4.3.3.2. Adverbs (1) {apr-} denotes approximation when combined with the < N 1 N I locative adv. akxw 'la1 : aprakiw 'par la* and asiw 'ici' : i aprasiw 'par ici'. (2) {ar-} occurring as /ar-/ <» /aryn-/ and meaning 'last' forms temporal adv. on the N bases sfi., m. ' soir' : arse 'hier soir' and an. m. 'an' : arynan 'l'an dernier'. 340 (3) {engw-} in the sense 'current' occurs once in the tem poral adv. encrwan ' cette annee' from the N base iua, m. 'an' 4.3.3.3. Adjectives {en-} is a negative prefix which is appended to: Adi. bases kapable 'capable' : enkapable 'incapable' pusible 'possible' : empusible 'impossible' zyste 'juste' : enzyste 'injuste' Adj. stems pasjent 'patient' : empasjent 'impatient' Allomorph /maw-/ occurs with: Adj. bases adret 'adroit' : mawadret 'maladroit' kuntent 'content' : mawkuntent 'mecontent' i ^64 a A oneste 'honnete' : mawonest 'malhonnete' Adj. stems anant 'facile a manier' : mawanant 'difficile a manier 4.3.3.4. Nouns Only the prefix {para-} deserves mention here. It ^ A learned word with initial unstressed /o/. 341 occurs in the sense 'device for protection against . . ■ v 65 in the common N paraplezo. m. 'parapluie', a masc. de- rivate of the fem. N base piezo 'pluie'. Other derivates containing this prefix are learned or technical loan-words, e.g. parasyto, m. 'parachute'. 4.4. Compounding Compared to derivation by affixation, compounding in this dialect plays a minor role. Although the total number of compounds is relatively small, the number of distinct types represented in the data is nevertheless rather large. Compounds are identified by means of several of the criteria listed earlier in section 3.2. Units entirely composed of free forms, as well as units which contain certain bound forms and which are structurally comparable to other compounds rather than to complex words, are treated as compounds. Not included here are unanalyzed borrowings from the French which have been integrated into Dzcs. merely by phonemic substitution, for instance, the N /su flyr/, m. 'chou-fleur1 (cf. Dzcs. /kawlet/, m. 'chou' and /flu/, f. ^Currently competing with the form parasol, m. which has identical meaning but is not felt to be bi-morphemic. 342 'fleur'), a French composite of two nouns which is of a type not represented among indigenous compounds. A few examples of recent introductions, usually hybrid compounds, will be given to show the extent of the French influence. A compound is classed as endocentric if the unit is of the same form-class and of the same gender as the head con stituent and if the referential meaning of the unit does not 66 contrast with that of the head. Compounds which do not fulfill these conditions are exocentric; these are somewhat more heavily represented in the data than are endocentric compounds and will therefore be treated first. It is to be noted that all units are of the subordinate type. Coordinate compounds, i.e. units wherein more than one constituent functions as head, are lacking. Compounds are written phonemically. The spaces are meant to indicate the division of a unit into smaller words, and have no phonological relevance. The stress patterns are marked on each unit; their significance has been discussed 66 ' Concerning the necessity of some semantic criterion for the classification of compounds otherwise identical, see L. Bloomfield, Language (London, 1935), p. 236 and E. A. Nida, Morphology (Ann Arbor, 1949), p. 94. By this criter ion alone, compounds which are comparable to English white- cap (person) or hotdog (food) are exocentric, whereas, say, blackbird (type of bird) is endocentric. 343 in section 1.3.1.II. With few exceptions, to be noted as they occur, the plural of a compound N is formed by adding /-s/ to the final component. Units given without mention of gender are of either gender. 4.4.1. Exocentric Compounds 4.4.1.1. Nouns I. The first element is a vb. These formations consist of at least two full words. The initial constituent is a vb. in 3 pres. Certain units are masc., others are of either gender. A. Vb. + N The constituents are: 3 pres, vb. + N in direct object function. These compounds are asyntactic in that a function word is lacking before the N. (1) Type/aasto sawso/ 'trouble-fete 1 . lit, 'gate sauce1 con sisting of 3 pres, of gasta 'gater' + sawso. f. Units of this type have agentive meaning, i.e. 'person or object performing an action'. Other examples: 344 Vb. bases Vb. + N pUrta 'porter' /porto fenan/, m. 'siege du charretier1, lit. 'porte-faineant' t v /porto felo/, m. 'porte-feuille' /porto manto/, m. 'porte-manteau' /porto munedo/, m. 'porte-monnaie' tira 'tirer' /tiro bu/un/, m. 'tire-bouchon' Vb. stems eskawfa 'echauffer' /eskawfo l£j/, m. 'chauffe-lit' Of this type are the modern hybrids, i.e. partially trans lated units, such as /grato sjel/, m. 'gratte-ciel' (cf. nat. sew, m. 'ciel') with the second element borrowed. The loan-compound /porto fylo/, m. 'porte-feuille' is a doublet of /porto felo/. (2) Tvne /tasto sawso/. m. ’tremoette'r lit. 1tate sauce' consisting of 3 pres, of tasta 'tater' + sawso. f. Units of this type mean 'item for a given purpose*. Other examples: Vb. bases Vb. + N esyga 'essuyer' /esygo mas/, 'essuie-mains' kupa 'casser' /kupo krusto/, m. 'petit dejeuner', lit. 'casse-croute' prEga 'prier' /prego diw/, m. 'prie-dieu' 345 (3) Type /kvro nlw/. m. 'dernier ne'. lit, 'cure nid1 con- ■ ■ I * sisting of 3 pres, of kyra + niw. m. ’nid1. These are bahuvrihi compounds, i.e. they denote some characteristic trait by which a person or object is identified. Of this type is also /grato kyw/, m. 'fruit de 1'eglantier', lit. 'gratte cul' : 3 pres, of orata 'gratter' + kyw. m. 'cul'. B. Vb. + function word + N These formations have clause structure and agentive meaning. The sub-types are: (1) Vb. + def. art. + N. e.g. /biro 1 el/ 'louchon', lit. 1 tourne l'oeil' : 3 pres, of bira 1 tourner' + l(u) + .el, m. ■ceil' . (2) Vb. + prep. + N. e.g. /krebo de fame/ 'meurt-de-faim' : 3 pres, of krEba 'mourir' + + fame. f. 'faim'. (3) The unit /fringo las tutos/, m. 'chercheur de femmes', lit. 'courtise les toutes' is asyntactic. It consists of 3 pres, of frinoa 'courtiser' + def. art. with fern. pi. in flection /las/ 'les' + inclusive pron. with matching in flection /tutos/. This is a unique occurrence of this pron. preceded by the def. art. 346 C. Vb. + adv. These are syntactic units with agentive meanings, e.g. /baw paw/ 1vaurien1, lit. 'vaut peu' made up of 3 pres, of bale Irr. 1valoir1 + the adv. paw 'peu' functioning as verb- modifier. Of this type is also /paso pertut/, m. 'scie passe-partout' : 3 pres, of pasa 'passer' + the adv. pertut 'partout'. D. Vb. + adi. Comparable to units of the preceding type is the com pound /sem bun/, m. 'parfum', lit. 'sent bon' which consists of 3 pres, of the copulative vb. sent! : /sen/ (with /n : m/ by assimilation) + the masc. adj. /bun/ 'bon'. II. The first element is a N A. N + adi. In these phrasal units, the adj. functions as N modi fier but does not necessarily show gender concord with the N constituent. The gender of the unit matches the inflec tion of the adj. These are bahuvrihi compounds. The type /damo blanko/, f. 'hibou', lit. 'dame blanche' : damo, f. + fern. adj. blanko is syntactic. Other examples are /bin agre/, m. 'vinaigre' : bi(n), m. + masc. 347 adj. /agre/; /pe nyt/, m. 'pauvre here1, lit. 'pied nu' s p£, m. with AA + masc. adj. /nyt/. Of this type is also the recent acquisition /pje nwar/, m. 'pied noir', an unassimi lated loan-word. The compound /pe nydo/, f. 'pauvresse', the fern, form of /pe nyt/, is asyntactic. The compound /san fret/, m. 'sang-froid' : san, f. + masc. adj. /fret/ may be reckoned here although it may be better regarded as a mistranslation of the French original which is syntactic and of masc. gen der . B. N + (function word + N) These units contain a N modified by a prepositional phrase consisting of the prep. 'de' + N. The gender of the unit is determined by the initial N. These formations are of the bahuvrihi type: /pan d awzet/, m. 'cenelle, fruit de l'aubepine', lit. 'pain d'oiseau' : pan, m. + d(e) + awz£t, m. /merdo de kukyt/, f. 'gomme du cerisier', lit. 'fiente de coucou' : merdo, f. + de + kukyt, m. /pumo de tero/, f. ’pomme de terre' : pumo, f. + de + 348 4 = 67 tero, f. The compound /ajgo de bijo/, f. 'eau-de-vie'— presently displacing the native /ajgorden/, f.— is a hybrid with the last element borrowed. Its constituents are ajgo. f. + de I I + bijo, f. (cf. nat. bito. f. 'vie'). C . N + N In these asyntactic formations the second N is attri butive to the first and has possessive meaning. The units are plant-names of masc. gender and of the bahuvrihi type: /pupo krabo/, m. 'Calabre, dit plant de Bordeaux', lit. \ i t 6 8 'mamelle (de) chevre' : pupo, f. + krabo, f. /lengo byws/, m. 'rumex acetosa', lit. ' langue (de) bceuf' : 1 ' 4T -L. T * 6 9 lengo, f. + byws, m . D. N + adv. There are two asyntactic compounds denoting direction ^The plural of this compound is atypically formed by the addition of /-s/ only to the initial component. Other compounds of this type are used only in the singular. 68The vernacular label, the name of a variety of grapevine, describes the shape of the fruit. 6^The final /s/ in /byws/ is not a plural marker but rather a relic of the Latin genitive ending. 349 by means of a N modified by an adv. They are limited to occurrence after the prep. a. 'a' and cannot be glossed independently from the phrases: /a kabbat/ 'vers le bas1 i 70 and /a katsys/ 'vers le haut'. The unit /kab bat/ con tains the N kaP. m. 'tete, face' (with /-b/ for /-p/ by i 71 assimilatory gemination) and the bound form /-bat/ of the adv. bas 1en bas'. The unit /kat sys/ consists of the N kaP. m. 'tete, face* (with /p : t/ before /s/) and the bound adverbial element -sys ' sus 1 III. The first element is an adi. The constituents are: adj. (modifier) + N. The N determines the gender of the unit. These formations are of the bahuvrihi type. (1) Syntactic units In these units the word-order is the same as in phras es, e.g.: /katre kamis/, m. 'carrefour', lit. 'quatre chemins' : 70The gender of these units cannot be determined. They have some of the syntactic characteristics of both N and adv. and could be regarded as adv. 7^For the recurrence of this bound form in compound adv., see 4.4.2.2.I below. 350 numeral katre + /kamis/, pi. of kami(n) . m. /mjezo nej/ f. 'minuit', lit. 1demie nuit' : fem. adj. /mjezo/ + nej, f. /mjed zur/, m. 'midi', lit. 'demi jour1 : masc. adj. /mjej/ (with -j 0 0 -d/) + zur, m. (2) Asyntactic units The constituents are adj. + N; in similarly " ‘ constituted phrases, the adj. would be postposed to the N: /rum pyn/, m. 'rond-point' : masc. adj. /run/ (with /n : m/) + py(n), m. 'point'. /blam bek/, m. 'blanc-bec' : masc. adj. /blan/ (with /n : m/) + bek, m. The compound /tuts sans/ 'Toussaint', lit. 'tous saints' has no identifiable gender and is asyntactic in that a function word is lacking before the N. The consti tuents are: the inclusive adj. in masc. pi. form /tuts/ + the bound pi. form /sans/ of the N sen. m. (with /e < ® a/). The unit /lungo byjo/, f. 'longue-vue' meaning 'instrument for1 is a hybrid comprising the fem. adj. /lungo/ and the phonologically adapted N /byjo/, f. (cf. nat. bisto. f. 'vue'). 351 IV. The first element is an adv. These compounds are formed of an adv. modifying a past participle. The units have masc. gender and the meaning ‘thing or person which or who is . . /maw entendyt/, m. 'malentendu' : maw + /entendyt/, past part, of entende 'entendre' /maw apres/, m. 'malappris' : maw + /apres/, past part, of aprenge Irr. 'apprendre' /bjen fej/, m. 'bienfait' : bjen + /fej/, past part, of fa_ Irr. 'faire'. V. Reduplicative compound N The N /triko trako/, f. 'tarare* is an imitative word designating a machine. 4.4.1.2. Adverbs These compounds are identifiable as adv. by their posi tion and function. The constituents may be: I. Adi. + N These are syntactic phrases: /lun tens/ 'longtemps' : masc. adj. /lun/ + N tens, m., /awtes kots/ 'autrefois' : masc. pi. adj. /awtes/ + pi. N /kots/, m . 352 II. Prep. + N The unit /en lok/ 'nulle part1 is a negative adv. con taining the prep, en 'en' and the bound N lok. m. (cannot be 72 glossed). III. Prep. + pron. E.g. /per tut/ 'partout' : prep. per, 'par' + inclusive pron. tut-C ' tout' . IV. Prep. + prep. E.g. /de din/, 'dedans' : prep. 'de’ + prep, din 'dans'. Its antonym /de foro/ 'dehors' contains the unique element /foro/, seemingly a prep. 4.4.1.3. Pronouns The indefinite pron. /gran kawzo/ 'grand' chose' is an asyntactic phrase comprising the masc. adj. /gran/ 'grand' and the N kawzo. f. 'chose' . 4.4.2. Endocentric Compounds 4.4.2.1. Nouns In endocentric compound N, the gender of a unit is ^This N recurs in the prepositional phrase /en kawke lok/ 'quelque part' where the masc. gender is apparent. 353 determined by the head N. The constituents appear in the following order. I. Adi, (attribute) + N (head) These formations are of two types: A. Syntactic units The type /belo maj/, f. 'belle-mere' consists of the fem. adj. /belo/ 'belle' (in sense 'in-law') + N maj., f. 'mere'. Other examples, /belo so/, f. 'belle-sceur ’ , /belo filo/, f. 'belle-fille', /petito filo/, f. 'petite-filie'. Of this type are also /bew paj/, 'beau-pere' : masc. adj. /bet/ 'beau' (with /t : w/ before /p/) + paj, m. 'pere' and /gram paj/, m. 'grand-pere' : masc. adj. /gran/ 'grand' (with /n : m/) + paj, m. 'pere'. B. Asyntactic units (1) Ty.pe /gran r.ytoA f■ 'grand' route' In units of this type there is gender discord between the constituents: here the masc. adj. /gran/ 'grand' and the N ruto, f. 'route'. Also of this type is /gram maj/ 'grand' mere' : masc. adj. /gran/ 'grand' (with /n : m/) + maj, f. 'mere'. 354 (2) Type , /ber zvs/r m. 'verius' In these units there occur in initial position adj. which normally follow the N in phrases: here the masc. adj. /ber/ 'vert' + the N zvs. m. Of this type is also the N /sazo fenno/, f. 'sage-femme' : fem. adj. /sazo/ 'sage' + N fenno. f. 'femme'. The compound /zyn ome/, m. 'jeune homme' is a hybrid with a borrowed first element (cf. nat. zwen 'jeune') + N ome. m. 'homme'. II. N (head) + adi. (attribute) These are syntactic units: /busun negre/, m. 'prunellier', lit. 'buisson noir' : busu(n), m. + masc. adj. /negre/ /busum blan/j m. 'aubepine1, lit. 'buisson blanc1 : busu(n), m. {with /n : m/) + masc. adj. /blan/. The compound /buro folo/, f. 'premier duvet, poil follet' contains the N buro. f. 'bourre' and the unique constituent /folo/, a fem. adj. formed on the hypothetical stem fol. III. N (head) + (function word + N) The attribute to the N head is a prepositional phrase in syntactic formations like /ben d awtan/ 'vent d'est', 355 lit. 'vent d'autan' : N ben. m. + prep, de 'de' (with eli- * 7 3 sion of /-e/) + N awtan. m . 'autan'. There are loan- translations of this type, e.g. /kamin de fer/, m. 1chemin 1 1 / l de fer' : N kami(n). m. + prep, de + N fer. m.; /sal a v | . ^ l minja/, f. 'salle a manger' : N salo. f. (with elision of /-o/) + prep, a, + vb. minna. The word /zandarmo/, m. 'gen darme' may be mentioned here as an example of unanalyzed loan-words. IV. N Plead) + N (attribute) The second N functions as modifier of the first in the asyntactic unit /ajgo saw/, f. 'saumure', lit. 'eau (de) sel1 : N ajgo. f. 'eau' + N saw, f. 'sel'. V. N (attribute) + N (head) The initial N denotes purpose in the asyntactic unit /kambo ligo/, f. ' jarretiere' : N kartibo, f. 'jambe' + N ligo. f. 'lie'. Of this type is the new word /oto ruto/, f. 'autoroute', a hybrid with initial N oto, f. 'auto' adopted intact and N ruto. f. 'route' translated. 73The Fr. noun autan was introduced in the 16th c. from Provencal. 356 VI. Adv. (attribute) + N (head) The asyntactic formation /apej mjedzur/, m. 'apr&s- midi' contains the adv. apei 'apres' and the compound N /mjed zur/ 'midi'. This unit is also an adv. 4.4.2.2. Adverbs I. Prep. + adverbial elements These formations consist of prep, followed by adverbial elements that occur only as bound constituents : /de la/ 1 1^,-bas 1 , /de sys/ 'dessus1, /en sys/ 'la-bas en haut', /em bat/ 'la-bas en bas', /en la/ 'de ce cote-la', /en sa/ 'de ce cote-ci'. The element /sa/ has unique occurrence in /en sa/. II. Adv. (head) + adv. (attribute) E.g. /aze dela/ 'avant-hier1 : adv. aze 'hier' + com pound adv. /de la/ 'la-bas'. III. Adv. (head) + past part. (attribute) E.g. /dumam pasat/ 1apres-demain1, an asyntactic com bination of the adv. duman 1demain' (with /n : m/) with the past part, /pasat/ 'passe' of pasa 'passer'. 357 IV. Adv. (attribute) + adv. (head) The unit /aban ze/ 'avant-hier' (rival form of /aze dela/) comprises the adv. aban 'avant1 and the adv. aze 'hier' in the reduced form /ze/. Also of this type is the hybrid /apredduman/ 'apres-demain' (rival form of /dumam pasat/) which is made up of the Fr. adv. apras borrowed without change and the adv. duman translated from Fr. de- main. There is linkage of the constituents by gemination of the initial /d-/ of /duman/. V. Pron. Lattribute) + adv. (head) E.g. /tut aro/ 'tout a l'heure, i.e. dans un moment, il y a un moment' : inclusive pron. tut" 'tout' + adv. aro 'maintenant'. 4.4.2.3. Verbs I. Adv. (attribute) + vb. (head) These are asyntactic formations with the adv. maw 'mal' as initial constituent in verb-modifying function, e.g.: /maw treta/ 'maltraiter', /maw mena/ 'malmener', /maw parla/ 'medire', /se maw fiza/ 'se metier'. 358 II. Vb. (head) + N (attribute) The asyntactic unit /tasta saw/ 'gouter un met pendant la cuisson', lit. 'tater sel' contains the vb. tasta di rectly followed by the N saw, f. in direct object function. 4.4.2.4. Pronouns Adi, (attribute) + pron. (he_ad) The indefinite pron. /kawk yn/ 'quelqu'un' (invariable) consists of the adj. kawke 'quelque' and the pron. ,yn 'un'. The numerative indefinite pron. /kawkez yns/ (masc.) 1quelques-uns' and /kawkoz ynos/ (fem.) 'quelques-unes' are syntactic phrases containing the same main constituents pluralized. The pron. /kad yn/ 'chacun1 consists of the adj. kado* 'chaque' (with elision of /-o/) and the pron. 'un'. 4.4.3. Problems of Classification Syntactically an adj., the combination /bum markat/ 'bon marche', the antonym of ser 'cher ' , is made up of the masc. adj. /bun/ and the N markat. m. 'marche'. Usage varies as to the formation of the fem. and the pi. For some speakers, the unit is invariable. For others, the adj. alone is invariable while the N carries the inflectional 359 characteristics; that is, the unit as a whole concords with the N modified, e.g. the fem. pi. form /bum markados/ 'bon marche'. In the compound N /li/o krabo/, m. 'chevre-feuille', the final component is the N krabo. f. 1chevre'. The mean ing of the unique initial element /li/o/ cannot be deter mined. This unit may be comparable to other plant-names, namely /grato kyw/, m. (vb. + N), lit. 'gratte cul' (4.4.1. l.type I.A.(3) above) or to /pupo krabo/, m. (N + N), lit. 'mamelle (de) chevre' (4.4.1.1. type II.C), or it may be of a type not represented elsewhere in the data. The N /ajgorden/, f. 'eau-de-vie' has as initial com ponent the N aioo. f. 'eau'. The final element is unique but may be regarded as a masc. adj. /arden/ 'ardent' showing aphaeresis of /a-/. If this is so, then the compound is to be classed with exocentric and asyntactic units of the bahuvrihi type. In the N /bun yr/, m. 'bonheur' the initial element is the masc. adj. /bun/ 'bon' modifying a bound N /-yr/, m. which recurs as base in the derived adj. yruzA 'heureux' and in the adv. vruzomen 'heureusement'. The word /bun yr/ can be analyzed as an endocentric and syntactic compound. If /bun yr/ is so treated, its antonym /mal yr/, m. 'malheur' 360 can also be regarded as a compound containing the initial bound element /mal-/, a residue which has no occurrence as an adj. outside of this compound. The N /mi mjot/ m. 'demi-muid' is constructed of the bound element /mi-/ and the N mjot, m. 'muid'. Since /mi-/ is not a free form, it may be set up as a unique prefix, in which case the unit is a derivate; or preferably /mi-/ may be regarded as a combining form of the masc. adj. /mjej/ . 74 'demi' and the unit itself as an endocentric compound. The negative pron. /digyn/ 'personne' may be inter preted as containing an obscure initial element /dig-/ and a pron. (head) £n by analogy to the units /kawk yn/ 'guelqu'un' and /kad yn/ 1chacun' which are clearly compound pron. Similarly, the indefinite pron. /kawkumet/ 'quelque chose' contains an obscure final element /-umet/. These two units are probably better treated as monomorphemic words. 4.5. Reduplication A few nouns which are commonly used in familiar style are formed by reduplication of all or part of the initial 74/mi-/ recurs as base in the derived N mitat. f. 'moitie'. syllable of a noun base; e.g. maj, f. 'mere' : mama 'maman' gram maj, f. 'grand' mere' : meme 'bonne maman'; gram paj, m. 'grand-pere' : pepe 'bon papa'; tanto, f. 1tante' : tata 'tantine'; unkle, m. 'oncle' : tutun 'tonton' (by analogy with the preceding); loro, f. (Christian name) 'Laure' : lolo. 4.6. Homophones There is in the Dzcs. dialect a rather small number of stems, nearly all base morphemes, which are phonemically identical but are members of two distinct word-classes, so for example the vb. minna 'manger' and the N minia. m. 'nourriture'. One may speak here of a zero process of word-formation since no derivational affix is present. According to the principles of this analysis, such homo- phonous stems are regarded as different morphemes although they would perhaps be better classed as single morphemes whose inflectional characteristics are determined by each word-class. In addition to those already noted in the descriptions of the various word-classes, a few other examples of homo phones have been found. Among these are: dezyna (vb.) 'dejeuner', dezuna, m. (N) 'dejeuner'; beWre (vb.) 'boire', 362 bewre, m. (N) 'boisson'; fred (adj.) 'froid', fret, m. (N) 'froid'; saduL (adj.) 1repu15 sadul, m. (N) 'indigestion'; gurmand (adj.) 'gourmand', gurraan, m. (N) 'rejeton'; patwez (adj.) 'patois', patwes, m. (N) 'patois'pryme[r] (adj.) 'premier', pryme (adv.) 'd'abord'; perdyd (past part.) 'perdu', perdyt, m. (N) 'perdu'. 5. Phrases 5.1. External Sandhi External sandhi modifications are phonological proc esses, specifically addition, loss, or substitution of one or more phonemes, which occur at word transition within breath-groups. A division into general and special sandhi is in order. Those phenomena which are regular in the sense that they apply to any component of a close-knit phrase in a given phonemic environment are called "general sandhi." Those phenomena which are irregular, in the sense that they apply under certain conditions to particular words or phrases, are assembled under "special sandhi." General sandhi modifications are universal and compul sory— i.e., consistently made— in close-knit phrases, op tional between two such phrases, not made at the end of a breath-group. The tendency is to omit optional sandhi, and thereby detach words, when the delivery is slow or when 363 364 various emotions come into play. For some examples of the omission of compulsory sandhi as a feature of contrastive accent, see above, 1.4.5.III. Certain special sandhi modi fications are compulsory, others optional, as will be speci fied in each case. By "close-knit phrase" is meant a syntactic unit com prising two or more words bound by close grammatical rela tionship, hence uttered, as is a single word, without in tervening pause. The constructions which have been termed "minimal phrases" (3.3) are examples of small close-knit phrases. Some of the larger phrases of the various types to be analyzed in section 5.2 are close-knit, others include a close-knit phrase. As a general rule, a phrase containing a full word and one or more function words is close-knit, e.g. /laz bykos/ 'les buches', /digozot./ 'dis-le1, /luz abem bis./ 'nous les avons vus1, /sinz yn so/ 'sans un sou'; so is any phrase wherein the head word is accompanied only by preposed attributes which may be either function words or full words, e.g. /mum petib busin/ 'mon petit morceau'. There is congruence between the head of such a phrase and all inflectable attributes, e.g. /akeloz grosoz bletos/ 'ces grosses betteraves'. The absolute form of a word (abbreviated abs.) is the 365 form which occurs before pause and, in the case of phrasally bound function words, the form which occurs before an un voiced consonant. Alterations of this form in included positions are sandhi-forms. 5.1.1. General Sandhi I. Assimilatorv voicing A. Before vowels As stated earlier in 1.2 in connection with word- linking, a word-final consonant joins the initial vowel of the next word, e.g. /ym bet aze/ 'un bel ane' is syllabified /ym-be-ta-ze/. While most consonants are thus carried over without phonemic substitution, /-s/ and /“// are voiced to /-z/ and /“// in anticipation of a vowel, e.g. (abs. /es/) /ezesputlt./ 'il est ecrase', (abs. /bene//) /bene/i./ 'venez-y'. B. Before consonants /-s/ is also subject to assimilatory voicing in anti cipation of an initial voiced consonant in the next word, e.g. (abs. /pas/) /pazlen/ 'pas loin'. This is the most widespread modification; however, as noted earlier in 1.5. III.A, implosive /~z/ can be optionally replaced by palatal 366 /— j/, e.g. /pazlen/ ~ /pajlin/ 'pas loin* . /-s/ remains before a voiceless consonant, e.g. /paspla/ 'pas beaucoup'. /~^/ undergoes the alterations described just below before both voiced and voiceless con sonants . II. Assimilatorv gemination and consonant mercers In word-final position the consonants /-p/, /-t/, and /-// either geminate by assimilation to certain following consonants or merge with other word-initial consonants. The various modifications undergone by /-p/ and /-t/ are shown in the table on p. 367. Similar examples with word-final /“//— not included in the table— are (abs. /abe//) /abeppuskyt/ 'vous avez pu', /abettor/ 'vous avez tort', /abikkranat/ 'vous avez fait le fier'. All modifications have this in common, that the oppo sitions between /-p/, /-t/, and /“// are neutralized, where as the initial consonant of the following word is preserved unchanged. This consonant, e.g. /p-/, is listed first in the left-hand column in the table; the resulting combination of this consonant with indistinctly word-final /-p/, /-t/, or /-//, e.g. /pp/, is given next. The central column in cludes two-word phrases by way of illustrations; the first V Word-final /-p/ or /-t/ + word-initial C p- > pp troppek 'trop idiot' 1 1 tuppetit 'tout petit' t- > tt 1 v 1 trottilus 'trop dur' tuttrawk&t 'tout troue' k- > kk trokkrane 'trop elegant' tukkapelat 'tout couvert b- > bb trobbiste 'trop vite' tubber 'tout vert' d- > dd troddybir 'trop ouvert' tuddys 'tous deux' g- > gg troggran 'trop grand' tugganat 'tout gagne' m- > mm trommasan 'trop mechant' tummustus 'tout souille' n- > nn trinnigaw 'trop nigaud' tunnegre 'tout noir' 1- > 11 trollew 'trop tot1 tullezit 'tout lu' 367 2. f- > [f:] s- > [5:] 3 . s- > ts v v ,V z- > dz A V v 4. c- > c r- > r troffenan 'trop faineant' trosser 'trop cher' trotsek 1trop sec' trodzwen 'trop jeune' trocuko 'trop sotte' l — ^ | trorunus 'trop crasseux' tuffij 'tout fait' tussalibat 'tout salive' tutsyzen 'tout suant' tudzejnat 'tout gene' tucykat 'tout suce' v turulat 'tout rouille' 368 369 word is /trop/ 'trop' throughout. In the right-hand column the first word of each phrase is /tut/ 'tout*. The sub-divisions correspond to the various types of changes undergone by the final consonants. In group one, the result is assimilation to the following consonant and subsequent gemination. This happens when the finals come into contact with any stop, the nasals /m-/, /n-/, or the dental lateral /l-/. With the fricatives /f-/, /s-/ of group two, the result is prolonged friction. The combina tions with the fricatives /s-/, /z-/ of group three yield not geminates, but respectively the clusters /ts/, /dz/. It will be noted that /~p/, /-t/, /“// all merge to voice less /t/ before /s-/, but to voiced /d/ before /z-/. In group four, the final consonants are absorbed by /c-/ and /r-/. Besides /-p/, /“t/, and /-//, only /-n/ and /-l/ recur often enough in word-final position to deserve mention. /-l/ geminates only in contact with /l-/, e.g. /bjellinsol/ 'vieux drap', but /bjelbydet/ 'vieux boyau'. There is no phonemic substitution of word-final /-n/ before consonants other than labials (see below), e.g. /ynnebut/ 'un neveu1, /yndit/ 'un doigt'. 370 III. Assimilation Word-final dental /-n/ is labialized to /-m/ before labial consonants, e.g. (abs. /yn/) /ymmustre/ 'un monstre1, /ympetas/ 'un chiffon', (abs. /un/) /umbuljo/ 'ou il vou- lait' . IV. Elision This is the loss of an unstressed word-final vowel before a word-initial vowel. The vowels canceled are /e a o u/, e.g. (with canceled vowels in parentheses) /yn(o)afa/ 'une affaire', /deraasant(o)imu/ 'de mauvaise humeur1, /dink(o)asiw/ 'jusqu'ici', /l(u)atyko./ 'il 1'assomme’, /s(e)akata/ 'se baisser1, /lupawr(e)awzet/ 'le pauvre oiseau', /l(a)unto/ 'la honte'. Words made up of a single vowel like /u/ 'ou' are not affected by elision, e.g. /buzuelo/ 1vous ou elle'. V. Palatalization of /i/ /i/ is palatalized to /j/ when in contact with an ini tial vowel in the following word, e.g. (abs. /i/) /jeri/ 'j'y etais’. 371 5.1.2. Special Sandhi I. Aphaeresis This is the loss of an unstressed word-initial vowel to avoid hiatus with an immediately preceding word-final vowel. This modification is a compulsory substitute for elision in specific constructions, such as /kadra(e)nsaza./ 'il faudra essayer1. See additional examples in 1.2.IV above. Either aphaeresis or elision can occur in the common phrase /kumoko/ ~ /kumako/ 'comme ga' (abs. /kumo/ + /ako/). II. Apocope The final consonant /-k/ of the numeral /sink/ 'cinq' is canceled before a word with initial consonant, e.g. /singrapaws/ 'cinq crapauds1, but /sinkomes/ 'cinq hommes', /njisink./ 'il y en a cinq'. III. Liaison This is the addition of a word-final consonant when a word with initial vowel immediately follows.'*' The linking consonant may be /-t/, /-?/, or /-n/. This modification is ^Liaison differs from sandhi 5.1.1.A above in that here the linking consonant is not present in the absolute form; it is actually added to a word for the liaison, so that the resulting form is a special sandhi-form of that word. 372 compulsory unless indicated otherwise. Specific words are affected only if they occur in certain constructions. The following list includes all compulsory sandhi-forms. A. /-t/ is added; (1) to the regular masc. sg. adj. forms /masan/ 'mechant' and /gran/ 'grand' only in proclitic (unstressed) position before a nominal, e.g. /masantome/ 'mechant homme', 2 /grantespitaw/ 'grand hopital'. (2) to the numerals /wej/ 'huit', /bin/ 'vingt1, and /sen/ 'cent' before a nominal, e.g. /bintagasos/ 'vingt pies', /sentawres/ 'cent arbres1, and to the numeral /bin/ 'vingt' also in compounds like /bintetres/ 'vingt trois'. (3) to the noun /sen/ 'saint', e.g. /sentemilo/ 'Saint Emile', and to the noun /ben/ 'vent' in the proverb — 3 /awboruzo : bentupluzo./ 'aube rouge, vent ou pluie'. ^It will be noted that the addition of the linking /_t/ does not produce unpermitted word-final consonant clusters since this /t/ is automatically shifted to the initial syl lable of the next word. ^Here /pluzo/ replaces Dzcs. /piezo/ 'pluie', either for the sake of the rhyme or because the proverb may have originated in western Gascon dialects. 373 b . /-?/ i,s (1) to the prepositions /din/ 'dans' and /sin/ 'sans1, e.g. /dinzynsule/ 'dans un grenier'. (2) to the numeral /katrebin/ 'quatre-vingt' before a nomi nal, e.g. /katrebinzans/ 'quatre-vingts ans'. (3) optionally to the coordinator /me/ 'mais', e.g. /mezalabe// 'mais alors'. C. /-n/ is added to the prepositions /a/ 'a' and /enta/ 'chez' only when a personal or a demonstrative pronoun fol lows, e.g. /anet/ 'a lui', /anakesto/ 'a celle-ci', /entane/is/ 'chez eux'. The addition of /-n/ in these types of phrases is optional but very frequent. In other close- knit phrases /-n/ does not appear and hiatus results, e.g. /aynawte/ 'a un autre', /aazen/ 'a Agen'. IV. Insertion of the connective morpheme /z/ The phonemic form of certain short object pronouns and the pronominal adverb /i/ 'y' is safeguarded in certain cases by the compulsory addition of the connective /z/. This connective, which is void of lexical meaning, is in serted in verb phrases where it joins a given short pronoun 374 to the head verb or to another given pronoun. The specific pronouns and combinations of pronouns accompanied by /z/ 4 are listed in full with examples. A. In proclitic position /awz-/ : /aw/ 'le' neuter d.o. pron., e.g. /awzeskuto./ 'il l'ecoute', /awzejfej./ 1je l’ai fait’, but /awmyso./ 'il le montre'. /jawz-/ : ij^ ' lui' 3rd pers. dative pron. followed by /aw/ I v ' le' , e.g. /jawzeskampilo./ 'il le lui eparpille’, /jawzejdit./ 'je le lui ai dit'. /jawz-/ : i2 'y' pron. adv. followed by /aw/ ' le' , e.g. /jawzeskuto./ 'il l'y ecoute', /jawzapurtat./ 'il l'y a porte'. ^That /z/ appears before a vowel or /j/ is not a defi nition of its occurrence. This is evident from the follow ing data, /me/ is normally subject to elision, so that /me/ 'me' + /aw/ 'le' > /maw/ in /maw porto./ 'il me le porte', also /me/ 'me' + /i/ 'y' > /mi/ in /mi porto./ 'il m'y porte', /porto mi./ 'porte m'y'; but /me/ 'me' + /ot/ 'le' > /mezot/ in /porto mezot./ 'porte-le-moi'. As to the shape /z/ of this connective, it is probably due to analogy with sandhi variants of several other pro nouns which have voicedv/-z/ for /-s/ before a vowel, e.g. (abs. /nus/) /nuz aw balo./ 'il nous le donne', /balo nuz ot./ 'donne-le-nous', (abs. /las/) /laz i bezi./ 'je les y vois' . B. In enclitic position /-zot/ : /ot/ 'le' neuter d.o. pron., e.g. /fajzot./ 'fais-le', /faze/zot./ 'faites-le', /perzot./ 'perds-le', /prumetzot./ 'promets-le1. /-mezot/ : /me/ 'moi' 1st pers. sg. ind. o. pron. joined to /ot/ 'le1, e.g. /balomezot./ 1donne-le-moi'. /-tezot/ : /te/ 'toi, toi-meme' 2nd pers. sg. ind. o. pron. joined to /ot/ 1le1, e.g. /fajtezot./ 'fais-le toi-meme'. /-zi/ : ix 'lui', e.g. /digozi./ 'dis-lui', /respundenzi./ 'repondons-lui'. /-zi/ : i3 'y', e.g. /damorozi ./ 'restes-y', /ananzi./ 'allons-y'. /-zjot/ : ix 'lui' followed by /ot/ 'le', e.g. /digozjot./ 'dis-le-lui', /kita/zjot./ 'enlevez-le-lui'. /-zjot/ : i2 'y' followed by /ot/ 'le', e.g. /desozjot./ 'laisse-l'y, /metzjot./ 'mets-l'y1. /mezjot/ : /me/ 'moi, pour moi' joined to /jot/ 'l'y', e.g. /memmezjot./ 'mets-l'y pour moi'. /tezjot/ : /te/ 'toi, toi-meme' joined to /jot/ 'l'y1, e.g. /portotezjot./ 'porte-1'y toi-meme'. 5.2. Phrase-Structure Like a word or a minimal phrase, a phrase may on 376 occasion be one immediate constituent of a meaningful ut terance, the other immediate constituent being an intonation contour. Ordinarily, however, phrases are found at lower levels of structure, as fractions of breath-groups, so that they occur with varying portions of intonation contours. In the analysis of phrases it will therefore not be necessary to mark intonation unless a phrase is at the same, time a complete sentence. Stress will be marked as it occurs in slow, deliberate speech. All compulsory sandhi modifica tions will be noted as well as those which are optional but frequently made. 5.2.1. Endocentric Phrases A phrase is endocentric if it has substantially the same range of privileges of occurrence as one of its con stituents (the head). Phrases which lack this property are exocentric. 5.2.1.1. Nominal Phrases These phrases occur mainly before a verbal, i.e. verb or equivalent phrase, as subjects, after a verbal as comple ments, after a preposition as axes. The head is a N. Attributes modifying the head are of 377 5 three types: nominal markers (NM), adj., and exocentric attributes. There is gender and number concord between the head and the inflectable attributes. For the forms of nomi nal markers, see adj. (4.2.2.4.II), numerals (4.2.2.5), and particles (4.2.5.7). Constituents appear in the general order: NM ± adj. + N (head) - adj. i exoc. attr. The possible combinations are illustrated below. A. NM + N. with NM position filled by: (1) the def. art. 1^ 'le, la1, e.g. /lu nas/ 'le nez', /la zano/ 'Jeanne', /l embezo/ 'l'envie', /luz esklots/ 'les sabots' (2) the indef. art. yn 'un', e.g. /ym balat/ 'un fosse', /yno padeno/ 'une poele1, /yn estelo/ 'une etoile' (3) the partitive art. cle 'du, de la, des' , e.g. /d ajgo/ ' de l'eau', /de byw/ ' du bceuf', /d oskos/ 'des entai lies' (4) poss. adj., e.g. /mun kutet/ 'mon couteau', /ta man/ ^These are function words which introduce nominal phrases.^ The term "nominal markers" is preferred to "noun markers" because, as will be noted below, the noun is not always expressed within the phrase. 378 'ta main', /nostos bestjos/ 'nos betes' (5) demonstr. adj., e.g. /aket tawre/ 1ce taureau-la', /akeste kop/ 'cette fois-ci', /akeloz bimos/ 1ces genisses- 1k' l (6) the interrogative-exclamative adj. kin 'quel', e.g. /kino fenno'./ 'quelle femme', /kim betetl/ 'quel veau' (7) the distributive adj. kado* 'chaque', e.g. /kado mes/ 'chaque mois', /kado porto/ 'chaque porte' (8) the quantitative adj. forso* 'bien des, bon nombre de' and kawke 'quelque' both occurring before count nouns, e.g. /forsoz awres/ 'bien des arbres', /forsoz lawzetos/ 'bien des alouettes', /kawkos fymelos/ 'quelques femelles', /kawkez liwses/ 'quelques eclairs' (9) the negative adj. nad 'aucun', e.g. /nat awzet/ 'aucun oiseau', /nado trezo/ 'aucune truie' (10) cardinal numerals, e.g. /djoz glejzos/ 'deux eglises', /brntewej testes/ 'vingt-huit auges' (11) two nominal markers, in which case the second is a numeral and the first is limited to: the def. art., a demonstr. adj., a poss. adj., or the distributive kado* 379 'chaque', e.g. /luz dyz endrets/ 'les deux endroits', /sas trez bordos/ 'ses trois fermes', /kado katre rens/ 'chaque quatre rangs' (12) a phrase containing: a. a numeral with the adv. kawke 'environ' as preposed modifier, e.g. /kawko kranto regos/ 'environ quarante sil- lons ' b. the partitive de_ 'du, le la, des' with quantitative ex pressions as preposed modifiers. These may be: i. a quantitative adv., e.g. /trod de blat/ 'trop de ble', /talomen de bermes/ 'tellement de vers', /pru de frumaje/ 'assez de fromage', /pla de sezes/ 'beaucoup de pois’, /betkod de tyros/ 'beaucoup de mottes de terre', /kumbjen/ de pesos/ 'combien de champs'. Where the negative adv. pas 'pas' or the negative phrase paz mei 'ne ... plus' are present, the entire nominal phrase functions as complement in larger constructions, e.g. /p4z mtj de sibado/ 1ne ... plus d'avoine', /paz d ajre/ 'pas d'air'. ii. the indef. art. yn 'un' + N, e.g. /ym moz de pastis/ 'un peu de pate', /yn busin de pan/ 'un morceau de pain', /ym paw de bin/ 'un peu de vin', /yno guto d ajgo/ ’une goutte d'eau', 380 B. NM + adi. + N. the adj. being an ordinal numeral or one of some 14 (lexically defined) proclitic adj. listed below. In proclitic position, adj. are normally unstressed, e.g. /lu gran zur/ 'le grand jour', /de pulik kases/ 'de jolis chenes'. Where two or at most three proclitic adj. are present, orders normally rank as follows: (1) adj. of first order, i.e. closest to the N head: bjeL I 1 v 1 'vieux', lung 'long', mjez ~ mjej 'demi' I , (2) adj. of second order: petit 'petit', grand 'grand', ' A i i gros gros (3) adj. of third order: pawre 'pauvre', bu(n) 'bon', masant 'mauvais, mechant', pulid 'joli', beL 'beau' (4) adj. of fourth order: memb* 'meme' , awte 'autre', dare[r] 'dernier', pryme[r] 'premier' and other ordinals, e.g. /la darero belo rekolto/ 'la derniere belle recolte', /lu pryme bum petig git/ 'le premier bon petit canard'. C. Nominal phrases of the preceding types may be followed by an enclitic adj., i.e. other than those listed as pro clitic, or an adjectival phrase having any adj. as head, e.g. /de fennoz ajmablos/ 'des femmes aimables', /de/ awkoz 381 blankos/ 'dix oies blanches', /yn drolle zwen/ 'un jeune gargon', /yn kapet pla pulit/ 'un tres joli chapeau', /laz grynoz eskampilados/ 'les grains eparpilles'. Here is a large nominal phrase including all the types of attributes described so far, i.e. NM + num. + 3 adj. + N (head) + adj., /laz djoz awtoz pawroz bjeloz bakoz malawzos/ 'les deux autres pauvres vieilles vaches malades'. D. The N head may be implied in a nominal phrase containing one of the common descriptive adj., either proclitic or en- g clitic. This adj. is then stressed and all constituents concord with a N in the immediate context, e.g. /yno masanto/ 'une mechante', /de negros/ 'des noires', /kin inusentoi/ 'quelle sotte', /dyz n£ws/ 'deux neufs', /yn awte petib bjel/ 'un autre petit vieux'. E. Nominal phrases are closed initially by the inclusive adj. tutj_ 'tout' occurring directly before one of the mark ers: def, art., demonstr. adj., or poss. adj., e.g. /tut akeste boj sek/ 'tout ce bois sec', /tutoz las sebos/ 'tous ^According to the analysis adopted in this study these words have been identified as adj. by their morphological characteristics and, if need be, by the supplementary cri terion that adj., unlike nouns, may be preceded by compara tive adv., e.g. /de mej masantos/ 'de plus mechantes'. 382 les oignons', /tul luz esarmens/ 1tous les sarments'. F. Nominal phrases are closed finally by exocentric attri butes of the types: (1) prepositional phrase a. with NM + N as axis, e.g. /ym manto damb yno sinto/ 1un manteau avec une ceinture'. In such phrases the introduc tory prep. dfi. 1de' carries diverse meanings, among them: possession, e.g. /lu sibaw d ym bezin/ 'le cheval d'un voisin', /la ped de 1 aze/ 'la peau de 1'ane1, purpose,- e.g. /la sazun de la pesko/ 'la saison de la peche', /la butelo de la lej/ 'la bouteille du lait1, origin, e.g. /laz zens de la bilo/ 'les gens de la ville'. b. with a N as axis. In such phrases the prep, de 'de' / 1 V I > carries such meanings as: contents, e.g. /la butelo de lej/ 'la bouteille de lait', substance, e.g. /yn ytiz de fer/ 'un outil en fer', instrument, e.g. /yn kod de pyn/ 'un coup de poing', ingredient, e.g. /la supo de kawlets/ 'la soupe aux choux', localization, e.g. /ym maw de kap/ 'un mal de tete'. The introductory prep, .a 'a' carries such meanings as: characteristic, e.g. /ym pane a traws/ 'un panier a trous', means, e.g. /yno kurso a pe/ 'une course a pied', purpose, e.g. /de tero a ordi/ 'de la terre a orge'. 383 c. with other types of axes (see 5.2.2.1 below), e.g. /de bestjoz a tja/ 'des animaux a tuer1. (2) relative clause opened by the relative pron. ke ~ ki in the form /ke/ 'que, qui' , e.g. /yno rawbo k ej fejto/ 'une robe que j'ai faite', /lu suflet k a atrapat/ 'la gifle qu'il a regue', /l annado ke ben/ '1'annee prochaine, lit. qui vient', /lu mes ke ben/ 'le mois prochain’. 5.2.1.2. Adjectival Phrases These are attributive phrases occurring mainly after a verb as subjective or objective complements and after a N as modifiers. The head is an adj. or a participial adj. Attributes may be: (1) the adverbial adj. tut-a 'tout', which precedes the head, e.g. /tub blan/ 'tout blanc', /tutos raputinados/ 'toutes rondelettes', /tuto destimburlado/ 'tout ecervelee'. (2) one or more adv. preceding the head and occurring in the order to be stated below in 5.2.1.4.IV.(l)c, e.g. /talomem biw/ 'tellement vif', /paz bjen rebiskulat/ 'pas bien ravive', /zamej pru gran/ 'jamais assez grand'. 384 (3) a comparative adv. preceding the head. In comparisons of superiority, inferiority, and equality, the adv. are respectively: mej 'plus', men 'moins', and ta(n) 1aussi', i * * i i 7 e.g. /mej petasat/ 'plus rapiece', /mem bun/ 'moins bon'. Complements may follow such phrases and close comparative constructions. Comparative constructions are discontinuous and enclose the adj. head. The complements are exocentric phrases of the types: a. clause opened by the con j. Ice 'que', e.g. /men ser Ice pensabi/ 'moins cher que je ne pensais' b. incomplete clause (i.e. with potentially present verbal) opened bv kumo 'que' in comparisons of equality, e.g. /tan kadyko kum et/ 'aussi idiote que lui' c. incomplete clause opened by ke 'que' in other compara tive constructions, e.g. /mej krano ke la dumajzelo/ 'plus elegante que la demoiselle'. The adj. bu(n) 'bon' does not occur preceded by mej 'plus'; the comparative of this adj. is a separate morpheme, V I . . . , | | V f the adj. milu Ln J 'meilleur', e.g. /akelo supo ez miluno ke 1 awto./ 'cette soupe est meilleure que 1’autre'. 7 A favorite substitute for men 'moins' is pas t4(n) 'pas si', e.g. /pas ta bun/ 'pas si bon'. 385 (4) the adv. ola 'tres' preceding the head and yielding an absolute superlative construction, e.g. /pla trende/ 'tres tendre'. (5) the def. art. 1^ 1le1 + mei 'plus' or men 'moins' pre ceding the adj. head. The resultant phrase is a relative superlative expression, e.g. /la mej benanto/ 'la plus vigoureuse', /luz mej rises/ 'les plus riches'. Complements may follow such a phrase and close the construction. Super lative constructions are discontinuous and enclose the adj. head. These complements are exocentric phrases of the types: I a. clause opened by the con j . ]£ £. 'que', e.g. /la mej pulido I I , k ej bisto/ 'la plus jolie que j'aie vue' b. phrase opened by the prep. ds. 'de' + nominal, i.e. N or I I 1 equivalent phrase, e.g. /lu mej ajmable du pais/ 'le plus aimable du pays', /lu mem bet de tutis/ 'le moins beau de tous1. (6) the negative adv. pas 'pas' preceding an adj. head or any of the above attributive phrases, e.g. /paz braj/ 'pas vrai', /paz mej ber/ 'pas plus vert', /pas tarn mady kum aket/ 'pas si mur que celui-la'. (7) exocentric attributes following the head. These are of two types: a. prepositional phrase with various axes (see 5.2.2.1), e.g. /bum per res/ 'bon a rien', /buno per punde/ 'pon- deuse', /klawfit de fryto/ 1recouvert de fruits', /preste a parti/ 'pret a partir'. When used as attribute to an adj. head which itself functions as subjective complement, a phrase opened by the prep. j i e . ' de' (with meanings such as 'a propos de, concernant') is replaceable by the pron. adv. ne 'en' placed in the verbal core, thus /(es) fjer de sum ben/ '(il est) fier de son bien': /n (es) fjer/ ’(il) en (est) fier', /(semblo) satisfej du bajlet/ '(il semble) satisfait du domestique' : /ne (semblo) satisfej/ '(il semble) satisfait de lui', /(sum) pawrez de blat/ '(ils sont) pauvres en ble' : /ne (sum) pawres/ '(ils sont) pauvres en cela'. In attributive phrases with infinitive vb. axes, the introductory function words de. 'de' and a. 'a', both nearly void of lexical content and often interchangeable, are bet ter regarded as connective particles than as prepositions, e.g. /empusible de (~ a) gari/ 'incurable', /defesille de (~ a) bende/ 'difficile a vendre', /masan de (~ a) dariga/ 'malaise a arracher'. 387 b. subordinate clause, e.g. /pasjentos kam bolon/ 'pati- entes quand elles veulent', /bestjo ke runko/ 1affreusement bete, lit. bete au point qu'elle ronfle', /let ke fa pow/ 'laid a faire peur1. When used as attribute to an adj. head which itself functions as subjective complement, a clause opened by the conj. ke 'que' is replaceable by the pron. adv. ne 'en* placed in the verbal core, thus /(sej) kunten ke sen ange/ '(je suis) content qu'il s'en aille' : /ne (sej) kunten/ '(je suis) content de cela', /(ero) facado ke turnese pas/ '(elle etait) fachee qu'il ne revint pas' : /n (ero) facado/ '(elle) en (etait) fdchee'. 5.2.1.3. Pronominal Phrases These phrases function essentially like nominal phras es. The head is a pron. Dependent pron., i.e. those having an antecedent, show gender and number agreement by cross- reference to a N in the linguistic or the situational con text. Inflected attributes agree with the pron. head. In phrases wherein the head is an independent pron., i.e. an uninflected pron. without antecedent, or a pron. with neuter inflection, attributes show masc. sing, inflection. Attributes occur with various types of pron. heads as follows. 388 (1) The def. art. precedes: a. poss. pron., e.g. /la mjo/ 'la mienne1, /luz bostes/ 'les votres'. These possessive phrases occur in positions other than after a linking verb, e.g. / (demandi) luz mews/ '(je demande) les miens', /lu noste (ez esabuzit)/ 'le notre (est abasourdi)'. b. the indefinite pron. yn 'un', e.g. /l yn/ 'l'un', /laz ynos/ 'les unes'. These phrases normally occur in correla tion with phrases containing the pron. awte 'autre'. c. the indefinite pron. awte 'autre' and memo* 'meme', e.g. /l awto/ '1'autre', /laz memos/ 'les memes'. (2) The adj. sul 'seul' and memo* 'meme' follow nominative pron., e.g. /zu sulo/ 'moi seule', /ty memo/ 'toi-meme'. (3) Descriptive adj. follow the inclusive pron. tut ^ 'tout', e.g. /tutoz bjokos/ 'toutes depareillees', /tutis petits/ 'tous petits'. These phrases function only as sub jective or objective complements. (4) The inclusive adj. tutx 'tout' precedes demonstr. pron. and numerals, e.g. /tutoz djos/ 'toutes les deux', /tud dys/ 'tous les deux', /tut ako/ 'tout cela'. (5) Nominals follow nominative pron. in close apposition, 389 e.g. /nuzaw pajzans/ 'nous paysans'. (6) The adv. me j 'd'autre' follows indefinite pron., e.g. /kawkym mej/ 'quelqu'un d'autre', /kawkumem m&j/ 'autre chose'. (7) The intensifier pas 'pas', when present, precedes a negative pron., e.g. /(j abj6) paz digyn/ '(il n'y avait) personne', / (kreze/) paj res/ '(vous ne croyez) rien', /(ne i i i 8 biskut) paz nado/ '(il n'en vit) aucune'. (8) Exocentric attributes may be: a. prepositional phrases consisting of: i. the prep, de 'de' with possessive meaning + NM + nomi nal. These phrases follow the phrase-bound demonstr. pron. I so (neuter) 'ce1 and 1^ 'celui, celle', e.g. /so de sun unkle/ 'ce qui appartient a son oncle', /lu de sa maj/ 'celui de sa mere', /laz de la bordo/ 'celles de la ferme'. In the absence of antecedent, the form /la/ 'celle' carries the special meaning 'wife', e.g. /la de 1 andre/ 'la femme d1 Andre' . ^The negative meaning of this pron. is not conditioned by the presence of pas 'pas', thus /(j abjo) digyn/ 1(il n'y avait) personne', /digyn (es kunesyt)/ 'personne (n'est connu)'. ii. the prep. de_ 'de' with possessive meaning + poss. pron. These phrases follow the pron. so 'ce', e.g. /so de mew/ 'ce qui m'appartient', /so de tun/ 'ce qui est a toi'. iii. various other prep, with infinitive verb axes. These phrases follow indefinite pron. other than on 'on', e.g. /r&z a fk/ 'rien a faire', /kawkym per adyza/ 'quelqu'un pour aider 1 . 9 iv. the connective particle de 1de' + adjectival. These phrases follow negative pron., indefinite pron. other than on 'on1, and the demonstr. ako 'cela', e.g. /kawkumet de pla bun/ 'quelque chose de tres bon', /digyn de preste/ 'personne de pret', /ako de pulit/ 1ga qui est joli*. b. relative clauses introduced by the relative pron. ke. ~ ki 'que, qui', following demonstr. pron., neg. pron., and indefinite pron. other than cm 'on', e.g. /so ke boli/ 1ce que je veux', /so ke te plaj/ 1ce qui te plait', /lu ki pluro/ 'celui qui pleure', /lus ke budras/ 'ceux que tu voudras', /kawkyn ke ben/ 'quelqu'un qui vient', /digyn ke sjoske balen/ 'personne qui soit travailleur', /res ke bale/ 'rien qui vaille'. 9This connective is absent when the head is a nominal phrase, e.g. /yno kar pla buno/ 'une tres bonne viande'. 391 The relative clause /ke sjoske/ yields indefinite pro- , I nominal phrases when preceded by the relative pron. ke. ~ ki ■que, qui' without antecedent, thus /ki ke sjoske/ 'qui que ce soit, n'importe qui', /ke ke sjoske/ 'quoi que ce soit, n1importe quoi'. 5.2.1.4. Verbal Phrases Actor-action constructions are of two favorite types. In the actor-action type the actor (subject), a nominal, and the action (predicate), a verbal, are distinct words or phrases separated by potential pause, e.g. /la fenno ez bengydo./ 'la femme est venue', /elo ez bengydo./ 'elle, elle est venue'. In the actor-action type the actor is a postposed nominal, e.g. /ez bengydo la fenno./ 'elle est venue, la femme', or it is expressed merely by means of the verb inflection, e.g. /ez bengydo./ 'elle est venue'. Any single verb in a finite form of any mode other than sub junctive may, therefore, occur as a sentence-word, e.g. /akabet./ 'il acheva*, and any verbal phrase containing such a finite form may occur as a sentence-phrase, e.g. /abjo akabat./ 'il avait acheve'. This section will deal with these subjectless finite verbals— single verbs or equivalent phrases— their complements, and their attributes, but the 392 analysis will not go beyond the frame-work of the indepen dent clause. I. The verbal core consists of a center, finite verb or equivalent phrase, with or without pronominal complements. A. Pronominal complements (pron. compl.) whose forms are listed above in 4.2.3.2 and 4.2.5.2 include: (1) reflexive pron., i.e. d.o. or ind. o. pron. of the same person as the actor. Verbal phrases containing one of these pron.— the so-called pronominal verbs— carry the meanings: i. reflexive proper, e.g. /se bano./ 'il se baigne', /se demando./ 'il se demande' ii. dative of interest, e.g. /me minji aso./ 1je mange ceci' iii. reciprocal, e.g. /nus parlan./ 'nous nous parIons' iv. passive, e.g. /se bej./ '9a se voit' v. special lexical meaning, e.g. /s ennyzo./ 'il s'ennuie'. (2) accusative and dative pron. and pron. adv. Pron. compl. precede the finite verb form (full verb or auxiliary) except in the affirmative imperative, e.g. /luz baki./ 'les voila', /lu porto./ 'il le porte', /luz a pres./ 'il les a pris’, /t afanes pas./ 'ne te depeche pas'. In 393 the affirmative imperative these compl. follow the head, e.g. /porto lu./ 1porte-le', /porto ne./ 'portes-en', /afano te./ 'depeche-toi'. In this position the fern. acc. pron. /la/ 'la' and the neuter acc. pron. /aw/ 1le' appear in their enclitic forms /lo/ 'la' and /ot/ 'le' respectively, e.g. /porto lo./ 'porte-la', /pirtozit./ 'porte-le'. Barring two exceptions to be noted below, pron. compl. are arranged in the general order: ind. o. pron. t d.o. pron. - ne 'en' - i2 'y' Examples: /me lu porto./ 'elle me le porte1, /nu laz a purtados./ 'il nous les a portees', /te laz a paz enkero purtados./ ’il ne te les a pas encore portees1, /j aw porto./ 'il le lui porte', /n i porto./ 'il y en porte', /porto me lu./ 'porte-le-moi', /porto nu los./ 'porte-les- 10 1 nous', /porto n i./ 'portes-y en', /p6rtozjot./ 'porte- le- lui'. Divergences from the prevailing order above are as follows: the ind.o. pron. ij. 'lui' follows d.o. pron. other than the neuter d.o. aw ~ ot 'le', e.g. /l i porto./ 'il le lui porte', /n i porto./ 'il lui en porte', and the pron. adv. ijg 'y' precedes the neuter d.o. aw ~ ot_ ' le' , e.g. /j •^But in colloquial RFr. the order of the pronouns in the last two examples coincides with that of the dialect, i.e. 'porte-me-le', 'porte-nous-les'. 394 aw porto./ 'il l'y porte', /m j aw porto./ 'il me l'y porte', /portozjot./ 'porte l'y'. As shown in the table on p. 395 (phonemic notation), coupled pron. compl. may be juxtaposed (e.g. /melu/ 'me le'), subject to general sandhi (e.g. /te/ + /aw/ > /taw/ 'te le'), or combined in various ways involving irregular phonemic shapes. The following pairs of pron. compl. occur combined: Combinations standing for non-occurring nun 'nous en' nus ne bun 'vous en' bus ne lazi 'la leur' la luzi lazi ' les leur' las lazi In these combinations: (1) the oblique pron. /nus/ 'nous', /bus/ 'vous' have re spectively the variants /nu/, /bu/ (2) the pron. adv. /ne/ has the variant /n/ (3) the acc. fern. pron. /la/ 'la' and /las/ 'les' are both represented by /a/ replacing /u/ of the dat. masc. /luzi/ 'leur', hence the single fern, shape /lazi/ 'la leur, les leur'. The following morphemes have variants determined by me te i i nus bus luzi ne 'me' ' te' ' lui' V 'nous' 'vous' 'leur' ' en' lu ' le' me lu telu li li nulu bulu luzi len la 'la' me la tela li li nula bula lazi len lus 1 les' me lus telus luzi luzi nulus bulus luzi luzen las ' les' me las telas lazi lazi nulas bulas lazi lazen aw ' le' maw taw jaw jaw nuzaw buz aw luzjaw _H ne ' i en men ten ni ni nun bun luzin HThis combination does not occur. LO L O Ln 396 position relative to other pron. compl.: (1) the oblique pron. /nus/ 'nous', /bus/ 'vous' have re spectively the variants /nu/, /bu/ before acc. pron. with 12 initial /l-/, e.g. /bulu/ 'vous le'. (2) the pron. adv. r i g . 'en' has the forms /ne ~ n ~ en/. The regular form /ne/ occurs when single in a verb phrase both in proclitic and enclitic positions and when coupled with other pron. compl. only in pair-initial position, thus /ni/ 'lui en', /ni/ 'y en'. The variants /n/ and /en/ both occur in pair-final position: /n/ in the combinations above and after /luzi/ 'leur', /en/ after other pron. compl., e.g. /men/ 'm'en', /lazen/ 'les en'. (3) the acc. masc. pron. /lu/ 'le' and /lus/ 'les' both have zero variants before the dative /luzi/ 'leur', hence the homophonous forms /luzi/ 'leur', /luzi/ 'le leur', and /luzi/ 'les leur'. As previously described in 5.1.2.IV, certain pron. compl. are joined, without alteration in phonemic shape, by means of the phrasal connective /z/. 12 The loss of /-s/ instead of its expected replacement by /-z/ is not phonologically conditioned since the sequence /zl/ occurs elsewhere at word border-points, e.g. /nuz leban./ 'nous nous levons'. 397 B. Center phrases These phrases fill the position of a single finite vb. They are of the following types: (1) Perfect phrases consisting of a form of auxiliary esta Irr. 1etre1 or abe Irr. 'avoir' (as required by the main 13 vb.) plus past. part, of the main vb. These phrases, or compound tense forms, signal perfect phase. All simple tense forms have matching perfect phrases (see 4.2.4.2.Ill). Auxiliary esta Irr. 'etre' is required by the following (lexically defined) vbs. when used intransitively, i.e. without a d.o.: ana Irr. 'aller1 dintra 'entrer1 ariba 'arriver' esta Irr. 'etre' benge Irr. 'venir, devenir' munta 'monter' damUra 'rester, demeurer' mUri Irr. 'mourir' debala 'descendre' n£se 'naitre' , ■ desende 'descendre' parti 'partir' l3While the main verb is always stressed, auxiliaries are sometimes stressed, sometimes unstressed. The only re mark which can be made at this time is that monosyllabic forms of auxiliaries are usually unstressed, whereas pluri- syllabic forms tend to be stressed. It is not possible without further study of this feature of accentuation to establish to what extent, if any, stress is here governed by rhythm. 398 pasa ‘passer1 tumba 1tomber1 • , sUrti 1sortir1 tUrna 1revenir1 The past part. shows gender and number concord with the subject, e.g. /sun turnats./ 'ils sont revenus1, /ez estado (untuzo)/ 'elle a ete (honteuse)1. Auxiliary esta Irr. 'etre' also forms perfect phrases of pronominal vbs. Concord of the past part, is here with a preceding direct object, e.g. /se sun afanados./ 'elles se sont depechees1, /nus sem bestits./ 'nous nous sommes ve- tus 1 . All other vbs. require auxiliary abe Irr. 'avoir' in perfect phrases. Concord of the past part, is as stated for pronominal vbs., e.g. /laz abjo kantados./ 'il les avait chantees'. (2) Passive phrases containing a form of auxiliary esta Irr. 'etre' plus past part, of the main vb. The past part, con cords with the subject— in this case distinct from the ac tor— e.g. /ez minjat./ 'il est mange', /(la kar) ez minjado/ '(la viande) est mangee'. Where auxiliary esta itself ap pears in a perfect phrase, the resulting phrase has the com bined meanings perfect and passive, e.g. /ez estab bir&t./ 399 'il a ete tourne'. In such phrases both past part, concord with the subject, e.g. /(la tawlo) ez estado birado./ '(la table) a ete tournee'. (3) Repetitive phrases made up of a form of auxiliary tUrna'a nouveau' plus inf. of the main vb., e.g. /turnap pasa./ 'repassez', /la torno lawra./ 'il la laboure a nou veau', /sun turnados pasa./ 'elles sont repassees'. (4) Immediate future phrases composed of a form of auxiliary ana Irr. 'aller' plus inf. of the main vb., e.g. /baz bate./ 'tu vas depiquer', /i ban demanda./ 'ils vont lui demander', /se ba kale leba./ 'il va falloir se lever'. (5) Immediate past phrases comprising a form of auxiliary bencre Irr. ' venir' plus prep, de 'de' plus inf. of the main vb., e.g. /benjoj de kawzi./ 'je venais de choisir', /la ben de kupa./ 'il vient de la casser'. (6) Causative phrases containing a form of auxiliary .fa Irr. 'faire' plus inf. of the main vb., e.g. /las fazen kreba./ 14Auxiliaries being tool-words semantically different from their full-verb homophones, they are to be entered in the lexicon as separate items, thus tUrnai 'a nouveau', tUrnao ’revenir'. 400 'nous les faisons mourir', /l a fejto rina./ 'il l'a fait gronder'. (7) Two-verb phrases constructed of a full vb. plus inf. of another full vb. , e.g. /podon ruziga./ 'ils peuvent ronger'. These phrases, which function as single units, are treated as center phrases because of the position of pron. compl. Any and all such compl. accompanying either of the two vbs. are preposed to the entire phrase, e.g. /aw dibem paga./ 'nous devons le payer', /ni podi bala./ 'je peux lui en donner', /laz benez beze./ 'tu viens les voir', /ti kadra ana./ 'il te faudra y aller', /sen kaw mawfiza./ 'il faut s'en mefier', /bu la bol fa./ 'il veut vous la faire'.^ IX. Complements in verbal phrases may be: (1) a nominal or a pronominal phrase used as direct object after transitive vbs. In this function a nominal phrase is replaceable by: -^Informants also accept and occasionally use an al ternate wording like /bol bu la fk./ 'elle veut vous la faire' with pronouns in the standard Fr. position. This is definitely not the Dzcs. pattern. Evidence is furnished by a good many common phrases such as /se kaw leba./ 'il faut se lever', /s i ba kale mete./ 'il va falloir s'y mettre' which have fixed word order. 401 a. the acc. pron. 1^ provided NM is a def. art., a demon- str. or a poss. adj., e.g. /regardaben lu sew./ 'nous re- gardions le ciel' : /lu regardaben./ 'nous le regardions', /semblo sa so./ 'elle ressemble k sa sceur' : /la semblo./ 'elle lui ressemble', /ajmi akelos kulus./ 'j1aime ces cou— leurs' : /laz ajmi./ 'je les aime'. b. the pron. adv. ne 'en' provided NM is the partitive art., e.g. /purtet de kukets./ 'elle porta des gateaux' : /ne purtet./ 'elle en porta'. Modifiers preposed to the parti tive art. are retained and placed in the verbal core, e.g. /an embykat pla d awkos./ 'elles ont gorge beaucoup d'oies' : /n an pla embykat./ 'elles en ont beaucoup gorge', /faskuri ym moz de sawsiso./ 'je fis un peu de saucisse' : /ne faskuri ym mos./ 'j'en fis un peu'. Only the N is re placeable by ne 'en' if NM is a numeral or one of the adj.: kawke 'quelque', nad 'aucun', forso* 'bien des1, e.g. /abjan tres klukos./ 'nous avions trois poules couveuses' : /n abjan tres./ 'nous en avions trois', /j a forsoz myrgetos./ 'il y a bien des souris1 : /n j a forsos./ 'il y en a bon nombre1, /krumpan nat remedi./ 'nous n'achetons aucun remede1 : /ne krumpan nat./ 'nous n'en achetons aucun1. The demonstr. pron. aso 'ceci' and ako 'cela' are re placeable by the acc. neuter pron. ~ ot 'le', e.g. /preni- 402 aso./ 'je prends ceci' : /aw preni./ 'je le prends'. (2) exocentric phrases with nominals as axes. Such phrases fulfill one of the following functions: a. dative complement. These phrases are opened by the prep, a 'a' and are replaceable by the dative pron. ii_ 'lui', e.g. /parlet aw patrun./ 'il parla au patron1 : /i parlet./ 'il lui par la', /eskribi a tus frajs./ 'j'ecris a <tes freres' : /luzj eskribi./ 'je leur ecris'. b. indirect object. These phrases are opened, as required by the transitive indirect vb. , by the prep, de 1de' or a 'a' (with little or no lexical meaning). Phrases opened by de 'de' are replaceable by the pron. adv. ne. 'en', e.g. /se ser d aket ytis./ 'il se sert de cet outil' : /sen ser./ 'il s'en sert', /buz rapele/ d elo./ 'vous vous la rappelez' : /bun rapel£/./ 'vous vous la rappelez'. Phrases opened by a 'a' are replaceable by the pron. adv. i2 'y', e.g. /pensan aw kyre./ 'nous pensons au cure1 : /i pensan./ 'nous pensons a lui', /respundi a laz letros./ 'je reponds aux lettres' : /i respundi./ 'j'y reponds'. c. agentive complement. These phrases occur after passive vb. phrases and they are opened by the prep. per3 'par', e.g. /ez estab blasat per yn kasajre./ 'il a ete blesse par 403 un chasseur', /sej estado swanado per de syrs./ 1j1ai ete soignee par des sceurs' . (3) a N, an adjectival, or a poss. pron. used as subjective complements of a potentially present subject. These comple ments follow linking vbs., e.g. /ez estaf fawre./ 'il a ete forgeron', /bem masanto./ 'elle devient mechante', /tumberi pla malaw./ 'je tombai tres malade', /ez mjo./ 'elle est a moi', /es pis trok kwej./ 'ce n'est pas trop cuit'. An un modified N or adj. is replaceable by the pron. adv. ne 'le', thus /n ez estat./ 'il l'a ete', /ne ben./ 'elle le de- , , 16 vient . (4) a N, an adjectival, or a poss. pron. used as objective complements, e.g. /truban ako untus./ 'nous trouvons cela honteux', /lu pensi oneste./ 'je pense qu'il est honnete', /la krezjoj sjo./ 'je croyais qu'elle etait a lui', /apelon sa drollo marselo./ 'ils appellent leur fille Marcelle'. (5) an infinitive verbal connected, as required by the main vb. with this type of complements, by the particles de 'de' ^Colloquial RFr. glosses for these phrases would read: 'il en a ete', 'il en devient', with 'en' for standard Fr. ' le' . 404 or a. 'a', both void of lexical meaning and often inter- * 17 changeable or omitted, e.g. /ijmi (a) kanta./ 'j1aime chanter', /krezjo (de) feni./ 'il croyait finir', /adyzan (a) nosto pij./ 'nous aidons notre pere', /s afanon de (~ a) brena./ 'ils se depechent de vendanger', /kumenso de (~ a) plewre./ 'il commence a pleuvoir'. Pron. compl. accompany ing the inf. precede the inf., e.g. /ijmi (a) la kanta./ 'j'aime la chanter', /krezjo (d) aw feni./ 'il croyait le finir'. Ill. Attributive complements There is a limited number of nominal phrases of high frequency which have adverbial meaning and position, but which are unlike prepositional phrases in that no introduc tory prep, is present, e.g. /(nus pawzan) lu dimese/ '(nous nous reposons) le dimanche', /(kuntyneren) dyz u trej zurs/ '(nous continuerons) deux ou trois jours', /(se l£bo) kado nej/ '(il se leve) chaque nuit', /(kosto) bin frans/ '(g:a coute) vingt francs', /(a lawrat) tuto sa bito/ '(il a la bour e) toute sa vie', /(sum bengyts) lu mes pasat/ '(ils sont venus) le mois dernier'. These phrases follow other 17There is no connective with a nominal complement, e.g. /ajmi la kansun./ 'j'aime la chanson'. 405 complements and modify the entire construction, thus /(netezi 1 estaple) lu matin/ '(je nettoie l'etable) le matin', /(zugan a las kartos) lu se/ '(nous jouons aux cartes) le soir'. IV. Attributes of verbals are of two main types: 18 (1) adverbials which normally follow the finite verb form, i.e. single vb. or auxiliary. Modifiers may be: a. the negative adv. pas 'pas', e.g. /pudep pas./ 'vous ne pouvez pas', /a paz brespalat./ 'il n'a pas fait quatre heures'.^ b. other negative adv., e.g. /turnep plys pasa./ 'il ne repassa plus', /parlo gajre./ 'il ne parle guere', /i bezi guto./ 'je n'y vois goutte'. The negative adv. zamej 'ja mais', ~ gajre 'guere', and guto 'goutte' may be reinforced ■^However, certain lexically defined adv. like aze 'hier', dusomen 'doucement' follow the entire verbal phrase, e.g. /s es tjat aze./ 'il s'est tue hier'. l9The occasional occurrence of /n/ before the verb as in, say, /(n) ej paz de pape./ 'je n'ai pas de papier' is due to the intrusion of the Fr. particle /ns/ ~ /n/ and is observable mostly in fictitious speech. Furthermore, this /n/ appears, if at all, only before a verb with initial vowel and when confusion with the Dzcs. pron. adv. ne 'en' is prevented by the context, for example by the presence of a nominal complement as in the statement just cited. 406 by the intensifier pas 'pas', e.g. /bew paz zamej./ 'il ne boit jamais', /a paz eskribyt gajre./ 'il n'a guere ecrit1. A negative adv. following a vb. inflected for persons 2, 4, or 5 pres. S. yields a prohibitive phrase, i.e. a construction which expresses a negative command. In such a phrase, the tense inflection /~e/ of the vb. has the free variant /-e/ in persons 4, 5, e.g. /benges pas./ 'ne viens I pas', /ajfp plys pow./ 'n'ayez plus peur'. c. other adv., e.g. /an tr&k kridat./ 'ils ont trop crie' . The position in a verbal phrase of one modifying adv. may be 20 filled by two to some three adv. in this order: i. adv. of first order (i.e. closest to the vb. head or first in a series): pas 'pas', plys 'ne ... plus' ii. adv. of second order: zamej ’ne ... jamais', tuzur 'toujours' iii. adv. of third order: pru 'assez', pla 'tres, beau coup', trop 'trop' I I | iv. adv. of fourth order: biste 'vite', bjen 'bien', maw 20a thorough treatment of this matter would be too lengthy to present here. The listing below is merely meant to substantiate the existence of a rigid order in strings of adv. It may be noted that semantically the adv. of first order seem to be negative adv., those of second order adv. of time, those of third order adv. of quantity, those of fourth order adv. of manner. 407 'mal', dusom^n 'doucement*, and other adv. in -omen, e.g. /marso plys pru biste./ 'elle ne marche plus assez vite', /ej tuzur pla maw kumpres./ 1j1ai toujours tres mal com- pris'. d. adverbial phrases (see below). (2) one or more (exocentric) prepositional phrases which follow the verbal (see 5.2.2.1), e.g. /s afano end akaba./ 'il se depeche pour terminer', /parlabi damb et./ 'je par- lais avec lui', /aw dizi per fire./ 'je le dis pour rire', /aribon aw mez de zyn./ 'elles arrivent au mois de juin', /sen tornon a 1 entrado de 1 iber./ 'elles repartent a 1'entree de l'hiver'. A phrase opened by the prep. £e_ 1de' indicating pro venience is replaceable by the pron. adv. ne 'en', e.g. /benez d akiw./ 'tu viens de la' : /ne benes./ 'tu en viens', /benen d entanet./ 'nous venons de chez lui' : /ne benen./ 'nous en venons', /tornon du bilaje./ ’ils revien- nent du village' : /ne tornon./ 'ils en reviennent1. A phrase opened by any prep, indicating location is replace able by the pron. adv. i_ 3 'y', e.g. /pundon debal 1 emban./ 'elles pondent sous le hangar' : /i pundon./ 'elles y pon- dent', /klaw din la tino./ 'il contient dans la cuve' : /i 408 klaw./ 'il y contient1, /dintreren enta la margerito./ 'nous entrerons chez Marguerite* : /i dintreren./ 'nous y entre- rons * . 5.2.1.5. Adverbial Phrases These are attributive phrases which occur chiefly be fore adv. and before adj. (see 5.2.1.2.[2] above), and, as described here, after verbs. The head is an adv. Preposed attributes include: (1) a comparative adv., e.g. /mej tar/ 'plus tard'. Com parative and superlative constructions with adv. heads have the same structure as those with adj. heads (see 5.2.1.2. [3-5] above), e.g. /pla masantomen/ 'tres mechamment', /pas pla len/ 'pas tres loin', /lu mej suben/ 'le plus souvent', /mej biste ke buljo/ 'plus vite qu'il ne voulait', /ta lew kurao luz awtes/ 'aussi tot que les autres'. The adv. bien 'bien' does not occur preceded by mej 'plus'. The comparative degree of this adv. is expressed by a different morpheme, the adv. milu 'mieux', e.g. /anam milu./ 'nous alions mieux'. (2) other adv., e.g. /tud dusomen/ 'tout doucement', /talomem prose/ 'tellement pres', /plys trob bj&n/ 'plus 409 trop bien'. For orders in a string of adv., see 5.2.1.4.IV. (1)c above. Postposed attributes are of the following types: (3) nouns, e.g. /dumam matin/ 'demain matin', /aze matin/ 1hier matin'. (4) exocentric attributes, which include: a. prepositional phrases, e.g. /anej a mjedzur/ 'aujour- d'hui a midi, /duman de se/ 'demain soir', /pru per yn kop/ 'assez pour une fois', /len du prat/ 'loin du pre', /awtur du klot/ 'autour de la mare' b. subordinate clauses opened by the conj. 'que', e.g. /aze k ero bandat/ 'hier qu'il etait ivre', /aro ke fa surel/ 'maintenant qu'il fait soleil', /tan ke kupet sa kadeno/ 'tant qu'elle cassa sa chaine'. 5.2.1.6. Phrasal Prepositions Like simple prepositions, these phrases serve as open ing constituents in exocentric phrases (see below 5.2.2.1). Some common phrasal prepositions are here arranged accord ing to the nature and number of their constituents. (1) Prep. + adv. de apej 'd'apres' perx desys 'par-dessus' per], dare 'par-derriere' (2) Prep. + N a par 'sauf' a trabes 'a travers1 (3) Particle + prep.. i afin de 'afin de' dinko a 1jusqu1 a 1 pramo de ’a cause de' (4) Prep. + prep. + adv. de perx dare 'vers l'arriere de' de per! debat 'en-dessous de' de per! daban 'a l'avant de' (5) Def. art. + N + prep, lu lun de 'le long de' 410 (6) Prep. + N + prep. a forso de 'a force de' a kustat de 'a cote de' a kawzo de 'a cause de' de fasun a 'de fagon a' de pow de ' de peur de' (de) kap a 'en direction de* en faso de 'en face de' (7) Prep. + def. art. + N + prep. t a la fin de 'a la fin de1 aw kap de 'au bout de' aw mjej de 'au milieu de' aw rebes de 'a l'envers de' aw inujen de ’au moyen de' su pyn de 'sur le point de'21 2 Alternately, a construction such as /aw mjej de la ruto/ 'au milieu de la route' may be cut before /de/ and regarded as containing two successive prepositional phrases, the second being then attributive to the first. It is per haps preferable, however, to analyze the word groups listed under (5), (6), and (7) as phrasal prepositions, that is, as single functional units, since at least some of these are fixed phrases, closed constructions with distinctive 411 Examples: /d apej lu sekretari/ 'd'apres le secretaire1, /per desyz la kyberto/ 'par-dessus la couverture1, /dink a la bi.no/ 'jusqu'a la vigne1 , /pramo du masan tens/ 'a cause du mauvais temps', /de per dare lu kot/ 'vers l'arriere du cou', /kap aw pun/ 'en direction du pont', /aw kad de trez mezis/ 'au bout de trois mois', /su pyn de s eskana/ 'sur le point de s'etrangler'. 5.2.1.7. Phrasal Conjunctions These phrases introduce subordinate clauses. All phrasal conjunctions have the con j . ' que' as last con stituent. Some common ones are arranged below according to the nature and number of other constituents. (1) Prop. aban ke 'avant que' dympej ke 'depuis que' ende ke 'afin que' penden ke 'pendant que' pera ke 'pour que' san ke 'sans que' (2) Past part, atendyt ke 'vu que' I purbyt ke 'pourvu que1 (3) Particle I afin ke 'afin que' pramo ke 'a cause de ce que' I dinko ke 'jusqu'k ce que' meanings, e.g. /de kap a/ 'en direction de' as opposed to, say, /aw kap/ 'a la tete', /pu kap/ 'sur, par la tete'. 412 (4) Prep. + adv. (6) Prep. + NM + N a mens ke 'a moins que' a la kundisjun ke condi tion que 1 per paw ke 'pour peu que' du mumen ke 1du moment que1 per tan ke 'pour autant que' (5) Prep. + n a prupursjun ke 'a mesure que' de fasun ke 1de facon que1 de pow ke 'de peur que' en kas ke 'en cas que' Examples: /ende ke trabalese/ 'afin qu'il travaillat',, /penden k eren deforo/ 'pendant que nous etions dehors', /atendyk k ero malawzo/ 'vu qu'elle etait malade', /de pow l I » ^ . | ke benguse asiw/ 'de peur qu'il vint ici', /a la kundisjun ke page/ 'a condition qu'il paie'. 5.2.2. Exocentric Phrases 5.2.2.1. Prepositional Phrases These phrases function mainly as attributes in nominal phrases, adjectival phrases, pronominal phrases, verbal phrases, and adverbial phrases, and as sentence-modifiers. The immediate constituents are prep. + axis. The ini tial constituent may be one of the simple prep, listed in 413 4.2.5.3 above or one of the phrasal prep, exemplified in 5.2.1.6. The following words or constructions occur as axes: (1) a bare N, e.g. /dambe plaze/ 'avec plaisir', /d abitydo/ 'd'habitude', /dympej dibendres/ 'depuis vendredi', /em patwes/ 'en patois', /a forso de trabal/ 'a force de tra vail ' . Quite a few nouns occur exclusively as axes of prep, in fixed phrases. They are identified as nouns by their in flectional or their derivational characteristics, by mega- tive criteria, or by a combination of these criteria. Here are some examples: /de kuntyn/ 'd'affilee', /de tiro/ 'd'un trait', /a pits e a pelados/ 'par saccades, tant bien que mal', /en rengileto/ 'a la queue leu leu', /de fame/ 'de faim (manque de nourriture)*, /de sjek/ 'a la suite'. (2) an adj. + N, e.g. /dim mjez uro/ 'dans une demi-heure', /de bun imu/ 'de bonne humeur', /penden aket tens/ 'pendant ce temps'. (3) a nominal phrase, e.g. /a la fjero/ 'a la foire', /d yn kop/ 'd'un coup', /syllu trestet/ 'sur le treteau', /pu kap/ 'sur la tete', /pes pots/ 'sur les levres', /aw kufin/ 'au 414 coin du feu', /per d arzen/ 'pour de l'argent', /en sim minytos/ 'en cinq minutes', /entaz awtes/ 'chez les autres1. (4) a numeral, e.g. /a dys/ 'a deux', /per dizen&w/ 'pour dix-neuf'. (5) a pron., e.g. /kum ako/ 'comme £a', /a par et/ 'sauf lui', /kuntro tutis/ 1contre tous', /de k1/ 'dont, duquel'. (6) a verb in pres. part, with or without its own comple ments or attributes. The prep, sn 'en' opens gerundive phrases, e.g. /en rizen/ 1en riant', /en se rebiran/ 'en se retournant', /en f&n la garbero/ 'en faisant le gerbier', /en se leban de matin/ 'en se levant de bon matin'. (7) a verb in past part., e.g. /a tengyt/ 'continuellement' (8) a verb in infinitive with or without its own complements or attributes, e.g. /a forso de trabala/ 'a force de tra- vailler', /ende se kawfa/ 'pour se chauffer', /de pow de bryka sum bejre/ 'de peur de heurter son verre'. The prep. aban 'avant' is linked to this type of axis by the connec tive deP'de', e.g. /aban de basela/ 'avant de faire la 415 1- . . 22 vaisselle . (9) an adv., /de len/ 'de loin', /per luntens/ 'pour long- temps', /dympej arse/ 'depuis hier soir', /aban duman/ 'avant demain1. (10) a prepositional phrase, e.g. /de din las pesos/ 'de dans les champs', /a par a zu/ 'sauf a moi', /d entaw fawre/ 1de chez le forgeron'. (11) a clause, e.g. /d apej so ke dizjo/ 'd'apres ce qu'il disait', /per urn budrej/ 'par ou je voudrai'. Most prep, can occur with several but not all types of axes. Prep, such as a 'a', de 'de', pern 'pour', sin 'sans' appear with the largest variety of axes. On the contrary, , I the prep, ende 'avec 1'intention de1, afin de 'afin de', de fasun a 'de fagon a', de pow de 'de peur de', su pyn de 'sur le point de' admit only infinitive verbs as axes. 5.2.2.2. Phrasal Sentence-Linkers The constituents of these phrases may be as follows: (1) prep. + N, e.g. en efet 'en effet', a prepaw 'a propos' 22The connective de ’de' is void of lexical meaning and absent when the axis is a nominal, e.g. /aban la baselo/ 'avant la vaisselle'. 416 (2) prep. + adv., e.g. de alyrs 'd'ailleurs' (3) prep. + def. art. + N, e.g. aw kuntriri 1au contraire'. 5.2.2.3. Phrasal Attributive U.ni.veE5aIs The phrase pas ... ke 'ne ... que' comprising the nega tive adv. pas 'pas' and the conj. ke 'que' follows a verbal but the construction is discontinuous when the verbal is a perfect phrase, thus /(ajmo) pas ke ty/ '(il) n'(aime) que toi', /(faj) pas ke kure/ '(il) ne (fait) que courir', /(es) pas ke fenanto/ '(elle) n'(est) que paresseuse', /(n j a) pas ke naw/ '(il) n'(y en a) que neuf', /(a) paz (minjit) ke de pan/ '(il) n' (a mange) que du pain', /(boli) pis ke demanda yn serbisi/ '(je) ne (veux) que de- mander un service', /(bem) pas ke kan a set/ '(il) ne (vient) que quand il a soif'. 5.2.2.4. Absolute Expresssions These expressions function as sentence-modifiers. The immediate constituents are a nominal plus a past part, showing concord with the nominal. Such a construction is followed by juncture and is equivalent to a subject-predi- cate construction, e.g. /la ly/ eskantido -* (sen angut de la krambo)/ 1 la lumiere eteinte, (il quitta la chambre)', 417 /laz bretoz mortos - » (kalut krumpa la lej)/ 'les vaches laitieres raortes, (il fallut acheter le lait)'. B I B L I O G R A P H Y 418 BIBLIOGRAPHY Agard, Frederick. Structural Sketch of Rumanian. Balti more: Linguistic Society of America, 1958. (Supple ment to Language. Language Monograph, No. 26.) Alarcos, Llorach, Emilio. Fonologia espanola. Madrid: Editorial Gredos, 1954. Austin, William M. "Criteria for Phonetic Similarity," Language. XXXIII (1957), 538-544. Barrenechea, Ana Maria. "Las clases de palabras en espanol, como clases funcionales," Romance Philology. XVII (1963), 301-309. Bazell, Charles Ernest. Linguistic Form. Istanbul: Istanbul Press, 1953. (Istanbul ttniversitesi, Edebiyat Faktlltesi, Yayinlarindan, No. 574.) ________________________. "The Choice of Criteria in Struc tural Linguistics," Word. X (1954), 126-135. Bloch, Bernard. "A Set of Postulates for Phonemic Analysis," Language. XXIV (1948), 3-46. --------------- , and George L. Trager. Outline of Linguistic Analysis. Baltimore: Linguistic Society of America, 1942. Bloomfield, Leonard. Language. London: G. Allen & Unwin Ltd., 1933, rev. 1935. Bolinger, Dwight L. "Complementation Should Complement," Studies in Linguistics. VIII (1950), 29-39. 419 420 Bolinger, Dwight L. "On Defining the Morpheme," Word. IV (1948), 18-23. . "Identity, Similarity, and Difference," Li tera. I (1954), 5-16. Bonfante, G. Rev. of Descriptive Italian Grammar by R. A. Hall, Word. V (1949), 90-91. Bourciez, Edouard. Elements de linguistigue romane. 4th ed. rev. Paris: C. Klincksieck, 1946. Bowen, J. D., and Robert P. Stockwell. Patterns of Spanish Pronunciation: A Drillbook. Chicago: University of Chicago Press, 1960. ______________________________________ _ "The Phonemic In terpretation of Semi-Vowels in Spanish," Language. XXXI (1955), 236-240. ________________________________________. "A Further Note on Spanish Semi-Vowels," Language. XXXII (1956), 290-296. Canet, Louis. Lectures d'histoire locale sur le Tarn-et- Garonne. 2 vols. Montauban: Paul Masson, 1925. Dauzat, Albert. Les Patois. Paris: Delagrave, 1927. Elcock, W. D. The Romance Languages. New York: Macmillan, 1960. Francis, W. Nelson. The Structure of American English. New York: Ronald Press, 1958. Fries, Charles C., and Kenneth L. Pike. "Coexistent Pho nemic Systems," Language. XXV (1949), 29-50. Gleason, Henry A. An Introduction to Descriptive Linguis tics . New York: Holt, Rinehart, and Winston, 1955, rev. 1961. Haas, W. "Zero in Linguistic Description," Studies in Lin guistic Analysis (Special Volume of the Philological Society). Oxford: B. Blackwell, 1957, 33-53. 421 Hall, Robert A., Jr. French. Baltimore, Md.: Linguistic Society of America, 1948. (Language Monograph No. 24, Structural Sketches, 1.) _____________________ Descriptive Italian Grammar. Ithaca, New York: Cornell University Press and Linguistic Society of America, 1948. (Cornell Romance Studies, Vol. II.) _____________________ "The Linguistic Structure of Taki- Taki," Language. XXIV (1948), 92-117. Harris, Zellig S. "Distributional Structure," Word. X (1954), 146-162. . "From Morpheme to Utterance," Language. XXII (1946), 161-183. _________________ . "Morpheme Alternants in Linguistic Anal ysis," Language. XVIII (1942), 169-180. . Methods in Structural Linguistics. Chi cago:__University of Chicago Press, 1951. ______________ . "Discontinuous Morphemes," Language. XXI (1945), 121-127. Haugen, Einar Ingvald. The Norwegian Language in America: a Study in Bilingual Behavior. 2 vols. Philadelphia: University of Pennsylvania Press, 1953. __________________. "Phoneme or Prosodeme?" Language. XXV (1949), 278-282. Heffner, Roe-Merrill S. General Phonetics. Madison, Wis.: University of Wisconsin Press, 1950. Hockett, Charles F. A Course in Modern Linguistics. New York: Macmillan, 1958. "Peiping Morphophonemics," Language. XXVI (1950), 63-85. "Problems of Morphemic Analysis," Language. XXIII (1947), 321-343. 422 Hockett, Charles F. "A Formal Statement of Morphemic Analysis," Studies in Linguistics. X (1952), 27-39. ___________________ . "Two Models of Grammatical Descrip tion," Word. X (1954), 210-234. ___________________ . "Two Fundamental Problems in Phone- mics," Studies in Linguistics. VII (1949), 29-51. ___________________ . Rev. of Morphology: The Descriptive Analysis Qf Words, by E. a. Nida, Language. XXIII (1947), 273-285. Householder, Fred W., Jr. Rev. of Methods in Structural Linguistics. by Z. S. Harris, International Journal of American Linguistics. XVIII (1952), 260-268. Katz, Jerrold J. "Mentalism in Linguistics," Language. XL (1964), 124-137. King, Harold V. "Outline of Mexican Spanish Phonology," s,tufli.es_in-Ling.ais.bics. , x (1952), 51-62. Marchand, Hans. The Categories and Types of Present-Dav English Word-Formation. Wiesbaden, Germany: Otto Harrassowitz, 1960. Martinet, Andre. La description phonologique avec applica tion au parler franco-provenyal d'Hauteville (Savoie). Geneve: Librairie Droz, 1956. ________________ . "About Structural Sketches," Word. V (1949), 13-35. ________________ . "Occlusives and Affricates with Reference to Some Problems of Romance Phonology," Word. V (1949), 116-122. ________________ . "Un ou deux phonemes," Acta Linguistics. I (1939), 94-103. Navarro Tomas, Tomas. Manual de pronunciacidn espanola. 4th ed. Madrid: Consejo Superior de Investigaciones Cientificas, 1961. 423 Nida, Eugene A. Morphology; The Descriptive Analysis of Words. 2nd ed. Ann Arbor; University of Michigan Press, 1949. (University of Michigan Publications, Linguistics, 2.) . "Field Techniques," International Journal of American Linguistics. XIII (1947), 138-146. . "A System for the Description of Semantic Elements," Word. VII (1951), 1-14. . "The Identification of Morphemes," Lan guage . XXIV (1948), 414-441. . "The Analysis of Grammatical Constituents," Language. XXIV (1948), 168-178. Pike, Kenneth L. Phonemics? A Technique for Reducing Languages to Writing. Ann Arbor; University of Michi gan Press, 1947. . "On the Phonemic Status of English Diph thongs," Language. XXIII (1947), 151-159. Pop, Sever. La dialectologie; aperyu historigue et methodes d'enquetes linguistiques. 2 vols. Louvain, Belgium: J. Duculot, 1950. Ronjat, Jules. Grammaire istorique fsicl des Parlers Pro- vengaux Modernes. 4 vols. Montpellier, France; Soci^te des langues romanes, 1930-1941. Roques, Gabriel. Grammaire Gasconne (Dialecte de 11age- Hais.). Bordeaux, France; Feret, 1913. Saporta, Sol. "Morpheme Alternants in Spanish," Structural Studies on Spanish Themes, ed. Henry Kahane and A. Pietrangeli. Salamanca and Madrid: S. A. Raycar, 1959. (=Acta Salmanticensia. Filosofia v Letras. XII, 15-162.) _______ . "A Note on Spanish Semi-Vowels," Language. XXXII (1956), 287-290. 424 Seguy, Jean. Atlas linguistique et ethnographigue de la Gascogne. 3 vols. Paris: Centre National de la Re cherche Scientifique, 1954- Smith, William B. S. Rev. of Descriptive Italian Grammar and French. by Robert A. Hall, Jr., Studies in Lin guistics . VIII (1952), 5-11. Swadesh, Morris. "On the Analysis of English Syllabics," Language. XXIII (1947), 137-150. Trager, George L. Rev. of Morphology: The Descriptive Analysis of Words, by E. A. Nida, International Jour nal of American Linguistics. XVII (1951), 126-131. __________________. "The Phonemes of Castilian Spanish," Travaux du Cercle Linguistique de Prague. VIII (1939), 217-222. __________________. Rev. of Phonemics: A Technique for Re ducing Languages to Writing, by K. L. Pike, Language. XXVI (1950), 152-158. _________________ . "The Phonemic Treatment of Semi-Vowels." Language. XVIII (1942), 220-223. __________________, and Henry Lee Smith, Jr. "An Outline of English Structure," Studies in Linguisticsr Occasional Papers. 3. Norman: Battenberg Press, 1951. Trubetskoi, Nikolai S. Grundziige der Phonologie. Prague: Cercle Linguistique, 1939. Trans. J. Cantineau, Principes de Phonologie. Paris: C. Klincksieck, 1949, Wells, Rulon S. "Immediate Constituents," Language, XXIII (1947), 81-117. "Automatic Alternation," Language. XXV (1949), 99-116. Zamora Vicente, Alonso. Dialectologia espanola. Madrid: Editorial Gredos, 1960.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Servius' Knowledge Of Juvenal: An Analysis Of The Juvenalian Quotations In Servius' Commentary On Vergil

PDF

Frank Wedekind And The Search For Morality

PDF

The Patriot In Exile: A Study Of Heinrich Mann'S Political Journalistic Activity 1933-1950

PDF

A Descriptive Syntax Of King Alfred'S Soliloquies

PDF

Temporal Structure Of Spoken Korean: An Acoustic Phonetic Study

PDF

Existentialism In The Theater Of Alfonso Sastre

PDF

Rhetoric And Fancy As A Basis For Narrative In The Novels Of Jean Giraudoux

PDF

The Proteges Of Lancelot: A Study Of Malory'S Characterization Of Lancelot In The 'Morte Darthur'

PDF

Albert Camus And The Kingdom Of Nature

PDF

A Syntactic Analysis Of The Writing Of Some Fourth-Grade, Fifth-Grade, And Sixth-Grade Caucasian Children In The Los Angeles Schools

PDF

Agrippa D'Aubigne'S Les 'Tragiques': The Conquest Of Profaned Time

PDF

Narrative And Lyric Originality In The Old French Versions Of "La Vie De Saint Eustache"

PDF

The Appositive In English: A Transformational Analysis

PDF

Theological And Dramatic Concepts Of The End Of Man In The Middle Ages

PDF

An Investigation Of Some Critical Factors In Language Synthesis And The Implications Of These Factors For Linguists As Language Engineers

PDF

A syntactic analysis of the written language of selected Black elementary school children with reference to sociological variables

PDF

The Grammar Of English Causative-Transitivity

PDF

Alfred Of Sareshel'S Commentary On The 'Metheora' Of Aristotle. (Latin Text)

PDF

The Authenticity Of David Of Augsburg'S German Works, With Particular Reference To His 'Paternoster'

PDF

A Syntactic Analysis Of Luther'S 'Adventspostille'

Asset Metadata

Creator Kelly, Reine Cardaillac (author)

Core Title A Descriptive Analysis Of The Gascon Dialect Spoken At Donzac (Tarn-Et-Garonne), France

Degree Doctor of Philosophy

Degree Program Linguistics

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag language, linguistics,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Waterman, John T. (committee chair), Berkey, Max Leslie, Jr. (committee member), Christensen, Francis (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-219941

Unique identifier UC11360615

Identifier 6611575.pdf (filename),usctheses-c18-219941 (legacy record id)

Legacy Identifier 6611575.pdf

Dmrecord 219941

Document Type Dissertation

Rights Kelly, Reine Cardaillac

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA