@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Arts, Faculty of"@en, "Linguistics, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Arkoh, Ruby Becky"@en ; dcterms:issued "2011-06-02T19:23:52Z"@en, "2011"@en ; vivo:relatedDegree "Master of Arts - MA"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """This thesis investigates two post-nominal morphemes, bi and nʊ, in Akan (a Kwa language spoken mainly in Ghana). I analyze bi as an indefinite marker and nʊ as a definite marker. Bi occurs in different environments; the main two environments are the pronominal environment and the determiner environment where it follows the noun within a DP. These environments correlate with certain tonal variations. I argue that in the pronominal use, bi is a specific indefinite pronoun and is toneless. Its specificity status is achieved through a weak anaphoric relationship it shares with a previously-mentioned noun in the discourse or a deictic element. In its determiner use, bi is a specific indefinite which is interpreted via a choice function. It bears a high tone. The Akan morpheme nʊ is used for different functions. It is used as a third person singular pronoun, as a definite article, as a distal demonstrative marker and as a clausal marker. These uses correlate with some tonal alternations. I argue that when the morpheme is used as a pronoun, it is toneless; however when it is used as a determiner or a clause marker it is marked with a high tone. I also argue that the morpheme in all its different functions encodes one semantic value which is familiarity. I take familiarity to entail both hearer old and discourse old along the lines of Prince (1988)."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/35083?expand=metadata"@en ; skos:note "     SEMANTICS OF AKAN BI AND NƱ by RUBY BECKY ARKOH B.A. (Hons.), University of Ghana, 2007 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES (Linguistics) UNIVERSITY OF BRITISH COLUMBIA (Vancouver) June 2011 © Ruby Becky Arkoh, 2011 ii   Abstract This thesis investigates two post-nominal morphemes, bi and nʊ, in Akan (a Kwa language spoken mainly in Ghana). I analyze bi as an indefinite marker and nʊ as a definite marker. Bi occurs in different environments; the main two environments are the pronominal environment and the determiner environment where it follows the noun within a DP. These environments correlate with certain tonal variations. I argue that in the pronominal use, bi is a specific indefinite pronoun and is toneless. Its specificity status is achieved through a weak anaphoric relationship it shares with a previously-mentioned noun in the discourse or a deictic element. In its determiner use, bi is a specific indefinite which is interpreted via a choice function. It bears a high tone. The Akan morpheme nʊ is used for different functions. It is used as a third person singular pronoun, as a definite article, as a distal demonstrative marker and as a clausal marker. These uses correlate with some tonal alternations. I argue that when the morpheme is used as a pronoun, it is toneless; however when it is used as a determiner or a clause marker it is marked with a high tone. I also argue that the morpheme in all its different functions encodes one semantic value which is familiarity. I take familiarity to entail both hearer old and discourse old along the lines of Prince (1988). iii   Preface The name of the UBC Research board for the ethic certificate is Behavioral Research Ethics Board. The following is the numbers for the ethic certificate obtained: Date of Approval: Certificate Number: July 11, 2007 H07-01365 June 9, 2008 H07-01365 May 22, 2009 H07-01365 May 4, 2010 H07-01365 May 6, 2011 H07-01365 iv   Table of Contents   Abstract ............................................................................................................................................... ii  Preface ................................................................................................................................................ iii  Table of Contents ................................................................................................................................ iv  List of Tables ..................................................................................................................................... vii  List of Figures ................................................................................................................................... viii  List of Abbreviations .......................................................................................................................... ix  Acknowledgements ............................................................................................................................ xi  Dedication ........................................................................................................................................ xiii  1  Introduction.................................................................................................................................. 1  1.1  Akan .............................................................................................................................................. 1  1.1.1  Sound inventory of Akan ..................................................................................................... 2  1.1.2  Plural marking in Akan ........................................................................................................ 4  1.1.3  Word order in Akan ............................................................................................................. 4  1.2  Methodology................................................................................................................................. 4  1.3  Organization of the thesis ............................................................................................................ 5  2  Internal structure of the Akan DP ................................................................................................. 6  2.1  Introduction .................................................................................................................................. 6  2.2  The basic phrase structure in the language ................................................................................. 6  2.3  The phrase structure for the nominal items in Akan .................................................................. 7  2.4  Conclusion .................................................................................................................................. 11  3  Semantics of the Akan bi ............................................................................................................ 12  3.1  Literature review ........................................................................................................................ 13  3.2  The two-way mechanism of marking indefiniteness in Akan................................................... 15  3.3  The two environments of bi ....................................................................................................... 17  3.4  Tonal alternations of the morpheme bi ...................................................................................... 19  v   3.4.1  Tonal realization of the indefinite pronominal bi .............................................................. 20  3.4.2  Tonal realization of the indefinite determiner bí ............................................................... 21  3.5  The specificity of bi .................................................................................................................... 23  3.5.1  Bi as a specific indefinite pronominal marker ................................................................... 23  3.5.1.1  Bi and one/some-anaphora ................................................................................................ 25  3.5.1.2  Specificity in Enç’s sense .................................................................................................. 29  3.5.2  The determiner bí ............................................................................................................... 33  3.5.2.1   Fodor and Sag’s sense of referentiality ......................................................................... 33  3.5.3  Bí as referential .................................................................................................................. 34  3.5.4  The scope of bí ................................................................................................................... 39  3.5.4.1  Bí and bare nouns .............................................................................................................. 40  3.5.4.2  Bí and kàkrá ‘few’ ............................................................................................................. 40  3.5.4.3  Bí and nyìnáá ‘all’ ............................................................................................................. 41  3.5.4.4  Bí and bíárá ‘every’ ........................................................................................................... 42  3.5.4.5  Bí does not allow intermediate scope with nyìnárá ‘all’ and kàkrá ‘few’ ....................... 49  3.6  Conclusion .................................................................................................................................. 50  4  Semantics of Akan nʊ ................................................................................................................. 51  4.1  Introduction ............................................................................................................................ 51  4.2  Literature review ........................................................................................................................ 52  4.2.1  Marking of definiteness in Akan ........................................................................................ 52  4.2.2  Amfo’s analysis of nʊ ...................................................................................................... 53  4.3  Tonal alternations for the morpheme nʊ ................................................................................... 56  4.3.1  Tonal realization of the pronominal nʊ .............................................................................. 56  4.3.2  Tonal realisation of the determiner nʊ .............................................................................. 58  4.4   Nʊ as a pronominal .................................................................................................................... 60  4.4.1  Akan pronouns ................................................................................................................... 60  4.4.2   Familiarity of the pronominal nʊ ....................................................................................... 64  vi   4.5  Nʊ́ as a determiner ..................................................................................................................... 68  4.5.1  Nʊ́ as a definite marker ...................................................................................................... 68  4.5.2   Definiteness: uniqueness or familiarity? ........................................................................... 68  4.5.3  Nʊ́ is obligatorily familiar .................................................................................................. 69  4.5.4  Akan definite nʊ́ is equivalent to a strong definite ........................................................... 72  4.5.4.1  Weak article definites ........................................................................................................ 72  4.5.4.2  Strong article definites ...................................................................................................... 76  4.5.4.3  Bridging ............................................................................................................................. 79  4.6   Nʊ́ as a distal demonstrative ...................................................................................................... 83  4.6.1  Demonstratives in Akan ..................................................................................................... 83  4.6.2  Dɛ́m is not necessarily familiar .......................................................................................... 84  4.6.3   Dɛ́m makes the NP salient ................................................................................................. 87  4.6.4   Conclusion .......................................................................................................................... 89  4.7  Nʊ́ as a dependent clause marker .............................................................................................. 89  4.8  Conclusion .................................................................................................................................. 91  5  Conclusion ................................................................................................................................. 92  References ......................................................................................................................................... 95    vii   List of Tables 1 Consonant inventory ................................................................................................................. 3    2 Full paradigm of Akan pronouns  ........................................................................................... 59   viii   List of Figures 1 Vowel inventory ........................................................................................................................ 2      ix   List of Abbreviations Acc Accusative Adj Adjective Ak Akuapem Ani Animate ATR Advanced Tongue Root As Asante CH Choice Function CM Clause Marker Comp Complementizer CP Complementizer Phrase Dat Dative DCM Dependent Clause Marker Dem Demonstrative DP Determiner Phrase e Empty Category f Function Fa Fante Fam Familiar Marker Foc Focus Marker Fut Future Marker H High Tone x   Inani Inanimate L Low Tone N Noun Num Numeral Obj Object Past Past Tense Marker Perf Perfective Marker Pl Plural PP Prepositional Phrase Poss Possessive Marker Prog Progressive Aspect Marker QP Question Particle Ref Referential Marker Rel Relative Clause Marker Sg Singular Spec Specificity Marker Subj Subject SVO Subject-Verb-Object Order VP Verb Phrase xi   Acknowledgements I am thankful to the Almighty God for his grace and favour upon me, for his strength and guidance to see me through and for the success he has granted me. My endless gratefulness goes to my thesis committee: Lisa Matthewson (chair), Rose-Marie Déchaine and Michael Rochemont. Lisa offered enormous encouragements and her guidance was invaluable. She directed the path the thesis should go, suggested many avenues to explore, spent ample time with me to see to it that the thesis becomes a success. I will forever be indebted to you Lisa. Rose-Marie, with her background in African languages offered suggestions that helped me to see the internal characteristics of the language. Michael also pointed out hidden assumptions and facts that needed attention. In all your unmeasurable contribution to this success will never be forgotten. I feel beholden to my consultants: Dr Emanuel Sackey, Mrs. Vida Sackey and Ms Solace Sam-Brew for the time they spared for my sake. Àsɪ̀dá á̀ ónyí kàbɪ̀á ńká hɔ̀n. I am thankful again to Lisa for her semantics class where the topic of the thesis was developed. I am also grateful to Rose-Marie and Martina for their interface syntax class which helped me to develop the pronoun aspect of the thesis. Special thanks go to all professors, staff and students I have had an encounter with here at UBC- Linguistics-It has really been an interesting experience been here. I will particularly thank Prof Pulleyblank and Maria Amelia Reis Silva (student) for their encouragements when the going got tough in my first year. I also would like to the Cross-Linguistic Pragmatics grant for supporting this research. My appreciation also goes to faculty members of the Department of Linguistics at the University of Ghana-Legon for introducing me to Linguistics and particularly to Mr. Clement Appah for all his help and advice and also for suggesting UBC to me. xii   Finally, words cannot express the depth of gratitude towards my family: my parents Mr. and Mrs Oremus Arko, my siblings; Michael, Anastacia, Judith and Isaac, Sylvester and Alice, Anthony, and Solomon for being there for me all the time. May the Lord God richly bless you all. xiii   Dedication Dedication To OREMUS 1   1 Introduction There has been a long-standing debate on the interpretation of definite and indefinite noun phrases. Some researchers argue that definites encode unique identifiablity (Abbott 1999 among many others), while another school of thought says that they encode familiarity (Heim 1982 among many others). Indefinites, on the other hand, have been said to either encode a quantificational reading or they are ambiguous between quantificational and referential (specific) readings (Fodor and Sag1982), or they introduce open sentences (Heim 1982), or are choice functions (Reinhart 1997 and Kratzer 1998 among others). This research adds to the discussion. It seeks to investigate the semantics of two post-nominal markers in Akan, namely bi and nʊ. I will argue that bi encodes specificity in the sense of Fodor and Sag and Kratzer, and nʊ encodes familiarity and is very similar to the German strong articles as analyzed by Schwarz (2010). 1.1 Akan The Akan language belongs to the Kwa sub-family of the Niger-Congo language family. It is spoken mainly in the southern part of Ghana and part of Ivory Coast. In Ghana about 44% of the population speaks Akan as a native language. This group comprises about 7,753,830 people (Osam 2004). The number of non-native speakers is unknown since it is mostly the medium of exchange for people who cannot communicate in the same language in the south and no research on that has been conducted. Akan has three main dialects and many sub-dialects. The three main dialects are: Akuapem Twi (Ak), spoken mainly in the Eastern Region of Ghana, Asante-Twi (As), spoken mainly in the Ashanti 2   Region, and Fante (Fa), spoken in the Central Region and some parts of the Western Region of Ghana. The dialect for the current research is Fante1. 1.1.1 Sound inventory for Akan Akan has nine vowels and numerous consonants. The chart in figure (1) shows the vowel inventory of the language. Fig.1 Vowel Inventory Front Back (Rounded) i u High ɪ ʊ e o ɛ ɔ a Low With the exception of the low vowel, the co-occurrence of the vowels must obey the ATR2 and rounding3 rules which state that different vowels occurring in a word must be of the same ATR and rounding features. Akan is a tonal language; it has two basic tones: high (H) marked with ( ́ ) and low (L) marked with ( ̀). The tone bearing unit in Akan is the syllable, therefore every syllable must bear a tone (Dolphyne 1988). Vowels can be distinguished based on the tones they bear. Minimal pairs are given in (1).                                                              1 The dialects are mutually intelligible with slight variation. Where necessary any variation will be highlighted. 2 ATR deals with whether the vowel is produced with an advanced tongue root or retracted tongue root. 3 While the ATR rule applies to all the dialects, the rounding rule applies only to verbs in the Fante dialect (the dialect for the research). 3   1. Low Tone High Tone /dà/ ‘never’ /dá/ ‘day’ /pàpà ‘fan’ /pápá/ ‘good’ /dàn/ ‘turn it over’ /dán/ ‘room/house’ /fìr/ ‘buy on credit’ /fír/ ‘went out’ /ɔ̀bɔ̀fʊ́/ ‘hunter /ɔ̀bɔ́fʊ́/ ‘creator’ Akan has downstepping tone melody where in a sequence of H-L-H, the pitch of the second high gets lowered because of the preceding low4 (Dolphyne 1988). Table 1 shows the sounds making up the consonant inventory of the language5. Table 1. Consonant Inventory Bilabial Labiodental Alveolar Postalveolar Palatal Velar Glottal Plosive p b - t d - - k; g; kʷ; gʷ - Nasal m n - ɲ; ɲʷ; ɲɥ ŋ; ŋʷ - Trill/ Lateral - ʀ (Ak) l; r - - - - Fricative - f s; sʷ; sɥ; sj ɕ;ɕɥ - - h; hʷ Affricate - - ts dz - tɕ; dʑ; tɕɥ; dʑɥ - - Approximant w - - - j; ɥ ɰ - (Dolphyne 1988:29).                                                              4 This feature of tones will not be explored further here as it has no consequence for the current research. 5 For simplicity’s sake, the orthography of the consonants (not the phonetic alphabet) will be used in this research, however the vowels will represent what is in the vowel chart. 4   1.1.2 Plural marking in Akan Akan overtly marks plurality on both nouns and adjectives and sometimes on the verb. On the verb reduplication is used to show intensive or multiple occurrences of the activity or the people involved in it. 2. Singular Gloss Plural Gloss Noun àbʊ̀frá ‘child’ m̀bʊ̀frá ‘children’ Adjective kàkrábá ‘little’ ǹkàkrámbá ‘little ones’ Verb kàsà ‘speak’ kásàkàsà ‘talking continuously’ 1.1.3 Word order in Akan The word order in Akan is SVO; modifiers can be interspersed. 3. a) Kwèsí hú-ù Ámá Kwesi see-Past Ama ‘Kwesi saw Ama b) Kwèsí bɔ́-ɔ̀ àbʊ̀frá kàkrábá ńdɪ́dà Kwesi beat-Past child little yesterday ‘Kwesi beat a little child yesterday’ 1.2 Methodology The data presented in this research are derived primarily from the author’s own judgments as a native speaker of the Fante dialect of Akan. However, other native speakers were consulted for their judgments. Over a period of one year, three native speakers of Akan (Two Fantes and one Ashanti) 5   were consulted from time to time as the research developed for their judgments on the grammaticality and felicity of sentences. In other cases the consultants were asked for judgments which dealt with the potential ambiguity of sentences. In those instances different contexts would be presented to the consultants to find out which readings are permitted or preferred in a particular context. Real life situations were also used some of the time. An example is that a sentence would be uttered out of the blue with an item that is infelicitous in an out blue of the context to see the response and reaction of the consultant. All sessions were conducted in Akan. 1.3 Organization of the thesis The thesis is organized as follows. Chapter 2 deals with the internal structure of the DP in Akan. Here I will present the phrase structures of the items that occur within the nominal domain and show the co- occurrence restrictions between these noun-phrase internal elements. Chapter 3 discusses the semantics of the morpheme bi, I will argue that the morpheme encodes specificity. I will show that bi- phrases are used when the speaker intends a specific referent, that bi-phrases in most cases are unable to take narrow scope, the full range of data are captured best by the choice function analysis of Kratzer (1998), which is in the spirit of Fodor and Sag (1982). I will also briefly mention the pronominal use of bi here. Chapter 4 discusses the semantics of the morpheme nʊ, where I show that the morpheme encodes familiarity. I will provide evidence that nʊ, when used as a determiner, requires that the hearer be familiar with the referent of the noun phrase, and not merely that there be a unique referent as argued by Amfo (2006 etc). I will argue that the Akan determiner nʊ is most similar to the German strong articles of Schwarz (2010). Again, I will show that nʊ has a pronominal counterpart which is distinguishable by its different tonal behaviour but which nevertheless shares a core familiar semantics with the determiner use. Chapter 5 then concludes the thesis. 6   2 Internal structure of the Akan DP 2.1 Introduction This chapter discusses the phrase structures in which the morphemes bi and nʊ in Akan occur. In this chapter I am going to propose a structure which fits with the basic behaviour of the elements in the Akan nominal domain. I will also discuss their co-occurrence restrictions. The chapter is organized as follows. Section 2.2 shows the basic phrase structure in the language. Section 2.3 shows the structure of the items that occur in the nominal domain and the restrictions on their co-occurrence, and section 2.4 concludes the chapter. 2.2 The basic phrase structure in the language For the phrase structure of the elements in the nominal domain in Akan I present surface representations of the surface-observable materials without investigating such questions as whether movement has taken place to achieve the surface word order. Akan has a split system in its phrase structure in that both head-initial and head-final orders exist in the language. In the following I show examples with the relevant phrase in italics and the heads highlighted. Phrases with head-initial order: 1. VP: Kòfí hú-ù Àmá Kofi see-Past Ama ‘Kofi saw Ama’ 2. CP: Kòfí ká-à dɛ́ɛ̀ ò-hú-ù Àmá Kofi say-Past Comp 3sg-subj-see-Past Ama ‘Kofi said that he saw Ama’ 7   Phrases with head-final order: 3. PP: Kòfí dà m̀pá dʊ̀ Kofi sleep bed on ‘Kofi is sleeping on a bed’ 4. AdjP: Mù-hú-ù bɔ̀dɔ́m tùntúm 1sg-subj-see-Past dog black ‘I saw a black dog’ As the data shows, the VP and CP have their heads in initial position while the PP and AdjP have their heads in final position. This indicates a split system on the surface level in the language. Below I show the various nominal markers and their distribution. In the structures (with exception of 5), (i) shows the obligatory items and (ii) also shows the items with optional ones. (The optional items are in parentheses). 2.3 The phrase structure for the nominal items in Akan In (5), (i) shows the surface observable materials and (ii) is an abstract analysis. 5) Kwámɪ̀ hwɪ́-ɪ̀ àbʊ̀frá/m̀-bʊ̀frá/*bʊ̀frám̀ Kwame cane-Past child/Pl-child ‘Kwame caned a child’ i) àbʊ̀frá ii) àbʊ̀frá iii) m̀bʊ̀frá NP NP NP N Num [-pl] N Num [+pl] N àbʊ̀frá Ø àbʊ̀frá m̀ àbʊ̀frá 8   For adjectival phrases, the adjective occurs after the noun, and it has to agree with the plurality of the noun. 6. a) Kwámɪ̀ hwɪ́-ɪ̀ àbʊ́frá tùntúm/m̀bʊ̀frá ètùntúm Kwame cane-Past child black/child-Pl Pl-black ‘Kwame caned a dark (in complexion) child/dark children’ b) * Kwámɪ̀ hwɪ́-ɪ̀ tùntúm àbʊ́frá/ètùntúm m̀bʊ̀frá Kwame cane-Past black child/Pl-dark Pl child i) AdjP ii) AdjP NP Adj NP Adj[+pl] Num N tùntúm Num N ètùntúm à bʊ̀frá m̀ bʊ̀frá The head of the referential phrase (a determiner phrase) is in the final position; it cannot precede the adjectival phrase. 7. a) Kwámɪ̀ hwɪ́-ɪ̀ àbʊ̀frá (tùntúm) bí Kwame cane-Past child black Ref ‘Kwame caned a certain (dark) child’ b) * Kwámɪ̀ hwɪ́-ɪ̀ bí àbʊ̀frá (tùntúm) Kwame cane-Past Ref child black i) RefP ii) RefP NP Ref AdjP Ref Num N bí NP (Adj) bí à bʊ̀frá Num N tùntúm à bʊ̀frá 9   The Familiar Phrase (a determiner phrase) has its head in the final position; it becomes bad if the familiar head appears before any other item as in (8b). 8. a) Kwámɪ̀ hwɪ́-ɪ̀ àbʊ̀frá (tùntúm) nʊ́ Kwame cane-Past child black Fam6 ‘Kwame caned the (dark) child’ b) *Kwámɪ̀ hwɪ́-ɪ̀ àbʊ̀frá nʊ́ (tùntúm) Kwame cane-Past child Fam black i) FamP ii) FamP7 N P Fam AdjP Fam Num N nʊ́ NP (Adj) nʊ́ à bʊ̀frá Num N tùntúm à bʊ̀frá The demonstrative marker occurs phrase initially, and cannot occur in phrase final position. 9. a) Kwámɪ̀ hwɪ́-ɪ̀ dɛ́m àbʊ̀frá (tùntúm) nʊ́ Kwame cane-Past Dem child black Fam ‘Kwame caned that (certain) dark child’ b) #8Kwámɪ̀ hwɪ́-ɪ̀ àbʊ̀frá (tùntúm) nʊ́ dɛ́m Kwame cane-Past child black Fam Dem                                                              6 ‘Fam’ means familiarity. I define familiarity along the lines of Prince (1988) where both discourse old and hearer old are familiar. See chapter 4 for discussion. 7In one special case (which I will not be addressing in the thesis), Fam and Ref can co-occur. This is where a previously mentioned new entity is being echoed (Amfo 2009). Fam in this case is phrase-final. 10   i) DemP ii) DemP Dem FamP Dem FamP dɛ́m NP Fam dɛ́m AdjP Fam Num N nʊ́ NP (Adj) nʊ́ à bʊ̀frá Num N tùntúm à bʊ̀frá The specific marker has its position at final edge of the phrase; it cannot occur elsewhere. 10. a) Kwámɪ̀ hwɪ́-ɪ̀ (dɛ́m) m̀bʊ̀frá (ètùntúm) nʊ́ bì9 Kwame cane-Past Dem Pl-child Pl-black Fam Spec ‘Kwame caned some of (those) (dark) children’ b) * Kwámɪ̀ hwɪ́-ɪ̀ bì m̀bʊ̀frá nʊ́ Kwame cane-Past Spec Pl-child Fam i) NP ii) NP FamP N DemP N NP Fam bì (Dem) FamP bì Num N nʊ́ dɛ́m AdjP Fam m̀ bʊ̀frá NP (Adj[+p]) nʊ́ Num N ètùntúm m̀ bʊ̀frá                                                                                                                                                                                                 8 # means either that the sentence is infelicitous, or that it receives a different interpretation from the one intended. 9 Bi here is a different instance of the Ref marker. This will be discussed in chapter 3. 11   2.4 Conclusion This chapter shows the internal structure of the DP in Akan with only the surface representations. It shows that on the surface level the phrase structure in Akan has a split system with some phrases being head initial and others head final. In the next chapter, I discuss the semantics of the morpheme bi. I will argue that bi encodes specificity. 12   3 Semantics of the Akan bi This chapter discusses the semantics of the Akan morpheme bi. Bi generally does not receive a tone. This is because it displays varying tonal properties depending on its function and position. Bi appears in a couple of different syntactic environments (which correlate with certain tonal differences), and has a corresponding set of different (but related) semantic functions. It can appear: a) As an indefinite pronominal:- 1) Ésí má-à Kòfí bì Esi give-Past Kofi some ‘Esi gave Kofi some’ b) Within a DP as a determiner following the NP:- 2) Pàpá bí bá-à há Man Ref come-Past here ‘A (certain) man came here’ c) Inside bíárá ‘every/any’, óbí ‘somebody/someone’:- 3) Àbʊ̀frá bíárá kɔ́-ɔ̀ fíé Child every go-Past home ‘Every child went home’ 4) Òbí frɛ́-ɛ̀ Kwámɪ̀ Someone call-past Kwame ‘Someone called Kwame’ 13   In this chapter I will set aside the use of bi in bíárá and óbí and discuss its determiner and pronominal uses. One of the environments bi appears is inside a DP following the NP. In this environment, it marks indefiniteness, and contrasts with bare nouns and noun phrases marked with the determiner nʊ́. I will argue that bi is a specific indefinite marker and that it conforms to an analysis along the lines of Kratzer (1998). I will also show that bi has a related use as specific indefinite pronoun. The chapter is organized as follows: Section 3.1 gives a brief review of the literature on indefiniteness; section 3.2 introduces the two mechanisms by which indefiniteness in Akan is marked. Section 3.3 shows the two environments in which bi occurs and introduces Amfo's (2009) analysis of the two uses. Section 3.4 discusses the tonal alternations for the two uses. Section 3.5 discusses the specificity of the pronominal and the determiner uses of the morpheme. Section 3.6 concludes the chapter. 3.1 Literature review Indefiniteness has been discussed extensively in the literature, largely in Indo-European languages. Different authors have described indefiniteness in different ways; see for example Heim (1982), Fodor and Sag (1982), Ward and Prince (1991), Diesing (1992), Abusch (1994), Haspelmath (1997), Farkas (1997, 2002), von Fintel (1998), Matthewson (1999) as well as Ionin (2009) to mention but a few. I adopt Heim’s (1982) novelty-based approach to indefiniteness which states that felicitous indefinites must introduce “novel” variables while felicitous definites must introduce “familiar” variables. In other words, the referent denoted by a definite NP must be familiar to the participants in the discourse while the referent introduced by an indefinite NP is a novel one. The following discourse in (5) illustrates this in Akan. (Context: A child comes home upset and crying so the mother inquires about what the problem could be. The child then utters either of the following the sentences) 14   5. a) Kyɪ̀rɛ̀kyɪŕɛ̀nyí á-hwɪ̀ mɪ̀ Teacher Perf-cane 1sg.obj ‘A teacher has caned me’ b) Kyɪ̀rɛ̀kyɪŕɛ̀nyí bi ́ á-hwɪ̀ mɪ̀ Teacher Ref Perf-cane 1sg.obj ‘A (certain) teacher has caned me’ In sentence (5) the noun phrases (kyɪ̀rɛ̀kyɪ́rɛ̀nyí in (a) and kyɪ̀rɛ̀kyɪ́rɛ̀nyí bí in (b)) introduce novel discourse referents. However if the mother continues and asks the child why s/he was caned, the child may add for instance the following in (6a) and not (6b): 6) a) Mɪ̀-sɪ́rɪ́ɪ̀w Kòdwó nà kyɪ̀rɛ̀kyɪŕɛ̀nyí nʊ́ hwɪ́-ɪ̀ mɪ́ 1sg.subj-laugh.Past Kojo and teacher Fam cane-Past 1sg.obj ‘I teased Kojo and the teacher beat me’ b) #Mɪ̀-sɪ́rɪ́ɪ̀w Kòdwó nà kyɪ̀rɛ̀kyɪŕɛ̀nyí/ kyɪr̀ɛ̀kyɪŕɛ̀nyí bí hwɪ́-ɪ̀ mɪ́ 1sg.subj-laugh.Past Kojo and teacher/teacher Ref cane-Past 1sg.obj ‘I teased Kojo and a/a (certain) teacher beat me’ In (5) the teacher is mentioned for the first time in the discourse, therefore the noun phrases kyɪ̀rɛ̀kyɪ́rɛ̀nyí (a) and kyɪ̀rɛ̀kyɪ́rɛ̀nyí bí (b) mark indefiniteness; however in (6a) the already introduced teacher is mentioned for the second time. Thus it is a familiar discourse referent and is marked as definite. This is to say that when an entity is newly introduced into the discourse (and is not already known to the hearer), it is encoded either with a bare noun or with the determiner bí following the noun. However if that entity is already mentioned in the discourse or the hearer can retrieve from her mental faculty a familiar referent, the determiner nʊ́ follows the noun. This makes the sentence (6b) - either with the bare noun or with bi- infelicitous in the above discourse. In the following sub-section I will go into more detail illustrating the two ways of marking indefiniteness in Akan. 15   3.2 The two-way mechanism of marking indefiniteness in Akan As was shown in (5), Akan has two distinct ways of marking indefiniteness: (a) By the use of the bare noun OR (b) by having the morpheme bi occurring after the nominal. Both the bare noun and the bí DP can occur in both subject and object positions: 7) a) Kyɪ̀rɛ̀kyɪŕɛ̀nyí hwɪ́-ɪ̀ Kòfí Teacher cane-Past Kofi ‘A teacher caned Kofi’ b) Kyɪ̀rɛ̀kyɪŕɛ̀nyí bí hwɪ́-ɪ̀ Kòfí Teacher Ref cane-Past Kofi ‘A (certain) teacher caned Kofi’ 8) a) Kyɪ̀rɛ̀kyɪ́rɛ̀nyí Bã̀ã́ hwɪ́-ɪ̀ àbʊ̀frá Teacher Baah cane-Past child ‘Teacher Baah caned a child’ b) Kyɪ̀rɛ̀kyɪ́rɛ̀nyí Bã̀ã́ hwɪ́-ɪ̀ àbʊ̀frá bí Teacher Baah cane-Past child Ref ‘Teacher Baah caned a (certain) child’ The context here will be where for instance, a child comes home from school upset and the mother asks what the problem is. The child’s response could be (7) where Kofi is perhaps his friend or someone he likes, or (8) where he likes teacher Baah and would prefer he never caned anyone because caning a child in that school for instance leads one into trouble. The difference between the above sentences in (7) and (8) is that in the (a’s) the child’s concern is not who the kyɪ̀rɛ̀kyɪ́rɛ̀nyí (7) or the àbʊ̀frá (8) is. However in the (b’s) the child has a particular kyɪ̀rɛ̀kyɪ́rɛ̀nyí (7) or a particular àbʊ̀frá (8) in mind. This goes to say that, whereas the child is making 16   an existential assertion that an individual in the set denoted by the NP is not empty in the (a’s), he has a particular referent in mind in the (b’s).Therefore in Akan bare nouns are used to encode a quantificational assertion while nouns followed by the determiner bí are used when the speaker has a specific referent in mind. The use of both a bare noun strategy and an overt marking strategy in marking indefiniteness is not restricted to Akan. A similar analysis is given by Ajíbóyè (2007) for Yoruba, another Niger-Congo language spoken in Nigeria. Ajíbóyè (2007) argues that in Yoruba, indefiniteness can be marked by (a) Bare nouns in the elsewhere case-where generic readings are unavailable (b) A post-nominal determiner kan attached to the noun The bare noun indefinites are interpreted as existential, while the kan DP indefinites are interpreted as specific. This is illustrated in the following sentences in (9): 9. a) Tàkúté Olú pa ajá Trap Olu kill dog ‘Olu’s trap caught a dog’ (Adapted from Ajíbóyè 2007:164). b) Tàkúté Olú pa ajá kan Trap Olu kill dog certain ‘Olu’s trap caught a certain dog’ (Adapted from Ajíbóyè 2007:164). Ajíbóyè (2007) argues that in appropriate contexts where a generic reading (which is obtained by the presence of the imperfective máa-ń which introduces a genitive operator (GEN))10 is not available, out of the blue bare nouns are interpreted as existential indefinites.                                                              10 I will not go further with this since the point being made here is that both the bare noun and the overt marker strategies of marking indefiniteness exist in Yoruba. 17   In this chapter, I am going to concentrate on the morpheme bi (and leave the bear noun for future research). I will show that although bi appears in two different syntactic contexts - either as a pronominal or as a determiner - in both of these, it conveys a similar semantics, namely specific indefiniteness. 3.3 The two environments of bi In this sub-section I lay out the two distinct environments in which the morpheme bi appears and introduce Amfo's (2009) analysis of the two uses. The double uses of bi in Akan were first documented by Amfo (2009). Amfo analyzes bi as an existential quantifier which functions as a determiner when it bears a high tone and a pronominal when it bears a low tone. 10. a) Ámá á-tɔ̀ èdzìbán bí Ama Perf-buy food Ref ‘Ama has bought a (certain) food’ b) Máá-só mɔ́-tɔ́ bì 1sg-too 1sg.subj-buy some’ ‘I will also buy some (of the food)’ Bì (the low tone version) occurs independently of a head noun and is restricted to object position as in (10b), making its occurrence in subject position bad (11b). Meanwhile bí (the high tone version) occurs post nominally in both subject (11a) and object positions (10a). 11. a) Bàsíá bí rɪ̀-fr̀ɛ́ Ésí Woman Ref Prog-call Esi ‘A (certain) woman is calling Esi’ b) *Bì rɪ̀-fr̀ɛ́ Ésí Some Prog-call Esi 18   In her paper Amfo states that the morpheme bi- irrespective of the tonal realization it gets, or the syntactic environment in which it occurs- fits the description of an existential quantifier. Her argument for this analysis relies on the claim that bi binds variables belonging to a conceptual domain restricted by a nominal which occurs to its left. Amfo furthermore argues that bi as an existential quantifier has the cognitive status of referential (Amfo 2009:2). However, it is difficult to understand what is meant by a referential existential quantifier, since quantifiers are usually taken to be non-referential (see e.g. Heim and Krazter 1998, among many others). Amfo states that the denotation of the pronominal bì depends on a linguistic stimulus in the preceding discourse which serves as an antecedent. However she states that the relationship between pronominal bì and its antecedent is a weak one. This is because the pronominal bì does not normally represent the same denotation as its antecedent. Bì always refers to a part, or the kind of, the entity the antecedent represents. Hence bì denotes a part, the same type or a sufficiently similar kind as the nominal predicate. Meanwhile, Amfo argues, the determiner bí introduces new referents into the discourse. When N-bí is used, the referent is an instantiation of a type of object, so the intention is for the hearer to focus on the properties of the head noun in question. I agree with Amfo's claim that bi appears in two environments, and that the determiner use marks an NP as referential and the pronominal use denotes a part or a kind of its antecedent. My analysis of bi will capture a similar intuition to those of Amfo’s. However Amfo’s analysis of the morpheme does not accord with the distinction between existential quantification and referentiality in the literature. It also does not capture the scope interaction between the morpheme bi and other scope taking operators, and the kinds of interpretation the morpheme gets. In this chapter, I am going to argue that the morpheme bi in Akan marks specificity. When it is used as a pronoun, it gets its specific interpretation via an anaphoric relationship with a previously mentioned 19   NP predicate (along the lines of Enç1991). However when it is used as a determiner, it is interpreted as referential along the lines of Fodor and Sag (1982) or more precisely, Kratzer's (1998) choice function implementation of Fodor and Sag's ambiguity hypothesis. 3.4 Tonal alternations of the morpheme bi In this section I am going to discuss the tonal alternations of the functions of bi as a pronominal and a determiner. Amfo (2009) argues that the determiner bi is marked with a high tone (a), while the pronominal bi is marked with a low tone, (b): 12. a) Mààmɪ́ Ésí á-tɔ̀ àtár bí Woman Esi Perf-buy dress Ref ‘Madam Esi has bought a (certain) dress’ b) Mɔ́-kɔ́ màà-kɔ́-tɔ́ bì 1 sg.subj-go 1 sg.subj-go-buy some ‘I will go and buy a similar one’ I will argue, following Amfo (2009), that the morpheme in its determiner use is marked with a high tone. However, I will argue that the morpheme in its pronominal use is actually toneless, contrary to Amfo’s claim that it is marked with a low tone. 20   3.4.1 Tonal realization of the indefinite pronominal bi I will argue for the toneless11 status of pronominal bi on the basis of the fact that the tone it surfaces with depends on the sentence type it occurs in. This follows since an underlyingly toneless element may not surface without a tone, so it gets its tone from the surrounding environment. In final position of sentences containing no A-bar movement, the pronoun gets a low tone irrespective of the tone preceding it: 13. a) Ésí tɔ̀n dɔ̀kʊ́n Esi sell kenkey ‘Esi sells kenkey’ b) Ńtsí mɔ́-tɔ́ bì So 1sg.subj.Fut-buy some ‘So I will buy some (of the kenkey)’ c) Ńtsí mɪ̀-tɔ́-ɔ̀ bì So 1sg.subj-buy-Past some ‘So I bought some (of the kenkey)’ However when the pronominal bi is not sentence final it has a high tone, irrespective of the preceding tone: 14. a) Ésí tɔ̀n dɔ̀kʊ́n Esi sell kekey ‘Esi sells kenkey’                                                              11 I assume toneless here to mean that the item has no lexical tone, therefore its tone is determined by the environment in which it occurs. See Déchaine (2001a, b) for discussion. 21   b) Ńtsí mɔ́-tɔ́ bí ámà Kwámɪ̀ So 1sg.subj.Fut-buy some give Kwame ‘So I will buy some for Kwame’ c) Ńtsí mɪ̀-tɔ́-ɔ̀ bí má-à Kwámɪ́ So 1sg.subj-buy-Past some give Kwame ‘So I bought some for Kwame’ In sentences containing an A-bar configuration, for example sentences containing a focus construction, the morpheme gets a high tone whether it is sentence final or not: 15. a) Ɔ́nʊ́ nà mɪ́ má-à nʊ̀ bí 3sg Foc 1sg.subj give-Past 3sg.obj some ‘It was him/her that I gave some to’ b) Ɔ́nʊ́ nà mɪ́-dzɪ ́ bí má-à nʊ́ 3sg Foc 1sg.subj-take some give-Past 3sg.obj ‘It was him/her that I gave some to (him/her)’ The above sentences in (13-15) support my claim that the pronominal bi is toneless, as the position it occurs in determines the type of tone it will have. The argument that it is toneless precisely rests on its variable tonal behavior, dependent on context. We only expect this of things which are toneless. The same thing cannot be said about the determiner bí. 3.4.2 Tonal realization of the indefinite determiner bí The morpheme bi, when it is used as a determiner, has a high tone. This is irrespective of the position it occupies or the type of sentence it occurs in. In the following sentences in (16-18) the determiner bí 22   has a high tone irrespective of the preceding tone (16), the position it occupies in the sentence (17) or the kind of sentence it appears in (18): 16. a) Pàpà bí dà hɔ́ Fan Ref sleep there ‘There is a (certain) fan over there’ b) Pàpá bí dà hɔ́ Man Ref sleep there ‘There is a (certain) man sleeping over there’ 17. a) Ésì tɔ̀n èdzìbán bí Esi sell food Ref ‘Esi sells a (certain) food’ b) Ésí tɔ̀n èdzìbán bí wɔ́ èkyír hɔ́ Esi sell food Ref at back there ‘Esi sells a (certain) food over there at the back’ 18) Émí nà mɪ́-wɪ́ǹ pàpà bí 1sg Foc 1sg.subj-weave-Past fun Ref ‘It was I who wove a (certain) fan’ The determiner bí always gets a high tone, thus it is lexically determined. In this section, I have shown that while the pronominal bi is toneless the determiner bí is always high. Further research is needed into the processes which derive the tonal realizations of the pronominal bi; such tonal research goes beyond the bounds of the current thesis. 23   3.5 The specificity of bi Indefinites are said to be specific when the speaker has a particular referent in mind at the time of the utterance or when the referent is linked to an NP in the preceding sentence. In this section, I am going to argue that the Akan morpheme bi is used to mark specificity. I am going to argue that the pronominal bi is interpreted as specific through an anaphoric relationship with a previously mentioned referent and when it is used as a determiner it is interpreted as referential. 3.5.1 Bi as a specific indefinite pronominal marker Bi as a specific indefinite pronominal marker in Akan is used to mark partitivity (this is because it may mark part of an entity mentioned previously in the discourse) as well as allowing an interpretation where it introduces an entity which is similar to or the same kind of entity mentioned previously in discourse. In the following sentences in (19) the bì in (b) signifies part of a whole, which in this case is èdzìbáǹ ‘food’. On the other hand, the bì in (20b) represents a similar/same kind of entity to that mentioned in (20a) which is búùkúù ‘book’. 19. a) Ésí à-yὲ èwìmbír èdzìbáǹ Esi Perf-do evening food ‘Esi has prepared dinner’ b) Mé-dzí bì 1sg.subj.Fut-eat some ‘I will eat some’ 20. a) Ámá tɔ́-ɔ̀ búùkùú bí Ama buy-Past book Ref ‘Ama bought a (certain) book’ 24   b) Mʊ̀-bɔ́-tɔ́ bì 1sg.subj-Fut-buy some ‘I will buy a similar one’ Indefinite pronouns are said to be pronouns that do not make the exact identity of their referent. The Akan indefinite pronoun as stated above does not get the exact denotation of its antecedent. This makes it different from a normal definite pronoun which gets the exact denotation of its antecedent. This function of bi is similar to the functions of the English indefinite pronouns one and some. While one is used to mark count nouns, some is used to mark mass nouns. 21. a) Mary bought a dress. b) I bought one too. 22. a) Mary bought food. b) I bought some too. Like in Akan, the one/some construction does not involve complete co-reference with the antecedent; in (21b) I didn't buy the same dress-token that Mary did; rather, I bought one of the same kind of entity. It is infelicitous to use some to mark count nouns or to use one for mass nouns. This is illustrated in (23): 23. a) Mary bought a dress. b) #I bought some too. 24. a) Mary bought food b) #I bought one too. 25   This count/mass distinction is not evident in Akan, as bi is used to mark both count and mass nouns. In the following sentences in (25) and (26), bi is used to mark both count nouns (25) and mass nouns (26). 25. a) Ésí tɔ́-ɔ̀ àsʊ̀màdzɪ́ Esi buy-Past earring ‘Esi bought earrings’ b) Máá-sʊ́ mɪ̀-tɔ́-ɔ̀ bì 1sg-too 1sg.subj-buy-Past some ‘I bought one too’ 26. a) Ésí kɔ̀-bú-ù ǹsú Esi go-break-Past water ‘Esi went to fetch water’ b) Máásʊ́ mɪ̀-kɔ̀-bú-ù bì 1sg.too 1sg.subj-go-break -Past some ‘I went to fetch some too’ As seen in the sentences in (25) and (26), the Akan bi does not distinguish between count and mass nouns. This amounts to saying that English one and some are subsets of the Akan bi. In the following sub-section, I am going to compare bi with the anaphoric one/some and show that in most cases bi functions in the same way as the anaphoric one/some. 3.5.1.1 Bi and one/some-anaphora Luperfoy (1991) (as cited in Dale 2003) describes the three discourse functions of one- anaphora as: (a) contrasting two sets of individuals, (b) denoting a representative sample of a set introduced by the 26   antecedent and (c) referring to a specimen of a type introduced by the antecedent. These three functions are illustrated in (27-29) respectively. 27. a) Steve has a gray Bentley b) Frank has a red one 28 a) Rose has a lot of cell phones b) I am going to ask her for (the Nokia) one 29. a) I saw John’s bike b) I think I want one (Adapted from Dale 2003:2) Mann and Thompson (1987) term the first discourse function as ‘one of ‘CONTRAST’ and the second and third as ‘instances of the ELABORATION relation’ (Dale 2003). Where English has contrastive one-anaphora, in Akan the bare noun (or just the modifier) is used and not one or bi. 30. a) Kwèsí wɔ̀ m̀pɔ̀bʊ̀wá tùmtúm Kwesi at shoes black ‘Kwesi has black shoes’ b) Yàw wɔ̀ fítáá Yaw at white ‘Yaw has white (ones)’ 31. a) Kwàsí wɔ̀ m̀pɔ̀bʊ̀wá tùmtúm Kwasi at shoes black Kwasi has black shoes’ 27   b) #Yàw wɔ̀ fítáá kʊ́r Yaw at white one ≠ ‘Yaw has white ones’ = ‘Yaw has a half pair of white shoes’ c) #Yàw wɔ̀ bì Yaw at some ‘Yaw has some’ Where in English one is used for the second and third use (elaboration function), in Akan bi and not a bare noun nor one is used. Bi can be used to denote a representative of the type introduced by the antecedent (the second use) and in functioning as such it marks partitivity (32). It can also denote a specimen of the type introduced by the antecedent (the third use), and in doing so it marks a similar or same kind of the entity denoted by the antecedent (33). This use of the pronominal bi is similar to the elaborative use of the English one-anaphora. 32. a) Ésí tɔ́-ɔ̀ èkùtú pìì Esi buy-Past orange(s) many ‘Esi bought a lot of oranges’ b) Ǹtsì mì-dzí-ì bì So 1sg.subj-eat-Past some ‘So I ate some’ 33. a) Ésí nɪ́ tsír nʊ́ yɛ̀ fɛ̀w Esi Poss head Fam do beautiful ‘Esi’s hairdo is beautiful’ 28   b) Ǹtsì mɔ́-bɔ́ bì So 1sg.subj-beat some ‘So I will do same (hairdo)’ It should be noted however that in the context in (32) the number of oranges is unspecified, and may be more than one. Bi here does not necessarily denote only one of the oranges. As such, bi does not correspond exactly to English one, but covers also the meaning of several. One-anaphora as seen in (27)-(29) is used to mark count nouns; however its treatment (with the exception of the contrastive function) can be extended to mass nouns by using the indefinite pronoun some. In the following sentences in (34) and (35), the some in (34b) picks out a representative sample of the type of entity denoted by the antecedent different dishes, while the some in (35b) introduces a new specimen of the type of entity denoted by the antecedent furniture. 34. a) Sam’s prepared different dishes for dinner. b) I am going to have some. 35. a) Sam’s furniture is so old. b) I am going to get him some by the end of the month. The pronominal bi in the same vein is used to mark mass nouns, as already discussed in the preceding part of the section. 36. a) Ésí á-ỳɛ èdzìbán áhʊ́rʊ́w Esi Perf-do food different ‘Esi has prepared different/several dishes’ 29   b) Mì-rì-ké-dzì bì 1sg.subj-Porg-go-eat some ‘I am going to eat some’ 37. a) Ésì n’angua nʊ yɛ huam Esi Poss.oil Fam do smell-good ‘Esi’s oil smells good’ b) Ǹtsì mɛ́-yɛ́ bì So 1sg.subj-do some ‘So I will prepare some (of the oil)’ In this sub-section, I have shown that the pronominal bi can function in similar ways as English one/some anaphora. Next, I am going to outline the specificity status of the pronominal bi along the lines of Enç’s analysis of specificity. 3.5.1.2 Specificity in Enç’s sense Enç (1991) argues that indefinites that form a partitive with some previously mentioned entities or events are specific. Enç argues that in Turkish, indefinites in object position with the overt accusative case marker (–(y)i) are specific, those without it are non-specific. The indefinites with the overt case marker form a weak link with a previously mentioned NP. In the following discourse in (38), the indefinite object in (b) which is marked with accusative case is linked to the NP in (a), therefore is specific, while the one in (c) is non-specific because it is not marked with accusative case. The girls in (b) form part of the set of the children in (a) and by virtue of this the noun phrase ‘two girls’ is specific. 30   38. a) Odam-a birkaç çocuk girdi 1sg.Poss-room-Dat several children entered ‘Several children entered my room’ b) Iki kɨz-ɨ tanɨyordum Two girl-Acc 1sg.subj-knew ‘I knew two girls’ c) Iki kɨz tanɨyordum Two girl 1sg.subj-knew ‘I knew two girls’ (Enç 1991:6) In Akan the pronominal bi can occur independently of a head noun in a sentence (Amfo 2009) as has been stated above, however in the vast majority of cases bi has to be anaphorically linked to an NP mentioned in the preceding sentence; otherwise a deictic gesture must be made at the time of the utterance in the direction of the referent. This in Enç’s (1991) sense is specificity. In a context where a speaker utters the sentence in (39) with no mention of food previously, she must make a pointing gesture otherwise the pronominal bi will be infelicitous. 39. Mé-dzí bì 1sg.subj-eat some ‘I will eat some’ This specificity status of the pronominal bi goes against Amfo (2009)’s analysis of it as an existential quantifier. The anaphoric feature of an indefinite pronoun is very distinct from the indefinite determiner bí, in that the determiner bí cannot be used in partitive constructions. 31   40. a) Skúll m̀bʊ̀frá pìì rɪ́-bá há School children more Prog-come here ‘A lot of school children are coming here’ b) Mì-nyìm bì 1sg.subj-know some ‘I know some of them’ c) *Mì-nyìm bí 1sg.subj-know Ref ‘I know (certain) of them’ d) #Mì-nyìm skúll m̀bʊ̀frá bí 1sg.subj-know school children Ref ‘I know certain school children’ (Adapted from Enç 1991:6) In (40b) the set denoted by the pronominal bì constitutes individuals from the previous entities mentioned in the discourse, in this case ‘school children’. As noted in section 2.1 above, the determiner bí obligatorily requires a preceding noun. Thus (40c) is syntactically ruled out. The sentence in (40d) cannot have the partitive reading of (40b); it cannot mean that I know some of the previously-mentioned school children. Instead, it only picks out a disjoint set of school children. As such, the sentence would be infelicitous in this context if uttered by the same speaker who uttered (40a). Another thing worth noting about the pronominal bi is its ability to mark events. Amfo (2009) argues that the function of the pronominal bi goes beyond marking entities; bi can also be used to mark events where it signifies part of an event or a similar kind of event. In the following sentences in (41), bi in (b) signifies part of the activity in (a), which the speaker says she is going to take part in. In (42), bi in (b) denotes a similar kind of whatever action Araba might have taken in (a). 32   41. a) Árábá nà A ́má rʊ́-bɔ̀ ámpé Araba and Ama Prog-beat ampe ‘Araba and Ama are playing ampe’ b) Mʊ̀-rʊ̀-kɔ́-bɔ́ bì 1sg.subj-Prog-go-beat some ‘I am going to take part (in it)’ 42. a) Dɛ́m àdzɪ́ yí áà Árábá a-yɛ́ yí Dem thing Dem Rel Araba Perf-do Dem ‘This thing that Araba has done’ b) Mɛ́-yɛ́ nʊ̀ bì dódòw 1sg.subj-do3sg.obj some much ‘I am going to do the same to her’ I would like to state that even though the pronominal bi is linked to events here, it is the same partitive or kind reading. Bi is unspecified as to whether it performs those functions on groups of individuals/masses, or on events. So far it has been shown that the toneless bi is an indefinite pronominal. However by virtue of its anaphoric relationship with a previously mentioned entity in the discourse, it gets specificity status thereby making it a specific indefinite pronoun in the sense of Enç (1991). This leads to the following section where I discuss the semantics of bí as a determiner. As mentioned in the introduction to this chapter, the determiner bí is an indefinite determiner which is used to mark referentiality. 33   3.5.2 The determiner bí One of the most noted works on indefiniteness is Fodor and Sag (1982). They describe the indefinite article in English as being ambiguous between quantificational and referential interpretations. In this research I am going to follow the fundamental insights of Fodor and Sag’s analysis of indefinites and argue that the Akan data are best captured along the lines of their analysis. 3.5.2.1 Fodor and Sag’s sense of referentiality Fodor and Sag (1982) describe the English article a as being ambiguous between referential and quantificational interpretations. They argue that the interpretation of the noun phrase will be referential if the speaker has a particular entity that satisfies the denotation of the NP in mind. On the other hand, if the speaker just makes an assertion of the mere existence of an entity or individual satisfying the NP predicate, then the interpretation will be quantificational. This is to say that, in the following sentence (43), the interpretation of a student in the syntax class is ambiguous between referential and quantificational. 43. A student in the syntax class cheated on the final exam. The two understandings of (43) will be; one, referential- when the speaker has a particular student in mind who cheated on the exam, and the noun phrase directly refers to this individual. The other understanding will be a quantificational one where the speaker asserts that the set of students who cheated on the exams is not empty. Fodor and Sag argue for the referential reading of indefinites on the basis of the ability of specific indefinites to appear to take scope outside of islands, something which an existential quantifier should not be able to do. They outline six features to distinguish the referential and quantificational readings of indefinites. These include; one, Rich Descriptive Context of the noun where the NP is described in more detail. They argue that this tends to favour a referential 34   interpretation of the noun. Two, Topicalization and Left Dislocation; they argue that this also favours a referential reading as one particular person being talked about could be the topic of what is being said. Three, the colloquial non-demonstrative use of this also favours a referential interpretation. Four, the use of the adjectives certain and particular tend to favour the referential reading of indefinites. Five, the use of there insertion, and finally six, the use of numerals favour the quantificational reading of indefinites. These findings support the claim that indefinites can have referential as well as quantificational readings. 3.5.3 Bí as referential Bí as discussed by Amfo (2009) is an existential quantifier which has the cognitive status referential. This analysis however is problematic as an existential quantifier is in direct opposition with that which is referential. An indefinite NP is said to have a referential status when a speaker makes reference to a particular entity or referent at the time of the utterance and expects the hearer to make a mental representation of that entity or referent. I hereby claim that the determiner bí when it occurs after a noun marks the noun phrase as specific in Fodor and Sag’s sense. Hence determiner bí in Akan marks an indefinite as referential. As shown in section 1.2 above, indefinite NPs in Akan are marked using two different mechanisms. One is by the use of the bare noun and the other by having the morpheme bí following the noun. These two mechanisms have two different interpretations- quantificational and referential- which cannot be used interchangeably. I thereby argue that indefinites in Akan are not ambiguous between referential and quantificational interpretations; rather these interpretations are marked in two distinct ways. In the following discourse in (44), (a) is infelicitous when the speaker has a specific entity or referent in mind and (b) cannot be used with the intention of reaching a quantificational interpretation. (44) contrasts with the discourse in (45), in which the (a) sentence contains a bí-marked noun phrase. This is seen 35   when the speaker continues the utterance with the noteworthy condition under which the NP in (a) would qualify as referential. 44. a) Kyɪ̀rɛ̀kyɪ́rɛ̀nyí Bã̀ã́ hwɪ́-ɪ̀ àbʊ̀frá Teacher Baah cane-Past child ‘Teacher Baah caned a child’ b) #Ɔ̀-yɛ̀ bùbúá-fʊ́ 3sg.subj-be cripple-Nom ‘S/he is a cripple’ 45. a) Kyɪ̀rɛ̀kyɪ́rɛ̀nyí Bàá hwɪ́-ɪ̀ àbʊ̀frá bí Teacher Baah cane-Past child Ref ‘Teacher Baah caned a (certain) child’ b) Ɔ̀-yɛ̀ bùbúá-fʊ́ 3sg.subj-be cripple-Nom ‘S/he is a cripple’ The interpretation for (45) is that there is a particular child, a cripple who is not supposed to be caned by anyone, but teacher Kofi caned her. However in (44) there is no demand for the speaker to continue with anything worth noting about the child in question. This is because the speaker is just making an assertion about the event. It is highly unlikely for the sequence in (44) to take place. Even if it does it will be highly marked. In the following sentence in (46) the NP in object position in (b) refers to some particular food while the one in (a) does not. 36   46. a) Mààmɪ́ Ámá yɛ́-ɛ̀ èdzìbán Woman Ama do-Past food ‘Madam Ama cooked (food)’ b) Mààmɪ́ Ámá yɛ́-ɛ̀ èdzìbán bi ́ Woman Ama do-Past food Ref ‘Madam Ama cooked (some specific) food’ Here the interpretation of the object in (a) is quantificational while that of (b) is referential. The context in (a) will be where Madam Ama does not normally cook and so the speaker is reporting the fact that she did cook this time. The interpretation will be that the speaker is just reporting an activity that has taken place namely, Madam Ama’s cooking. In contrast however in (b), the context will be one where there is something special about the food. For instance the food was delicious or the speaker has never had such a meal before, so she attaches bí. It will then be expected of her to continue the utterance with the reason why the food should be specific, i.e. why the speaker used the morpheme bí as a determiner for the NP. Similarly in the following sentences, we see the referential/quantificational contrast between determiner bí and the bare noun. The context here will be where the hearer comes home from work or something like that and then the speaker utters (a)/ (b): 47. a) Bànyín bí bá-à fíé há Man Ref come-Past house here ‘A (certain) man came to this house’ b) Bànyín bá-à fíé há Man come-Past house here ‘A man came to this house’ 37   Here the interpretation for the subject bànyín bí in (a) will be that of referential. This is because the speaker has a particular/specific man in mind at the time of the utterance. It could be that there is something special12 about that man. Something that is special to either the speaker or the hearer, or something that is unusual. This sentence can be contrasted with the (b) where the speaker just asserts the existence of a man who came to the house. Here the speaker recognizes the oddity of the man coming to the house (for instance a house where girls being groomed to enter the puberty stage are kept) and is just ‘reporting’ it without necessarily having a particular man in mind. The special features of the referent of the subject NP in (47a) are what qualify it to have bí attached to it. This is to say that bí cannot be used if there is nothing special about the individual being referred to. This condition Ionin (2006) terms ‘noteworthiness’- it is that which is important and which gives more information on the referent to make it more specific. In her paper, Ionin argues that, in spoken English, this can be used as indefinite. It must however satisfy the felicity condition of noteworthiness (which can come before or after the utterance containing the referent). This is exemplified in the following sentence: 48. I found this blue apple on my plate (Ionin 2006:185) Ionin explains that in (48), the adjective blue being noteworthy makes the referent a specific apple and so the indefinite this satisfies the felicity condition of noteworthiness. Another example is given in (49): 49. a) I want to see this movie that my friends have been recommending to me for ages. b) #I want to see this movie. Ionin argues that (b) alone without the subordinate clause does not satisfy the felicity condition of noteworthiness. Therefore indefinite this is infelicitous in this environment.                                                              12 This is in line with Ionin’s (2006) description of specificity. 38   Similar sentences to those in (48) and (49) in Akan will have the determiner bí following the NPs. 50. a) Mù-hú-ù ànʊ̀má bí wɔ̀ àsɔ́rdán mù 1sg.subj-see-Past bird Ref at church-building in ‘I saw a (certain) bird inside the church building’ b) Mù-hú-ù ànʊ̀má wɔ̀ àsɔ́rdán mù 1sg.subj-see-Past bird at church-building in ‘I saw a bird inside the church building’ A sentence like (50a) will demand some kind of noteworthiness just like its English counterpart. This is because a bird being inside the church building might not be news so the bare noun will be enough as in (50b). However if the speaker does not use just the bare noun but accompanies the NP with the determiner bí, then there should be something worth noting about the bird that permits the speaker to make the utterance. Hence, the speaker should continue the utterance in (50a) with something worth mentioning, or the hearer may demand that, by asking for instance a question like the following in (51a). However if the speaker uses the bare noun the hearer’s response could be something like in (51b), not requiring anything special about the referent. 51. a) Ééh dɛ́m áà, ɪ́bɛ̀n ànʊ̀má á? Eeh really which bird ‘Really, what kind of bird?’ b) Óh dɛ́m ‘Oh really’ With a question like (51a) the hearer will expect some kind of special feature about the bird. For instance in this discourse context, the bird could be an owl which is perceived to be evil, and so its 39   presence in the church will be news. This amounts to saying that Akan bí demands that the relevant individual be noteworthy. 3.5.4 The scope of bí Referential indefinites as described in the literature are said to be scopeless. Fodor and Sag (1982) claim that the referential indefinite is forbidden to fall under the scope of any scope-bearing operator. This is illustrated in the following sentence: 52. Sandy did not see a squirrel that chased its tail around the oak tree. (Fodor and Sag 1982:359) Fodor and Sag argue that the referent a squirrel that chased its tail around the oak tree cannot fall within the scope of the negation. The interpretation will be that Sandy might have seen different squirrels, some flying, some singing and others running, but the particular one that chased its tail around the oak tree, Sandy did not see. Ionin (2006) has also argued in favor of Fodor and Sag by saying that the referential indefinite cannot fall under the scope of another operator. Enҫ (1991) also states that an indefinite NP becomes specific when it takes scope over operators like negation, modals or prepositional attitude verbs. However, Enҫ’s analysis is different because indefinite NPs that do not have a really referential analysis can also take scope over most things but if they are inside an island, they will still scope inside the island. What is different about Fodor and Sag’s proposal is that referential indefinites could appear to scope outside islands. In what follows I am going to discuss the scope of the Akan bí with respect to its interaction with the bare noun as well as other scope-bearing operators like kàkrá ‘few’ and the universal quantifier nyìná ‘all’. My hypothesis is that in all these instances, the determiner bí is not allowed to fall within the scope of these operators therefore is specific in line with Fodor and Sag, Ionin and Enҫ. 40   3.5.4.1 Bí and bare nouns Bí cannot scope under the bare noun whether in subject or object position. 53. a) Kyɪ̀rὲkyɪ́rέnyí bí hwɪ́-ɪ̀ àbʊ̀frá Teacher Ref cane-Past child ‘A (certain) teacher caned a child’ b) Kyɪ̀rὲkyɪ́rέnyí hwɪ́-ɪ̀ àbʊ̀frá bí Teacher cane-Past child Ref ‘A teacher caned a (certain) child’ The interpretation in (a) is referential, where a particular teacher canes a child. Here the child could be just any; the sentence merely asserts that the set denoted by the NP predicate child is not empty. In (b) on the other hand we have just any teacher caning a particular child. This amounts to saying, when the determiner bí co-occurs with the bare noun, it is always interpreted as referential whether in subject or object position. 3.5.4.2 Bí and kàkrá ‘few’ In the following sentences in (54), bí is attached to the noun kyɪ̀rὲkyɪ́rέnyí ‘teacher’ in subject position in (a) and àbʊ̀frá ‘child’ in object position. 54. a) Kyɪ̀rɛ̀kyɪ́rɛ̀nyí bí hwɪ́-ɪ̀ m̀bʊ̀frá kàkrá Teacher Ref cane-Past children few ‘A (certain) teacher caned a few children’ b) A ̀kyɪ̀rɛ̀kyɪ́rɛ̀nfʊ́ kàkrá hwɪ́-ɪ̀ àbʊ̀frá bí Teachers few cane-Past child Ref ‘Few teachers caned a (certain) child’ 41   The interpretation in (a) will be that a particular teacher (where the speaker has a specific teacher in mind, for instance teacher Kofi and not any other) caned a small proportion of a group of children. It cannot be that there are different groups of few children having potentially different teachers caning them. Also in (b), the referent denoted by the object DP is still specific. The interpretation here will be a few teachers caning a child who is specific (i.e. the speaker having a particular child in mind at the time of the utterance) and not potentially different children to different groups of few teachers. This goes to say that bí whether it is attached to the NP in subject or object position- where scope interactions with the scope operator kàkrá are concerned- is specific. Bí does not fall under the scope of kàkrá.    3.5.4.3 Bí and nyìnáá ‘all’ The story is the same for the universal quantifier nyìnáá ‘all’ when it interacts with determiner bí. Consider the following sentences: 55. a) Kyɪ̀rɛ̀kyɪ́rɛ̀nyí bí hwɪ́-ɪ̀ m̀bʊ̀frá nyìnáá Teacher Ref cane-Past children all ‘A (certain) teacher caned all (the) children’ b) Àkyɪ̀rɛ̀kyɪ́rɛ̀fʊ́ nyìnáá hwɪ́-ɪ̀ àbʊ̀frá bí Teachers all cane-Past child Ref ‘All (the) teachers caned a (certain) child’ The bí DPs in both subject position (a) and object position (b) are specific. The interpretation of (a) will be a particular teacher caning all the children. In a situation where all the children were caned, possibly not by the same teacher but for instance by different ones for each, the sentence will be 42   infelicitous. Also in (b) the interpretation will be all teachers caning one particular child. There cannot be different teachers caning different children. If the situation is one where different teachers caned different specific children, the sentence in (b) will be infelicitous. This amounts to saying that bí never falls within the scope of nyìnáá. The data in (53-55) support the argument that bí is not allowed to fall within the scope of other scope- bearing operators, as argued for referential and specific indefinites by Fodor and Sag and Ionin as outlined above. This argumentation that bí does not fall under the scope of kàkrá ‘few’ and nyìnáá ‘all’ is not the end of the story however. This is because the scope behaviour of bí changes when it interacts with the universal quantifier bíárá ‘every’ in the language. In the following sub-section, I am going to discuss the scope interaction of bí and bíárá and the problem that comes about when bí is in object position. 3.5.4.4 Bí and bíárá ‘every’ In this subsection, I am going to discuss the scopal interaction between determiner bí and the universal quantifier bíárá. The claim so far has been that bí has a referential status irrespective of its position in the sentence or which other scope bearing operators are there in the sentence. However, when bí occurs with bíárá in an utterance and bí occupies object position, something interesting happens. I argue that, with bíárá in subject position and bí in object position, bí appears to have two interpretations. The most readily available interpretation will have bí apparently falling under the scope of bíárá. The other interpretation is the referential one. This poses a problem for my analysis so far that bí has a referential status, as referentials must NOT fall under the scope of other logical operators. Consider the following sentences: 43   56. a) Kyɪ̀rɛ̀kyɪ́rɛ̀nyí bí hwɪ́-ɪ̀ àbʊ̀frá bíárá Teacher Ref cane-Past child every ‘A (certain) teacher caned every child’ b) Kyɪ̀rɛ̀kyɪ́rɛ̀nyí bíárá hwɪ́-ɪ̀ àbʊ̀frá bí13 Teacher every cane-Past child Ref ‘Every teacher caned a (certain/specific) child’ Sentence (56a) has the referential interpretation for bí, where a particular teacher canes all contextually salient members of the set denoted by the NP àbʊ̀frá ‘child’ (i.e., the children present at the time of caning). Sentence (56b) cannot be as simple as that because it has two different interpretations. The first interpretation is that each teacher has a particular child that he caned, while the second one involves a particular single child being caned by each teacher. A solution to the problem with cases like the one in (56b) will be to say, following Kratzer (1998), that there is an implicit variable on the bí DP which can get bound by the quantifier in this environment. Hence in a sentence like (56b) there is a particular/specific child for each teacher, yielding an apparent narrow scope reading. Adopting Kratzer's analysis of specific indefinites enables me to maintain the claim that bí has a unified denotation, in spite of the apparent variability shown in (56b). I will spell this out in what follows. Kratzer (1998), like Fodor and Sag, treats indefinites as being ambiguous between quantificational and specific readings. Kratzer however argues that the specific reading for indefinites may not necessarily be referential, but rather depends on a choice function (also see Reinhart 1997, Winter 1997, Romero 1997 and Mathewson 1999 among others for choice function analyses of indefinites). Thus, specific indefinites are interpreted through contextually determined choice functions which “are often intended                                                              13 For one speaker, only the referential interpretation is available here. 44   by the speaker, but not revealed to the audience” (Kratzer 1998). A choice function is defined as the following: “A function f is a choice function (CH(f)) if it applies to any non-empty set and yields a member of that set” (Reinhart 1997:372). In the following sentence in (57a), the indefinite NP can be interpreted as a choice function as in (b). 57. a) Richard is dating a certain woman. (Kratzer, 1998:8) b) (Richard is dating f(woman)). The f in (57b) is a contextually salient choice function which picks one woman out of the set of women. Thus, while the noun phrase itself is not directly referential, we obtain the effect that it picks out one particular woman. We also predict, in the ordinary case, that the specific indefinite cannot scope under any quantifier, since it will always still pick out a single woman as in (58). 58) Every professor rewarded every student who read some book I had reviewed for the New York Times. (Kratzer, 1998:10) Even though in (58) there is a quantifier in the sentence, the specific indefinite still picks out a single individual, in this case one particular book (it cannot be different books for each professor). So far, Kratzer's analysis makes the same predictions as Fodor and Sag's does. However, in (59) below, the indefinite does seem to allow an intermediate scope reading, (i.e. being able to scope under the universal quantifier). The relevant reading is where for each professor, there is a certain book she reviewed (a different book for each professor, such that the professor rewarded every student who read that book). 59) [Every professor]ᵢ rewarded every student who read some book sheᵢ had reviewed for the New York Times. (Kratzer, 1997:10) 45   Kratzer argues that the presence or absence of the anaphoric pronouns within the indefinite NP in these sentences14 facilitates the two different interpretations. The presence of the bound variable pronoun in (59) as opposed to the referential one in (58) gives an intermediate scope reading to (59). This is basically because there are different professors being quantified over, and there is a pronoun bound by every professor that ensures that there is a different set which is given to the choice function for each professor. So for each professor, we take the set of the people who read the book that that professor reviewed, and then the choice function picks one of them. But in (58), there is one set: the book I had reviewed for the New York Times. Since there is just one set all the time, there can only be one book picked by the choice function. In a nutshell, indefinites are ambiguous between quantificational and specific readings, and the specific interpretation can either give rise to a single individual being picked out, or different individuals being picked out if the relevant NP itself contains a bound variable pronoun. Similarly in Akan, the presence or absence of a bound variable pronoun results in different interpretations of the indefinite NP. In sentences like those in (58) and (59) above, the bound pronoun will facilitate the intermediate scope reading while a referential one facilitates a referential reading. This is illustrated in the following sentences in (60). 60. a) [Kyɪ̀rɛ̀kyɪ́rɛ̀nyí bíárá]ᵢ kyɛ́-ɛ̀ àbʊ̀frá bíárá áà ɔ̀-káǹ búúkùú bí Teacher every give-Past child any CM 3sg-read-Past book Ref áà ɔ́ᵢ-kyɪ́rɛ́wɪ́-ɪ̀ àdzɪ́ CM 3sg-write-Past thing ‘Every teacher gave any child who read a (certain) book s/he had written a gift’                                                              14 Both sentences (58) and (59) also allow the narrowest scope reading, where each professor rewards every student who read ANY book she/I reviewed. 46   b) Kyɪ̀rɛ̀kyɪ́rɛ̀nyí bíárá kyɛ́-ɛ̀ àbʊ̀frá bíárá áà ɔ̀-káǹ búúkùú bí Teacher every give-Past child any CM 3sg-read-Past book Ref áà mɪ-́kyɪ́rɛ́wɪ́-ɪ̀ àdzɪ́ CM 1sg-write-Past thing ‘Every teacher gave any child who read a (certain) book I had written a gift’ (Adapted from Kratzer, 1998:10) The interpretation in (a) will be where each of the teachers gives a gift to any student who read a specific book that teacher had written. On the other hand, (b) will have one interpretation- referential- where all the students read one specific book which is one the speaker wrote15. Interestingly, however, an overt bound variable pronoun is not necessary to facilitate intermediate scope readings in Akan. A sentence similar to those in (60a, b), but with no pronoun in the restrictive clause, will also have both the referential and the intermediate readings. Consider the following sentence in (60c): c) 16Kyɪ̀rɛ̀kyɪ́rɛ̀nyí bíárá kyɛ́-ɛ̀ àbʊ̀frá bíárá áà ɔ̀-káǹ búúkùú bí àdzɪ́ Teacher every give-Past child any CM 3sg-read-Past book Ref thing ‘Every teacher gave any child who read a (certain/specific) book a gift’. (Adapted from Kratzer, 1998:10) The interpretation for the indefinite NP will be a specific single book or different specific books for each teacher. The intermediate scope reading of sentence (60c) without an overt bound pronoun is not a problem for the current analysis as Kratzer herself does not require an overt pronoun to get the bound reading - she allows implicit variables. The intermediate interpretation of sentences like (60c) is                                                              15 A co-varying reading could be forced here, where the students could have each read a different book the speaker wrote which is specific to each teacher. 16 Again for the same speaker who gets only a referential interpretation for sentence (56b) above, (60c) also can get only a referential interpretation. 47   however lost when there is more description of the indefinite NP. In the following sentences in (61), (a) has two interpretations, a referential and an intermediate scope reading, while (b) has only the referential interpretation. 61. a) Bànyín bíárá tú-ù kɔ̀-tsɪ̀ná-à Kùmásɪ́ ɔ́nám bàsíá bí ńtsí Man every uproot-Past go-sit-Past Kumasi because woman Ref so ‘Every man moved to settle in Kumasi because of a (specific) woman’ b) Bànyín bíárá tú-ù kɔ̀-tsɪ̀ná-à Kùmásɪ́ ɔ́námdɛ́ɛ̀ bàsíá bí tsɪ̀ hɔ́ Man every uproot-Past go-sit-Past Kumasi because woman Ref sit there ‘Every man moved to settle in Kumasi because of a (certain) woman who lived there’ (Adapted from Kratzer, 1998:2) The presence of a modifier (like kɔ̀kɔ̀ɔ́ ‘red’, tuntum ‘dark/black’) in (60c) or (61a) will also result in a referential interpretation. 62. a) Kyɪ̀rɛ̀kyɪ́rɛ̀nyí bíárá kyɛ́-ɛ̀ àbʊ̀frá bíárá áà ɔ̀-káǹ búúkùú kɔ̀kɔ̀ɔ́ bí àdzɪ́ Teacher every give-Past child any CM 3sg-read-Past book red Ref thing ‘Every teacher gave any child who read a (certain) red book a gift’. b) Bànyín bíárá tú-ù kɔ̀-tsɪ̀ná-à Kùmásɪ́ ɔ́nám bàsíá tùntúm bí ńtsí Man every uproot-Past go-sit-Past Kumasi because woman black/dark Ref so ‘Every man moved to settle in Kumasi because of a (certain) dark woman’ (Adapted from Kratzer, 1998:5) 48   In (a) here, the only interpretation available is one specific red book that all the students read and (b) will be one specific dark woman for all the men. This is due to the rich description 17of the NP which Fodor and Sag (1982) argue tends to favor a referential interpretation of an indefinite NP. When a bí DP with a rich description occurs with a bound pronoun the two interpretations-referential and intermediate scope readings- resurface. The analysis predicts that even when there is a rich description, a bound pronoun can still save the intermediate scope reading. This is illustrated in (63). 63. Bànyín bíárá tú-ù kɔ̀-tsɪ̀ná-à Kùmásɪ́ ɔ́nám bàsíá tùntúm bí Man every uproot-Past go-sit-Past Kumasi because woman black/dark Ref áà ɔ́-dɔ̀ nʊ ńtsí Rel 3sg.subj-love 3sg.obj so ‘Every man moved to settle in Kumasi because of a (specific) dark woman he loves’ As the Akan data show, when an NP containing the universal quantifier bíárá ‘every’ occurs in subject position with a bí DP occurring in object position, there are two possible readings for bí: referential and narrow (or intermediate scope) readings. However, my analysis is NOT that bí is ambiguous: it is always a choice function determiner. Following Kratzer, there is a free choice function variable, and this predicts that in cases where there is an (overt or implicit) bound variable inside the bí-phrase, the reading is an apparently narrower-than-widest scope reading18. It should be noted that for Kratzer, some is ambiguous between a quantifier (which can take narrowest scope and be completely non- specific), and a choice function (which can be purely referential or take apparent intermediate scope). The Akan bí is UN-ambiguous; it is always a choice function, with the predicted two readings.                                                              17 The intermediate scope reading can be forced here too, but it is dispreferred as predicted. 18 Kratzer says the specific indefinite does not actually take scope below, it just LOOKS like it. That’s why she calls this effect ‘pseudo-scope’. 49   3.5.4.5 Bí does not allow intermediate scope with nyìnárá ‘all’ and kàkrá ‘few’ It is worth noting that the intermediate scope reading is peculiar to only bíárá ‘every’. This is because the other quantifiers- nyìnáá ‘all’ and kàkrá ‘few’- do not allow an intermediate scope reading. 64. a) Àkyɪ̀rɛ̀kyɪ́rɛ̀fʊ́ nyìnáá kyɛ́-ɛ̀ àbʊ̀frá bíárá áà wɔ̀-káǹ búúkùú bí àdzɪ́ Teachers all give-Past children any CM 3pl-read-Past book Ref thing ‘All teachers gave any child who read a (certain) book a gift’. b) Àkyɪ̀rɛ̀kyɪ́rɛ̀fʊ́ kàkrá kyɛ́-ɛ̀ àbʊ̀frá bíárá áà wɔ̀-káǹ búúkùú bí àdzɪ́ Teachers few give-Past children any CM 3pl-read-Past book Ref thing ‘Few teachers gave any child who read a (certain) book a gift’. In (a), the bí is interpreted as referential, with any child having to read one particular book for all teachers to receive a gift. An explanation for this will be the fact that all might not be a true quantifier as has been argued in the literature (see Brisson 1998, Ferch 2009 and Davis (to appear) among others). This is because all is not able to bind a variable pronoun. Also in (b), bí is interpreted as referential because there is only one particular book for the few teachers so that any child who read that book was given a gift by those few teachers. Few can also be described as not being a true quantifier. This is in lines with the argument made by Davis (to appear) that, in St'át'imcets (a Salish language of British Columbia) which lacks generalized quantifier, there is no scopal asymmetry between DPs containing strong quantifiers or weak quantifiers with proportional reading for example, half, both and many. Since few has a proportional reading, it falls into this category and therefore cannot bind a variable pronoun in St'át'imcets. The same may be true of Akan kàkrá; further research is required. 50   3.6 Conclusion In this chapter I have argued that the morpheme bi as an indefinite marker encodes specificity. I have shown that the morpheme is used as a pronominal (which is toneless) and a determiner (which bears a high tone). In its pronominal use, it can occur alone without an NP but must be linked to an NP mentioned previously in the discourse. By virtue of this anaphoric relation, the pronominal bi gets its specificity status. This is along the lines of Enç (1991). The determiner bí on the other hand cannot stand independently of a noun. It must occur after a noun. It introduces novel discourse referents into the discourse and is interpreted as a choice function along the lines of Kratzer (1998). 51   4 Semantics of Akan nʊ   4.1 Introduction This chapter is a discussion of the semantics of the morpheme nʊ in Akan. Nʊ appears in a couple of different syntactic environments (Amfo and Fretheim (2005), Fretheim and Amfo (2005) and Amfo (2006)). These syntactic environments correlate with certain tonal differences; however nʊ has a corresponding set of different (but related) semantic functions. Nʊ can appear: a) As a pronominal:- 1) Ésí ká-à nʊ̀ Esi bite-Past 3sg.obj ‘Esi bit him/her’ b) Within a DP as a determiner following the NP:- 2) Pàpá nʊ́ bá-à há Man Fam19 come-Past here ‘The man came here’ c) Within a CP as a dependent clause marker:- 3) Kòfí hú-ù máàmɪ́ nʊ́ áà ɔ̀-tɔ̀n tám nʊ́ Kofi see-Past woman Fam Rel 3sg.subj-sell cloth DCM ‘Kofi saw the woman who sells cloth’                                                              19 I am glossing nʊ́ as 'Familiar' in accordance with my analysis of the morpheme, to be given in section 4.5 below. 52   Within the DP domain, the morpheme is used as a definite determiner, while in the CP domain it is used to mark the dependent clause. I am going to discuss the various functions of the morpheme and show the differences and similarities between the other functions and its function as a definite marker. Also I will show that though this morpheme is used to mark definiteness in the language, it is quite different from the definite marker the in English. I am going to argue, based on comparison between the Akan data and German data, that the morpheme nʊ when used as a definite article is equivalent to what is termed ‘strong definite’ marking in the literature (Schwarz 2010). 4.2 Literature review Definiteness has been discussed extensively in the literature, especially in European languages like German and English. Definiteness has been described differently by different authors, including for example Heim (1990), Kadmon (1990), Hawkins (1991), Birner and Ward (1994), Lyons (1999), Abbott (2001) and Elbourne (2001). One famous author on definiteness is Heim. Heim (1982) describes definites as having to introduce familiar variables as against indefinites which introduce novel variables. 4.2.1 Marking of definiteness in Akan In Akan, the definite marker nʊ́ can be distinguished from the indefinite marker bí (4) or the bare noun (5). 4. a) Mʊ̀tɔ́-ɔ̀ èkùtú bí 1sg.subj-buy-Past orange Ref ‘I bought a (certain) orange’ 53   b) Èkùtú nʊ́ yɛ̀ dɛ̀w dɛ́ɛ̀ Orange Fam be nice so ‘The orange is/was so nice’ 5. a) Mʊ̀-tɔ́-ɔ̀ èdzìbán 1sg.subj-buy-Past food ‘I bought food’ b) Èdzìbán nʊ́ yɛ̀ dɛ̀w pápá Food Fam be nice good ‘The food is/was really tasty’ In (4) the referent èkùtú bi ‘a certain orange’ in (a), is mentioned for the first time in the discourse; however in (b) the referent is being mentioned for the second time so nʊ́ not bí has to be used. Also in (5), sentence (a) has a discourse-new referent so the bare noun is used, but in (b) because the referent is mentioned for the second time, the bare noun cannot be used and nʊ́ is used. This is to say that the definite marker nʊ́ is not exchangeable for bí and/or the bare noun. 4.2.2 Amfo’s analysis of nʊ Previous work done on nʊ includes Amfo and Fretheim (2005), Fretheim and Amfo (2005) and Amfo (2006). Amfo (2006) argues that nʊ as a multifunctional marker has different syntactic positions but only one semantic value, which is “uniquely identifiable”. She argues for two tonal alternations for the marker: the low tone nʊ is used for the 3rd person singular pronominal object and the high tone nʊ for post-nominal and post-clausal marking. Amfo states that the high tone nʊ when it occurs in post- nominal position functions as a determiner, and in post-clausal position as a dependent clause marker (DCM). In the following sentences, the nʊ in (a) is a pronoun, in (b) a determiner and in (c) a DCM. 54   6. a) Kwámɪ̀ dzɪ̀ èdzìbán nʊ́ má-à nʊ̀ Kwame take food Fam give-Past 3sg.obj ‘Kwame gave the food to him/her’ b) Kwámɪ̀ dzɪ̀ èdzìbán nʊ́ má-à àbʊ̀frá nʊ́ Kwame take food Fam give-Past child Fam ‘Kwame gave the food to the child’ c) Ànàpá nʊ́ nà Kwámɪ̀ dzɪ̀ èdzìbán nʊ́ má-à nʊ́ Morning DCM Foc Kwame take food Fam give-Past 3sg.obj ‘It was in the morning that Kwame gave the food to him/her’ In the determiner use of the morpheme, Amfo states that it can be used as a definite marker as well as a demonstrative marker. In the following sentence the nʊ́-marked NP has two interpretations. It can have either a definite or a demonstrative interpretation with the distinction coming either only from the context, or from the addition of the demonstrative marker dɛ́m pre-nominally as in (b). 7. a) Èdùàbá nʊ́ yɛ̀ dɛ̀w Fruit Fam be nice ‘The/that fruit is nice’ b) Dɛ́m èdùàbá nʊ́ yɛ̀ dɛ́w Dem Fruit Fam be nice ‘That fruit is nice’ Amfo argues that when the NP is interpreted as definite it has the cognitive status of uniquely identifiable-though in many instances it is familiar. However when it is interpreted as demonstrative by having the pre-nominal demonstrative marker co-occurring with it, it gets a familiar interpretation. Amfo (2006) agrees with the cognitive status given to ‘the N’ by Gundel et al, and argues for the 55   same for Akan. This makes her model Akan determiners after Gundel et al.’s cognitive status hierarchy as follows: 8) in focus ɔ̀-s/he; ɛ̀- ‘it’; nʊ̀ ‘him/her’ activated20 ɔ́nʊ́ ‘S/HE’; ɛ́nʊ́ ‘IT; ɔ̀-s/heTHAT’; N-yí ‘this N’; éyí21‘this’ familiar uniquely identifiable sáá22 N-nʊ́ ‘that N’ N-nʊ́ ‘the N’ referential type identifiable23 N-bi ‘a certain NØ24 (Fretheim and Amfo 2005:110) I am going to argue against Amfo’s claim that nʊ́ is only uniquely identifiable and not familiar, and that it gets its familiarity status from the demonstrative marker dɛ́m. The rest of the chapter is organized as follows: section 4.3 discusses the tonal alternations of the pronominal and the determiner functions of the morpheme. Section 4.4 discusses the function of nʊ as a pronominal. Sections 4.5 discusses the function of nʊ as a definite article and 4.6 discusses the function of nʊ as a distal demonstrative marker. Section 4.7 touches on its function as a clausal marker while section 4.8 concludes the chapter.                                                              20 ‘Activated’ means the discourse referent is represented in current short-term memory though it can be retrieved from long-term memory (Gundel et el 1993). 21 In Fante (the dialect used for this research) this form is íyí. 22 The Fante version of this is dɛ́m. 23 This means the addressee is expected to access the representation of the type of object under discussion (Gundel et el 1993). 24 This does not necessarily mean there is a null determiner. It shows that the noun occurs by itself with no overt determiner. Whether bare nouns in Akan have a null determiner or not, is not discussed in this thesis and this does not have consequences for the current research. 56   4.3 Tonal alternations for the morpheme nʊ In this section I am going to discuss the tonal alternations between the functions of nʊ as a pronominal and as a determiner. I am going to argue that the morpheme when it is used as a pronominal is toneless but when it is used as a determiner it is marked with a high tone. 4.3.1 Tonal realization of the pronominal nʊ Nʊ as a 3rd person singular object pronoun is toneless. The tone of the pronominal morpheme is dependent on a number of factors including the phonology, the type of structure in which it occurs as well as its position in the sentence. There is tone spreading from the preceding tone to the morpheme in its argument position in non-A- bar configurations. Thus the phonology determines the tone of the pronominal nʊ in this sentence type. This is illustrated in the following sentences in (9) and (10). 9. Kòfí rʊ́-kɔ́-hù Ésí Kofi Prog-go-see Esi ‘Kofi is going to see Esi’ 10. a) Kòfí rʊ̀-kɔ̀-hù nʊ̀ Kofi Prog-go-see 3sg.obj ‘Kofi is going to see her’ b) Kòfí rʊ̀-kɔ́-hú nʊ́ Kofi Prog-go-see 3sg.obj ‘Kofi is going to see her’ 57   It can be noted that the tone the nʊ receives depends on the tone the preceding tone bearing unit bears. In (10a), the tone on the preceding tone bearing unit hu25 ‘see’ is low, thus the tone on the pronoun is also low. However in (10b) the preceding word hu ‘see’ bears a high tone, therefore the pronoun gets a high tone. This amounts to saying that in the same position in the same sentence type, there is tonal alternation (low or high) for the morpheme which is determined by the preceding tone. In sentences containing an A-bar configuration, where the morpheme is used as a resumptive pronoun however, the story is different. The phonology alone does not determine the tone the pronoun bears; different factors come into play here: 11. a) Wɔ́áná nà Kòfí hú-ù nʊ́ Who Foc Kofi see-Past 3sg.obj ‘Who was it that Kofi saw?’ b) Wɔ́áná nà Kòfí hù-ù nʊ̀ ńdɪ́dà Who Foc Kofi see-Past 3sg.obj yesterday ‘Who was it that Kofi saw yesterday?’ 12. a) Wɔ́áná nà Kòfí bé-hyìá nʊ́ Who Foc Kofi Fut-meet 3sg.obj ‘Who is it that Kofi will meet?’ b) Wɔ́áná nà Kòfí bé-hyìá nʊ̀ ɔ̀kyɪ́ná Who Foc Kofi Fut-meet 3sg.obj yesterday ‘Who is it that Kofi will meet tomorrow?’ In (11a) the pronoun bears a high tone in sentence final position though the preceding tone is low. Also in (12a), the tone on the pronoun is high just like the preceding tone. However in the (b’s) where                                                              25 Speakers have the option of choosing between the two forms of hu ‘see’. It is not clear why this is so. 58   the modifiers ńdɪ́dà ‘yesterday’ and ɔ̀kyɪ́ná ‘tomorrow’ are sentence final, the pronoun gets a low tone. This goes to say that in an A-bar movement construction, the preceding tone does not determine the tone of the pronoun; however the position of the pronoun in the structure determines it. When the pronoun occurs in sentence final position, it receives a high tone irrespective of the tone preceding it, but when something else follows, it receives a low tone irrespective of the tone preceding it. Note that factors other than the phonology play a role in determining the tone the pronominal nʊ bears in these environments. However further investigation (which is beyond the scope of the current research) needs to be conducted for a full analysis of the factors that influence the tonal realization of the Akan pronominal nʊ. As items with unstable tone can be assumed to be toneless (Déchaine 2001a, b), I adopt a toneless analysis of pronominal nʊ based on the data in (9)-(12). 4.3.2 Tonal realisation of the determiner nʊ The determiner nʊ́ bears a high tone. This is irrespective of the preceding tone, the position or the type of sentence the morpheme is occurring in. In the following sentences in (13), nʊ́ occurs after a low tone bearing unit in (a), and a high tone bearing unit in (b). 13. a) Fà pàpà nʊ́ Take fan Fam ‘Take the fan’ b) Frɛ̀ pàpá nʊ́ Call man Fam ‘Call the man’ 59   As (13) shows, the determiner nʊ́ bears a high tone whether the preceding tone is low or high. Consider also the following sentences in (14) and (15) with an A-bar configuration: 14. a) Kwèsí nà ɔ́-fá-à pàpà nʊ́ Kwesi Foc 3sg.subj-call-Past fan Fam ‘It was Kwesi who took the fan’ b) Kwèsí nà ɔ́-fá-à pàpà nʊ́ ńdɪ́dá Kwesi Foc 3sg.subj-call-Past fan Fam yeaterday ‘It was Kwesi who took the fan yesterday’ 15. a) Ɔ́nʊ́ nà ɔ́-fr̀ɛ́-ɛ̀ pàpá nʊ́ 3sg Foc 3sg.sub-call-Past man Fam ‘It was him who called the man’ b) Ɔ́nʊ́ nà ɔ́-fr̀ɛ́-ɛ̀ pàpá nʊ́ ńdɪ́dá 3sg Foc 3sg.subj-call-Past man Fam yesterday ‘It was him who called the man yesterday’ Also in A-bar configurations like the sentences in (14) and (15), the determiner nʊ́ has a high tone whether it is sentence final or not. This suggests that the tone on the determiner nʊ́ is fixed therefore can be assumed to be lexical. In this section, I have argued that while the pronominal nʊ is toneless, the determiner nʊ has a lexical tone which is high. In what follows I am going to discuss the functions of the pronominal (section 4.4) and the determiner nʊ́ (section 4.5) and what these two functions have in common. 60   4.4 Nʊ as a pronominal In this section, I am going to discuss the function of nʊ as a pronominal. I am going to argue that the morpheme in its pronominal use is interpreted as familiar. 4.4.1 Akan pronouns Pronouns in Akan can be grouped into strong and weak forms, where the strong forms are more morphologically complex than the weak ones. The following chart in table 2 (adapted from Abor 1998) shows the full pronominal paradigm in the language. Table 2 Full pronominal paradigm Strong Forms (H) Weak Forms (Toneless) Possessive Person Number Nominative Accusative 1sg a ́mɪ́/e ́mi ́ mɪ26 mɪ /-m mɪ́ 2sg Ɔ́wʊ́ ɪ- wʊ/w wʊ ́ 3sg (animate) Ɔ́nʊ́ ɔ- nʊ/n nɪ́ 1pl Ɛ̀hɛ́n yɛ hɛn/yɛn hɛ́n 2pl Íhʊ́m/Íhɔ́n wɔ/hʊm hʊm/ hɔ́n hʊ́m/hɔ́n 3pl Íhɔ́n wɔ hɔn hɔ́n 3rd Inanimate Sg Ɔ́nʊ́ ɔ-/o- Ø nɪ́ Pl Íhɔ́n wɔ Ø hɔ́n                                                              26 The tones on the nominative pronouns are prosodically determined; they have a low tone in non-A-bar constructions and a high tone in A-bar constructions. The accusative weak pronouns pattern in the same way as the pronoun nʊ. 61   The strong pronominal forms which bear high tones occur at the left periphery in A-bar configurations e.g. in an A-bar focus position. The weak forms however occur in argument position, and/or as resumptive pronouns in A-bar movement constructions. In the following sentences in (16), (b) is an answer to the question in (a). In (bii) the strong pronoun occurs to the left of the focus marker nà while the weak form appears in sentence final position as a resumptive pronoun. 16. a) Ì-dzɪ̀ èdzìbán nʊ́ má-à wɔ̀áná, Àbɪ́ná áà? 2sg.subj-take food Fam give-Past who Abena QP ‘You did give the food to who, Abena?’ (lit) Who did you give the food to, was it Abena?’ b) Ńyèw! Ɔ́nʊ́ nà mɪ́-dzɪ́ má-à nʊ́ Yes 3sg Foc 1sg.subj-take give-Past 3sg.obj ‘Yes! It was her that I gave the food to’ In (17) however, (b) as an answer to (a) has the weak pronoun nʊ in accusative position in a non-A-bar movement construction. 17. a) Ì-má-à Ésí ɪ́bɛ̀nàdzɪ́? 2sg.subj-give-Past Esi what ‘What did you give to Esi?’ b) Mɪ̀-má-à nʊ̀ èdzìbán 2sg.subj.give-Past 3sg.obj food ‘I gave her food’ The weak pronominal form cannot occur in A-bar position as the strong form does. 62   18. a) *Nʊ́ nà mɪ́-dzɪ́ má-à nʊ́ 3sg.obj Foc 2sg.subj-take give-Past 3sg.obj ‘It was her that I gave the food to’ However the strong form can occur in a non-A-bar position for contrastive purposes, sometimes with contrastive modifiers like mbom ‘rather/instead’. (See Adesola 2005, 2006 for a similar discussion for Yoruba pronouns). b) Mɪ̀-dzɪ̀ má-á ɔ́nʊ́/ ɔ́nʊ́-ara/ ɔ́nʊ́ mbom 1sg.subj-take give-Past 3sg/ 3sg-just/3sg rather(instead) ‘I gave it to her/her herself/her instead’ It should be noted that the strong pronominal does not distinguish between nominative and accusative forms. (19) illustrates a strong nominative form. 19. a) Wɔ̀áná nà ɔ́-frɛ́-ɛ̀ wʊ́, Ésí áà? Who Foc 3sg.subj-call-Past 2sg.obj Esi QP ‘Who called you, was it Esi?’ b) Nyew. Ɔ́nʊ́ nà ɔ́-frɛ́-ɛ̀ mɪ́ Yes 3sg Foc 3sg.subj-call-Past 1sg.obj ‘Yes. It was her who called me’ The weak pronoun nʊ, as already discussed, is the 3rd person object pronoun; therefore, it cannot occur in subject position hence the ungrammaticality of (20a). 20. a) *Nʊ dzí-ì èkùtú 3sg-(ani) obj eat-Past orange ‘He ate orange(s)’ 63   b) Ò27-dzí-ì èkùtú 3sg-subj-eat-Past orange ‘He ate orange(s)’ As can be noted from the pronoun chart, there is an animacy distinction in the 3rd person singular accusative pronominals. Hence the pronoun nʊ which marks a 3rd person singular animate object cannot be used for a 3rd person singular inanimate object. In the following sentences (21) and (22) the (b)’s have as their objects pronouns substituted for the nouns in the (a)’s: 21. a) Kwèsí fá-à àbʊ́wátsán Kwesi take-Past snake ‘Kwesi took a snake’ b) Kwèsí fá-à nʊ̀ Kwesi take-Past 3sg.(ani.).obj ‘Kwesi took it’ 22. a) Kwèsí fá-à sɪ́kán Kwesi take-Past knife ‘Kwesi took a knife’ b) #Kwèsí fá-à nʊ̀28 Kwesi take Past 3sg.(ani).obj ‘Kwesi took it’                                                              27 The vowels used as clitics are realized with the same ATR status as the vowel of the verb root. 28(22b) is acceptable when the referent of the pronoun is animate. 64   c) Kwèsi f á-ɪ̀29 Kwesi take Past-(inani obj) ‘Kwesi took it’ (22a) has as its object the NP sɪ́kán ‘knife’ which is an inanimate object, so when it is substituted with an object pronoun it has to be covertly marked. This is because Akan’s inanimate 3rd person pronoun is not marked overtly. In the next sub-section, I will show that the pronominal nʊ has familiarity status. 4.4.2 Familiarity of the pronominal nʊ Pronouns are argued to be definite descriptions in the literature. One of the most notable authors is Postal (1969) who argues that pronouns are definite. Roberts (2005) also treats pronouns as definite. Prince (1988) also argues that pronouns are definite descriptions because their referents are salient and therefore hearer old30. Something is said to be salient when it is “appropriately in the hearer’s consciousness at that point in the construction of the discourse model” (Prince1988:8). In the following sentences in (23) and (24), both the strong pronoun in (23b) and the weak pronoun in (24b) refer back to Esi and Ama respectively. 23. a) Íbɛ̀nàdzɪ́ ńtsí nà Ésí nà ɔ́-rɪ́-bà há? What reason and Esi Foc 3sg.subj-Prog-come here ‘Why is it that it is Esi who is coming here?’                                                              29 Intransitive verbs in Akan end in a high front vowel in their past tense. The quality of this vowel depends on the ATR status of the vowels of the verb root. This mechanism is also used for verbs within sentences which have inanimate object pronouns as their objects. 30 I take hearer old to include discourse old. 65   b) Ɔ́nʊ́ nà mɪ̀-fɛ́-ɛ̀ nʊ́31 3sg Foc 1sg-call-Past 3sg.obj ‘It was her that I called’ 24. a) Íbɛ̀nàdzɪ́ ńtsí nà Ámá rú-sṹ? What reason and Ama Prog-cry ‘Why is it that Ama is crying? b) Kòfí bɔ́-ɔ̀ nʊ̀ Kofi beat-Past 3sg.obj ‘Kofi beat her’ The context in (24) will be where Ama is seen crying and so the question in (24a) is asked. The reference of the pronouns ɔ́nʊ́ in (23b) and nʊ in (24b) to the individuals in the questions in (23a) and (24a) qualifies them to be hearer old (along the lines of Prince 1988) since their antecedents have already been mentioned in the discourse. I argue that the use of the pronouns will be infelicitous in a context where their referents are not already known to the hearer. This argumentation is in line with Saah (1994) where he states that “nʊ can be construed with an animate referent whose identity has already been established in the previous discourse or by designation” (Saah 1994:107). Pronouns whose referents are not already known to the hearer can attain hearer old status when they are accompanied by a deictic gesture at the time of the utterance. In the following sentence in (25), the pronoun nʊ can be used to refer to a man walking by if the speaker points at him or makes any form of deictic gesture towards the man’s direction, especially when there is no one else in that direction. This amounts to saying that (25) is infelicitous if there is no prior mention of a person and no deictic gesture.                                                              31 Saah (1994) argues that the resumptive pronoun is not an anaphoric pronoun or a spell-out trace. This is because it is base generated and not bound in its binding category; therefore it is an ordinary pronoun. 66   25. Mé-bísà nʊ̀ 1sg.subj-ask 3sg.obj ‘I will ask him/her’ So far it has been established that the pronoun nʊ has a familiarity requirement; its referent must be hearer old (either by being known by the hearer prior to the utterance, or by being made hearer-old by means of a pointing gesture). I now turn to another aspect of the semantics of pronouns, namely whether they are interpreted only referentially, or may also function as bound variables. Evans (1977) discusses four types of pronouns, with three types being referential and one type a bound variable. In Akan, the strong pronominal form is referential; however the weak pronoun can be interpreted as referential as well as supporting bound variable anaphora. The pronominal use of nʊ in the above sentences in (24) and (25) has a referential interpretation. In the following sentence in (26), the pronoun nʊ can be interpreted as referential as well as a bound variable. (See Matthewson 2008 for a similar phenomenon in St'at'imcets). 26) Bànyín bíárà ɔ́-wɔ́ yɪ́r tɔ̀ àdzɪ́ kyɛ̀ nʊ̀ Man every 3sg.subj-at wife buy thing give 3sg.obj ‘Every man who has a wife buy gifts for her’ On the referential interpretation, the quantifier expression does not bind the pronoun, whereas on the bound variable interpretation, the pronoun does get bound by the quantifier expression and does not refer to anything (along the lines of Evans 1977 among many others). The question that may arise is why familiar pronouns (or hearer old in the sense of Prince) be able to be interpreted as bound variables and thus fail to refer. This problem is only an apparent one, as there exist analyses of bound variable readings of definite/familiar variables. One such analysis is that of Elbourne, which I now briefly outline. Elbourne (2005) analyzes pronouns as definite articles. In the following example, the NP shirt has 67   undergone deletion. 27. My shirt is the same as his. (Elbourne 2005:42) Elbourne argues that the sentence could be paraphrased as “My shirt is the same as his shirt”. Just as the shirt in (28) has undergone deletion because of the presence of an identical NP, D-type pronouns32 can have their NP complements deleted in environments where there is an identical NP, as in the following: 28. a) Every trader who owns a camel sits on it. b) Every trader who owns a camel sits on the camel. c) Every trader who owns a camel sits on it camel. (Adapted from Elbourne 2005:48) In (a), Elbourne argues that there could be the NP camel occurring after it -because of the overt identical NP in this environment, camel is deleted. Therefore sentence (a) is identical to (b) and by this virtue it sits in the position of the determiner, as in (c). Elbourne states that this analysis can be extended to the other 3rd person pronominals. For bound and referential pronouns, Elbourne argues that they are definite articles which take indices (which are phonologically null NPs) as arguments. Hence all 3rd person pronouns, including those which are bound, are definites. Following this view, I argue that the D-type pronouns in (26) as well as the bound or referential ones in (25) etc. can be definite articles therefore can be familiar. Adopting an Elbourne-style analysis does not necessarily predict that the same form will surface for both the pronominal use of the definite article, and the determiner use - witness English, where they are pronounced differently (it vs. the). However, Akan provides interesting potential strong support for the Elbourne-style view, because the same morpheme (albeit with different tonal properties) serves both the pronominal as well as the determiner functions of the definite marker. I turn in the next section to the determiner function of nʊ.                                                              32 ‘D-type pronouns’ are pronouns that have definite description interpretation. 68   4.5 Nʊ́ as a determiner Nʊ́ as a determiner has two main functions: as a definite marker and as a distal demonstrative marker. In this section, I am going to discuss the definite marker use of the determiner nʊ́ and show that though the morpheme nʊ́ is used to mark definiteness, it is not the same as the definite article the in English as has been analyzed by other authors. I will go further to argue that the function of nʊ́ as a definite article is equivalent to what is described in the literature as a strong definite. I will do this by comparing Akan data with data from German. 4.5.1 Nʊ́ as a definite marker Definiteness is said to have its prototype in English as the -the definite article- as well as the NPs that are determined by this article (Abbott 2004). The definite article has been analyzed differently by different authors. Prince (1988) states that definiteness can be described in terms of ‘old’ information, meaning the information about the entities or referents may be considered ‘old’ where the speaker’s beliefs about the hearer’s beliefs are concerned, in that the hearer’s old entities or referents are definite. 4.5.2 Definiteness: uniqueness or familiarity? There has been a long-standing argument in the literature about whether definite descriptions signal or encode uniqueness or familiarity. Some of the people who argue in favour of a uniqueness analysis of defnites include Hawkins (1978, 1991), Gundel, Hedberg and Zacharski (1993), Birner and Ward (1994), Abbott (1999, 2000, 2003, 2004) and Szabo (2000, 2003). Some of the familiarity people are Heim (1982, 1983), Kadmon (1987), Prince (1981 and 1988) and Roberts (2003). Gundel, Hedberg and Zacharski (1993) describe a definite NP of the form ‘the N’ as uniquely identifiable, which means that the speaker’s intended referent can be identified by the hearer on the basis of the nominal alone, as in I couldn’t sleep last night. The dog (next door) kept me awake. They argue that identifiability 69   can be based on a representation that already exists in the hearer’s memory, as in the sentence here without what is in the brackets. However they state that Hawkins (1978) and others pointed out that the English definite article the can be used even when the NP is new information to the hearer, therefore unfamiliar. This can be so as long as there is enough description of the NP in the discourse and the hearer is able to make a mental representation of a unique individual which satisfies the predicate denoted by the NP. For instance in the utterances below, the hearer need not be familiar with the particular food/studio the speaker is referring to; so long as the food/studio is uniquely identified by the hearer by forming a mental representation of the possible food that could be on the table/ the studio I am supposed to be going to, the definiteness of the NP is satisfied. 29) The food is on the table / I am going to the studio. This goes to say that in English the presence of the definite article does not necessarily demand that the NP complement be familiar. Those who adopt a familiarity analysis have an answer to this, namely that a discourse referent can be accommodated in cases like (29) (see for instance Heim 1982 among others). 4.5.3 Nʊ́ is obligatorily familiar In Akan, the determiner nʊ́ is used as a definite marker (aside its demonstrative function mentioned above which I return to shortly in section 4.6 below). This morpheme follows the noun to make it definite as in the following: 30) Kwèsí è-dzì èdzìbán nʊ́ Kwesi Perf-eat food Fam ‘Kwesi has eaten the food’ Fretheim and Amfo (2005) and Amfo (2006) argue that the use of nʊ́ has the cognitive status of uniquely identifiable. They argue that though the referent of the nʊ́-phrase is most of the time already 70   known to the hearer, it cannot be said that it is familiar. This is because nʊ́ can be used attributively without it referring to any particular individual but rather to anyone who can fit into the position being described. I argue that the Akan definite marker nʊ́ is not just uniquely identifiable but necessarily familiar. Taking all the data into consideration, it will be infelicitous to use nʊ́ when the information is not hearer old but hearer new, even if the individual satisfying the NP is uniquely identified. I argue in line with Prince’s (1981, 1988) analysis, that nʊ́ is used when the hearer has prior information about the entity or referent. When (30) above is uttered, it presupposes that the èdzìbán ‘food’ had already been mentioned previously, making it hearer old. This amounts to saying that the hearer has knowledge particularly about that food the speaker is making reference to. This prior information makes the NP hearer old. A similar point is illustrated in (31). These sentences will attract a ‘what/which food/cloth?’ kind of response from the hearer, if there has not been a prior mention of the NP in the discourse or the referent is not shared information between the interlocutors. Thus in Akan the definite article nʊ́ is used only when the information is familiar. 31. a) Èdzìbán nʊ́ é-gù fàmù Food Fam Perf-fall ground ‘The food has poured down’ b) Kòfí á-fà tám nʊ́ Kofi Perf-take cloth Fam ‘Kofi has taken the cloth’ The sentences in (31) are only felicitous if the food/cloth is hearer old. Amfo (2007) argues that in an example like the following, the individual satisfying the definite NP need not be familiar to both the speaker and the hearer, because the noun phrase can be used to refer to anyone who fits the description. 71   32) Sɔ́fʊ́ nʊ́ bé-dzí-kán àbɔ̀ m̀páá Preist Fam Fut-take-first beat prayer ‘The priest will pray first (before anything taking place).’ The definite NP in the above sentence may be unfamiliar to the interlocutors in terms of name or real identity, but it is familiar in terms of the function it is performing, as in being a priest and being expected to pray for whatever function that is going on. An utterance like this will not be felicitous in a place where no prayer from any priest is needed or expected, say at a night club. My argument here is that so long as it is the norm that at that particular function a priest is expected to pray, the audience may take the presence of the priest for granted though it might not have been mentioned in the immediate past. The fact that the information is registered in the mental faculty of the people present as the usual norm this means that the familiarity of the priest may be accommodated. Fretheim and Amfo (2005) argue that nʊ́ without the demonstrative marker sáá (dɛ́m for Fante) makes an NP only uniquely identifiable with no familiarity status. In the following sentences in (33), they argue that pictures in (a) are familiar due to the presence of the demonstrative marker dɛm33. However they are only uniquely identifiable in (b) where the demonstrative marker is absent. 33. a) Mɛ́-kyɪ̀rɛ́ wʊ̀ dɛ́m m̀fònyí nʊ́ áà ò-fír Ghánà nʊ́ 1sg.subj-show you Dem picture Fam Rel it-come Ghana DCM ‘I will show you those pictures from Ghana’ b) Mɛ́-kyɪ̀rɛ́ wʊ̀ m̀fònyí nʊ́ áà ò-fír Ghánà nʊ́ 1sg.subj-show you picture Fam Rel it-come Ghana DCM ‘I will show you the pictures from Ghana’ (Fretheim and Amfo 2005:109) I argue that neither of the sentences can be uttered unless the hearer is aware of the fact that there will be pictures arriving from Ghana, and so even (33b) still presupposes that m̀fònyí nʊ́ the pictures                                                              33 The demonstrative function of dɛ́m will be discussed shortly below. 72   are hearer old therefore familiar. I therefore conclude that the Akan definite marker nʊ́ demands that the NP complement be familiar to the hearer, along the lines of Prince’s hearer old. This quality however distinguishes the Akan definite marker from its counterpart in English and other languages which can have only a uniqueness status for the referent in certain cases. 4.5.4 Akan definite nʊ́ is equivalent to a strong definite Schwarz (2010) describes two main types of definite article in German: weak and strong articles. The weak ones are the ones which contract after certain prepositions. 34. Hans ging zum/ zu dem Haus Hans went to.the.weak/the.strong house ‘Hans went to the house’ (Schwarz 2010:12) In this section, I am going to compare the two types of German definite with the Akan nʊ́. I am going to argue that the strong familiarity status of the Akan definite article makes it equivalent to the strong definite article in German. 4.5.4.1 Weak article definites This type of article is described by Schwarz as a classical uniqueness definite which involves a uniqueness requirement on the NP. The data in (35) is from the Frisian dialect (cited from Ebert 1971:159). 35) Ik skal deel tu a/*di kuupmaan. I must down to the (weak)/ *the (strong) grocer ‘I have to go down to the grocer’ (Schwarz 2010:15) 36) Der Postbote kommt (Context: Was ist los? ‘What is going on?’) The(weak) mailman comes ‘The mailman is coming’ (Schwarz 2010:28) 73   The grocer and the mailman are only uniquely identifiable so the weak article is used; hence the ungrammaticality of the strong article in (35). Schwarz follows Krifka (1984) in characterizing the weak articles - which are non-anaphoric - as being tied to the shared world knowledge of speakers and hearers. This shared world knowledge could be knowledge of the world on the basis of which the hearer is able to form a mental representation of the referent. If this is right then it is in line with the argument made by Gundel et al (1993) that in such situations the referent need not be familiar. In (35) and (36) above, the grocer or the mailman need not be a particular one which is familiar to the hearer. Within Akan, nʊ́ is the only determiner which encodes any kind of definiteness in the language. However the only environment nʊ́ is allowed in is the strong definite environment. This amounts to saying that though Akan does not have two separate definite determiners- one of which is weak and the other strong, like in German- it marks the strong with the definite marker nʊ́ while the weak is marked with bare nouns which are not unique uniquely identifiable but type identifiable. A sentence similar to (36) above in Akan will not require a definite article attached to the NP complement. 37. a) Mʊ̀-rʊ́-kɔ̀ gúá mù 1sg.subj-Prog-go market in ‘I am going to (the) market’ b) Mʊ̀-rʊ́-kɔ̀ gúá mù nʊ́ 1sg.subj-Prog-go market in Fam ‘I am doing the ‘going to (the) market’’ Gúá mù ‘in (the) market’ without the definite article nʊ́, could be any market at all. As long as it is a place where specific activities such as buying and selling take place, it fits the description and the sentence is felicitous. When nʊ́ is attached to the NP, it makes the act of going to the market familiar, 74   it does not make the NP complement definite. Nʊ́ here qualifies the whole event of going to the market in the clause and not just the NP complement. The interpretation will be that the speaker had already mentioned to the hearer about her intention of going to the market. Hence the speaker is undertaking the action already talked about at some point in the past by the interlocutors. The comments by my consultants support the claim that when the definite article is present in sentences like (37) familiarity is required. They made the following comments when (37) with nʊ́ was uttered with no prior mention of ‘going to the market’; one consultant said ‘but you have not mentioned that you will be going to the market’. The second consultant said ‘this sentence is funny, no it’s not good’, and the third said ‘the sentence is odd, it will be good if you told me you would be going to the market’. These comments from the consultants support the fact that the activity of going to the market needs to be prior information for a statement like (37) with nʊ́ to be felicitous. Sentences similar to the mailman sentence in (36) above will have quite a different interpretation in Akan, depending on whether the definite article is attached to the NP complement or not. 38. a) Kòfí hú-ù ɔ̀tʊ̀mfʊ́ nʊ́ Kofi see-Past blacksmith Fam ‘Kofi saw the blacksmith (man)’ b) Kòfí hú-ù ɔ̀tʊ̀mfʊ́ Kofi see-Past blacksmith ‘Kofi saw a/≠the blacksmith’ In (38a), the definite article demands that the ɔtʊmfo ‘blacksmith’ be familiar to the hearer; otherwise the sentence will be infelicitous. However in (38b), the interpretation of the NP complement is quantificational where the existence of the blacksmith is merely asserted. The English sentences in (39) are examples of typical ‘weak definite’ cases where the NP complements need not be familiar. As predicted by my analysis of nʊ́ as a familiarity marker/strong 75   definite article, the Akan equivalents of these sentences do not demand the presence of nʊ́. When nʊ is attached the NPs will have different interpretations. This is because the definite article will not be determining the NPs but rather the entire clause within which the NPs are found. 39. a) This afternoon I went to the park b) John went to the hospital c) I withdrew money from the bank 40. i) a) Èwìàbír yí mʊ̀-kɔ́-ɔ̀ páák dʊ́ Afternoon this 1sg.subj-go-Past park on ‘This afternoon, I went to (the) park’ b) Èwìàbír yí mʊ̀-kɔ́-ɔ̀ páák dʊ́ nʊ́ Afternoon this 1sg.subj-go-Past park-on Fam ‘This afternoon, I went to the park’ c) Park nʊ́ áà mʊ̀-kɔ́-ɔ̀ nʊ́ èwìàbír yí á-hyɪ̀w Park Fam Rel 1sg.subj-go-Past Fam afternoon this Perf-burn ‘The park which I visited this afternoon got/is burnt’ ii) a) Kwámɪ̀ kɔ́-ɔ̀ ásùpítsì Kwame go-Past hospital ‘Kwame went to (the) hospital’ b) Kwámɪ̀ kɔ́-ɔ̀ ásùpítsì nʊ́ Kwame go-Past hospital Fam ‘Kwame went to the hospital’ c) Mé-é-hù ásùpítsì nʊ́ áà Kwámɪ̀ kɔ́-ɪ nʊ́ 1sg.subj-Perf-see hospital Fam Rel Kwame go-Past Fam ‘I have seen the hospital Kwame visited’ 76   iii. a) Mì-yí-ì sìká wɔ̀ sìkàkʊ̀rábɪ̀á 1sg-subj-take-Past money at bank ‘I took money from (the) bank’ b) Mì-yí-ì sìká wɔ̀ sìkàkʊ̀rábɪ̀á nʊ́ 1sg-subj-take-Past money at bank Fam ‘I took money from the bank’ c) Sìkàkʊ̀rábɪ̀á nʊ́ áà mì-yí-ì sìká wɔ̀ nʊ́ á-hyɪ̀w Bank Fam Rel 1sg-subj-take-Past money at Fam Perf-burn ‘The bank which I took money from is burnt’ The presence of the definite marker nʊ́ in sentences (b) in (40) demands that the hearer be aware of the activities taking place in the utterances, just as in the market with nʊ́ case in (38) above. However, the familiarity of the nouns can be forced in special contexts where there is detailed description of the referents. This is what the sentences in (40c) show. 4.5.4.2 Strong article definites Schwarz describes the strong article definite as anaphoric in nature, in that the interpretation of its NP generally depends on a preceding expression. 41) Oki hee an hingst keeft. Di hingst haaltet Oki has a horse bought. The (strong) horse limps ‘Oki has bought a horse. The horse limps’. (Schwarz 2010:15) This is to say that the referent in the second clause is familiar because it has already been mentioned in the preceding clause. This is in line with my analysis of the Akan definite article except that the Akan nʊ́ can be used even for referents mentioned in the remote past for instance a year ago. In Akan, the NP complement need not be discourse old to be familiar. As Prince (1988) argues, an entity or referent can be discourse new, but hearer old and therefore definite. The requirement is that the hearer be familiar with the referent. This could be either because it was mentioned in the 77   preceding utterance or because the hearer can retrieve the referent from their mental faculty. When interlocutors have a common ground- that is when there is shared knowledge where the entity under discussion is concerned- the referent is hearer old, hence familiarity status is established. Schwarz claims that sometimes the distribution of both the weak and the strong articles can overlap in that some utterances can have either of the articles. This is seen in the following sentences. 42. A/Di h\\nj hee tuswark The (weak)/ (strong) dog has tooth ache ‘The dog has tooth ache’ (Schwarz 2010:27) Schwarz cites Ebert (1971a) as saying that sentence (42) with the weak article presupposes that the intended dog does not need to be specified further since there is only one dog at the time and place of the speech so it uniquely identifiable. In the Akan system, the bare noun is used when the hearer has no knowledge about the said dog. The interpretation could be accommodated as a generic reading34 rather than uniquely identified. This is because the distribution of bare singulars is restricted; in subject position they are sometimes not accepted (see Schmitt and Munn 1999 for discussion on bare nouns in Brazilian Portuguese). 43) ??Bɔ̀dɔ́m wɔ̀ kákáw Dog has toothache ‘A dog has tooth ache’ However when the dog is in view the definite article can be used with a deictic gesture accompanying the utterance (this will be more of a demonstrative interpretation which will be discussed shortly below in 4.6). Otherwise the sentence will trigger a question like ‘Which dog?’ This is illustrated in (44) in an out of the blue context.                                                              34 The marginal state of the generic use of (43) is because the generic readings have plural discourse referents rather than singular. 78   44. a) Bɔ̀dɔ́m nʊ́ wɔ̀ kákáw Dog Fam has toothache ‘The dog has tooth ache’ b) Íbɛ́n bɔ̀dɔ́m á? Which dog QP ‘Which dog?’ This is compatible with the German strong article use in (42), as Schwarz argues that the strong article in German can be used with discourse new referents when the utterance is accompanied with a deictic gesture. For any definite NP -i.e. any DP which contains the definite determiner nʊ́ - to be felicitous in Akan, there should be shared knowledge between the speaker and the hearer. This shared knowledge being discourse old, hearer old, or facilitated by a deictic gesture, is what I argue to be the familiarity status of the definite article in Akan. The Akan determiner nʊ́ differs from the English definite article the. A sentence in English like: Can you please bring me the pack of noodles I left in the shopping bag? might not need familiarity of the definite NP to be felicitous; however in Akan if the hearer has no knowledge whatsoever about the referent of the NP the sentence will be odd. For instance, consider a context where Esi visits her friend Ama and in conversation, Ama utters (45) with no prior knowledge of the said cassava: 45) Ésì fá bànkyɪ́ nʊ́ áà ó-gú kɛ́ntsɛ́n mù nʊ́ br ̀à Esi take cassava Fam Rel it-pour basket in Fam come ‘Esi, bring the cassava that is in the basket’ This statement is odd in this context; however, if Ama had already talked about the cassava in the basket, say that it is what they are going to use for lunch when Esi visits, then (45) will be accepted. 79   4.5.4.3 Bridging Another mechanism used by Schwarz (2010) to distinguish between weak and strong definites is bridging. There is a distinction between part/whole bridging which involves weak definites, and inference bridging (where the interpretation of an NP is inferred from the preceding utterance or discourse), which involves strong definites. In these contexts, the articles cannot alternate with each other. The weak article cannot be used in inference bridging, and the strong cannot be used in part/whole bridging. The following sentences35 illustrate this: 46. a) Wi foom a sark uun a maden faan’t taarep. A/#Dem We found the church in the middle of the village the(weak)/the(strong) törem stän wat skiaf tower stood a-little crooked ‘We found the church in the middle of the village. The tower was a little crooked’ b) Das Theaterstück missfiel dem Kritiker so sehr, dass er in seiner The play displeased the critic so much that he in his Besprechung kein gutes Haar an dem #am Autor lieβ review no good hair on the(strong) /the(weak) author left ‘The play displeased the critic so much that he tore the author into pieces in his review’. (Schwarz 2010:34) In similar sentences in Akan, only NPs with an inference interpretation will have the definite article nʊ́ attached to them. This is illustrated in (47) - (49): Part/whole relation:                                                              35 The data here are from both Frisian and standard German dialects. 80   47. a) #Iguan nʊ́ sʊ̀ ara ma nyimpa ebiasa na wo-dzi-i na-asr kʊr nʊ́ Sheep Fam big just Comp person three and 3pl.subj-eat-Past thigh one Fam ‘The sheep was so big that it was three people that ate (the) one thigh.’ (Adapted from Schwarz 2010:34) (Consultant’s comment: This is odd unless there was something peculiar with that particular thigh. For instance, it was crooked or something). b) Iguan nʊ́ sʊ ara ma nyimpa ebiasa na wo-dzi-i na-asrɛ kʊr Sheep Fam big just Comp person three and 3pl.subj-eat-Past Poss-thigh one ‘The sheep was so big that it was three people that ate one of its thighs.’ In (47), (a) cannot be accepted because it is uninterpretable. This is due to the presence of the definite article, which is not licensed in part-whole bridging situations. However, (b) is accepted as the presence of the possessive marker nɪ́ 36 signifies that the NP complement is part of the whole mentioned in the preceding clause. Another example is given in (48): 48. a) #Ye-hu-u dan dadaw bi wɔ ekurasɪ hɔ nkyɛnsɪdan nʊ́ ehodwow 1pl.subj-see-Past building old Ref at village there roof Fam worn-out ‘We saw an old building in the village; the roof was worn out.’ (Adapted from Schwarz 2010:34) (Consultant corrected the sentence by replacing nʊ́ with nɪ́ before roof and said (a) is two statements put together). b) Ye-hu-u dan dadaw bi wɔ ekurasɪ hɔ nɪ́ nkyɛnsɪdan ehodwow 1pl.subj-see-Past building old Ref at village there Poss roof worn-out ‘We saw an old building in the village; its roof was worn out.’ In (48), (a) is acceptable but it has a different interpretation from the intended one which is to show                                                              36 The vowel following the nasal gets assimilated if the following word begins with a vowel. 81   the part/whole relationship between the NP in the preceding clause and that of the second clause. Here there are two different utterances; one being the speaker’s seeing an old building in the village and the other being a roof familiar to the hearer which has become worn out. In (b), nʊ́ is replaced by a possessive and the interpretation shows the relationship between part and a whole. One final example of the failure of nʊ́ to appear in part-whole bridging cases is given in (49). 49. a) Asɔrdan nʊ́ a o-si ekyir hɔ nʊ, adɔnbaa nʊ́ a-stɪw Church Fam Rel 3sg.subj-stand back there DDM bell Fam Perf-tear ‘The church at the back there, the bell is torn.’ (Adapted from Schwarz 2010:34) b) Asɔrdan nʊ́ a o-si ekyir hɔ nʊ́, na-adɔnbaa a-stɪw Church Fam Rel 3sg.subj-stand back there DDM Poss-bell Perf-tear ‘The church at the back there, its bell is torn’ (Consultant’s comment: This (sentence (a)) is also odd because there are two different statements put together).37 The sentence in (49) has similar interpretations to (48). There are two different utterances going on in (a), and in (b) the relationship between the matrix subject and the subordinate subject is a part /whole one. This amounts to saying that where the object is only uniquely identifiable the definite article nʊ́ cannot be used. The inference type of bridging is illustrated in (50) - (52) below. 50. a) Àsáw nʊ́ yɛ́-ɛ̀ ɔ̀hɪ́n nʊ́ fɛ̀w árá mà ɔ̀-kyɛ́-ɛ̀ ɔ̀kyɪ̀rɛ́fʊ́ nʊ́ àdzɪ́ Dance Fam do-Past chief Fam beautiful just Comp 3sg.subj-give-Past trainer Fam thing ‘The dance was so beautiful that the chief gave the trainer a gift.’ b) Àsáw nʊ́ yɛ́-ɛ̀ ɔ̀hɪ́n nʊ́ fɛ̀w árá mà ɔ̀-kyɛ́-ɛ̀ ɔ̀kyɪ̀rɛ́fʊ́ àdzɪ́                                                              37 For one consultant, though the sentence is odd, she will accept it because she belongs to a church that uses bells. 82   Dance Fam do-Past chief Fam beautiful just Comp 3sg.subj-give trainer thing ‘The dance was so beautiful that the chief gave a trainer a gift.’ 51. a) Kwámɪ̀ yɛ́-ɛ̀ dɛ́ɛ̀ ó-rú-twà dùá nʊ́, àdár nʊ́ è-n-túm è-n-twá Kwame do-Past Comp 3sg.subj-Pross-cut tree Fam cutlass Fam Perf-Neg-power Perf-Neg-cut ‘Kwame tried cutting the tree but the cutlass could not cut it’ b) Kwámɪ̀ yɛ́-ɛ̀ dɛ́ɛ̀ ó-rú-twà dùá nʊ́, àdár è-n-túm è-n-twá Kwame do-Past Comp 3sg.subj-Pross-cut tree Fam cutlass Perf-Neg-power Perf-Neg-cut ‘Kwame tried cutting the tree but a cutlass could not cut it.’ 52. a) Ǹsɔ́hwɛ́ ǹsʊ̀sʊ̀ándʊ́ nʊ́ mmó árá mà kyɪ̀rɛ̀kyɪ́rɛ̀nyí pènyín nʊ́ é-gyá-à àdzɪ̀-kyɪ̀rɛ́ Exams results Fam bad just Comp teacher elder Fam Perf-stop thing-teach ‘The results of the exams are so bad that the teacher has stopped teaching.’ b) Ǹsɔ́hwɛ́ ǹsʊ̀sʊ̀ándʊ́ nʊ́ mmó árá mà kyɪ̀rɛ̀kyɪ́rɛ̀nyí pènyín é-gyá-à àdzɪ̀-kyɪ̀rɛ́ Exams results Fam bad just Comp teacher elder Perf-stop thing-teach ‘The results of the exams are so bad that a teacher has stopped teaching.’ In all the sentences in (50) – (52), the interpretations of the highlighted DPs in the (a)’s are inferred from the preceding clauses, where their definiteness (in other words familiarity) comes about as a result of the presence of items mentioned in the preceding clause. The interpretations of the bare nouns in the (b)’s will be quantificational. For instance in (50b) the NP trainer could be any at all, not necessarily a definite one. The same goes for the other NPs in (51b) and (52b). The data discussed in (47-52) strongly support my claim that only the strong article in German can pattern with the Akan definite article nʊ́ and hence my claim that nʊ́ encodes familiarity. In this sub-section, I have argued that the Akan definite marker nʊ́ is equivalent to what is termed a strong definite article in the literature. I have compared data from German and Akan to show that where the German data demands the presence of the weak article, Akan uses the bare noun, and 83   where the strong article is used, Akan uses the definite article nʊ́. Also with the English data, it can be noted that where the NP is not familiar, Akan uses the bare noun. It is only the NPs with familiarity status that get nʊ́ attached to them. In the following sub-section I am going to discuss the demonstrative use of the determiner nʊ́. I will argue that the demonstrative interpretation of nʊ́ can be felicitous only when NPs are familiar. 4.6 Nʊ́ as a distal demonstrative In this section, I will discuss the function of nʊ́ as a distal demonstrative marker and show that the determiner nʊ́ can sometimes co-occur with the demonstrative morpheme dɛ́m to form a demonstrative determiner marker along the lines of Amfo (2007). I will however argue that the presence of the pre-nominal demonstrative marker is not what gives the NP the familiarity interpretation. This will contrast with Amfo's analysis, according to which dɛ́m encodes the familiarity status of the NP complement. I will show that dɛ́m only distinguishes the demonstrative use of nʊ́ from the definite article one. Also along the lines of Ajíbóyè (2007) I will say that the presence of the morpheme dɛ́m makes the NP complement salient. 4.6.1 Demonstratives in Akan Akan has two main demonstrative determiners: proximal yí and distal nʊ́. Both occur post-nominally (Amfo 2007). 53. a) Àtár yí yɛ̀ fɛ̀w Dress PDD be beautiful ‘This dress is beautiful’ 84   b) Àtár nʊ́ yɛ̀ fɛ̀w Dress DDD be beautiful ‘That dress is beautiful’ In (53a) the morpheme yí signifies the immediate presence of the entity denoted by the DP while in (53b) nʊ́ signifies that the entity denoted by the NP is relatively far away either in location or time from the interlocutors at the time the utterance is made. 4.6.2 Dɛ́m is not necessarily familiar Amfo (2007) rightly argues that the demonstrative interpretation of nʊ́ is not ‘part of its encoded semantic value’. This is because, while yí is a dedicated demonstrative determiner, nʊ́ is not. Amfo suggests that the distal demonstrative and the definite determiner can only be distinguished contextually; she goes further to argue that the differentiation between the distal demonstrative determiner and the definite marker can be achieved by attaching the distal demonstrative marker dɛ́m to the noun which is followed by the definite marker nʊ́. I will argue along the lines of Amfo that the use of nʊ́ as a demonstrative can be distinguished from its dedicated use as a definite marker when the demonstrative marker dɛ́m attaches in DP initial position. Consider the following context where two friends go to a shoe store to look for a pair of shoes to buy, and on their arrival at the shop they see a dress (one and only one) hanging in the shop. After leaving the shop one friend makes the utterance in (53b) above repeated here: 54. a) Àtár nʊ́ yɛ̀ fɛ̀w Dress DDD be beautiful ‘That/the dress is beautiful’ There is absolutely nothing here in the semantics that can prevent the interpretation of the NP as a plain definite, though the speaker might have an NP with a demonstrative interpretation in mind. The mechanism used to distinguish between the demonstrative determiner and the definite article, as 85   pointed out by Amfo, is by attaching the demonstrative morpheme dɛ́m to the noun making the distal demonstrative determiner as determiner dɛ́m….. nʊ́ in (54b), or by making a pointing gesture along with the utterance. b) Dɛ́m àtàŕ nʊ́ yɛ̀ fɛ̀w Dem dress DDD be beautiful ‘That dress is beautiful’ Dɛ́m…..nʊ́ can be used both with and without a pointing gesture; in both cases, the interpretation is that of a demonstrative. When the referent is new, dɛ́m….. nʊ́ is used alongside a pointing gesture when making the utterance, but when the referent is hearer old a pointing gesture is not necessary. This is equivalent to English stressed vs. unstressed that as in the following. 55. a) That DRESS is beautiful b) THAT dress is beautiful. Sentence (56a) is felicitous when the dress is hearer old, and sentence (55b) is the ostensive use. It is however necessary to make a deictic gesture when making the utterance in (55a) if the referent is hearer new. This mechanism may facilitate the demonstrative interpretation. Amfo argues that when dɛ́m is attached to N nʊ́, the cognitive status of the referent becomes familiar rather than just uniquely identifiable. This is illustrated in (56). 56. a) Mɪ̀-pɛ̀ (dɛ́m) àtàŕ nʊ́ 1sg.subj-like Dem cloth Fam ‘I like the (that) dress’ (Amfo 2007:146) However I have argued above in section 4.3. that nʊ́ by itself is always already familiar, therefore it cannot be the case that it is the dɛ́m which is providing the familiarity in (57). 86   It is worth noting that dɛ́m38 has other uses which could be interpreted as emphasis or as a degree word. In the following discourse dɛ́m is used for different functions: 57. A: Ɔ̀kyɪ́ná yɛ̀-bɔ́-kɔ́ hàbán mù Tomorrow 1pl.subj-Fut-go farm in ‘Tomorrow we will go to (the) farm’ B: Ì-sɪ̀ dɛ́n? Ɔ̀kyɪ́ná wɔ̀-bɔ́-kɔ́ hàbán mù à? 2s.subj-say what tomorrow 2pl.subj-Fut-go farm in QP ‘What are you saying? You will go to farm tomorrow?’ A: Nyèw yɛ́-yɛ́ èkwàfʊ́ yí dɛ́m árá na ɔ́-wɔ́ dɛ́ɛ̀ yɛ́-yɛ́ Yes 1pl.subj-do farmers this Dem just and it-has Comp 1pl.subj-do ‘Yes as farmers that is what we must do’ B: Nyèw dɛ́m árá Yes Dem just ‘Yes just like that/Yes that is so/Exactly so’ The above discourse shows that the function of dɛ́m is more of emphasizing something rather than encoding familiarity, so the familiarity of the sentences (56) and (57) above with the demonstrative determiner is not due to the morpheme dɛ́m but the definite marker nʊ́ which familiarity in its semantics. Along the lines of Ajíbóyè (2007), I propose here the status Salience for the morpheme dɛ́m.                                                              38 Amfo (2007) describes another use of dɛ́m as a manner demonstrative adverb. I set these uses aside here for reasons of space. 87   4.6.3 Dɛ́m makes the NP salient According to Ajíbóyè (2007: 170), “saliency implies significant and striking”. He argues that an NP complement becomes salient “when it is particularly important to the speaker” so the speaker draws the hearer’s attention to it. In the following example, the morpheme náà marks the NP ‘the house’ as salient (this data is from Yoruba, a Kwa language spoken mainly in Nigeria). 58) Ilé náà ni yàrá méje House Sal have room seven ‘The house has seven rooms’ (Ajíbóyè 2007:172) Ajíbóyè argues that náà does not mark an NP complement as definite. This is because Yoruba does not overtly mark definiteness. This is illustrated in his example cited here: 59) Olú gbé ọyá sínú àpó Olu put grass-cutter inside bag ‘Olu put the grass-cutter in the bag.’ (Ajíbóyè 2007:165) As can be seen in sentence (59), the morpheme náà is not present here; however the NP ọyá ‘grass- cutter’ is interpreted as definite in Yoruba unlike in Akan. This is to say that bare nouns39 are sometimes interpreted as definite NPs in Yoruba. In the following sentences the NP complements are interpreted as indefinite in (a), definite in (b) and definite salient in (c). 60. a) Tàkúté Olú mú ọyà Trap Olu hold grass-cutter ‘Olu’s trap caught a grass-cutter’                                                              39 Bare nouns can also be interpreted as non-specific indefinites in the language. This is however not discussed here. The main aim is to state that náà does not mark definiteness. 88   b) Olú gbé ọyà lọlé Olu carry grass-cutter go.house ‘Olu carried the grass-cutter home’ c) Ọ̀yà náà ni wọ́n fi jẹyán Grass-cutter Sal Foc 3pl use eat.pounded yam ‘The very grass-cutter served as meat with which they ate pounded yam’ (Ajíbóyè 2007:172) The equivalent of these sentences in Akan will be the following with (a) being indefinite, (b) definite and (c) definite and salient. Notice that (a) contains a bare noun, (b) a nʊ́-marked DP and (c) a dɛ́m…nʊ́-marked DP. 61. a) Kòfí n’éfír pá-à kràntsɪ́ Kofi Poss-machine caught grass-cutter ‘Kofi’s trap caught a grass-cutter’ b) Kòfí dzɪ̀ kràntsɪ́ nʊ́ bá-à fíé Kofi take grass-cutter Fam come-Past home ‘Kofi brought the grass-cutter home’ c) Dɛ́m kràntsɪ́ nʊ́ nà wɔ́-dzɪ́ yɛ́-ɛ̀ ǹkwán dzí-ì fùfú Dem grass-cutter Fam Foc 3pl.subj-take do-Past soup eat-Past fufu ‘It was that grass-cutter they used for soup to eat fufu’ Along the lines of Ajíbóyè, I will assume the status of Saliency for the morpheme dɛ́m, in that the morpheme makes a noun phrase (which is already definite in terms of my analysis) salient. This might be the reason why the morpheme cannot co-occur with indefinite NPs as below: 62) #Kòfí n’éfír pá-à dɛ́m kràntsɪ́/kràntsɪ́ bí Kofi Poss.machine caught Dem grass-cutter/grass-cutter Reef ‘Kofi’s trap caught a grass-cutter/a (certain) grass-cutter’ 89   The sentence here is unacceptable because an indefinite NP cannot be made salient. It is worth noting that Ajíbóyè (2007) states that the salience marker in Yoruba can co-occur with specific indefinites. It is however not clear why the salience marker is allowed with indefinites in some languages and not others. I leave this for future research. I have argued in this sub-section, following Amfo (2007), that Akan does not have a dedicated distal demonstrative so nʊ́ is used as a demonstrative. In most cases however users accompany the utterance with a deictic gesture or attach the pre-nominal demonstrative morpheme dɛ́m to distinguish it from its use as the definite article. 4.6.4 Conclusion In a nutshell, the Akan determiner nʊ́ has a strong familiarity semantics which is obligatory. The absence of the familiarity40 status of a nʊ́ DP where the hearer is concerned renders the sentence odd, or gives it a different interpretation. I have made comparison with data from other languages including German and English to support this point. Also, I have argued that the morpheme dɛ́m does not make the NP familiar as Amfo (2006, 2007) argues, but it makes it (the NP) salient in line with Ajíbóyè’s analysis of the Yoruba element náà. 4.7 Nʊ́ as a dependent clause marker This section concentrates on nʊ́ as a dependent clause marker. Here only the function will be mentioned with no detailed analysis. I leave the analysis of the clause-marking use for future research. Nʊ́ as discussed by Amfo (2006) bears a high tone and it can function as a dependent clause marker in temporal clauses, relative clauses and in clauses where two options are given and one is chosen over                                                              40 I am taking familiar to include inference bridging but not part-whole bridging as outlined in section 4.5.4.3 above, following Schwarz (2010). 90   the other. (Amfo terms this last case ‘a substitutive clause’). This nʊ́ is preceded by the dependent clause whose role is to give more information on the main clause. In the following sentences in (63), (a), (b), (c) are temporal, relative and substitutive clauses respectively. 63. a) Ǹsú tɔ́-ɪ̀ nʊ́, nnà má-àdà Water fall DCM and 1sg.subj-sleep ‘I was asleep when it rained’ b) M̀pɔ̀bʊ̀wá nʊ́ áà mʊ̀-tɔ́-ɪ̀ nʊ́ á-yɪ̀w Shoe Fam Rel 1sg-subj-buy-Past DCM Perf-miss ‘The shoe I bought is missing’ c) Yé-bé-bísà mà wɔ̀-áhám nʊ́, yɛ́-nká hɛ́n ánʊ̀ ńtúm 1pl.subj-Fut-ask Compl 3sg.subj-quarrel DCM 1pl-say our mouth close ‘Instead of asking her to upset her, let’s keep quiet’ The sentence in (63a) can be uttered in a context where the speaker wants the hearer to know the exact time at which the speaker’s action of sleeping took place, as well as the exact action she took at the time of raining. In (63b) the DCM marks the relative clause, which gives more information on the particular shoe which is missing, in this case the one the speaker bought. In (63c) on the other hand, a choice is made between asking a question and keeping quiet. The speaker urges the rest of the interlocutors to choose keeping quiet over asking the question in order to avoid any quarrel. The DCM is followed by the choice made. In the DCM use of the determiner, it marks a relative clause, a temporal clause, clauses where an option is made from two possible things or the whole clause as familiar. This multifunction of the determiner is not peculiar to only Akan. Lefebvre (1998) describes the [+definite] determiners in Fongbe (a Kwa language) and Haitian as being multifunctional heads. She argues that apart from the determiner’s use in the nominal domain, they can function as assertive markers and event determiners in the clausal domain. In all these uses the determiners mark the 91   things they scope over (be it a noun, an assertion or an event) as things that the interlocutors have knowledge about. The familiarity marker in Akan functions in a similar way as it can be a nominal marker or a clause marker. 4.8 Conclusion This chapter discussed the semantics of the Akan morpheme nʊ. It has shown all the functions the morpheme performs, including being pronominal (toneless), a determiner and a dependent clause marker (the last two marked with a high tone). I have also shown in this chapter that the morpheme as a determiner can be used in two ways; as a definite marker and as a distal demonstrative marker. I argued that the morpheme’s definiteness function is mostly like the German strong definite article and a subset of English definites. I have also discussed its use as a clausal marker, where the clause it occurs after is familiar. In conclusion, I have argued that, in the various functions of nʊ which correspond with certain tonal variations, there is only one semantic value which is familiarity. 92   5 Conclusion In this thesis, I have described the semantics of the Akan morphemes bi and nʊ. I have shown that the morphemes can be used as determiners as well as pronominals and these uses correlate with their tonal variations. I have argued that bi encodes specificity in the sense of Fodor and Sag and Kratzer, and nʊ encodes familiarity and patterns like the German strong articles as analyzed by Schwarz (2010). Chapters 1 and 2 gave some background on the Akan language and the structure of the elements that occur in the nominal domain and their co-occurrence restrictions. I provided phrase structure analyses of the surface representations of the nominal elements. In chapter 3, I discussed the semantics of the morpheme bi, and argued that the morpheme encodes specificity. I showed that bi determiner phrases are used when the speaker intends a specific referent, and that in most cases the bi-phrases are unable to take narrow scope. This behaviour is predicted by the choice function analysis proposed by Kratzer (1998), which is in the spirit of Fodor and Sag (1982) and as such, my findings provide cross-linguistic support for Kratzer's and Fodor and Sag's proposals. I also showed that the pronominal bi is an indefinite specific one, encoding specificity in Enç’s sense. In chapter 4, I discussed the semantics of the morpheme nʊ; I showed that the morpheme encodes familiarity. I provided evidence to show that nʊ as a determiner requires that the hearer be familiar with the referent of the noun phrase, and not merely that there be a unique referent as argued by Amfo (2006) and others. I compared the Akan data with data from German and argued that nʊ is most similar to the German strong articles analyzed by Schwarz (2010). Again, I showed that nʊ in its pronominal usage, though distinct in tonal behavior, nonetheless shares a core familiar semantics with the determiner use. I provided evidence that though the uses of the morpheme nʊ go beyond the determiner and the pronominal, in the other use there is also one semantic value the morpheme encodes which is familiarity. 93   In the course of this thesis, I have added to the discussion on the specificity of indefinites and the familiarity of definite descriptions. My research has provided cross-linguistic support for certain analyses available in the current literature, while shedding doubt on other potential analyses. I have also touched on several questions which deserve further investigation. For instance in chapter 2, I stated that Akan has a split system in its phrase structure, thus both head initial and head final orders exist in the language. Further research is needed in order for a unified account for the phrase structure in the language to be given. Also in chapters 3 and 4, we saw that both the morphemes bi and nʊ are used for determiners as well as pronominals. Both bi and nʊ determiners have high tones and their pronominal counterparts are toneless. One question that arises will be if this is accidental. There are two hypotheses that could be pursued. One is that the determiners and the pronouns are separate lexical entities (hypothesis A) and the other is that they are the same lexical entities (hypothesis B). Hypothesis A Hypothesis B DP NP DP NP D pronoun NP D noun bí /nʊ́ bi/nʊ noun/Ø(pro) bi/nʊ Elbourne (2005) argues that the English 3rd person pronouns are definite articles which have their overt nouns elided. This could be illustrated as follows: 1. a) [D [NP]] b) [D [e]]41 the N it hidden                                                              41 ‘e’ means elided or covert NP 94   Elbourne argues that when there is an overt noun, the is used but when the there is no overt noun it is used (see section 4.4.2 for more details). Applying this to Akan will mean that when there is an overt noun, the high tone nʊ́ which is the determiner is used but when there is no overt noun the toneless nʊ which is the pronoun is used. In English, there is a series of determiner elements called the th-series. As I mentioned in the introduction section of chapter 3, apart from the determiner and the pronominal forms of bi, it occurs in òbí ‘someone’. Could there be a bi-series so that when there is an overt noun the determiner (high tone) bí is used, but when there is no overt noun the pronominal (toneless) bi is used? Further research is required for these questions to be addressed. 95   References Abboh, Enoch, O. (1998). On the Syntax of Gungbe Noun Phrases. A paper presented at the annual meeting of the African Language and Linguistics Colloquium. Abbott, Barbara (1999) Support for a Unique Theory of Definite Descriptions. Proceedings of the Ninth Conference on Semantics and Linguistic Theory, Cornell University. Abbott, Barbara (2000). Presupposition as Nonassertions. Journal of Pragmatics Vol. 32:10, 1419-1437 Abbott, Barbara. (2001). Definiteness and Identification in English. Németh T. Enikö, (ed.), Pragmatics in 2000: Selected papers from the 7th International Pragmatics Conference. Vol. 2: 1-15. Antwerp: International Pragmatics Association. Abbott, Barbara (2003). Definiteness and Indefiniteness. In Horn, L. and Ward, G. (eds.), Handbook of Pragmatics. Blackwell. Abbott, Barbara. (2004). Definiteness and Indefiniteness. Handbook of Pragmatics. In Horn, L. and Ward, G. (eds.). 122-149. Blackwell, Oxford. Abusch, Dorit. (1994). The Scope of Indefinites. Natural Language Semantics.Vol.2.83-135. Academic Publishers. Netherlands. Adesola, Oluseye, P. (2005). Pronouns and Null Operators – A-bar Dependencies and Relations in Yoruba. PhD. Dissertation, Rutgers University. Adesola, Oluseye, P. (2006). A-bar Dependencies in the Yoruba Reference Tracking System. Lingua 116: 2068- 2106. Ajiboye, Oladiipo. (2007). Syntax and Semantics of Yoruba Nominal Expressions. M&J Grand Orbit Communications Ltd. Port Harcourt, Nigeria. 96   Amfo, Nana Aba Appiah and Fretheim, Thorstien. (2005). A Relevance-theoretic to Linguistics Signalling of the Cognitive Status of Discourse Referents. MilaVulchanova and Tor A. Afarli (eds.). Grammar and Beyong-Essay in honour of Lars Hellan. Amfo, Nana Aba Appiah (2006). Syntactic Variation versus Semantic Uniqueness. New Voice in Linguistics 211-223. Thue Vold E., Lyse G., and Müller A. (eds.).Cambridge Scholars Publishing. Amfo, Nana Aba Appiah. (2007). Akan Demonstratives. Selected Proceedings of the 37th Annual Conference on African Linguistics, 134-148. Doris L. Payne and Jaime Peña, (eds.). Somerville, MA: Cascadilla Proceedings Project. Amfo, Nana Aba Appiah. (2009). Indefiniteness Marking and Akan Bi. Journal of Pragmatics. Birner, Betty and Ward, Gregory. (1994). Uniqueness, Familiarity, and the Definite Article in English. Proceedings of the twentieth annual meeting of the Berkeley Linguistics Society, 93-102. Berkeley, CA: Berkeley Linguistics Society. Brisson, C. (1998). Distributivity, Maximality, and Floating Quantifiers. Ph. D. Dissertation. Rutgers University. Davis, Henry. (2010).Salish Languages Lack Generalized Quantifiers After All! (to appear) SALT XXI Proceedings. Dale, Robert. (2003). One-Anaphora and the Case of Discourse Driven Referring Expression Generation. Centre for Language and Technology, Macquarie University, Sydney. Déchaine, Rose-Marie. (2001a). On the Left Edge of Yorùbá Complements. Ms, University of British Columbia. Déchaine, Rose-Marie. (2001b). On the Left Edge of Yorùbá Complements. Lingua 111:81130. Diesing, Molly. (1992). Indefinites. MIT Press, Cambridge Massachusetts, London England. 97   Dolphyne, A. Florence. (1988). The Akan (Twi-Fante) Language: Its Sound System and Tonal Structure. Ghana Universities Press. Elbourne, Paul. (2001). E-type Anaphora as NP-deletion. Natural Language Semantics 9(3): 241–288. Elbourne, Paul. (2005). Situations and Individuals. MIT Press, Cambridge Massachusetts, London England. Enҫ, Mrüvet. (1991). The Semantics of Specificity. Linguistics Inquiry Vol.22, No.1:1-25.The MIT Press. Evans, Gareth. (1977). Pronouns, Quantifiers and Relative Clauses (I). Canadian Journal of Philosophy 7:467-536. Reprinted in Mark Platts (ed.), Reference, Truth and Reality: Essays on the Philosophy of Language. Routledge and Kegan Paul, 255-317. Farkas, Donka (1997). Dependent indefinites. In Corblin, F., Godard, D., and Maradin, J.-M., editors, Empirical Issues in Formal Syntax and Semantics, 243–267. Peter Lang. Farkas, Donka (2002). Specificity Distinction. Journal of Semantics 19, 213-243. Ferch, Elizabeth Ann. (2009). Number and the Scope of Indefinites. A Thesis submitted to the Department of Linguistics at the University of British Columbia, Vancouver. Fintel, Kai von. (1998). Evidence for Presupposition Indefinites. Ms., MIT. Massachusetts. Fodor, Janet Dean and Sag, Ivan A. (1982). Referential and Quantificational Indefinites. Linguistics and Philosophy. D. Reidel Publishing Co., Dordrecht, Holland, and Boston, U.S.A. 5:355-398. Fretheim, Thorstien and Amfo, Nana Aba Appiah. (2005). Reference, Determiner and Descriptive Content. Working Papers. ISK 2:87-112. 98   Gundel, K. Jeanette, Hedberg, Nancy, Zacharski, Ron. (1993). Cognitive Status and the Form of Referring Expressions in Discourse. Language Vol. 69 No.2: 274-307. Linguistics Society of America. Haspelmath, Martin. (1997). Indefinite Pronouns. Oxford Studies in Typology and Linguistics. Clarendon Press. Hawkins, John A. (1978). Definiteness and Indefiniteness. London: Croom Helm. Hawkins, John A. (1991). On (In)definite Articles: Implicatures and (Un)grammaticality Prediction. Journal of Linguistics 27:405–442. Heim, Irene. (1982). The Semantics of Definite and Indefinite Noun Phrases. Ph.D. Dissertation, University of Massachusetts at Amherst, Published in 1988 by Garland Pub. Heim, Irene. (1983). On the Projection Problem for Presuppositions. WCCFL 2: 114-125. Heim, Irene. (1990). E-Type Pronouns and Donkey Anaphora. Linguistics and Philosophy 13:137-177. Heim, Irene and Angelika Kratzer. (1998). Semantics in Generative Grammar. Oxford: Blackwell Publishers. Ionin, Tania. (2006). This is Definitely Specific: Specificity and Definiteness in Article Systems. Natural Language Semantics. Vol.14, No.2:175-234. Kadmon, Nirit. (1987). On Unique and Non-Unique Reference and Asymmetric Quantification. Ph.D. Dissertation, University of Massachusetts at Amherst, Published in 1992 by Garland Pub. Kadmon, Nirit. (1990). Uniqueness. Linguistics and Philosophy 13: 237-324. Kratzer, Angelika. (1998). Scope or Pseudo Scope? Are There Wide Scope Indefinites? In Events in grammar. Rothstein S. (ed.) 163– 196. Dordrecht: Kluwer. 99   Krifka, Manfred. (1984). Fokus, Topik, Syntaktische Struktur und Semantischem Interpretation. Ms., Universität Tübingen. Kripke, Saul (1977). Speaker's Reference and Semantic Reference. In French, P., Uehling, T., and Wettstein, H., (eds.) Perspectives in the Philosophy of Language, No.2 in Midwest Studies in Philosophy, 6–27. University of Minnesota Press. Lefebvre, Claire. (1998). Multifunctional and Variation among Grammars: the Case of the Determiner in Haitian and Fongbe. Journal of Pidgin and Creole Languages, Vol. 13, No. 1: 93-150(58). Lyons, Christopher. (1999). Definiteness. Cambridge Textbooks in Linguistics, Cambridge University Press. Matthewson, Lisa. (1999). On the Interpretation of Wide-Scope Indefinites. Natural Language Semantics 7(1), 79-134. Matthewson, Lisa. (2008). Pronouns, Presuppositions, and Semantic Variation. Proceedings of SALT XVIII. Osam, E. Kweku. (2004). An introduction to the Structure of Akan: Its Verbal and Multiverbal System. University of Ghana-Legon. Accra- Ghana. Prince, Ellen F. (1981) Toward a Taxonomy of Given/New Information, in Cole P. (ed.), Radical Pragmatics: 223–255. Academic Press, New York. Prince, Ellen F. (1988). The ZPG Letter: Subjects, Definiteness and Information Status. University of Pennsylvania. Pennsylvania. Postal, Paul. (1966). On So-Called “Pronouns” in English. In Dinneen F. (ed.), Report on the Seventeenth Annual Round Table Meeting on Linguistics and Languages Studies, Georgetown University Press, Washington, DC, 177-206. 100   Reinhart, T. (1997). Quantifier Scope: How Labor is Divided Between QR and Choice Function. Linguistics and Philosophy. 20: 335-397. Kluwer Academic Publishers. Netherlands. Roberts, Craige. (2003).Uniqueness in Definite Noun Phrases. Linguistics and Philosophy. Vol.26:287-350. Kluwer Academic Publishers, Netherlands. Roberts, Craige. (2005). Pronouns as Definites. In Reimer M. and Bezuidenhout A. (eds.) Description and Beyond. Oxford University Press. Romeo, Maribel.1997. A Problem for Choice Functions: Local Presupposition Contexts. In Dekker P., Stokhof M. and Venema Y. (eds.), Proceedings of the Eleventh Amsterdam Colloquium, 259-264. Amsterdam: ILLC/Department of Philosophy, University of Amsterdam. Russell, Bertrand 1905. On Denoting. Mind 14: 479-493. Saah K.Kofi. (1994). Studies in Akan Syntax. PhD. Dissertation. University of Ottawa, Canada. Schmitt, Cristina and Munn, Alan. (1999) Against the Nominal Mapping Parameter: Bare Nouns in Brazilian Portuguese. NELS 29 Proceedings. Schwarz, Florian. (2010). Two Types of Definites in Natural Language. PhD. Dissertation. University of Massachusetts, Amherst. Szabo, Zoltan. (2000). Descriptions and Uniqueness. Philosophical Studies 101:29-57. Szabo, Zoltan. (2003). Definite Descriptions Without Uniqueness: A Reply to Abbott. Philosophical Studies 114:279-291. Ward, Gregory L. and Prince, Ellen F. (1991). On the topicalization of Indefinite NPs. Journal of Pragmatics Vol.16, 2:167-177.  "@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "2011-11"@en ; edm:isShownAt "10.14288/1.0071866"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Linguistics"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "Attribution-NonCommercial-NoDerivatives 4.0 International"@en ; ns0:rightsURI "http://creativecommons.org/licenses/by-nc-nd/4.0/"@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Semantics of Akan bi and nʊ"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/35083"@en .