Patterns of noun class attrition in Kwa

Maria Konoshenko, Daria ShavarinaRussian State University for the Humanities, Moscow

Why Kwa classes?

Kwa languages are usually viewed as being close to the isolating

prototype (Aboh & Essegbey 2010: 40), cf. also the notions of Bantu-

type with robust inflection vs. Kwa-type with minimal inflection in

Hyman (2004), Good (2012)

class systems in Kwa are understudied in typological as well as historical

perspectives

this study focuses on some synchronic properties of noun class systems

in modern Kwa in the broader context of “canonical” vs. “reduced”

(Katamba 2003: 108) class systems in Niger-Congo 2

Where are classes in Kwa?

Classes are attested in genetically remote groupings:

Ka-Togo and Na-Togo aka Ghana-Togo Mountain languages (GTM)

the Guang languages

Akanic lects, Ega

usually described as “rudimentary”, “vestigial” (Welmers 1973: 184),

“residual” (Aboh 2010: 11), “defunct” (Osam 1994: 117), “undergoing

decay” (Essegbey 2009: 37), but many systems are in fact quite

elaborate 3

Ka-Togo Na-TogoGbe

Kwa

Potou-Tano

Potou Tano

Guang

(Williamson and Blench 2000)

Proto-Kwa as having a full-fledged class system

full-fledged sets of noun class markers have been reconstructed for

proto-GTM language(s) (Heine 1968) and proto-Guang (Manessy

1987; Snider 1988)

noun class prefixes in modern Kwa are cognate to class markers in other

Niger-Congo subfamilies and can be traced back to proto-Niger-

Congo (Blench 2009; Kiessling 2013: 46)

Hence, we may safely assume that a complex noun class system was

present in proto-Kwa as well, but it was lost in most modern groups of

the family. 5

Our assumptions about classes in proto-Kwa

Proto-Kwa as having a “typical” Niger-Congo class system:

(1) Class marking on nouns: most nouns are assigned obligatory class

markers from a limited set, different in singular and plural forms;

(2) Class agreement: nominal modifiers, verbs and anaphoric

pronominals obligatory agree with nouns in class (and number)

Cf. (Kiesling 2013:44) for similar generalizations about typical noun class

systems in Niger-Congo

6

Languages in the sample

A sample of 15 languages limited by the availability of descriptions

but still rather representative:

GTM: Logba, Lelemi, Akebu, Avatime, Nyangbo, Igo, Tuwuli, Ikposo

Guang: Chumburung, Foodo, Gonja, Nkonya, Nawuri, Efutu

Akan (Fante and Twi lects)

Note: a database of class marking in Kwa is being prepared via Google Drive and

can be shared with anyone who would like to get access to our data.7

The data

CLass systems in Kwa quantified by the number of unique class

markers on nouns:

8

Number of class markers System type

≥ 8 markers Elaborate

< 8 markers with singular-plural pairing preserved for most nouns

Moderate

traceable markers, but singular-plural pairing lost for most nouns

Vestigial

Languages Grouping Main source Class marking on nouns

Logba GTM: NaT Dorvlo 2009 Moderate

Lelemi GTM: NaT Allan 1973 Elaborate

Akebu GTM: KaT>KeA Makeeva, Shluinsky 2015 Elaborate

Avatime GTM: KaT>AvN Schuh 1995 Elaborate

Nyangbo GTM: KaT>AvN Essegbey 2009 Elaborate

Igo GTM: KaT>KpAB Gblem 1995 Elaborate

Tuwuli GTM: KaT>KpAB Harley 2005 Elaborate

Ikposo GTM: KaT>KpAB Soubrier 2013 Vestigial

Chumburung Guang>Northern Hansford 1990 Moderate

Foodo Guang>Northern Plunckett 2009 Elaborate

Gonja Guang>Northern Painter 1970 Moderate

Nkonya Guang>Northern Reineke 1971 Elaborate

Nawuri Guang>Northern Sherwood 1982 Moderate

Efutu Guang>Southern Obeng 2008 Vestigial

Akan Central Tano Osam 1993ab, Osam 1994 Moderate9

Class marking on nouns (1)

The correspondence of several singular markers to one plural marker (a)

as a result of (apparent) generalization:

attested in many languages, but could already be present in the

proto-system...

so only “gross” convergence conditioned by external factors is

treated as innovative simplification

E.g. animacy-based affix overgeneralization in plural formation in Lelemi

(Allan 1973) 10

Animacy-based plural formation in Lelemi (Allan 1973)

11

Singular Plural

Animate humans o- ba-

animals a- ba-

tribes, birds, bats le- ba-

animals living in the bush ka- ba-

animate objects ko- ba-

onomatopoeic or borrowed animates ∅- ba-

Inanimate artefacts and domesticated crops, body parts o- le-

kitchen utensils, body parts, natural events le- a-

Class marking on nouns (2)

(b) non-segmental class marking, e.g. in Akebu (Makeeva, Shluinsky

2015)

(c) fossilization of class marking, e.g. in Ikposo (Soubrier 2013)

(d) loss of class marking, e.g. in Akan (Osam 1993), Nyangbo (Essegbey

2009)

12

Non-segmental class marking in Akebu

13

In Akebu, some singular classes are marked with tone rather than

segmental prefix (Makeeva & Shluinsky 2015):

Singular Plural Semantics

Segmental marking kɨ-ʈoo-kǝ wə-ʈoo-kpǝ ‘feather’

Tonal marking lee-ʈǝ e-lee-yǝ ‘calabash’

Fossilization of class marking in Ikposo

Most nouns have V-CV structure with initial vowels as remnants of

ancient class markers; only some human nouns change initial vowel in

plural (Soubrier 2013: 75-83)

14

Singular Plural Semantics

Humans ɔ-sɩ

ɔ-lʊ

a-sɩ

a-lʊ

‘woman’‘person’

Non-humans utʃu

ʊɣɔ

utʃu nɩ

ʊɣɔ nɩ

‘fly’‘hand’

Animacy-based restructuring in Kwa and Niger-Congo

Human nouns in Kwa have more conservative class marking as opposed

to non-human, e.g. they retain prefixal plural formation in Ikposo, they

never lose prefixes in singular in Akan etc.

Animacy-based restructuring of class systems is a well-known process in

Niger-Congo (Maho 1999, Good 2012)

Cf. similar generalization by Faraclas (1986: 51): “class/gender distinctions

distinctions are lost for [-human] nouns before [+human] nouns” for

Cross-River15

Class agreement

A complex parameter comprising:

features

targets

markers

conditions

cf. Corbett (2006: 4-5)

16

Agreement features

Types of agreement systems attested:

“conservative” class (and number) agreement - multiple semantic cores,

multiple agreement markers:

Logba, Lelemi, Akebu, Tuwuli, Chumburung, Foodo

animacy (and number) agreement - a single semantic parameter, few

agreement markers:

Igo, Gonja, Nkonya, Nawuri, Akan

vocalic and tonal agreement - semantically not motivated:

Avatime, Nyangbo

no agreement: Ikposo, Efutu 17

Animacy-based agreement: Nkonya

Numerals from 2 to 9 (Reineke 1972: 45-46)

18

Humans aba- a-tse aba-sa ‘3 women’

n-yebi aba-sa ‘3 children’

Animate non-humans

a- m-boe a-sa ‘3 animals’

a-kpokpo a-sa ‘3 buffalos’

Inanimate ∅- e-kolo ∅-sa ‘3 boats’

a-wulu ∅-sa ‘3 towns’

Animacy-based agreement: Igo (Gblem 1995: 153-157)

Numerals: human vs. non-human distinction

Indefinite markers: animate vs. inanimate

19

Singular Plural

Human i- iŋo-

Non-human i- ini-

Singular Plural

Animate u- bu-

Inanimate ki- ke-

Vocalic agreement: Nyangbo (Essegbey 2009)

Most targets have tV- prefix copying its vowel from noun prefix:

20

Feature Noun prefix: CVi Prefixes on indefinite pronouns and numerals: tVi

Singular A- tA-

Plural bA- tA-

Singular O- tO-

Plural I- tI-

Singular kA- tA-

Plural bU- tU-

Agreement targets

Most common targets:

pronouns (subject, object, possessive, independent/emphatic)

numerals, demonstratives, indefinite markers (‘some’), but not

adjectives!

21

22

Numerals Adjectives

Logba YES NO

Lelemi YES NO

Akebu YES NO

Avatime YES NO

Nyangbo YES NO

Igo YES NO

Tuwuli YES YES*

Chumburung YES YES*

Foodo YES YES

Gonja YES NO

Nkonya YES NO

Nawuri NO NO

Akan YES YES*

Agreement on adjectives in Kwa

the only language with (apparently) unrestricted class agreement on

adjectives is Foodo (Plunckett 2009: 124-125)

in the other three languages with adjective agreement (Akan, Tuwuli,

Chumburung) only a few adjectives agree, and in number rather than

class (Osam 1993: 97-98; Harley 2005: 111; Hansford 1990: 266)

23

Adjective agreement in Akan

In Akan, some adjectives get class marker, but only in plural. Strangely,

the class form of the adjective does not depend on the class form of

noun and is unpredictable (Osam 1993: 97):

(1) a-tar tuntum ‘black dress’

Cl2-dress black

(2) n-tar e-tuntum ‘black dresses’

Cl5-dress PL-black

24

When was adjective agreement lost? (1)

Two possible scenarios:

(1) adjective agreement was present in the proto-system, but is the first

to disappear in modern languages

provides a better explanation for lexical restrictions and

idiosyncrasies in Akan, Tuwuli and Chumburung

25

When was adjective agreement lost? (2)

Two possible scenarios:

(2) adjective agreement had already been lost in Proto-Kwa, and modern

cases are recent innovations

better explains why languages from different subgroups pattern

similarly

why otherwise rather conservative GTM almost lack adjective

agreement

26

Adjective and numeral agreement in Niger-Congo

Numeral agreement is very robust in Kwa as opposed to adjectival agreement

which seems to have been lost either in the proto-Kwa or in most its

daughter languages;

This pattern is opposite to what is attested in neighboring Niger-Congo

families, e.g. in Kru and Cross River languages, which are more likely to

preserve adjective rather than numeral agreement (Demuth et al. 1986;

Faraclas 1986)

Cf. Hierarchy of persistence in agreement patterns for Cross River: Strongest: subject-verb

Strong: adjective-nounWeak: other targets, especially numerals (Faraclas 1986: 51) 27

Conclusions

patterns of simplification in noun class systems in Kwa, e.g.

animacy-based restructuring, largely correspond to diachronic

phenomena attested in other Niger-Congo subfamilies

however, numerals and adjectives pattern differently in Kwa as

opposed to (some) other Niger-Congo subfamilies: numerals

preserve class agreement much better than adjectives

the idea that there was no adjective agreement in proto-Kwa is

not impossible 28

References (1)

Aboh, Enoch O. 2010. The Morphosyntax of the Noun Phrase. In: E. Aboh & J. Essegbey (eds). Topics in Kwa Syntax. Studies in Natural Language and Linguistic

Theory 78. 11-37.

Allan, E. J. 1974. A Grammar of Buem. Ph.D . dissertation, University of London.

Blench, R. 2009. Do the Ghana-Togo mountain languages constitute a genetic group? The Journal of West African Languages 36.1/2: 19-36.

Demuth, K. A.; Faraclas, N. G.; Marchese, L. 1986. Niger-Congo noun classes and agreement systems in language acquisition and historical change. In: Noun

classes and categorization, p. 453-471. Ed. by Colette Grinevald Craig. Typological studies in language, v. 7. Amsterdam & Philadelphia: John Benjamins Publ.

Essegbey, James. 2009. Noun classes in Tutrugbu. Journal of West African Languages. 37-56.

Faraclas, Nicholas G. 1986. Cross River as a model for the evolution of Benue-Congo nominal class/concord systems. In Studies in African Linguistics 17 (1): 39-54.

Gblem, H. M. 1995. Description systematique de l'Igo. Thèse de doctorat. Grenoble: univ. de Grenoble III.

Good, J. 2012. How to become a “Kwa” noun." Morphology 22.2.

Harley, M. 2005. A descriptive grammar of Tuwuli. Ph.D. dissertation. London: SOAS.

Heine, B. 1968. Die Verbreitung und Gliederung der Togorestsprachen. Berlin: Dietrich Reimer Reimer

Hyman, L. 2004. How to become a “Kwa” verb. Journal of West African Languages 30:69–88.

Katamba, F X. 2003. Bantu nominal morphology. In: The Bantu languages, p. 103-120. Derek Nurse & Gérard Philippson (eds.). Language family series, v. 4.

London & New York: Routledge.

Kießling, R. 2013. On the origin of Niger-Congo nominal classification. In Historical Linguistics 2011: Selected papers from the 20th International Conference on

Historical Linguistics, Osaka, 25-30 July 2011 (Vol. 326, p. 43). John Benjamins Publishing.

Maho, J. F. 1999. A comparative study of Bantu noun classes. Orientalia et africana gothoburgensia, n. 13. Göteborg: Acta Universitatis Gothoburgensis.

29

References (2)

Макеева Н.В., Шлуинский А.Б. Именные классы и согласование в языке акебу // Исследования по языкам Африки, вып. 5. М., 2015. С. 174-200.

Manessy, Gabriel. 1987. La Classification nominale en Proto-Guang. Afrikanistische Arbeitspapiere 9:5-49.

Obeng, S. G. 2008. Efutu grammar. Munchen: Lincom Europa.

Osam, E. K. 1993. The loss of the noun class system in Akan. Acta Linguistics Hafniensia, z, 81 - 106.

Reineke B. 1972. The Structure of the Nkonya Language. Leipzig: VEB.

Schuh, R. 1995. Avatime noun class and concord. Studies in African Linguistics Vol. 24, No. 2: 123-149.

Snider, K. L. 1988. The noun class system of proto-Guang and its implications for internal classification. Journal of African languages and linguistics, v. 10, 2, p.

137-163.

Storch, A, Yao K. 2000. Noun Classes and Consonant Alternation in Akebu (Kəgbərəkə) // Meissner, Antje, & Anne Storch (eds.). Nominal Classification in African

Languages. Frankfurter Afrikanistische Blätter 12. Köln: Rüdiger Köppe Verlag.. 79-98.

Subrier, A. 2013. L’ikposso uwi. Phonologie, grammaire, textes, lexique. Thèse de doctorat. Lyon: Université Lumière Lyon 2.

Welmers, W. 1973. African language structures. Berkeley: University of California.

Williamson, K, R. Blench.. 2000. Niger-Congo. In Heine, Bernd, & Derek Nurse (eds.). African languages: an introduction. Cambridge: Cambridge University press.

11–42.

de Wolf, P. 1971. The noun class system of Proto-Benue-Congo. The Hague: Mouton.

30