Lugo 1 The University of the West Indies St. Augustine Campus Faculty of Humanities and Education Department of Modern Languages and Linguistics LING 3099 Special Project in Linguistics COVER PAGE Student names​: Laurisa Lugo Student ID no.​ : ___811003304________ Degree Programme:​ Spanish and Linguistics Supervisors​: Jo-Anne S. Ferreira Title of Projec​t: On the Phonemic Inventory and Syllable Structure of Trinidadian French Creole Declaration 1. I declare that this project has been composed solely by myself and that it has not been submitted, in whole or in part, in any previous application for a degree. Except where stated otherwise by reference or acknowledgment, the work presented is entirely my own. 2. I authorise The University of the West Indies to make a physical or digital copy of my thesis/research paper/project report for its preservation, for public reference, and for the purpose of interlibrary loan. 3. I consent to have my attached project used in any publication comprising Linguistics Projects by The University of the West Indies. I understand that I will receive no compensation. I hereby assign publishing rights for the contribution to The University of the West Indies, including all copyrights. Signature of Student: ___Laurisa Lugo_______ Date: ______09/08/2017________ Lugo 2 On the Phonemic Inventory and Syllable Structure of Trinidadian French Creole Laurisa Lugo LING 3099: Special Project in Linguistics ABSTRACT Trinidadian French Creole (TFC) or Patois is an endangered heritage language of Trinidad which belongs to a group of language varieties referred to as Lesser Antillean French Creole. While TFC has previously been the subject of linguistic study, there has been relatively little investigation into its phonology, thus far. Goodman's contribution to the field is important, but his study of the TFC phonemic inventory, published in 1958, almost 60 years ago, needs to be compared with present-day TFC to determine if any changes have taken place. This study verifies and updates the phonemic inventory described by Goodman. Additionally, it provides an analysis of the syllable structure of TFC, focusing on its syllable template and phonotactic constraints. Data were collected from three native speakers from the villages of Paramin and Talparo and were elicited through a word list, folklore and traditional stories and original narrations from participants. The sole noted change to the phonemic inventory of TFC was the addition of the central alveolar approximant, /ɹ/ which was attributed to English and Trinidadian English Creole influence. The syllable template was determined to be (C)(C)V(C), which indicates that TFC allows complex onsets but only permits simple codas. It was also noted that TFC does not display a tendency toward the CV syllable structure, since phonological restructuring of borrowed lexical items exhibits a pattern of creating closed syllables. These findings are significant since they provide evidence of change in the phonemic inventory of TFC, and it is vital that such change is recorded, due to TFC’s status as an endangered language. Additionally, they illustrate that the syllable structure of TFC is relatively diverse which supports the claim that Antillean French Creoles do not tend only toward a CV syllable structure. Lugo 3 Supervisor: Dr. Jo-Anne S. Ferreira Word Count: 6,725 Table of Contents Chapter Page Number List of Abbreviations 3 List of Tables 4 Chapter 1: Introduction 5 Chapter 2: Literature Review 7 2.1 Greater Antillean French Creole 9 2.2 Lesser Antillean French Creoles 9 2.3 Trinidadian French Creole 11 Chapter 3: Methodology 13 Chapter 4: Results and Discussion 16 4.1 Consonants 16 4.2 The Alveolar Central Approximant 17 4.3 Free Variation 19 4.4 Vowels 19 4.5 The Syllable 20 4.6 Nuclei 21 4.7 Vowel Hiatus 21 4.8 Onsets 22 4.9 Codas 23 4.10 Resyllabification 24 Chapter 5: Conclusion 28 References 30 Lugo 4 Appendices 32 List of Abbreviations C Consonant G Glide GC Guadeloupean Creole HC Haitian Creole LAFC Lesser Antillean French Creole MC Martiniquan Creole SLFC St. Lucian French Creole SSP Sonority Sequencing Principle TrinEC Trinidadian English Creole TFC Trinidadian French Creole V Vowel Lugo 5 List of Tables Title Page Number Table 1: Markedness scale of syllable typology 8 Table 2: Consonant phoneme inventory of TFC 16 Table 3: Phonemic contrast among TFC consonants 17 Table 4: Vowel phoneme inventory of TFC 19 Table 5: Phonemic contrast among TFC vowels 19 Table 6: Types of syllables attested in TFC 20 Table 7: Syllable nuclei in TFC 21 Table 8: Complex onsets in TFC 23 Table 9: Single codas in TFC 23 Table 10: Resyllabified TFC lexical borrowings of Spanish origin 24 Lugo 6 Chapter 1: Introduction Trinidadian French Creole (TFC) or Patois (as it is called by its speakers) was introduced to Trinidad by settlers from other French-speaking Caribbean territories, and has been spoken here since 1783. It once served as the island’s lingua franca but is now endangered, and survives only in a few rural areas (Ferreira and Holbrook 3). TFC belongs to a group of language varieties referred to as Lesser Antillean French Creole (LAFC) which also encompasses the French Creoles of Guadeloupe, Martinique, Dominica, Grenada, St. Lucia and Venezuela, but is differentiated from Greater Antillean French Creole which refers to Haitian Creole (HC). Many linguists would agree that the phonology of Creole languages is an understudied area. Additionally, studies on Creole phonology tend to focus on a small number of themes and are usually diachronic in nature (Klein 3). This certainly seems to be the case in the Caribbean, especially regarding French Creoles. While studies of Antillean French Creole phonology are by no means exhaustive, fewer still have been conducted on TFC (see Goodman 1958, Thomas 1869). Additionally, previous attempts have been made to describe the inventory of sounds of TFC, but with the exception of Ferreira, Drayton and Hodge (in progress), there have been no recent studies in this area. Similarly, while reference has been made to the syllable in previous studies on TFC, its syllable structure has never been discussed at length. In light of this, the current study aims to verify the phonemic inventory of TFC proposed by Goodman (206), and propose an updated one, if necessary. It also seeks to provide a more in-depth analysis of the syllable structure of TFC, including its syllable template and phonotactic constraints. Therefore, the present research will seek to answer the following questions: Lugo 7 1. What is the current phonemic inventory of TFC? 2. Does it differ from the one proposed by Goodman, and how? 3. What is the syllable structure of TFC? This study will contribute toward the documentation of TFC, which is important due to its status as an endangered language. Additionally, the findings of this study can be used as a tool for teaching TFC to non-native speakers in aid of ongoing revitalisation efforts. The remainder of this study is divided as follows. Chapter Two presents a review of the existing literature regarding the phonemic inventories and syllable structure of Antillean French Creoles. Chapter Three outlines the methodological procedures followed in collecting and analysing the data for the current study. Chapter Four contains a presentation and discussion of the findings of the current research and Chapter Five summarises the salient findings and highlights some areas for possible future study. Lugo 8 Chapter 2: Literature Review There has been much debate as to whether Creoles are a typologically distinct set of languages which can be subjected to synchronic study rather than just being described in terms of their historic development. As Klein describes, linguistic typology demonstrates that the grammars of different languages can be shown to be similar in many respects (155-156). Klein has attempted to develop a typological classification of the phoneme inventories of a representative sample of Creole languages. This study provides evidence that Creoles are not less complex than non-Creoles, but typologically average. Based on Maddieson’s study of the typology of phoneme systems in non-Creole languages, which serves as the methodological base for Klein’s study, non-Creole languages typically have an inventory which consists of 20 to 37 phonemes (Klein 163). Klein establishes that the phoneme inventories of the majority of creole languages (97%) fall within this average (Klein 164). Additionally, the vowel inventory of non-Creole languages usually comprises 5 or 6 phonemes (Maddieson qtd. in Klein 168). However, most Creoles have an inventory just above average at 7, and Creole languages with 5 to 7 phonemic vowels make up 75% of the languages studied (Klein 168-169). Moreover, the majority of Creole languages in Klein’s sample exhibit two plosive series, that is, one of voiced and another of voiceless plosives. Therefore, Creole languages show a narrower range of plosive series than Lugo 9 non-Creoles, most of which display between one and six series of plosives1 (Klein 171-172). As a result, Klein concludes that Creole languages appear to be more phonologically uniform than non-Creoles since most of their phoneme inventories fall within the average and their vowel and plosive systems display relatively less variation (184). Therefore, it is expected that the phoneme inventories of Antillean French Creoles follow this pattern. There have been many claims that Creoles, much like their substrate languages, show a tendency toward a CV syllable structure (e.g., Holm 144). Phonological restructuring of European lexical items in Creole languages is often attributed to their preference for CV sequences. It is noted that phonological processes such as vowel insertion, consonant deletion or metathesis do produce forms that are closer to a CV structure than the corresponding source words (Klein 173). However, phonological restructuring does not uniformly produce CV syllables since restructured words can contain more diverse syllable structures (Klein 174). ​Furthermore, ​it has been acknowledged that some groups of Atlantic Creoles do not exhibit a tendency toward a CV syllable structure, with Atlantic French Creoles being cited in this context (e.g., Holm 144, McWhorter 156). Holm also notes that French-based Creoles show few phonotactic constraints, one notable exception being the absence of word-final consonant clusters (144). In an attempt to analyse their synchronic surface syllable types, Klein uses CV-notation to present syllable templates of a representative sample of Creole languages based on the markedness scale proposed by Levelt and Van der Vijver. It identifies five levels of syllable markedness which increases depending on the number of elements which can be added to the CV baseline (which is unmarked because it is present in all languages). The scale is presented in Table 1 below. Markedness Syllable Template Unmarked CV Marked I (C)V CV(C) C(C)V 1 The six most common plosive series found among non-Creoles include plain voiceless, plain voiced, aspirated voiceless, voiceless ejective. voiced implosive and prenasalized voiced plosives (Klein 171). Lugo 10 Marked II C(C)V(C) (C)(C)V (C)V(C) CV(C)(C) Marked III C(C)V(C)(C) (C)(C)V(C) (C)V(C)(C) Marked IV (C)(C)V(C)(C) Table 1: Markedness scale of syllable typology (adapted from Levelt and Van der Vijver qtd. in Klein 181) Based on Klein’s analysis, the core syllable in Creoles emerges as (C)(C)V(C) or markedness III, with just over half of the languages in the sample, including SLFC, displaying it. This provides evidence that Creole syllable structure is more diverse than previously claimed by Holm (144). In fact, none of the Creoles in Klein’s sample exhibit an exclusively CV syllable structure; rather, a large number of them feature complex onsets or single codas, and a few also exhibit complex codas, including HC. G​ iven that Creoles, inclusive of another LAFC variety, appear to cluster around the upper mid-point of the markedness scale, it is expected that TFC will display a similar syllable structure. 2.1 Greater Antillean French Creole The phonemic inventory of HC consists of six plosives /p, b, t, d, k, ɡ/; seven fricatives /f, v, s, z, ʃ, ʒ, ɣ/; two nasals /m, n/; four approximants /l, j, w, ɥ/; seven oral vowels comprising four front vowels /i, e, ɛ, a/ and three back vowels /o, u, ɔ/ and three nasal vowels /ẽ, ã, õ/ (Fattier ​APiCS Online​; Valdman 62, 68-69). Valdman also includes two phonemic affricates /ʧ, ʤ/ which Fattier lists as minor allophones of /t/ and /dj/, respectively. However, Valdman acknowledges the ongoing debate about whether they constitute unitary phonemes or biphonemic sequences since they alternate freely with the the sequences /tj/ and /dj/, respectively (65). Regarding the syllable, Valdman a​ ttests the syllabic consonants /m, n, l/ in HC, which represent the short forms of the first person singular, first person plural and third person Lugo 11 singular pronouns, respectively (64). It is noted that of the three syllabic consonants, /m/ occurs most frequently in HC (Valdman 65). It would be interesting to note if this also occurs in TFC. 2.2 Lesser Antillean French Creole Corne uses Guadeloupean French Creole (GC) as a point of departure for describing the phonemic inventories of LAFC (130). LAC usually has an inventory of seven oral vowels: /i, e, ɛ, a, u, o, ɔ/; and three nasal vowels: /ɛ̃, ã, õ/. Additionally, in contrast with HC, the glottal fricative, /h/ is established as a phoneme in LAFC. It is also posited that, unlike HC, many varieties of LAFC have two phonemic affricates /ʧ/ and /ʤ/ (/​c, ɉ/2) (Corne 131). Moreover, it is established that there is a relative degree of phonological uniformity among LAFC (Corne 131). Therefore, it should follow that the phonemic inventory of TFC display no significant differences from that of other LAFC varieties. Colot and Ludwig identify identical phonemic inventories for GC and Martiniquan Creole (MC) (A​ PiCS Online)​ . T​ he vowel inventory is the same as that identified by Corne with the exception of the nasal vowel /õ/ which is replaced by /ɔ/̃ . However, Brousseau explains that it is possible that /ɔ̃/ is a phonetic realisation of the underlying phoneme, /õ/ (110). Therefore, it is likely that Corne chose /ɔ̃/ because he considered the open-mid allophone to be more salient. ​The consonantal inventory comprises six plosives /p, b, t, d, k, ɡ/; two affricates /ʧ, ʤ/ (/c, ɟ/); eight fricatives /f, v, s, z, ʃ, ʒ, ɣ, h/; four nasals /m, n, ɲ, ŋ/; and three approximants /l, j, w/ (Colot and Ludwig​ APiCS Online​). With regard to syllable structure, Colot and Ludwig indicate that GC and MC allow onsets that are moderately complex. From the data provided3, it is possible to surmise that onsets can contain two consonants, but the data are too limited to ascertain the phonotactic constraints which govern onset clusters. Regarding codas, both languages permit only simple codas, that is, codas which contain a single consonant (​APiCS Online​). Carrington’s study of St. Lucian French Creole (SLFC) phonology has identified a phonemic inventory comprising 33 phonemes, almost identical to those of GC and MC. However, SLFC’s inventory consists of three nasals /m, n, ŋ/ as opposed to four (Carrington 2 This paper uses modern IPA conventions, therefore, the symbols within the brackets, (/x/), are the ones which are attributed to the original authors. 3 The data provided by Colot and Ludwig to exemplify onset complexity in GC and MC are /klɛ/ ‘bright’ and /kɣas/ ‘grime’ (​APiCS Online​). Lugo 12 56). Moreover, Carrington identifies eight oral vowels in contrast with the seven identified by Corne, Colot and Ludwig. In this regard, he establishes /ɑ/ (/à/) as part of SLFC’s vowel inventory (Carrington 38). A​ dditionally, its nasal vowel inventory is different in that it contains /ẽ/ as opposed to /ɛ̃/ as suggested by Corne. As above, Brousseau acknowledges the possibility that /ɛ̃/ represents a phonetic realisation of the underlying phoneme, /ẽ/ (110). Once more, it is likely that Corne chose the allophone he considered to be more salient. Regarding the syllable, Carrington attests twelve syllable types in SLFC but describes the CV sequence as the core of the language's word structure since it is the most frequently occurring syllable type. While SLFC does not contain exclusively CV sequences, Carrington’s data seem to suggest that SLFC does display a tendency toward the unmarked syllable which has traditionally been associated with Creole languages. Carrington states that the nucleus of the SLFC syllable can constitute a single vowel, a single vowel preceded by the semi-vowels /w/ or /j/, a single vowel followed by the semi-vowel /j/ or a single consonant (38). Similar to HC, the consonants /m, n, s/ constitute syllabic consonants in SLFC (Carrington 42-43). Furthermore, all consonants can appear in onset position in simple onsets (Carrington 25-26). Additionally, with the exception of /h, ɣ (r)4, ʧ, ʤ/ all single consonants can appear in coda position. It is noted that the only cluster which appears in coda position is /tɣ/ (/tr/) which is only attested in one form, /katɣ/ (/katr/) which is a variant of /kat/ ‘four’ (Carrington 30-31). Carrington adds that all vowels can occur in open syllables and all may occur in closed syllables, with the exception of /e/ (32). These findings can serve a point of reference for TFC since work on its syllable structure is not as extensive. 2.3 Trinidadian French Creole Thus far, there has been relatively little work in the area of TFC phonology. Noteworthy contributions the field are Goodman (1958) and Thomas (1869). Much of Thomas’ work focuses on other aspects of TFC, therefore, only a small portion is related to phonology. Additionally, Thomas describes the phonology of TFC from an orthographic perspective, that is, sounds are attributed to letters of the LAFC alphabet used for TFC. As a 4 According to Carrington the velar fricative, /ɣ/ (for which he uses the symbol /r/) has two allophones [ɣ] and [ɣʷ]. Carrington states that, “[l]abialisation of this consonant is a prominent feature of St. Lucian [French] Creole, to such an extent that more often than not [ɣʷ] resembles [the labial velar approximant] /w/.” Carrington also adds that the allophone [ɣʷ] is more widespread than the allophone [ɣ], which is limited to older, rural dwellers (27). Lugo 13 result, it is difficult to ascertain which sounds have phonemic status in TFC’s inventory, according to Thomas. Goodman identifies 31 TFC phonemes; six plosives /p, b, t, d, k, ɡ/; two affricates /ʧ (č), ʤ (j)/; seven fricatives /f, v, s, z, ʃ (š), ʒ (ž), h/; three nasals /m, n, ŋ/; three approximants5 /l, j (y), w/; seven oral vowels; three front vowels /i, e, ɛ/; one central vowel /a/; and three back vowels /u, o, ɔ/; in addition to three nasal vowels /ɛ,̃ ã, ɔ/̃ (208). TFC’s inventory is quite similar to those of GC and MC as identified by Colot and Ludwig, the only difference being the lack of the phonemes /ɣ/ and /ɲ/. There is also very little distinction between TFC’s inventory as posited by Goodman and that of LAFC proposed by Corne. This is in keeping with Klein’s and Corne’s findings regarding Creole uniformity. It is important to note, however, that the phonemic inventory described by Goodman was of the language as it was spoken 59 years ago. Therefore, this inventory may require an update as sustained language contact with English and TrinEC could have resulted in the modification of TFC’s phonemic inventory (see n5). Although, Goodman does present some information on the TFC syllable, it is quite limited in its scope. TFC permits syllables which comprise only the vocalic nucleus and also allows both simple and complex onsets, but onset clusters are limited to two consonants. In the case of complex onsets, the second consonant in the cluster must be /j (y)/, /w/, or /l/ (Goodman 211). Additionally, TFC does not permit consonant clusters in coda position, but all consonants are allowed in simple codas, except /ʧ (č), ʤ (j)/ (Goodman 212). Goodman also states that /e/ and /o/ occur only in open syllables, while all others occur also in closed ones (208). Moreover, Thomas provides information on some phonological restructuring in TFC. However, Thomas’ discussion of the syllable is limited to identifying several processes, including apocope, epenthesis and metathesis, and acknowledging that they affect the number and order of elements in a word (6-8). Therefore, to my knowledge, there has never been a detailed analysis and description of syllable structure in TFC. Consequently, the current study seeks to verify the phonemic inventory of TFC described by Goodman and to update this inventory, if needed. Additionally, it aims to provide a synchronic analysis of the TFC syllable in order to determine its syllable template and present a preliminary description of its phonotactic constraints. 5 Ferreira, Drayton and Hodge (in progress) account for four approximants in TFC (three central and one lateral), and include the alveolar central approximant, /ɹ/ in their minimal set. Lugo 14 This study will contribute to the documentation of TFC, which is necessary due to its status as an endangered language. As Crystal notes, “[d]ocumentation is a s​ ine qua non of language maintenance” (199). Although a language cannot be saved by documentation alone, it is an early priority in all investigations and a top priority in cases where a language is facing impending death (Crystal 199). Crystal notes that “... the preservation of linguistic diversity is essential, for language lies at the heart of what it means to be human” (44). Additionally, language preservation is important since language loss impacts not only the individual speakers (through the loss of their personal history), but also the wider society through the loss of inherited knowledge (Crystal 44). Moreover, the findings of this research can also provide valuable information which can be used in teaching TFC to those with little or no knowledge of the language, in aid of revitalisation efforts. I​ n this regard, a well-established phonemic inventory can be used a basis for teaching native-speaker pronunciation to non-native speakers of TFC. Additionally, this study will aid in an understanding of the phonotactic constraints of TFC which is also useful for teaching the language. As Cook notes, knowledge of the syllable structure of the target language is vital to language acquisition since phonotactic constraints can pose difficulty to foreign language acquisition, as learners often apply the constraints of their native language to the target language (74-75). Chapter 3: Methodology This study is situated within the field of qualitative research. The process of collecting qualitative data includes identifying and selecting participants and sites to be studied and obtaining permission to access them, selecting appropriate sampling strategies, determining Lugo 15 the type of data needed to answer the research question, designing protocols and instruments for data collection and ensuring that ethical considerations are made during the collection process (Creswell 205). For the purpose of data collection, purposive sampling was used. This type of sampling involves researchers intentionally selecting the participants and sites best able to provide data to describe a particular phenomenon (Creswell 206). In the context of this study, native speakers who fit specific criteria were chosen as participants (to be discussed below). Additionally, data were collected in communities where TFC is still spoken; the villages of Paramin and Talparo. Paramin was chosen since it is the largest community of practice of TFC in Trinidad (Ferreira and Holbrook 6). Moreover, Talparo was included because relatively little research has been conducted there. Data were collected over two sessions in Paramin and in one session in Talparo. With regard to participant selection, Bowern identifies the ‘ideal’ speaker as a monolingual speaker of the language under investigation who is neither too old nor too young. However, it is admitted that, in practice, there may be no speakers who fit this criterion since multilingualism is widespread in many parts of the world and in the case of an endangered language, there may not be any young speakers (Bowern 76). This situation exists with regard to TFC since it is generally only spoken by the elderly and there are no monolingual speakers (Ferreira and Holbrook 4, 17). Since the concept of the ‘ideal’ speaker cannot be applied to TFC, the criteria for participant selection were native speaker status and clarity of enunciation. The sample constituted three native speakers, two residents of Paramin and one of Talparo, between the ages of seventy and eighty. A revised Swadesh list (see Swadesh) of 145 words was used to ascertain the inventory of sounds of TFC. The list was amended as follows. Firstly, words which did not apply to a Caribbean context (e.g. snow) were removed. Additionally, using SLFC as a point of departure (based on Frank et al 2001) words which yielded homophones were eliminated and words which did not have a French Creole-origin equivalent in Frank et al, were also removed. Chelliah and de Reuse indicate that developing an inventory of sounds for the target language would require a list of about 500-700 words unless the inventory of a related language is known, in which case a shorter list would suffice (252). Given that the inventories of other LAFC varieties are known and both Goodman and Thomas have Lugo 16 attempted to describe the inventory of TFC, a more compact list was sufficient to achieve this aim. In addition to recording words in isolation, longer sequences were elicited since, “… articulation is affected by adjacent segments through assimilation” (Chelliah and de Reuse 252). These longer sequences took the form of stories: folklore, traditional stories such as “Konpè Lapin” (Br’er Rabbit), and original narrations from the participants. Data were collected from participants during one-on-one interviews using Reverse Translation Elicitation (RTE). This involves asking participants to translate lexical items or sentences into the target language (Chelliah and de Reuse 377). In the context of this study, participants were asked to translate lexical items from English into TFC. The following is a sample interaction between the interviewer and participant. Interviewer: Can you tell me the word for ‘pig’ in Patois? Participant: … Interviewer: Do you know any “Konpè Lapin” stories in Patois? Participant: … Based on the procedures proposed by Bowern (49), data were collected and analysed as follows. Phonetic data were elicited and recorded during one-on-one interviews with participants by means of a word list and stories, as described above. Narrow transcriptions were made during the interviews and the recordings were revisited in order to correct any transcription errors that may have occurred. Subsequently, all transcribed sounds were recorded on a phonemic chart, their environments noted and suspicious pairs identified. These were then tested to determine if they have phonemic status by identifying minimal or near-minimal pairs which contained them. If contrast was found in identical or analogous environments, these sounds were confirmed as phonemes. However, if sounds were found to be in identical environments but no contrast was noted, these sounds were determined to be in free variation. The phonemic inventory attested by the current study was then compared with the one proposed by Goodman (208) and similarities and differences between them were noted. The transcribed surface forms of the lexical items on the word list were then analysed to determine their syllable structure. Each lexical item was separated into syllables using CV- notation, and phonemes were examined to ascertain their distribution within the syllable. Additionally, all possible syllable combinations were noted in order to determine the syllable Lugo 17 template of TFC. Finally, data (Ferreira 2016) which illustrate the reanalysis of the syllable structure of lexical items of Spanish origin in TFC were analysed. This entailed comparing lexical items against their source words (Winer 2009) to determine the phonological process involved and the type of syllable structure which resulted from the reanalysis. Lugo 18 Chapter 4: Results and Discussion As discussed in Chapter 2, it is confirmed that there is little distinction between the phonemic inventories of TFC and other Antillean French Creole varieties which is in keeping with Klein’s and Corne’s findings regarding phonological uniformity among Creole languages. Additionally, there has been minimal change in the phonemic inventory of TFC proposed by Goodman (208), with one change being noted in the consonant phoneme inventory; the addition of the alveolar central approximant /ɹ/ (to be discussed further below). 4.1 Consonants The phonemic inventory of modern TFC comprises twenty-two phonemic consonants including three pairs of plosives, one pair of affricates, three nasals, seven fricatives and four approximants, three of which are central and one lateral, as shown in Table 2. Bilabial Labio- Alveolar Post- Palatal Velar Glottal dental Alveolar Plosive p b t d k ɡ Affricate ʧ ʤ Nasal m n ŋ Fricative f v s z ʃ ʒ h Central (w) ɹ j w Approximant Lateral l approximant Table 2: Consonant phoneme inventory of TFC The following near-minimal set exemplifies the phonemic contrasts among TFC consonants. The consonant phonemes are illustrated in Table 3 in word-initial position in monosyllabic words, except for /ɹ/ which is shown in a disyllabic word and /ŋ/, which cannot appear word-initially in TFC. Lugo 19 TFC Gloss /bɛ/ bè ‘butter’ /pɛ/ pè ‘pair’ /tɛ/ tè ‘dirt, ground’ /dat/ dat ‘date’ /ka/ ka ‘progressive marker’ /ɡad/ gad ‘guard’ /ʧɛ/ tjè ‘heart’ /ʤɛ/ djè ‘hardly’ /mɛ/ mè ‘but’ /nɛt/ nèt ‘clean’ /sãŋ/ sanng ‘belt’ /fɛ/ fè ‘to do’ /vɛ/ vè ‘glass’ /sɛl/ sèl ‘salt’ /zɛl/ zèl ‘wing’ /ʃɛ/ chè ‘expensive’ /ʒɛ/ jè ‘jar’ /had/ had ‘clothes’ /ɹadjo/ radyo ‘radio’ /jɛ/ yè ‘yesterday’ /wɛ/ wè ‘to see’ /lɛ/ lè ‘when’ Table 3: Phonemic contrast among TFC consonants. Lugo 20 4.2 The Alveolar Central Approximant As mentioned above, the only noted change to the consonant phoneme inventory of TFC is the addition of the alveolar central approximant, /ɹ/. However, it appears to be a phoneme which is considerably limited in its distribution. The HC phoneme, /ɣ/ will be used as a point comparison in discussing /ɹ/ in TFC6. Valdman indicates that /ɣ/ has limited distribution in HC, as it does not occur in a postvocalic position, except in the regional dialect of Northern Haiti, and it never appears word-finally. However, it is acknowledged that this may not be an autonomous development of HC since a weakened or deleted /​ ʁ/ (​ its corresponding French phoneme) in these positions is a well-attested phenomenon in French dialects (65). Additionally, it does not generally occur before rounded vowels7 and therefore it does not contrast with /w/ in this environment (Valdman 66). The distribution of /ɹ/ is similar in TFC in that it does not occur in a word-final position or after a vowel in the same syllable. However, it can occur before a rounded vowel and does contrast with /w/ in this position as evidenced by the following minimal pair8: /ɹoti/ ‘roti’ /woti/ ‘to roast.’ The distribution of /ɹ/ in TFC seems to be limited to onset position, as either the sole consonant in a simple onset or as the second element in a consonant cluster, which is preceded by an obstruent. Additionally, although it a​ ppears in toponyms such as /paɹamin/ ‘Paramin’ and /maɹaval/ ‘Maraval​, it seems to appear most frequently in recent lexical borrowings, such as, /fɹiʤ/ ‘refrigerator’ /kɹikɛt/ ‘cricket’ /tɹakta/ ‘tractor’ /ɹadjo/ ‘radio’ /ɡaɹaʒ/ ‘garage.’ It is possible that the addition of this phoneme is due to the influence of English and TrinEC. As Ferreira and Holbrook state, even elderly speakers of TFC are not monolingual and also speak English or TEC (4). Sankoff indicates that “... languages spoken by bilinguals influence each other in various ways” (638). One type of influence is phonological interference or transfer, especially in lexical borrowings (Sankoff 644). In this context, the speaker usually applies native phonological patterns to the borrowed lexical items. However, the patterns of the source language can sometimes be retained, and can act as a catalyst for 6 Brousseau indicates that in HC, /ɣ/ behaves as the language’s rhotic and may behave both as a fricative and sonorant (liquid or semivowel) (115). In light of this, together with the the parallels in its distribution described by Valdman (64-66), this phoneme will be used as a point of comparison in discussing the voiced alveolar central approximant, /ɹ/ in TFC. 7 Valdman indicates that some speakers of Northern Haitian Creole use /r/ before a rounded vowel, but it is difficult to differentiate between /ɣ/ and /w/ in this context (77). 8 It should be noted that this was the sole recorded instance of contrast between /ɹ/ and /w/ before a rounded vowel. Lugo 21 phonological change in the receiving language (Sankoff 647). Therefore, it is possible that the alveolar central approximant, ​/ɹ/ ​in TFC has been retained from English and TrinEC. 4.3 Free Variation The phonemes /ɹ/ and /w/ seem to be in free variation in lexical borrowings in TFC. Regarding free variation between /ɣ/ and /w/ in HC, Valdman indicates that in the regional dialect of Northern Haiti these phonemes are in free variation before unrounded vowels and after labial and labiodental consonants. In this regard, it is stated that speakers of this dialect freely substitute /w/ for /ɣ/ and vice versa in this position, which is illustrated by /bɣa/ ~ /bwa/ ‘arm’ and /bwa/ ~ /bɣa/ ‘wood,’ respectively (Valdman 66-67). However, the free variation between /ɹ/ and /w/ in TFC can only be attested in word-initial position before an unrounded vowel, in an English lexical borrowing /ɹadjo/ ~ /wadjo/ ‘radio.’ 4.4 Vowels The vowel phoneme inventory of TFC includes seven oral vowels (four front and three back), three nasal vowels (two front and and one back) and one diphthong, as shown by Table 4, below. Front Back Close i u Close-mid e o, õ Open-mid ɛ, ɛ ̃ ɔ Open a, ã, aɪ Table 4: Vowel phoneme inventory of TFC The following near-minimal set shows the phonemic contrasts among TFC vowels. In Table 5, vowel phonemes are illustrated in a word-final position, in monosyllabic words, after a plosive. Lugo 22 TFC Gloss /ti/ ti ‘variant of piti ‘small’’ /te/ té anterior verb marker /tɛ/ tè ‘dirt, ground’ /pɛ̃/ pen ‘bread’ TFC Gloss /tã/ tan ‘time’ /tu/ tou ‘also’ /do/ do ‘back’ /põ/ pon ‘bridge’ /pɔ/ pò ‘port’ /baɪ/ bay ‘to give’ Table 5: Phonemic contrast among TFC vowels It should be noted that Goodman’s inventory did not account for the diphthong, /aɪ/. However, this is likely because, rather than considering it a unitary phoneme, it was analysed as a biphonemic sequence which comprises the open front unrounded vowel, /a/ and the voiced palatal approximant, /j/ (/aj/). 4.5 The Syllable Table 6 illustrates the types of syllables which are attested in TFC. Syllable Type Example Gloss V /u/ ‘you’ CV /jo/9 ‘them, they, their’ CCV /flɛ/ ‘flower’ VC /iʃ/ ‘child’ 9 Although glides are ambiguous segments, and this syllable could also be analysed as GV, they are classified as consonants in the current study. This allowed them to fit a CV syllable template with unambiguous segments, as opposed to creating a GV one. Lugo 23 CVC /sis/ ‘six’ CCVC /gwɛ̃n/ ‘seed’ Table 6: Types of syllables attested in TFC The data in Table 6, illustrate that TFC permits both open and closed syllables, with simple and complex onsets and simple codas. This provides evidence that TFC does not feature exclusively simple and unmarked syllables; rather, its syllable template can be posited as (C)(C)V(C), which encapsulates all six syllable types shown in Table 6. This is in keeping with previous findings that Atlantic French Creoles are among the groups of Creoles which do not exhibit a tendency toward the unmarked CV structure (see Holm 144, McWhorter 156). Additionally, this further exemplifies the uniformity among Creoles languages and situates TFC around the typological middle with regard to syllable structure. 4.6 Nuclei The nucleus of the TFC syllable can comprise a monophthong or a diphthong,10 as exemplified in Table 7. The current study confirms Goodman’s assertion that all TFC vowels can occur in open syllables and all, with the exception of /e, o/, can also occur in closed ones. The forms in (a) - (d) below illustrate syllables which consist of monophthongs while (e) demonstrates a diphthong as the nucleus of the syllable. Additionally, (a) - (d) provides data which illustrate syllables comprising only a nucleus in word-initial and word-final positions. Gloss (a) /u/ ‘you’ (b) /e.pi/ ‘and, with’ (c) /ku.to.a/ ‘the knife’ (d) /sa.a/ ‘this, that’ 10 Although Carrington (42-43) and Valdman (64-65) note the presence of syllabic consonants in SLFC and HC, respectively, this study cannot attest them in TFC. However, this may be an accident of the data collected for this study and not proof that a consonant cannot function as the nucleus of a syllable in TFC. Lugo 24 (e) /baɪ/ ‘to give’ Table 7: Syllable nuclei in TFC 4.7 Vowel Hiatus The form in (c) comprises the noun k​ outo ‘knife’ and an allomorph [-a] of the definite determiner /-la/. In TFC, the postposed definite determiner appears as a CV or V allomorph depending on the final sound of the preceding noun. The V allomorph appears after a noun ending in a vowel whereas the CV allomorph appears after one which ends in a consonant. The prevailing view that Creoles exhibit a strong tendency toward CV sequences would predict the opposite distribution since the observed pattern creates sequences which deviate from this syllable structure. Additionally, it would be expected that the vowel hiatus created by the choice of allophone in (c) would be ungrammatical in TFC, but this is not the case. Therefore, TFC definite determiner allomorphy provides evidence in support of Holm’s and McWhorter’s claim that Atlantic French Creoles do not tend toward a CV structure. Similarly, (d) represents a variation of ​sala11 ​‘this, that’ which has an unmarked CVCV syllable structure. However, the omission of the alveolar lateral approximant, /l/ which occupies the onset position of the second syllable, once more, creates a vowel hiatus which deviates from this unmarked structure. That the resulting CV.V structure is in more frequent use in TFC further illustrates that the TFC syllable is not limited to CV sequences. 4.8 Onsets As Klein states, all Creole languages have CV syllables (176), that is, a syllable which comprises a simple onset of a single consonant which precedes the vocalic nucleus. With regard to TFC, all consonants, with the exception of the velar nasal /ŋ/, can appear in onset position in simple onsets which is illustrated in Table 3 above. TFC also allows complex onsets, not only in word-initial position but also word-medially. The formation of complex onsets is guided by the Sonority Sequencing Principle (SSP), which requires the sonority of the consonants in the onset to be rising toward 11 The form s​ ala was not recorded during data collection for the current study. However, the variation between sala and s​ a’a is noted in SLFC in Frank et al (195). Given the similarity between both varieties, the current study also analysed ​sa’a as a variant of ​sala​. Additionally, consultation with researcher N. Hodge, confirmed that ​sala,​ while present in TFC, has limited usage, usually only appearing in songs (personal communication). Lugo 25 the nucleus. There are two types of complex onsets in TFC; those which comprise an obstruent and a sonorant (Type 1) and those which consist of two sonorants (Type 2). With reference to Type 1, as per the SSP, the obstruent must precede the sonorant in the cluster. In TFC, plosives, affricates and fricatives (beside /h/)12 are followed by liquids and glides in the onset cluster. Regarding Type 2, the alveolar lateral approximant, /l/ and nasals (except /ŋ/) appear first in the cluster and are followed by glides, since liquids and nasals are less sonorous than glides, thereby upholding the SSP. However, with regard to nasals in initial position in sonorant-sonorant clusters, it was noted that /m, n/ are only followed by the labial-velar approximant, /w/. Table 8 provides a sample of some possible complex onsets in TFC. Gloss /ple.we/ ‘to laugh’ /fwɛ/ ‘brother’ /ʧɹak.ta/ ‘tractor’ /lwɛ̃/ ‘far’ /mwa/ ‘month’ Table 8: Complex onsets in TFC 4.9 Codas Simple codas, consisting of a single consonant are permitted in TFC, both word-medially and in a word-final position. All TFC consonants can occupy this position, except /ʧ, h, ɹ/ and glides13. Therefore, codas are generally obstruents since the only sonorants which appear in this position are nasals and the alveolar lateral approximant, /l/. 12 The voiced postalveolar affricate, /ʤ​ / and the voiceless postalveolar fricative, /ʃ/ were not attested in initial position in obstruent-sonorant clusters by this study, however, this may not be a reflection of the phonotactic constraints of TFC, but rather, an accident of the data, since both these phonemes are permitted in this position in SLFC. For example, /ze.ʤwi/ ‘needle’ (Frank et al 243), and /ʃwa.zi/ ‘to choose’ (Frank et al 41). 13 As previously indicated, Goodman notes that /​ ʤ​ / is not permitted in coda position in TFC (212). However, in the current study, there was one noted instance of this phoneme in coda position, in the lexical borrowing, /fɹiʤ/ ‘refrigerator.’ ​This may also be due to influence from English and TrinEC. Lugo 26 Moreover, TFC only permits simple codas since consonant clusters are not allowed in coda position. Table 9 illustrates a few possible codas in TFC. Gloss /mɛɡ/ ‘thin’ /tãn/ ‘to hear’ /ɡal.wi/ ‘gallery, porch’ /ɡa.ɹaʒ/ ‘garage’ Table 9: Simple codas in TFC 4.10 Resyllabification14 As further evidence that TFC does not favour a CV syllable structure, some attention will be given to the reanalysis of the syllable structure of lexical items of Spanish origin in TFC. Table 10 provides a list of resyllabified TFC lexical items of Spanish origin (Ferreira 2016) and their source words, along with a gloss (Winer 2009). Source Word TFC Gloss /al.pa.ɡa.ta/ /al.pa.ɡat/ ‘a type of sandal’ /a.re.pa/ /a.ɹep/ ‘a semi-circular fried patty made of cornmeal’ /ba.ʧa.ko/ /ba.ʧak/ ‘a large ant’ /bo.ni.to/ /bo.nit/ ‘an edible marine fish’ /bo.ʎo/ /boj/ ‘a steamed cornmeal dumpling’ /bu.ro.ki.to/ /bu.ɹo.kit/ ‘a traditional Carnival character’ /ka.ʧa.pa/ /ka.ʧap/ ‘a type of cornmeal dumpling baked in a pot’ /kaɪ.mi.to/ /kaɪ.mɛt/ ‘a type of tree’ /kas.ka.ra.du.ra/ /kas.ka.du/ ‘an armoured catfish’ 14 Although this study intended to examine the TFC syllable synchronically, this section was included in order to further illustrate that TFC does not tend only toward the CV syllable. Lugo 27 /dwen.de/ /dwɛn/ ‘a folklore character’ /ɡa.ʎe.ra/ /ɡa.jɛl/ ‘a ring for cock-fighting’ /la.ɲa.pa/ /lan.jap/ ‘an additional bit which a vendor gives free of cost’ /la.pa/ /lap/ ‘a large rodent’ /ma.mar.ga.ʎo/ /ma.ma.gaɪ/ ‘to try to get something by flattery’ /pa.ran.da/ /pa.ɹaŋ/ ‘a traditional Venezuelan-derived type of Christmas singing’ /pi.ko.pla.ta/ /pi.ko.plat/ ‘a type of songbird’ /pla.na.zo/ /pla.nas/ ‘a blow with the flat side of a cutlass blade’ /po.ko.a.po.ko/ /pok.a.pok/ ‘little by little’ /pon.ʧe.kre.ma/ /pon.ʧa.kɹɛm/ ‘an eggnog type drink’ Source Word TFC Gloss /san.ko.ʧo/ /saŋ.koʧ/ ‘a thick soup’ /sa.pa.to/ /sa.pat/ ‘a type of sandal’ Table 10: Resyllabified TFC lexical borrowings of Spanish origin. The data in Table 9 show that lexical items were adopted into TFC using the process of apocope. Phonological restructuring in Creole languages, via such a process, is often regarded as being driven by the desire to achieve a syllable structure which is closer to a CV structure than the corresponding European lexical item (Klein 173). As illustrated in Table 9, this is not the case in TFC. Holm’s discussion on deletion in Creole languages hinges on the omission of consonants, which has the effect of breaking up consonant clusters in the European source words (141). However, the target of deletion in TFC is not the consonant but the vocalic nucleus which instead creates a syllable with deviates from the CV structure. As Harris indicates, while a case can be made for viewing final consonant deletion as motivated by pressure to simplify syllable structure, deletion of a final vowel cannot be rationalised in the same way since it almost always increases syllable markedness (1,613). This is because vowel apocopation triggers resyllabification, as it forces the preceding consonant, originally an onset, into the coda position of the previous syllable (Harris 1,614). Lugo 28 In TFC, as Harris describes, the nucleus of a final syllable is deleted, and its onset is reanalysed as the coda of the preceding syllable producing a word-final closed syllable, as in (1) below. (1) Input /al.pa.ɡa.ta/ Vowel Apocopation al.pa.ɡa.t Resyllabification al.pa.ɡat Surface Form [al.pa.ɡat] The type of resyllabification shown in (1) is the most common among the data in Table 9, accounting for more than half of the TFC forms presented. It is also attested in cases such as /pok.a.pok/ (Sp p​ oco a poco ‘little by little’), in which the source item comprises multiple words. Closer analysis reveals that apocope targeted the final syllables of both disyllabic words (​poco​) in the source phrase. That is, their nuclei were deleted and their onsets reanalysed as the codas of the previous syllables, generating closed syllables. This type of apocope in TFC is summarised by the following rule which states that a vowel (nucleus) is deleted after a consonant (onset) when that vowel occurs in a word-final position. V --> Ø / C__# However, in some cases, as with /kas.ka.du/ (Sp ​cascara dura ‘​ hard shell’), deriving TFC surface forms required an additional process, as shown in (2). (2) Input /kas.ka.ra.du.ra/ Vowel Apocopation kas.ka.r.du.r Resyllabification kas.kar.dur Consonant Apocopation kas.ka.du Surface Form [kas.ka.du] Regarding (2), apocope targets the final nucleus of both source words, triggering resyllabification. However, since resyllabification produces codas which comprise a rhotic, Lugo 29 which is not permitted by TFC’s phonotactic constraints, t​ he rhotic is also deleted. Therefore, although the resulting structure is CV, this does not indicate a preference for the unmarked syllable, but rather reflects TFC’s phonotactic constraints. I​ n light of this, it is expected that this also occurs with any phoneme which the phonotactic constraints of TFC do not allow in coda position. A similar situation is that of /dwɛn/​ illustrated in (3) below. (3) Input /dwen.de/ Vowel Apocopation dwen.d Resyllabification dwend Consonant Apocopation dwen Surface Form [dwɛn] In this case, after apocope and resyllabification occurs, the form contains a consonant cluster in coda position. As discussed in a previous section, TFC does not permit coda consonant clusters. Since the structure which results from resyllabification violates this phonotactic constraint, the final consonant in the cluster is deleted to create a simple coda. This is summed up in the following rule which states that a consonant is deleted word-finally when it occurs after another consonant, that is, in a consonant cluster. C --> Ø / C__# The resyllabification of these lexical items provides additional evidence that TFC does not display a tendency only toward the unmarked CV syllable since the syllables targeted by apocope were already of this type. Although based on a limited amount of data, the pattern of phonological restructuring to create a CVC syllable, except when prevented by the phonotactic constraints of the language, implies that TFC tends toward a closed syllable instead. Lugo 30 Chapter 5: Conclusion This study has been concerned with identifying the current phonemic inventory of TFC and establishing whether there have been any significant changes to the inventory posited by Goodman in 1958. Additionally, it sought to determine the syllable template of TFC and provide a preliminary description of its phonotactic constraints. The current phonemic inventory was found to be identical to that described by Goodman (208) except for the addition of the central alveolar approximant, /ɹ/ as a phoneme. It was established that this phoneme has a limited distribution with parallels to the distribution of HC’s /ɣ/. The addition of this phoneme was considered to be the result of contact with English and TrinEC, since it is found most frequently in lexical items borrowed from these languages and given that all TFC speakers are also speakers of English or TrinEC or both. Furthermore, the possibility of /ɹ/ and /w/ occurring in free variation in borrowed lexical items was also raised but the data collected for this study were insufficient to conduct Lugo 31 a thorough analysis of this phenomenon. This, therefore, is a matter which may be addressed in future work. With regard to the syllable structure of TFC, its syllable template was determined to be (C)(C)V(C) which indicates that TFC permits consonant clusters in an onset position but only allows simple codas and that the vocalic nucleus can stand on its own as a syllable. The nucleus of the TFC syllable can comprise a single vowel or a diphthong. Moreover, all vowels can occur in open syllables and all, with the exception of /e, o/, can also occur in closed ones. Vowel hiatus was also attested in TFC. It was determined that all consonants can occupy onset position in simple onsets, with the exception of the velar nasal, /ŋ/ which occurs only in coda position. Additionally, it was deduced that consonant clusters in onset position are governed by the SSP. As a result, only specific combinations are allowed; obstruent-sonorant combinations are permitted where plosives, affricates and fricatives precede liquids or glides, and sonorant-sonorant clusters whose elements constitute a rise in sonority toward the nucleus are also allowed. In these combinations, liquids and nasals must be followed by a glide. With regard to the coda, all TFC consonants can appear in coda position except glides and /ʧ, h, ɹ/. In light of these findings, it was determined that the TFC syllable is relatively varied and does not display any significant tendency toward the CV syllable structure. As further evidence in support of this claim, the resyllabification of lexical borrowings of Spanish origin in TFC was analysed. This led to the conclusion that TFC exhibited a pattern of deleting the vocalic nucleus in an open syllable and reanalysing the onset of that syllable as the coda of the previous syllable, except when the phonotactic constraints of the language prevented it, for example, when the resyllabification produced a coda such as a rhotic or a consonant cluster. This was posited as evidence that TFC generally prefers a closed syllable. However, it must be acknowledged that due to time constraints, it was only possible to analyse phonological restructuring through the process of apocope. However, an analysis of how other types of phonological restructuring processes affect the syllable may provide further insight into TFC’s preferred syllable structure, and therefore, is a possible area of investigation for future work. These findings are significant since they provide evidence of a change in the phonemic inventory of TFC, by way of the addition of the alveolar central approximant, /ɹ/, and given TFC’s status as an endangered language, it is vital that such change is recorded. Lugo 32 Additionally, they illustrate that TFC has a relatively diverse syllable structure which lends support to the claim that Antillean French Creoles do not display a toward a CV syllable structure. Works Cited Bowern, Claire. ​Linguistic Fieldwork: A Practical Guide.​ 2​nd​ ed., Palgrave Macmillan, 2015. Brousseau, Anne-Marie. “One Substrate, Two Creoles: The Development of Segmental Inventories in St. Lucian and Haitian.” ​Creoles, their Substrates, and Language Typology​, edited by Claire Lefebvre, John Benjamins, 2011, pp. 105-125. Carrington, Lawrence D. S​ t. Lucian Creole: A Descriptive Analysis of its Phonology and Morpho-syntax.​ PhD Thesis, University of the West Indies, Mona, 1967. Carrington, Lawrence D. S​ t. Lucian Creole: A Descriptive Analysis of its Phonology and Morpho-syntax.​ Helmut Buske Verlag, 1984. Lugo 33 Chelliah, Shobhana L, and Willem J. de Reuse. H​ andbook of Descriptive Linguistic Fieldwork​. Springer, 2011. Colot, Serge, and Ralph Ludwig. “Martinican Creole Structure Dataset.” A​ tlas of Pidgin and Creole Language Structures Online​, edited by Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber, Max Planck Institute for Evolutionary Anthropology, 2013. Cook, Vivian. ​Second Language Learning and Language Teaching​. 4t​ h ed., Hodder Education, 2008. Corne, Chris. F​ rom French to Creole: The Development of New Vernaculars in the French Colonial World​. University of Westminster Press, 1999. Creswell, John W. ​Educational Research: Planning, Conducting, and Evaluating Quantitative and Qualitative Research. ​4t​ h ​ ed., Pearson, 2012. Crystal, David. ​Language Death​. Cambridge UP, 2014. Fattier, Dominique. “Haitian Creole Structure Dataset.” A​ tlas of Pidgin and Creole Language Structures Online​, edited by Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, and Magnus Huber, Max Planck Institute for Evolutionary Anthropology, 2013. Ferreira, Jo-Anne. “Last Spheres of a Dying Language.” Paper presented at the International Caribbean Conference, UWI, St. Augustine, 13-15 October 2016. Ferreira, Jo-Anne, and David Holbrook. “Are They Dying? The Case of Some French-lexifier Creoles.” Islands in Between Conference, Sir Arthur Lewis Community College, Castries, St. Lucia, 8-10 November 2001. Conference Presentation. Ferreira, Jo-Anne, Kathy-Ann Drayton, and Nnamdi Hodge. “Trinidadian French Creole: A Phonetic Illustration.” (In progress). Frank, David, et al. ​Kwéyòl Dictionary​. MOE GOSL, and SIL International, 2001. Goodman, Morris F. “On the Phonemics of the French Creole of Trinidad.” W​ ord,​ vol. 14, no. 2-3, 1958, pp. 208-212. Harris, John. “Deletion” ​Phonological Processes​, edited by Marc van Oostendorp, Colin J. Ewen, Elizabeth V. Hume, and Keren Rice, Wiley-Blackwell, 2011, pp. 1,597-1,621. Holm, John. ​An Introduction to Pidgins and Creoles​. Cambridge UP, 2000. Lugo 34 Klein, Thomas. “Creole Phonology Typology: Phoneme Inventory Size, Vowel Quality Distinctions and Stop Consonant Series.” T​ he Structure of Creole Words,​ edited by Parth Bhatt and Ingo Plag, De Gruyter, 2006, pp. 3-22. Klein, Thomas. “Typology of Creole Phonology: Phoneme Inventories and Syllable Templates.” ​Journal of Pidgin and Creole Languages​, vol. 26, no. 1, 2011, pp. 155-193. Levelt, Clara C., and Ruben Van der Vijver. “Syllable Types in Cross-Linguistic and Developmental Grammars.” ​Constraints in Phonological Acquisition,​ edited by René Kager, Joe Pater, and Wim Zonneveld, Cambridge UP, 2004, pp. 204–218. Maddieson, Ian. ​Patterns of Sounds​. Cambridge UP, 1984. McWhorter, John. ​The Missing Spanish Creoles​. University of Berkeley Press, 2000. Sankoff, Gillian. “Linguistic Outcomes of Language Contact.” ​The Handbook of Language Variation and Change​, e​ dited by J. K. Chambers, Peter Trudgill, and Natalie Schilling-Estes, Wiley-Blackwell, 2006, pp. 638-668. Swadesh, Morris. “Toward Greater Accuracy in Lexicostatistic Dating.” I​ nternational Journal of American Linguistics​, vol. 21, no. 2, 1955, pp. 121-137. Thomas, J. J. T​ he Theory and Practice of Creole Grammar​. 1869. New Beacon Books, 1969. Valdman, Albert. ​Haitian Creole: Structure, Variation, Status, Origin​. Equinox, 2015. Winer, Lise, editor. D​ ictionary of the English/Creole of Trinidad & Tobago. McGill-Queen’s UP, 2009. Appendix A Data Collection Instrument Lugo 35 A.​ ​Biodata: 1. Name 2. Age 3. Sex 4. Languages 5. Place of birth/Residence B. Word List 1. I/Me: Mwen [mwɛ̃] 35. Yesterday: yè [jɛ] 2. You: ou [u] 36. One: yonn [jõn] 3. He/She/It: li [li] 37. Two: dé [de] 4. We: nou [nu] 38. Three: twa [twa] 5. They: yo [jo] 39. Four: kat [kat] 6. And: épi [epi] 40. Five: senk [sɛk̃ ] 7. Because: pas [pas] 41. Six: sis [sis] 8. Far: lwen [lwɛ]̃ 42. Seven: sèt [sɛt] 9. Near: pwé [pwe] 43. Eight: wit [wit] 10. Right: dwèt [dwɛt] 44. Nine: nèf [nɛf] 11. Here: isi [isi] 45. Ten: dis [dis] 12. To Come: vini [vini] 46. All: tout [tut] 13. To Sit: asiz [asiz] 47. Many/Plenty: an pil [ãpil] 14. To Give: bay [baɪ] 48. Thin: mèg [mɛɡ] 15. To Stand: doubout [dubut] 49. Big: gwo [ɡwo] 16. To Fall: tonbé [tõbe] 50. Short: kout [kut] 17. To Walk: maché [maʃe] 51. Tall: ho [ho] 18. To Throw: voyé [voje] 52. Fat: gwa [ɡwa] 19. To Pull: halé [hale] 53. Salt: sèl [sɛl] 20. To Lie (Down): kouché [kuʃe] 54. Sea: lanmè [lãmɛ] 21. To Push: pousé [puse] 55. Stone: w​ òch [w​ɔʃ] 22. To Take: pwan [pwã] 56. Ashes: sann [sãn] 23. To Wash: lavé [lave] 57. Fire: difé [dife] 24. To Tie: mawé [mawe] 58. Tree: pyébwa [pyebwa] 25. To Hit: fouté [foute] 59. Flower: flè [flɛ] 26. To Cut: koupé [kupe] 60. Forest: gwo bwa [ɡwobwa] 27. To Rub: fwoté [fwote] 61. Seed: gwenn [ɡwɛñ ] 28. To Dig: fouyé [fuje] 62. Peas: pwa [pwa] 29. Day: jou [ʒu] 63. Snake: sèpan [sɛpã] 30. Night: lannwit [lãnwit] 64. Pig: kochon [koʃõ] 31. Year: lanné [lãne] 65. Lizard: zanndoli [zãndoli] 32. Today: jòdi [ʒɔdi] 66. Blood: san [sã] 33. Tomorrow: denmen [dɛ̃mɛ̃] 67. Hand: lanmen [lãmɛ]̃ 34. Month: mwa [mwa] 68. Tongue: lanng [lãŋ] Lugo 36 69. Tooth: dan [dã] 108. To Play: jwé [ʒwe] 70. Foot: pyé [pje] 109. To Dance: dansé [dãse] 71. Egg: zé [ze] 110. Knife: kouto [kuto] 72. Back: do [do] 111. Name: non [nõ] 73. Eye: zyé [zje] 112. Street: lawi [lawi] 74. Head: tèt [tɛt] 113. To Break: kasé [kase] 75. Nose: né [ne] 114. Word: mo [mo] 76. Heart: tjè [tʃɛ] 115. To Kill: tjwé [tʃwe] 77. Belly: bouden [budɛ]̃ 116. Truck: tròk [tɹɔk] 78. Neck: kou [ku] 117. Rice: diwi [diwi] 79. Hair: chivé [ʃive] 118. To Let Go: lagé [laɡe] 80. Leg: janm [jãm] 119. Belt: sanng [sãŋ] 81. Wing: zèl [zɛl] 120. To Resemble: sanm [sãm] 82. Tail: latjé [latʃe] 121. Bedsheet: dwa [dwa] 83. Arm: bwa [bwa] 122. Cross: kwa [kwa] 84. To Drink: bwè [bwɛ] 123. But: mé [me] 85. To Die: mò [mɔ] 124. Jar: jè [ʒɛ] 86. To Hear: tann [tãn] 125. Time/Occasion: fwa [fwa] 87. To See: wè [wɛ] 126. When: lè [lɛ] 88. To Eat: manjé [mãʒe] 127. Radio [ɹadjo] 89. To Laugh: wi [wi] 128. Pair: pè [pɛ] 90. To Sing: chanté [ʃãte] 129. Butter: bè [bɛ] 91. To Cry: plewe [plewe] 130. Clean: nèt [nɛt] 92. To Speak: palé [pale] 131. Hardly any: djè [dʒɛ] 93. To Kiss: bo [bo] 132. Worm/Glass: vè [vɛ] 94. Black: nwè [nwɛ] 133. Late: ta [ta] 95. Green: vè [vɛ] 134. Bread: pen [pɛ]̃ 96. Blue: blé [ble] 135. Bridge: pon [põ] 97. White: blan [blã] 136. Roti: roti [ɹoti] 98. Dry: sèk [sɛk] 137. Garage: [ɡaɹaʒ] 99. Good: bon [bõ] 138. Clothes: had [had] 100. Hot: cho [ʃo] 139. Guard: gad [ɡad] 101. Cold: fwèt [fwɛt] 140. Gallery/Porch: [ɡalwi] 102. Bad: mové [move] 141. Cricket [kɹikɛt] 103. Expensive: chè [ʃɛ] 142. Tractor [tɹakta] 104. Brother: fwè [fwɛ] 143. Refrigerator [fɹidʒ] 105. Husband: mawi [mawi] 144. Port [pɔ] 106. To Hunt: chasé [ʃase] 145. Date [dat] 107. To Cook: tjwit [tʃwit] Lugo 37 C. Elicitation of Longer Sequences 1. Can you share any ​Konpè Lapen ​or other old time stories in Patois? 2. Original Narrations: e.g. Can you tell me a little about growing up in Paramin in Patois? Lugo 38 Appendix B Sample Data A.​ ​Biodata: 1. Name: K. Romain 2. Age: 75 3. Sex: Male 4. Languages: L1: TFC; L2: English 5. Place of birth/Residence: Paramin B. Word List 1. I/Me: [mwɛ]̃ 30. Night: [lãnwit] 2. You: [u] 31. Year: [lãne] 3. He/She/It: [li] 32. Today: [ʒɔdi] 4. We: [nu] 33. Tomorrow: [dɛm̃ ɛ]̃ 5. They: [jo] 34. Month: [mwa] 6. And: [epi] 35. Yesterday: [jɛ] 7. Because: [pas] 36. One: [jõn] 8. Far: [lwɛ]̃ 37. Two: [de] 9. Near: [pwe] 38. Three: [twa] 10. Right: [dwɛt] 39. Four: [kat] 11. Here: [isi] 40. Five: [sɛk̃ ] 12. To Come: [vini] 41. Six: [sis] 13. To Sit: [asiz] 42. Seven: [sɛt] 14. To Give: [baɪ] 43. Eight: [wit] 15. To Stand: [dubut] 44. Nine: [nɛf] 16. To Fall: [tõbe] 45. Ten: [dis] 17. To Walk: [maʃe] 46. All: [tut] 18. To Throw: [voje] 47. Many/Plenty: [ãpil] 19. To Pull: [hale] 48. Thin: [mɛɡ] 20. To Lie (Down): [kuʃe] 49. Big: [ɡwo] 21. To Push: [puse] 50. Short: [kut] 22. To Take: [pwã] 51. Tall: [ho] 23. To Wash: [lave] 52. Fat: [ɡwa] 24. To Tie: [mawe] 53. Salt: [sɛl] 25. To Hit: [fute] 54. Sea: [lãmɛ] 26. To Cut: [kupe] 55. Stone: [wɒʃ] 27. To Rub: [fwote] 56. Ashes: [sãn] 28. To Dig: [fuje] 57. Fire: [dife] 29. Day: [ʒu] 58. Tree: [pye bwa] Lugo 39 59. Flower: [flɛ] 101. Cold: [fwɛt] 60. Forest: [ɡwo bwa] 102. Bad: [move] 61. Seed: [ɡwɛ̃n] 103. Expensive: [ʃɛ] 62. Peas: [pwa] 104. Brother: [fwɛ] 63. Snake: [sɛpã] 105. Husband: [mawi] 64. Pig: [koʃõ] 106. To Hunt: [laʃas] 65. Lizard: [zãndoli] 107. To Cook: [tʃwit] 66. Blood: [sã] 108. To Play: [ʒwe] 67. Hand: [lãmɛ]̃ 109. To Dance: [dãse] 68. Tongue: [lãŋ] 110. Knife: [kuto] 69. Tooth: [dã] 111. Name: [nõ] 70. Foot: [pje] 112. Street: [lawi] 71. Egg: [ze] 113. To Break: [kase] 72. Back: [do] 114. Word: [pawɔl] 73. Eye: [zje] 115. To Kill: [tʃwe] 74. Head: [tɛt] 116. Truck: [loto] 75. Nose: [nɛ] 117. Rice: [diwi] 76. Heart: [tʃɛ] 118. To Let Go: [lage] 77. Belly: [budɛ]̃ 119. Belt: [sãŋ] 78. Neck: [ku] 120. To Resemble: [sãm] 79. Hair: [ʃive] 121. Bedsheet: [kuɛvti] 80. Leg: [jãm] 122. Cross: [kwa] 81. Wing: [zɛl] 123. But: [mɛ] 82. Tail: [latʃe] 124. Jar: [ʒɛ] 83. Arm: [bwa] 125. Time/Occasion: [fwa] 84. To Drink: [bwɛ] 126. When: [lɛ] 85. To Die: [mɔ] 127. Radio: [ɹadyo] 86. To Hear: [tãn] 128. Pair: pè [pɛ] 87. To See: [wɛ] 129. Butter: bè [bɛ] 88. To Eat: [mãʒe] 130. Clean: nèt [nɛt] 89. To Laugh: wi [wi] 131. Hardly any: djè [dʒɛ] 90. To Sing: [ʃãte] 132. Glass: vè [vɛ] 91. To Cry: [plewe] 133. Late: ta [ta] 92. To Speak: [pale] 134. Bread: pen [pɛ]̃ 93. To Kiss: [bo] 135. Bridge: pon [põ] 94. Black: [nwɛ] 136. Roti: roti [ɹoti] 95. Green: [vɛt] 137. Garage: [ɡaɹaʒ] 96. Blue: [ble] 138. Clothes: had [had] 97. White: [blã] 139. Guard: gad [ɡad] 98. Dry: [sɛk] [ʃɛs] 140. Gallery/Porch: [ɡalwi] 99. Good: [bõ] 141. Cricket [kɹikɛt] 100. Hot: [ʃo] 142. Tractor [tɹakta] Lugo 40 143. Refrigerator [fɹidʒ] 144. Port [pɔ] 145. Date [dat]