SYNTAGMATIC RELATIONS OF NOUN MODIFIERS IN AKAN BY JULIANA AFRIYIE THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL FULFILLMENT OF THE REQUIREMENT FOR THE AWARD OF MPHIL LINGUISTICS DEGREE. JUNE, 2014 University of Ghana http://ugspace.ug.edu.gh ii ABSTRACT Relationship in linguistics is very essential. There are two types of relationship in linguistics: paradigmatic and syntagmatic. Syntagmatic relations are based on the linear character of speech. They enable language to function as a means of communication. The syntagmatics of noun modifiers are therefore the relation of noun modifiers with the HN and each other. Adopting Dependency Theory from Hays (1972), we base our argument on the hypothesis that the occurrence of any modifier in the noun phrase is more related to and conditioned by the HN. Noun modifiers in Akan therefore depend on the HN for meaning and form. We also argue that the combining properties of modifiers are their potential ability to get into syntagmatic relation and to pattern with other modifiers of appropriate types. In this study we combine modifiers of the same type together in one NP and test how grammatical the construction turns out. The various kinds of determiners are put in one construction to test their grammaticality. The data collected shows that certain determiners can occur in a sequence. We also test plural marking of all the types of noun modifiers in Akan. We find out that Prenominal modifiers do not have plural markers. This study also tests how acceptable an NP is when the order of its modifiers is altered. Certain constructions with altered arrangement of modifiers are accepted to some level. With the use of data we also examine the movement of NPs, that is, the HN, the adjective, and the determiner are “moved about” in a clause. Consequently, we conclude that syntagmatic studies are good basis for language communication. University of Ghana http://ugspace.ug.edu.gh iii DECLARATION I, Juliana Afriyie do hereby declare that with the exception of the references which have been duly acknowledged, this dissertation is the result of my personal research carried out under the supervision of Professor Kofi Saah and Doctor Charles O. Marfo of the Department of Linguistics, University of Ghana, Legon and Department of Modern Languages, Kwame Nkrumah University of Science and Technology, Kumasi respectively. I also declare that this dissertation has not been submitted, either in part or whole for another degree elsewhere. ……………………………… Date……………………….......... Juliana Afriyie (Candidate) ……………………………… Date…………………………….. Prof. Kofi Saah (Supervisor) ……………………………… Date…………………………... Dr. Charles O. Marfo (Supervisor) University of Ghana http://ugspace.ug.edu.gh iv DEDICATION This research is dedicated to my father, Nana Kwame Appiah II: my backbone. University of Ghana http://ugspace.ug.edu.gh v ACKNOWLEDGEMENT I thank God for seeing me through this period of thesis writing. I could not have done it without Him. I thank my supervisors Prof. Kofi Saah and Dr. Charles Marfo for guiding me through to the end. I also thank the lecturers of the Department of Linguistics, Legon: Dr. Amuzu, Prof. Nana Aba Amfo, Prof. Agyekum, Dr. Akanlig-Pare, Dr. Agbedor, Dr. Saanchi, Dr. Hudu, Prof. Duthie, a big gratitude goes to you for your assistance and valuable comments during my progress report presentation. I also owe loads of gratitude to Mr. Frimpong for making the Florence Dolphyne Library available for use, and to Ante Agartha for seeing to it that my letter of change of topic gets to the Dean of Faculty of Arts. My sincere thanksgiving goes to all my coursemates: Abdul-Razak Sulemana, Esther Dogbe, Vera Gyamera, Dominic Frimpong, Muhammed Sumaila and Christorph Amekpordi for your support. I also thank all my friends who supported me in diverse ways. Finally, my sincere gratitude goes my father Nana Kwame Appiah for his love, support and encouragement. Despite all the help I received in the course of this research, I am responsible for all the remaining errors of this research. University of Ghana http://ugspace.ug.edu.gh vi TABLE OF CONTENT Title Page…………………………………………………………………………...i Abstract…………………………………………………………………………….ii Declaration………………………………………………………………………...iii Dedication…………………………………………………………………………iv Acknowledgement……………………………………………………………........v Table of Content…………………………………………………..………………vi List of Tables……………………………………………………………………...xi List of Figures………………………………………………………………….....xii List of Abbreviation……………………………………………………………...xiii CHAPTER ONE: GENERAL INTRODUCTION 1.0 Introduction…………………………...............................................................1 1.1 Background to Study…………………………………………………………..1 1.1.1 Paradigmatic Level of Constituents Arrangement…………………............1 1.1.2 Syntagmatic Level of Constituents Arrangement………………………….2 1.2 Akan: The Language and the People………………………………………….5 1.3 Nominal………………………………………………………………….…….6 1.3.1 Noun Phrase in Akan…………………………………………………….......6 1.3.1.1 Prenominal Modifiers………………………………………....……...........7 1.3.1.2 Postnominal Modifiers………………………………………….................9 1.4 Hypothesis……………………………………………………………...........11 1.5 Aims of Study…………………………………………………….…….…...12 1.6 Research Questions………………………………………………………….12 University of Ghana http://ugspace.ug.edu.gh vii 1.7 Significance of Study……………………………………………………......12 1.8 Data Collection Method………...…………………………………………...13 1.9 Related Works ………………………………………………………………15 1.10 Theoretical Framework……………………………………………………...16 1.11 Organization of the Study………………………………………………….17 CHAPTER TWO: TYPES OF SYNTAGMATIC RELATIONS, THEIR APPROCHES AND THEORETICAL FRAMEWORK. 2.0 Introduction……………………………………………………..……….…...20 2.1 Types of Syntagmatic Relations in Linguistics……………………………...21 2.1.1 Constructional Relations………………………..…………………...……...21 2.1.2 Functional Dependency Relations………………..…………………..…….22 2.1.3 Occurrence Dependency Relations…………………..……………......……23 2.1.4 Realizational Relations…………………………………..………...……….24 2.2 Types of Approaches used in Syntagmatic Studies.........................................25 2.2.1 Dependency-only Approach..........................................................................25 2.2.2 Constituency-only Approach………………………………………….........26 2.2.3 Systemic Functional Grammar Approach……………………………........26 2.3 Theoretical Framework……………………………………………………....28 2.3.1 Dependency Theory…………………………………………………..…….28 2.3.2 D-Trees…………………………………………………………..…….......30 2.4 The Significance of Dependency Theory……………………………...…….32 2.5 Summary………………………………………………………………..........33 University of Ghana http://ugspace.ug.edu.gh viii CHAPTER THREE: NOUN MODIFIERS IN AKAN 3. 0 Introduction……………………………………………………………........34 3.1 Determiners in Akan……………………………………………………........34 3.1.1 The Determiner no ‘the’……………………………………………………35 3.1.2 The Determiner yi ‘this’………………………………………………........36 3.1.3 The Determiner bi ‘a certain’………………………………………….......37 3.1.4 The Zero Determiner……………………………………………………….37 3.1.5 The Determiner kō/korɔ ‘a particular’……………………………………..39 3.1.6 The determiner saa ‘this/that’……………………………………….……...42 3.2 Adjectives……………………………………………………………………44 3.2.1 Morphology of Adjectives…………………………………………………45 3.2.1.1 Pluralisation of Adjectives………………………………………….......45 3.2.1.2 Adjective Diminution …………………………………………………..47 3.2.1.3 Indicating Comparison with Degree Modifiers………………………...47 3.2.1.4 Adjective Reduplication………………………………………………….48 3.2.2 Nominal Adjectives…………………………………………………….......49 3.3 Quantifiers in Akan………………………………………………………….51 3.4 The Relative Clauses in Akan………………………………………………..53 3.4.1 A Head/Antecedent NP………………………..……………………….......54 3.4.2 A Obligatory Relative Clause Marker áà…………………………………..54 3.4.3 A Resumptive Pronoun in the Relativized Position………………………..55 3.4.4 A Clause-final Determiner…………………………….……………………55 3.5 Summary…………………………………………………..…………………56 University of Ghana http://ugspace.ug.edu.gh ix Endnotes…………………………………………………………………………..57 CHAPTER FOUR: ANALYSIS OF NOUN MODIFIERS IN AKAN 4.0 Introduction……………………………………………………………...........59 4.1 The Head of the Noun Phrase…………………………………………….......59 4.2 Sequencing of Modifiers………………………………………………...........60 4.3 The Relation of Postnominals and the HN……………………………………61 4.3.1 The Relation of Adjective and the HN……………………………………...61 4.3.1.1 A Dependency Tree of an Adjectival Phrase……………………………..62 4.3.2 Food items as HN and their modifiers……………………………………..63 4.4 Prenominal Modifiers and the HN………………………………………........66 4.4.1 HNs and Possessive Pronouns/Morphemes………………………………...67 4.4.2 Saa and its HNs…………………………...………....…………….….........68 4.4.3 Nouns as Modifiers……………………………………………………........69 4.4.4 Post and Prenominal…………………………………………………….......71 4.5 Linearity and Possessive Morpheme.……………………………………........72 4.6 Adjectives, Numerals and Quantifiers………………………………………..74 4.6.1 Observation……………………………………………………………........77 4.7. Sequencing of Determiners in Akan…………………………………………77 4.7.1. Observation………………………………………………………………...79 4.8 Scattered NPs…………………………………………………………………79 4.9 D-Trees and the NP…………………………………………………………...83 4.9 Summary……………………………………………………………………...85 University of Ghana http://ugspace.ug.edu.gh x Endnotes………………………………………………………………………...85 CHAPTER FIVE: CONCLUSION 5.0. Introduction………………………………………………………………......87 5.1. Major Findings……………………………………………………………….87 5.2. Summary of the Chapters………………………………………………….....89 5.3 Some Future Research Areas…………………………………………………90 Appendix……………………………..…………………………………………...91 Bibliography………………………………………………………………...…….94 University of Ghana http://ugspace.ug.edu.gh xi LIST OF TABLES Table 1: A list of human and non-human quantifier in Akan…………………...52 Table 2: A Table of Possessive Pronouns/morphemes in Akan…………………67 Table 3: Acceptability Levels in the ordering of Adjectives, Quantifiers and Numerals in the Noun Phrase……………………………………….76 University of Ghana http://ugspace.ug.edu.gh xii LIST OF FIGURES Fig 1: A Tree of an NP having prenominal and Postnominal Modifiers……….…7 Fig 2: A Dependency Tree of (23)…………….………………………………….31 Fig 3: A Dependency Tree of (24)……………………………………………….31 Fig 4: A Dependency Tree of (25)……………………………………………….32 Fig 5: The Arrangement of Post Nominal Modifiers in Akan…………………...60 Fig 6: Dependency Tree of (77)…………………………….…………………….63 Fig 7: Dependency Tree of (116)……………………...……………………….....84 University of Ghana http://ugspace.ug.edu.gh xiii LIST OF ABBREVIATIONS ABBREVIATIONS FULL FORM 1PL First Person Plural 3PLU Third Person Plural 1SG First Person Singular 2SG Second Person Singular 3SG Third Person Singular ADJ Adjective ADJP Adjective Phrase ART Article CFGs Context-Free Phrase Structure Grammars DEM Demonstrative DET Determiner DG Dependent Grammars DP Determiner Phrase DT Dependency Theory FUT Future Aspect Marker H Head HAB Habitual Aspect Marker HN Head Noun MOD N Modified Noun MOD Modifier(s) MTT Meaning-Text Theory University of Ghana http://ugspace.ug.edu.gh xiv N Noun ND No Date NEG Negative NP No Page NP Noun Phrase NUM Numeral OBJ Object PERF Perfective Aspect marker PLU Plural PROG Progressive Marker PST Past Aspect Marker QUANT Quantifier(s) REL Relative Clauses SG Singular STAT Stative Aspect Marker University of Ghana http://ugspace.ug.edu.gh CHAPTER ONE GENERAL INTRODUCTION 1.0 Introduction This chapter gives the general introduction to Syntagmatic Relations of Noun Modifiers. The introduction consists of the background which explains syntagmatic relations and how different it is from paradigmatic relations. The chapter further explains the terms that come together to form the categories under discussion; nominal and modifiers. The discussion of the modifiers is carried out by looking into the various post-nominal and pre-nominal modifiers as have also been discussed in Akan and English where the need arises. The hypothesis, aims significance of the study, and study questions are all discussed in this chapter. Furthermore, this chapter introduces works that will be reviewed and the theory for data analyses. It finally concludes with how the rest of this study is organized. 1.1. Background to Study Generally, words are combined to form phrases, phrases are used to form clauses and then to sentences. It is important to understand the arrangement of constituents, but it is equally important to understand the relationship between them. This brings to the fore the two levels of word relations: paradigmatic level and syntagmatic level. 1.1.1 Paradigmatic Level of Constituents Arrangement At the paradigmatic level, consideration is given to the relationship of a category with other words in the vocabulary system. “Paradigmatic relations are not University of Ghana http://ugspace.ug.edu.gh 2 displayed in constituent structure diagrams but are captured in a more indirect fashion in dictionaries, where the entry for each word specifies which diagram or form of class the word belongs to” (Brown et al 1991: 257-258). According to Van Vallin (2001: 87), “paradigmatic relation refers to the relation holding among elements which can be substituted for each other in a specific environment.” Considering the phrase in (1) below, it could be noticed that very can be substituted by words like really, incredibly and amazingly; Nice can be replaced by words such as kind, friendly, cordial or nasty. These two sets of words, that is very and its possible substitutes and nice and its substitutes comprise two form classes or paradigms and they are said to be paradigmatically related (Brown et al 1991). Cox (1982) explains that paradigmatic relations involve a horizontal relations or associations. In example (2) below, Cox continued that pint would be in paradigmatic relations with words like glass, quart, or gallon. 1. very nice boy 2. A pint of milk (Cox 1982:379) Knowing what paradigmatic relations is, we will move to the discussion of syntagmatic relations. 1.1.2 Syntagmatic Level of Constituents Arrangement Brown et al (1991:258) trace the origin of the word syntagmatic from Greek. According to them, the word means “placed together” and it is certainly related to syntax which also means ‘placing together’ in Greek. Even though it is not University of Ghana http://ugspace.ug.edu.gh 3 uncommon to consider a syntagmatic relation as just placing together of constituents, it must be noted that that kind of relation goes beyond that. Brown et al (1991: 258) point it out that “on a more complex note, syntagmatic relation is not just the placing of constituents together in a linear order or what types of constituent can go together and in what order they can occur”. Syntagmatic studies analyze the semantic structure of a word in its linear relationship with neighbouring words. Lviv (2010: 1) places syntagmatic on the plane of speech meaning. To him, speech should be understood as the manifestation of the system of language in the process of communication. Thus, the structure of a word in its linear relations with a neighbouring word is based on the meaning that structure communicates. Solntsev (1983:60) highlights it by saying that “syntagmatic relations are based on the linear character of speech. They enable language to function as a means of communication. They are therefore the functional relations of language”. According to Cox (1982: 379), Lyons (1964: 74) states that “the context of a linguistic unit is specifiable in terms of its syntagmatic relations.” Cox makes his point clear by using example (2). He explains that pint contracts syntagmatic relations with a, of and milk in the context given in (2). Therefore, when accessing a particular word in a sentence, we have to deal with interference from other words in the sentence (Dell et al 2008: 2). There are different kinds/forms of syntagmatic relations in syntax. According to Lieberman and Truberzkoy (2001:78), “the relation between the element modified and its modifier is one of the most common of all syntagmatic relations, but others like the subject and predicate relation which cannot be considered as a modifier University of Ghana http://ugspace.ug.edu.gh 4 and the modified also exist”. He continues to say that there are many languages with a special means of marking the relation of modified to modifier, and in most of them, this marker is not applied to the subject-predicate relation. There is also another form of syntagmatic relations which is presented by associative phrase (Lieberman and Truberzkoy 2001: 79). In such relations, terms in association stand in syntagmatic relation to some other member of the utterance. This could be two subjects with the same predicate, two predicate with same subject, two modifiers modifying the same element (Lieberman and Truberzkoy 2001: 79). Gibbon (1997: 2) also gives five kinds of syntagmatic relations in syntax. According to him, “complex linguistic units are defined by syntagmatic relations, that is, relations between complex objects and their parts, or between the parts of these objects and other parts”. The kinds of relations he gives are; immediate dominance, linear precedence, sentence functions, government relations, dependency relations, binding relations. There are three general types of syntagmatic relations: coordination, subordination and predicative. In coordination, linguistic units are considered to be on equal rank and are therefore independent of each other. Example of coordinate relation is (3). Subordination is a relation of dependence. Subordinate syntagmatic relation has three levels, morphological (4), word-group level (5) and predicative and subordinate clauses, also known as sentence level. The third type of syntagmatic University of Ghana http://ugspace.ug.edu.gh 5 relation is predicative. Predicative is a relation of interdependence: primary and secondary predication. 3. Kofi  ne pa pa kɔ -ɔ afuom Kofi and father go-PST farm ‘Kofi and father went to farm 4. ku!a + ni = kuani  farming + affix = ‘farmer’ 5. Abaaye wa kɔkɔɔ no girl fair the ‘the fair girl’ This study concentrates on the dependency relation of modifiers and their HNs which is subordination, specifically, the word-group level of subordination like (5). At this point, I will briefly talk about the language in question, Akan, the study questions, organization of the rest of this study among others. 1.2. Akan: The Language and the People Akan is the name of an ethnic group and the language spoken by this group. It is a language of the Kwa branch of the Niger Congo family. The subdivisions of Akan people are Asante, Brong, Kwahu (Kwawu), Akwamu, Akyem, Anum, Awutu, Efutu, Akuapem, Agona, Fante, Wassa, Nzema, Ahanta, Aowin, Sehwi (Sefwi), and Baoule, Anyi, Abron, and Nzema in Cote d’Ivoire as well as other minor groups. According to Dolphyne (1988: xi), three of the Akan dialects, Akuapem, Asante, and Fante, have acquired literary status. For the purpose of this study, the Asante dialect would be focused on. University of Ghana http://ugspace.ug.edu.gh 6 1.3. Nominal The term nominal is used to describe any structure that is headed by a noun or pronoun. In this study, the term nominal is used to refer to the noun which is being modified. The simplest noun phrase contains only the noun. Syntactically, a noun phrase in Akan is made up of the head noun, determiner, adjectives, numerals, quantifiers and among others. Of all these elements, the only obligatory candidate is the head noun, whilst the rest are optional linguistic elements performing variable grammatical functions like specifiers or complements of the head noun within the structure of the noun phrase (NP). Now let us discuss what Akan noun phrase is. 1.3.1 Noun Phrase in Akan According to Abakah (2005: 276), in Akan, two sub-types of NPs are distinguished. “One type has the Head Noun (HN) preceded by one or more phrase modifiers in specifier position (premodification) while in the other, the HN is followed by one or more modifying word classes (post modification)”. Example (6) below is a noun phrase in which the pre- and post-modifier components of the NP in Akan occur in relation to the HN in a tree diagram. In the following sections, we discuss prenominal and postnominal modifiers in Akan. 6.Kɔbena no wɔfa ne akonta ne akyerɛkyerɛfo atsentsen baanan Kɔbena his uncle his brother-in-law his teachers tall four no nyinaa the all ‘All the four tall teachers of Kɔbena’s uncle’s brother-in-law’ (Abakah 2005: 277, ex 6.5) University of Ghana http://ugspace.ug.edu.gh 7 Fig 1: A Tree of an NP having prenominal and Postnominal Modifiers (Abakah 2005:227) NP PreMod HN PostMod NP Poss PreMod NP Adj. Num. Det. Quant. NP Poss PreMod NP RE NP Poss Kɔbena no wɔfa ne akonta ne akyerɛkyerɛfo atsentsen baanan no nyinaa (Abakah 2005: 277, ex 6.5) 1.3.1.1. Prenominal Modifiers Three types of premodifiers have been identified in Akan. One of these is the possessive morpheme which is joined to the HN. According to Abakah (2005), this morpheme is covertly expressed by the Akuapem and Asante speakers of the language but the Fante speakers overtly express it. However, in some cases, the Asante and the Akuapem speakers also express this possessive marker. In isolation, Akuapem and Asante speakers covertly express possession when the noun is a possessive pronoun. It is therefore uncommon to find expressions like University of Ghana http://ugspace.ug.edu.gh 8 (7) – (9). Before we go further, let us note that “in Akan, possessive pronouns do not stand alone; they always occur with the head noun” (Abrefa 2011: 13). If possessive pronouns always occur with the head nouns, then they are premodifiers in Asante and Akuapem too. Comparing (10) with (11), we can observe that in (11) pa pa ‘father’ has its premodifier me, ‘my’, wɔfa ‘uncle’ has a premodifier being no ‘his’ and adamfo, ‘friend’ n(e) ‘his’. The Asante version makes us understand that papa, wɔfa and ada m fo are all modified by m e but wɔfa and adam fo have ne. 7. me m !pa 1SG bed ‘my bed’ 8. yɛ(n ) akokɔ 1PL chicken ‘our chicken’ 9. ne wɔfa 3SG uncle ‘his/her uncle’ 10. Me papa wɔfa adam fo….(As.) 1SG father uncle friend ‘A friend of my father’s uncle’ 11. Me papa ne wɔ fa n’-a dam fo…(Fa) 1SG father 3SG uncle 3SG-friend ‘The friend of my father’s uncle’ The second prenominal modifier in Akan is the determiner saa, which Boadi (2005:123) refers to as article. This article can occur with determiners like yi and no as shown in (12). University of Ghana http://ugspace.ug.edu.gh 9 12. Saa  onipa {no/yi} As-identified-earlier person {this/that} ‘{this/that} person as earlier referred to or characterized’ (Boadi 2005: 134, ex 29). The third type of prenominal modifier is a noun functioning as a modifier. Osam (2002) says that a noun can be used as a non-adjectival modifier when it precedes the HN. This makes the non-adjectival noun modifier a prenominal modifier with regards to its position in the structure of the NP. The following examples are taken from Osam (2002: 177). If the head of (13) and (14) as discussed by Osam (2002) are anoma ‘bird’ and mo ‘rice’ respectively, then the modifiers (wuramu) and (sarmu) are preceding it, then a conclusion could be drawn that nouns can function as premodifiers. 13. Wu ramu anoma Forest bird (Osam 2002:117, ex. 20a) 14. Sarmu mo Savannah rice (Osam 2002:177, ex. 20b) 1.3.1.2. Postnominal Modifiers Postnominal modifiers in Akan include determiners, adjectives, relative clauses, and quantifiers. In the chapter three of this study, we will deal with these modifiers, however, we will give a brief account of each modifier here. According to Boadi (2005: 124), all determiners occur in post-head position in Akan. Determiners are modifiers of head noun in the noun phrase. University of Ghana http://ugspace.ug.edu.gh 10 Adjectives are post modifiers in Akan. Osam (1999: 189-191) uses Dixon’s (1982) semantic classification to identify seven semantic types of adjectives in Akan. These types are extensively discussed in the chapter three of this study, but let us note them in passing as follows: i. Dimension: kɛseɛ ‘big’, tenten ‘tall’ ii. Physical property: de n ‘hard’, me rɛ ‘soft’ iii. Colour: tuntu m ‘black’, kɔkɔɔ ‘red’ iv. Human propensity: tirimuɔdenfoɔ ‘wicked’, huuhuu ‘boastful’ v. Age: da!da ‘old’, kumaa ‘young’ vi. Value: bɔne ‘bad’, fɛɛfɛ ‘beautiful’ vii. Speed: nya a ‘slow’, ha(re ) ‘fast’ All the afore-mentioned types of adjectives can occur in a particular phrase. However, “the choice of a particular order will be based on what property of the referent the user wants to emphasize” (Pokuaa et al 2007: 147). Another modifier which occurs post-head is relative clause. Saah (2010) defines relative clauses as “embedded/subordinate clauses that typically serve as noun modifiers within an NP structure”. He gives four salient features of Akan relative clauses. i. A head/antecedent NP ii. An obligatory relative clause marker áà University of Ghana http://ugspace.ug.edu.gh 11 iii. A resumptive pronoun in the relativized position iv. A clause-final determiner All these features are discussed in chapter three. The rest of the post-head modifiers which will be dealt with in chapter three of this study are quantifiers, numerals and cardinals. These modifiers are all “partakers of the noun modifiers that code notions of quantity, extent, number or serial order” (Givon 2001: 4). The main focus of this study is not to describe modifiers, even though that would not be overlooked, the study proposes to theoretically explain the structure of the modifiers in relation to the HN. Along with the explanation of the relation, all these afore-mentioned modifiers are thoroughly discussed specifically in chapter three of this study. 1.4. Hypothesis The study is based on the hypothesis that, in maximal NP construction, all the modifiers must be in the domain of the HN, no matter how distant the modifier is from the HN (Bunt 1996). Therefore, if a modifier is not under the domain or projection of the HN, the construction becomes ungrammatical. This hypothesis is based on the assumption that the presence of a modifier in a given NP is conditioned by the HN. This hypothesis will be proven by discussing in detail the postnominals and the prenominals in Akan of the HN and where they occur in the morphosyntax of the NP. University of Ghana http://ugspace.ug.edu.gh 12 1.5. Aims of Study The study aims at investigating the syntagmatics of noun modifiers in Akan, the dependency of the modifiers on the noun they modify. The study adopts Gibbon’s (1997) fourth kind of syntagmatic relations (Government Relations: Head and Modifier) in analyzing modifications of the HN in Akan. Even though the primary aim of the study is not to talk about modifiers of the HN, it seeks to give a comprehensive study of the various modifiers of the HN i.e., to provide a structural analysis of the NP. 1.6. Research Questions The following are the questions this study seeks to answer: i. Does a change of a specific modifier from its assigned position to a different position in the noun phrase give a different meaning to the phrase or does it make the phrase ungrammatical? ii. How many modifiers can the head noun have? iii. How many of a particular modifier can occur in the noun phrase? iv. Can the modifiers be anything else besides being nominal modifiers? 1.7. Significance of Study The following are the contributions this study hopes to add to the literature in linguistics: i. This study contributes to the knowledge of syntactic structures in the Akan language. University of Ghana http://ugspace.ug.edu.gh 13 ii. This study also contributes to the communicative effectiveness of the language as language learners and speakers of the language would be cautious of the ordering of the modifiers of the HN. The study brings to view as to how communication is distorted if modifiers are not arranged properly. iii. It contributes to the studies of modifiers in Akan in general as all the modifiers beside their syntagmatic relation, are extensively discussed. iv. It serves as a document of the Akan language, thereby preserving the language. v. It will help researchers who would want to do similar research on Akan modifiers. 1.8. Data Collection Method Data for the study was mainly collected in Akan. Data for the study was collected by using the two main types of data collection; primary and secondary. The primary data was collected by the use of questionnaires. The questionnaires were in two parts: the acceptability part and YES or NO part. The acceptability part was used to test the arrangement of modifiers. Respondents were given a set of constructions in Akan using three or more modifiers of the head noun. The construction in each set differs from the others by the order in which the modifiers appear. Respondents were asked to rank the construction in each set using the following scale: 0 = Completely Unacceptable 1= Highly Unacceptable 2 = Quite Unacceptable 3 = Quite Acceptable 4 = Highly Acceptable 5 = Completely Acceptable University of Ghana http://ugspace.ug.edu.gh 14 The acceptability scale was carried out on the arrangement of modifiers. A look at the questionnaires at the appendix shows that the HN was moved from its position throughout (106a) to (106e) in 4.6. In each set, the total number of respondents who gave a particular order of specific rank was scored. A percentage of what was scored for that group was calculated over the total number of respondents for that question. The higher percentage scored was interpreted as the more preferred order. With the second part of the questionnaires, respondents were asked to rate the constructions with either a YES or a NO. The second part was carried out to test the sequencing of determiners and the arrangement of the modifiers and as well as the HN. With the YES or NO part, respondents were asked to rate the constructions with either YES or NO, where YES means the construction is grammatical and actualized in speech and a NO-rated construction is ungrammatical and therefore, not said at all. The respondents were native speakers of Asante dialect of Akan who have lived in Kumasi for more than 20 years. Twenty respondents were selected with their age ranging from 30 and above. Fourteen out of the twenty were illiterate whilst the rest, (6) were literate. The sentences were read to the respondents who were illiterate. The analysis was based on the respondent’s intuitive judgments regarding the sets of constructions that were presented or read to them. University of Ghana http://ugspace.ug.edu.gh 15 1.9. Related Works Works on noun modifiers are very extensive and in Akan, it is no different. There have been studies on the individual modifiers like adjectives (Osam 1999; Saah and Osam 2010; Pokuaa et al 2007; Amfo et al 2007; Osam 2002). These works describe adjectives, how they are derived and how they occur in a phrase. Boadi (2005) discusses determiners, quantifiers and relative clauses in Akan. He discusses the unmodified HN and then goes further to give the various types of determiners of the HN in Akan. Boadi (2005) gives the types of relative clauses and strategies for deriving them. Saah (2010) also extensively discusses relative clauses in Akan. These studies do not provide the relationship aspect of these modifiers with their HNs they occur with. This study seeks to tackle the relationship aspect, that is, the dependency aspect of these modifiers. Research on syntagmatic relations has also been carried out. Work such as Green (1995) looks at syntagmatic relation in information retrieving. According to Green (1995), syntagmatic relation is essential in information retrieving. She says that “conceptual syntagmatic relationships can narrow subject expressions in meaningful ways and thus promote greater precision” (Green 1995: 371). She continues that in order to limit the set of information being retrieved, there is the need for precision. Therefore, narrowing by specifying the relationships between concepts, which is relational narrowing, would often be better than Boolean narrowing. She cites example that the retrieval output “libraries IN schools” would be smaller in size than “libraries AND schools” University of Ghana http://ugspace.ug.edu.gh 16 In Lviv (2010), the syntagmatic and paradigmatic peculiarities of adverbs in English are discussed. The issues Lviv (2010) deals with include the syntagmatic valency of adverbs and its actualization in speech, syntactic valency and combinability patterns of adverbs, and the semantic and syntactic properties of adverbs of degree. However, these works do not talk about what happens to modifiers and that is what this study deals with. 1.10. Theoretical Framework The study uses Dependency Grammar, hereafter DG, in analyzing the data. This study will use Dependency Grammar and Dependency Theory (DT) interchangeably. The basic idea of DG is that syntactic structure consists of lexical items, linked by binary asymmetric relations called dependencies. DG can be primarily traced back to the work of Lucien Tesnière, a French linguist. This quote is a translation of Tesnière (1959) in Kübler and Dickenson (2010:4) The sentence is an organized whole, the constituent elements of which are words. Every word that belongs to a sentence ceases by itself to be isolated as in the dictionary. Between the word and its neighbors, the mind perceives connections, the totality of which forms the structure of the sentence. The structural connections establish dependency relations between the words. Each connection in principle unites a superior term and an inferior term. The superior term receives the name governor. The inferior term receives the name subordinate. University of Ghana http://ugspace.ug.edu.gh 17 Hays (1972: 223) throws light on the core issue of dependency theory as follows. “The insight formalized by dependency theory is that particular occurrences of minimal units are directly related to one another. In English the occurrence of an adjective before a noun, for example, is more closely related to the noun occurrence than to anything else”. He further exemplifies his point by using the occurrence of prepositions. He explains that “an occurrence of a preposition is generally most directly related to two other occurrences, its object and the unit modified by the prepositional phrase. Prepositions are therefore regarded as connectors, and their distributions are most easily stated by reference to classes of objects and classes of governing or modified units”. Following the words of Hays, it could also be realized that in Akan the occurrence of a modifier before or after a noun more closely relates to the noun than anything else in a given phrase. The occurrence is so dependent that in “describing the distribution of a class of adjectives, therefore, one naturally refers to the classes of nouns with which the adjectives occur” (Hays 1972: 223). A modifier; determiner, adjective, quantifier, relative clause, etc. occurs in a NP solely because of the modified. 1.11. Organization of the Study This thesis is organized into five chapters. Each chapter has an introduction, which captures the core issues of the chapter. Within every chapter are University of Ghana http://ugspace.ug.edu.gh 18 subdivisions. These subdivisions have various headings. The following is the summary of the organization of this study. Chapter one is the general introduction to this study. It talks about the background to the study, syntagmatic and paradigmatic levels of word relation. The chapter also touches on the Akan Noun Phrase, prenominal and postnominal modifiers of the Head Noun. It further highlights the aims and significance of the study. The method of data collection is also discussed in this chapter as well as the hypothesis of the study. It concludes with a summary of the organisation of chapters in the work. Chapter two deals with types of syntagmatic relations in linguistics, types of approaches used in syntagmatic representation and theoretical framework. In this chapter, we discuss the four types of syntagmatic relations. These types are based on Rastall (1999)’s work. We then go further to deal with some models Teich (1998) talks about. We also talk about the theory for the data analysis; Dependency Theory in this chapter. Chapter four is devoted to noun modifiers in Akan. Chapter looks at the syntagmatic relations of the noun modifiers discussed in chapter three of this study. University of Ghana http://ugspace.ug.edu.gh 19 Chapter five, the conclusion chapter of this study provides the key findings of the study. It also gives a brief summary of the study and provides some recommendations for future research. University of Ghana http://ugspace.ug.edu.gh 20 CHAPTER TWO TYPES OF SYNTAGMATIC RELATIONS IN LINGUISTICS, THEIR APPROACHES AND THEORETICAL FRAMEWORK 2.0 Introduction There are different types of syntagmatic relations in Linguistics. In section 1.1.2 of this study, we mentioned three general types of syntagmatic relations. Rastall (1999) also gives four types of syntagmatic relations: Constructional Relations, Functional Dependency Relations, Occurrence Dependency Relations, and Realizational Relations. In this chapter, we review all these four types of relation and then conclude that the HN and its modifiers fall under Occurrence Dependency Relation. As this study will show in chapter four, the occurrence of a modifier solely depends on the presence of the HN in a given NP. For that reason, this study adopts Dependency Theory (DT) which was originally proposed by Tesniere (1959). The key claim of Dependency Grammar is that “every phrase has a most prominent element (the head) which determines its syntactic properties. Modifiers and complements of the head are called dependents” (Covington 1992:1). At this point it is essential to give how the rest of the chapter is organized. This chapter is divided into five sections. In section 2.1, I discuss Rastall’s four types of syntagmatic Relations. In 2.2, I discuss types of approaches used in syntagmatic studies. The approaches discussed in 2.2 are based on Teich (1998). The University of Ghana http://ugspace.ug.edu.gh 21 approaches are three and they are as follows: Dependency-only Approach, Constituency-only Approach, and Systemic Functional Grammar Approach. Section 2.3 deals with the theoretical framework for the study. Section 2.4 deals with the significance of DT to this study. Section 2.5, summarizes this chapter. 2.1 Types of Syntagmatic Relations in Linguistics Different types of syntagmatic relations exist in linguistics. This section of the study discusses the syntagmatic relations that Rastall (1999) gives. Rastall (1999:80) states that “it is syntagmatic relations which are normally regarded as the main concern of grammar, although it is clear that, at the level of sentences, we must account for combination of syntagms (and other signs) with parasyntactic features.” He identifies four different types of grammatical relations. They are: constructional relations, functional dependency relations, occurrence relations and realization relations. 2.1.1 Constructional Relations Constructional relation explains that, a pair which is either an autonomous syntagm may directly join with appropriate parasyntactic features to make up a sentence (a model for an actual utterance) (Rastall 1999). The same pair can also commute with other signs to function as units in further constructions. For instance, at can be in a construction relation with home to form the pair at home, which can commute with here. The pair at home can further conjoin with parasyntactic features to form a sentence, enter into further construction like I University of Ghana http://ugspace.ug.edu.gh 22 amat home. Let us come to Akan. Below is a construction consisting of Ama and kɔ-ɔ‘went’. The construction (15) can go into further construction to form (16). 15. Am!a kɔ-e Ama go-PST ‘Ama went’ 16. Am!a kɔ-ɔ Ko fi  fi e Ama go-PST Kofi house ‘Ama went to Kofi’s house’ 2.1.2 Functional Dependency Relations A pair of signs, (say x and y) are said to be in a functional dependency relation when x depends on y for its communicational function or y depends on x for its communicational function or the two depend on each other (mutual dependency) in the syntagm (Rastall 1999). For instance in the construction we stay at Accra, we depends on stay for its function in we stay and Accra on at in the syntagm at Accra. In Akan, an area where functional dependency is used is the function of pronouns in a sentence. The form of pronouns in the subject or object position do not change. The only way by which the function of a pronoun could be known is by its relation with the verb as well as its position. In (17), the subject of the verb bo -o ‘beat’ is as the same as that of the object of bo-o ‘beat’ in (18). They could only be differentiated by their function in the construction. The subject depends on the verb for its communicational function, so is the object. University of Ghana http://ugspace.ug.edu.gh 23 17. Me-bo-o Ko fi  1SG-beat-PST Kofi ‘I beat Kofi’ 18. Kofi  bo-o me Kofi beat-PST me ‘Kofi beat me’ 2.1.3 Occurrence Dependency Relations The third type of relation discussed by Rastall (1999) is occurrence dependency. In occurrence relations, the presence of a sign in a construction calls for the presence of the other sign. That is to say, if two signs a and b are in an occurrence relations, the presence of a calls for the presence of b or the occurrence of b demands the occurrence of a or “that the presence of each is necessary for the presence of each other” (Rastall 1999: 84). He gives example in support of this point by saying that “a noun girls, is necessary condition for an article in English (those girls) and in preposition phrase, the preposition and the nominal mutually presuppose each other from the point of view of occurrence (as opposed to function)” (Rastall 1999: 84). Likewise in Akan, a noun conditions the occurrence of the determiner, adjective, quantifier or any of the noun modifiers. Without the HN, noun modifiers cannot stand or occur on their own in a NP. Even when they do occur alone in a sentence, we presuppose that the noun or whatever is being referred to is already known. 19. Mmeɛnsa na ɛ-ba-eɛ Three FOC it-come-PST ‘It is three that came’ University of Ghana http://ugspace.ug.edu.gh 24 20. M-mɔfra mmeɛnsa na ɛ-ba-eɛ PLU-child three FOC it-come-PST ‘It is three children that came’ 21. M-mɔfra nnan ba-a ha PLU-child four come-PST here ‘Four children came here’ 22. Ɛnnan ba-a ha  Four come-PST here ‘Four came here’ 2.1.4 Realizational Relations This type of relations involves the “sequence of units in utterances and with the contextual selection of allomorphic variants (function)” (Rastall 1999: 84). Realizational relations are concerned with the ordering of grammatical units. The ordering of grammatical units may show their grammatical relations (Rastall 2004). Rastal (2004) gives more information on Realisational Relations in Grammar. All the aforementioned types of relations Rastall (1999) gives are of different order of importance in linguistic studies, specifically in grammar. Constructional relations show which signs may be combined into syntagms, functional dependency relations tell us how syntagms are internally organized as well as how one syntagm is related to the other. Occurrence dependency and realizational relations are solely interested in the actual sequences in utterances. The study concentrates on dependency relations between the HN and its modifiers. Therefore, the syntagmatic relations of noun or nominal modifiers are considered University of Ghana http://ugspace.ug.edu.gh 25 to be an occurrence dependency relation. Let us review some of the approaches used in syntagmatic studies in linguistics. 2.2 Types of Approaches used in Syntagmatic Studies This section of the study discusses the various approaches used in syntagmatic studies. The section solely concentrates on approaches discussed by Teich (1998). Teich (1998) discusses the different kinds of approaches used in syntagmatic relation in linguistics, specifically, in computational linguistics. The types of representations or approaches he talks about are Dependency-only approach, Constituency-only approach, Hybrid constituency-dependency approach and Systemic Functional Grammar approach. Teich argues that the more diversified model of syntagmatic relations which does not encounter problems of dependency- only approach and constituency-only approach is the one employed in systemic functional grammar. The purpose of this section is to bring to light that there are different types of approaches used in syntagmatic relations analysis. In 2.1, we dealt with Rastall’s four kinds of syntagmatic relations. In this section, we will show that every type of syntagmatic relations requires a distinct approach to analyse it. Let us, however, observe all of these approaches as follows. 2.2.1 Dependency-only Approach Dependency-only approach gives importance to the category head. This kind of approach according to Teich (1998: 39), Tesniere (1959), “maintains that it is sufficient to account for the relation between words for a syntactic description to be adequate, the word being the only syntactic unit acknowledged”. This approach University of Ghana http://ugspace.ug.edu.gh 26 is comprehensively discussed in section 2.3. The head determines the presence of the other elements in a construction. 2.2.2 Constituency-only Approach Another type of approach Teich talks about is Constituency-only Approach. Constituency approach argues that the category head is not a necessary element. This approach uses five candidates for the concept of head: the subcategorizand, the semantic argument, the morphosyntactic locus, the determinant of control, and the constituent determining government. Zwicky (1985), a strong advocate of this approach says that the above listed candidates have to be included in any grammar model if it is going to be interfaced with semantics, the lexicon and the morphology. Zwicky’s concept of head seems to be dependent on the ‘morphosyntactic locus’. Hudson (1986:110), points out that the “notions ‘syntactic head’ and ‘morphosyntactic locus’ appear to be one and the same”. Constituency approach considers head as not only superfluous, but as “a completely different additional category whose use for a grammar model is doubtful” (Teich 1998:40). 2.2.3 Systemic Functional Grammar Approach The third type of approach for syntagmatic studies this study would talk about is Systemic Functional grammar (Henceforth SFG). Teich (1998: 40) considers it as “a way of looking at syntagmatic relations that is not a priori committed to a strict notion of dependency or constituency.” As Teich (1998) following Halliday University of Ghana http://ugspace.ug.edu.gh 27 (1979) notes, SFG maintains that there are four different kinds of syntagmatic patterning. These are as follows: Prosodic structure, Periodic structure, Interdependency and Constituency structure. The first syntagmatic patterning Teich talks about is prosodic structure. Teich (1998) explains that agreement is a syntagmatic phenomenon that is prosodic in nature. Therefore, a particular realization effect spreads over more than one constituent. SFG takes care of the prosodic nature of the relation. The second kind of patterning is periodic structure. Teich (1998) points it out that there are constructions that dependency-only and constituency only approach cannot be used to analyze. One of such construction is information structure. Teich explains that information structure for instance belongs to a class of structure considered to be periodic and dependency-only or constituency-only approach cannot explain it. The third patterning Teich brings to the fore is interdependency patterning. Structures like coordinate structures are paratactic and they belong to interdependency structure. Teich (1998:44) argues that though dependency-only approach can handle hypotactic structures it cannot do so about the former because there is no identifiable head element in paratactic structures and that their elements are rather mutually dependent. Hence, SFG becomes the ideal approach. The fourth syntagmatic patterning found in SFG is structures with elements like Subject, Object, Predicate or Actor, Goal Process. Teich’s argument is that “the University of Ghana http://ugspace.ug.edu.gh 28 categorical values (NP, PP etc.) are simply attributes associated with functional constituents. For further information on all these types of syntagmatic approaches and the four patterning of SFG, see Teich (1998) After dealing with all aforementioned approaches, we could realize that the choice of an approach to any syntagmatic analysis would be based on the kind of construction being dealt with. Even though the study concentrates on dependent- only approach, other approaches will be used when we get to the chapter four of the study. For instance, in the discussion of coordination, paratactic aspect of SFG is used, hence the need to mention it under this section. Again, the section’s discussion of the other approaches shows why the selection of Dependency Theory is necessary for the data analysis. We will now deal with the theory for the data analysis. 2.3 Theoretical Framework This section of the chapter concentrates on the theory for data analysis, Dependency Theory. The section explains the general view of the theory, it then goes further to discuss what Dependency Trees, the trees used for dependency relation in DT are. This section also gives the reason why this study adopts DT for the data analysis. 2.3.1 Dependency Theory The general view of dependency theory according to Hays (1972: 223) is the idea of “tactic relations among elementary units in utterances.” The relation is in such University of Ghana http://ugspace.ug.edu.gh 29 a way that “particular occurrences of minimal units are directly related to one another” (Hays 1972: 223). Hays (1972) uses adjectives and prepositions to explain dependency theory. His point is that “the occurrence of an adjective before a noun in English is more closely related to the noun occurrence than to anything else” in a given expression. Therefore, “in describing the distribution of a class of adjectives, one has to refer to the classes of nouns with which the adjectives occur” (Hays 1972: 223). Likewise, the occurrence of a preposition is generally most directly related to two other occurrences, its object and the unit modified by the prepositional phrase. Prepositions are therefore regarded as connectors and their distributions are most easily stated by the reference to the classes of objects and classes of governing or modified units (Hays 1972:224). Hays argument is that the reason for the occurrence of a syntactic unit is because of the presence of another unit in a given utterance. Hence, there is an association of that unit to another. Hudson (2008:6) explains this by saying that DG is based on relationships between words, i.e., dependency relations. Therefore, A → B means A governs B or B depends on A. Ninio (1996:5) clarifies this by saying that Syntax is handled in terms of grammatical relations between pairs of individual words, such as the relation of the subject to the predicate or of a modifier to a common noun. Grammatical relations are seen as subtypes of a general, asymmetrical dependency relation: one of the University of Ghana http://ugspace.ug.edu.gh 30 words (the head) determines the syntactic and semantic features of the combination, as well as controls the characteristics and placement of the other word (the dependent). Ninio (1996:5) After having known what DT is, we are going to talk about what Dependency Trees (D-trees) are in the following section. 2.3.2 D-Trees A tree is set of elements on which a relation is defined. DT is a “theory of grammars with abstract mechanisms for characterizing sets of utterances and for assigning to them certain structural descriptions called D-trees” (Hays 1972: 224). Zhekova in his presentation of the Mel’c uk (1988) chapter one, explains that D- trees are based on relations. These relations are hierarchical links between the elements. The main operation of the trees is to establish binary relations. They consist only of terminal nodes and they give full specification of all syntactic relations between the word forms. Yli-jyra (2004: 2) explains that “a D-tree consists of dependency links (directed arcs) drawn above sentence – and implicitly, of the sentence itself. The D-tree shows which word is related to which and in what way.” They are lines drawn from heads to dependents. The direction of dependency is marked by an arrow from the head to the dependent (Ninio 1996). Let us consider (23) below. It could be seen that the verb (eats) is the head of the clause. The other elements (Kofi and fufu) are joined to the head. In phrases, the head of phrase functions as the peak of the tree. For instance, in (24) below, University of Ghana http://ugspace.ug.edu.gh 31 (pictures) functioning as the head, is the element to which the modifiers are connected. . 23. Kofi eats fufu. Fig 2: A Dependency Tree of (23) eats NP.SUB NP.OBJ. Kofi eats fufu 24. Some new pictures. Fig 3: A Dependency Tree of (24) pictures some new Some new pictures Another way of representing D-tree is by the use of arcs. The arcs are two- dimensional, directed acyclic graph with a single root. “Each node of the tree represents a word, and directional arcs between the nodes represent the dependency relation, leading from head to dependent” (Ninio 1996:6). A directed tree is a completely connected, two-dimensional, directed acyclic graph with a single root. Each node of the tree represents a word, and directional arches University of Ghana http://ugspace.ug.edu.gh 32 between the nodes represent the dependency relation, leading from head to dependent. The tree is headed by the highest word in the sentence, the root, which is the word that does not possess a head of its own. Sentence (25) is represented with dependency arcs. 25. Old cars make loud noise. Fig 4: A Dependency Tree of (25) noun modifier subject object noun modifier Old cars make loud noise. 2.4 The Significance of Dependency Theory Kruijff (ND) explains that the core concepts of DG as a perspective, namely the relational view arising from the head/dependent asymmetry, has proven useful (and even necessary) for cross-linguistic accounts of grammar, particularly in the explanation of word order and the relation between surface structure and meaning. In this study, our concentration is on the relation between the categories and not the structure of the categories. According to Hudson (2000:1), “grammars, and theories of grammar, can be classified according to whether the basic unit of sentence structure is the phrase (PSG) or the dependency between two words (DG)”. This study concentrates on noun modifiers, their relationship with the HN University of Ghana http://ugspace.ug.edu.gh 33 and how the structure of the HN affects that of the modifiers and their occurrences. So, this study considers this relation as a dependency relation, hence, the adoption of DT or DG for the data analysis. 2.5 Summary This chapter has reviewed the various types of syntagmatic relations in linguistics. The types discussed were taken from Rastall (1999). The types discussed are constructional relations, functional dependency relations, occurrence dependency, and realizational relations. All these relations are about how words relate to each other in constructions. The chapter also discussed the types of syntagmatic representations or approaches in linguistics. This section concentrated on the types of approaches discussed by Teich (1998). However, because this study is using Dependency approach for the data analyses, the third section of this chapter concentrated on Dependency theory. This is because, the relation between the HN and its modifiers is seen as a dependent one. University of Ghana http://ugspace.ug.edu.gh 34 CHAPTER THREE NOUN MODIFIERS IN AKAN 3. 0 Introduction For a full understanding of the syntagmatic relations of noun modifiers in Akan, it is important that we understand the various modifiers in Akan. In this chapter we discuss noun modifiers in Akan. The various noun modifiers have been extensively researched into. Boadi (2005 and 2010) discuss determiners very well, Osam 1999; Amfo et al (2005); Dzameshie 2005and Pokuah et al (2005) also deal with adjectives and Saah (2010) deals with relative clauses very well. Modifiers discussed under this section are based on these afore-mentioned works. This chapter is a review of what others have said about the various noun modifiers in Akan. This chapter is divided into five sections. Section 3.1 concentrates on determiners in Akan, section 3.2 talks about adjectives, in 3.3 we deal with quantifiers, 3.4 deals with relative clauses and section 3.5 summarizes this chapter. As we stated in section 1.3.1.2, post modifiers of the HN will be discussed in this section. Therefore, we must note that, even though the heading of this chapter is Noun Modifiers in Akan, we focus on the post modifiers. 3.1 Determiners in Akan The constitution of the noun phrase is the Head Noun (HN) and its modifiers. Determiners are modifiers of the HN. There are different types of determiners in Akan and other languages in general. Each type of determiner conveys a different meaning. Akan has six determiners which occur after the HN in modification. University of Ghana http://ugspace.ug.edu.gh 35 These determiners are no, bi, , yi, ko , and saa. The only predeterminer in Akan is saa (Boadi 2005). The discussion of determiners in this section is based on Boadi (2005) and Boadi (2010). The selection of these two works is based on the fact that they discuss determiners deeply. 3.1.1 The Determiner no ‘the’ The determiner no ‘the’ indicates that the speaker is far away from the speaker in space and/or time, however, the speaker believes he has grounds to assume that the identity of the thing being talked about is known to both his addressee and himself, thus, it has a definite feature (Boadi 2005). “The reason for his belief may be due to an earlier reference to the item in a conversation or to its visible appearance in the neighbourhood” (Boadi 2005: 127). According to Boadi (2005), no has two important linguistic properties. The first is the property of presupposition. Presupposition means that the item being talk about is known by the speaker and the hearer, hence definiteness. Considering sentence (20), it could be noted that the subject (akokɔ ‘chicken’) as well as the object (a bu ro‘corn’) is definite, and this attribute is because of the presence of the determiner no. 26 A-ko kɔ no mene-e aburo  no SG-chicken the swallow-PST corn the ‘The chicken swallowed the corn.’ University of Ghana http://ugspace.ug.edu.gh 36 The second linguistic property of no, according to Boadi (2005), is its paradigmatic relation with yi ‘this/these’. As we discussed in section 1.1.1, paradigmatic relations is about substitution. The determiner yi has a definiteness nature and like no, its presence in a sentence presupposes that the identity of the noun is known to the speaker and the addressee. These two determiners serve the same deictic function. Boadi (2005: 128) explains that “the deictic system (of no and yi) encodes extra linguistic features of the context of speech; in particular, those features which have to do with temporal and spatial locations of events relative to each other, and to the physical body of the speaker at the time of speaking.” It could be noted that the difference between sentence (26) and (27) is conditioned by the spatial position in relation to the object. The presence of no communicates that the object being talked about is distant from the speaker while yi indicates nearness of the object (Boadi 2005). 27. a-ko kɔ yi  mene-e aburo no SG-chicken this swallow-PST corn the ‘This chicken swallowed the corn.’ 3.1.2 The Determiner yi ‘this’ Another non-generic determiner in Akan is yi ‘this’. The use of yi according to Boadi (2005: 128) “presupposes that the identity of the noun is known to both speaker and hearer.” The occurrence of yi also signifies the noun is relatively near at the time of speaking. In (28), ɔkra man ‘dog’ with the occurrence of yi ‘this’ interprets that it is near the speaker and hearer, and they both know that particular that is being talked about. Apparently they could see it. University of Ghana http://ugspace.ug.edu.gh 37 28. ɔ-kra man yi  wi a-a nam no  bi  SG-dog this steal-PST meat the some ‘This dog stole some of the meat’ 3.1.3 The Determiner bi ‘a certain’ The determiner bi ‘a certain’ encodes specificity. When used in a construction, “the hearer has good reason to assume that theidentity of the item being referred to is known to the speaker, although not to himself necessarily” (Boadi 2005: 127). It should be noted that the absence of bi does not make the noun generic. In sentence (29), kra taa ‘paper’ is non­generic because the speaker has in mind a specific paper. However, the presence of bi in (30) also gives the noun a “specificity-of-reference interpretation in addition to its non-generic reading,…making the hearer assume that a specific identifiable paper is being referred to” (Boadi 2005: 127). But the speaker does not commit himself to claim that there was a specific and identifiable paper in (29). 29. krataa  da a ­konwa no so Paper STAT­lie SG­chair the top ‘There is a paper on the chair.’ 30. krataa bi  da a­konwa no so Paper a certain STAT-lie SG.­chair the top ‘There is a certain paper on the chair.’ 3.1.4 The Zero Determiner The zero determiner in Akan gives a generic and a non-generic meaning to the noun it modifies. That is to say that if a noun occurs with a zero determiner, that University of Ghana http://ugspace.ug.edu.gh 38 noun is not necessarily given a generic reading. This is because “a noun has a generic reading if it refers to the whole class or universe of objects which it denotes rather than any single individual of the class” (Boadi 2005: 124). Considering sentences (31) and (32), it could be noted that the subject of (31) is generic (referring to the whole class of chickens) but that of (32) is non-generic because the speaker has a certain or specific chicken in mind. 31. a-ko kɔ di  aburo  SG-chicken HAB-eat corn ‘A chicken eats corn.’ 32. a-ko kɔ re-di  abu ro SG-chicken PROG-eat corn ‘A chicken is eating corn.’ From the above examples, it could be noted that the generic nature of the noun is not determined by the noun. Boadi (2005:126) explains that “the assignment of the negative or positive value of the semantic feature generic is not determined by the inherent lexical properties of the noun but by the semantic properties of the predicate and a number of other factors like verbal aspect, syntactic environment and pragmatic considerations.” The predicate of (31), di  abu ro denotes general rather than particular actions or events, [but], the predicate of (32) redi  abu ro denotes a particular action (Boadi 2005:126). University of Ghana http://ugspace.ug.edu.gh 39 3.1.5 The Determiner kō/korɔ ‘a particular’ The use of ko ‘‘a particular’ in a sentence presupposes that “the identity of the referent is known to either the speaker or the hearer or both, depending on the situational context” (Boadi 2005: 128). Even though kō has a definiteness nature like no and yi, it however gives additional role by stressing the speaker’s or hearer’s interest in the “sort, individuality, or particularity of the identity” under discussion. Boadi (2005: 128) explains that this is a feature of “the entity’s quiddityi, haeccityii or qualeiii” which no and yi do not have. Distributionally, kō occurs with “NPs which are constituents of a relative clause in surface structure or NPs whose relative clause have been deleted.” Intuitively, one could notice that (34) is the reduced form of (33) and could connect them. 33. Me-bi sa-a a-duane ko rɔ aa Ko fi  tɔn 1SG-ask-PST SG-food particular that Kofi HAB-sell ‘I enquired to find out what particular food Kofi sells.’ 34. Me-bi sa-a a-duane ko rɔ 1SG-ask-PST SG-food particular ‘I asked to find out what particular food’ ‘I asked to find out the particular food’ When ko occurs with a noun as the only modifier, that NP is always a complement. Otherwise the construction is ungrammatical (Boadi 2005). So, sentence (35) according to Boadi (2005) is ungrammatical but (36) is grammatical. 35. *asɛm ko yɛ hu (Boadi 2005: 129) Matter particular STAT-be frightening University of Ghana http://ugspace.ug.edu.gh 40 36. asɛ m ko no yɛ hu Matter particular the STAT-be frightening ‘That particular matter is frightening.’ There are grammatical constructions which have ko /ko rɔ as the only modifier of the HN. If we consider the sentence below, it could be observed that the speaker intends to be derogatory. In a relative clause like (37) below, it could be observed that ko /korɔ is the only modifier of the HN yet, the construction is grammatical. The no in (38) does not modify ɔkra man ‘dog’ but ɛnam ‘fish/meat’. . 37. Maame korɔ ho yɛ hu Woman particular self Stat-be frightening ‘That particular woman is frightening/weird.’ 38. Ɔ -kraman ko /korɔ aa  ɔ-we-e nam no  SG-Dog particular REL 3SG-chew-PST fish the ‘The particular dog that ate the fish.’ Boadi (2005: 130) gives another property of the determiner ko . He points it out that the verb governing a ko -determined NP is selected from a limited semantic class of verbs of complementation. The verbs are those used to make enquiries about the specific nature of phenomena as well as assertions and declarations of known facts about such phenomena." These verbs include bi sa (enquire to find out); ti e (listen to ascertain); hwehwɛ (search with a view to discovering); pɛ (seek to find out); ni m (get to know after a search); kyerɛ (point out, indicate what is known for certain); ka (assert, declare to remove doubt or uncertainty). If Boadi University of Ghana http://ugspace.ug.edu.gh 41 claims that the above-listed verbs only can occur with k , then how will (39) and (40) be accounted for. Obviously, the verb ku m  ‘to kill’ does not fall into the class of verbs used for making enquiries about the specific nature of phenomena, neither is it used for making assertions and declarations of known facts. The verb hu/hu nu ‘to see/find out’ can fall into that category. 39. M-ɛ-ku m ntonto m ko rɔ aa ɔ-we me wɔ ɛda n mu  1SG-FUT-kill mosquito particular that 3SG-HAB-chew 1SG be room in ha yi  here this ‘I will kill the particular mosquito that bites me in this room.’ 40. Da koro pɛ m-ɛ-kye korɔm fo ko rɔ no day one only 1SG-FUT-catch thief particular the ‘I will catch that particular thief one day’ Another property of k is individuation. This property according to Boadi (2005: 130) “stresses the quale of referents and definitizes that its occurrence with nouns of unique reference is unacceptable in Relative Clauses.” Boadi’s claim is that like no and yi (which are definite) k does not occur with proper nouns like Asamoah and “nouns referring to the only one of its kind in existence and known to the hearer like o wi a ‘sun’, ɔsra m/ɔsramo ‘moon’ and personal pronouns.” He therefore considers sentence (41) ungrammatical. The sentence labeled (41) is taken from Boadi (2005: 130). 41. *Me-ni m Asamoah k aa ɔ-wɔ  ha 1SG-STAT-know Asamoah particular REL 3SG-STAT-be here ‘I know the particular Asamoah who is here.’ (Boadi 2005: 130, ex 18a) University of Ghana http://ugspace.ug.edu.gh 42 His point is that “one of the most important functions of k is to stress the uniqueness and particularity of referents, a property which proper names and personal pronouns already incorporate in their reference” (Boadi 2005: 130). What about cases where the speaker wants to stress the fact that he/she knows the particular A samoah being referred to? In that instance, sentence (41) becomes normal. Because the uses ko only to emphasize his knowledge of A samoah but not to render A samoah common, the name still remains distinctive. It does not refute the fact that “the name A samoah does not have the property of individuation and quiddity which are semantic constituents of proper nouns and personal pronouns with human reference” (Boadi 2005: 131). 3.1.6 The determiner saa ‘this/that’ Saa is referred to as the identity article in Akan by Boadi (2005). It has a Fante version which is dɛm, which can substitute the Twi version without any contrast in conceptual meaning. This determiner can co-occur with either of these two determiners yiand no. When this happens, this article receives the following interpretations; “as afore-mentioned, in the manner or capacity specified, characterized, identified or alluded to earlier” (Boadi 2005: 135). Sentence (42) below presupposes shared speaker and hearer’s knowledge about the identity of the referent (Boadi 2005). Unlike the other determiners discussed in sections 3.1.1. to 3.1.5, saa occurs post-nominally. According to (Boadi 2005: 135), “it has non-operator functions in adjunct and complement positions.” See (42) and (43). University of Ghana http://ugspace.ug.edu.gh 43 42. Saa  a-bɔfra {no /yi } As-identified-earlier SG-child {that/this} ‘This/that/the child as earlier referred to or characterized.’ 43. Me-n-twe rɛ saa ISG-NEG-HAB-write the-manner-specified-earlier ‘I do not write like that.’ Boadi also talks about the restrictions on the occurrence of saa  in the NP. The first of these restrictions is that “it can occur as a constituent of the noun phrase only if the NP contains a deictic determiner” (Boadi 2005: 135). The following examples best explain how ungrammatical an expression can be if saa does not co-occur with a deictic determiner. 44.*Sa a maame bi  a-tukwan Article woman a certain PERF-travel ‘Such a certain woman has travelled.’ 45. *Sa a a-duane ko sɛe-e Article SG-food particular go bad-PST ‘Such a particular food went bad.’ However, the ungrammaticality can be rectified only when the HN of the expressions are modified by a deictic determiner as well as the saa article. That is why the sentences below are grammatical. 46.Saa  maame bi  no  a-tu kwa n Article woman a certain the PERF-travel ‘A certain woman (known to you and me, being referred to) has travelled.’ 47.Saa  a-duane ko no sɛe-e Article SG-food particular the go bad-PST ‘That particular food being referred to went bad.’ University of Ghana http://ugspace.ug.edu.gh 44 The second restriction discussed by Boadi (2005) is selectional restriction on the sequence of determiners modifying the HN. That is, there is a restriction on the sequence of any of these three determiners (bi, no and ko ) in the NP, therefore, “sequences of bi, no, ko are allowed but not bi, yi or ko , yi” Boadi (2005: 136). Boadi explains that there should be sequences of distal determiners or a non- definite determiner followed by a definite determiner. With the latter, the definite one neutralizes the non-definite determiner making the expression definite. So, one can say the sentences (48) and (49) but not (50). 48. Sa a  maame bi  no te efi e  no  akyi  (article) woman a certain the STAT-sit house the behind ‘The man referred to is sitting at the back of the house.’ 49. Saa  a-baayewa ko no gyi na  hɔ (article) SG-girl particular the STAT-stand there ‘That particular girl is standing over there.’ 50. *(saa) ɔbarima bi  yi  baa ha (Boadi 2010: 136, Ex. 33c). This section has discussed the six kinds of determiners Boadi (2005) and Boadi (2010) talk about. The determiners are no, bi, yi, saa , ko and . The following section deals with adjectives in Akan. 3.2 Adjectives Adjective as a modifier of the HN in a NP according to Shopen (2007: 168), has two senses when used in linguistics. University of Ghana http://ugspace.ug.edu.gh 45 On the first usage, it is used semantically to denote a set of words on the basis of their meaning, regardless of their grammatical properties in particular languages. On its second usage, it is used as a label for a word class in a particular language defined by grammatical characteristics which distinguish it from other words in that language. Shopen (2007: 168). Before we move further, let us note that adjectives in Akan have been thoroughly dealt with. This section talks about what the various works on adjectives have said. Issues discussed in this section are morphology of adjectives, adjectival reduplication and nominal adjectives. We will begin our discussion with the morphology of adjectives. 3.2.1 Morphology of Adjectives This section deals with the morphological characteristics of adjectives in Akan. Most of the issues we will talk about are taken from Dzameshie (2005). 3.2.1.1 Pluralisation of Adjectives We will begin with pluralisation of adjectives. In section 1.3.1.2, we listed Osam’s semantic class of adjectives in Akan. However, because we will need this in the discussion of adjectives pluralisation, we will bring the class here again. Below is the semantic class of adjectives. University of Ghana http://ugspace.ug.edu.gh 46 i. Dimension: kɛseɛ ‘big’, tenten ‘tall’ ii. Physical Properpeties: de n ‘hard’, mrɛ ‘soft’ iii. Colour: tuntum  ‘black’, kɔkɔɔ ‘red’ iv. Human propensity: ti ri moɔdenfoɔ ‘wicked’, hu uhuu  ‘boastful’ v. Age: da!da ‘old’, kumaa ‘young’ vi. Value: bɔne ‘bad’, fɛɛfɛ ‘beautiful’ vii. Speed: nya a ‘slow’, ha(re ) ‘fast’ Adjectives can be pluralized in Akan. According to Dzameshie (2005: 75), Osam (1999) reports that “adjectives from certain semantic classes may be inflected for plural”. Among this class, colour, age, dimension and some of human propensity adjectives can be pluralized. Human propensity adjectives like hu uhu u  ‘boastful, ani amɔn ho‘‘rude’ cannot be pluralized. 51. Colour: kɔkɔɔ- a-kɔkɔɔ ‘red’ red PLU-red 52. Age: kakraba – n-kakraba‘young’ Dzameshie (2005: 75). young - PLU-young 53. Dimension: kɛ seɛ - a-kɛseɛ  ‘big’ big PLU-big 54. Human Propensity: ti ri moɔdenfo ɔ - a-ti ri moɔdenfo ɔ ‘stingy’ Stingy PLU-stingy University of Ghana http://ugspace.ug.edu.gh 47 3.2.1.2 Adjective Diminution Another morphological characteristic of adjectives is diminution. In Akan, adjectives in the dimension class take diminutive marker which is –wa in Twi and –ba in Fante. Ewe also has a dimunitive marker which can be marked on dimension class of adjectives. Dzameshie says that this marker is a suffix –i which is attached to the adjective to indicate smallness. 55. koro koro + wa korokoro wa short and small + DIM ‘round and small’ 56. kakra + wa  kakrawa small + DIM ‘small’ 57. xoxo + I xoxui/xoxoe Dzameshie (2005: 75) Old + DIM ‘old’ (expressive) Looking at examples (55) to (57), we could observe that all the adjectives the diminutive occurs indicate smallness. However, the diminutive does not intensify the smallness of the adjectives in anyway. 3.2.1.3 Indicating Comparison with Degree Modifiers Adjectives can be marked to show comparative and superlative degrees. In English, for instance, adjectives have affixes indicating comparison (comparative degree, superlative and degree). In Akan, comparison is done by using degree modifiers paa ara ‘very’ or kyɛn/sen(e) ‘than’. It must be noted that Akan adjectives do not show superlative and comparative degrees. Akan adjectives can be used in comparison by the above given degree modifiers. University of Ghana http://ugspace.ug.edu.gh 48 58. ten ten sen(e) tall than ‘taller than’ 59. ketewa paa ara small very ‘smaller than’ 3.2.1.4 Adjectival Reduplication The fourth morphological characteristic this study will discuss is adjectival reduplication. Adjectival reduplication has syntactic and semantics reasons. Semantically, adjectives are replicated for the purpose of indicating intensity or emphasis. In Akan, moreover, adjectives in all the semantic varieties can undergo this phenomenon. According to Ofori (2006: 123), “from the Akan perspective, words with such reduplicated forms could be described as having more/much/’many’ (or ‘large quantities’ as if countable) of the quality contained by the root adjective. The second reason for reduplication is syntactic. According to Ofori (2006), syntactic reduplication allows for predicative CViv and CVC adjectives to have variants that can be used attributively. Ofori (2006) continues that these variants while having the same meaning tend to appear in exclusive environments, the unreduplicated (i.e. [CV] or [CVC] in predicative usage and the reduplicated (i.e. [CV-CV] or [CVC-CVC]) used attributively. University of Ghana http://ugspace.ug.edu.gh 49 We have looked at the morphological aspects of adjectives, we will go to how certain nouns can function as adjectives in Akan in the next section. 3.2.2 Nominal Adjectives In Akan, certain nouns can function as adjectives. Dzameshie states that Osam (1999) found out that when this occurs all these adjectives “semantically mark HUMAN PROPENSITY”. Examples of such nouns include osu ani ‘learner’, oh iani ‘poor person’. Originally, these are nouns but they can occur in the slot of adjectives and function as adjectives. There are ways of distinguishing between ordinary nouns and nominal adjectives. Syntactically, nominal adjectives succeed the nouns they modify when they are used attributively. However, ordinary nouns precede the nouns they modify (Dzameshie 2005). 60. ɔ-baa busuefoɔ SG-woman devilish ‘a devilish woman’ 61. ɔ-bari m!a ti ri moɔdenfoɔ SG-man stingy ‘a stingy man’ The following examples, (62) and (63) are ordinary nouns modifying another noun. None of the nouns are functioning as adjectives here. 62. efi e a-dua ne home SG-food ‘homemade food’ University of Ghana http://ugspace.ug.edu.gh 50 63. Mpo ano anhwea (Dzameshie 2005:77) Beach sand A way by which one can determine whether a noun is an ordinary noun or a nominal adjective in an NP is by using that noun predicatively. We can use bu su efo ɔ and ti ri moɔden foɔ predicatively, like sentences (64) and (65). 64. ɔ-baa no yɛ busu efoɔ  SG-woman the be devilish ‘the woman is devilish’ 65. ɔ-barima yi  yɛ ti ri moɔdenfoɔ SG-man this be stingy ‘this man is stingy’ However, we cannot say sentences like (66) and (67) because unlike bu su efo ɔ‘devilish’ and ti ri mo ɔden foɔ ‘stingy’, adu a ne food’ cannot be used predicatively to modify efi e‘house’ neither can anhwe a‘sand’ be used to modify mpoano  ‘beach’. The reason being that they are not adjectives. 66. *efi e no yɛ a-duane house the be SG-food *The house is food 67. *mpo ano yɛ anhwea beach 3SG-be sand *sand that is on the beach This study will concentrate on the attributive use of adjectives in relation to the HN. The next section is about quantifiers in Akan. University of Ghana http://ugspace.ug.edu.gh 51 3.3 Quantifiers in Akan The discussion of nominal modifiers cannot be done without the mention of quantifiers. Works have been done on quantifiers as modifiers of the HN in the NP. According to Boadi (2005: 171), quantifiers belong to a closed word class that “express quantities, amount, fractions and the percentage of wholes of the entity denoted by the preceding noun”. The following are some of the quantifiers Boadi (2005) gives: kakra  ‘a little, a few’, bebree ‘many’, bi  ‘some’ ɛfa ‘half’, nyi naa ‘all, the whole of’, pi i ‘many, a lot’, e si n ‘half’, n ko ‘only’. Givon (2001) includes numerals to quantifiers. From the definition Boadi (2005) gives, we could also say that Akan quantifiers include numerals, as they (numerals) also express quantity or amount. Boadi (2010) groups quantifiers into definite and indefinite. Cardinal numerals are the definite quantifiers in Akan. Their definiteness is due to the fact that they express “precise numerical quantities of individuals which are lexicalized as count” (Boadi 2010: 143). This type of quantifiers include baako, ‘one’ m mienu  ‘two’ ɛ nnan ‘four’ etc. The list of such quantifiers as stated above include ka kr a ‘a little, a few’, bebree ‘many’, bi  ‘some’ etc. Amfo (2010: 105) also gives a table of quantifiers in Akan. Her table differentiates between human and non-human quantifiers in Akan. University of Ghana http://ugspace.ug.edu.gh 52 Table 1: A list of human and non-human quantifier in Akan. Amfo (2010: 105) QUANTIFIERS GLOSSES Human Non-Human Existential Universal Obi Ebi Someone/ Some Everyone/ Anyone (of) Obiara ebiara - - N bi NP bi A certain/Some (of) - N biara Nbiara - Any/Every -Biribi Something - NP nyinaa NP nyinaa - All - Biribiara - Everything/Anything Human or non-human, numeral or ordinary, quantifiers are post-nominal modifiers in Akan as shown in (68) – (70). 68. ɔ-baa bi ara ho yɛ fɛ SG-woman every self be beautiful ‘Every woman is beautiful’ 69. ankaa bebree gu dan no  mu orange a lot lie room the inside ‘There are a lot of oranges in the room’ 70. a-su kuu-foɔ mmienu  a -twa nsɔhwɛ no  PLU-student-PLU two PERF-pass exams the ‘Two students passed their exams’ University of Ghana http://ugspace.ug.edu.gh 53 Quantifiers can occur in both attributive and predicative positions in Akan. All the sentences we have seen so far show the distribution of quantifiers in the attributive position. 71. a-su kuu-foɔ no yɛ mmi enu PLU-student-PLU the be two ‘The students are two’ 72. ankaa no yɛ  bebree Orange the be a lot ‘The oranges are a lot’ However, none of the quantifiers in table 1 above can be used predicatively. They all follow their HNs and cannot be separated from them by a verb. Hence, (73) is ungrammatical. 73.*Ɔ-baa no yɛ biara SG-woman the be every In (71) and (72) above, we observe that the determiner no directly follows the HN before the predicate. When quantifiers (numeral or quantifiers in general) are used predicatively, the HN should have the determiner no closely following it. 3.4 Relative Clauses in Akan This section deals with relative clauses in Akan. Our discussion is based on Saah (2010). Saah’s work on relative clauses is very extensive, thus our reason for choosing it. Relative clauses are modifiers of the HN in the NP. According to Saah (2010), they typically function as noun modifiers within the NP structure. Semantically, a relative clause may be characterized as a clause that “incorporates, University of Ghana http://ugspace.ug.edu.gh 54 as one of its terms, a nominal which is co-referential with a nominal outside of the clause” (Saah 2010: 92). He gives the following four salient features that a relative clause must have in Akan. i. A head/antecedent NP ii. An obligatory relative clause marker áà iii. A resumptive pronoun in the relativized position iv. A clause-final determiner In this section, we will give a brief explanation of all these features. For further knowledge on the above mentioned features, see Saah (2010). 3.4.1 A Head/Antecedent NP According to Saah (2010), in Akan, however, both definite/specific NPs and proper names can occur as heads/antecedents of relative clauses. “Though the names refer to specific individuals, the relative clauses do provide additional information about the head” (Saah 2010: 103). 3.4.2 A Obligatory Relative Clause Marker áà Saah (2010: 93) says that “Akan uses a particle áà(said with a falling tone) to mark the beginning of the relative clause”. Saah analyses it as a relative complementizer (REL. COMP) for the following reasons: a. It does not occur anywhere else in Akan, neither is it utilized for any other purpose in the language. University of Ghana http://ugspace.ug.edu.gh 55 b. There is no person, number, gender or case agreement between this relative complementizers and the head noun. Thus, no matter the nature or the type of the head noun, the complementizer is always the same. The relative complementizer is compulsory in Akan. Its absence in a sentence with relative clause renders the sentence ungrammatical. 3.4.3 A Resumptive Pronoun in the Relativized Position According to Saah (2010: 97), Schachter (1973: 23) states that “the relativization rules of Hausa, like those of Akan, provide that a special pronoun replaces the NP within the underlying sentence which corresponds to the antecedent.”Saah (2010: 98) explains that “the resumptive pronoun in Akan relative clauses is obligatory. However, it is realized as null or overt depending on whether its antecedent is inanimate or animate.” 3.4.4 A Clause-final Determiner In discussing the clause-final feature of relative clauses in Akan, Saah (2010: 94) says that “two demonstrative elements occur in the position of the clause-final determiner, and when they do, their semantics has to be compatible with the determiner that modifies the head, if there is any.” This clause-final determiner, which occurs at the end of the clause, is obligatory. We have talked about all the four features of relative clauses in Akan that Saah (2010) gives. However, for extensive knowledge on these features, see Saah (2010). Now let us give an example that has all these features. University of Ghana http://ugspace.ug.edu.gh 56 74. maame no áà ɔ -kyerɛ  ade ɛ no ba-a ha Woman the REL COM 3SG-teach something CD come-PST here ‘the woman who teaches came here’ The above sentence contains the head/antecedent NP which is maame‘woman’, a relative clause marker áà, a resumptive pronoun in the relativised position Ɔ ‘3SG, she’ and a clause-final determiner no.All these features are italicized and boldened in (74). 3.5 Summary This chapter has looked at noun modifiers in Akan. The modifiers discussed were determiners, adjectives, quantifiers and relative clauses in Akan. Using Boadi (2005), this chapter has given the six types of determiners in Akan. This chapter has found out that Boadi (2010)’ assertion that ko /ko rɔ cannot occur as the only modifier of the HN, and that when that happens, the construction becomes ungrammatical is not always the case. The occurrence of ko/ko rɔ as the only modifier communicates derogation. This chapter has also shown that saa can co- occur with a non-deictic determiner like yi and that it is not always the case that the co-occurrence of these two determiners is ungrammatical. This chapter has discussed adjectives in Akan. This chapter has shown the types of adjectives Osam (1999) gives. From Dzameshie (2005), this chapter has discussed adjectives in Akan. Quantifiers are also discussed in this chapter. In our University of Ghana http://ugspace.ug.edu.gh 57 discussion of quantifiers, we gave a table of human and non-human quantifiers in Akan that Amfo (2010) gives. With the predicative use of quantifiers, we found out that when quantifiers (numeral or quantifiers in general) are used predicatively, the HN should have the determiner no closely following it. The chapter also discussed relative clause in Akan by reviewing Saah (2010). i Quiddity is the real nature or character of the determiner. ii Haecceity is the essential property that makes k unique from the other determiners. iii Quale is the property of the determiner. This is not the determiner itself but the property. ivCV in phonology is a type of syllable structure. Basically it is Consonant and vowel representing the Onset and Nucleus in syllable structure. University of Ghana http://ugspace.ug.edu.gh CHAPTER FOUR ANALYSIS OF NOUN MODIFIERS IN AKAN 4.0 Introduction This chapter analyses the syntagmatic relations of the postnominal and prenominal modifiers discussed in chapter three and section 1.3.1.1. of this study. This chapter is sectioned into nine. Section 4.1 discusses the head of the NP. In 4.2, sequencing of the modifiers is discussed. In section 4.3, we will deal with the relation of postnominals and the HN. We will deal with prenominal modifiers and the HN in section 4.4. In section 4.5, linearity and possessive morpheme will be our focus. The next section, 4.6 will deal with adjectives, numerals and quantifiers. We will then talk about scattered NPs in section 4.7. Section 4.8 deals with D-trees and the NPs. We will summarise the chapter in section 4.9. This chapter stands on the hypothesis that the occurrence of any modifier in the noun phrase is conditioned by the HN, and it is dependent on it for meaning and form. 4.1 The Head of the Noun Phrase The first sign of any noun phrase when it comes to syntagmatic relations is the head noun. The presence of the HN calls for the presence of the modifiers. Ninio (1996: 5) clarifies this by saying that “grammatical relations are seen as subtypes of a general, asymmetrical dependency relation: one of the words (the head) determines the syntactic and semantic features of the combination, as well as University of Ghana http://ugspace.ug.edu.gh 60 controls the characteristics and placement of the other word (the dependent)”. The head is the absolutely independent part of the phrase. “For any syntagmatic structure, there exists a unique head word, called the absolutely independent part, which does not contain the meaning of the part of the sentence” (Yakovishin 2011: 7). Therefore head is an important notion or element in the noun phrase description. Due to their central role in the phrase, HNs are the most syntactically active elements. 4.2 Sequencing of Modifiers In talking about modifiers, we take into consideration their arrangement in the NP. Appah (2003: 24) gives a diagram of how the modifiers are expected to be arranged when they are all present in the NP. Labeled as Figure 5 below is found in Appah (2003: 24) as (2-29). Standard wise, the arrangement of modifiers is supposed to occur in the order they are represented on Fig. 5 Fig. 5. The Arrangement of Post Nominal Modifiers in Akan N Adj – Num – (Quant) – Det – Quant – Rel. Cl Head We have established the arrangement of modifiers and what we consider the HN. The rest of the chapter will look at the relations of the modifiers and the HN. It must be noted that we are looking at the attributive use of modifiers in Akan. University of Ghana http://ugspace.ug.edu.gh 61 4.3 The Relation of Postnominals and the HN In this section, we will deal with the relation of postnominals and the HN. We know that the relationship is a HN-modifier relationship. We will begin with adjectives and how their position influence what HN they modify. 4.3.1 The Relation of Adjective and the HN In Akan, the sequence of units may show the grammatical relations between them. In Akan, adjectives occur after the noun they modify. The following NPs have adjectives in them. Our focus is on what the adjective is modifying in these two phrases. 75. M-mari ma m paboa a-kɛseɛ PLU-man shoe PLU-big ‘Men’s big shoes’ 76. M-mari ma a-kɛseɛ mpabo a PLU-man PLU-big shoe ‘big men’s shoes’ By virtue of occurrence, we could say that a-kɛseɛ ‘big’ modifies m paboa ‘shoe’ in (75) whilst a-kɛse ɛ ‘big’ modifies mmar im!a ‘men’ in (76). Linearity is very important characteristics of syntagmatic relations and we know that attributively, adjectives occur after the nouns they modify. The structure of a word in its linear relations with a neighbouring word is based on the meaning that structure communicates. From the two, we could see that the adjective directly follows m pabo a in (75) and mmar ima in (76), showing the nouns they are modifying in each phrase. Looking at the standard sequence in Fig. 5, we have no difficulty finding out which noun the adjective modifies. University of Ghana http://ugspace.ug.edu.gh 62 We will deal with a phrase containing an adjective, and a numeral and put it on a D-tree in the next section. 4.3.1.1 A Dependency Tree of an Adjectival Phrase This section deals with how a NP is displayed on a D-Tree. In this section, our focus is on the inflections (the plural markers) on the modifiers. At the base of the tree, is the NP without the inflections. All the plural markers are shown by the use of the tree. Below is D-tree of (77) 77. m -maa a-kɔkɔɔ du  PLU-woman PLU-fair ten ‘Ten fair women’ The HN of (77) is m maa ‘women’ and it is a plural. From the root of m maa to NUM+ maa is an arrow. This arrow points to the number marker prefixed to maa indicating that the prefix M is showing number. From NUM+maa is anther pointer. This pointer says that the number marker is not any other marker but plural. The plural marker is then indicated as PLU+maa. Let us go to the first modifier, the one that occurs immediately after the HN, which is the adjective akɔkɔɔ  ‘red’. Before we go further with explanation of the tree, we should have it in mind that Amfo et al (2005: 67) says that with regard to plural marking, the adjective may take an overt plural marker when the head noun is plural in Ga and Akan.” In our case, the tree shows that the number a, prefixed to kɔkɔɔ ‘red’ is marked with NUM showing number. We then move up to the next arrow where we are told that the number marker is plural. University of Ghana http://ugspace.ug.edu.gh 63 The last modifier in the diagram below is du  ‘ten’. With HN being more than one, none of the modifier can be singular. This is reflected in the numeral in this NP. Fig.6 A D-tree of (77) HN ADJ NUM PLU+maa PLU+kɔkɔɔ PLU NUM+maa NUM+Kɔkɔɔ M-maa a-kɔkɔɔ du Affixation 4.3.2 Food items as HN and their modifiers We have observed in 4.3.1.1 that modifiers morphologically inflect for number and that their inflections are dependent on the HNs. There are different types of nouns in Akan of which most inflect for number. Those that do not inflect for number include mass nouns. Mass nouns are conceptually plural and most of them University of Ghana http://ugspace.ug.edu.gh 64 inflect for the nasal prefix (Bodomo and Marfo 2002). Another type of nouns that do not inflect for number is “food items noun”. We use the expression food item nouns” to refer to the nouns of foodstuff. The nouns here include Akan lexicons for fruits, vegetables, roots and tubers, cereals among others. Tubers in Akan include bankye ‘cassava’, mankani ‘cocoyam’, bayerɛ ‘yam’, and abrɔdwomaa ‘potatoes’. Examples of fruits are ankaa (As), ekutu (Fa) ‘orange(s)’, kwadu ‘banana(s)’, borɔferɛ ‘pawpaw’, and aborɔbɛ ‘pineapple(s)’. None of the foodstuff in the above given list have plural names. When they are used as HNs of an NP, determining their plurality is obscure. One has to consider the modifiers occurring with it before it can be concluded that the HN is more than one. 78. Bankye  a-kɛseɛ mmi enu pɛ na a-ka gu  ɛ-dan no mu hɔ. Cassava PLU-big two only FOC PERF-leave lying SG-room the in there. There are only two cassavas in the room. 79. Bankye  no a-ka asi ni asi ni mmi enu pɛ. Cassava the PERF-leave pieces two only It is left with only two pieces of cassavas. 80. Bankye  bɔtɔ baako na ɔ-de ba-eɛ Cassava sack one FOC 3SG.-use come-PST He\she brought only one sack of cassava. Looking at (78), it can be observed that we are talking about two big tubers or pieces of cassavas. The number of the cassava is determined by their modifiers. Although the HN does not inflect for number, it co-occurs with number (i.e. mmienu). This informs the plural status of the HN and for which reason the adjective could inflect for number in (78). In sentence (79), the HN, bankye has asiniasini and mmienu to show its number. Other items like plantain and banana University of Ghana http://ugspace.ug.edu.gh 65 have special lexicons they occur with. Bɛtem ‘finger’ and esa ‘bunch’ are expressions that occur with banana and plantain specifically. These expressions encode number. 81. Bo rɔdeɛ bɛte m n ti  wo-re-bo a-kɔdaa yi ? Plantain finger why 2SG-Prog-beat SG-child this Are you beating the child because of a finger of plantain? 82. Bo rɔdeɛ m-mɛtem mmeɛnsa bɛ-so fu fuo no. Plantain PLU-finger three FUT-enough fufuo the Three fingers of plantain will be enough for the fufu As can be seen from (81) and (82), bɛtem has a plural form, mmɛtem. Therefore we have can easily determine that the item being talked about is either plural or singular, even if there is no morphological marker to show for it. Vegetables are another food items which do not have plurals. Perhaps it is because they are often measured in containers. Items like pepper, onions, garden eggs and tomatoes are measured in baskets or tins. Often, these measuring items are used to modify them. In cases where any of the above-mentioned vegetable is referred to in a sentence as a HN, one will need the modifiers to determine whether the HN is plural or single. 83. Nya adowa fi ta a no a-sa Garden eggs white the Perf-finish The white garden eggs are finished 84. Ama fa-a nya ado wa a-kɛseɛ mmi enu no aa ɛgu pono  no so no  Ama take-PST garden eggs PLU-big two the Rel. lying table on top CD Ama took the two big garden eggs lying on the table University of Ghana http://ugspace.ug.edu.gh 66 85. Ama fa-a nya ado wa no aa ɛgu pono no so  no Ama take-Pst garden eggs the Rel. lying table the top CD Ama took the garden eggs lying on the table. In (83), the HN is conceptually viewed as more than one even though it has no modifier to show for it. In (84) however, its modifiers, being the adjective, numeral and the relative clause communicate plurality. Looking at the verb in the relative clause, we can say the HN of (85) is also more than one. Looking at vegetables in Akan, specifically Asante, most of them have nasal prefix. In Akan, nasal affixed nouns are underlyingly plural. It is clear that food item nouns like mass nouns do not inflect for number morphologically. It is therefore difficult to determine their number when they are used as HNs unless they are measured like (80), their modifiers show plurality or the sentence they occur in communicates that the item is plural. Without plural morphological inflections, we should note that these nouns are considered plural. Thus, they are conceptually defined as plural than singular nouns. We cannot therefore say that they depend on the modifiers semantically in the above given phrases. 4.4 Prenominal Modifiers and the HN In our discussion of modifiers in Akan in section 1.3.1.1, we identified three different prenominal modifiers in Akan. The possessive morpheme/pronoun, the saa article and nouns being used as non-adjectives. In this section, we talk about the relation of these three modifiers to the HN. In our quest to discuss the relation University of Ghana http://ugspace.ug.edu.gh 67 of Akan prenominals and their HNs, we will seek to answer the question: Do these modifiers become pluralized when the HN they occur with are plural? 4.4.1 HNs and Possessive Pronouns/Morphemes We will begin our discussion by looking at possessive morpheme/pronouns. Below is a table of the Akan pronouns which are also used as possessive pronouns. The table below is from Osam (2003). Table 2: A Table of Possessive Pronouns/morphemes in Akan GLOSSES PRONOUNS/MORPHEMES FULL FORMS 1SG me- me 2SG wo- /i- (Fa) wo 3SG Ɔ- ɔno 3 (inanimate) ɛ (Ak, As, and some Fa dialects) ɛno 1PLU yɛ- yɛn/hɛn (Fa) 2PLU mo- /wɔ- (Fa) mo/hom (Fa) 3PLU wɔ- /yɛ- (As) wɔn/hɔn (Fa) As the table above shows, all the singular pronouns have their plural counterparts. Let us go through the examples (86) – (89) below. In example (86) below, we could observe that the HN akonnwa ‘chair’ is singular so is the possessive pronoun/morpheme. Again, in (87), the HN is singular but the possessive pronoun/morpheme is plural. (88) has its HN being plural but possessive is University of Ghana http://ugspace.ug.edu.gh 68 singular. In (89), both the HN and its modifier are plural. It should be noted that all these examples are grammatical. Perhaps, the plurality of possessive pronouns/morphemes is independent of the plurality of the HNs they occur with. 86. M’-a-kon nwa 1SG-SG-chair ‘my chairs’ 87. Yɛ(n) a-konnwa 1PLU SG-chair ‘Our chair’ 88 Me n-konnwa 1SG PLU-chair ‘my chairs’ 89 Yɛ(n) n-konn wa 1PLU PLU-chair ‘our chairs’ 4.4.2 Sa a and its HNs Our next prenominal modifier is saa, ‘article’. It has been identified as the only determiner that occurs post nominally. The form of saa in all the NPs from (90) to (93) is the same. Saa has no plural counterpart unlike the possessive morphemes/pronouns. Perhaps, determiners in Akan do not indicate plurality. We are making this claim because, in (92) the HN abɔfra  ‘child’ is singular so yi is glossed as ‘this/that’. However, in (93) because the HN m mɔfra ‘children’ is plural, yi is glossed as ‘those/these’. There is a plural version of yi which is eyinom ‘these/those’. The usage of plural determiner, if there is any, to correlate University of Ghana http://ugspace.ug.edu.gh 69 with a plural HN is not obligatory. That is, it is not obligatory that a determiner should carry the number morpheme like that of the HN it occurs with. 90. Saa m -mɔfra no Article PLU-child the ‘The children being referred to’ 91. Saa a-bɔfra no Article SG-child the ‘the child being referred to’ 92. Saa a-bɔfra yi  Article SG-child this/that ‘this/that child being referred to’ 93. Saa m -mɔfra yi  Article PLU-child those/these Those/these children being referred to’ 4.4.3 Noun as a Modifier We have talked about plural marking of possessive morphemes and saa. In this section, we will deal with nouns functioning as non-adjectives to modify the HN it occurs with. We mentioned in 1.2.1.1. that nouns can also modify other nouns and when they do, they precede the HNs. In our discussion of plural marking in non- adjectival nouns functioning as modifiers, let us observe the following examples: 94. Sarmu mo Savannah rice (Osam 2002: 177, ex. 20b) 95. amanɔne n-sɛm  Foreign PLU-issue ‘Foreign news’ University of Ghana http://ugspace.ug.edu.gh 70 96. amanɔne a-sɛm  Foreign SG-issue ‘Foreign issue’ 97. *n/asarmu mo *savannas rice 98. mmɔfra  a -ta deɛ  bi  sɛn aho ma no so PLU-child SG-cloth a certan hang-STAT clothesline the top ‘a certain clothing for children is hanging on the clothesline’ 99. maame no tɔn m-mɔfra  n -ta deɛ Woman the sells PLU-child PLU-cloth ‘the woman sells children clothing’ In (94), both the modifier and the modified are all singular. In (95), we have a plural HN but a singular modifier. (96) has a singular HN and singular modifier, the same modifier like that of (95). The modifier of (95) and (96) is a place, that is, a generic term for all countries besides Ghana. (95) is an expression mostly used in Akan News. The unchangeability of amanɔne ‘foreign’ is because it is a place and nouns indicating place are not marked for plurality. If we should take a critical look at Osam’s example at (94), we cannot have an NP with sarmu ‘savannah’ being plural. It is ungrammatical, thus (97). When we come to (98) and (99), whilst the HN of (98) is singular, its modifier is plural. In (99), both the HN and the modified are all plural. We could not draw a consistency in all these examples, so, we will conclude with the claim that, like the other two prenominal modifiers, saa and possessive pronouns/morphemes, nouns functioning as modifiers do not reflect the plurality of the HN they occur with. University of Ghana http://ugspace.ug.edu.gh 71 4.4.4 Post and Prenominal We have found out that prenominals do not become pluralized even when their HN are plural. In this section, we want to find out whether prenominals can take plural when they have plural HNs and plural postnominals in an NP. We will use the following sentences (100) to (103) for our analysis. 100. Saa n-nua a-kɛseɛ  no Article PLU-tree PLU-big the ‘The aforementioned big trees/the big trees being referred to’ 101. Kofi  ne n-ko kɔ pi i  a-wuwu Kofi his PLU-chicken a lot PERF-die ‘Most of Kofi’s chickens are dead’ 102. Wu ramu m -moa a-kɛseɛ ho  yɛ hu Forest PLU-animals PLU-big self be dangerous/scary ‘Big forest animals are dangerous/scary’ 103. Kofi papa n-nwan a-kɛseɛ mmeɛnsa aa wɔ-year-eɛ no Kofi father PLU-sheep PLU-big three REL. 3PLU-lost-PST the The three fat sheep of Kofi’s father which were missing Looking at (100) to (103), none of the prenominals is marked for plurality even though all the postnominals are plural. We should not overlook the fact that it is the HN that determines the number marking of its modifiers. Prenominals are autonomous with regards to plural marking whereas postnominals depend on their HNs for number. We would like to state that as far as pluralization is concerned, adjectives and relative clauses are the only modifiers that acquire the plural markers of their HNs. Relative clauses display plural adoption as the plurality or the singularity of their HNs is reflected in the modifying clause. University of Ghana http://ugspace.ug.edu.gh 72 In the next section, we discuss the occurrence of prenominal modifier, specifically the possessive morpheme, in relation to the HN it occurs with. 4.5 Linearity and Possessive Morpheme. We have noted in 4.3.1. that arrangement is very essential in modification. We have observed that post nominal modifiers directly follow their HNs. In this section, we will test and see if sequence is essential when it comes to possessive morphemes and their HNs. We will consider (104) and (105) to show how occurrence or sequence of grammatical units correlates with the function of the grammatical units. In (104) we can note that m’(me) asu ku u foɔv ‘students’ whilst in (105), it prefixes adu ane ‘food’. By reference to sequence, we could identify which HN is being modified by the possessive morpheme m’(me) ‘my’. 104. Ko fi  ma-a m’-a-suku u-foɔ  no a-du ane no bi  Kofi give-PST my-PLU-student-PLU the SG-food it some ‘Kofi gave my students some of the food’ 105. Kofi  ma-a a-sukuu -foɔ no  m’-a-du ane no bi  Kofi give-PST PLU-student-PLU the my-SG-food it some ‘Kofi gave the students some of my food’ We can note that m’(me) ‘my’ is in construction with asu ku u foɔ ‘students’ in (104) whilst in (105) it is with adu ane ‘food’. The constructional relations and the dependency relations are clear in these two sentences. The statement is pre- nominals precede HN. Therefore, m’(me) ‘my’ modifies asu ku u foɔ vi ‘students’ but not adu ane ‘food’ in (104) whilst it modifies adu ane ‘food’ in (105). We University of Ghana http://ugspace.ug.edu.gh 73 would like to however state that it is not the sequence that necessarily expresses the modification but it is the expression of the relation that expresses the relation. This affirms Rastall (2004: 46) that says that “The potential mistake lies in thinking that the sequence is the grammatical relation, when it is merely the expression of that relation” According to Green (1995: 369), “words acquire relations based on the linear nature of language because they are chained together…the elements are arranged in a sequence on the chain of speaking. Combinations supported by linearity are syntagms.” Because of this perception, the only thing that comes to mind when dependency is mentioned is precedence and linear orderings of elements or words in a construction. It is not always the case that dependency occurs in linearity. Looking at linearity in dependency theory, Brody (1994) gives a formal explanation of precedence and totality which are linear correspondence axioms in Dependency theory. It must be observed that actual occurrence of syntactic units is not dependent on linearity. It must be noted that sometimes syntagmatics operate in parallel process too. When that happens, the two nouns are conceived as one noun. So when a speaker of Akan speaks of ‘mako kɔkɔɔ’ versus an English speaker’s ‘red pepper’, neither perceives the object separate from the colour in a linear sequence. Structure is made up of elements which are graphically represented as being in a linear progression; but the theoretical relation among them is one of ‘order’. Halliday (1966) says that order may, but does not necessarily have as its realization, sequence, the formal relation carried by linear progression; sequence is at a lower University of Ghana http://ugspace.ug.edu.gh 74 degree of abstraction than order, and is one possible formal exponent of it. A structure is thus an arrangement of elements ordered in ‘places’ We have stated in 1.8. that questionnaires are used in the data collection process. As the study is a relation-based research, we contacted native speakers of the Asante dialect of the Akan language for their views on some constructions which have been altered. The following two sections, 4.6 and 4.7 will discuss others view on the constructions we gave in the questionnaires. 4.6 Adjectives, Numerals and Quantifiers This section examines the ordering of Adjectives, Numerals and Quantifiers in a noun phrase and how their interchangeability alters the meaning of the NP. Data was collected by the use of questionnaires. Respondents were given a set of constructions in Akan using three or more modifiers of the HN. A construction in each set differs from the others by the order in which the modifiers appear. Respondents were asked to rank the construction in each set using the following scale: 0 = Completely Unacceptable 1= Highly Unacceptable 2 = Quite Unacceptable 3 = Quite Acceptable 4 = Highly Acceptable 5 = Completely Acceptable The response of the above questions is put in a table. The table below (Table 3) shows the responses to the ordering (106a – e). University of Ghana http://ugspace.ug.edu.gh 75 106a. M -maa a-kɔkɔɔ m meɛnsa nko ara PLU-women PLU-women three only Noun Adjective Numeral Quantifier A N Q ‘Only three fair women’ b. M -maa a-kɔkɔɔ nko ara m meɛnsa PLU-women PLU-fair only three Noun Adjective Quantifier Numeral A Q N c. M -maa nko ara a-kɔkɔɔ m meɛnsa PLU-women only PLU-fair three Noun Quantifier Adjective Numeral Q A N d. M -maa m meɛnsa nko ara  a-kɔkɔɔ PLU-women three only PLU-fair Noun Numeral Quantifier Adjective N Q A e. M -maa m meɛnsa a-kɔkɔɔ nko ara PLU-women three PLU-fair only Noun Numeral Adjective Quantifier N A Q University of Ghana http://ugspace.ug.edu.gh 76 Table 3. Acceptability Levels in the ordering of Adjectives, Quantifiers and Numerals in the Noun Phrase. Level of Acceptabilit y (a)A-N-Q Freq. % (b) A-Q-N Freq. % (c) Q-A-N Freq. % (d) N-Q-A Freq. % (e) N-A-Q Freq. % Completely Acceptable 18 90% 13 65% 3 15% 0 0% 0 0% Highly Acceptable 1 5% 3 15% 3 15% 0 0% 1 5% Quite Acceptable 1 5% 2 10% 5 25% 0 0% 2 10% Completely Unacceptable 0 0% 0 0% 0 0% 16 80% 14 70% Highly Unacceptable 0 0% 1 5% 1 5% 3 15% 2 10% Quite Unacceptable 0 % 1 5% 8 40% 1 5% 1 5% TOTAL 20 100% 20 100% 20 100% 20 100% 20 100% (106a) contains the prototypical Noun Phrase in which the HN, the adjective, the numeral and the quantifier are contiguous. This ordering received the highest level of acceptability. Out of the 20 respondents, 90% evaluated (106a) as completely acceptable. In (106b), the HN and the adjective are contiguous but the numeral occurs after the quantifier. This construction was rated the second highest as completely acceptable with 65%. 0% rated (106d) and (106e) as completely acceptable as against completely unacceptable. Out of the 20 respondents, 80% considered (106d) as completely unacceptable whilst 70% rated (106e) as completely unacceptable. University of Ghana http://ugspace.ug.edu.gh 77 4.6.1 Observation The above table indicates that there is a certain level of acceptability even when it comes to NPs with unacceptable arrangement of modifiers. Probably, (106b) and (106c) are used, despite the fact that they are not the standard NPs. Again, responses show that the more adjectives become farther from the HN, the more unacceptable the construction becomes. Adjectives should then be close to their HNs. 4.7. Sequencing of Determiners in Akan This section deals with how the various determiners order with respect to each other. We have talked about determiners in Akan in section 3.1. We gave six different kinds of determiners Boadi (2005) gives. In this section, we are looking at why the sequence of certain determiners is acceptable and why others are not. In our discussion of the saa determiner, we realized that, it (sa a ) can co-occur with the other determiners in Akan. Again Boadi (2005) gives a restriction on the sequence of the following bi, no, ko determiners. In section 3.1.6, we discussed that the sequence of bi, no, ko is allowed but not bi, yi or ko , yi. The sequence should be distal determiners or a non-definite determiner followed by a definite determiner. We will use the section B of the questionnaires to look at the sequencing of determiners in Akan. This section will also show whether or not there has been a change in determiner sequencing in Akan. Respondents were asked to rate the constructions with a yes or a no, where yes indicates it is grammatical and it is actualized in speech, and no indicates ungrammatical. University of Ghana http://ugspace.ug.edu.gh 78 All constructions which were rated no, thus ungrammatical by more than twelve (12) respondents are marked ungrammatical. Whilst those rated yes by more than twelve (12) people are analyzed as grammatical in the study. The following NPs (107a) – (107c) have bi occurring as the final determiner. They are ungrammatical. 107a. *A-bɔfra  no bi vii SG-child the a certan *’The certain child’ b. *A-bɔfra yi bi  SG-child this/that a certain *’this/that certain child’ c.*A-bɔfra  ko bi  SG-child particular a certain * a certain particular child’ The following NPs numbered (108a) – (108c) have no as their final determiner. 108 a. A-bɔfra bi  no SG-child a certain the ‘The certain child’ b. *A-bɔfra yi  no SG-child this/that the *The this/that child c. A-bɔfra  ko/ko rɔ no  SG-child particular the ‘the particular child’ In (109a) – (109c), we decided to use ko ‘particular’ as the final determiner in the construction. University of Ghana http://ugspace.ug.edu.gh 79 109a.* A-bɔfra  bi  ko /ko rɔ SG-child a certain particular *’A certain particular child’ b. *A-bɔfra no ko/ko rɔ SG-child the particular ‘the particular child’ c. *A-bɔfra  yi  ko/ko rɔ SG-child this/that particular 4.7.1. Observation From the above NPs, we can observe that all NPs that are rated grammatical are those that have no occurring as the final determiner. (108a) and (108c) were rated grammatical at the end of the data collection. (108b) was not rated grammatical by respondents, even though no occurs as the final determiner. From the analyses, we can conclude that in a construction with determiner sequencing, the final determiner should be no. However yi no sequence is not accepted because yi is a distal determiner whilst no is a proximal determiner. Hence, determiner sequence has to agree in proximity too. Again, only bi no and ko no sequence is accepted. This affirms Boadi (2005) which says the sequence should be distal determiners or a non-definite determiner followed by a definite determiner, so that its definiteness neutralizes the non-definiteness in the other determiners. 4.8 Scattered NPs So far, all the NPs we have used in discussions have modifiers which occur attributively. Thus, all the modifiers directly follow the HNs without any other unit occurring between them. This section discusses what happens when members of University of Ghana http://ugspace.ug.edu.gh 80 the NP are scattered in a simple clause and positioned at a different environment in the clause. Saah (2010) discusses extraposition of relative clause in Akan. Therefore, it is common to come across examples like (110) and (111). 110. Ɔ -bar m b [áà ne d de Nya m !kyɛ n ten -a ase SG-man INDEF REL 3SG.POSS name be called N. CD sit-PST under “There lived a man whose name was Nyamekyɛ.” 111.Ɔ bar m b ten -a ase [áà ne d de Nyam kyɛ ] SG-man INDEF sit-PST under REL 3SG.POSS name be.called N. “There lived a man whose name was Nyamekyɛ” (Saah 2010:102, ex 17a,b) However, when it comes to the other modifiers of the NP such as the determiners, adjectives, quantifiers, numerals etc, what has been talked about is their arrangements in a fixed or restrictive order. We seek to look at what happens when these modifiers are taken out of their appointed environment in the NP and positioned at a different place. This section tests the versatility of the modifiers of the HN. Givon (2001) explains that scattered NPs are reported to be common in languages with flexible word-order. Respondents were asked questions like; is this sentence used in Akan? Is it grammatical to use such expressions? We want to point it out that we did not use the acceptability scale for this section. If a sentence is used in Akan, respondents were asked to answer YES and NO, if the sentence is not used at all. All sentences that were rated NO by the respondents are marked as ungrammatical in this study. University of Ghana http://ugspace.ug.edu.gh 81 The following are the sentences which were read out to the respondents. All constructions which were rated no, thus ungrammatical by more than twelve (12) respondents are marked ungrammatical. Whilst those rated yes by more than twelve (12) people are analyzed as grammatical in the study. 112a. Me-pɛ a-konwa kɛseɛ no 1SG-want SG-chair big the ‘I want the big chair.’ b. * a-konwa no me-pɛ kɛseɛ SG-Chair the 1SG-want big ‘The chair, I want the big one’ ‘I want the big chair.’ c. * Me-pɛ kɛseɛ a-konwa no 1SG-want big SG-chair the *I want big chair the’ d. * kɛseɛ no me-pɛ a-konwa Big the 1SG-want SG-chair e. * kɛseɛ me-pɛ a-ko nwa no Big 1SG-want SG-chair the f. Me-pɛ a-konwa no kɛseɛ 1SG-want SG-chair the big I want the chair big Sentences (112a-d) contain a HN which is modified by an adjective and a determiner as the object. (112a) has the fixed arrangement of the HN and its modifiers, but (112b-d) have the modifiers misarranged. Looking at (112b) closely, we could see that the HN is taken away from the adjective and fronted. Sentences (112b-e) are all ungrammatical. However, (112f) is different in the University of Ghana http://ugspace.ug.edu.gh 82 sense that, though the adjective occurs after the determiner, the sentence is grammatical. It was observed that six (6) of the respondents who rated (112b) ungrammatical explained that though they do not use such expression, they believe that people sometimes use it. For instance when a child is asked to bring a chair, the sender might use (112b). However, the sender is expected to bring no after kɛseɛ. Therefore, one could say (113). 113. a-konwa no me-pɛ kɛseɛ  no PLU-chair the 1SG-want big the I want the big chair However if we topicalised akonwa no by placing a comma after the determiner, (112b) becomes grammatical. See (114). 114. a-konwa no, me-pɛ kɛseɛ  SG-Chair the 1SG-want big ‘The chair, I want the big one’ ‘I want the big chair.’ According to Lviv (2010), the establishment of syntagmatic relations is conditioned by the combining properties of the units entering into the syntagmatic relation. Hence, the combining properties of modifiers are their potential ability to get into syntagmatic relation and to pattern with other modifiers of appropriate types. A respondent even gave an example like (114). University of Ghana http://ugspace.ug.edu.gh 83 115. Me-pɛ fufuo no merɛmerɛɛmerɛ 1SG-want fufu the very soft ‘I like the fufu very soft’ 4.9 D-Rules and the NP This section of the study discusses the application of the D-trees on NPs. As we have said in the chapter two of this study, we will use dependency theory to analyse our NPs. Returning to our theory, we will state again that DT is a relation theory. These relations are hierarchical links between the elements. Trees are used to show this relation. The main operation of the trees is to establish binary relations. We have employed two types of D-trees: the arcs and the trees. We used the arcs in 4.3.1.1. to show the dependency of adjectives on their HNs. One may note that adjective is not the only modifier the study is interested in. We have not talked about all the modifiers coming together in a phrase with a D-tree showing relations. We will now proceed to illustrate a D-tree of a NP with a relative clause. Using (116), let us look at the tree below. 116. m foni n kɛseɛ no aa me-hu-u wɔ efi e hɔ no Picture big the Rel. 1SG.-see-PST at house there the The big picture that I saw at the house. University of Ghana http://ugspace.ug.edu.gh 84 Fig. 7: A D-tree of (116) mfonin kɛseɛ no aa huu me wɔ hɔ efie no mfonin kɛseɛ no aa me-huu wɔ efie hɔ no In the phrase, mfonin projects into the maximal node/projection. Hence, everything starts with it. The modifiers then occur under it. The tree shows that the adjective (kɛseɛ), the determiner (no) and the relative clause marker (aa) are sisters to the mother node mfonin. This is they all occur on the same level. Under the relative clause marker aa is the modifying clause. The modifying clause exists because of the relative marker, so its occurrence is dependent on it. Hence, it is positioned under it. The mother node of the modifying clause is the relative marker. University of Ghana http://ugspace.ug.edu.gh 85 4.10 Summary This chapter has discussed a number of issues. In our quest to discuss the syntagmatics of noun modifiers in Akan, the issues we have dealt with include the relation of postnominals and the HN. This led us to the discussion of adjective and the HN. We found out that in a NP the position of adjective determines the noun it is modifying. We used (75) and (76) to support our argument. We also discussed plural marking with the use of a D-tree in 4.3.1.1. We have also discussed NPs with food items nouns functioning as HN and found out that these though these nouns do not inflect for number morphologically, they are inherently plural. However, when they are used singularly, their modifiers communicate singularity. Further, the chapter looked at prenominals and their HNs and found out that all prenominal modifiers do not adopt the plural markers of their HNs. We then came to the conclusion that as far as pluralization is concerned, adjectives and relative clauses are the only modifiers that acquire the plural markers of their HNs. Relative clauses display plural adoption as the plurality or the singularity of their HNs is reflected in the modifying clause. Again we discussed linearity and found out that even though the linear occurrence of words helps us to determine their relation with their HNs, there are times syntagmatics operate in a parallel process. When it happens that way, words occurring together are not perceived to be in a linear order. Moreover, we tested the arrangement of modifiers and found out that sometimes the accepted adjective - numeral – quantifier ordering is altered. Table 3 is the University of Ghana http://ugspace.ug.edu.gh 86 result of the acceptability levels of the adjective - numeral – quantifier orderings. We observed that the more adjectives are positioned farther from the HNs, the more the NPs become unacceptable. The chapter also discussed determiner sequence in Akan. This chapter also discussed dislocation of NPs. We adopted the term “scattered NPs” from Givon (2001) to refer to this phenomena. We tested whether any of the categories of noun – adjective - determiner NP in (112a) could be moved. The result showed that apart from the prototypical (112a), only (112f) is grammatical. Respondents informed us that one can use (112b) but the person should add no after the adjective to make the determiners two. One is modifying akonwa and the other, kɛseɛ. The last section of this chapter illustrated a NP on a D-tree. We have shown that all modifiers of the HN are sisters of the HN vi Tones on words in isolation may differ from when those words occur in a phrase or a clause. This is what is happening to a sukuufo ɔ  and a dua ne . When they occurred in the associative phrases, the tone of the second syllable of a suku ufo ɔ, changed from high (  ) to low (  ) making it (m’-) a sukuufo ɔ  whilst the first syllable of a dua ne  changed from low to high, making it (m’-)a dua ne  in (104) and (105) respectively. vii One might confuse the determiner bi with the quantifier bi. The former is marked with a high tone (  ) and the latter, a low tone (  ). What is being used here is the determiner bi. See Amfo (2010). University of Ghana http://ugspace.ug.edu.gh CHAPTER FIVE CONCLUSION 5.0. Introduction In this study, an attempt has been made to explain the syntagmatic relations of noun modifiers in Akan by combining some modifiers together in one NP. In many instances, we have proved that some modifiers can occur together whilst others cannot, especially with regards to determiners. We must always keep in mind that the combining properties of modifiers are their potential ability to get into syntagmatic relation and to pattern with other modifiers of appropriate types. In the following sections we outline the key findings of this research and give a summary of the chapters in this study. We conclude by providing some recommendations for future related studies. 5.1. Major Findings A number of issues have been discussed in this study which will contribute to the literature in Akan Syntax. This study has shown that in any determiner sequence, the final determiner must be the definite determiner. The definiteness of no neutralizes the indefiniteness of the determiners occurring before it. We also found out that though yi, and no are all definite determiners, it is unacceptable to put them in a sequence. This is because they do not agree in the distal feature specification. No is a +distal determiner whilst yi is –distal. We have also found out that unlike postnominals, prenominals cannot be inflected when it comes to plural marking. Thus, prenominal do not take the plural markers University of Ghana http://ugspace.ug.edu.gh 88 of the HNs in NP constructions. Adjectives and relative clauses may take overt plural markers when their HNs are plural. Determiners do not take plural markers. Quantifiers have their plurality embedded in them. It is clear that food item nouns like mass nouns do not inflect for number morphologically. Thus, they are conceptually defined as plural than singular nouns. We cannot therefore say that they depend on the modifiers semantically in the above given phrases. Another observation is the more adjectives are moved away from the HNs, the more ungrammatical an NP becomes. In 4.3.1., we found out that by virtue of occurrence, adjectives modify the nouns immediately succeed. Therefore in a NP with two nouns occurring in a sequence, the first noun becomes a premodifier, and the second the HN. In (103), for instance, we could see that the HN immediately precedes the postnominals. We also found out that food item noun like mass nouns do not inflect for number morphologically. In a NP the number is determined by their modifiers. We also noted that these nouns are considered plural. Thus, they are conceptually defined as plural than singular nouns. Whenever we have an NP which has a relative clause as a modifier, the modifying clause is not located on the same level as the relative clause marker. Neither does it occur on the same level as the other modifiers in the NP it occurs in. It becomes University of Ghana http://ugspace.ug.edu.gh 89 a daughter to the relative marker. This is because all modifiers in a particular NP are sisters. 5.2. Summary of the Chapters This thesis is organized into five chapters. Each chapter has an introduction, which captures the core issues of the chapter. Within every chapter are subdivisions. These subdivisions have various headings. The following is the summary of the various chapters of this study. In chapter one, we introduced the study. Brief information on Akan and its speakers are provided. The aims, and the significance of the study are also outlined in chapter one. The research questions, significance of study and the organization of the work are also outlined in this chapter. In chapter two, we discussed the types of syntagmatic relations in linguistics. We also discussed the types of approaches used in syntagmatic relations. The theoretical Framework for the study is also discussed in this chapter. Chapter three was devoted to noun modifiers in Akan. We concentrated on works that have dealt with modifiers in Akan, specifically, postnominals. In the chapter four, we examined the syntagmatic relations of noun modifiers, postnominals and prenominals, are discussed. The arrangement of the modifiers is dealt with. We also discussed the plural marking of prenominals in Akan. We also University of Ghana http://ugspace.ug.edu.gh 90 analysed determiner sequencing in Akan. We talked about all the various prenominals in Akan. Chapter five is the concluding chapter of this study. We outline the key findings of the study in this chapter and conclude that syntagmatic studies are good basis for language communication. 5.3 Some Future Research Areas We hope that this study serves as a good reference for other studies in syntagmatic relations in other Ghanaian languages. The following recommendations are made for future research in the area: a. Future works may have to look at determiner sequencing in other Ghanaian Languages. b. Other syntactic theories could be used for syntagmatic studies in the future. c. Syntagmatic studies of other areas of syntax like the prosodic structures, coordinate structure among others of Akan could also be researched into. University of Ghana http://ugspace.ug.edu.gh 91 APPENDIX DEPARTMENT OF LINGUISTICS QUESTIONNAIRES This questionnaire is intended to help in a research aimed at describing the syntagmatic relations of noun modifiers in Akan. The information gathered is strictly for academic purpose. All information gathered will be treated confidentially. Please give me a little of your time to answer the following questions for me. SECTION A With the use of the numbers attached to the acceptability rank, indicate how acceptable the following constructions are. Below is the scale for rating. For instance, if a construction is completely unacceptable, indicate (0) in the bracket beside it. 0 = Completely Unacceptable 1= Highly Unacceptable 2 = Quite Unacceptable 3 = Quite Acceptable 4 = Highly Acceptable 5 = Completely Acceptable 1. Mmaa akɔkɔɔ mmeɛnsa nko ara [ ] Noun Adjective Numeral Quantifier A N Q 2. Mmaa akɔkɔɔ nko ara mmeɛnsa [ ] Noun Adjective Quantifier Numeral A Q N 3. Mmaa nko ara akɔkɔɔ mmɛnsa [ ] Noun Quantifier Adjective Numeral Q A N University of Ghana http://ugspace.ug.edu.gh 92 4. Mmaa mmeɛnsa nko ara akɔkɔɔ [ ] Noun Numeral Quantifier Adjective N Q A 5. Mmaa mmeɛnsa a-kɔkɔɔ nko ara [ ] Noun Numeral Adjective Quantifier N A Q SECTION B In this section, please indicate with either YES or NO whether the following constructions are actually used. The question is, are the following expressions actualized in Akan? Their actualization tells us that they are grammatical. 6. a). abɔfra no bi  YES [ ]/NO [ ] b) abɔfra yi  bi  YES [ ]/NO [ ] c) abɔfra ko bi  YES [ ]/NO [ ] 7. a) abɔfra bi  no YES [ ]/NO [ ] b) abɔfra yi  no YES [ ]/NO [ ] c) abɔfra ko no YES [ ]/NO [ ] 8. a.) abɔfra bi  ko YES [ ]/NO [ ] University of Ghana http://ugspace.ug.edu.gh 93 b) a-bɔfra no ko YES [ ]/NO [ ] c) a-bɔfra yi  ko YES [ ]/NO [ ] 9. a) Mepɛ akonwa  kɛseɛ no YES [ ]/NO [ ] b) akonwa no mepɛ kɛseɛ YES [ ]/NO [ ] c) Mepɛ kɛseɛ akonwa no YES [ ]/NO [ ] d) kɛseɛ no mepɛ akonwa YES [ ]/NO [ ] e) kɛseɛ me pɛ akonwa no YES [ ]/NO [ ] f) Mepɛ a konwa no kɛ seɛ YES [ ]/NO [ ] University of Ghana http://ugspace.ug.edu.gh 94 BIBLIOGRAPHY Abakah, Emmanuel Nicholas (2005). Tone in Akan Nominals. Studies in the Languages of the Volta Basin 3, 193-218. Abrefa, K. Busia (2011). Tone Patterns of the Possessive Pronouns in Asante- Twi: A preliminary study. A paper presented at the 2011 LAG conference. Amfo, N. A. A., Boateng S. A., Otoo Y. A. (2005). A Comparative Study of the Morphosyntactic Properties of Adjectives in Three Kwa Languages. Studies in the Languages of the Volta Basin 4 (1), 60-71 Amfo, N. A. Appiah (2010) Negation in Akan: Linguistic Convention verses Pragmatic Inference. Journal of Asian and African Studies (80), 97-112 Appah, C. K. I. (2003). Nominal derivation in Akan; A descriptive analysis. MPhil Thesis, NTNU, Norway. Boadi, L. A. (2005). Three Major Syntactic Structures in Akan: Interrogative, Complementation and Relativisation. Accra: Black Mask Limited. Boadi, Lawrence A. (2010). The Akan Noun Phrase: Its Meaning and Structure. Accra: Black Mask Limited. Bodomo, A., and Marfo, C. (2002). The morphophonology of noun classes in Dagaare and Akan. In 33rd Annual Conference on African Linguistics (ACAL 33), Ohio University, March 2002. Brody, Michael (1994). Phrase Structure and Dependence. http://www.phon.ucl.ac.uk/publications/WPL/94papers/BRODY.pdf Brown, E. K., Brown K., and Miller J. (1991). Syntax: A Linguistic Introduction to Sentence Structure. London: Routledge. University of Ghana http://ugspace.ug.edu.gh 95 Covington, Michael (1992). GB as Dependency Grammar. Presented at the International Congress of Linguists, Québec, August 1992. Cox, Jerry L. (1982). Teaching Adjective Endings in German: A Syntagmatic Approach. The Modern Language Journal 66 (4), 379-382 Dell, Gary S., Oppenheim, Gary M., and Kittredge Audrey K., (2008). Saying the Right Word at the Right Time: Syntagmatic and Paradigmatic Interference in Sentence Production. Lang Cogn Process 23(4), 583–608. Dependency Dixon, R. M. W. (1982). Where Have All the Adjectives Gone? And Other Essays in Semantics. Berlin: Mouton Dolphyne, A. Florence (1988). The Akan (Twi-Fante) Language: its Sound Systems and Tonal structure. Accra: University of Ghana Press. Dzameshie, Alex. (2005). A Comparative Study of Adjectives in Ewe and Akan. Studies in the Languages of the Volta Basin 4. Accra: Department of Linguistics. 72-84. Gibbon, Dafyd (1997). Semantics: The meaning of Verbs. (Printed May 25, 1997),http://coral.lili.unibielefeld.de/classes/summer97/verbsem/vsemss97 /node8.html Retrieved on 2nd June 2013. Givon. T. (2001). Syntax: An Introduction. Amsterdam: John Benjamins Publishing Company. Green Rebecca (1995). Syntagmatic Relationships in Index Languages: A Reassessment. The Library Quarterly, 65(4). 365-385 Halliday, Michael A. K. (1966). Some Notes on 'Deep' Grammar. Journal of Linguistics 2 (1), 57-67 University of Ghana http://ugspace.ug.edu.gh 96 Hays, D. G. (1972). Dependency Theory: A Formalism and Some Observations. In Fred W. Householder (ed.), Syntactic Theory one: Structuralist. England: Penguin Books Limited. Pp. 223-240. Hudson, Dick (2000). Dependency Grammar. Course summary and reading of a five-session course July 14-18 2000, at ESSLLI2000 in Birmingham. http://www.phon.ucl.ac.uk/home/dick/esslli.htm Hudson, Richard A. (1987). Zwicky on Heads. Journal of Linguistics 23(1), 109- 132 Kruijff, Geert-Jan M. (No Date). Formal and Computational Aspects of Dependency Grammar: Historical Development of DG. GJ@COLI.UNI- SB.DE. Kübler Sandra and Dickinson Markus (2010). Dependency Grammar. NASSLLI short course on Dependency Parsing, Summer 2010. http://cl.indiana.edu/~md7/nasslli10/01/01-grammar.pdf Lieberman, A. and Truberzkoy N.  S. (2001). N. S. Trubetzkoy: Studies in General Lingistics and Language. In Lieberman Anatoly (ed.) Language Arts and Disciplines. Duke University Press. Lviv (2010). Syntagmatic and Paradigmatic Peculiarities of Adverbs in English. http://revolution.allbest.ru/languages/00352974_0.html Mel’čuk, I. A. (1988). Dependency Syntax: Theory and Practice, Chapter I. New York: The State University of New York Press. University of Ghana http://ugspace.ug.edu.gh 97 Ninio, Anat (1996). A proposal for the adoption of dependency grammar as the f ramework for the study of language acquisition. In G. Ben Shakhar & A. Lieblich (Eds.), Volume in Honor of Shlomo Kugelmass. Jerusalem: Magnes. Pp. 85-103. Ofori, Seth A. (2006). Topics in Akan Grammar. Bloomington: Indiana University. Osam, E. Kweku (2002). Derived Adjectives in Akan and Ewe: a comparative Analysis. Studies in the Languages of the Volta Basin 171-179. Osam, E. Kweku (2003). An Introduction to the Verbal and Multi-Verbal System of Akan. Dorothee Beermann and Lars Hellan (Eds.), Proceedings of the Workshop on Multi-Verb Constructions. Trondheim Summer School 2003 Osam, E. Kweku. (1999). Adjectives in Akan. Afrika und Ũbersee (82), 189-211 Pokuaa, C., Osam E. K., Saah K. (2005). Adjectives Sequencing in Akan: A Preliminary Investigation. Studies in the Languages of the Volta Basin 4 (1), 143-156 Rastall, Paul (1999). Towards a Logic of Grammatical Relations. La Linguistique 35(2), 79-96 Rastall, Paul (2004). Realisational Sequences in Grammar. La Linguistique (40), 43-52 University of Ghana http://ugspace.ug.edu.gh 98 Saah, K. K. (2010). Relative Clauses in Akan. In Aboh, O. Enoch and Essegbey James (eds.) Topics in Kwa Syntax. Studies in Natural Language and Linguistic Theory. Springer, 90-107. Saah, K. K. and Osam, E. Kweku (2009).Adjectives and Qualifications in the Languages of the Volta basin. Studies in the Languages of the Volta Basin 6 (1). 105-112 Schachter, P. (1973) Focus and Relativization. Language 49 (1), 19–46 Solntsev, V. M. (1983). Language: A System and A Structure. Moscow: Nauka Publishing House Teich, Elke (1999). Types of Syntagmatic Grammatical Relations and their Representations. Institute of Applied Linguistics, Translation and Interpreting. University of Saarland, Germany. www.aclweb.org/../w980500.pdf Van Vallin, R. D. Jnr. (2001). An Introduction to Syntax. United Kingdom: Cambridge.University Press Yakovishi, V. S. (2011). Algebraic Representation of Syntagmatic Structures. fccl.ksu.ru/issue11/FCCL_YAKOVISHIN_09.doc. Yli-jyra, Anssi (2004). Axiomatization of Restricted Non-projective Dependency Trees through Finite-State Constraints that Analyze Crossing Bracketing. www.aclweb.org/../w04-1504.pdf Zhekova, Desislava (No Date). Basic Elements of Dependency Representation in Syntax.www.sfs.unituebingen.de/files/Kursmaterialien/.../presentationPrint. pdf. Zwicky, Arnold (1985). Heads. Journal of Linguistics 21, 1-30 University of Ghana http://ugspace.ug.edu.gh 99 University of Ghana http://ugspace.ug.edu.gh