On the reconstruction of person marking in Proto-Kwa

30
1 ON THE RECONSTRUCTION OF PERSON MARKING IN PROTO-KWA Kirill V. Babaev Center of Comparative Studies, Russian State University for the Humanities, Moscow [email protected] This paper aims at conducting a comparative analysis of personal pronouns and bound verbal and nominal person markers in Kwa languages and to propose a hypothesis on the reconstruction of the system of person marking in Proto-Kwa. This will cover both the personal pronouns, and verbal and nominal person affixes. Existing hypotheses will be verified and new ones will be proposed for the ways of transformation of the original person marking system into the systems of the present-day languages of the family. Cet article essaie de conduire une analyse comparative des pronoms personels est les marques des personnes fixes des langues kwa et de proposer une hypothèse de la reconstruction de la système des marques personelles dans la langue proto-kwa. Ce va couvrir tous les deux les pronoms personnels independants, et les marques de personne nominales et verbales. Les hypothèses existant vont être vérifiées, et les idées nouvelles vont être proposées pour les routes de transformation du système originelle dans les systèmes des langues de la famille du present. 0. Introduction This paper is not at all an ultimate reconstruction of the proto-language system of personal pronouns, rather just a first step towards it which it yet to be made. Reaching the objective of the paper is also an important step in the complex work on reconstructing the Niger-Congo (NC) system of person marking. In the recent decades, successful tentative reconstructions have been made for a number of family proto- languages within NC: Kordofanian [Schadeberg 1981], Adamaua [Boyd 1989], Ijoid [Williamson, manuscript], Ubangi [Moniño 1995], Gur [Miehe 2004], and Atlantic [Pozdniakov & Segerer 2004]. An attempt of reconstructing person marking for Proto- Benue-Congo (PBC) was recently undertaken by the author of the present paper [Babaev 2008; forthcoming]. However, such a reconstruction is yet to appear for the Kwa languages. Since Kwa is genetically the closest relative for the BC family, the reconstruction of this subsystem of morphology can raise the research to a higher level. The Kwa family of languages (name proposed by Krause [1895]) includes nearly eighty languages (79, according to the latest edition of Ethnologue [Lewis 2009]) which are spoken in southern West Africa, along the northern coast of the Guinean Bay. Kwa speakers populate the eastern part of Cote d’Ivoire (mostly to the east of the Bandama river), southern and central areas of Ghana, they are also widespread in Togo and Benin, with their western boundary approximately coinciding with the Benin - Nigerian state border. To the west, Kwa speakers neighbour the areas populated by Kru- and Mande-speaking populations, to the north it is the Gur family, and the BC family to the east. All the above language stocks are branches of the NC or Niger- Kordofanian macrofamily as classified by Joseph Greenberg [1955]. The external and internal classification of the Kwa languages is still an issue of discussion in African linguistics. Upon the classification by Diedrich Westermann [1927], the family was divided into eight groups to include a great number of languages some of which were later excluded from Kwa to separate language stocks, namely the language families of Ijoid and Kru languages. Besides, it was proposed in the 70s of the 20th century to regard the eastern Kwa languages as belonging in fact to the BC family [Bennett & Sterk 1977], the idea which was welcomed by scholars [Williamson 1989]. A new term ‘Benue-Kwa’ was generated (called also ‘East Volta- Congo’ by some) to characterise a new taxonomic level within the NC macrofamily

Transcript of On the reconstruction of person marking in Proto-Kwa

1

ON THE RECONSTRUCTION OF PERSON MARKING IN PROTO-KWA

Kirill V. Babaev Center of Comparative Studies, Russian State University for the Humanities, Moscow

[email protected]

This paper aims at conducting a comparative analysis of personal pronouns and bound verbal and nominal person markers in Kwa languages and to propose a hypothesis on the reconstruction of the system of person marking in Proto-Kwa. This will cover both the personal pronouns, and verbal and nominal person affixes. Existing hypotheses will be verified and new ones will be proposed for the ways of transformation of the original person marking system into the systems of the present-day languages of the family. Cet article essaie de conduire une analyse comparative des pronoms personels est les marques des personnes fixes des langues kwa et de proposer une hypothèse de la reconstruction de la système des marques personelles dans la langue proto-kwa. Ce va couvrir tous les deux les pronoms personnels independants, et les marques de personne nominales et verbales. Les hypothèses existant vont être vérifiées, et les idées nouvelles vont être proposées pour les routes de transformation du système originelle dans les systèmes des langues de la famille du present.

0. Introduction

This paper is not at all an ultimate reconstruction of the proto-language system of personal pronouns, rather just a first step towards it which it yet to be made. Reaching the objective of the paper is also an important step in the complex work on reconstructing the Niger-Congo (NC) system of person marking. In the recent decades, successful tentative reconstructions have been made for a number of family proto-languages within NC: Kordofanian [Schadeberg 1981], Adamaua [Boyd 1989], Ijoid [Williamson, manuscript], Ubangi [Moniño 1995], Gur [Miehe 2004], and Atlantic [Pozdniakov & Segerer 2004]. An attempt of reconstructing person marking for Proto-Benue-Congo (PBC) was recently undertaken by the author of the present paper [Babaev 2008; forthcoming]. However, such a reconstruction is yet to appear for the Kwa languages. Since Kwa is genetically the closest relative for the BC family, the reconstruction of this subsystem of morphology can raise the research to a higher level. The Kwa family of languages (name proposed by Krause [1895]) includes nearly eighty languages (79, according to the latest edition of Ethnologue [Lewis 2009]) which are spoken in southern West Africa, along the northern coast of the Guinean Bay. Kwa speakers populate the eastern part of Cote d’Ivoire (mostly to the east of the Bandama river), southern and central areas of Ghana, they are also widespread in Togo and Benin, with their western boundary approximately coinciding with the Benin - Nigerian state border. To the west, Kwa speakers neighbour the areas populated by Kru- and Mande-speaking populations, to the north it is the Gur family, and the BC family to the east. All the above language stocks are branches of the NC or Niger-Kordofanian macrofamily as classified by Joseph Greenberg [1955]. The external and internal classification of the Kwa languages is still an issue of discussion in African linguistics. Upon the classification by Diedrich Westermann [1927], the family was divided into eight groups to include a great number of languages some of which were later excluded from Kwa to separate language stocks, namely the language families of Ijoid and Kru languages. Besides, it was proposed in the 70s of the 20th century to regard the eastern Kwa languages as belonging in fact to the BC family [Bennett & Sterk 1977], the idea which was welcomed by scholars [Williamson 1989]. A new term ‘Benue-Kwa’ was generated (called also ‘East Volta-Congo’ by some) to characterise a new taxonomic level within the NC macrofamily

2

unifying two geographically and genetically close families: Benue-Congo and Kwa. Therefore, ‘New Kwa’ appeared as the name for those of ‘Old Kwa’ languages left after this reclassification. These languages, formerly known as Western Kwa, are exactly the subject of the present paper. Nevertheless, the border between BC and Kwa is still not quite exact according to the latest research, and this allows to consider Benue-Kwa as the dialect continuum with a taxonomic division which is not yet clear enough [Williamson & Blench 2000: 17-18]. The internal classification of the Kwa language family (Pic. 1, groups and subgroups are marked bold) is based on [Stewart 1989; 2001a] and its modification in [Williamson & Blench 2000]. It was mainly repeated in the latest language tree published by Guillaume Segerer [2008].

(Pic. 1)

It is necessary to admit, however, that the given classification cannot be treated as a final one for multiple reasons including the fact that most of it is based on just a few lexical isoglosses. For example, the Nyo branch is postulated mainly according to the presence of the item ɲo ‘two’ in a number of Kwa languages [Bennett & Sterk 1977]. Similarly, the LB group is a mostly geographical unity rather than genetic. The same kind of unit is the so called ‘Central Togo’ language stock defined as such by Heine [1968]. According to him, it comprises two large branches: Ka-Togo (groups Avatime-Nyangbo, Kebu-Animere and Kposo in Pic. 1) and Na-Togo (groups Lelemi, Basila-Adele and the Logba languages). This binary classification is based on a number of lexical isoglosses such as the word ‘flesh’ used in the names of the branches. The Central Togo language unity was later supported by Blench [2001]. The genetic relationship of the Kwa languages are not studied deeply enough: there is still no reliable system of phonetic or lexical correspondences between all the branched, many taxonomic conclusions are built on the typological data rather than historical comparison. A number of languages are still not properly recorded to judge on their external genetic links.

3

There are more or less solid comparative evidence for a number of groups which have properly defined genetic shape and borders. Among these are the Potou-Tano languages which account for over a half of all Kwa languages and were carefully studied in the recent decades. The Avikam-Alladian languages [Bôle-Richard & Lavage 1983], the the Ga-Dangme languages and especially the Gbe group [Capo 1991; Stewart 1994], are reliably grouped as well as the Akan dialect cluster and some other small stocks; however, the interrelationship of these groups are not that well studied. Some basic characteristic features of the phonetic system of the Kwa languages include the vowel harmony based on the ±ATR (advanced tongue root) criteria, and the opposition of fortis and lenis consonants. Another interesting trait is the complementary distribution of plain and nasal consonant phonemes. All these features are quite widespread and should be traced back to the proto-languages [Stewart 1989: 231]. However, the above are by no means uniform within the family: e.g., the ATR feature does not apply to Gbe and Ga-Dangme languages, and the same goes for prefixed nominal classes. It is beyond the scope of the present paper, though, to go into details. The morphology of Kwa is also characterised by a number of common phenomena. The system of nominal classes is represented by class prefixes which demonstrate different levels of preservation across the family. Verbal extensions are quite common, denoting a variety of morphological categories, from reflectivity to modality. The word order is mostly SVO; attributes follow the noun. Nominal postpositions are widespread [Williamson & Blench 2000: 30]. There are no monographs devoted to the reconstruction of the Proto-Kwa language stage. Numerous articles on reconstructing the phonetic inventory, phonology and the tonal system of Proto-Potou-Tano were published by Stewart [1973, 1993, 1999, 2000, 2001b, 2002, 2004], who compared this proto-tongue with that of Bantu in an attempt to use the resulting Benue-Kwa proto-language1

We used data from about fifty languages representing all major branches and subgroups of Kwa, and including tongues which are considered isolated. Multiple dialect forms are given for some large languages (e.g., Akan), these are using a dash:

as a model for the analysis of Proto-NC. Stewart also published over a hundred of lexical reconstructions [Stewart 2002: 214-223, 2004], including two person markers. Another list of tentative lexical correspondences was published by Blench [2001], which does not however contain any proposed reconstructed proto-language forms. The comparative and historical analysis of personal pronouns in the Kwa languages as a complex morphological system is conducted here for the first time. It will be based on the data provided by numerous articles, monographs and dissertation papers (some of which are still unpublished) devoted to the synchronic description of certain languages and dialects, their morphology and syntax. Dictionaries, vocabularies, glossaries and wordlists gathered and (partly) published in the 19th and the 20th century as well as the present decade will also be used. Some forms are extracted from the extensive online database collected by Guillaume Segerer [Segerer 2002-2007]; all such forms were verified in the source papers.

1 He called it Proto-Potou-Akanic-Bantu (PPAB).

4

Akan-Fanti, Akan-Akwapim, Ewe-Adangbe, etc. Where the forms of two or more sources contradict each other, all of them are given with the respective footnote references. The Kwa languages are mostly analytic in their morphological structure. Therefore, the system of person markers is represented in most languages by independent or clitic personal pronouns which mark the subject and the object of the verb without being strictly bound to it. This has to be examined carefully: at least some languages (e.g., Sekpele, Logba, Lelemi etc.) use bound prefix markers, and in some it is more the matter of orthography rather than the language structure. The most common situation is the use of several series of subject pronouns: verbal (these can be prefixed or serve as clitics), independent or nominal (called also emphatic or ‘stressed’) and, more rarely, reflexive. Personal pronouns in Kwa have cumulative meanings of both person and number within one form, as well as other NC languages show. The pronouns of the 1st and 2nd person constitute are different from the 3rd person by origin: it is easy to prove that the latter are derived from nominal class prefixes. This phenomenon is typical for both the neghbouring BC languages [Babaev 2008] and the NC language family as a whole. Typologically, this kind of opposition is observed in the majority of the world’s languages [Бабаев 2009], including the Indo-European tongues where the 3rd person pronouns originated from the demonstrative ones. A number of linguists therefore treat the 3rd person as ‘non-person’ [Benveniste 1971] and exclude it from the paradigm of personal pronouns in diachronic perspective [Creissels 1993]. Though it seems advisable to draw a line between them in the present paper as well, we find it useful to give a full chart of the third person pronouns for the reference as an Appendix, for the sake of comparison which is in many cases quite valuable. We will also attract the data of the 3rd person marking in the course of the paper in order to demonstrate various diachronic structure processes such as analogy or paradigmatic levelling. It is known, for instance, that in Kwa, as well as in BC, the 2nd person plural markers are often replaced by the 3rd person ones [Delafosse 1904: 29], which will significantly complicate the reconstruction of the former (cf. below). Some Kwa languages demonstrate extended plurality: inclusive and exclusive pronouns function in the 1st person plural (e.g., in Anyi, Baulé, Aburé, Adioukru). Pronouns in Kwa can accept some morphological categories of the adjacent verb: they may vary in tense (Adioukru ya ‘thou’ (Present) vs. i ‘thou’ (Past)) or aspect (Adele tí ‘we’ (Perfect)). Sometimes these categories are only marked by the tone (e.g., in Abron and Anyi for perfective and imperfective forms). In some tongues, affirmative pronominal forms are opposed to negative and sometimes also interrogative (e.g., in Sele, Abron). In Adioukru, there is an opposition between pronoun in direct and indirect speech [Hill 1995]. All these characteristics are not universally adopted among the Kwa languages, and we will analyse below whether they can be derived from the proto-language or not. For presentation of data, we here accept a model of four series of personal pronouns, which is traditionally used under the influence of Bantu linguistics: subject, object, possessive and emphatic (or independent) pronouns. This model does not obviously reflect the reality in the Kwa languages correctly: since, for instance, in the majority of

5

tongues the subject markers are syntactically independent, they coincide in shape with independent / emphatic pronouns. However, the traditional Bantu-style model is easier for visual analysis. Languages in the comparative charts below are sorted according to the scheme of Pic.1. Names of groups and, where necessary, subgroups are given in brackets in the first chart and omitted in the next ones to avoid complexity. Forms divided by commas mark morphological modification (indicated in brackets following them) or are taken from different sources: some of the latter can represent different dialects, but may also be a result of incorrect recording. Forms divided by a slash sign denote phonetic modifications such as the vowel harmony. Those pronouns or markers which change their tone according to the verb meaning will be italicised.

1. The 1st person singular (Chart 1)

subject object possessive emphatic Ewe (Gbe)2

me, ma (juss.) m ɲe ɲe Ewe-Adangbe (Gbe)3 mē ŋ -ɲē, àʄē- ēɲē

Gen (Gbe)4mu mu ape , ɲɛ eɲɛ, ɲɛ

Gun (Gbe)5(u)n -mi cé ɲε

Fon (Gbe)6un ɲε , mì ce, ɲε ɲε

Nyangbo-Tafi (Avatime-Nyangbo)7

é / ε (pres., fut., aor.), mε (perf., hab.), í

mε mε / m- mεšú (refl.)

Avatime (Avatime-Nyangbo)8 me / ma

mē / mε

mε / m- mé

Bowili (Kposo)9

m, mí, ɲie дат.п.

yĩ n-

Igo (Kposo)10 ɔmī

Kebu (Kebu-Animere)11 mə

Animere (Kebu-Animere)12 mʊ

Abé (Agneby)13 mə mə

2 [Schadeberg 1985; Agbedor 1996] 3 [Sprigge 1976] 4 [Bôle-Richard 1983c] 5 [Aboh 1997: 31] 6 [Lefebvre & Brousseau 2002] 7 [Funke 1910: 173-176] 8 [Funke 1909; Kropp Dakubu 1967: 59-60; Schuh 1995, 1995a] 9 [Westermann 1922: 28; Starwalt 2008: 29] 10 [Stewart, manuscript: nos. 585-588] 11 [Blench 2001: 35] 12 [Casali 2006] 13 [N’Guessan 1983]

6

subject object possessive emphatic Abidji (Agneby)14

m- -mú ɛmɛ Adioukru (Agneby)15 ma / m- (pres.), me (past) èm / am / εm ɛ m, mes,

me, mi Alladian (Avikam-Alladian)16

mɛ / me , m (+V), ni (rare) me mɛ

Avikam (Avikam-Alladian)17 m- / me, ma (aff. only) -m m

amɛ, mi, me

Ga (Ga-Dangme)18N- (pres.), mí- (hab., subj., aor.), ma- (fut.), mm- (prog.), mì

mĩ mĩ mi

Dangme (Ga-Dangme)19 i mĩ yée ámĩ

Adele (Potou-Tano, Basila-Adele)20

n, perf. na, fut. na mé/me (neg.)

me me

Ega (PT)21nɩ nɔ n- ɩnɔ

Logba (PT)22m- / n- / n - / un- -m amu- amú

Sekpele (PT, Lelemi)23 ni- / n-, me- / m-, N- mɛ mɔ / mɔ mɔ / mɔ

Sele (PT, Lelemi)24n- (hab. aff., past neg.), ko (prog. aff.), noo (prog./fut. neg.), nin (neg.), le / la (past)

mi nıi a mi

Siwu (PT, Lelemi)25 lo -, le / lo mɛ ū m mɛɛ

Lelemi (PT, Lelemi)26 mɔ- / N-, lε- / li-, nε- / ni-

-mì, mi

múna ami, ami

Attié27

mε mε ,

mε mε / mε mε / mε

14 [Tresbarats 1983] 15 [Delafosse 1904: 53; Hérault 1976; Hill 1995] 16 [Delafosse 1904: 53; Duponchel & Bertin Mel 1983] 17 [Delafosse 1904: 53; Segerer 2002-2007] 18 [Kropp Dakubu 1976, 2008] 19 [Apronti 1976] 20 [Westermann 1922: 49] 21 [Bôle-Richard 1983b] 22 [Westermann 1903: 29] 23 [Westermann 1922: 8; Allan 1980] 24 [Allen 1980] 25 [Westermann 1922; Iddah 1980] 26 [Kropp Dakubu 1967: 7; Höftmann 1971; Allan 1976] 27 [Delafosse 1904: 28; Kutsch, Connie & Hood 1983]

7

subject object possessive emphatic Ebrié (PT, Potou)28

n / me mi,

me mɛ

Mbato (PT, Potou)29 n / uñ, mã mó n , mi, me mē, mõ

Anyi (PT, C.Tano, N.Bia)30 mé / N-, mi mé mi, mé / m-

mĩ, mé,momi

Anyi-Sanvi (PT, C.Tano, N.Bia)31 mĩ / N, mi (fut.) mĩ mi, me mĩ, me

Baulé (PT, C.Tano, N.Bia)32

me, m- (+V), N- (+C), mi (fut.)

mĩ (+V)

n+C,mĩ+V mĩ, mí

Nzema (PT, C.Tano, S.Bia)33 mi- / mɪ-, me me mi

Abron (PT, C.Tano, Akan)34 m / me, mĩ (fut.) m, mì m, mi, me mĩ

Akan (PT, C.Tano, Akan)35 me- / mi- / ma- me me m(e) / mi

Akan-Fanti36mi- / ma- -mi mí- / ma- ími

Gichode (PT, Tano, N.Guang)37 nɛ / nã , N, mi me mĩ / me / m’ mĩ

Ginyanga (PT, Tano, N.Guang)38 nɛ / n’, N mĩ mĩ / me / m’ mĩ

Dompo (PT, Tano, N.Guang)39 mi

Gonja (PT, Tano, N.Guang)40 mé, N ma ma / m- ma

Nkonya (PT, Tano, N.Guang)41 no- / nɔ- / ne- / nε-, m- me me me

Nchumuru-Banda (PT, Tano, N.Guang)42

m-

m / mʊ

Nchumuru-Bejamso-Grubi

me / mi (non-past aff.), ø- (non-past neg.), mε (past

me mε / me, N (inalien.), m-

me

28 [Delafosse 1904: 28; Bôle-Richard 1983a] 29 [Delafosse 1904: 28; Grassias & Bôle-Richard 1983] 30 [Delafosse 1900: 50-54; Pyne 1976] 31 [Delafosse 1904: 122; Burmeister 1983b] 32 [Delafosse 1904: 122; Timyan-Ravenhill 1980; N’Guessan 1983] 33 [Delafosse 1904: 122; Clopper 2001: 2] 34 [Delafosse 1904: 122; Timyan-Ravenhill 1983] 35 [Christaller 1933] 36 [Welmers 1946: 33-34] 37 [Cleal 1976] 38 [Cleal 1976a] 39 [Blench 2007] 40 [Painter 1980a] 41 [Reineke 1972] 42 [Cleal & Price 1980]

8

subject object possessive emphatic (PT, Tano, N.Guang)43

neg.), N- (cond.)

(+V)

Cherepon (PT, Tano, S.Guang)44 mɩ mɩ mɩ mɩ

Aburé (PT, W.Tano)45 m mí mí, (me, m’)

me, mɩ / mι

Eotilé (PT, W.Tano)46 N-, me mɛ me, m’, mi mɩɛ / mi

Krobu (PT, Tano)47 mɛ, N mɛ mɛ It is extremely complicated (and not always possible) to draw a line between the pronominal element itself and a topic / tense-aspect or other predicative marker. Therefore, we have to present data as precise as it is used in the language, with an opportunity for further analysis. From the chart above, one can extract two clusters of pronominal forms most widespread across the family for marking the subject: 1) me / mi / ma / mu / m- / N-; 2) ne / ɲε / ni / na / no / n-. The N symbol marks the so called ‘homorganic’ nasal frequently met across the languages of the region: its articulation depends on the following sound, so it becomes m- before labials, n- before dentals and ŋ- before velar consonants. Still, in Kwa this phoneme sounds as m- when followed by a vowel, we can suppose that it was originally derived from the m-group of forms. The ma pronoun is also conditioned by some specific phonetic environment. Otherwise, the unmarked form is either mi or me / mε. In a number of languages (as in Avikam) ma denotes morphologically marked forms of polarity, tense or aspect: this phenomenon is characteristic also for BC languages, where the ma 1st person singular pronoun usually marks the recent past or the perfect (e.g., in Cross River languages, cf. [Babaev, in press]). The -a- affix is a standard perfect marker in a lot of Kwa languages too [Timyan-Ravenhill 1980]. The Lelemi prefixed mɔ- emerged by analogy with the 1st person pl. bɔ-. Single consonant forms m- result from the reduction of the vowel in the marker. Therefore, we can conclude that the mi and me / mε are the main candidates for the proto-language pronoun. Mi contains the vowel which in many Kwa languages (and beyond, cf. Bantoid) is marking the unmarked subject in general, e.g. in Dangme [Kropp Dakubu 1987], in most cases the present or future tenses. The two variants of vocalism do not demonstrate any geographical distribution: the only interesting feature is in Central Tano where mi marks the future, while me is unmarked.

43 [Cleal & Price 1980] 44 [Painter 1980b] 45 [Delafosse 1904: 28; Burmeister 1983a] 46 [Delafosse 1904: 28; Hérault 1983] 47 [Mensah 1983]

9

Let us now switch to the n- / nV48

The object series does not give us a chance to extract any specific markers different from the subject forms. The dominating pronoun is the same mi / me / mε as well as the reduced variant m- which appears in those tongues which mark object by a bound affix.

group of pronouns. These forms are discovered in all major branches of Kwa and thus cannot be treated as an innovation of a certain group. It is noteworthy that the n-forms, as opposed to pronouns in m-, are mostly used in the subject series and are very rarely used for marking object, possession or emphasis. Such rare examples may witness the analogical spread of n-forms from the subject to other series, the process happened in certain tongues: Ega, Bowili and the Gbe group of languages. The Sele pronoun nīi contains a possessive marker -n- / -l- which is used in all person / number slots. Bearing this in mind, it is possible to call n-forms subject markers by origin. Moreover, this is confirmed by the data of external comparison with the BC languages, including Bantu where the reflexes of the proto-language*ɲι- are mostly subject prefixes, while the other three pronoun series are built on the basis of the non-subject independent pronoun *me. This binary opposition must go back to Proto-Bantoid [Babaev 2008]. In Kwa, the vowel of the proto-language *nV / *nV- pronoun is by no means clear, after complicated phonetic mutations and vowel harmony which have definitely led to a number of vocalisation variants. Analogical forms are also sometimes created across the paradigm: e.g., in Nkonya the 1st person sg. pronouns no- / nɔ- / ne- / nε- share vocalism with the 2nd person sg. forms fo- / fɔ- / fe- / fε-. However, it seems that a front vowel should be reconstructed for the proto-language: the subject pronoun must have been *nI. The Gbe languages show *ɲ which is the result of the front vowel influence. To conclude, we can reconstruct two subject markers for Proto-Kwa: *nI and *mi / *me / *mε, of which the former was limited to marking the verb subject. The latter can then be defined as ‘non-subject’: the m-form used to mark nominal predicates, object, possessive and emphatic forms and was later projected to subject forms by analogy. The original distribution or its traces have nevertheless been preserved in a number of tongues (e.g., in N.Guang). Typologically this situation is quite similar to the one we observe in Indo-European where the reflexes of the 1st person sg. pronoun *me, which used to mark indirect case forms and possession in Proto-Indo-European, start to squeeze out the old subject *eg’h(om / ō) pronoun (in Celtic, Iranian, Indo-Aryan languages). Where the binary opposition is preserved, it fuinctions in a similar way with what we have just seen in Kwa: cf. French je (verb subject) vs. me (direct object), mon (possessive), moi (indirect object / nominal predicate / emphasis). It is hardly possible to recover the original vocalism of *mi / *me by means of comparison. It should be noted, however, that the mi pronoun does not act together with any n-forms in the subject slot. Where they coexist, it is rather me / mε. But for Proto-Kwa, we still have to reconstruct the independent pronoun as *me / *mε / *mi. These variants correspond to Proto-BC [Babaev, in press]. Stewart [2002: 222] also postulates *mī / *me for Benue-Kwa.

48 It is possible to refer le / li / lo pronouns in Lelemi languages to the reflexes of original *nV: this is a regular phonetic change in Lelemi.

10

Sometimes a tone variation is observed between the subject and the object, but it is not systematic. For the object proto-form, we reconstruct *me / *mε identical to the independent form we discussed above. Therefore, the object did not constitute a separate series in Proto-Kwa. The possessive pronoun must have looked the same: across Kwa, it is overwhelmingly identical with the independent pronoun. Their only difference lies in syntax: the possessive meaning acts in the preposition to the noun: Ewe ɲè ‘exactly me’, ɲè ágɔ ‘my coconut’ [Schadeberg 1985: 16-18]. The possessive pronoun can contain an extension affix which is spread to the whole paradigm, as in Lelemi 1st sg. múná, 2nd sg. fúlá, 1st pl. búlá, 2nd pl. míná. In Gen and Ewe-Adangbe, this possessive marker is prefixed: apê ‘my’, àpo ‘thy’. Such constructions are widely known across Niger-Congo and are sometimes explained as the combinations of frozen noun class markers adjusted to both the noun and all its attributes including possessives, cf. in the Esimbi languages (Bantoid family):

fì-ŋkōbī f-ā f-yā 19.CL-statue 19.CL-my 19.CL-DEF ‘my statue’ [Tamanji 2006].

The pronoun in this construction was an independent one, which is yet another confirmation of the reason for the material resemblance between independent and possessive pronouns: in the proto-language (Proto-Kwa or even Proto-Benue-Kwa) there was no distinct class of possessive pronouns, and independent forms functioned instead to mark possession. Finally, let us analyse the emphatic (independent) personal pronouns of the 1st person singular. Syntactically, the meaning of these forms across the family is similar: they are used for emphasising the subject (in phrases like ‘It was me who broke the window’) or where the pronoun acts as a nominal predicate (‘This is me’). It is possible to note three basic morphological models for building the emphatic pronoun in Kwa:

1) the pronoun is identical to the non-subject form going back to *me / *mε / *mi; 2) «V- + non-subject form», i.e. *Vme / *Vmi: Igo ɔmī, Abidji εmε, Avikam ámɛ, Logba amú, Lelemi ámi, Akan-Fanti ími; 3) «non-subject form + extension»: Akan-Akwapim mi-ara (where ara is a nominal topic marker and intensifier), Siwu mmɛɛ etc.

The third type carries the most evident origins: suffixed extensions add specific meanings to the older independent pronoun. Sometimes this is an emphatic meaning, but also reflexive (as in Nyangbo-Tafi mešú, which may be related to Adioukru mes). Such extensions act throughout the whole paradigm of persons / numbers, so there is no doubt about their original meaning. They are, however, later innovations and never go back to Proto-Kwa. The second type has a remarkable parallel in BC languages, including Bantu and Bantoid, where we reconstruct the 1st sg. independent non-subject pronoun as

11

*Vme. There are several hypotheses on the origins of the prefixed vowel: it is sometimes traced back to the demonstrative particle *γa- [Meinhof 1906: 35; Werner 1919: 86-87]. It is also called ‘associated marker’ which is rather widespread across BC. It is quite logical to suggest that the proto-language form in Kwa was not only built according to the same principle, but can be reconstructed for Proto-Benue-Kwa as *Vme / *Vmε / *Vmi, generalised as *VmI where the initial vowel is back, and the final one is front. The analysis conducted above leads us to the reconstruction of the following three forms of personal markers for the 1st person singular in Proto-Kwa. They were, however, based on two pronominal roots: *nI- chronologically the most ancient subject verb prefix; *me / *mε / *mi the independent (non-subject) pronoun which started to shift

to mark the verb subject already in the proto-language; *Vme / *Vmε / *Vmi the independent emphatic pronoun. Unfortunately, at this stage we have to limit our reconstruction to the phonetic shape without marking the prosodic structure. Some conclusion can obviously be made on the tone of the proto-language forms, but a much deeper research is needed in order to do it.

2. The 2nd person singular Pronominal forms for the 2nd person sg. are given in Chart 2 below.

(Chart 2)

subject object possessive emphatic Ewe e, ne (dep.) wo wo wo Ewe-Adangbe ē wō àʄē-, -wō ēwō Gen o o apo, o eo Gun a -we -towe jε Fon a we, hwε tówé, hwε hwε Nyangbo-Tafi wɔ / wo wɔ / wo wɔ / w- wɔšú (refl.) Avatime wo / wɔ wō / wɔ wɔ / w- wɔ Bowili ō- / ɔ-, no Igo ɔwε Kebu lə Animere wo Abé fə fə Abidji f- -fú ɛfɛ Adioukru ø-, ya (pres.), i / e (past) -ɲ ɲi, ɲ ŋ, i Alladian ɛ / e / a (neg.),

wo / w’ (aff.)

vé (topic.)

vē, e vè, vé (topic.), e

12

subject object possessive emphatic Avikam a- -a a aɣɛ, a Ga ó- (perf., subj.), o- (aor.),

oo- (prog.) bò / ò ó bò

Adele a, a (fut.) ai awo Ega ɔ ɣɔ ɣ- ɩɣɔ Logba a- -u au- aú Sekpele a-, e- fœ, fe fé / fē fē / fé Sele fan- (hab. aff.), fa- (past

neg.), foo- (prog. aff.), fo- / fɔ- (neg.) и др.

fɔ lɛfɔ a fɔ

Siwu a , fō ɔ fɔ fɔɔ Lelemi a-, é- -ɔ fúla, fua awɔ Attié bu / bɔ bu, bó bu, bó bu Ebrié ɛ e e, e hɛ, e Mbato a / ã a hɛ, a Anyi

wɔ, w- (+V), o, ε / e (dial.) wo, o, wɔ / w-

Anyi-Sanvi ɛ wɔ wo, e wɔ Baulé

a, wo, e wɔ (+V) ɔ (+C), wɔ (+V)

wɔ (+V), wɔ

Nzema e- / ε-, ho e we Abron ɷ, wo, u (int., neg.) ɷ ɷ, wo wɷ Akan wo- , i- / a-, wu-, woa- wo wo wo Akan-Fanti i- / a-, wu / wa -wu íwu Dompo wo Gonja fé, fó fó fó fó Nkonya fo- / fɔ- / fe- / fε- fo fo fo Nchumuru-Banda fɩ- fʊ fʊ

Nchumuru-Bejamso-Grubi

fe / fi (non-past aff.), fʊ / fε (non-past neg.), fε (past neg.), fé (cond.)

fʊ (inalien.)

fε , f- (+V), fʊ (inalien.)

Cherepon wʊ wʊ wʊ wʊ Aburé ɛ wɔ wɔ, w’ wɔ Eotilé e e, e fo, e Krobu bo bó bo It is typologically well known that the pronouns of the 1st person singular are diachronically more stable in the language and therefore easier to reconstruct. They may remain intact through thousands of years of the language development. This rule

13

must work for Kwa as well: as we see from Chart 2, the amplitude of forms for the 2nd person singular is much wider than for the 1st person singular. Still, it is possible to pick out a few common models:

1) Forms of the type wV, with a back vowel following a bilabial sonant. We should probably name here several non-subject forms such as Avikam and Ega (V)ɣɔ, which demonstrate a regular phonetic correspondence with -w- of other Kwa languages. 2) Forms of the type fV, with a voiceless labial fricative, distinguish two subgroups of Kwa, namely Lelemi and North Guang. This -f- seems to regularly correspond to -w- in other Kwa languages (cf. Anyi ku-we ‘belly’ vs. Lelemi ká-fɔ [Allan 1976; Pyne 1976]). The common origin of these two pronominal roots is evident. The retention of *f in those languages may be a result of their geographical contact with Gur languages where fo is universally the 2nd person singular pronoun. 3) Forms of the type bo / bu in two isolated languages of Attié and Krobu are descending again from the same root49

.

4) Pure vocalic forms dotted across various subgroups of Kwa. Among them, the most widespread are: a / a- (Gun, Avikam, Adele, Logba, Sekpele, Siwu, Lelemi, Mbato, Baulé, Akan), e / e- / ε (Ewe, Adioukru, Alladian, Ebrié, Anyi, Lelemi, Nzema, Aburé, Eotilé), o / o- / ɔ- (Gen, Bowili, Ga, Ega, Anyi). Aspirated forms such as Nzema ho should probably be attributed here too.

Thus, we actually have to deal with two groups of pronominal forms of the 2nd person singular: we can mark them *wV и *V, for convenience. A morphological distribution is seen between them resembling the one we saw above for the 1st person singular: V-forms are mostly used in the subject series, while wV-forms are more or less equally distributed among the four sets. This binary opposition of *V (verb subject) vs. *wV (other series) exists in Ewe, Gun, Fon, Alladian, Avikam, Adele, Ega, Sekpele, Lelemi, Anyi-Sanvi, Akan, Aburé, i.e. in all principal groups of the Kwa family, which makes it possible to reconstruct a similar situation in Proto-Kwa as well. This hypothesis can be supported by external data: there was an almost identical opposition in Proto-BC [Babaev 2008; in press]. The diffusion of *wV to subject slots which we see in some tongues was caused by paradigmatic levelling which we have already discussed above. In the emphatic series, we can easily reconstruct *Vwe / *Vwo which is structurally analogous to what we saw in the same series of the previous section: the independent pronoun with a prefixed vowel, the so called ‘associated marker’. This form is reconstructed for Proto-Kwa according to Igo, Abidji, Avikam, Adele, Ega, Sele, Lelemi, Akan-Fanti.

49 This isogloss makes it tempting to think over a genetic proximity of the two languages; this, however, will need more correspondences from both lexicon and morphology.

14

It seems much more difficult to define the phonetic, especially vocalic composition of the proto-language forms, for the lack of any system of phonetic correspondences between the languages and any proto-language phonetic reconstruction. It is natural that *V and *wV have influenced each other a lot, levelling vocalism within the paradigm. External data can be of some assistance here: for Proto-BC, we reconstruct forms *υ- (verb subject) and *(à)wo (independent pronoun) [Babaev, in press]. There is one more tentative hypothesis here. According to the system of phonetic correspondences between Proto-Potou-Akanic and Bantu, produced by Stewart [2002], the PT *wυ corresponds to Bantu *kυ. This poses a question of possible kinship between Kwa forms and the Bantu object suffix *-kυ-: by Stewart, the Proto-Benue-Kwa form should be given as *kwu. However, this curious view does not seem viable: apart from *-kυ-, we see a great number of 2nd person sg. pronouns in Bantu and other BC languages which definitely descend from *wo / *we, in all series including the object, and we cannot prove their relationship with Bantu *-kυ-. Other forms in Chart 2 can be called solitary pieces of innovation, such as Kebu lə, Bowili no, Adioukru (possessive) ɲi. The origins of these forms is in most cases analogical. The result of our reconstruction is given below as three forms of the 2nd person singular pronouns which genetically descend from two different roots: *o- / u- verb subject (bound) *wo / *we non-subject (independent) *Vwe / Vwo emphatic (independent)

3. 1st person plural Selected pronominal forms for the 1st person plural are given in Chart 3.

(Chart 3)

subject object possessive emphatic Ewe mı e mı mı a mı awó Ewe-Adangbe mí mí míá mé- míá wó Gen mı mı mı a mı aó Gun mí -mí -mítɔn mílε Fon mí mí mí mí Nyangbo-Tafi blɔ- / bɔ- (before l, d) blɔ, lɔ blɔ / bl- blɔšú Avatime

blɔ / bl- bε (excl.), blɔ (incl.)

Bowili bo- / bu- bue Igo ówo Kebu lə Animere wo Abé elə Abidji r- ɛrɛ

15

subject object possessive emphatic Adioukru s- / se (incl.), we (excl.) εy epum, εy ɛy, si, εs Alladian

bɔ / bo bɔ bɔ / bo bɔ, bɔ (topic.)

Avikam o- / ɔ- (+C), w- / wa (+V) -wa o / ɔ, wa awa, wa Ga wɔ- (prog., fut., aor.), wɔ-

(perf.) wɔ wɔ wɔ

Adele tée , tí (perf.), te (fut.) ti e té Ega wa wɛ w- ɩwɛ Logba ati -tu atu atú Sekpele bo- bu, bo bō, bó bō, bó Sele mun- (hab. aff.), bwo- / bwa-

(prog./past aff.), bu- (prog./past neg., fut.), mu- (past neg.)

-wo lo o a wo

Siwu bo bo bo bo o Lelemi bɔ-, bú- -bu búla, bua abu Attié

hã, a, an hā, ha, a, an

hā, ha, a, an hā, ha, a, an

Ebrié o lo lo Mbato lo / nõ, do lɔ, do lɔ, do Anyi

jɛ, ye / e (excl.), ame (incl.) jε jε / j-, ye, ame

jε / yɛ, ye, ame

Anyi-Sanvi jɛ, ye (excl.), ame (incl.) jɛ ye, ame jɛ Baulé

e, ye (excl.), ame (incl.) ye, ame

ye, ame é, ye, ame

Nzema ye (excl.), yame (incl.) ye, yame ye, yame Abron ye (excl.), ame (incl.) jɛ ? ye, ame ye, ame Akan yɛ- / ye- / yɛa- (excl.), ame

(incl.) yɛŋ yɛŋ yɛŋ , hen

Akan-Fanti ye- / ya-, hen hén- íhen Gichode à àyé àyí / àyé àyé Ginyanga à àɲi àɲi / àɲe àɲi Dompo aniya Gonja aN, aɲé aɲé aɲé ba aɲé Nkonya a- / ã- ane ane ane Nchumuru-Banda anɩ anɩ anɩ

Nchumuru-Bejamso-Grubi an- (+V), ane (+C) ane

an- +V, ane+C

ane

Cherepon ɛnɩ ɛnɩ ɛnɩ ɛnɩ Aburé jε / ye / e (excl.), ame (incl.) jɛ, e, e, ame e, ame

16

subject object possessive emphatic

ame Eotilé ye (excl.), ame (incl.) ye, ame ye, ame Krobu nɛ nɛ nɛ We will again start from identifying the most characteristic and widespread types of pronouns which could be traced back to the proto-language. The form *mí can be reconstructed for Proto-Gbe, but not higher than that since it is not used anywhere else in Kwa. We shall also note phonetic resemblance between this pronoun and the 2nd person plural forms in Gbe (see below): their only differ in tone and are most likely the product of analogical development. The influence of the plural class prefix can be suggested as the reason; in any case the link with the 1st person singular *me is unlikely. Gbe suffixes -á, -áwó, -áó all have nominal origins. Forms bo / bu / wo / wa are observed in all major branches of Kwa except Gbe and Tano. The genesis of this pronoun is undoubtedly linked with the common NC noun class prefix *bV- [Williamson & Blench 2000: 40] denoting animated plurality. In a number of Kwa languages this form marks plural forms of all three persons (Sekpele bo ‘we’, ba ‘you’, ba ‘they’). The correspondence between initial b- and w- is widely spread in Kwa (cf. Anyi wu-n ‘to sleep’ vs. Lelemi bɔ-nu id. [Allan 1976; Pyne 1976]). This feature is well known in BC as well [Babaev, in press] where the class prefix is bound to the old pronominal root (cf. section for the 2nd plural forms below). Later processes of phonetic simplification may have led to the 1st person plural pronoun *bV: cf. bà ‘we’ in Ukaan [Jungraithmayr 1973: 47], wa / awa id. in Yoruba and Igala (Defoid) [Philpot 1935, Ohiri-Aniche 1999]. We may call this process an areal typological similarity, and most probably an old one. Ye is the 1st person plural exclusive in most Tano languages; it was formerly treated as dual by some scholars [Delafosse 1900: 49]. Forms marked as je / jε should be analysed here as well; at least some of them are products of incorrect transliteration. Since ye is not found beyond Tano, we cannot postulate its Proto-Kwa origins. It is possible, however, to trace the phonetic transformation of this pronoun as *ɲe > ye > e, whose nasality was dropped in the initial position but is still seen in the intervocalic position in emphatic pronouns with a prefixed *a-: Dompo aniya, Gonja àɲé, Nkonya ane, Nchumuru ànɩ / ane, Cherepon ɛnɩ. We can clearly see different stages of the process in closely related Ginyanga (àɲe) and Gichode (àyé). All these forms have common origins in Proto-Tano and should not be regarded as a recent local borrowing, as is proven by identical forms in Krobu and Nkonya, the two language located on the opposite poles of the Kwa-speaking area. The external comparison data gives us more material to support the ancient roots of Tano *(V)ɲe. We should first of all mention some West BC languages. In Proto-Igboid, the 1st person plural pronoun is reconstructed as *(a)ɲì [Babaev, in press], in some tongues such as Ekpeye or Ogba it was transformed to ye. In Nupoid, most languages demonstrate ye / yi [Scholz 1976; Smith 1980; Hyman & Magaji 1970; Blench &

17

Doma, рукопись]. In Ibibio and Efik (Cross River family) there is nɲìn [Cook 1976]. All these BC pronouns do not bear clusivity category, but can be considered as evidence for ancient origins of the Tano pronoun. In the same Tano group we find *ame, the 1st person plural inclusive form. It might be connected with *me of the first person singular extended by a plural prefix, but such a prefix cannot be discovered. But, we do discover a possibility of another version of its origin: *ame could originate from two Tano pronouns combined: *a- ‘thou’ and *me ‘I, me’. This hypothesis enjoys wide typological evidence in the world’s languages: it is indeed one of the most typical means of building the inclusive pronouns. We can recall Tungus-Manchu forms bi-ti, mi-ti, mün-ti [Rédei 1988: 294] or Mongolian bide, bida ‘we (incl.)’ < *bi ‘I’ + *ta ‘you’ [Рамстедт 1957: 71]. In the Ilocano language of the Philippines, the inclusive pronoun tayo is a mere composition of ta ‘we two’ and yo ‘you’ [Cysouw 2003: 90]. The inclusive form yumi of Tok-Pisin, the pidgin tongue of New Guinea, comes from English you-me. There are typological evidence from neighbouring BC languages: in Yemba (Bantoid) the combination of pèk ‘we’ и pε ‘you’ produces the inclusive pronoun mpε [Harro & Haynes 1991: 18-19]. Beyond Tano, the clusivity category is only fixed in Avatime and Adioukru, where inclusive forms are clear innovations. We are confident that there was no clusivity in Proto-Kwa. Finally, we should analyse pronominal forms of Chart 3 which are based on the dental consonant: Adele tí, te, eté, Logba -tu, atu, Mbato do / lo, as well as Adioukru (incl.) s- / se. The Adioukru inclusive pronouns are definitely cognate to the phonetically modified forms of genetically related Abe èlə and Abidji r- / ɛrɛ . External evidence is solid and important to note. The comparison of the Kwa forms with Proto-BC 1st person plural *tυ- (verb subject prefix) and *(à)ce (independent pronoun) [Babaev 2008] will be correct, as well as with the neighbouring Gur languages which demonstrate tu as a normal 1st person plural pronoun. An exact parallel with Adioukru can be found in Bantoid: in Limbum the 1st person plural pronouns are sèè (incl.) and wèr (excl.) [Fiore & Peck 1980]. Extending our view, we can attract evidence from Gur languages which are geographically close to Kwa: cf. Bassari tí, táá, -tι [Cox 1998]. All that makes it possible to treat *(V)tu / *(V)ti in Proto-Kwa as a relic pronoun, even though it is now witnessed in all branches of contemporary Kwa. We will reconstruct the following proto-language forms of the 1st person plural pronouns: *(V)tu / *(V)ti 1st pl. (Proto-Kwa?) *(a)ɲe 1st pl. exclusive (Proto-Tano) *ame 1st pl. inclusive (Proto-Tano) *mí 1st pl. (Proto-Gbe) It is amazing that even though all the given proto-language forms are limited by certain branches of Kwa, three of the five (except *ame и *mí) have external parallels. Of these three, *(a)ɲe can possibly also be an areal innovation linking West BC and Tano languages. *bo / *bu as a 1st person plural marker is undoubtedly an ancient form which is widespread not only in Benue-Kwa but also beyond. However, whether it can

18

be called Proto-Kwa is a question. The genetically reliable form should be *(V)tu / *(V)ti. 4. 2nd person plural

(Chart 4)

subject object possessive emphatic Ewe mı e mı mı a mı awó Ewe-Adangbe mī mī mīā mé- mīā wó Gen mı mı mı a mı aó Gun mì -mì -mìtɔn mìlε Fon mi mi mi mi Nyangbo-Tafi

wɔnɔ-, nɔ- wɔnɔ, nɔ

wɔnɔ / wɔn-

wɔnɔšú (refl.)

Avatime mlɔ / ml- Bowili mì mi Kebu nə Animere jɪ Abé éɲə Abidji f-..-nı n-ɛfɛ Adioukru ø-..-r ɔɲ ɔɲ ɔɲ Alladian

ɲɛ / ɲe ɲé (topic.)

ɲē ɲè, ɲé (topic.)

Avikam ɔ- (+C), w- (+V), un -wa ɔ, un awa, unõ Ga ɲε - ɲε ɲε ɲε Adele yée , yí (perf.), ye (fut.) yi e yé Ega ŋʊ ŋɛ ŋ- ɩŋɛ Logba ani -nu anu- amú Sekpele ba mí mimí mimí Sele min- (hab.aff.), bye- / bya-

(prog./past aff.), bi- (prog./past neg., fut.), mi- (past neg.)

-ye le e a ye

Siwu mı mı mı mıı Lelemi

bí- -mí, -mì

mína, bia amî, amì

Attié mu / mɔ, mun(e) mú, une mú, mune mú, mun(e) Ebrié o lo, o hɔ, lo Mbato nã, u na , bo hé, bo Anyi

amu, amwĩ ɛmɔ amu, ɛmɔ ɛmɔ, amwĩ, amu / amõ

Anyi-Sanvi ɛmɔ, be ɛmɔ be ɛmɔ, be

19

subject object possessive emphatic Baulé a, amu amu amu amu, amu Nzema be- / bε-, yame-e yame-e, be yame-e, be Abron hɷ, amene hɷ amene hɷ Akan mo-, mu-, moa-, amene mo mo, amene mo, hum Akan-Fanti hum- -hum hum- íhum Dompo aniide Gonja be , beɲé beɲé beɲé ba beɲé Nkonya mlε- mlε mlε mlε Nchumuru-Banda mɩnɩ mɩnɩ mɩnɩ

Nchumuru-Bejamso-Grubi mε n, mãn, mane mane mane mane

Cherepon ɛnɩ↓ ɛnɩ↓ ɛnɩ↓ ɛnɩ↓ Aburé vɛ, ama vɛ vɛ, ama vɛ, ama Eotilé ama ama mo, ama Krobu be bé be For Kwa, as it was mentioned below, it is typical to use the 3rd person plural pronoun (itself originating from a plural noun class prefix) in the 2nd person plural. In the great variety of languages, this leads to their direct unification. Let us for example consider subject forms of the two plural persons in certain Kwa languages in Chart 5.

(Chart 5)

2nd pl. 3rd pl. Sekpele ba ba Sele min- / bye- / bi- man- / bee- / bo- Mbato nã, u, bo u, bo Anyi-Sanvi ɛmɔ, be be Nzema be- / bε- be- / bε- Gonja be, beɲé be Aburé vɛ, ama vɛ Krobu be, bē be, bē

This specific feature is not only characteristic for Kwa: it is quite widespread in the languages of West and Central Sub-Saharan Africa and can be considered a morphological isogloss of a huge area. In Edoid languages (West BC), the ‘you’ pronouns have been almost universally replaced by ‘they’ forms: cf. Engenni ɓa, Yekhee βā, Edo (ù)wà, Okpamheri we, etc. In Cross River, we have Tee bò, Legbo ba-, Kukele be ba ‘you’ and many others [Babaev, in press]. Even in the Bantu languages which are considered morphologically conservative compared to other NC families, this phenomenon is widely found in Northwest and Northeast zones: e.g., Akoose (А15) uses both ɲê and bé to denote ‘you’, the latter being the 3rd person pronoun [Nurse 2007]. In Babole (С101), Makaa (А83), Bila (D30) and other Bantu tongues the

20

2nd and 3rd person plural pronouns are only different in vocalism [Heath 2003, Leitch 2003]. There are examples of similar expansion of the 3rd person plural forms in many other languages of the region: e.g., Adamaua languages (such as Dai [Nougayrol 1979]) or Mande (such as Ligbi [Persson & Persson 1980]). The reason for this contraction must lie in the semantic field (in understanding ‘you pl.’ as an exclusive pronoun covering both the addressee of the speech act and those with him), and typologically this is one of the commonest processes of merger within the pronominal paradigm in the languages of the world [Cysouw 2003]. In German, for instance, Sie ‘you (polite)’ does originate from the same source. Without going into much typological detail, we tend to regard the 2nd person plural forms ba / wa in Kwa as the reflection of the same process, which could take place in Proto-Kwa or its dialects. Of the other forms, two types are necessary to mention, both seemingly rooting from the proto-language: mu / mo and ne / nε. The former of them (used in Attié, Anyi, Logba, Baulé, Akan, Eotilé and, possibly, also in Aburé ama) is not witnessed in Left Bank Kwa languages, but is widespread across Nyo. Forms with a prefixed vowel such as Baulé amũ help to prove that these forms were initially emphatic. *mu does have a correspondence in BC (Bantu and Bantoid *mυ-, Ukaan mɔ ), its reconstruction for Proto-Benue-Kwa (as *mυ) was proposed by Stewart [2002: 223], with whom we need to agree. The other proto-language form should be reconstructed in two varieties: *(V)ne and *mI-ne / *mI-ni < *bI-ni, of which the one with a prefixed vowel is emphatic by origin, and the latter one is a contamination with a plural noun class prefix we have widely discussed. We can only tentatively suggest the vocalism of these pronouns: it is clear to have suffered much due to intensive assimilation. Here are the reflexes:

*ne: Nyangbo-Tafi nɔ-, Kebu nə, Alladian ɲε / ɲē, Ga ɲε-, Ega ŋυ, ŋε , Logba -nu, Mbato nã; *Vne: Abé éɲə, Adioukru ɔɲ, Avikam unõ, Ega īŋε, Logba ani, anu-, Attié une, Dompo ani-ide, Cherepon εní; *mI-ne / *bI-ne: Gun mìlε, Nyangbo-Tafi wonɔ-, Lelemi míná, Attié mun(e), Abron amene, Akan amene, Gonja bèɲé, Nchumuru-Banda mìnι, Nchumuru-Bejamso-Grubi mεn.

It would be tempting to compare these genetically close forms with those with phonetic modifications, e.g. Avatime mlɔ, most probably from *bɔ-nɔ, compared with its close cognate wɔnɔ in Nyangbo-Tafi. Among a number of the 2nd person plural pronouns of the North Guang subgroup (Gonja bèɲé, Nchumuru-Banda mɩnɩ, Nchumuru-Bejamso-Grubi man e ) we note a phonetically different Nkonya mlε-: their common descendance from *be-nI or *bε-nI is without doubt. Adele forms like emphatic e yé < *Vné are the result of the development of old *n > *ɲ > *y, cf. ɩŋɛ in close Ega. Therefore, we can suggest three Proto-Kwa pronouns for the 2nd person plural slot: *ne, *Vne, *bI-ne. Paradigmatically, it would be logical to proceed that *ne was the non-subject independent pronoun, while the latter two were emphatic. It can also be

21

supported by the fact that *ne could be shaped with the nominal prefix *bI-, so it acted as the nominal predicate as well. In Chart 5 we give several innovative forms. For Gbe languages, Bowili and Lelemi dialects we can reconstruct *mì, which was subject to mutual influence with the 1st person plural *mí: in both of them, traces of the plural suffix should be guessed. In possessive and emphatic series, these pronouns add standard nominal suffixes. The origins of Abron hɷ and its likes in Ebrié and Mbato are not clear, but the Baron form should be somehow connected with the 2nd person singular ɷ. The distribution between two proposed variants *mu and *(V)ne / *bI-ne is also unclear. Basing on the distribution of proto-language pronouns in the singular, we can only propose that one of them was marking the verb subject, while the other was independent and nominal. This is at least what we see in Proto-Bantoid: *mυ- is reconstructed as the subject verb prefix, and *n(u)e / *ben(u)e as the independent non-subject pronoun [Babaev 2008]. This data of external comparison can help us to support conclusions which are sometimes hard to prove by means of the internal Kwa analysis: such as the vocalism of Proto-Kwa *ne above, which we could only guess unless the external evidence. 5. Conclusions The conducted analysis has led us to the conclusion that the Proto-Kwa paradigm of personal pronouns should be reconstructed as two main original series: verbal (subject) and independent (nominal, non-subject) which marked possession and a nominal predicate. Another set of pronouns which must have existed in Proto-Kwa was the emphatic one with a prefixed vowel, formed on the basis of the independent pronoun and sometimes replacing it.

(Chart 6)

subject non-subject 1st sg. *nI- *(V)me / *(V)mε 2nd sg. *o- / *u- *(V)wo / *(V)we 1st pl. *(V)tu / *(V)ti 2nd pl. *mu *(V)ne / *bI-ne

Let us now compare this result with the reconstruction of the Proto-BC personal pronouns / person markers made by the author in [Babaev 2008; in print]:

(Chart 7) subject non-subject

1st sg. *ɲi- *(à)me 2nd sg. *υ- *(à)wo 1st pl. *tυ- *(à)ce / *bèce 2nd pl. *mυ- *(à)nue / *bènue

The considerable resemblance between the two reconstructions of the relative families is naturally a picture yet to be recovered in more detail, bearing in mind a number of individual characteristics of various groups and subgroups of Kwa which we have touched in the present paper. However, the brief analysis we have attempted to make creates a real opportunity to make a significant step forward in reconstructing the

22

grammatical system of the Benue-Kwa proto-language, one of the major branches of the Niger-Congo macrofamily of languages.

Appendix 1. Third person pronouns in the Kwa languages. 3rd person singular

(Chart 8)

subject object possessive independent / emphatic

Ewe é / wò, né (juss.)

e é / è éya

Gen é, bě (dep.) è / ε é éyà Gun é e / ε / i é-tɔ n éɔ / úɔ Fon é(yε) é(yε) é-tɔn é(yε) Nyangbo-Tafi ye / yε yε yε ye / yε Avatime é / á yē / yε ye- yε Bowili o / yè, e- / ε- o / ye Kebu ŋwə Animere ye ye Abé nə / ə nə Abidji n- -nì εnε Adioukru

o, nā, lu o, nān li (anim.), ow, in

(dep. anim.), ow’n (dep.)

Alladian l- / n- (V), li- / ni- (C), ne / no /

o -r in, ne e, ne

Avikam e / è è e / è Ga

e lε (anim.), ø

(inanim.) e lε

Dangme e ε, a, ɔ e lε Adele ε, ε (fut.,

perf.neg.), ya (perf.)

o εwɔ

Ega lι jí j-ʹ ιjí Logba

ɔ ε / o, lε

(anim.), yε (inanim.)

o (anim.), alε (inanim.)

ɔlε (anim.), mε (inanim.)

Sekpele o / u (anim.),

a / e (inanim.) wε (anim.) wε wε

Sele o ɔ nɔɔ ŋwɔɔ Siwu ɔ u õ õ

23

subject object possessive independent / emphatic

Lelemi o- -ŋ ŋwa àŋu Attié e e o, e Ebrié a / e mi a / ā / ĩ Mbato e ē hē Anyi o- / ɔ- (C), y-

(V) -y

o- / yé (C), y-, i (V)

yε, i / èi

Anyi-Sanvi o / a i i / èi Baulé ɔ i i (y)i Nzema o- / ɔ-, ye- o, i èi / a Abron wo / u we wo Akan-Asante ɔ- / o- / wa-

(anim.),

ε- / e- / a- (inanim.)

ɔno (anim.),

εno (inanim.)

Akan-Fanti o- / wa-, nu / na nú- owo Gichode o / ɔ mɔ me / m- / mɔ mɔ Ginyanga ó / ɔ / ò / ɔ è / à nε wa Dompo nyinε Gonja è (anim.),

kì (inanim.)

mò (anim.), kúmú

(inanim.)

m bá (anim.),

kúmú bá (inanim.)

mò (anim.),

kúmú (inanim.)

Nkonya ɔ- (anim.), e- / i- (inanim.)

mõ (anim.) (e)mõ

Nchumuru-Banda ɔ- mυ / m mυ

Nchumuru-Bejamso-Grubi

mu (anim.),

kεm (inanim.) mu / me / mε mu (anim.), kéyà

(inanim.)

Cherepon á (anim.), ε (inanim.) mυ mɔ mυ

Aburé o / a / e i / u e Eotilé o o wo Krobu o ó 3rd person plural

(Chart 9) subject object possessive independent /

emphatic Ewe wó, wóné (juss.) wó wó wó / áwó

24

subject object possessive independent / emphatic

Gen ó / ò ó óá óáó Gun yé yé yé-tɔn yélε Fon yé yé yé-tɔn yé Nyangbo-Tafi

balε balε / alε balε ba / be / balε, balεšú (refl.)

Avatime bā ba- bā Bowili ba / bε, bi- bε Kebu m pə Animere bε bε Abé kεlə Abidji n-...-ní n-εnε Adioukru

bè, li, wè en bè, -li (anim.),

wεl, εl (indirect)

Alladian l- / n- (V), li- / ni- (C)

εl εl, ye

Avikam aõ ā aõ, ā Ga àme àme àme àme Dangme a me à me Adele béε (pres.), bεε

(progr.), bá (perf.), bε (fut., perf.neg.)

bε bε

Ega mυ mε m-ʹ ιmε Logba e a a alε Sekpele

ba, má nyε / mε

(anim.), má máma má

Sele boo ma lεmā m m ā Siwu mā, mã mā mā mā` Lelemi ba- -mà bànà / mànà àma Attié ba ba ba Ebrié o o / lo lo Mbato ɓo / mo bɔ bɔ Anyi bε / b- bε bε / b- bε Anyi-Sanvi be be be Baulé be be be be Nzema be- / bε- be be Abron be be be Akan-Asante wɔ- / wo- / wɔa-

(anim.), ε- / e- / a- (inanim.)

wɔŋ (anim.),

εno(nom)

wɔŋ (anim.),

no (inanim.)

wɔŋ (anim.),

n(e) (inanim.)

25

subject object possessive independent / emphatic

(inanim.) Akan-Fanti wo- / wa-, hon óhon Gichode mã mo mõ mo mõ mo mõ Ginyanga

me / m- / mε mɔ (anim.), á

(inanim.) mɔ mɔ

Dompo bam Gonja

bǎ / bò (anim.),

à (inanim.)

bòmò (anim.),

ámú

(inanim.)

bómó bá (anim.),

ámú bá

(inanim.)

bòmò (anim.),

ámú (inanim.)

Nkonya be- / bε- / bo- / bɔ- / bu- (anim.), e-

(inanim.) amõ amõ

Nchumuru-Banda bι- bamυ / bam bamυ

Nchumuru-Bejamso-Grubi

bám báme / bám- báme

Cherepon ε mυ ε mυ ε mυ ε mυ Aburé υe υe / υ- υwe Eotilé wa Krobu a á

Appendix 2. Alternate Language Names

language alternate names Abe Abbé, Abbey, Abi Abron Bron, Brong, Doma, Gyaman Aburé Abonwa, Abule, Akaplass Adele Bedere, Bidire, Gadre, Gidire Adioukru Adjukru, Adyoukrou, Adyukru, Ajukru Akan Twi Akan-Fanti Fante, Twi Alladian Alladyan, Allagia, Allagian Animere Anyimere, Kunda Anyi Agni, Anyin Anyi-Sanvi Afema, Sanwi Attié Akie, Akye, Atche, Atie, Atshe Avatime Afatime, Siya, Sia, Sideme Avikam Avekom, Brignan, Brinya, Gbanda, Kwakwa, Lahu Baulé Bawule Bowili Bawuli, Bowiri, Liwuli, Siwuri, Tuwili, Tuwuli Cherepon Cherepong, Chiripon, Chiripong, Kyerepong, Okere

26

Dompo Dumpo, Ndmpo Ebrié Cama, Caman, Kyama, Tchaman, Tsama, Tyama Ega Diés, Egwa Eotilé Beti Ewe Ebwe, Efe, Eibe, Eue, Eve, Gbe, Krepe, Krepi, Popo, Vhe Fon Dahomeen, Djedji, Fo, Fogbe, Fongbe, Fonnu Ga Accra, Acra Gen Ge, Gebe, Gen-Gbe, Guin, Mina, Mina-Gen, Popo Ginyanga Agnagan, Anyanga, Genyanga Gichode Gechode, Chode, Kyode, Gikyode Gun Alada, Alada-Gbe, Egun, Goun, Gu, Gugbe, Gun-Alada, Gun-

Gbe, Seto-Gbe, Toli-Gbe Igo Ahlo Kebu Akebou, Ekpeebhe, Gakagba, Kabu, Akebu, Kegberike Lelemi Buem Logba Ikpana Mbato Goaa, Gwa, M’bato, Mbatto, Mgbato, N-Batto, Ogwia, Potu Nchumuru Kyongborong, Nchimburu, Nchumburu Nyangbo-Tafi Tutrugbu Nzema Zema, Appolo, Nzima Sekpele Likpe Sele Santrokofi Siwu Akpafu

List of Abbreviations. aff. affirmative alien. alienable possession anim. animate aor. aorist cond. conditional mood dep. dependent clause dial. dialectal excl. exclusive fut. future hab. habitual mood imperf. imperfect inalien. inalienable possession inanim. inanimate incl. inclusive int. interrogative juss. jussive neg. negative perf. perfect pl. plural pres. present tense progr. progressive refl. reflexive sg. singular topic. topicalised

27

REFERENCES

Aboh, Enoch. 1997. On the Syntax of Gungbe Noun Phrases. Paper presented at the Annual Meeting of

the African Language and Linguistics Colloquium, Leiden. Agbedor, Paul Kofi. 1996. The syntax of Ewe personal pronouns. // Linguistique africaine 16, pp. 19-

52. Allan, Edward Jay. 1976 [1973]. Buem. // Kropp Dakubu, M.E. (ed.), West African language data

sheets 1, Legon & Leiden: West African Linguistic Society; African Studies Centre. Allan, Edward Jay. 1980 [1974]. Likpe. // Kropp Dakubu, M.E. (ed.), West African language data

sheets 2, Legon & Leiden: West African Linguistic Society; African Studies Centre. Allen, Christine Anne. 1980 [1973]. Sele. // Kropp Dakubu, M.E. (ed.), West African language data

sheets 2, Legon & Leiden: West African Linguistic Society; African Studies Centre. Apronti, E.O. 1976 [1972]. Dangme. // Kropp Dakubu, M.E. (ed.), West African language data sheets

1, Legon & Leiden: West African Linguistic Society; African Studies Centre. Babaev, Kirill. 2008. Reconstructing Niger-Congo Personal Pronouns I: Proto-Bantoid. // Journal of

West African Linguistics. Vol. XXXV: 1-2. Pp. 131-183. [http://www.nostratic.ru/benue-congo1.pdf] Babaev, Kirill. 2010. Reconstructing Niger-Congo Personal Pronouns II. // Journal of Language

Relationship, no. 4 (2010). [http://www.nostratic.ru/benue-congo2.pdf] Bamgbose, Ayo. 1966. A Grammar of Yoruba. // West African Languages Monographs, 5. Cambridge:

Cambridge University Press. Bennett, P. & Sterk, J. 1977. South-Central Niger-Congo: a Reclassification. // Studies in African

Linguistics, 8, pp. 241-273. Benveniste, Emile. 1971. Problems in General Linguistics. Miami. Blench, Roger. 2001. Comparative Central Togo: What Have We Learnt Since Heine? Paper presented

at the 32nd Annual Conference on African Linguistics. Berkeley, 23-25th March. Blench, Roger. 2007. The Dompo Language of Central Ghana and Its Affinities. Manuscript

[http://www.rogerblench.info]. Blench, Roger & Doma, Musa. A Dictionary of the Gbari Language. Manuscript

[http://www.rogerblench.info]. Bôle-Richard, Rémy. 1983a. L’ébrié. // Hérault, G. (ed.), Atlas des langues kwa de Côte d’Ivoire 1,

Paris & Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 307-357.

Bôle-Richard, Rémy. 1983b. L’éga. // Hérault, G. (ed.), Atlas des langues kwa de Côte d’Ivoire 1, Paris & Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 359-401.

Bôle-Richard, Rémy. 1983c. Systématique phonologique et grammaticale d’un parler ewe: le gen-mina du sud-Togo et sud-Bénin. Paris : L’Harmattan.

Boyd, Raymond. 1989. Adamawa-Ubangi. // Bendor-Samuel, John (ed.) Niger-Congo Languages. NY - London: Lanham, pp. 178-215.

Burmeister, Jonathan L. 1983a. L’abouré. // Hérault, G. (ed.), Atlas des langues kwa de Côte d’Ivoire 1, Paris & Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 67-81.

Burmeister, Jonathan L. 1983b. L’agni. // Hérault, G. (ed.), Atlas des langues kwa de Côte d’Ivoire 1, Paris & Abidjan : Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 155-179.

Capo, Hounkpati B.C. 1991. A comparative phonology of Gbe. Berlin: Walter de Gruyter. Casali, Rod. 2006. Animere Phonetic Word List. Draft. SIL. Christaller, Johann Gottlieb. 1933. Dictionary of the Asante and Fante language called Tshi (Twi).

Basel : Evangelical Missionary Society, vol. 32. Cleal, Alizon. 1976 [1973]. Gechode. // Kropp Dakubu, M.E. (ed.), West African language data sheets

1, Legon & Leiden: West African Linguistic Society; African Studies Centre. Cleal, Alizon. 1976a [1973]. Genyanga. // Kropp Dakubu, M.E. (ed.), West African language data

sheets 1, Legon & Leiden: West African Linguistic Society; African Studies Centre. Cleal, Alizon & Norman Price. 1980 [1975]. Nchumuru. // Kropp Dakubu, M.E. (ed.), West African

language data sheets 2, Legon & Leiden: West African Linguistic Society; African Studies Centre. Clopper, Cynthia. 2001. The Nzema Verbal Phrase: An Optimality Theoretic Account. // IULC

Working Papers Online, vol. 1. Cook, T.L. 1976 [1968]. Efik. // Kropp Dakubu, M.E. (ed.) West African Language Data Sheets. Vol.

1. West African Linguistic Society. Cox, Monica. 1998. Description grammaticale du ncam (bassar), langue gurma du Togo et du Ghana.

Paris: Ecole Pratique des Hautes Etudes, Université Paris-Sorbonne. Cysouw, Michael. 2003. The Paradygmatic Structure of Person Marking. Oxford: Oxford University

Press.

28

Delafosse, Maurice. 1900. Essai de manuel de la langue agni. Paris: J.André. Delafosse, Maurice. 1904. Vocabulaires comparatifs de plus de 60 langues ou dialectes parlés à la Cote

d’Ivoire et dans la régions limitrophes. Paris: E.Leroux. Duponchel, Laurent & Gnamba Bertin Mel. 1983. L’alladian. // Hérault, G. (ed.), Atlas des langues kwa

de Côte d’Ivoire 1, Paris & Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 181-225.

Fiore, Lynne & Peck, Patricia. 1980 [1973]. Limbum. // Kropp Dakubu, M.E. (ed.) West African Language Data Sheets. Vol. 2. West African Linguistic Society.

Funke, E. 1909. Versuch einer Grammatik der Avatimesprache. // Mitteilungen des Seminars für orientalische Sprachen, 12(3), S. 287-336.

Funke, E. 1910. Die Nyangbo-Táfi Sprache: Ein Betrag zur Kenntnis der Sprachen Togos. // Mitteilungen des Seminars für orientalische Sprachen, 13(3), S.166-201.

Grassias, Alain & Rémy Bôle-Richard. 1983. Le m’batto (ngula). // Hérault, G. (ed.), Atlas des langues kwa de Côte d’Ivoire 1, Paris & Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 465-506.

Greenberg, Joseph. 1955. Studies in African Linguistic Classification. New Haven: Compass. Harro, Gretchen, & Haynes, Nancy. 1991. Grammar Sketch of Yemba. Yaoundé: SIL Cameroon. Heath, Teresa. 2003. Makaa. // Nurse, D. & Phillipson, G. (eds.), The Bantu Languages. London &

New York: Routledge, pp. 335-348. Heine, Bernd. 1968. Die Verbreitung und Gliederung der Togorestsprachen. Berlin: Reimer. Hérault, Georges. 1976 [1973]. Adioukrou. // Kropp Dakubu, M.E. (ed.), West African language data

sheets 1, Legon & Leiden: West African Linguistic Society; African Studies Centre. Hérault, Georges. 1983. L’éotilé. // Hérault, G. (ed.), Atlas des langues kwa de Côte d’Ivoire 1, Paris &

Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 403-424.

Hill, Harriet. 1995. Pronouns and Reported Speech in Adioukrou. // Journal of West African Languages, 15 (1), pp. 87-106.

Höftmann, Hildegard. 1971. The Structure of Lelemi Language. Leipzig: VEB. Hyman, Larry & Magaji, Daniel. 1970. Essentials of Gwari Grammar. Occasional Publication, 27. IAS:

University of Ibadan. Iddah, Robert Komra. 1980 [1975]. Siwu. // Kropp Dakubu, M.E. (ed.), West African language data

sheets 2, Legon & Leiden: West African Linguistic Society; African Studies Centre. Jungraithmayr, H. 1973. Notes on the Ishe dialect of Ukaan (Akoko, Western State, Nigeria). //

Africana Marburgensia, 6,1, pp. 39-57. Kouadio N’Guessan, Jérémie. 1983. L’abbey. // Hérault, G. (ed.), Atlas des langues kwa de Côte

d’Ivoire 1, Paris & Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 13-39.

Krause, Gottlieb A. 1895. Die Stellung des Temne innerhalb der Bantusprachen. // Zeitschrift für afrikanische, ostasiatische und ozeanische Sprachen, 1. S. 250-267.

Kropp Dakubu, M.E. 1967. Lefana, Akpafu and Avatime with English Glosses. // Comparative African Wordlists, 3. Legon: Institute of African Studies.

Kropp Dakubu, M.R. 1976 [1973]. Ga. // Kropp Dakubu, M.E. (ed.), West African language data sheets 1, Legon & Leiden: West African Linguistic Society; African Studies Centre.

Kropp-Dakubu, M.E. 1987. The Dangme language: an introductory survey. London-Accra: Macmillan/Unimax Publishers Ltd.

Kropp Dakubu, M.E. 2008. Ga Verb Features. // Ameka F., Kropp Dakubu M.E. (eds.) Aspect and Modality in Kwa Languages. Amsterdam: J.Benjamins, pp. 91-134.

Kutsch Lojenga, Connie & Elizabeth Hood. 1983. L’attié. // Hérault, G. (ed.), Atlas des langues kwa de Côte d’Ivoire 1, Paris & Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 227-253.

Lefebvre, Claire & Anne-Marie Brousseau. 2002. A Grammar of Fongbe. The Hague: Mouton de Gruyter.

Leitch, Myles. 2003. Babole. // Nurse, D. & Phillipson, G. (eds.), The Bantu Languages. London & New York: Routledge, pp. 392-421.

Lewis, M. Paul (ed.). 2009. Ethnologue: Languages of the World, 16th edition. Dallas: SIL International. Online version: http://www.ethnologue.com/

Meinhof, Carl. 1906. Grundzüge einer vergleichende Grammatik der Bantusprachen. Berlin. Mensah, Emmanuel N.A. 1983. Le krobou. // Hérault, G. (ed.), Atlas des langues kwa de Côte d’Ivoire

1, Paris & Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 425-463.

Miehe, Gudrun. 2004. Les pronoms personelles dans les langues gur. // Ibriszimow, D. & Segerer, G. (eds.) Systèmes de marques personnelles en Afrique. Collection «Afrique et Langage», 8, p. 151-162. Louvain – Paris: Peeters.

Moniño, Y. 1995. Le Proto-Gbaya: Essai de linguistique comparative historique sur vingt-et-une langues d’Afrique centrale. Louvain-Paris : Peeters.

29

N’Guessan, Jérémie Kouadio. 1983. Le baoulé. // Hérault, G. (ed.), Atlas des langues kwa de Côte d’Ivoire 1, Paris & Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 277-306.

Nougayrol, Pierre. 1979. Le day de Bouna (Tchad), 1: phonologie, syntagmatique nominale, synthématique. Paris: SELAF.

Nurse, Derek. Tense and Aspect in Bantu. 2007. Oxford: Oxford University Press. Appendices 1-2: http://www.ucs.mun.ca/~dnurse/tabantu.html

Ohiri-Aniche, Chinyere. 1999. Language Diversification in the Akoko Area of Western Nigeria. // R.Blench, M.Spriggs (eds.) Archaeology and Language IV: Language Change and Cultural Transformation. London: Routledge, pp. 79-94.

Painter, Colin. 1980a [1974]. Gonja. // Kropp Dakubu, M.E. (ed.), West African language data sheets 2, Legon & Leiden: West African Linguistic Society; African Studies Centre.

Painter, Colin. 1980b [1974]. Hill Guang. // Kropp Dakubu, M.E. (ed.), West African language data sheets 2, Legon & Leiden: West African Linguistic Society; African Studies Centre.

Persson, Andrew & Janet Persson. 1980. Ligbi. // Kropp Dakubu, M.E. (ed.), West African language data sheets 2, Legon & Leiden: West African Linguistic Society; African Studies Centre.

Philpot, W.T.A. 1935. Notes on the Igala Language. // Bulletin of the School of Oriental and African Studies, 7, pp. 897-912.

Pozdniakov, Konstantin & Guillaume Segerer. 2004. Reconstruction des pronoms atlantiques et typologie des systèmes pronominaux. In: Ibriszimow, D. & Segerer, G. (eds.) Systèmes de marques personnelles en Afrique. Collection «Afrique et Langage», 8, Louvain – Paris: Peeters, p. 151-162.

Pyne, P.C. 1976 [1972]. Anyi. // Kropp Dakubu, M.E. (ed.), West African language data sheets 1, Legon & Leiden: West African Linguistic Society; African Studies Centre.

Rédei K. (ed.) 1988. Uralisches etymologisches Wörterbuch. Wiesbaden. Reineke, Brigitte. 1972. The Structure of the Nkonya Language. Leipzig: VEB. Riis, A. 1853. Elemente des Akwapim-Dialekts der Odschi-Sprache. Basel, Bahnmeier. Schadeberg, Thilo. 1981. The Heiban Group. Vol. 1: The Survey of Kordofanian. Hamburg: Helmut

Buske. Schadeberg, Thilo. 1985. A Small Sketch of Ewe. // Afrikanistische Arbeitspapiere, Universität zu

Köln. Scholz, Hans-Jürgen. 1976 [1973]. Igbirra. // Kropp Dakubu, M.E. (ed.) West African Language Data

Sheets. Vol. 1. West African Linguistic Society. Schuh, Russell. 1995a. Aspects of Avatime phonology. // Studies in African Linguistics, 24(1), pp. 31-

67. Schuh, Russell. 1995b. Avatime noun classes and concord. // Studies in African Linguistics, 24(2), pp.

123-149. Segerer, Guillaume. 2008. Les langues Niger-Congo. // Желтов А.Ю. (ред.), Языки нигер-конго:

структурно-динамическая типология. СПб: Издательство Санкт-Петербургского университета. Segerer, Guillaume. 2002-2007. Les marques personnelles dans les langues africaines, base de données

en ligne. [http://sumale.vjf.cnrs.fr/pronoms] Smith, Neil. 1980 [1967]. Nupe. // Kropp Dakubu, M.E. (ed.) West African Language Data Sheets. Vol.

2. West African Linguistic Society. Sprigge, R.G.S. 1976 [1973]. Ewe (Adangbe dialect). // Kropp Dakubu, M.E. (ed.) West African

Language Data Sheets. Vol. 1. West African Linguistic Society. Starwalt, Coleen. 2008. The Acoustic Correlates of ATR Harmony in 7- and 9-vowel African

Languages: a Phonetic Inquiry into Phonological Structure. Ph.D. thesis. Arlington: University of Texas. Stewart, J. Word List for African Languages. Manuscript. Stewart, J. 1973. The lenis stops of the Potou Lagoon languages and their significance for pre-Bantu

reconstruction. // M.E. Kropp Dakubu (ed.), Papers in Ghanaian Linguistics, 2, pp. 1-49. Research Review supp. 4. Legon: Institute of African Studies, University of Ghana.

Stewart, J. 1989. Kwa. // Bendor-Samuel, J. (ed.) The Niger-Congo Languages. Lanham: University Press of America, pp. 217-245.

Stewart, J. 1993. The second Tano consonant shift and its likeness to Grimm’s law. // Journal of West African Languages 23 (1), pp. 3-39.

Stewart, J. 1994. The Comparative Phonology of Gbe and Its Significance for That of Kwa and Volta-Congo. // Journal of African Languages and Linguistics, vol. 15, pp. 175-193.

Stewart, J. 1999. An explanation of Bantu vowel height harmony in terms of a pre-Bantu nasalized vowel lowering. // Journal of African Languages and Linguistics, 21, pp. 161-178.

Stewart, J. 2000. Symmetric versus asymmetric vowel height harmony and e, o versus I, U in Proto-Bantu and Proto-Savanna Bantu. // Journal of West African Languages, 28 (1), pp. 45-58.

Stewart, J. 2001a. South Volta-Congo (Benue-Kwa) subclassification: The position of Tano (Akanoid). Paper read at the 32nd Annual Conference on African Linguistics, University of California at Berkeley.

Stewart, J. 2001b. The stem-initial consonant system of Proto-Potou-Tano-Bantu: An update. Paper read at the Benue-Congo Workshop at the 32nd Annual Conference on African Linguistics, University of California at Berkeley.

30

Stewart, J. 2002. The potential of Proto-Potou-Akanic-Bantu as a pilot Proto-Niger-Congo, and the reconstructions updated. // Journal of African Languages and Linguistics, 23, pp. 197-224.

Stewart, J. 2004. The Proto-Potou-Akanic-Bantu Reconstructions Updated. Manuscript. Tamanji, Pius. 2006. Concord and DP structure in Bafut. Afrikanistik online, no. 595. Timyan-Ravenhill, Judith. 1980 [1975]. Baule. // Kropp Dakubu, M.E. (ed.), West African language

data sheets 2, Legon & Leiden: West African Linguistic Society; African Studies Centre. Timyan-Ravenhill, Judith. 1983. L’abron. // Hérault, G. (ed.), Atlas des langues kwa de Côte d’Ivoire 1,

Paris & Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 83-128.

Tresbarats, Chantal. 1983. L’abidji. // Hérault, G. (ed.), Atlas des langues kwa de Côte d’Ivoire 1, Paris & Abidjan: Agence de Coopération Culturelle et Technique; Institut de Linguistique Appliquée, Université d’Abidjan, pp. 41-65.

Welmers, William. 1946. A Descriptive Grammar of Fanti. // Language, vol. 22, no. 3, Language Dissertation no. 39, pp. 3-78.

Werner, Alice. 1919. Introductory Sketch of the Bantu Languages. London: Kegan Paul, Trench, Trubner & Co.; New York: E.P.Dutton & Co.

Westermann, Diedrich. 1903. Die Logbasprache in Togo: Kurzer Abriss der Grammatik und Texte. // Zeitschrift für Afrikanische und Ozeanische Sprachen, 7, pp. 23-39.

Westermann, Diedrich. 1922. Vier Sprachen aus Mitteltogo: Likpe, Bowili, Akpafu und Adele, nebst einigen Resten der Borosprachen. // Mitteilungen des Seminars für orientalische Sprachen, 25, S. 1-59.

Westermann, Diedrich. 1927. Die westernen Sudansprachen und ihre Beziehungen zum Bantu. Berlin: de Gruyter.

Williamson, Kay. 1989. Benue-Congo Overview. // The Niger-Congo languages, ed. by John Bendor-Samuel. NY - London: Lanham, pp. 247-274.

Williamson, Kay. Proto-Ijoid Reconstructions. Manuscript. Williamson, Kay & Roger Blench. 2000. Niger-Congo. // Heine, B. & D.Nurse (eds.) African

languages: an introduction. Cambridge: Cambridge University Press, pp. 11-42. Бабаев К.В. [Babaev K.V.]. 2009. О происхождении личных местоимений в языках мира [On the

Origins of Personal Pronouns in the World’s Languages]. // Вопросы языкознания [Voprosy yazykoznaniya], no. 4 (2009), pp. 119-138.

Рамстедт Г.И. [Ramstedt G.I.]. 1957. Введение в алтайское языкознание [Introduction to the Altaic Linguistics]. Мoscow.