Heine Bernd, Nurse Derek. A Linguistic Geography of Africa
Подождите немного. Документ загружается.
in verb forms which in German tend to be expressed with markers such as
weg-orhi n-(i.e.wegtragen ‘carry away,’ hinlegen ‘put down’). Some
sixty-five years later, Masica (1976) pointed out that this property is
commoninmanyverb-finalsouth-Asian and central-Asian languages, as
well as in the Ethiopian area. And, as illustrated below, verbal com-
pounding is also common in other African languages with a predominantly
verb-final syntax. In order to illustrate the nature of this type of con-
struction, I will first present a more detailed analysis of one Om otic lan-
guage, Wolaitta, followed by a compar ison with genetically related as wel l
as unrelated languages.
From a formal point of view, Wolaitta verbal compounds are best treated as
a specialized (lexicalized) type of converb plus main verb construction.
Wolaitta uses a number of converb markers, as shown above, alternating for
gender (masculine versus feminine) as well as for coreferential versus switch
reference marking. In the compound type the subject of the converb and the
main verb are always identical; the converb in such cases is marked by either -a
(feminine) or -i (masculine). Moreover, objects never intervene between
converb and main verbs in verbal compounds. An example:
(64) ?ı
´
ba keett-aa [baizz-i
?ekk-iisi]
3MSG:NOM LOG house-M:ABS sell-CNV take-3MSG:PERF
‘He sold his house’
There is also a prosodic clue that the converb plus main verb constitute a
lexical unit in Wolaitta, because the main verb does not carry a high-tone pitch
accent (65), as would be the case in a freely generated conver b plus main verb
construction (66):
(65) ?i
maay-u
´
wa meec’c’-ı
´
dı
´
mı
´
c’c’-iisi
3MSG:NOM cloth-M:ABS wash-CNV hang-3MSG:PERF
‘Having washed the cloth, he hung it up’
(66) bita
´
nee ?oo
´
s-uwa wurs-ı
´
dı
´
R
emp-ee
´
si
man:NOM work-M:ACC finish-CNV rest-3MSG:IPF
‘Having finished work, the man is resting’
We thus observe a discrepancy in Wolaitta between the grammatical and the
phonological word in the case of verbal compounds. Interestingly, however,
the prosodic structure of the latter is identical to that of nominal compounds in
Wolaitta, where there is also one high pitch only within the phrase, rather than
each word or stem having its own pitch accent (Amha 1996:133).
It is useful in Wolaitta to distinguish between symmetrical and asymmet-
rical compoundin g. In both types the main verb bears the full range of
inflectional information (person, number, gender, tense–aspect, modality) as
well as negation of verbs. With the former symmetrical type the converb and
Gerrit J. Dimmendaal302
the main verb may be argued to contribute an equal amount of semantic
information, as with ‘be too tight’, lit. slap-hold, below.
(67) hagesu
´
re ne
´
na bak’k’i
?oyk’-iisi
this.M trousers:NOM 2SG:ABS slap-CNV hold-3.
MSG.PF
‘These trousers are too tight for you’
With the so-called asymmetrical type on the other hand, the main verb is
selected from a closed set, whereas the converb is freely generated. Although
formally it is clear that the converb is a dependent verb form with asymmetrical
(as well as with symmetrical) verbal compounds, the main verb appears to
modify the meaning of the preceding converb, more specifically the Aktion-
sart, in the case of asymmetrica l compounds.These main verbs may also occur
as an independent main predication. (For a full list of asymmetrical verbs in
Wolaitta, see Amha & Dimmendaal 2006b.)
(68) harga
´
nc-ı
´
ya ?aa
´
??-a
´
wo
´
??-aasu
patient-F:NOM pass-CNV descend-3FSG:PERF
‘The patient (F) turned over’
(69) ma
´
rz-iya ?u
´
y-i
di
hai
k’k’-i ?agg-iisi
poison-ACC drink-CNV die-CNV give up-3SG:PERF
‘Having taken the poison, he died instantly’
(70) siy-i
?ak’-ibeenna
hear-CONV spend.the.night-3M:PERF:NEG:DECL
‘He did not hear receptively’
It is also common in Wolaitta to form lexical compounds with a verb ‘say,’ as
in the following example:
(71) bay-i
g-iisi
disappear-CNV say-3M:PERF
‘He suddenly disappeared’
Wolaitta uses ‘say’ in combination with intransitive verbs, whereas transitive
verbs are combined with a verb ‘do’:
(72) fu
´
ttu g- ‘to come, to appear unexpectedly’
k’o
´
rc’u ?oott- ‘to swallow something quickly’
‘Say’ constructions are mentioned as an areal feature of Ethiopian lan-
guages by Ferguson (1970). But the same property is in fact widespread in
the typological zone to which Wolaitta and other Afroasiatic as well as
Nilo-Saharan languages belong, as pointed out above. The prosodic clue for
the composite structure of such constructions does not necessarily involve
tonal reduction. In the Nilo-S aharan language Nyimang, a language with
three register tones, the verb root ‘say’ alternates for vowel harmony,
depending on whether the preceding verb root or ideophonic adjective
Africa’s verb-final languages 303
contains [ATR] or [þATR] vowels, showing that ‘say’ is phonologically
bound:
(73) u
`
n
€
a
˛-
se
`
e
`
bow-say
bend’
‘bow, bend’
(74) j
Erj
Er-
s
E
E
IDEO-say
‘scatter’
Interestingly, verbal compounding is not a prominent property of all lan-
guages belonging to this typological zone in northeastern Africa, in spite of
the fact that they com monly appear to have converb plus main verb con-
structions. In Haro, which also belongs to the Ometo group within Omotic,
only compounds with ‘say’ appear to be common . As argued above, the
verbs from the so-called asymmetrical set in Wolaitta affect the Aktionsart
(or “derivational aspect”) of verbs. In Haro, these symbolic functions are
expressed through other strategies, e.g. the use of manner adverbs. Th is
latter language also has a more extensive system of tense–aspect as well as
modality marking than Wolaitta does. The presence of these properties
would seem to make the use of asymmetrical compounding in Haro
superfluous.
Verbal compounding is also common in Nilo-Saharan languages with
converbs; see Ch. and M. le Coeur (1956) on the Saharan language Teda, or
Armbruster (1960) on Dongolese Nubian, or Tucker and Bryan (1966: 344), as
well as Bender (1996: 34–5, et passim) on Kunama. Bender (1996: 36) also
makes reference to compound verbs in Kunama, which differ from converb
plus main verbs in that the former “act like a single verb, i.e. they take person/
number and T[ense]M[ood]A[ spect] affixes like single verbs. Converbs ...are
subordinate verbs marked for TMA ...”
Verbal compounding also appears to be attested in Ijoid languages.
Jenewari (1977: 535) discusse s so-called “auxiliary verbs” in Kalab
_
ari
_
which
seem to affect the Aktionsart expressed by the preceding verb. For example,
the verb ‘complete, finish,’ which is a derived form consisting of a root fa
´
- plus
a causative marker -m
a¯
, may be combined with verbs in order to express
totality:
(75) o si fa
´
ma¯mri
_
he be.bad finish.GEN
‘He is becoming totally bad’
(76) o
_
mangi
_
la
´
ma¯
m
he run reach.FAC
‘He ran very well’
Gerrit J. Dimmendaal304
From the existing data on Ijo it is not clear to what extent transitivity, more
specifically the presence of coreferential objects, plays a role, since all
examples appear to be with intransitive verbs.
Kilian-Hatz (2006) discusses verbal compounds in Khoe involving motion
verbs as the first component. Unlike other types of serial verbs forming a
complex predicate, verbal compounds are necessarily contiguous in Central
Khoisan languages like Khoe:
(77) c--xu
`
u
´
take-I-leave
‘take away’ (xuu ‘leave’)
(78) ya
´
a-a- ko
˜
a
˜
climb-I-go.down
‘climb down’ (high tone on oa)
(79) mu
˜
u
˜
-a-a
˜
see-I-know
‘understand, realize, recognize (high tone on nasal vowels)
(The epenthetic vowel, harmonizing with the prec eding root vowel, is a
transitivity marker.)
Another domain where lexicalization appears to have taken place in Khoe
(to such an extent that the verbs from the closed list may be considered der-
ivational suffixes, according to Kilian-Hatz 2006) is with two verbs affecting
Aktionsart (‘leave’ and ‘remain’) and valency (‘distribute to’):
(80) xuu ‘leave’ -xu ‘do completely, be done already’
ei ‘remain’ -ei ‘firmly, tightly, well’
We thus observe in language after language that grammaticalized ways of
expressing meaning may emerge from the semantic interaction between
adjacent verbs if coreferential subjects (or agents) are involved, and if no
other constituent (e.g. an object) intervenes. These latter conditions pre-
sumably s et the switch for semantic interaction between t he main verb and
the adjacent dependent verb. The emergence of verbal compounding
accordingly would seem to follow from so-called self-organizing principles.
Verbal compounding in languages of northeastern Africa as well as southern
Asia shows that such systems may emerge independently of each other.
Accordingly, one may observe the emer gence of parallel s truc tures between
languages without either genetic inheritance or areal diffusion being neces-
sarily involved. This process, where the organization of a system spontan-
eously increases without this increase be ingcontrolledbyanexternalsystem
(e.g. the environment) is known as “self-organization” in the natural sci-
ences. One may consequently observe the emergence and development of
Africa’s verb-final languages 305
verbal compounding as new, complex s tructures taking place in and through
the system itself, i.e. from internal or self-organizing principles; the lin-
guistic parallel t o the external or environmental factor would be areal contact
with other languages. This is not to say t hat constructions require adjacency
in order to build a complex predicate, since morphological coindexing or a
fixed order may accomplish the same. But in the absence of such properties,
adjacency is the more likely trigger for semantic interact ion. The semantic
range covered by such verbal compounds will virtually never be identical
even between closely related languages, as we saw for Wolaitta and Haro
above. Whether such specialized uses emerge and become conventionalized
or grammaticalize d would seem to depend on the frequency with which such
collocations are used. This in turn would seem to be related to the question to
what extent other construction types (e.g. manner adver bs, tense –aspe ct
markers) may render the same meaning. Clearly, then, interaction with other
grammatical subsystems determines the productivity of verbal compounding
and the lexicalization of such structures.
9.6 Extending the typology or accepting the limits of
typological research?
In a somewhat destructive spirit, I have dissolved the concept of verb-final
languages as a type, not out of intention, but because this seems to be the only
logical conclusion to be drawn from descriptions of so-called verb-final lan-
guages in Africa which have become available after the seminal contribution
by Heine (1976).
First, I observed that there is an important difference typologically between
languages in their degree of rigidity with respect to constituent order. I
observed a similar problem with respect to so-called type B languages, some of
which indeed follow a strict SVO/SAUXOV order, whereas in other languages
the preverbal position may be occupied by any constituent. Similarly, so-called
verb-initial languages differ with respect to rigidity of constituent order. In
Nilotic languages like Nandi, which allows for postverbal scrambling, any
syntactic constituent may occur immediately after the verb, the actual order
being governed by pragmat ic principles. But the position immediately after the
verb is used for focused elements (Creider & Creider 1983). In the Nilotic
language Turkana, on the other hand, the syntactic order is far more rigid and
focus marking is less strict than in Nandi, i.e. not fixed with respect to one
position within the sentence (Dimmendaal 1983b).
Of course, the same kind of variation in terms of constituent order can be
observed between languages which have been claimed to be basically SVO
languages. Compare the following examples from the Bantu language Rundi,
Gerrit J. Dimmendaal306
which is usually characterized as an SVO language. But, as shown by
Ndayiragije (1999), there are various alternatives possible:
(81) a
´
ba
ˆ
na ba-a
´
-ra-nyo
ˆ
ye amata
´
2:children 2-PAST-AFF-drink:PERF milk
‘Children drank milk’
(82) amata
´
y-a
´
-nyo
ˆ
ye aba
ˆ
na
milk S-PAST-drink:PERF 2:children
‘Children (not parents) drank milk’
(83) pro
exp
ha-a
´
-nyoye amata
´
aba
ˆ
na
16-PAST-drink:PERF milk children
‘Children (not parents) drank milk’
For the same reason, i.e. because syntactic order may be governed entirely by
information structure, it may be impossible to determine a basic constituent
order, or it may be claimed that OVS is the basic order, as for a number of
Wester n Nilotic languages (Ander sen 19 88 ; Reh 1996 ; Miller & Gilley 2001).
The important descriptive task to be addressed when analyzing such languages
is the question to what extent syntax adapts to discourse struct ure or vice versa,
and the question which (morphosyntactic and/or constituent order) strategies
these languages use in order to accomplish such interactions.
Second, when studying parts of speech from a crosslinguistic perspective,
we again observe rather dramatic differences. Whereas the categorical dis-
tinction between noun and verb appears to be fundamental to all of the lan-
guages and language groups discussed here (though not necessarily to other
African languages), it is also clear that the status of these categories in a
specific language is determined by the constructions in which they may occur.
The structural behavior of a verb in a verb-final Omotic language is not the
same as in a verb-final Ijoid or Central Khoisan language, as argued above, and
these differences are far from trivial. They relate to inflectional properties of
verbs and their syntactic behavior ; the central status of case marking (as a
dependent-marking strategy) in Omotic for the expression of semantic rela-
tions, for example, also has its impact upon the behavior of finite verbs. Alter-
natively, the conceptual and etymological link between nouns and adpositions,
or instead, between verbs and adpositions, may have consequences for
the structural behavior of these respective categories synchronically (Bernd
Heine, p.c.).
Third, at the clausal level we are in need of more subtle representations than
the traditional, and somewhat outdated, coordination–subordination distinc-
tion, regardless of the type to which a particular language may belong. This
latter position has, of course, been argued for extensively by Van Valin and
LaPolla (1997).
Africa’s verb-final languages 307
On the basis of the dynamic (intragenetic and intergenetic) comparison, I
arrive at the following conclusions: where typological similarities can be
observed between Africa’s verb-final languages, these are due not to some
presumed universal typological principle, but rather to genetic inheritance on
one hand and to areal diffusion on the other. The typological link between
Nilo-Saharan languages in the eastern Sahel region and Afroasiatic languages
of Ethiopia is probably explained best along these lines. In addition, what we
perceive of as a language type or a Gestalt appears to be due to so-called self-
organizing principles of languages; the interaction between different sub-
systems (adjacency and subsequent semantic interaction between constituents;
absence of alternative strategies in a language) may indeed result in typolo-
gically similar phenomena. This applies to verbal compounding, for example,
which may emerge independently without one having to assume genetic
inheritance or area l contact.
Even if we accept the old descrip tivist attitude that every language is unique
in its structure, this does not imply that one’s approach to language description
and analysis is necessarily naı
¨
ve. Knowledge about crosslinguistic variation is
indeed an important heuristic device in describing and analyzing hitherto
undescribed or poorly described languages. And there are still many such
languages on the African continent waiting to be investigated in more detail
along these lines.
The success of future investigations in language typology would seem to
depend on the question to what extent we succeed in showing how linguistic
subsystems in a specific language intera ct. Obviously, such an understanding is
accomplished best through in-depth comparative studies of languages which
are closely related genetically, in tandem with intergenetic comparisons. It is
only through such a dynamicization of subtypologies that one may arrive at an
ultimate understanding of the portrait of a language as well as of important
similarities or differences in the organizational structure of languages.
Gerrit J. Dimmendaal308
Notes
1 INTRODUCTION
1 In a more technical format, this tool for identifying instances of contact-induced
transfer has been described in the following way: ‘‘ If there is a linguistic property x
shared by two languages M and R, and these languages are immediate neighbors
and/or are known to have been in contact with each other for an extended period of
time, and x is also found in languages genetically related to M but not in languages
genetically related to R, then we hypothesize that this is an instance of contact-
induced transfer, more specifically, that x has been transferred from M to R’’ (Heine
& Kuteva 2005: 1.4.6).
2 Since the Hadza language of Tanzania has basic VSO order, one might say that all
three word-order types are found within the Khoisan phylum. However, Hadza’s
genetic position is controversial; as we will see in chapter 6, there are good reasons
to treat it as a genetic isolate.
3 We are ignoring here the fact that verb-final negation is also found in the fourth
phylum, Khoisan: the Central Khoisan language Khwe places its negation marker
be
´
at the end of the clause. The reason for ignoring Khoisan is that it does not
appear to be genetically or areally related to the Niger-Nile belt of verb-final
negation.
4 There is a partial exception: the Central Khoisan language kAni is reported to lack
subject agreement on the verb but to commonly exhibit object agreement (Voßen
1985).
2 IS AFRICA A LINGUISTIC AREA?
1 We are grateful to Felix Ameka, Joachim Crass, Gerrit Dimmendaal, Ludwig
Gerhardt, Tania Kuteva, Martine Vanhove, Derek Nurse, Marie-Claude Vandame,
and Erhard Voeltz for valuable comments and information on the subject matter of
this chapter. Our gratitude is also due to the roughly fifty colleagues who have
provided us with valuable data on a wider range of African languages.
2 In a number of works discussing linguistic areas, Joseph (1983) is cited as the
primary reference work, or at least as one of the primary works on the Balkan
sprachbund. However, while this work constitutes the most detailed study on the
sprachbund, it deals essentially only with one linguistic property making up this
area.
3 For a critical review, see Tosco (2000b).
309
4 Cf. also the following statement according to which areal linguistics ‘‘deals with
the results of the diffusion of structural features across linguistic boundaries’’
(Campbell et al. 1986: 530).
5 ‘‘This term ‘linguistic area’ may be defined as meaning an area which includes
languages belonging to more than one family but showing traits in common which
are found not to belong to the other members of (at least) one of the families’’
(Emeneau 1956: 16, n. 28).
6 It is possible that the presence of gender systems in the Eastern Nilotic languages
(Maa, Teso-Turkana, Lotuxo, Bari) is the result of language contact with Cushitic
languages, but the evidence on this issue is far from conclusive.
7 According to Mike Cahill (p.c.), ‘‘perhaps 20 languages of Papua New Guinea
have labial-velar stops, including Kate, Dedua, Kube, Ono, Fulumu, Amele, as
well as Yeletnye, which uniquely not only has /kp, gb/, but also phonemic post-
alveolar /tp, db/ as well. ’’ Furthermore, he adds that Santa Ana of the Solomon
Islands has /gb/.
8 Greenberg (1983:16) says, however, that this is not always so: ‘‘The most
conspicuous exception is the Grasslands languages where the form bep or the like
is found in many languages with the meaning ‘meat’ while the nama root survives
as ‘ animal.’‘‘It would seem that this fact does not invalidate the hypothesis of a
directionality ‘meat’ > ‘animal’; rather, it might suggest that – for whatever
reasons – an earlier meaning ‘meat’ received a new form of expression.
9 This is to express our gratitude to all our colleagues who have contributed to the
survey.
10 ‘‘In these maps the demarcation lines, called isopleths, mark off areas with
languages displaying the same number or plethora of features, but not necessarily
the same features’’ (van der Auwera 1998: 260).
3 AFRICA AS A PHONOLOGICAL AREA
We would like to thank a number of colleagues for their generous help in responding
to queries and contributing language data, in particular Derek Nurse, Gerrit Dim-
mendaal and Mohamed Elmedlaoui, as well as Zygmunt Frajzyngier, Tom Gu
¨
lde-
mann, and Larry Hyman for reading and sending comments on the final draft version.
We give special thanks to Raphae
¨
l Kabore
´
and the Bibliothe
`
que Africaine of the
Universite
´
de Paris 3, whose resources greatly aided us in compiling data on lesser-
known languages.
1 The term ‘‘Sudanic languages’’ is used here and below as a convenient shorthand
for ‘‘languages spoken in the Sudanic belt,’’ and does not refer to any presumed
genetic grouping. Sudan as a geographical term enjoys a long tradition, predating
later linguistic adaptations. The term Sudanic in this historical sense is not to be
confused with Sudanese, referring to the land and people of the Republic of the
Sudan, nor with Central Sudanic and East Sudanic, designating genetic subunits of
the Nilo-Saharan language family.
2 We identify languages in terms of the traditional Greenberg-derived classification
for convenience, though not all its proposals are accepted by all scholars. In
particular, many specialists prefer to treat the ‘‘Khoisan’’ languages as a grouping of
as many as five unrelated families.
Notes to pages 16–39310
3 The full database draws on the (mostly non-Bantu) African phoneme systems
collected in UPSID (UCLA Phonological Segment Inventory Database,
Maddieson & Precoda 1989), the Bantu phoneme systems collected in Nurse
and Philippson (2003), and a variety of other published sources. All non-African
languages are drawn from UPSID. We regret that space limitations preclude our
citing sources for all language data mentioned in this survey. Standard sources
have been used whenever they were available to us, and the best reliable sources
were used otherwise.
4 Shona, a Bantu language spoken in Zimbabwe.
5 Amele, a language of Papua New Guinea, and Iai, a Malayo-Polynesian language
spoken on Ouve
´
a Island in New Caledonia.
6 Alphanumeric codes such as S10 refer to Guthrie’s system of Bantu language
classification, as updated and amplified by Maho (2003). We follow the current
preference for referring to Bantu languages without their prefixes, e.g. Ganda
rather than Luganda, Swahili rather than Kiswahili. In citing languages here and
below we use the following conventions:‘‘X/Y’’ indicates alternate names for the
same language, ‘‘ X-Y’’ indicates closely related languages or members of a dialect
chain (exception: Diola-Fogny is a single member of the Diola cluster), and dialect
names precede language names: Dendi Songay, Owere Igbo.
7 An exception to this generalization occurs in the variety of Ma’di described by
Blackings and Fabb (2003), where the prenasalized stop [m(n)gb] begins with
labial closure; the existence of a velar closure is reported as uncertain.
8 It is mainly in languages which lack voiced stops that we find /kp/ to the exclusion
of /gb/. Excluding such languages, Cahill (1999) finds that languages having [gb]
alone outnumber those with [kp] alone by a significant margin. One consideration
that might explain such a trend is the occasional tendency for labial-velar stops to
have implosive realizations, as in the case of the Nigerian languages Idoma, Isoko,
and Igbo (Ladefoged 1968; Ladefoged & Maddieson 1996); implosives are of
course normally voiced.
9 Our sources include Richardson ( 1975), Guthrie (1967–71), Tylleska
¨
r(1986–7),
Mutaka and Ebobisse
´
(1996–7), Gre
´
goire (2003), and Mangulu (2003), among
others.
10 Guthrie (1967–71: vol. 3, 303–4, vol. 4, 16), Nurse and Hinnebusch (1993: 171–3).
11 Our sources are Schachter and Fromkin (1968), Le Saout ( 1973), Bentinck (1975),
Singler (1979), Capo (1981, 1991 ), Bole-Richard (1983a,b, 1984), Ihionu (1984),
Creissels (1994), and Clements and Osu (2005), to which we have added
languages drawn from inventories in Bole-Richard (1985), Maddieson (1984), and
Cohn (1993a,b). The zone in question is a sprachbund, characterized by a complex
of other features such as a strong tendency toward monosyllabism, ‘‘horizontal’’
(that is, front–back) root harmony, three or more distinctive tone levels, and
certain ‘‘lax ’’ question markers. We discuss the latter two features in sections 3.3.2
and 3.3.3.
12 Our transcriptions are phonemic; stops are voiced intervocalically.
13 Languages like Yoruba or Kikuyu, with one high vowel series, two mid vowel
series and constraints requiring consecutive mid vowels to be of the same height,
have also been described as having ATR harmony. In such systems, unlike those
with two series of high vowels, [ATR] is usually the active value, at least in
African languages (Casali 2003). See below for illustrations from Kikuyu.
Notes to pages 40–51 311