Content uploaded by Gerrit Jan Dimmendaal
Author content
All content in this area was uploaded by Gerrit Jan Dimmendaal on Feb 18, 2019
Content may be subject to copyright.
1
To appear in: Oxford Research Encyclopedia of Linguistics
A typological perspective on the morphology of Nilo-Saharan languages
Gerrit J. Dimmendaal
Summary
Nilo-Saharan, a phylum spread mainly across an area south of the Afroasiatic phylum and north
of the Niger-Congo, was established as a genetic grouping by Greenberg (1963). In his earlier,
continent-wide classification of African languages in articles published between 1949 and 1954
(also published as Greenberg 1955), Greenberg had proposed a Macro-Sudanic family
(renamed Chari-Nile in subsequent studies), consisting of a Central Sudanic and an Eastern
Sudanic branch plus two isolated members, Berta and Kunama. This family formed the core of
the Nilo-Saharan phylum as postulated in Greenberg (1963), where a number of groups were
added which had been treated as isolated units in his earlier classificatory work: Songhay,
Eastern Saharan (now called Saharan), Maban and Mimi, Nyangian (now called Kuliak or Rub),
Temainian (Temeinian), Coman (Koman) and Gumuz.
the more than 140
languages belonging to this phylum is impossible in such a brief study, particularly given the
tremendous genetic distance between some of the major subgroups. Instead, typological
variation in the morphological structure of these genetically related languages will be central.
In concrete terms, this involves synchronic and diachronic observations on the formal properties
of these languages (section 2), followed by an introduction to the nature of derivation,
inflection, and compounding properties in Nilo-Saharan (section 3). This traditional
compartmentalization has its limits because it misses out on the interaction between lexical
structures and morphosyntactic properties in the members of the group, as argued in section 4.
As pointed out in section 5, language contact must also have played an important role in the
geographical spreading of several of these typological properties.
Keywords: alignment, alloying, associated motion, contextual inflection, evidentiality,
frequency of usage, increment, inherent inflection, learnability, multi-verb construction, non-
concatenative morphology, (non-)configurational, noun incorporation, paradigmatic
displacement, univerbation
2
1. Introduction
-Saharan family in 1963, scholars have
disagreed on the subgrouping or the inclusion or exclusion of specific language families. This
applies, for example, to the Kadu languages, spoken along the southern range of the Nuba
Mountains, Sudan, which had been classified as Kordofanian, i.e. as members of the Niger-
Kordofanian phylum by Greenberg (1963), but which has been argued by Bender (1996) to be
part of the Nilo-Saharan phylum. Ehret (2001: 68) assumes that the Kadu languages, as well as
the Shabo language in southwestern Ethiopia (usually treated as a linguistic isolate),
possibly be related at some deeper remove to Nilo-
Map 1 The spreading of Nilo-Saharan languages
3
One cannot prove that languages are unrelated presumably, all human languages are.
But genetic classifications are only interesting if they help us to understand historically related
but diverging systems where a common ancestry provides the most plausible hypothesis. It is
for this reason that the Kadu group is not further discussed below, and why a number of other
families which may or may not be part of Nilo-Saharan are discussed only briefly. This applies
to Songhay (spoken mainly in the vicinity of the River Niger in West Africa) as well as to
Koman and Gumuz (in the border area between Ethiopia and Sudan). More recently, two
languages have been identified which are closely related to Gumuz and Kadallu. This
Again, because the Kuliak
languages constitute an isolated unit (and authors such as Sands (2009) claim that they are not
part of Nilo-Saharan), they are not further discussed here; see Dimmendaal et al. (2019), who
give additional details, and who provide arguments for the following subgrouping as a basis for
a typological survey of Nilo-Saharan.
Figure 1 The subclassification of Nilo-Saharan.
Maban
Fur and
Amdang
Kunama
Saharan
Eastern
Sudanic
Kuliak
Central Sudanic
Nilo-Saharan
Northeastern
Koman,
Songhay
4
The largest subgroup in terms of number of languages and subgroups is the Eastern Sudanic
branch of Northeastern Nilo-Saharan, which also manifests an interesting typological break at
the morphosyntactic level, as illustrated in section 4.
Figure 2 The subclassification of Eastern Sudanic
2. Formal characteristics
In terms of traditional morphological typology, Central Sudanic languages may be
characterized as more analytic in nature compared to their Northeastern Nilo-Saharan relatives,
which tend to be more synthetic; polysynthesis is not attested in this phylum. A number of free
morphemes in Central Sudanic appear to be cognate with bound morphemes in Northeastern
Nilo-Saharan, as illustrated in Dimmendaal (2018), where it is further argued that the common
proto-language appears to have been typologically closer to modern Central Sudanic languages.
Taman
Meroitic (extinct)
Nubian
Nara
Nyimang plus Afitti
Berta
Eastern Sudanic
Northern
Southern
Jebel
Daju
Temeinian
Surmic
Nilotic
5
Cross-linguistically, there is a preference for suffixation as against prefixation, as
observed by Himmelmann (2014: 927), prosody, specifically prosodic
phrasing, plays an important part in bringing about this asymmetry. Nilo-Saharan languages
confirm these observations statistically. Nevertheless, there are also inflectional and
derivational prefixes, which interestingly turn out to be less subject to phonetic erosion or fusion
with lexical roots than are suffixes, thereby also providing robust evidence for the genetic
relationship between these languages.
Among the most stable verbal derivational markers is a causative prefix i- (sometimes
alternating with ɪ-), which has been retained as such in both of the primary branches of Nilo-
Saharan, Northeastern Nilo-Saharan and Central Sudanic. Unseth (1998) presents a survey of
this archaic feature in the Surmic branch of Eastern Sudanic, giving the following examples
from Majang (Ethiopia):
(1)
dibis-
full
i-dibis
(2)
bod-
-bod
Blacking and Fabb (2003: 76) give the following examples for Madi, a Central Sudanic
language spoken in South Sudan and Northern Uganda:
(3)
ś
-ś
In addition, the first and second person singular markers, realized as proclitics in most
Central Sudanic languages and as prefixes in Northeastern Nilo-Saharan, form a fairly stable
structural property of Nilo-Saharan, as already pointed out by Greenberg (1963: 109).
i
Whereas
in Central Sudanic languages the first person singular tends to be a- and the second person
singular i- (as in Lugbara; Tucker and Bryan 1966: 42), reflexes of the first person singular
pronoun in Northeastern Nilo-Saharan languages suggest the presence of an original (probably
uvular) consonant in these pronouns, as in Fur ʔ(a)-, or Temein ŋa- and Baale ka-, the latter two
being Eastern Sudanic languages.
Derivational prefixes other than the causative lost their status as a productive strategy
of word formation in Northeastern Nilo-Saharan (for reasons further discussed in section 5).
These mainly vocalic prefixes were incorporated into verb stems, i.e. increment occurred, in
Maban and Saharan, as well as in the Fur language, where they trigger complex
6
morphophonemic alternations when combined with person-marking subject prefixes. In Fur,
this involves consonant change and deletion, vowel insertion, metathesis, and tonal alternations,
thereby giving rise to a large number of conjugational classes, as shown in the detailed survey
by Waag (2010: 114-134, 308-319). Examples from Waag illustrate some of these alternations
(2010: 119):
Third person singular
First person singular
paarl
ʔ-aarl
wə̆ýt
ʔ-ə̆ýt
tiinl
ʔ-iinl
jʊ́ʊ́l
ʔ-ʊʊ́l
ḱt́
ʔ-́t́
ɲ́
ʔ-́ɲ
Table 1 Third and first person present tense verb forms in Fur
The transformation of verbs borrowed from Arabic into (non-conjugated) coverb plus
(conjugated) light verb constructions (as in Fur jrrɪ́bɛŋ pɪ
jarrab ), instead of their conjugation as verbs (both in Fur and in Maban and Saharan
languages,), suggests that most of these conjugational classes are simply memorized by
speakers. ssue, and the question of the extent to which speakers produce
again below in the discussion of number marking in Western
Nilotic languages.
Across Nilo-Saharan, nominal features like number and case marking for nouns are
expressed by suffixes or internal changes. In addition, there are inflectional and derivational
prefixes whose etymological source or synchronic function is not always clear, as first pointed
-
Africa. Such morphemes are found in Central Sudanic languages (pp. 36-37, 67-68), Saharan
(pp. 174-176), Maban (pp. 197-198), Fur (pp. 221-222), Temeinian (p. 156), Kunama (p. 338),
-359), Surmic (pp. 374-376), and
Nilotic (pp. 411-414). Examples from the Nilotic language Luo, with gender-sensitive
7
derivational prefixes, used for example to describe colours and patterns of cattle, are shown in
(4) (Tucker and Bryan 1966: 412).
Singular
Plural
(4)
r̀-t̀ˑ
ro-t̂̀
blac
r̀-c̀ˑr
ro-ceˆˑr̀
d̀-cɔ́ˆl
di-cɔ̂ˑnd̀
d̀-bɔ̂
di-bɔ̂ˑỳ
Incrementation of such prefixes in nouns or nominal attributive phrases is common across the
phylum. Dimmendaal (1983: 25-
incorporated into nominal roots in the Nilotic language Turkana: -kɪ-/-ki-, -ka-, -sɪ-/-si-, -ma-,
ŋa-. Several of these are attested as function morphemes in Nilotic and beyond, for example as
prepositions (e.g. instrumental kɪ-/ki-, locative ka-), as a derivational prefix for masculine
gender (ma- in Western Nilotic languages like Dinka: ma-caar
ŋa-.
There is a widespread and rather permanent tendency within the wider Nilo-Saharan
phylum to use such headless noun phrases as a circumlocution strategy for animal names. (5)
is an example from the Nilotic language Maa (Vossen 2015: 180):
(5)
n-k-nya+mamisha
Passer rufocinctus; lit. the one
FSG-ki-eat.PRES+bits.and.pieces
Greenberg (1981: 106), in his survey of moveable-kin Nilo-Saharan, assumed that the ki-
element, as in (5), goes back to a demonstrative. But it is more likely to be a reflex of a
widespread and extremely stable preposition *ki in Nilo-Saharan (Dimmendaal 2018), which
was (and is) also used to introduce attributive phrases in noun phrases. The incremented forms
in nouns in Turkana and elsewhere provide historical evidence that such elliptic (headless)
attributive or relative clauses
independent (i.e. as head) nouns themselves.
Whereas prefixes or proclitics are clearly more stable than suffixes or enclitics in Nilo-
Saharan, the former are not entirely exempted from deletion or fusion with lexical roots. The
Central Sudanic language Lendu (spoken in eastern Congo) is one of the few among the
8
languages of the world without any (apparent) affixational morphology, although closely
related languages do have such bound morphemes (Dimmendaal et al., 2019). The following
as shown by the rising tone.
(6)
b
b
It is fairly common across Nilo-Saharan to use tone as an exponent of inflection (e.g. to mark
case, tense-aspect, person, number) or derivation (both with and without category shift). In
other words, internal morphology involving tonal modification occurs, not only in Nilo-Saharan
languages with strongly reduced segmental structures, but also in languages with opulent
segmental morphology, as in the Saharan language Kanuri, where abstract nouns may be
derived from basic nouns by means of a fixed Low-High tone melody, as in these examples
from Hutchison (1981: 65):
(7)
km
km
knr
knr
Rounding (or fronting) harmony or advanced tongue root (ATR) harmony for vowels in
adjacent morphemes may also be an exponent of wordhood. ATR harmony is either controlled
by the lexical root, or alternatively by [+ATR] suffixes. In a classicalsystem, all vowels
within a (phonological) word belong to one of two harmony sets: a set of five [-ATR] vowels,
, , a, , , and a set of five [+ATR] vowels, i, e, , o, u. As a general rule, [+ATR] roots or
suffixes tend to be dominant, although there may be irregularities because of opaque (or
transparent) bound morphemes; see Casali (2008) for a discussion. In languages with opulent
morphologies, like the Nilotic language Kipsikiis, feature control ([+ATR) dominance) occurs;
otherwise, the lexical root tends to control vowel harmony processes. The examples in (8) are
from Toweett (1979: 422):
Verb
Noun
(8)
-kr
kr-n
Suffixational morphology is far more common as a formal strategy across Nilo-Saharan than
prefixation. However, due to the anticipatory nature of speech there is a permanent tendency
for such bound morphemes, belonging to closed sets, to fuse or be absorbed by lexical roots.
9
In a number of Western Nilotic languages, specifically Atuot, Dinka, Nuer and Shilluk, internal
morphology has been taken to an extreme, affecting consonants and tone, as well as vocalic
qualities in stems. Derived stems in Dinka are essentially monosyllabic, but still manifest
different layers of derivational and inflectional morphology formally expressed through
modifications in vowel length and voice quality (creaky versus breathy voice). Andersen (1992-
h the
then by an
inflectional layer.
(9a)
-w
ec
D-kick:3SG
(9b)
-w
ec
hither
D-kick:VEN:3SG
(9c)
-w
ec
D-kick:IT:3SG
Rottland (1982: 228) reconstructs a [+ATR) ventive suffix *-u (expressing movement
towards the deictic centre), and a [-ATR] itive suffix *-ta (expressing movement away from the
deictic centre). But as shown in (9b), ventive marking in Dinka is realized as
the root vowel (changing creaky vowels to corresponding breathy vowels), whereas itive
marking is realized as creaky voice in addition to tonal modification (9c).
ii
Word-final phonetic erosion (resulting from the anticipatory nature of speech)
frequently results in the transfer of phonetic properties onto preceding segments or morphemes,
as shown for Dinka above. Alternatively, paradigmatic displacement may occur,
as shown by the following cognate second person singular and plural imperative paradigms for
a group of Eastern Nilotic languages. Part of an original voice marker, still found in Lotuxo,
was reinterpreted as part of the plural marker in Teso, and was then shortened in four other
languages, in line with their predominant -(C)V template for second and third person plural
marking suffixes (Dimmendaal 1991: 296).
10
Singular
Plural
Lotuxo
ɔwax- a
o-wax-a-ta
Teso
kɔ-bɔk
kɔ-bɔk-ata
Karimojong
tɔ-bɔk
tɔ-bɔk-a
Nyangatom
tɔ-bɔk
tɔ-bɔk-a
Toposa
tɔ-bɔk
tɔ-bɔk-a
Turkana
tɔ-bɔk
tɔ-bɔk-a
Table 2 Imperatives in six Eastern Nilotic languages
Such reinterpretations also illustrate the importance of paradigmatic contrasts, and also that
inflectional properties like person, number, tense, and aspect are associated with whole words
rather than with slots (morphemes).
Additional cases of , resulting in the creation of new number suffixes as a
result of paradigmatic displacement, are presented for Turkana in Dimmendaal (1983: 246-247)
and Dimmendaal (1991: 292, 295-298).
The dialect cluster within the Nilotic branch of Eastern Sudanic to which Turkana
belongs also has morphologically conditioned allomorphy (alongside phonologically
conditioned allomorphy, involving ATR harmony and fronting of vowels). The kind of
derivational suffix a derived abstract noun takes, for example, is determined by the canonical
shape of the adjectival or stative verb root (Dimmendaal 1983: 50-51, 223-258, 270-274). -CVC
roots take a suffix ɪ̀s(ɪ́)-, partially reduplicated roots take a nominal suffix -ù, whereas non-
reduplicated -CVCVC roots copy the last vowel and consonant from the root; a- is a gender
marker for feminine nouns in the singular.
11
Adjective
Derived noun
(10)
-jk
-jk-s̥
-r
a-r-̀ś̥
-gogo
a-gogo-̥̀
-babar
a-babar-̥̀
‘saltiness’
-plr
-plr-r
-ckl
-ckl-l
The prefixes in these examples are masculine and feminine gender markers. Number
inflection in Turkana is governed by mora-counting principles to a large extent. Monomoraic
nominal roots (i.e. -CVC roots containing one vowel), for example, almost always take the
number suffix -ɪn/-in (the alternation being determined by vowel harmony rules in the
language); bimoraic roots ending in a vowel (i.e. -CV(C)V roots) tend to take the number suffix
-ɪ/-i (depending on vowel harmony rules), whereas bimoraic roots ending in a consonant (i.e. -
CV(C)VC roots) tend to take a suffix -a/-o (depending on vowel harmony rules).
Singular
Plural
(11)
-cm
-cm-n
(s)
-pm
-pm-n
-p
-p-
-wn
-wn-
́-b́kok
́-bòkòk-́
́-cùkút
́-cùkùt-́
The presence of specific moraic structures in nominal roots, e.g. -CVC, -CVCV, -CVCVC,
determines the choice between different number suffixes, like -in, -i, and -o, whereas the choice
between derivational suffixes like -ɪ̀sɪ
or -u
̀ depends on whether a -CVC root or a partially
reduplicated root is involved. Given their predictability, these distinct number or derivational
morphemes may be argued to be allomorphs (rather than
phonologically conditioned allomorphs).
12
This constellation differs rather dramatically from that described for Western Nilotic
Dinka by Ladd et al. (2009). Parallel to the system described for the Dinka verb system above,
there is a considerable degree of vertical morphology for number inflection with nouns, a corpus
of four hundred Dinka nouns yielding eighty-two combinations. Ladd et al. (2009) consequently
raise the learnability question for such languages, and point out that
to identify any phonological or semantic motivation for the choice of number marking pattern
-derived nouns. As further pointed out by the same authors (p. 669),
crosslinguistically, there may be more functional pressure toward regularity in verbs which
typically have many forms in a highly inflected language than in nouns which may be
Similar observations on paradigmatic complexity are made by Baerman (2012) with
respect to case-number alternations in the closely related language Nuer, for which twenty-five
suffix patterns and forty-eight stem alternation patterns were identified. Without questioning
the presumed complexity of such systems, the monosystemic character of such alternations in
these Western Nilotic languages should be emphasized as well. Similar modifications (vowel
lengthening, creaky voice vowels becoming breathy voice, tonal modifications) occur as part
of the derivation or inflection of verbs as a major category, and of other categories like
adjectives, as well as with locative case marking, for example.
Apart from affixation, reduplication is common across the Nilo-Saharan phylum,
expressing semantic properties like plurality, intensity or repetition, but also category shift, as
with adjectives or adverbs derived from nouns. This latter strategy is found in Central Sudanic
languages like Avokaya (data from the author):
(12)
c
cc
Such widespread patterns of reduplication are not necessarily indicative of a common origin,
but instead appear to have emerged independently again and again in individual languages due
to the high degree of diagrammatic iconicity involved.
3. The tripartite division between derivation, inflection, and compounding
The traditional compartmentalization between derivation, inflection, and compounding in
morphology, though still useful in and by itself with respect to many languages, may also
obliterate important analytical issues: first, their link with the lexical structure of languages,
and second, their link with syntactic structure. Moreover, the boundaries between these
13
strategies are occasionally fuzzy. The survey of these three domains in the section 3 is therefore
followed by an illustration of their interrelationship and interaction in section 4.
3.1 Derivation
The distinction between nouns and verbs as distinct basic categories in Nilo-Saharan
languages is usually fairly clear from an analytical point of view, due to their morphological
features and syntactic positions, as well as their functions. Other categories, such as adjectives
or adverbs, do not necessarily exist as basic, non-derived categories, or may occur as small,
closed sets, additional forms being derived from other categories (such as nouns or verbs).
iii
Derivation as a lexical process contrasting with inflection is most obvious if the
transposition involves category shift. Reh (1996: 144-159) describes a prototypical system of
nominalizations for a Nilo-Saharan language where verbs are common, the Nilotic language
Anywa. Derived nouns express semantic roles such as agentives, instruments, action nouns,
abstract nouns and other types. From a formal point of view, suffixation and internal change
are involved, sometimes combined with prefixation (as elsewhere in Northeastern Nilo-
Saharan):
iv
detransitivized verb
noun
(13)
kr
kr (sg) / kyy (pl)
r
r (sg) / r (pl)
Nilotic languages like Turkana allow derivational suffixes such as the causative, dative, itive,
or ventive from the verb in corresponding nominalized forms (Dimmendaal
1983: 269-292).
(14)
a-cam-ʊ̀n-t
(--ʊn ‘agree’)
Habitual marking tends to be an aspectual (i.e. inflectional) category cross-linguistically
(Comrie 1976: 26-32). But in Nilotic languages like Turkana it is treated as a derivational
phenomenon, since it may also be part of a nominalized form, unlike progressive or perfective
aspect markers, as with the following form (with the habitual marker -aan-, which has a range
of additional allomorphs).
14
(15)
a-dak-aan-̥̀
‘satiation’
(-dak ‘graze’, -dak-aan- ‘eat till satisfied’)
INF-graze-HAB-NOML
Moreover, the same habitual marker precedes (other) derivational suffixes like the dative, itive
and ventive in derived verbs (as in a-boŋ-oon-ri -oon- is the habitual
and - ri the itive marker). Consequently, what is inflectional in one language may be
derivational in another.
Whereas in Eastern Sudanic languages verbs are common as predicates expressing
constructions as a lexical strategy. Hutchison (1981: 97) points out that in the Saharan language
Kanuri, there is a closed set of less than 150 verbs, the remaining verbal predications being built
with the light -n-). Verb to noun derivations consequently are extremely rare in this
language, although a few do exist (for example: d- , d-n
). Instead, noun to noun derivations are much more common in Kanuri and elsewhere in
Saharan, as is also shown by Jakobi and Crass (2004: 112-120) for Beria:
(16)
k
k-
snd
snd-r
kr
kr-n
tm
tm-n
Deverbal derivation in the Central Sudanic branch of Nilo-Saharan is rare for another reason.
While there is a certain degree of verbal derivation by means of prefixes in Central Sudanic (as
shown next), derivational affixes for nouns (or other syntactic categories) appear to be rare in
most Central Sudanic languages (Tucker and Bryan 1966: 35-37, 67), with compounding being
used as a dominant alternative lexical strategy for word formation with nouns.
v
Central Sudanic languages have different CV-/V-shaped verbal prefixes expressing
derivational relations with other verbs (whereas verbal derivation in Northeastern Nilo-Saharan
tends to be expressed through suffixation and/or root-internal changes, apart from the causative,
as already illustrated in section 2). In Madi, one finds a ventive (directional) prefix e-, or an
iterative prefix u- or a-. In Ngiti, derivational prefixes may also express singular action and
diminutive (i.e. an action which is carried out for a brief period of time or a little bit), or
pluractionality. As shown by the following examples from Kutsch Lojenga (1994: 299), such
derivations also tend to be accompanied by tonal modifications:
15
(17)
-drta
-drta
-drt
-drta
In addition to such vocalic prefixes, there are CV- prefixes in Central Sudanic, as with
pluractional marking in Moru (Tucker and Bryan 1966: 35).
(18)
tɔ̍ga̍
‘to chop to pieces’
ɔ-g
‘to chop’
lɔ̍vɔ̍
‘to throw one by one’
ɔ-vɔ́
‘to throw’
Pluractional marking on the verb is widespread in Nilo-Saharan (and other language families
on the continent; see Newman (1990) for a discussion of the Chadic branch within Afroasiatic).
Prototypically, it expresses an event performed several times by a singular subject, in the case
of intransitive predications, whereas with transitive predications the event affects multiple
objects or the object multiple times. The actual formal strategies, however, often differ even
between closely related languages (as the examples from Central Sudanic above illustrate).
There is also a common tendency towards the development of formally suppletive basic
versus pluractional verb stems, as with arta owuta (pluractional)
in the Central Sudanic language Ngiti (Kutsch Lojenga 1994: 283).
Apart from pluractional marking, directionality marking on the verb, in particular
movement towards the deictic centre (or the speaker), is frequently subject to renewal across
Nilo-Saharan.
vi
This ventive marking is again expressed by means of prefixes in Central
Sudanic and by means of suffixes in Northeastern Nilo-Saharan languages. Blacking and Fabb
(2003: 73-82) describe a ventive prefix ɛ-/e- for Madi:
(19)
-
k
l
r
3-eat
PL
food
AFF
16
(20)
-
k
l
r
(3)-VEN-eat
PL
food
AFF
As pointed out by Blackings and Fabb (2003: 73), the inferred meaning or conversational
implicature in example (20) as against example (19) is that the persons referred to are present
with the speaker at the moment of speaking, but the eating took place elsewhere.
Ventive marking in Nilo-Saharan (and elsewhere) is also interesting for another reason.
In an early, highly detailed analysis of the Nilotic language Maasai by Tucker and Mpaayei
(1955: 124), it was shown that verbs with a ventive suffix may express (in more recent terms)
a macro event involving two separate events, one of which is motion associated with another
event (expressed by the verbal root).
vii
(21)
---
3-see-VEN-PASS
This cross-linguistically widespread phenomenon is referred to as
study of Pama-Nyungan languages in Australia or languages in South America. As shown in
Belkadi (2016), this semantic notion is different from deictic directionality (usually referred to
as ventive marking in African linguistics), although the boundary between the two is sometimes
fuzzy due to the semantics of the main verb (Belkadi 2016: 65).
(22)
aa-rɛ-ʊn-ɪ
3>1SG-send-VEN-PASS
‘I (was) am driven this way’
These phenomena, illustrated for Maasai but equally common elsewhere in Nilo-
Saharan and other African language families, is also an instance of what Dimmendaal (2015)
, the conceptual conflation or expression of two events (expressing one macro-
event) in one grammatical word, and the morphological merger in one phonological word.
17
Otero (2018) describes a further interesting dimension of direction marking for the
Koman language Komo, namely its grammaticalization into a temporal/aspectual domain with
non-motion verbs, and into hearer-based pragmatics. As shown by the following examples, the
three deictic directional markers in Komo, which are in paradigmatic opposition to each other,
speech act time or a distinct time (Otero 2018: 175),
(23a)
h
p--
w
m
3SG.F
drink.SG-DD∅ -3SG.F
beer
there
(23b)
h
p--
w
m
3SG.F
drink.SG-DD1-3SG.F
beer
there
(23c)
h
p-k-
w
m
3SG.F
drink.SG-DD2-3SG.F
beer
there
Example (23a) refers to a situation where the referent is at the location of the
8: 172). Example (23b) is used when
the referent is no longer at the location of the verbal event and has come to the speech act
location. Finally, (23c) is appropriate when the referent is no longer at the location of the verbal
event, nor at the speech act location; moreover, the verbal event has been realized prior to the
utterance time. Deictic directional or associated motion marking thus expresses the fact that the
speaker (or protagonist) observed the event with his own eyes; it is therefore part of
evidentiality marking strategies in these languages (Dimmendaal 2014a).
Basic adverbs are usually small in number, as are adjectives, derivation from major
categories like nouns and verbs serving as a tool to create additional modifiers. Not well known
in the literature, but probably quite common in Nilo-Saharan as well as other language families,
is the derivation of verbs from ideophonic adverbs. One common strategy in Kanuri (and other
18
Saharan languages) is to combine these with a
example from Hutchison (1981: 107):
(24)
kl
klkl-
It is also common in Nilo-Saharan (as well as cross-linguistically) to derive adpositions
from nouns referring to body part terms. The position of these markers relative to the noun does
not necessarily correlate with the position of the verb relative to other constituents. Thus, verb-
initial Surmic languages like Tennet have postpositions (as well as a few prepositions), as
shown by Randal (1998: 223)
(25)
̱tṯ
Ḻ̱ḏ
c̱̱z
̱c̱ṯ
PFV:enter
Loudo:NOM
house
inside
Similar possessor-possessed constructions occur in verb-initial Southern Nilotic
languages like Nandi. Creider and Creider (1989: 70) give an example with
, ke:t-it kɛ:l, wherein -it. This example
and examples from other Nilotic and Surmic languages strongly suggest that these possessor-
possessed phrases are historical reflexes and synchronic instances of construct state
constructions (as also found in Koman and Mande, as well as Semitic, for example).
Whereas the historical link with body-part terminology is not always clear from these
languages, more distantly related Eastern Sudanic languages like Gaahmg (Jebel group) clearly
show this derivational link; for a description see Stirtz (2011: 266), who calls these nominal
modifiers following the reference Special purpose languages such
as secret languages or avoidance registers have been reported to manifest some unique
derivational features. Storch (2011: 67-68) gives examples from the Nilotic language Lango in
Uganda, where archaic prefixes are added to nouns, replacing affixless nouns, as with o-duk
tol
married woman among the Datooga in Tanzania, who speak a Nilotic language. Instead of using
éawéeda, married women use a word consisting of an adjectival root
19
mìsán-, and a suffix -úumèeda, which is otherwise used with nouns in order to create an
abstract state, mìsán-úumèeda
3.2. Inflection
Cross-linguistic data indicate that the distinction between inflection and derivation is gradient,
and therefore that this dual mechanism model is sometimes problematic. Authors like Booij
(1994) consequently distinguish between contextual and inherent inflection, the latter being
closer to derivation, contextual inflection being determined by the syntactic context. A range
of phenomena from Nilo-Saharan languages appear to support this gradient notion (as
illustrated in the present section and in section 4).
Contextual inflection for Nilo-Saharan languages involves definiteness marking or case
on nouns and noun phrases and tense-aspect marking on verbs. Tense-aspect marking appears
to constitute a relatively unstable property of this phylum. While it is common in Central
Sudanic to use auxiliary verbs and primarily bound morphemes in Northeastern Nilo-Saharan
languages, even relatively closely related languages may vary in terms of their tense-aspect (or
mood) system. Given the space limitations of the present survey, these inflectional properties
are not further discussed here (nor is definiteness marking for noun phrases). Inherent inflection
in Nilo-Saharan involves number marking for nouns and diathesis for verbs.
In Central Sudanic languages, the interpretation of argument relations is determined by
a strict constituent order (with the main verb or auxiliary verb occurring in second position and
the subject preceding it). Northeastern Nilo-Saharan languages on the other hand tend to have
case marking. In Fur, Kunama, Maban, Saharan, and the Northern branch of Eastern Sudanic
this is combined with a verb-final order, whereas in the Southern branch of Eastern Sudanic
(and Kuliak) a verb-initial or verb-second structure is found (with reduced case marking, except
in Kuliak). These Southern members of the Eastern Sudanic subbranch, more specifically Berta,
Gaahmg and various Nilotic and Surmic languages, allow for postverbal subjects, which are
marked for case, whereas objects are not inflected for case (in contrast with other Northeastern
Nilo-Saharan languages with case, which have Accusative case marking; see Dimmendaal et
al. (2019) for further details).
viii
In Eastern Sudanic languages which only allow the subject (or A-role) of transitive
predications in postverbal position, i.e. with ergative systems, the verb also receives an ergative
marker, as in Gaahmg (Stirtz 2014: 247):
20
(26)
mn
nmn=
lg=ì
m
w
goat
certain
beat.CONT=ERG
thirst.ERG=ERG
GP
well
NEG
Eastern Sudanic languages also allowing for postverbal subjects with intransitive predications,
i.e. those with marked nominative systems, sometimes make a formal difference on the verb
(presumably as a reflex of a former suffix, as still found in Gaahmg) when the subject follows
rather than precedes a transitive or intransitive verb, as shown by the following examples from
Western Nilotic Dinka (Andersen 1991: 297; interlinear glossing added by the present author).
(27)
t̰̰̀̀k
̰̀-t̰̀t
m̰̀th
woman: ABS
D-beat
child
(28)
m̰̀th
̀-thɛ́ɛ́t
t̰̰̀̀k
child:ABS
D-beat:NTS
woman:NOM
Western Nilotic languages like Dinka also show the limits of syntactically relevant non-
concatenative morphology. As shown by Andersen (1992-94), and as illustrated in example
(9a-c) above, three layers of internal morphology can be identified for the verb (including
indexing for subject and changes in valency). Syntactically, Dinka has a rigid verb-second
constituent order, whereby the initial position is reserved for topics (syntactic subjects, objects
or other constituents). -means
of tonal inflection, whereas inflection on the verb also shows that a constituent other than the
subject functions as a topic. Moreover, only topics can be heads of relative clauses (Andersen
1991).
Southern Nilotic languages like Kipsikiis on the other hand are strongly non-
configurational. Both Dinka and Kipsikiis have Marked Nominative case systems, whereby the
postverbal (as against the preverbal) subjects (S-roles and A-roles) are inflected for case by way
of tonal inflection. But in contrast to Dinka, Kipsikiis has a rich affixational morphology,
indexing for subject and object as well as for a range of additional functions or semantic roles
(instrumental, benefactive, comitative, movement towards or away from the deictic centre,
21
mediative, the latter term being used in Nilotic studies for associated motion), can be observed,
as illustrated in (46) below. Moreover, postverbal constituent order is largely free syntactically,
the actual order being conditioned by discourse to a large extent. Creider and Creider (1983: 3)
for the closely related language Nandi (interlinear glossing by the
present author).
(29a)
ki:rì:p
la:kw:t
a:rt̂:t
PAST:3.watch
child:NOM
lamb:ABS
(29b)
ki:rì:p
a:rt̂:t
la:kw:t
PAST:3.watch
lamb:ABS
child:NOM
While (29a) WWho did
WWg rise to O or VO as
Both in languages with Ergative/Absolutive and Nominative/Absolutive or
Nominative/Accusative case systems in Northeastern Nilo-Saharan, active alignment (i.e. split
S-systems) may occur. The argument structure interpretation follows from case marking (if
independent core arguments occur) and/or from cross-reference marking on the verb, as in the
following examples from Beria, where pronominal subjects are expressed by suffixes and
objects by prefixes (Jakobi 2009):
(30)
kì-g-ì
come-1SG:SU-IPV
(31)
nɛ́-gɛ́-Ø-ɪ̄
you slept
2SG:OB-sleep-3:SU-PFV
22
(32)
nɔ́-rɔ́-g-ɪ̄
2SG:OB-marry-1SG:SU-IPV
With respect to indexing for subject and object on the verb, the only argument of the verb in
(31) is treated like the object of the transitive verb in (32) (the alternation between nɛ
́- and nɔ
́-
being determined by rounding harmony for vowels, which is common in the area).
Other diathesis features marked on verbs include passive (or impersonal active) and
middle voice marking by means of suffixes (or sometimes internal morphology), as in the
Nilotic language Turkana (Dimmendaal 1983: 157), where the middle voice marker -a is
probably cognate with a widespread marker with the same form elsewhere in Nilo-Saharan
(Greenberg 1963: 133).
(33)
à-yn-à
ayɔ̀
1SG-know-MI
1SG:NOM
(34)
à-mn-à
ayɔ̀
1SG-like-MI
1SG:NOM
Whereas -yɛn -mɪn- only occurs in
combination with middle voice marking in Turkana. Lexicalized forms such as the latter show
that the transition between inflection and derivation in languages is sometimes fuzzy, a position
also defended by Spencer (2013).
In Central Sudanic, number marking on nouns is restricted to nouns referring to animate
entities. Many Northeastern Nilo-Saharan languages on the other hand have rich number-
marking systems, whereby either the plural or the collective form is morphologically unmarked
(with corresponding singulative marking), or the singular (with corresponding plural marking
on the non-basic form); alternatively, both the singular and plural are marked for number (i.e.
replacement occurs). A recently identified language, Ngaalam, spoken by a few hundred people
in southwestern Ethiopia, for example, also manifests this system division for number
marking.
ix
23
Singular
Plural
(35)
émé-ni
ɛ́mɛ́
‘bone’
ɛ́lɛ́
l-ti
‘body’
oroð-a
oroð-e
‘dog’
This tripartite division manifests an amazing historical stability, except in the peripheral zones
of this Nilo-Saharan branch (Dimmendaal 2000). The local markedness for specific singular
forms (such
Dimmendaal 2000). Nouns referring to entities normally occurring in pairs or larger quantities
tend to take singulative marking.
There appears to be no direct link or interaction between such complex and rich systems
of number marking on nouns (involving singulatives, pluratives, replacement, and sometimes
distributives, greater plurals, greater singulars, and transnumeral forms) with pluractional or
singulative marking on verbs in these languages.
x
For example, Maba (a Maban language of
Chad described by Weiss 2009) combines this extensive tripartite number-marking system for
nouns with pluractional as well as singulative marking on the verb.
(36)
-w:n-
-w:k-r-
In languages like Maba, number marking on nouns is associated with more permanent states,
whereas the marking on verbs expresses temporary or fluid states with respect to certain events,
as further discussed in Dimmendaal (2014b).
3.3 Compounding
A certain degree of compounding, particularly exocentric compounding involving nouns, is
found in most Nilo-Saharan languages, regardless of their overall morphological typology.
Thus, even languages with opulent derivational verbal and nominal morphology usually have a
range of exocentric or endocentric nominal compounds referring to flora andfauna, as well as
names for human beings or domestic animals and their physical or behavioural characteristics,
as in the following examples from Turkana (Dimmendaal 1983: 292-295), where the prefixes
24
e- and lɔ- are masculine gender markers, whereas -u-
epenthetic vowel.
(37a)
e-wur-̀-pey
Olea
MG-smell-EP-one
(37b)
lɔ̀-mr̀-pʊs̥̀
-
MG-blotched-mauve
However, there are four genetically more restricted types of compounding, whose
presence does seem to be related to other structural properties of the languages involved: the
first type is the extensive use of nominal compounding in Central Sudanic (as mentioned in
section 2); second, verbal compounding in one group of Central Sudanic languages; third,
verbal compounding resulting from the juxtaposition of converb plus main verb in a range of
Northeastern Nilo-Saharan languages; and fourth, noun incorporation. Each of these is
discussed briefly below.
Boyeldieu (2007) discusses this second type of compounding, involving up to four
verbs, which have a high frequency in one group of Central Sudanic language spoken in the
Central African Republic and Sudan. (38) is an example with three verb stems from a folktale
in one of these languages, Yulu (Boyeldieu 2007: 25).
(38)
̏m̏aɗə́
yə̄.ltə̀.layə́
ja-ngè
də̏
samə̀ gomə̄
mud wasp
bends-brings out.lets come
heat-its
of
nose hare
Similar multi-verb constructions occur in Gula and Modo, which belong to the same subgroup
as Yulu within the Western (Bongo-Bagirmi) branch of Central Sudanic, as shown by
Boyeldieu (2007), who also notes similarities and differences with multi-verb constructions in
languages in West Africa, the Oceanic branch of Austronesian, and New Guinea.
Multi-verb constructions in languages such as Yulu are characterized by specific tonal
and segmental interactions, and by the fact that a single subject (agent) is involved (as well as
a single object in the case of transitive verb constructions), whereas no other morpheme may
25
be inserted between.
xi
-
multi-verb constructions in Yulu on semantic grounds. The second type bears no
formal signs of dependency or hierarchical ranking, and involves around 40 verbs expressing
orientation, aspect, and quality or quantity, as in (34).
(39)
ɲ̀tə́ + ˚.joɗə̄
= ɲ̀tə́joɗə̄
‘(do [a]) work again’
work doing again
(40)
nd̀oɗə̏ + lə̀
= nd̀oɗlə̀
‘sleep (soundly)’
lie die
A formally different type of multi-verb construction is found in Northeastern Nilo-
Saharan languages with converb plus main verb constructions, which also have extensive case
marking and a verb-final constituent order. There is a rather permanent strategy in these
languages to develop lexical compounding from sequences of converb, i.e. a dependent verb
form with restricted inflectional features for person, number, tense or aspect, and a following
(syntactically) main verb, which carries all these relevant indexical features, as in Taglennaa, a
Nubian language in the Nuba Mountains in Sudan. The following examples from Gulfan (2013)
illustrate a freely generated sequence of verbs in (41) and an idiomatic (lexicalized)
construction involving compounding in (42):
(41)
Ahmed
kɔy-g
kel-i-kɔ
out-gi
di-n
Ahmed
meat-ACC
eat-SSC-SIM
water-ACC
drink-PRS
(42)
ol-i
-
come.out-SSC
walk-IMP
If compounding is defined as a lexical strategy involving two or more lexical roots, noun
incorporation in Nilo-Saharan needs to be listed as another lexical strategy. Berta is the only
Eastern Sudanic language with noun incorporation (as far as present knowledge goes); see
Neudorf (2015) for a description. This latter phenomenon is widespread in the neighbouring
26
-Saharan), areal spreading
probably played a key role, especially because Berta also shares the presence of ejective
consonants with these languages.
xii
In Gumuz, for example, nouns referring to body parts occur
as grammaticalized verbal classifiers (examples from Ahland 2010: 180-181):
(43)
ʔaf-ə́c
məχa
wash-eye
blood
(44)
ʔaf-́w
lɨ-c-́́
wash-head
POSS-eye-2SG
Noun incorporation has also been described for the Songhay language Koyraboro Senni in Mali
by Heath (1999b: 170-172):
(45)
-jer
mee-haw
ear-lift
mouth-tie
Whether Songhay is part of Nilo-Saharan is not clear; the genetic significance of noun
incorporation as a common typological property remains to be determined. Songhay, Koman,
thus show instances of non-inflectional concatenation of a verb plus
complement noun (or the inverse order). But these should not to be treated as cases of
polysynthesis, following the definitions of de Reuse (2009), as they do not involve category
changes or interact with syntax.
4. Going beyond the traditional compartmentalization of morphology
In the preceding discussion, it was pointed out at several points that morphological
procedures interact in non-trivial ways with lexical and morphosyntactic structures. The
dominance of noun-to-noun derivation in Saharan, for example, is linked to the fact that there
are relatively few verbs in several Saharan languages. Also, verbal compounding in various
Northeastern Nilo-Saharan languages is a direct consequence of the frequent juxtaposition of
27
converbs (as dependent verbs) and main verbs as a morphosyntactic property. But such
interactions are even more compelling, as the following cases should help to illustrate.
Whereas Northeastern Nilo-Saharan languages like Fur or Kunama, those belonging to
the Maban and Saharan groups, and the Northern branch of Eastern Sudanic are verb-final and
strongly dependent-marking at the clause level (with extensive case-marking), languages
belonging to the Southern branch of Eastern Sudanic are verb-second or verb-initial.
Corresponding to this latter property, one finds a reduced case marking system (or absence of
case), and a much more extensive system of head marking at the clausal level. The head-marked
clause pattern favours verb-initial constituent order, as argued in Nichols (1986); this is exactly
what can be observed in Southern Eastern Sudanic subgroups. Nilotic languages like Kipsikiis,
for example, distinguish only between nominative and absolute (morphologically unmarked)
case, but express semantic roles like agent, patient, beneficiary/recipient, ventive, itive,
instrument, or mediative on the verb. The following is an example from Toweett (1979: 129-
195):
(46)
k̀̀-p̀r-t̀̀t-̀̀-chí
̀ryéèt
INF-dust-IT-MED-DAT
dress:ABS
Rottland (1982) shows that the restricted flagging strategy (due to reductions in the case-
marking system) and the extensive verbal indexing strategy (through stacking of inflectional
and derivational morphemes) in languages like Kipsikiis can be reconstructed for Proto-
Southern Nilotic. The Proto-Southern Nilotic dative extension *-ci, for example, is cognate
with the Eastern Nilotic dative marker -kɪ/-ki, itself probably a reflex of the widespread Nilo-
Saharan preposition ki. Similarly, the Proto-Southern Nilotic itive (or allative) marker *-ta is
reminiscent of an adposition with the same deictic function elsewhere in Nilo-Saharan (Ehret
2001: 205).
The univerbation of verbs and following prepositions as phrasal units, and subsequently
as phonological units, presupposes the frequent occurrence of such constructions (or
collocations) in day-to-day interactions.
xiii
This may be expected in a language where case
marking for more peripheral syntactic arguments (expressing semantic roles like location,
direction, instrument and the like) is absent or manifests a high degree of syncretism.
28
Goldberg (2006: 85) has argued that construction learning is facilitated by high token
frequency. This in turn leads to a cognitive anchoring effect (standard of comparison), and a
, p. 89),
whereby the degree of semantic relatedness of the new instances is crucial. In other words,
generalization of new instances does not occur on the basis of type frequency alone. A
construction grammar approach consequently provides a natural explanation for the drift
towards clausal head marking in the Southern branch of Eastern Sudanic.
The same process presumably also resulted in a reinterpretation of coverb plus light verb
constructions (usually n(i)), which are common elsewhere in Northeastern Nilo-Saharan
branches (as stated in section 2). Tucker and Bryan (1966: 330) give an example of such a
construction in the Southern Eastern Sudanic language Berta, where the former light verb n(i)-
is used as a suffix deriving verbs from nouns.
(47)
kere
kere-n
The same marker is found as a derivational suffix marking habitual (or sometimes inchoative,
depending on the tense-aspect system of the language involved) elsewhere in the Southern
branch of Eastern Sudanic, as illustrated for Turkana in example (14) above. Its status as a
derivational marker (rather than as an inflectional feature marking aspect) may be due to its
former status as a light verb which was used to construct event structures in combination with
coverbs.
Apart from the (historical) interaction between syntactic structures and morphological
strategies, inherent inflection and derivation in languages sometimes interact, as shown by the
historical development of gender marking from a derivational phenomenon into an inflectional
property of nouns in the Eastern Nilotic branch of Nilotic. Gender marking is not widespread
in Nilo-Saharan. It occurs in the Kadu and
genetic affiliation with Nilo-Saharan is disputed), and also in Nilotic languages. In the Western
and Southern branches of Nilotic, gender marking is a derivational property of nouns primarily
referring to human beings and animals, as illustrated in example (3) above for the Western
Nilotic language Luo. Rottland (1982: 224) reconstructs a masculine prefix *kɪp- and a
feminine prefix *cɛ:p for Proto-Southern Nilotic. But in Eastern Nilotic, the marking of
masculine and feminine gender (as well as neuter gender in the Teso-Turkana cluster) has
become an obligatory inflectional feature of nouns. In the Bari group, gender is covert
(manifesting itself on agreement markers within the noun phrase), but in the remaining (non-
29
Bari) languages (for example in Turkana) it is marked obligatorily on the noun (data from
Vossen 1982).
Bari
Turkana
kwn
-kɲ
‘bird’ (neuter noun)
kt
a-t
‘cow’ (feminine noun)
tɔ́m-́
́-tɔm
‘elephant’ (masculine noun)
Table 3 Nominal gender in two Eastern Nilotic languages
Nouns referring to animate entities and in particular domestic animals (cows, sheep,
goats, and camels) presumably have a high frequency, especially among pastoralists (as most
Eastern Nilotic speech communities are traditionally). This, combined with system pressure,
i.e. a tendency among speakers to target entire grammatical classes of items, i.e. to treat and
code whole lexeme classes in a uniform way (as observed by Haspelmath 2014) presumably
resulted in the change from a derivational into an inflectional system.
Another example of a former derivational strategy being reinterpreted as an exponent of
inflection comes from the Eastern Nilotic language Maasai, where the pluractional verb,
involving reduplication of the verb root, has been reinterpreted as a conjugational (inflectional)
form expressing plural arguments, specifically for the second person plural (Dimmendaal
2014b: 146), as in the imperfective paradigm for the verb -tum
Singular
Plural
1
-tm
ki-tum
2
-tm
-tm--tm
3
-tm
-tm
Table 4 The imperfective paradigm in Maasai
As illustrated in the next section, former members of compounds may also be reinterpreted as
affixes, and thereby become part of inflectional or derivational paradigms.
30
5. Areal dimensions of morphological typologies
The importance of multilingualism and therefore also of language contact for our
understanding of morphological divergence between Nilo-Saharan languages is central to this
final section.
The genetic division between the two primary branches of Nilo-Saharan, Central
Sudanic and Northeastern Nilo-Saharan, coincides with a number of typological differences.
Central Sudanic languages share areal properties with Ubangian (which may or may not be part
of Niger-Congo) phonologically and morphosyntactically, as well as pragmatically, as
illustrated in Dimmendaal (2014b). Northeastern Nilo-Saharan languages, on the other hand,
share a range of features with Ethiopian Afroasiatic languages.
xiv
These areal features include
the tripartite number-marking system, the use of case combined with a verb-final syntax,
converbs, and coverb plus light verb constructions. As already pointed out above, these
properties are found in Fur, Maban, Saharan, and the Northern branch of Eastern Sudanic.
Languages in the geographically more peripheral zones, such as Kunama in Ethiopia and
Eritrea, and some Saharan languages in Chad, did not necessarily retain all these areal features.
Languages belonging to the Southern branch of Eastern Sudanic retained the tripartite
number-marking system as a lexical property, but lost other features such as extensive case
marking and a verb-final syntax, although the areal source for this loss is not obvious. The
n a languages also tend to be verb-second (often allowing
postverbal subjects), but differ in other respects (as further discussed in Dimmendaal et al.
2019). The dramatic phonological and morphosyntactic restructuring discussed above for one
group of Western Nilotic languages, Dinka, Nuer, Atwot, and Shilluk, is even more mysterious,
as no areal source for this strong tendency towards internal morphology can be found today.
In her comparative study of Western Nilotic morphological systems, Storch (2005) did
identified several layers of chronologically organized stages of historical influence on the
morphology of another Western Nilotic subgroup, Southern Lwoo, whose extant members were
strongly influenced by Ubangian and Central Sudanic, as reflected in the strong reduction of
nominal number marking by means of suffixation. This latter reduced system formed the basis
for an even more elaborate system of morphological restructuring due to language contact
affecting one member, Luo, a language whose speakers have had long term contacts with
speakers of Bantu (i.e. Niger-Congo) languages in Kenya and Tanzania. Heavy (unadapted)
lexical borrowing from Bantu (as in mɪsʊmba (sg)/wa-sumb-n , and
31
reinterpretations of former nominal compounds as prefix plus lexical roots, resulted in the
gradual emergence of a noun class system (as is typical for Bantu) in Luo.
(48)
dh-l
j-l
j-l
For further details on convergence in the Nilotic-Bantu borderland, the interested reader is
referred to Dimmendaal (2011: 193-196).
More recently, Barasa (2017: 47) presented another case involving a reinterpretation of
a gender system as a noun class system in the Eastern Nilotic language Ateso, which has been
strongly influenced by the neighbouring Bantu language Soga. While maintaining the typical
Northeastern Nilo-Saharan tripartite number-marking suffixation system, Ateso speakers are
apparently in the process of reinterpreting the gender prefixes as alternating noun classes.
Ateso
Turkana
Singular Plural
Singular Plural
Feminine
a- a-
-, e- -, e-
-, i- -, i-
a- -
-, e- --
-, i- --
Masculine
Diminutive
Table 5 Gender prefixes in Ateso and Turkana
Whereas elsewhere in the Eastern Nilotic branch of Nilotic masculine and feminine nouns keep
their inherent gender in the plural, they may shift in Ateso, thereby acquiring features of
alternating noun class systems, as in Bantu.
32
Ateso
Turkana
Singular
Plural
Singular
Plural
(49)
a-pese
a-pesur(u)
a-pese
-pesur
a-ratɔ
-ratɔ-n
a-rait
-rait-o
-sap-at
-sap
-sapat
-sapa
-ktabɔ
-ktabɔ-
a-ktab
-ktab-a
e-kofja
e-kofja-i
a-kopya
-kopya-
hat
a-bakasa
i-bakasa-i
a-baaca
-baaca-
Bantu languages or English, but not uniquely so (as with semantic
motivation for the reallocation of nouns as masculine, feminine or neuter/diminutive nouns, or
for this class shift in the plural for that matter, consequently remains a topic on the research
agenda for contact phenomena in Nilo-Saharan languages.
The present author would like to express his gratitude to two anonymous reviewers, as
well as to Mary Chambers and Jan Knipping, for their extensive and helpful comments.
Abbreviations
ABS = absolutive
ACC = accusative
AFF = affirmative
CONT = continuous
D = declarative
DAT = dative
DD = deictic directional
EP = epenthetic segment
ERG = ergative
F = feminine
33
GP = general preposition
HAB = habitual
IMP = imperative
INF = infinitive
IPV = imperfective
IT = itive
MED = mediative
MG = masculine gender
MI = middle voice
NEG = negative
NOM = nominative
NOML = nominalizer
NTS = non-topical subject
OB = object
PASS = passive
PL = plural
POSS = possessive
PFV = perfective
PRS = present
SG = singular
SSC = same subject converb
SIM = simultaneous
SU = subject
VEN = ventive
References
Ahland, Colleen. 2010. Noun incorporation and predicate classifiers in Gumuz. Journal of
African Languages and Linguistics 31(2): 159-203.
Ameka, Felix K. 2005. Multiverb constructions on the West Africa littoral: microvariation and
areal typology. In Mila Dimitrova-Vulchanova and Tor A. Åfarli (eds.), Grammar and Beyond:
Essays in Honour of Lars Hellan, pp. 15-42. Oslo: Novus Press.
34
Andersen, Torben. 1990. Vowel length in Western Nilotic languages. Acta Linguistica
Hafniensia 22(1): 5-26.
Andersen, Torben. 1991. Subject and topic in Dinka. Studies in Language 15(2): 265-294.
Andersen, Torben. 1992-94. Morphological stratification in Dinka: On the alternations of voice
quality, vowel length and tone in the morphology of transitive verbal roots in a monosyllabic
language. Studies in African Linguistics 23: 1-63.
Baerman, Matthew. 2012. Paradigmatic chaos in Nuer. Language 88(3): 467-494.
Barasa, David. 2017. A Descriptive Grammar of Ateso. Munich: LINCOM.
Belkadi, Aicha. 2016. Associated motion constructions in African languages. Africana
Linguistica 22: 43-70.
Bender, M. Lionel. 1996. The Nilo-Saharan Languages: A Comparative Essay. Munich:
LINCOM.
Blacking, Mairi, and Nigel Fabb. 2003. A Grammar of Ma’di. Berlin and New York: Mouton
de Gruyter.
Booij, Geert. 1994. Against split morphology. In Geert Booij and Jaap Van Marle (eds.),
Yearbook of Morphology 1993, 2749. Dordrecht: Kluwer.
Booij, Geert. 2010. Construction morphology. Language and Linguistics Compass 4(7): 543-
555.
Boyeldieu, Pascal. 2007. Compound verbs and modalities of process in Yulu (Central Sudanic).
In M. Reh and D.L. Payne (eds.), Advances in Nilo-Saharan Linguistics. Proceedings of the 8th
Nilo-Saharan Linguistics Colloquium, University of Hamburg, August 22-25, 2001 . Cologne:
Rüdiger Köppe, pp. 25-39.
35
Casali, Roderic F. 2008. ATR harmony in African languages. Language and Linguistics
Compass 2(3): 496-549.
Comrie, Bernard. 1976. Aspect. Cambridge University Press.
Corbett, Greville. 2000. Number. Cambridge: Cambridge University Press.
Creider, Chet A., and Jane T. Creider. 1983. Topic comment relations in a verb-initial
language. Journal of African Languages and Linguistics 5(1): 1-15.
Creider, Chet A., and Jane T. Creider. 1989. A Grammar of Nandi. Hamburg: Helmut Buske.
De Reuse, Willem. 2009. Polysynthesis as a typological feature: An attempt at a
characterization from Eskimo and Athabaskan perspectives. In Marc-Antoine Mahieu and
Nicole Tersis (eds.), Variation on Polysynthesis: The Eskaleut Languages, pp. 19-34.
Amsterdam and Philadelphia: John Benjamins.
Dimmendaal, Gerrit J. 1983. The Turkana Language. Dordrecht: Foris.
Dimmendaal, Gerrit J. 1991. The geometry of verb paradigms in Teso-Turkana. In Franz
Plank (ed.), Paradigms: The Economy of Inflection, pp. 275-306. Berlin: Mouton de Gruyter
Dimmendaal, Gerrit J. 2000. Number marking and noun categorization in Nilo-Saharan
languages. Anthropological Linguistics 42: 214-261.
Dimmendaal, Gerrit J. 2011. Historical Linguistics and the Comparative Study of African
Languages. Amsterdam: John Benjamins.
Dimmendaal, Gerrit J. 2014a. The grammar of knowledge in Tima. In Alexandra Y. Aikhenvald
and R. M. W. Dixon (eds.), The Grammar of Knowledge, pp. 245-259. Leiden: Brill.
Dimmendaal, Gerrit J. 2014b. Pluractionality and the distribution of number marking across
categories. In Anne Storch and Gerrit J. Dimmendaal (eds.), Number: Constructions and
36
Semantics. Case studies from Africa, Amazonia, India and Oceania. pp. 57-75. Amsterdam and
Philadelphia: John Benjamins.
Dimmendaal, Gerrit J. 2014c. Derivation in Nilo-Saharan. In Rochel Lieber and Pavol
(eds.), Oxford Handbook of Derivation, pp. 591-608. Oxford: Oxford University Press.
Dimmendaal, Gerrit J. 2015. Alloying as an economy principle in morphology. SKASE Journal
of Theoretical Linguistics 13(2): 2-14.
Dimmendaal, Gerrit J. 2018. On stable and unstable features in Nilo-Saharan. In Helga
Schröder and Prisca Jerono (eds.), Nilo-Saharan Issues and Perspectives, pp. 9-23. Cologne:
Rüdiger Köppe.
Dimmendaal, Gerrit J., Colleen Ahland, Angelika Jakobi, and Constance Kutsch Lojenga.
2019. -
In H. Ekkehard Wolff (ed.), The Cambridge Handbook of African Linguistics, pp. 326-381.
Cambridge: Cambridge University Press.
Ehret, Christopher. 2001. A Historical-Comparative Reconstruction of Nilo-Saharan. Cologne:
Rüdiger Köppe.
Goldberg, Adele E. 2006. Constructions at Work: The Nature of Generalization in Language.
Oxford: Oxford University Press.
Greenberg, Joseph H. 1955. Studies in African Linguistic Classification. New Haven CT:
Compass Publishing Co.
Greenberg, Joseph H. 1963. The Languages of Africa. Bloomington & The Hague: Indiana
University Press and Research Center in Anthropology, Folklore and Linguistics & Mouton.
Greenberg, Joseph H. 1981. Nilo-Saharan moveable-k as a Stage III article (with a Penutian
typological parallel). Journal of African Languages and Linguistics 3: 105-112.
37
Gulfan, Gumma Ibrahim. 2013. Converbs in Taglennaa (Kordofan Nubian). In Thilo C.
Schadeberg and Roger Blench (eds.), Nuba Mountain Language Studies, pp. 371-379. Cologne:
Rüdiger Köppe.
Haspelmath, Martin. 2014. On system pressure competing with economic motivation.
In Bryan MacWhinney, Andrej Malchokov, and Edith Moravcsic (eds.), Competing
Motivations in Grammar and Usage, pp. 197-208. Oxford: Oxford University Press.
Heath, Jeffrey. 1999. A Grammar of Koyra Chiini: The Songhay of Timbuktu. Berlin: Mouton
de Gruyter.
Himmelmann, Nikolaus. 2014. Asymmetries in the prosodic phrasing of function words:
Another look at the suffixing preference. Language 90(4): 927-960.
Hutchison, John. 1981. The Kanuri Language: A Reference Grammar. Madison: African
Studies Program.
Jakobi, Angelika. 2009. Focus in an active/agentive alignment system the case of Beria
(Saharan). ZAS Papers in Linguistics 46: 129-142.
Grammaire du Beria (langue saharienne).
Jakobi, Angelika, and Gerrit J. Dimmendaal. To appear. Number in Karko and Nilo-Saharan.
In Paolo Acquaviva and Michael Daniel (eds.), Mouton Handbook on Number. Berlin: Mouton
de Gruyter.
Kutsch Lojenga, Constance. 1994. Ngiti – A Central Sudanic Language of Zaire. Cologne:
Rüdiger Köppe.
Ladd, D.R, Bert Remijsen, and C. Adong Manyang. 2009. On the distinction between regular
and irregular inflectional morphology: evidence from Dinka. Language 85(3): 659-670.
Mitchell, Alice. 2015. Extra-ordinary morphology in an avoidance register of Datooga. In Ruth
Kramer Elizabeth C. Zsiga, and One Tlale Boyer (eds.), Selected Proceedings of the 44th
38
Annual Conference on African Linguistics, pp. 188-198. Somerville, MA: Cascadilla
Proceedings Project.
Neudorf, Susanne. 2015. The use of body part lexemes in Berta. In Anne Storch and Angelika
Mietzner (eds.), Nilo-Saharan – Models and Descriptions, pp. 93-112. Cologne: Rüdiger
Köppe.
Newman, Paul. 1990. Nominal and Verbal Plurality in Chadic. Dordrecht: Foris.
Nichols, Johanna. 1986. Head-marking and dependent-marking grammar. Language 62(1): 56-
119.
Otero, Manuel A. 2018. Directional verb morphology in Ethiopian Komo. In Helga Schröder
and Prisca Jerono (eds.), Nilo-Saharan Issues and Perspectives, pp. 165-177. Cologne: Rüdiger
Köppe.
Payne, Doris L., and Leonard ole Kotikash. 2008. Maa Dictionary.
https://pages.uoregon.edu/maasai/Maa%20Lexicon/lexicon/main.htm
Randal, Scott. 1998. A grammatical sketch of Tennet. In Gerrit J. Dimmendaal and Marco Last
(eds.), Surmic Languages and Cultures, pp. 218-272. Cologne: Rüdiger Köppe.
Rottland, Franz. 1982. Die südnilotischen Sprachen: Beschreibung, Vergleichung und
Rekonstruktion. Berlin: Dietrich Reimer.
Sands, Bonny. 2009. Language and Linguistics Compass
3(2): 559-580.
Schrock, Terrill. 2014. A Grammar of Ik (Icé-td): Northeast Uganda’s Last Thriving Kuliak
Language. Utrecht: LOT dissertations 374.
Spencer, Andrew. 2013. Lexical Relatedness – A Paradigm-based Model. Oxford: Oxford
University Press.
39
Stirtz, Timothy M. 2011. A Grammar of Gaahmg. Utrecht: Landelijke Onderzoekschool
Taalwetenschap.
Stirtz, Timothy M. 2014. 2014. Ergative, antipassive and other verb derivational morphemes in
Gaahmg. Journal of African Languages and Linguistics 35: 243-272.
Storch, Anne. 2005. A Morphological Study of the Noun in Western Nilotic. Cologne: Rüdiger
Köppe.
Storch, Anne. 2011. Secret Manipulations: Language and Context in Africa. Oxford: Oxford
University Press.
Toweett, Taaitta. 1979. A Study of Kalenjin Linguistics. Nairobi: Kenya Literature Bureau.
Tucker, A. N., and M.A. Bryan. 1966. Linguistic Analyses: The Non-Bantu Languages of
North-Eastern Africa. London: Oxford University Press for the International African Institute.
Tucker, Archibald Norman, and John Tompo ole Mpaayei. 1955. A Maasai Grammar with
Vocabulary. London: Longmans, Green & Co.
Unseth, Peter. 1998. Two old causative affixes in Surmic languages. In Gerrit J. Dimmendaal
and Marco Last (eds.), Surmic Languages and Cultures, pp. 113-126. Cologne: Rüdiger Köppe.
Vossen, Rainer. 1982. The Eastern Nilotes – Linguistic and Historical Reconstructions. Berlin:
Dietrich Reimer.
Vossen, Rainer. 2015. Faunal terminology and word formation in Nilotic languages: A
typological approach. In Osamu Hieda (ed.), Information Structure and Nilotic Languages, pp.
175-187. Kyoto: Research Institute for Languages and Cultures of Asia and Africa.
Waag, Christine. 2010. The Fur Verb and Its Context. Cologne: Rüdiger Köppe.
40
Phonologie et morphosyntaxe du maba
Yigezu, Moges. 2018. Ngaalam: an endangered Nilo-Saharan language of southwest Ethiopia
a sociolinguistic survey on language vitality and endangerment. In Helga Schröder and Prisca
Jerono (eds.), Nilo-Saharan Issues and Perspectives, pp. 25-42. Cologne: Rüdiger Köppe.
Suggested further reading:
Bender, M. Lionel. 1996. The Nilo-Saharan Languages: A Comparative Essay. Munich:
LINCOM.
Boyeldieu, Pascal. 2000. Identité tonale et filiation des langues sara-bongo-baguirmiennes
(Afrique Centrale). Cologne: Rüdiger Köppe.
Dimmendaal, Gerrit J. 2011. Historical Linguistics and the Comparative Study of African
Languages. Amsterdam: John Benjamins.
Dimmendaal, Gerrit J., Colleen Ahland, Angelika Jakobi, and Constance Kutsch Lojenga.
2019. -
In H. Ekkehard Wolff (ed.), The Cambridge Handbook of African Linguistics, pp. 326-381.
Cambridge: Cambridge University Press.
Ehret, Christopher. 2001. A Historical-Comparative Reconstruction of Nilo-Saharan. Cologne:
Rüdiger Köppe.
Greenberg, Joseph H. 1963. The Languages of Africa. Bloomington & The Hague: Indiana
University Press and Research Center in Anthropology, Folklore and Linguistics & Mouton.
Notes
41
i
An additional, and historically probably secondary process, involving encliticization or
suffixation of person markers, found in Kuliak, Saharan and a number of Eastern Sudanic
languages, is illustrated in Dimmendaal et al. (2019).
ii
The historical change from [-ATR] to creaky voice and from [+ATR] to breathy voice in Dinka
is regular (Andersen 1990).
iii
This section presents a brief summary of Dimmendaal (2014c), where a more extensive
description of derivation in Nilo-Saharan can be found.
iv
Storch (2005: 386-390) presents such markers as sometimes oscillating between lexical root
(in a compound) and prefix in her survey of nominal morphology of Western Nilotic languages.
v
Our current state of knowledge does not allow us to determine whether suffixation as a
derivational strategy for nouns was lost in Central Sudanic, or whether the suffixes in the other
primary branch, Northeastern Nilo-Saharan, are due to reinterpretations of former lexical roots
in compounds as affixes.
vi
One reason for the frequent renewal of ventive markers (which are sometimes in paradigmatic
contrast with itive markers, expressing movement away from the deictic centre) may be their
extended use in other domains, for example as aspectual markers.
vii
Maasai is one of the few Nilo-Saharan languages for which an online dictionary is now
available (Payne and ole Kotikash 2008), which opens new ways of doing research on
morphology.
viii
Outside Eastern Sudanic, marked nominative case systems are found in Sinyar, whose
genetic position within Northeastern Nilo-Saharan however is unclear, and in Kadu languages,
which may or may not be related to Nilo-Saharan.
ix
Data collected and kindly provided by Moges Yigezu, whose help is gratefully acknowledged
here. Yigezu (2018) classifies Ngaalam as a member of the Southwestern branch of Surmic.
x
The usually rich number-marking system as an exponent of inflection or derivation, in
particular in Northeastern Nilo-Saharan languages, cannot be elaborated upon here for reasons
of space. For a more detailed survey the interested reader is referred to Jakobi and Dimmendaal
42
(2000).
xi
See Ameka (2005) for a discussion of such constructions in West African languages.
xii
The absence of evidence for convergence with neighbouring Omotic (Afroasiatic) languages
in the area may be an indication of the absence of multilingualism (and thereby of social
contact) between these groups, but this needs to be confirmed by sociolinguistic research.
xiii
The importance of constructions for morphology has been emphasized by authors like Booij
(2010).
xiv
The palaeoclimatological conditions for this areal diffusion, and more specifically the
presence of a former riverine system, the Yellow Nile, are discussed in Dimmendaal (2011:
340-343).
