Thesis: A grammatical description of the Noun Phrase in the English-lexicon Creole 

of St Vincent and the Grenadines

Defence date: 07 December 2004

Result: High pass with distinction

• Professor Claude DELMAS, Université Paris III, Sorbonne Nouvelle
• Professor Emeritus Mervyn C. ALLEYNE, University of the West Indies, St-Augustin, now at the Universidad de Puerto Rico, Río Piedras
• Professor Salikoko MUFWENE, University of Chicago, Frank J. McLoraine Distinguished Service Professor
• Laurence GOURY, PhD, Research Officer, Institut de Recherche pour le Développement
• Professor Daniel VERONIQUE –Université Paris III, Sorbonne Nouvelle, Thesis Advisor 

Abstract: This study proposes a grammatical description of the noun phrase (NP) of the English-lexicon Creole of St. Vincent and the Grenadines (VinC).  VinC, which coexists with its superstrate, is at present linguistically unaccounted for. As a result, prior to the syntactic analysis, we reflect on the process of creolisation that could account for the origin of VinC, based on historical documents. We then provide a phonological analysis, which gives a concise idea of the phonemic and accentual features of VinC. The alphabet recommended, is intended to render the transcription of the examples provided in the study conform to the phonological nature of the creole. The grammatical description itself, covered in three chapters, hinges on the functionalist and structuralist frameworks, as it seeks to describe the syntactic relations existing between the constituents of the NP. Firstly, we study the features of the head of the NP, i.e. the noun and its substitutes.  We then analyse the left expansions of nouns, i.e. determiners as well as modifiers. The penultimate chapter offers a unified analysis of the post-modifiers of the VinC NP, i.e. relative clauses, complementizer clauses, prepositional clauses and noun complements. The final chapter focuses on the concept of reference and how it influences the choice of definite and indefinite determiners.  Here, we advocate an analysis based on the theory of familiarity and degree of identification from the hearer’s point of view that could account for definiteness and specificity: grammatical-semantic notions that often fall short in their analysis of the null determiner.

UNIVERSITY OF PARIS III—SORBONNE NOUVELLE

GRADUATE SECTION 268: LANGAGES ET LANGUES

INSTITUTE OF GENERAL AND APPLIED LINGUISTICS AND PHONETICS

Summary of Thesis

Defended publicly

Tuesday 7 December 2004

In fulfilment of the requirements for the Degree of

DOCTOR OF PHILOSOPHY IN LINGUISTICS

UNIVERSITY OF PARIS III

by

Paula PRESCOD

A Grammatical Description of the

Noun Phrase in the English-Lexified Creole of

St. Vincent and the Grenadines

xi, 324 p, 32 p annex

Thesis Advisor: Daniel Véronique

Choice of topic: This thesis sets out to provide a grammatical description of the English-lexicon creole of St. Vincent and the Grenadines (VinC). As a teacher of French foreign language in a secondary school which enrolled 12+year old students with the best passes in the Common Entrance Examination, it became apparent that the status of English as the official language was purely symbolic since the de facto the mother tongue of many Vincentians is the creole. As a result, the creole created interferences not only in the use of standard English but also in learning French. The choice of topic for this study is a corollary of my teaching experience.

Theoretical framework: In this study, I opt for a functional and structuralist perspective. First of all, I describe the grammatical and morphological properties of elements of the noun phrase then classify them according to their functional roles. I then analyse the relations between noun phrase components and the way in which expansions i.e. determiners and modifiers, are organised around the noun and its substitutes. I adopt Martinet’s (1985: 83)[1] definition of the syntagm: l’ensemble d’unités significatives plus étroitement reliées entre elles qu’avec le reste de l’énoncé, plus, éventuellement, l’élément qui le relie à cet énoncé. To this, I associate Leech & Svartvik’s (1978: 251)[2] assumption that the noun phrase (NP) is so called because the word which is head,(its main part) is typically a noun. The noun can act as subject, object, or complement of a clause or as prepositional complement. This study also takes into consideration the role played by pronominal components which are referred to here as noun substitutes since they act like nouns in different syntactic positions.

Procedure: In the light of the fact that VinC is grammatically unaccounted for, it was necessary to select and critically examine the literature which has been a reference for VinC in the Vincentian community over the decades. These writings in VinC include the folkloric literature of Esther Edwards[3] and The New Artists Movement.[4] Oral recordings, via radio programmes and face-to-face contact, which present noteworthy features of VinC, were added to the corpus. These recordings made it possible to observe the whole array of registers used by speakers of VinC. Interestingly, the radio programmes offered authentic examples of natural VinC expression. Since the introduction of interactive radio in St. Vincent and the Grenadines (SVG), the moderators have been able to put callers at their ease, encouraging the participants to express themselves in the language they feel comfortable speaking. This relative freedom of expression has encouraged the use of diverse registers in VinC and callers are in no way coerced to use standard English, a language they do not necessarily “master”. This explains why the referenced utterances in this study exemplify a mixture of all the registers of English and VinC.

The examples thus collected were then transcribed. This is where a lack of uniformity in the transcriptions became evident, due no doubt to the influence exerted on VinC by English orthography. After comparing the transcriptions with work published in Vincentian parlance, it became clear that the issue of phonological representation needed to be resolved. Objectively speaking, VinC writers tend to use a multitude of orthographies for the same sound. This lack of uniformity can also be attributed to the inexistence of a system of homogenised orthography, thus the importance of proposing a uniform writing system made possible from acoustic measurements and analyses done using the automatic speech analyser PRAAT.[5] This functional orthography was applied to all unpublished transcriptions used in the corpus. Examples from published sources were cited unchanged.

Theoretical problems: A major problem faced was the question of how to process the orthography, in the light of the linguistic diversity of the Vincentian community. None of the utterances retained in the corpus were solicited. The majority of utterances were obtained in authentic face-to-face exchanges on a variety of topics, or upon request to relate a story or an experience. A number of utterances were also recorded during interactive radio and television programmes. My knowledge of VinC, enabled me to formulate a small number of utterances to complete the corpus. As the majority of speakers display an ability to express themselves in a variety of registers, the choice of utterances was another major concern. Was it enough to simply give a description of basilectal VinC, or was it legitimate to retain acrolectal utterances, which unequivocally indicate that speakers of VinC do not resist the structural and lexical influences of the official language and that VinC can be analysed in terms of a variation of “lects”?

A second concern related to the theoretical framework adopted for this study. Attempting at all costs to mold the grammar of VinC into a hard and fast theoretical framework would no doubt have provided a rigid analysis. From the moment a theoretical framework in which one sets out to make a grammatical description is adopted, one runs the risk of confining oneself within the limits of the framework. In addition to using the functionalist and structuralist models as a starting point for the grammatical and morphological descriptions of the components of the NP, this study also takes advantage of the generative approach to language analysis. For instance, in an attempt to provide an in-depth study of NPs modified by relative clauses, terms such as movement, syntactic gap and extraposition were applied insofar as they provided a coherent description.

Lastly, this descriptive work represented a major challenge. The linguistic literature, which referred to the dialectal varieties spoken in SVG, tends to do so in a rather anecdotal fashion. The absence of descriptive literature on VinC meant therefore that everything was yet to be accomplished as far as grammar is concerned. Succumbing to the temptation of basing a pilot grammar of VinC on the grammar of largely documented English-lexicon creoles, namely those of Jamaica, Guyana, Trinidad would imply that all English-creole speakers of the Caribbean spoke the same dialectal variety, without in the least seeking to lay down rules of usage and pointing out idiosyncrasies which exist. During this study, it turned out that VinC was a linguistic territory to discover and this ensured a certain measure of freedom with respect to the terms of analysis to be applied, the results to be anticipated and the conclusions to be drawn. Indubitably, providing a linguistic description of the speech of Vincentians means accepting the responsibility of making this description conform to usage insofar as it is based on empirical data.

Summary of the thesis: The thesis is divided into seven chapters, the first of which provides historical and demo-linguistic elements aimed at accounting for the presence of an English-lexicon creole on SVG. I first take a look back at the seventeenth century, a flashback which is necessary if we are to understand the demo-linguistic complexity of the Vincentian community. Even today, historical and tourism literature still proclaim erroneously, that a French patois is spoken on SVG. The content of this first chapter serve to clarify this popular belief by tracing the settlement history of the Arawaks, the Caribs and the Garifuna, the race resulting from the intermarriage between the Caribs and Africans who took refuge on SVG before the rise of plantation slavery. This mixed population, which held the Europeans at bay, refusing assimilation and domination for many decades, were able to thwart the introduction plantation slavery until the end of the eighteenth century.

In this first chapter, I show that the creolisation models proposed by Bickerton (1981)[6] and Chaudenson (1992)[7] only partially explain the presence of a creole in SVG. The use of a language which shares but few morphosyntactic and phonological features with modern English, is, as I see it, the result of a conscious effort made by the Africans and their descendants to remain linguistically set apart from the European. In spite of the lack of authentic testimonies from the inhabitants during the era of European occupation which will no doubt help to formulate solid facts and not mere extrapolations about the structural characteristics of the language of the non Europeans at the time when linguistic contact took root, I do not resolve to view the genesis of a creole as the servile population’s inability to learn the superstrate, nor as the imperfect appropriation of a second language. With regard to modern day Vincentian speech psycholinguistic and social considerations must be examined.

From a psycholinguistic point of view, the initial “differences” of the creole vis-à-vis its lexifier can be explained by the fact that any individual who finds him/herself in a situation of unguided language learning, employs a number of processes to acquire this language. Among these processes, we can mention borrowing, transfer and calquing, which, in psycholinguistic terms, offers the individual who finds him/herself in a situation of language contact, an immediate reward in filling the gaps in his/her knowledge of the target language. These gaps, if one admits that the term is cogent given the socio-historical context of colonisation, are the sign that learning is in progress, learning which readjusts itself by trial and error, ellipse of semantic redundancies or by meaning transfer: a situation where semantics takes precedence over syntax.

From the social perspective, it is important to stress that the société d’habitation, a key concept in Chaudenson’s[8] theory, did not last much more than seventy (70) years in SVG. Despite the cession of the territory to the British Crown in 1763, no official British occupation was set up prior to 1783, the year in which lands disputed between the British and the French were restored to the British. Massive arrivals of Africans date back to that period. Surprisingly, this situation did not encourage the complete assimilation of the African population remaining on the territory after the deportation of the Garifunas to Roatan, off the coast of Honduras, Central America. The emancipation of slavery was proclaimed in 1834. At this point, one of two things happened. On the one hand, non Europeans' lack of motivation to identify themselves linguistically with the Europeans was a refusal of European acculturation and a willingness to set themselves apart: indeed quite representative of New World image. On the other hand, the desire to obscure their way of speaking may have been a conscious effort made by the Africans during their secret meetings: a deliberate choice to create a language which the slave masters could not easily fathom and which could be termed as distinctly their own. These elements point to the fact that the desire to learn the European language, which may have been ubiquitous during the initial stages of colonization, changed as non-Europeans strengthened their new linguistic identity during the post-abolition era.

The second chapter provides an analysis of the sounds of VinC. This phonological description was essential for several reasons. Most importantly, although some comparative work has in the past referred to certain phonological characteristics of VinC, albeit from a diachronic perspective (c.f. Alleyne (1980)[9] and Le Page (1972))[10], this creole has not been subjected to a full analysis which could account for the phonological value of the sounds used. The goal of this second chapter is threefold. It first offers an articulatory and acoustic description of the sounds used by speakers of VinC, before outlining the accentual patterns shaped by the morphology of the noun. However, the focal point of this chapter is the orthography proposed, based on the results obtained. This alphabet offers an empirical writing system which could be accessed by anyone who needs to use VinC in its written form.

What comes out of this phonological analysis is that the sound system of VinC comprises more high vowels than it does low vowels, high vowels being longer during articulation. Consequently, there is a widespread use of long vowels /ii/, /aa/, /uu/, so that vowel length is a distinctive feature in VinC. Vowel tenseness also plays a role in the double phoneme feature as vowels which tend to close progressively during articulation, also tend to be articulated as rising diphthongs. This feature is in sharp contrast to Jamaican creole, which has falling diphthongs. There are no centralizing diphthongs in VinC since the schwa /«/ does not have phonemic status.

The alphabet proposed reveals six (6) pure vowels and four (4) long vowels. Three (3) compound vowels or diphthongs complete the list of vowels. There are two (2) nasal sounds whose distribution is highly restricted: [e)] appears only as the echo question (isn’t that so?); [a)] appears only in the sequence /aan/. In general terms, nasalization is observed only on long vowel sounds and is therefore not a distinctive feature in morphemes that are in the same paradigm. These thirteen (13) vowel sounds combine with two (2) glides and twenty (20) consonants to complete a table of sounds with distinctive phonemic values in VinC.

The three chapters which follow (3,4,5) make up the description per se of the VinC NP. Chapter 3 bears out that the central element of the NP is the noun. This stance is taken on the theoretical grounds that predeterminers or qualifiers (definite and indefinite articles, quantifiers and adjectives) and post-determiners and postposed modifiers (deixis, relative clauses and possessive phrases) are not requisite for the syntactic composition of the NP. They are in actual fact expansions, which operate primarily to provide grammatical or semantic information about the noun or its substitute. This amounts to postulating that the noun which has no overt determiner gives sufficient information about itself, thus enabling us to identify its reference. The referential perspective is developed in Chapter 6.

The third chapter proposes a grammatical and semantic analysis of the NP geared towards establishing the morphosyntactic differences between the simple, complex and compound nouns. This distinction allows us to analyse the semantic relations between the components of compound nouns. In this chapter, the grammatical characteristics of nouns are also examined. The underlying nominal is unmarked for gender. Conversely, inherent semantic features marking gender become apparent when nouns combine with other morphemes to form compound nouns. The processes identified in the formation of compound nouns are juxtaposition and suffixation. This chapter also proposes a development on the prosody of the NP, which serves to highlight the effect of suffixation on the accentual patterns of nouns.

Pronominal morphemes (personal, possessive, demonstrative, reflexive, interrogative, relative and indefinite) also function as NPs. They can replace nouns in all syntactic positions. The term pro-form was preferred to the traditional one of pronouns in this study since it is generally agreed upon that the term pronoun is tantamount to noun substitute. This study shows that pronominal morphemes can replace not only substantives but also clauses. The term pro-form appeared therefore to be more appropriate, given the syntactic autonomy of the pronominal paradigm in VinC.

Chapter 4 is concerned with NPs which exemplify the qualifier+qualified and determiner+determined syntactic patterns. In practical terms, preposed qualifier expansions are of the adjective+noun type. In the case of the components of compound nouns, also analysed as qualifiers, the previous chapter had already established that the lexical relations highlight the same syntactic distribution i.e. qualifier+qualified. The grammatical elements which act as predeterminers of nouns in a determiner+determined type syntactic relation are the definite article, the indefinite articles and demonstratives.

With regards to noun morphology, the distinction between singular and plural is ineffectual in the majority of contexts. In general terms, nouns are unmarked for number. However, a nominal lexeme which bears /s/ or /z/ (plural marker in the lexifier) does not necessarily require a plural reading. For instance, in VinC /ants/ may refer to one or more ‘ants’. There is no morphological defect if this kind of noun is preposed by the singulative article wan since wan ants is attested in basilectal VinC, unlike wan ant.

A significant number of lexemes do not have lexical entries without final /s/. This mainly concerns nouns referring to entities which are not directly countable albeit intrinsically massive or non-singulative. It is essential to stress that these items cannot combine with the singulative determiner wan. Interestingly, nominal items which are not directly countable can be preposed by wan, thus the need to analyse this morpheme as having a double morphosyntactic role: as an indefinite article and as a cardinal numeral. In addition, a third role, this time semantic, can be attributed to wan. Indeed, abstract items can be preposed by wan, in which case it operates as an intensifier. Som marks a semantic plural, not a grammatical plural since the noun it qualifies is not overtly plural. The concept of number marking is thus an unpredictable one given the morphosyntactic arguments developed in this thesis.

This study pays close attention to possessives and quantifiers. A quantifier may appear juxtaposed to the entity it quantifies or it may be connected by the partitive morpheme a. It has been shown that quantifiers are not in the same paradigm as the definite article since the latter can precede the former. On the other hand, there are a number of constraints which influence the use of quantifiers with possessive and demonstrative pro-forms. These pro-forms cannot act as preposed qualifiers to indefinite determiners or generalised quantifiers whereas they can act as preposed qualifiers to numerals, fractions and the morpheme hu motch. In the majority of possessive syntagms, the syntactic ordering is of the possessor+possessum type. Only possessives combining fo+pN or fo+cN[11] exhibit reversed syntactic ordering to display the possessum+possessor type. The possessum can be preposed or postposed by fo+possessive pro-form.

Chapter 5 sets out to examine the way relative clauses operate in VinC and the relativisation strategies used. Basing the analysis on Keenan & Comrie’s (1977,[12] 1987[13]) Accessibility Hierarchy, I identify seven (7) processes of relativisation which are not in themselves independent since VinC, like many languages, tends to associate various strategies, so that embedding one clause into another (the matrix clause) often operates with the deletion of the relativised NP, which in turn, may be reappear in the utterance as a relative pro-form. In some cases a pronominal trace may signal the underlying position of the deleted relativised NP (cf. pronominal copying). Other relativisation processes include clause final preposition stranding, left movement of the possessive relative pro-form which drags the relativised NP along with it (cf. pied-piping) and finally, noun complementation introduced by fo.

The syntactic concept of noun complementation is particularly interesting in the analysis of postposed qualifiers. It enables to provide a unified description of all syntagms that are positioned after the noun. Consequently, modality clauses, prepositional clauses and noun complements can be accounted for using the same terms of analysis as for reduced relative clauses. More formally, it is shown that between the NP and its complement, there is an elided position, which can be potentially filled by the relative pro-form. In these terms, a semantic parallel is drawn between postposed qualifiers and the relative clauses as well as a syntactic or functional mapping with reduced relative clauses.

Chapter 6 posits that semantic concepts like referentiality, genericity, specificity and definiteness can be related to the grammatical category of determiners. Contrary to the hypothesis put forward by Bickerton (1981)[14] which surmises that determinerless NPs are exclusively non-specific, this critical analysis reveals that determinerless NPs can have a variety of readings ranging from generic to non generic, from unique specific reference to non unique specific reference in as much the same way as nouns which are preceded by definite determiners. VinC therefore has no separate morphosyntactic marker which coincides with the semantic distinction between generic and specific. The overt use of the determiner makes it possible to situate the NP on a scale of identifiability: NPs which are predetermined by overt definite markers are more easily identifiable than those which bear overt indefinite markers. The question that arises is how to account for bare NPs felicitously. 

What this analysis shows is that the absence of determiners does not imply absence of identifiability. On the contrary, the examination of situational references reveals that whenever a bare NP appears, the entity to which it refers is easily identifiable. In other words, a null determiner does not correspond to zero specificity but rather a lower degree of definiteness with respect to specified NPs. Preposed and postposed quanlifiers simply serve to narrow the intensional definition of the referent. This allows us to account for the duality of bare NPs which appear either in fossilised expressions or in contexts where the noun forms a collocation with the preposition or verb displaying deictic semantic features.

This thesis recommends an analysis in terms of familiarity and the ability of interlocutors to identify NP reference in order to fully grasp the concepts of definiteness and specificity, grammatical and semantic terms which become inoperative in the analysis of determinerless NPs. It is thus posited that determinerless NPs be characterised as previously thought of, known, foreseeable or easily localisable entities. Their use emphasises the consciousness of the speaker with respect to the listener’s degree of familiarity vis-à-vis the referent, which allows the latter to successfully identify it. The null determiner+N combination refers to a wide range of entities: 

Ø      A NP that is previously thought of by the interlocutor and which appears as a second mention or anaphoric cross reference;

Ø      A NP that is not yet thought of by the interlocutor but about which the speaker will be careful to provide sufficient information in first mentions with cataphoric force;

Ø      A unique NP which is easily identifiable since it is the only representative existing in the cultural or geographic space of the listener.

Ø      A generic NP which has several representatives. The listener will make the necessary selection in order to reconstruct the meaning of the utterance if need be. This phase of meaning reconstruction is necessary as it is omitted by the speaker who prefers to rely on the intrinsic capacity of the linguistic sign to refer;

Ø      A NP which the speaker hopes the listener will eclipse in order to focus his/her attention on the clause as a whole and not on the predicate which is geographically anchored in situ and which forms one with the verb or the preposition collocated with the noun.

Conclusion and prospects: This thesis presents some obvious limitations since it focuses almost exclusively on the description of the NP. Consequently, many aspects remain unaccounted for, namely, the ways in which the VP may influence the semantics or the syntax of the NP. This aspect will be examined in subsequent studies. Likewise, this study does not claim to offer a comparative approach to creolistics. However, interesting parallels are drawn between VinC and other English-lexicon creoles of the Caribbean, albeit in simple terms. Since those creoles are well documented, the reader can make appropriate comparisons, and consequently, draw conclusions about the similarities or divergences between VinC and these creoles.

In the immediate future, my intention is to make a translated copy of this grammatical description available to the partners in education and culture in SVG so that those who wish to formalise their knowledge of VinC will find, in this thesis, the tools to do so. More formally, a contrastive approach vis-à-vis standard English will enable us, as educators, to make the population of SVG cognizant of the differences between VinC and standard English. Such an approach could better prepare teachers of English to deal with problems hindering bilingualism in their learners. This study of NPs is meant to offer a wealth of didactic prospects. More generally, it could contribute to the field of comparative creolistics, a useful area in linguistic typology.