- 1 It should be noted that the data presented in this paper was collected within the framework of a l (...)
- 2 Natural language data is here understood as recorded texts of spontaneously produced speech, inclu (...)
1Recent years have seen a big leap forward in the documentation of non-Bantu click languages from southern Africa commonly subsumed under the typological label “Khoisan”. Comprehensive descriptions as well as sketch grammars have become available for a considerable number of languages from the Kx’a (Dickens 2005; Collins & Gruber 2014; Heine & König 2015; Pratchett 2020a), Tuu (Güldemann 2005; 2013; Collins & Namaseb 2011) and Khoe-Kwadi families (e.g., Kilian-Hatz 2008; Haacke 2013; Visser 2013; Widlok 2013; Collins 2023; Fehn 2016; Phiri 2021), covering (almost) the entire geographical and genealogical spectrum (Map 1). Given the amount of available linguistic descriptions, surprisingly little is known about the expression of commands. Beyond occasional mentions of an imperative clause type in descriptive works, no in-depth study exploring the encoding of directive modality in southern African “Khoisan” has yet been undertaken. In this article, I intend to provide a starting point for future research by offering preliminary insights into the expression of commands and prohibitions in the Khoe-Kwadi language Ts’ixa.1 My aim is two-fold: First, I will provide a language-specific overview using both elicited and natural language data2 which not only makes reference to the grammatical encoding of commands and obligations, but also to the ethnographic background. Second, I will adopt a comparative perspective and consider observations made for Ts’ixa in the context of publicly available descriptions of commands in related Khoe-Kwadi languages. I will thereby consider both closely related varieties from the Khwe and Shua dialect clusters, as well as more remote relatives from across the family, including the extinct Angolan language Kwadi.
2The present article is structured as follows: I start out by setting up the preliminaries, including a review of Khoe-Kwadi classification (§1.1), an introduction to the Ts’ixa speech community (§1.2), and a brief description of the typological profile of the Ts’ixa language (§1.3). In §2, I present the Ts’ixa data, starting from “canonical” (2nd person) imperatives with different levels of politeness (§2.1). I then turn to non-canonical imperative expressions directed at individuals encoded as 1st (§2.2) and 3rd person (§2.3) subjects. §3 focuses on comparative aspects (§3.1) and diachronic relations (§3.2) in the morphosyntax of imperative clauses across Khoe-Kwadi. §4 provides a short conclusion and outlook on future research.
Figure 1 — Map displaying the historical extension of the typological unit “Southern African Khoisan”, including major languages of the Kx’a, Tuu and Khoe-Kwadi families (cf. Güldemann 2014).
The Ts’ixa language, which forms the main subject of this study, is marked by a star.
3In his influential classification of the languages of Africa, Greenberg (1963) postulated a genealogical unit “Khoisan”, encompassing all non-Bantu, non-Cushitic click languages of southern and eastern Africa. This proposal has since been rejected in favour of three independent language families, Kx’a (Heine & Honken 2010), Tuu (Güldemann 2005) and Khoe-Kwadi (Vossen 1997; Güldemann 2004; Fehn & Rocha 2023), that co-exist in a linguistic area roughly delimited by the Kalahari Basin of southern Africa (Güldemann 1998; 2014). While the two eastern African click languages Hadza and Sandawe are usually considered isolates (Sands 1998), a link between Sandawe and Khoe-Kwadi has been suggested on the basis of lexical, morphological and typological evidence (Güldemann & Elderkin 2010).
4The Khoe-Kwadi family (Figure 2) consists of two sister branches, Khoe and Kwadi, whereas the Kwadi branch is represented by a single language which used to be spoken around the Curoca River Basin in southwestern Angola (Westphal 1965; 1971; Güldemann 2004; Fehn & Rocha 2023). The Khoe or “Central Khoisan” languages were recognized as a family by Maingard (Maingard 1961; 1963), Westphal (Westphal 1962a; 1962b; 1971; 1980), Köhler (Köhler 1962; 1963; 1966; 1971; 1981), Baucom (1974) and Winter (1981; 1986) before becoming the subject of a thorough historical-comparative reconstruction by Vossen (1997). While the subdivision of Khoe into a Khoekhoe and a Kalahari Khoe branch is relatively undisputed, the further existence of an eastern and a western Kalahari Khoe unit as assumed by Vossen (1997) is doubtful (cf. Güldemann & Fehn 2017; Fehn 2020; Pratchett 2020b) and therefore not accounted for in Figure 2. Likewise, Vossen’s (1997) classification of Ts’ixa as a dialect of Shua is rejected here on the basis of my own research (Fehn 2016; 2019).
Figure 2 — The Khoe-Kwadi language family.
Extinct languages are marked by †.
5Due to the association between speakers of Khoekhoe and pastoralism, a foreign origin of the herders (Bleek 1851; Lepsius 1880: 188; Meinhof 1912) and, later on, of the entire language family (Westphal 1963; 1980; Güldemann 2008; 2020) has frequently been evoked. Based on typological (Güldemann 2008; 2020) as well as extra-linguistic evidence from archaeology (e.g., Sadr 2015; Lander & Russell 2018) and human genetics (e.g., Henn et al. 2008; Pickrell et al. 2014; Pinto et al. 2016), it is now widely assumed that Khoe-Kwadi languages were introduced in southern Africa by sheep herders from eastern Africa, around 2,000 BP. The details of this migration, as well as the historical relationship between the original herders and modern Khoe-Kwadi speakers – most of whom are foragers – are topics of ongoing research (Fehn et al. 2022).
6Ts’ixa is a Kalahari Khoe language spoken by ca. 200 individuals in the village of Mababe in the Ngamiland region of northern Botswana. The Ts’ixa speech community exclusively consists of former hunter-gatherers who belong to the ethnographic macro-unit commonly referred to as “San”, “Bushmen” or “Basarwa”. The term Ts’ixa transparently traces back to the expression ts’íí.xà ‘having buttocks’ and is commonly understood to be a xenonym coined by a closely related neighbouring community, the Bugakhwe, to refer to the inhabitants of Mababe. Speakers of Ts’ixa do not consider this term derogatory and agreed to its use for the language and speech community in scholarly publications. Their preferred term of self-reference is Xuukhoe ‘people left behind’ which is also in use among various Khwe-speaking communities of the Okavango River Basin. The unit Ts’ixa is further subdivided into three subgroups or extended families referred to as Handakhoe ‘people of the plains’, Hiyokhoe ‘people of the forest’, and Danisi, with the latter term referring to the descendants of formerly Shua-speaking migrants from the east.
7The Ts’ixa led a fully nomadic life until their hunting grounds were turned into conservation areas during the 1960s. Village elders were still born and socialized in a hunter-gatherer lifestyle and retain a profound knowledge of their natural environment, traditions and social values. While the Ts’ixa community has a history of headmen going back to the time before sedentarism, their social organization is predominantly egalitarian, with few differences between the individual rights of men and women. Although the younger generation is progressively adapting to a westernized lifestyle and has partially gained access to higher education, many traditions are still observed, including joking and taboo relationships, female initiation rites, traditional dances and healing.
8Ts’ixa is still used as a medium of everyday communication by most villagers over 40, and during my fieldwork (2010-2013), even children had at least a passive knowledge of the language. All villagers are bilingual in Setswana, and those younger than 50 are usually fluent in both spoken and written English. Speakers of Ts’ixa also understand and speak the neighbouring Khoe language Bugakhwe, and those having relatives in the east often have a passive knowledge of Shua. Hence, Ts’ixa is exposed to contact influence from both related (Shua, Khwe) and unrelated (Setswana, English) languages.
9Languages of the Khoe-Kwadi family generally display a preference for verb-final constituent order in pragmatically unmarked clauses (Heine 1976). While speakers of Ts’ixa clearly favour SOV in elicitation contexts, OSV and especially SVO appear with almost equal frequency in narrative texts, suggesting that the constituent order is affected by the information structure of the clause. While the basic constituent order in the noun phrase is head-final, pragmatic considerations may trigger inversion. Ts’ixa nouns are optionally marked by portmanteau morphemes encoding person, gender and number (PGN). These clitics serve the function of specific articles and attach to about 75% of the language’s noun phrases.
10Ts’ixa distinguishes between three syntactic verb classes, according to the number of core participants they may take: intransitives, transitives and S/O-ambitransitives. Ditransitives in the sense of Kittilä (2006), i.e., with two objects treated like the single object of a transitive predication, are not attested. Subject and direct object are optionally case-marked by means of the above-mentioned PGN clitics, which (in singular and plural, but not in dual) display a nominative-accusative distinction. The direct object is optionally marked by the postposition (ʔ)à, which interacts with the argument’s information-structural properties (Fehn 2017; McGregor 2018). Oblique (peripheral) participants are obligatorily marked by a set of semantically specified postpositions.
11Like other Khoe languages, Ts’ixa has a rich suffixing morphology: Derivational affixes attach to both verbs and nouns, and a subset of the language’s tense-aspect morphemes is linked to the verb stem via the so-called juncture morpheme (Elderkin 1986; Heine 1986; Vossen 2010). Predicates may be simple or complex. Complex predicates display semantic features similar to serial verb constructions and involve two or more verbs, which are connected by the juncture morpheme (Fehn 2016; Güldemann & Fehn 2017; Fehn & Phiri 2022).
12Some features like a preference of SVO over SOV in narrative texts, inversed constituent order in the noun phrase, verb-like modifiers and the absence of ditransitive constructions suggest past substrate influence from a dialect of the Kx’a language !Xun (Fehn 2016).
13Ts’ixa expresses commands and obligations, but also advice, pleas and wishes, by using the imperative. The imperative is here understood as a specific clause type with morphosyntactic properties not shared by other clause types, i.e., declarative and interrogative (see, e.g., König & Siemund 2007 on basic clause types). The imperative can be oriented towards 1st, 2nd and 3rd person subjects. Therefore, Ts’ixa does not formally differentiate between “canonical” imperatives (§2.1), and hortative (§2.2), jussive or optative (§2.3) expressions.
14A finite verb phrase in declarative and interrogative clauses requires the presence of a grammatical formative encoding tense, aspect or modality (tam). This formative may either be a particle preceding the verb stem, or a suffix (1a-c). The majority of tense-aspect markers are portmanteau morphemes encoding temporal and aspectual notions, as well as mood. While Ts’ixa has a dedicated interrogative morpheme, the clause final particle rè (1d), declarative clauses do not require a clause type marker.
| (1) |
Examples of declarative and interrogative clauses in Ts’ixa |
|
a. |
Declarative: Imperfective |
|
|
tí |
kò |
gǀã̄ã̀-kàà |
táó.táò=mà |
|
|
1sg |
ipfv |
try-vol |
drive=sg.m |
|
|
‘I want to try driving.’ (elicited) |
|
b. |
Declarative: Hodiernal Past |
|
|
|
ʔájā.kʰóè |
ʔã̄ã́-káxū-nā-tā |
kʰōē=nā |
ʔà |
kʼóā=m̄ |
móláò |
|
|
headman |
know-caus-j-pst |
person=pl.c |
acc |
new=sg.m |
law |
|
|
kà |
kà |
|
|
|
|
|
|
attr |
mpo |
|
|
|
|
|
|
‘The headman informed the people about the new law.’ |
|
c. |
Declarative: Subjunctive |
|
|
kōlóí=sà |
ʔé.m̀ |
xà |
hĩ̄ĩ́ |
ʔíté |
nāá=ḿ |
|
|
car=sg.m |
3sg.m |
sbjv |
do |
neg |
dem.ref=sg.m |
|
|
‘He is not capable of fixing the car himself.’ (elicited) |
|
d. |
Interrogative |
|
|
ǂˀórī |
tú |
ʔāná-há |
rè? |
|
|
sourplum |
2pl.c |
know:j-prf |
q |
|
|
‘Do you know sourplums?’ |
- 3 An anonymous reviewer rightfully observes that imperative expressions are by definition context-sp (...)
15The Ts’ixa imperative follows a cross-linguistically common pattern in using the simple verb stem, which may consist of the bare root or of a root plus derivational affix (Aikhenvald 2010: 18; Aikhenvald 2017: 5; Devos & Van Olmen 2013: 44; König & Siemund 2007: 303). Imperative clauses are incompatible with tam marking and therefore clearly distinct from declarative clauses which are defined by the presence of at least one tam morpheme or a copula-like element.3 As will be seen, the distinction between interrogative and imperative is less clear-cut: the interrogative particle rè may be part of an (unmarked) imperative clause if the speaker wishes to be more polite (§2.1). While Ts’ixa does have a dedicated imperative morphology which includes a conjunction nà, a prohibitive particle tíí, and a subordination marker kè, an imperative morpheme as such is absent. Infrequently, the particle (ʔ)è occurs, possibly as a result of contact influence from Khwe or Shua (see §3.1).
16Canonical (2nd person) imperatives in Ts’ixa are expressed by means of the bare (2a-b) or derived (2c) verb root. The subject pronoun is considered implicit and usually dropped (2a-b), but tends to be expressed when multiple individuals are addressed (2c). Word order in imperative clauses is flexible and aligns with the intended information-structural properties of the utterance. Imperative clauses may be verb-final (2a, 2c) or verb-initial (2b). In (2c), given as reply to the question ‘What shall I boil?’, the object has been fronted and occurs in the focus slot before the openly expressed subject.
| (2) |
Examples for “canonical” imperatives in Ts’ixa |
|
a. |
sēkúélé |
ʔò |
kṹũ̀ |
|
|
school |
dir |
go |
|
|
‘Go to school!’ (elicited) |
|
b. |
ǁám̀ |
ʔʲũ̄ṹ=sí |
ń=sà |
ʔà |
|
|
taste |
food=sg.f |
dem.prox=sg.f |
acc |
|
|
‘Taste this food!’ |
|
c. |
k’ōxú=mà |
tó |
kxābá.kxábā-kàxù |
|
|
meat=sg.m |
2pl.c |
boil-caus |
|
|
‘Boil the meat!’ |
17Ts’ixa does not have a dedicated clause type marker signalling imperative clauses. Very rarely, a particle (ʔ)è appears, which was presumably borrowed from Khwe or Shua where it functions as a vocative and imperative marker (3a) (cf. §3.1: (16)). The Ts’ixa vocative rè is formally identical with the interrogative marker and may even share the same historical source (see §3.2). However, when used as vocative, rè appears right after the subject, and not in clause-final position (compare (1d)). The vocative occasionally appears with overt 2nd person subjects in imperative clauses (3b). Imperatives with a vocative subject pronoun are considered more polite than those without a (vocative) subject pronoun.
| (3) |
“Marked” imperatives in Ts’ixa |
|
a. |
Imperatives with the (borrowed) particle (ʔ)è |
|
|
gáò |
è |
nà |
mũ̄ũ̀ |
|
|
look |
imp |
conj.imp |
see |
|
|
‘Look and see!’ |
|
b. |
Imperatives with a vocative pronoun |
|
|
tsá |
rè |
gáò |
|
|
2sg.m |
voc |
look |
|
|
‘Look!’ |
- 4 A conjunction na is widespread in languages of the Bantu family (see, e.g., Nurse 2006). However, (...)
18Ts’ixa uses a dedicated conjunction nà to link verbs in imperative clauses (4a).4 In declarative and interrogative clauses, other strategies are used, including complex predicates in which individual verb stems are linked by means of the juncture morpheme (4b-c), the conjunction ʔà (4c), and the discourse referential marker tʰí.à (4d).
| (4) |
Strategies of verb linkage in Ts’ixa |
|
a. |
Imperative: nà |
|
|
hāà |
nà |
dzāá |
|
|
come |
conj.imp |
taste |
|
|
‘Come and taste!’ |
|
b. |
Declarative: complex predicate |
|
|
cʰɛ̄ɛ́-tã̀ |
kónò |
sá |
kò |
gūī-ā |
kʰāà |
ǁṹũ̄=sī |
ʔò |
|
|
fall.off-neg.ipfv |
when |
2sg.f |
ipfv |
pick.up-j |
give |
parent=sg.f |
dir |
|
|
ŋǁá=sì |
ǀúḿ-káxù |
|
|
|
|
|
|
|
dem.ref=sg.f |
nurse-caus |
|
|
|
|
|
|
|
‘When (the umbilical cord) does not fall off, you pick up (the child) and give it to the mother, (so that) she makes it nurse.’ |
|
c. |
Declarative: ʔà |
|
|
kʼáá.kʰóè=mà |
ǁkʼāḿ-á |
kérè.kérè |
ʔà |
gǁāā.kʰóè=sà |
tsʼã́ã̄-tā |
|
|
man=sg.m |
beat-j |
repeat |
conj |
woman=sg.f |
steal:j-pst |
|
|
‘(They) tortured the man and kidnapped the woman.’ |
|
d. |
Declarative: tʰí.à |
|
|
lēbótélè=m̀ |
tè |
ǀkʼɛ́ɛ́ |
tʰí.à |
!ˀārá.!ˀárā-sī |
|
|
bottle=sg.m |
pst |
fall |
drm |
break-refl |
|
|
‘The bottle fell and broke.’ (elicited) |
19Along similar lines, the dedicated prohibitive morpheme tíí is restricted to imperative clauses (5a). In declarative and interrogative clauses, the default negation particle ʔíté (5b), and the negated imperfective -tã̀ (5c) appear.
| (5) |
Negation strategies in Ts’ixa |
|
a. |
Imperative: tíí |
|
|
kʰōē=mā |
céé.céè |
tíí, |
cxṹũ̄ |
kò |
cīì |
|
|
person=sg.m |
listen |
proh |
brain |
ipfv |
be.sick |
|
|
‘Don’t listen to this man, he is mad!’ |
|
b. |
Declarative (default): ʔíté |
|
|
bāā.tsʰáā |
tí |
gérè |
k’áà |
ʔíté |
|
|
beer |
1sg |
fut |
drink |
neg |
|
|
‘I will never drink beer.’ |
|
c. |
Declarative (Imperfective): -tã̀ |
|
|
ʔābá=ḿ |
kúè |
ǀkʼáǹ=sè |
gáò-sī |
mēé |
páá-tã̀ |
|
|
dog=sg.m |
ipfv |
bad=advz |
look-refl |
but |
bite-neg.ipfv |
|
|
‘The dog looks dangerous, but it does not bite.’ (elicited) |
20In addition to a dedicated conjunction and negation particle, Ts’ixa imperative morphology also includes a particle kè (6), which is tentatively interpreted as a clause-final subordination marker (but see §3.2).
| (6) |
The subordination particle kè |
|
a. |
xúú |
ʔé.ǹ |
k’ōó |
k’ōxú=mà |
ǁˀābà-nà |
kè |
|
|
let |
3pl.c |
eat.meat |
meat=sg.m |
get.hungry-pst |
sub.imp |
|
|
‘Let them eat the meat, (because) they are hungry.’ |
|
b. |
xúú |
ʔé.m̀ |
kṹũ̀ |
nĩ̄ĩ́.xúè |
ʔé.m̀ |
hĩ̄ĩ́-á-tá |
ʔíté |
kè |
|
|
let |
3sg.m |
go |
that.thing |
3sg.m |
do-j-pst |
neg |
sub.imp |
|
|
‘Let him go, (because) he is innocent (lit. he didn’t do that thing).’ |
21The genuine way to express obligation, i.e., ‘you must do X’, is a simple imperative clause (7a). Especially younger speakers also tend to use the Tswana verb tsónèlà ~ tsónàlà which is grammatically treated as an accomplishment verb (Fehn 2016: 134–136 on lexical verb classes), i.e., its unmarked meaning is ‘to become obliged’. It therefore requires use of the past/perfect suffixes -nà (7b) or -hã ∼ -ha (7c). In the available data, use of tsónèlà ~ tsónàlà is restricted to 2nd person subjects and it is not clear whether it can be used with imperatives directed at 1st or 3rd person subjects. To express obligation for 1st person subjects, the volition suffix -kàà appears to be the preferred choice (10c-d).
| (7) |
The expression of obligation in Ts’ixa |
|
a. |
sá |
kà |
ǁṹũ̄=na |
ʔà |
bɛ̄ɛ̄ |
|
|
2sg.f |
poss |
parent=pl.c |
acc |
obey |
|
|
‘You must obey your parents.’ |
|
b. |
tsá |
tí |
ʔà |
k’ōxú |
kà |
cũ̄ũ̄-ā-mā-tsónèlà-nà |
|
|
2sg.m |
1sg |
acc |
meat |
mpo |
buy-j-ben-must-prf |
|
|
‘You must buy meat for me.’ (elicited) |
|
c. |
tsá |
kṹũ̀-tsónàlà-nà-hà |
Mã̄ṹ |
ʔò |
|
|
2sg.m |
go-must-j-prf |
Maun |
dir |
|
|
‘You must go to Maun.’ |
22While there are various means to mitigate the effect of commands in Ts’ixa, all of them are firmly rooted within the imperative clause type. One possibility to appear more polite, the use of a vocative pronoun, has already been discussed above (3b). Other options include the addition of tsá/sá tí ɟírà ‘I ask you (m/f) to’, commonly translated as ‘please’ (8a), and the formulation of the command as a question, either solely by intonation (8b) or by adding the interrogative particle rè (8c). Following cross-linguistically common patterns (Aikhenvald 2017: 3), advice also falls within the semantic spectrum of the imperative, but requires addition of the clause-initial particle hĩ́ĩ̄.nā.
| (8) |
Polite commands and advice in Ts’ixa |
|
a. |
with ɟírà ‘ask for’ |
|
|
tsá |
tí |
ɟírà |
ǁk’āḿ.ǁk’ám̄ |
tsá |
cã́ã̄-tã̀=sè |
|
|
2sg.m |
1sg |
ask.for |
knock |
2sg.m |
enter-neg.ipfv=advz |
|
|
‘Please knock before you enter (lit. I am asking you to knock before entering).’ (elicited) |
|
b. |
Question without rè (but with rising intonation) |
|
|
tí |
kà |
ǀˀṹũ̄=mā |
dzūú? |
|
|
1sg |
poss |
hair=sg.m |
cut.hair |
|
|
‘Could you cut my hair (lit. Cut my hair)?’ (elicited) |
|
c. |
Question with rè and rising intonation |
|
|
tsá |
hāá.ká-nà-mà |
rè |
tí |
|
|
2sg.m |
bring-j-ben |
q |
1sg |
|
|
‘Could you bring me (some coins)?’ |
|
d. |
with clause-initial hĩ́ĩ̄.nā |
|
|
hĩ́ĩ̄.nā |
ǁxáà.kūm̄.ǁxáà |
ʔà |
tāǹ |
|
|
part |
early.morning |
loc |
get.up |
|
|
‘You should get up early in the morning to see lions.’ |
23The preferred way to formulate polite commands is by what may be considered a fusion of imperative and interrogative clause types (8b-c). While intonation and even clause type marking match the interrogative (cf. 1d), the absence of tam marking is typical of the imperative. It is not clear whether the politeness considerations in Ts’ixa are a genuine feature of the speech community’s ethnolinguistic repertoire, or a recent contact feature introduced through westernized schooling and interactions with tourists. Cross-linguistically, it appears that societies with a mainly egalitarian social organization either lack dedicated imperative expressions (Aikhenvald 2017: 11–12) or use them without associations of dominance and impoliteness (Aikhenvald 2017: 34–35). The latter is certainly also true for Ts’ixa where formulaic expressions like greetings and well-wishes are commonly encoded by the imperative. Ex. (9) provides the widely used formulaic exchange used to say goodbye.
| (9) |
Saying goodbye in Ts’ixa |
|
a. |
kã́ĩ̀=sè |
kṹũ̀ |
|
|
good=advz |
go |
|
|
‘Go well!’ (said by person who is staying behind) |
|
b. |
kã́ĩ̀=sè |
káū |
|
|
good=advz |
stay |
|
|
‘Stay well!’ (said by person who is leaving) |
24For commands directed at a first person (inclusive) subject, the term “hortative” is sometimes used, particularly in languages with dedicated morphosyntactic means to express canonical and non-canonical imperatives (e.g., Overall 2017 on the Peruvian Chicham language Aguaruna). Here, hortative is used more generally for a semantic subtype of commands commonly translated with ‘let’s do X’ which is formally expressed by the imperative. The 1st person dual or plural subject pronoun is always overt and cannot be omitted (10a-b).
| (10) |
Hortative expressions in Ts’ixa |
|
a. |
ǁé |
gūnī |
ʔò |
kṹũ̀ |
|
|
1pl.m |
hunt |
dir |
go |
|
|
‘Let’s go hunting!’ (elicited) |
|
b. |
kʰúḿ |
Kʰʷáí |
ʔò |
kṹũ̀ |
tíí |
ń.tsʰéè |
|
|
1du.c |
Khwai |
dir |
go |
proh |
today |
|
|
‘Let’s not go to Khwai today!’ |
25Weaker hortative expressions or suggestions best translated with ‘we should do X’ are usually expressed as questions, i.e., with the interrogative particle rè (11a-b). The expression of an obligation (‘we need to do X’) involves the volition suffix -kàà following the derived or underived verb (11c-d). In other respects, the morphosyntax of all clauses given under (10) follows the imperative scheme, including lack of tam marking and use of the dedicated imperative morphology, such as the conjunction nà (11b).
| (11) |
Suggestions and obligations |
|
a. |
ŋ́.xúà |
ʔà |
tsé |
ɲṹũ̀ |
rè? |
|
|
here |
loc |
1pl.c |
stay |
q |
|
|
‘Should we stay here?’ |
|
b. |
tsé |
c’ūá |
nà |
kṹũ̀ |
rè? |
|
|
1pl.c |
exit |
conj.imp |
go |
q |
|
|
‘Should we go away?’ |
|
c. |
ǁé |
kà |
kɛ́ɛ̄=mā |
ǁé |
ǁʰāí-á |
céè-tì-kū-kàà |
|
|
1pl.m |
poss |
belt=sg.m |
1pl.m |
pull-j |
pull-dur-coll-vol |
|
|
‘We need to tighten our belts.’ (elicited) |
|
d. |
gǀāná=dzì |
mĩ̄ĩ́=dzà |
tsé |
tsxɛ́ɛ́-xù-tì-kū-kàà |
|
|
leaf=pl.f |
dem.dist=pl.f |
1pl.c |
tear-compl-dur-coll-vol |
|
|
‘We need to tear off those leaves.’ |
26In some imperative clauses with 1st person subjects, irrespective of whether the intended meaning is a command (12a), an obligation (12b) or a goal (12c), the suffix -tígèrè appears. When co-occurring with the volitional suffix, -tígérè precedes -kàà (12b). The origin, semantic scope, and usage frame of -tígérè will require further research.
| (12) |
The suffix -tígèrè |
|
a. |
ǁé |
ǁxōɛ́-tígérè |
kálāsā |
ʔà |
nà |
ǀˀēé=mà |
tābú-káxù |
|
|
1pl.m |
lop-sfx |
branch |
acc |
conj.imp |
fire=sg.m |
burn-caus |
|
|
‘Let's lop (some) branches and get the fire burning.’ |
|
b. |
ǀˀáò |
kà |
ŋgǁōɛ́ |
kà |
tsé |
kò |
hĩ̄ĩ́-ká=dzà |
|
|
money |
attr |
month |
mpo |
1pl.c |
ipfv |
do-inst=pl.f |
|
|
tsé |
ǁūḿ-káxù-tígérè-kàà |
|
|
|
|
|
1pl.c |
short-caus-sfx-vol |
|
|
|
|
|
‘We need to reduce the money we are spending each month.’ (elicited) |
|
c. |
ŋṹṹ=mà |
tsé |
ǁʰāá-xú-tígérè |
ŋgūú |
tsé |
kò |
ŋǁānī-kàà |
kónò |
|
|
place=sg.m |
1pl.c |
clear-compl-sfx |
house |
1pl.c |
ipfv |
build-vol |
when |
|
|
‘We have to clear the place if we want to build a house.’ |
- 5 There erroneously given as tíí.
27The imperative clause type in Ts’ixa also permits for 3rd person subjects, usually when a jussive or optative reading is intended (13a). Imperative clauses with 3rd person subjects are most common in reported speech featuring the quotative particle tʰī (13b-d). tʰī usually follows the complement clause and most likely derives from a verb tʰii ‘be thus’ (cf. also Fehn 2016: 263).5 The verb tʰī retains certain properties of a verb, such as the ability to receive tense-aspect marking as in (13c) where the verbum dicendi has been elided. Embedded imperative clauses cannot receive tam marking and display imperative morphology, such as the prohibitive particle tíí (13d).
| (13) |
Jussive and optative expressions in Ts’ixa |
|
a. |
ʔé.m̀ |
kṹũ̀ |
|
|
3sg.m |
go |
|
|
‘He should go/ He may go.’ |
|
b. |
ǁṹũ̄=m̄ |
kà |
ǁṹũ̄=m̄ |
k’ārō-ǀṹã̀ |
tè |
ɟírà |
ʔé.ǁù |
|
|
parent=sg.m |
poss |
parent=sg.m |
boy-dim |
pst |
ask.for |
3pl.m |
|
|
ǀˀēé=dzà |
ʔà |
ǀˀēé=ḿ |
ʔà |
cã́ã̄ |
tʰī |
|
|
|
firewood=pl.f |
acc |
fire=sg.m |
loc |
enter |
quot |
|
|
|
‘The grandfather asked the little boys to add firewood to the fire.’ |
|
c. |
Tʰelatʰile=sì |
kúú |
tʰī-nā-hā |
|
|
Thelathile=sg.f |
gather.firewood |
quot-j-prf |
|
|
‘Thelathile (told the children) to gather firewood.’ |
|
|
d. |
tsé |
kà |
dáò=dzà |
tsé |
ǀk’ũ̄ṹ-nà |
ʔé.ǹ |
tsé |
ʔà |
kámà |
|
|
1pl.c |
poss |
track=pl.f |
1pl.c |
kill-pst |
3pl.c |
1pl.c |
acc |
pursue |
|
|
tíí |
tʰī |
|
|
|
|
|
|
|
|
|
proh |
quot |
|
|
|
|
|
|
|
|
|
‘We covered our tracks, so that they would not follow us.’ |
28In the following two sections, an expanded perspective on the expression of imperatives in the Khoe-Kwadi language family will be provided. As many known Khoe-Kwadi languages are already extinct or severely endangered, an attempt was made to assemble as many examples as possible from a maximally wide range of sources. In part owing to the exploratory nature of the present article, in part to the limited amount of data available, I will restrict myself to the morphological representation of canonical imperatives directed at 2nd person subjects. §3.1 presents a general overview of imperative clauses and morphology, and §3.2 looks at Khoe-Kwadi imperative morphology from a diachronic perspective while also considering contact influence from languages of the Kx’a and Tuu families within the Kalahari Basin typological area (Güldemann 1998; Güldemann & Fehn 2017). Although occasional reference to typological features attested elsewhere is made, a thorough assessment of the observed patterns in a cross-linguistic framework is beyond the scope of this work and will hopefully form the subject of future studies.
29To facilitate comparison, all data taken from published or archival sources has been transliterated to match IPA standards, and glossing has been adapted. When deemed necessary, the original transcription is provided in brackets < >.
30From the available data, it seems clear that the use of zero marking to express canonical imperatives is a cross-Khoe feature. All varieties surveyed displayed imperatives directed at 2nd person subjects with the following features:
- absence of tam marking
- optional use of a dedicated clause type marker
- optional expression of pronominal subjects
31Examples of unmarked imperatives for all major subgroups and language complexes of the Khoe branch of Khoe-Kwadi (see Figure 2) are provided in (14).
| (14) |
Unmarked imperatives across Khoe |
|
Kalahari Khoe: Khwe |
|
a. |
á.à, |
tʃá |
tí |
ʔà |
kx’ṍã́ |
|
|
no |
2sg.m |
1sg |
acc |
wait |
|
|
‘No, you have to wait for me!’ (Khwe; Kilian-Hatz 2008: 122) |
|
b. |
tsá |
ǂí |
ʔā=m̄ |
ǀȭã̄=mà |
|
|
2sg.m |
call |
dem=sg.m |
child=sg.m |
|
|
‘Call the boy!’ (ǁAni; Heine 1999: 63) |
|
Kalahari Khoe: Shua |
|
c. |
ʔũ̀ã́ |
màrí |
cí |
ʔò |
|
|
bring |
money |
1sg |
dir |
|
|
‘Bring the money to me.’ (Gǁoro; Westphal n.d. a) |
|
d. |
tí |
càā.hú |
ŋǀè |
ǁxáó |
ʔúú. |
|
|
1sg |
brother |
dir |
spear |
take.to |
|
|
‘Take the spear to my brother.’ (Deti; Westphal n.d. b) |
|
Kalahari Khoe: Tshwa |
|
e. |
yii |
k’oo.xo |
kaxu |
ka |
ǁx’um-a-xu. |
|
|
dem |
meat |
knife |
with |
cut-j-compl |
|
|
‘Cut this meat with a knife.’ (Tjwao; Phiri n.d.) |
|
f. |
<khonwa nyo> |
|
|
xom |
wa |
ɲũũ |
|
|
ground |
loc |
sit |
|
|
‘Sit on the ground!’ (Hiechware; Dornan 1917: 71) |
- 6 The Gǀui-Gǁana variety referred to as “Kua” by Collins (2023) is not identical with the Tshwa dial (...)
|
Kalahari Khoe: Gǀui-Gǁana6 |
|
g. |
c’ṹĩ̀-χà |
kṹũ̀ |
|
|
beautiful-advz |
go |
|
|
‘Go well!’ (Kua; Collins 2023: 36) |
|
Kalahari Khoe: Naro |
|
h. |
ǂˀõõ |
|
|
eat |
|
|
‘Eat!’ (Naro; Visser 2013: 198) |
|
Khoekhoe: Nama-Damara |
|
i. |
mũ̀ũ̀ |
|
|
see |
|
|
‘See!’ (Nama; Hagman 1977: 146) |
|
Khoekhoe: Haiǁom |
|
j. |
ma |
te |
tsi |
ko |
|
|
give |
1sg |
2sg.m |
emph |
|
|
‘Do give it to me [now]!’ (Haiǁom; Widlok 2013: 349) |
|
Khoekhoe: !Ora |
|
k. |
!ae |
|
|
wait |
|
|
‘Wait!’ (!Ora; Meinhof 1930: 60) |
32Westphal’s Kwadi field notes (Westphal n.d. c) include only few examples for imperative clauses. In two of the examples available (15a-b), the bare, unmarked verb-stem appears. The third example contains a postverbal element -ʔà which may be an optional imperative marker, but is most likely linked to (indirect) object-indexing. While the addressees in (15a) and (15c) are not overtly expressed, (15b) features a pronominal subject of the 2nd person dual.
| (15) |
Unmarked imperatives in Kwadi |
|
a. |
<xwã!> |
|
|
xũã |
|
|
‘Dance!’ (Westphal n.d. c) |
|
b. |
<uwa xwã!> |
|
|
u.wa |
xũã |
|
|
2du |
dance |
|
|
‘Dance (you two)!’ (Westphal n.d. c) |
|
c. |
<tanǀ’óká’a> |
|
|
ta |
ǀˀó |
ká-ʔà |
|
|
1sg |
fire:?sg.m |
give-? |
|
|
‘Give me fire!’ (Westphal n.d. c) |
33While all Khoe-Kwadi languages have zero-marked imperatives, not all languages automatically link absence of tam morphology with this particular clause type. For example, zero-marked verb forms in Khwe are also used to express thetic events (Kilian-Hatz 2008: 121), although imperative is still considered the “default interpretation” (Kilian-Hatz 2008: 123).
34Some of the Khoe languages surveyed also make optional use of a morpheme which could be interpreted as an imperative clause type marker (16). While the most common position for the imperative marker is clause final, it may also occur in clause-second (Gǁoro (16c)) or clause-initial position (Kua (16g)). An imperative marker ʔè, è or wè also used as vocative is found in Khwe and Shua (16a-e) (Kilian-Hatz 2008: 124; Heine 1999: 64; Westphal n.d. a; n.d. d). The very limited Tshwa data did not reveal any overt imperative marking, save for the singular appearance of an element ho in Hiechware (16f) which may be cognate with Naro o (16h). The Khoekhoe varieties Nama and Haiǁom both use a particle re (16i-k), which has additional connotations of politeness (i.e., ‘please’) in Nama (Hagman 1977: 146; Haacke 2013: 335–336).
| (16) |
Marked imperatives across Khoe |
|
Kalahari Khoe: Khwe (ʔè) |
|
a. |
jāá |
ʔè |
|
|
come |
imp |
|
|
‘Come!’ (Khwe; Kilian-Hatz 2008: 126) |
|
b. |
kṹũ̀ |
(è) |
ǁˀáé |
(kà) |
|
|
go |
imp |
home |
loc |
|
|
‘Go home!’ (ǁAni; Heine 1999: 63) |
|
Kalahari Khoe: Shua (wè ~ è) |
|
c. |
ʔěmérè |
wē |
tí |
ʔò |
ʔṹã́ |
|
|
bucket |
imp |
1sg |
dir |
bring |
|
|
‘Bring the bucket to me.’ (Gǁoro; Westphal n.d. a) |
|
d. |
mṹũ̀ |
gǁōô |
xáḿ |
wê |
|
|
see |
big |
lion |
imp |
|
|
‘Look at the big lion!’ (Nata-Shua; Westphal n.d. d) |
|
e. |
ǀíí |
tó |
wè |
|
|
sing |
2pl.c |
imp |
|
|
‘Sing!’ (Deti; Westphal n.d. b) |
|
Kalahari Khoe: Tshwa (ho) |
|
f. |
<ama ‖khaho> |
|
|
ʔama |
‖x’aa |
ho |
|
|
pot |
wash |
?imp |
|
|
‘Wash the pot!’ (Hiechware; Dornan 1917: 71) |
|
Kalahari Khoe: Gǀui-Gǁana (nà) |
|
g. |
nà |
cúá |
kṹũ̀ |
|
|
imp |
neg |
go |
|
|
‘Don’t go!’ (Kua; Collins 2023: 36) |
|
Kalahari Khoe: Naro (o) |
|
h. |
ǂxobe.kx’am |
te |
o |
|
|
open |
1sg |
imp |
|
|
‘Open (for) me!’ (Naro; Visser 2013: 198) |
|
Khoekhoe: Nama-Damara (re) |
|
i. |
ŋǂaa |
re |
|
|
dance |
imp |
|
|
‘Please dance!’ (Nama; Haacke 2013: 336) |
|
Khoekhoe: Haiǁom (re) |
|
j. |
ha |
re |
|
|
come |
imp |
|
|
‘Come!’ (Haiǁom; Widlok 2013: 349) |
|
Khoekhoe: !Ora (re) |
|
k. |
ʔau |
re |
|
|
give |
imp |
|
|
‘Give (me)!’ (!Ora; Meinhof 1930: 60) |
35While the Zimbabwean Tshwa variety Tjwao does not seem to have a dedicated imperative clause type marker, a special set of 2nd person pronouns different from declarative clauses is used whenever the subject is overtly expressed (17a-b). A similar phenomenon has been observed in the Gǀui-Gǁana variety Kua by Collins (2023: 36), who dubs this special paradigm “subjunctive” (17c-d).
| (17) |
2nd person declarative and “subjunctive” pronouns in Tjwao and Kua |
|
a. |
tʃa |
kua |
ti.a |
tʰũũ-kaxu |
|
|
2sg.m |
ipfv |
1sg.acc |
injure-caus |
|
|
‘You are making me (meaning: my head) hurt.’ (Tjwao; Phiri n.d.) |
|
b. |
tʃi |
kx’ui-a |
kũũ |
tii, |
tʃi |
tʃee.tʃee |
na |
kũũ |
|
|
2sg.m |
talk-j |
go |
proh |
2sg.m |
listen |
conj.imp |
go |
|
|
‘Don’t walk talking, walk and listen!’ (Tjwao; Phiri n.d.) |
|
c. |
sá |
è |
kwà |
kṹũ̀ |
|
|
2sg.f |
sbj |
prog |
go |
|
|
‘You are going.’ (Kua; Collins 2023: 172) |
|
d. |
sì |
cúá |
kṹũ̀ |
|
|
2sg.f |
neg |
go |
|
|
‘You don’t go!’ (Kua; Collins 2023: 38) |
36Negated imperative or prohibitive clauses are found across the Khoe subgroup of Khoe-Kwadi, but not all languages actually display a dedicated prohibitive marker. The most widespread prohibitive morpheme, the clause-final particle tii ~ cii, is found in Khwe, Ts’ixa (5), Shua, Tshwa and Gǀui-Gǁana. In some languages, tii ~ cii can optionally be combined with the imperative marker (e.g., Gǁoro (18c); Deti (18e)). The Tshwa dialect Hiechware has an additional prohibitive marker, the clause-initial particle g|ana, which can appear on its own (18g), but is also found combining with tii (18h). In the Gǀui-Gǁana variety Kua, the generic negation morpheme cúá (18j) may be used instead of cīī (18i). It appears that cúá (but not cīī) can co-occur with the clause-initial imperative marker nà (18j). In Naro (18k) and Khoekhoe (18n-p), the imperative is negated by means of the clause-initial negation particle táá. While táá is a dedicated prohibitive morpheme in Nama, it is a general negation particle in Naro (Visser 2013: 199–200) and Haiǁom (Widlok 2013: 163; 350), and a negator of imperative and subordinate clauses in !Ora (Meinhof 1930: 55). Naro makes use of an additional prohibitive marker xūū which follows the verb and can either replace (18l) or accompany táá (18m). It may be assumed that xūū derives from the verb xūū ‘leave behind, come from, stop’ (Visser 2001: 30). While táá co-occurs with the imperative rè in both Nama (Hagman 1977: 147) and Haiǁom (Widlok 2013: 351), none of the prohibitive morphemes in Naro can co-occur with the imperative o (Visser 2013: 200).
| (18) |
Prohibitive clauses across Khoe |
|
Kalahari Khoe: Khwe (tí) |
|
a. |
tává-è, |
m̀.ḿ, |
tʃá |
ǂˀṹṹ |
tí |
|
|
uncle-voc |
no |
2sg.m |
eat |
proh |
|
|
‘Uncle, no, don’t eat it!’ (Khwe; Kilian-Hatz 2008: 128) |
|
b. |
ɦàâ |
tí |
|
|
come |
proh |
|
|
‘Don’t come!’ (ǁAni; Heine 1999: 64) |
|
Kalahari Khoe: Shua (tíí) |
|
c. |
jìī |
ʔà |
é |
céè |
tíí |
|
|
wood |
acc |
imp |
pull |
proh |
|
|
‘Don’t pull the log!’ (Gǁoro; Westphal n.d. a) |
|
d. |
ǀ’áǹ |
k’òxú |
tâ |
kʰàà |
tíí |
ná |
t’úɲè |
tâ |
kʰàà |
|
|
bad |
meat |
1sg.acc |
give |
proh |
conj.imp |
good:cop |
1sg.acc |
give |
|
|
‘Don’t give me bad meat, give me good!’ (Nata Shua; Westphal n.d. d) |
|
e. |
ǀíí |
tíì |
tó |
wè |
|
|
sing |
proh |
2pl.c |
imp |
|
|
‘Don’t sing now!’ (Deti; Westphal n.d. b) |
|
Kalahari Khoe: Tshwa (tíí) |
|
f. |
tʃi |
karo-ra |
gam-a-xu |
tii |
|
|
2sg.m |
stone-pl |
throw-j-compl |
proh |
|
|
‘Don’t throw the stones!’ (Tjwao; Phiri n.d.) |
|
g. |
<|gana tsaa hwee> |
|
|
g|ana |
tsʰaa |
xue |
|
|
proh |
water |
run.away |
|
|
‘Do not run away from the water!’ (Hiechware; Dornan 1917: 73) |
|
h. |
<ǀgana chi kwi thii> |
|
|
gǀana |
tʃi |
k’ui |
t(ʰ)ii |
|
|
proh |
1sg |
speak |
proh |
|
|
‘Don’t talk to me!’ (Hiechware; Dornan 1917: 73) |
|
Kalahari Khoe: Gǀui-Gǁana (cīī, cúá) |
|
i. |
Thabo=m̀ |
ǀχòà |
qχ’úī |
cīī |
|
|
Thabo=sg.m |
com |
talk |
proh |
|
|
‘Don’t talk to Thabo!’ (Kua; Collins 2023: 36) |
|
j. |
(nà) |
cúá |
Thabo=m̀ |
ǀχòà |
qχ’úī |
|
|
imp |
neg |
Thabo=sg.m |
com |
talk |
|
|
‘Don’t talk to Thabo!’ (Kua; Collins 2023: 36) |
|
Kalahari Khoe: Naro (táá, xūū) |
|
k. |
táá |
tséé |
|
|
neg |
work |
|
|
‘Don’t work!’ (Naro; Visser 2013: 200) |
|
l. |
tséé |
xūū |
|
|
work |
proh |
|
|
‘Don’t work!’ (Naro; Visser 2013: 200) |
|
m. |
táá |
tséé |
xūū |
|
|
neg |
work |
proh |
|
|
‘Don’t work!’ (Naro; Visser 2013: 200) |
|
Khoekhoe: Nama-Damara (táá) |
|
n. |
táá |
!úũ |
|
|
proh |
go |
|
|
‘Don’t go!’ (Nama; Hagman 1977: 147) |
|
Khoekhoe: Haiǁom (ta) |
|
o. |
ta |
tu |
!ao |
re |
|
|
neg |
2pl.c |
fear |
imp |
|
|
‘Don’t be afraid!’ (Haiǁom; Widlok 2013) |
|
Khoekhoe: !Ora (ta) |
|
p. |
ta |
ǀχ’a |
|
|
proh |
steal |
|
|
‘Don’t steal!’ (!Ora; Meinhof 1930) |
37Data on conjoined imperatives is fragmented and does not allow for a comprehensive assessment of the situation across Khoe. While the Ts’ixa imperative conjunction nà (4) is widely attested in Kalahari Khoe, including Khwe (19a), Shua (19c-e), Tshwa (19f) and Gǀui-Gǁana (19g), the data is not always sufficient to say whether the particle is strictly dedicated to imperative contexts or may also appear elsewhere. In the Gǀui-Gǁana variety Kua, nà stands in complementary distribution with the default verbal coordination marker à: nà appears in imperative clauses, and à elsewhere (Collins 2023: 40). A similar analysis may also hold for Ts’ixa, and possibly for various dialects of Khwe, Shua and Tshwa. Other strategies of coordinating verbs or clauses in the imperative include use of generic conjunctions (Khwe (19b); Naro (19h)), or simple juxtaposition (Naro (19i)). No data on conjoined imperatives could be retrieved from languages of the Khoekhoe cluster, i.e., Nama-Damara, Haiǁom and !Ora.
| (19) |
Conjoined imperatives in Kalahari Khoe |
|
Kalahari Khoe: Khwe |
|
a. |
jāá |
nà |
ǁx’áō |
tí |
à |
|
|
come |
conj.imp |
help |
1sg |
acc |
|
|
‘Come and help me!’ (Khwe; Kilian-Hatz 2008: 287) |
|
b. |
jāá |
nò |
ǁx’áō |
tí |
à |
|
|
come |
conj |
help |
1sg |
acc |
|
|
‘Come and (then) help me!’ (Khwe; Kilian-Hatz 2008: 287) |
|
Kalahari Khoe: Shua |
|
c. |
ʔāū |
ŋgǀóá |
tíí |
ná |
k’òxú |
ŋgǀóá |
|
|
fish |
cook |
proh |
conj.imp |
meat |
cook |
|
|
‘No, don’t cook fish, cook meat.’ (Gǁoro; Westphal n.d. a) |
|
d. |
ǀ’áǹ |
k’òxú |
tâ |
kʰàà |
tíí |
ná |
t’úɲè |
tâ |
kʰàà |
|
|
bad |
meat |
1sg.acc |
give |
proh |
conj.imp |
good:cop |
1sg.acc |
give |
|
|
‘Don’t give me bad meat, give me good!’ (Nata-Shua; Westphal n.d. d) |
|
e. |
ʔáū |
ŋǀòá |
tíì |
nā |
k’òxú |
ŋǀóá |
è |
|
|
fish |
cook |
proh |
conj.imp |
meat |
cook |
imp |
|
|
‘No, don’t cook fish, cook meat.’ (Deti; Westphal n.d. b) |
|
Kalahari Khoe: Tshwa |
|
f. |
tʃãã |
na |
xum |
ǀxe |
ɲũũ |
|
|
enter |
conj.imp |
ground |
loc |
sit |
|
|
‘Come in and sit down.’ (Tjwao; Phiri n.d.) |
|
Kalahari Khoe: Gǀui-Gǁana |
|
g. |
ʔè |
qχ’úō |
nà |
kṹũ̀ |
|
|
3sg.c |
eat |
conj.imp |
go |
|
|
‘Eat and go!’ (Kua; Collins 2023: 40) |
|
Kalahari Khoe: Naro |
|
h. |
hàà |
nākā |
tēē |
kx’áà |
|
|
come |
conj |
tea |
drink |
|
|
‘Come and drink tea!’ (Naro; Visser 2013: 391) |
|
i. |
!ṍõ̀ |
síí |
tséé |
|
|
walk |
go |
work |
|
|
‘Go and start working!’ (Naro; Visser 2013: 388) |
38There is not enough data to elaborate further on cross-Khoe patterns of expressing commands, including dedicated morphology for subordinate clauses. It is, however, worth mentioning that Naro has a clause-final particle kē which marks subordinate clauses that provide a reason for the command expressed in the accompanying imperative clause (Visser 2013: 393) (20). As shown in (6), ke also appears in Ts’ixa and probably displays full form-function correspondence with its counterpart in Naro.
| (20) |
The subordination particle kē in Naro |
|
xūū |
tē |
tsàā=r |
kō |
kē |
|
leave |
1sg |
be.sick=1sg |
cont |
sub.imp |
|
‘Leave me, (because) I am sick!’ (Visser 2013: 393) |
39The results of the survey presented in §3.1 reveal that not only Ts’ixa, but all Khoe-Kwadi languages use a cross-linguistically common pattern by expressing canonical imperatives through the unmarked verb stem (cf. also Aikhenvald 2010: 18; Aikhenvald 2017: 5; Devos & Van Olmen 2013: 44; König & Siemund 2007: 303). However, it is not clear for all languages whether zero-marking is inextricably linked to the imperative clause type (as in Ts’ixa), or also allows for other readings, such as theticity or perfective (cf. Khwe, Kilian-Hatz 2008: 121). A historical assessment of the relationship between absence of tam morphology and the imperative would therefore require a better synchronic appraisal than is presently available. Along similar lines, more data from a large number of languages would be needed to evaluate the extension of imperative morphology to commands directed at 1st and 3rd person subjects. The available information from Ts’ixa (this study), Khwe (Kilian-Hatz 2008: 121) and Nama (Hagman 1977: 146–148; Haacke 2013: 335–336) suggests that the Khoe imperative can be used with 1st, 2nd and 3rd person subjects.
40No dedicated imperative marker can be reconstructed for proto-Khoe-Kwadi or proto-Khoe. However, most Khoe languages make optional use of a clause-initial, clause-second or clause-final particle which may be interpreted as an imperative clause type marker. From a comparative perspective, the attested imperative markers partially overlap with vocative and interrogative morphology. The cross-linguistically common “intimate connection” between imperatives and vocatives (Aikhenvald 2010: 48) may therefore be expanded to all addressee-oriented forms in Khoe, i.e., imperative, vocative and interrogative (see also Overall 2017 for similar observations on Aguaruna). In this semantic framework, at least two particles with a link to imperative morphology, (ʔ)e and re, are attested across Khoe. A third particle or suffix, (-ʔ)o, either has a more limited distribution, or displays a broader semantic scope including declarative and subordinating morphology (Table 1).
Table 1 — Cross-Khoe morphology displaying a link to imperative clause type marking (Sources: Kilian-Hatz (2008) for Khwe, Fehn (2016) for Ts’ixa, Westphal (Westphal n.d. a; n.d. b; n.d. d) for Shua, Phiri (2021) and Dornan (1917) for Tshwa, Collins (2023) for G|ui-Gǁana, Visser (2013) for Naro, Hagman (1977) and Haacke (2013) for Nama-Damara, Meinhof (1930) for !Ora, and Widlok (2013) for Haiǁom)
|
Khwe |
Ts’ixa |
Shua |
Tshwa |
Gǀui-Gǁana |
Naro |
Nama |
Haiǁom |
!Ora |
| (ʔ)e |
voc, imp |
|
voc, imp |
voc, inter |
? |
voc |
voc |
? |
? |
| re |
inter |
inter, voc |
inter |
|
n/a |
|
imp |
imp |
imp |
| (-ʔ)o |
?sub |
?sub |
?decl (Deti) |
?imp |
n/a |
imp |
?sub |
inter, ?sub |
?sub |
41The particle (ʔ)e is most widely attested as a vocative marker following nouns and pronouns, i.e., in Khwe (Kilian-Hatz 2008: 82–83), Shua (Westphal n.d. a; n.d. b; n.d. d; own data), Tshwa (Phiri n.d.), Naro (Visser 2013: 387), and Nama (Hagman 1977: 149; Haacke 2013: 339). If the expression of vocatives is assumed to be its main function, the marking of imperative (Khwe, Shua) and interrogative (Tshwa) clauses are best considered semantic extensions and therefore secondary functions of the particle. While the distribution of (ʔ)e would technically allow for a proto-Khoe reconstruction as vocative, borrowing from an unrelated language within the Kalahari Basin linguistic area (Güldemann 1998; Güldemann & Fehn 2017) cannot be excluded. For example, the Groot Laagte variety of the !Xun dialect ǂKx’aoǁ’ae displays a vocative suffix -è (Pratchett 2020a: 67), but lack of data on !Xun vocatives in general makes it difficult to assess whether ǂKx’aoǁ’ae borrowed from Khoe or the other way around. Following a different trail, all !Xun dialects display a proximal demonstrative *ē which follows the head noun (Heine & König 2015: 153–155) and may have been reinterpreted by Khoe speakers as vocative in the sense of (e.g.) ‘you here!’. Yet another candidate for a link with the Khoe form is the imperative suffix -we which appears with 2nd person plural subjects in the Tuu language Nǀuu (Collins & Namaseb 2011: 57) and shows full formal correspondence with the Shua imperative marker we (16c-e).
42The particle re marks interrogatives in the Kalahari Khoe languages Khwe (Kilian-Hatz 2008: 297–299), Ts’ixa (Fehn 2016: 281–283) and Shua (Westphal n.d. a; n.d. b; n.d. d; own data), and imperative in Khoekhoe (Meinhof 1930: 60; Hagman 1977: 146–148; Widlok 2013: 349). Ts’ixa shows an extension from interrogative to vocative (Fehn 2016: 120), which probably constitutes a language-internal development. It will be remembered that the most common strategy to formulate polite commands in Ts’ixa is by way of adding the interrogative marker re to the imperative clause (8c). While not attested with questions in Nama, the appearance of re still carries connotations of politeness, suggesting that the particle’s interrogative function is the more basic one. This is also supported by data from !Xun, a language with a long history of areal contact with Khoe-Kwadi languages (Güldemann 1998; Güldemann & Fehn 2017), where clause-final *re is widely attested as an interrogative marker (Heine & König 2015: 245).
43Following a parallel line of argumentation, the suffix -o which appears as an imperative marker in Naro may be related to the interrogative particle o attested in Haiǁom (Widlok 2013: 350). An interrogative o has not been described for other Khoekhoe varieties, but a clause-final interrogative particle wò exists in the W2 dialect of !Xun (Heine & König 2015: 245). While the !Xun, Haiǁom and Naro forms could be conveniently linked through language contact, the argument is difficult to extend to the imperative marker ho observed in the geographically distant Tshwa dialect Hiechware (Dornan 1917). It therefore remains unclear whether the Naro and Hiechware imperatives are actually related, and, if yes, whether a link between the Haiǁom and !Xun interrogatives should be assumed. The issue is further complicated by the wide distribution of particles of the form (ʔ)oo which display a link to clause type marking. In Khwe, Ts’ixa, and Khoekhoe, a particle (ʔ)oo appears as marker of subordinate clauses, transmitting the notion of ‘when, while’ in Khwe (Kilian-Hatz 2003: 152; Kilian-Hatz 2008: 318), Nama (Hagman 1977: 128–129), Haiǁom (Widlok 2013: 165) and !Ora (Meinhof 1930: 80), and of ‘because’ in Ts’ixa (Fehn 2016: 266–268). In the Shua dialect Deti, a presumably unrelated element (ʔ)o acts as declarative marker (Westphal n.d. b; own data).
44As the above discussion has shown, it is difficult to reconstruct any of the attested imperative markers to the proto-Khoe stage. The interweaving of imperatives with other addressee-oriented morphology, as well as the density of contact between unrelated languages in the Kalahari Basin linguistic area creates a fuzzy situation which may only be resolved when more data from a greater variety of languages becomes available.
45A less complex picture emerges with regards to the prohibitive morphology: the available data shows a quite clear-cut separation between Kalahari Khoe *tii and Khoekhoe *taa, whereas Naro aligns with Khoekhoe, possibly due to contact (Table 2). As no Kx’a, Tuu or Bantu language displays formally similar particles, it may be assumed that both *tii and *taa are genuine Khoe morphemes which may be reconstructed to the proto-Kalahari and proto-Khoekhoe stages, respectively.
Table 2 — Cross-Khoe prohibitive morphology (Sources: Kilian-Hatz (2008) for Khwe, Fehn (2016) for Ts’ixa, Westphal (Westphal n.d. a; n.d. b; n.d. d) for Shua, Phiri (2021) and Dornan (1917) for Tshwa, Collins (2023) for G|ui-Gǁana, Visser (2013) for Naro, Hagman (1977) and Haacke (2013) for Nama-Damara, Meinhof (1930) for !Ora, and Widlok (2013) for Haiǁom)
|
Khwe |
Ts’ixa |
Shua |
Tshwa |
Gǀui-Gǁana |
Naro |
Nama |
Haiǁom |
!Ora |
| *tii |
proh |
proh |
proh |
proh |
proh |
|
|
|
|
| *taa |
|
|
|
|
|
proh, neg |
proh |
proh, neg |
proh, neg.sub |
46While Kalahari Khoe *tii appears to be a dedicated prohibitive marker, Khoekhoe *taa has a more general profile in Naro (Visser 2013: 199–200), Haiǁom (Widlok 2013: 351), and !Ora (Meinhof 1930: 55). Beyond the mere expression of imperative morphology, there is some reason to assume that the Kalahari Khoe prohibitive *tii is further related to an element *ti ~ *te which appears as part of future tense negators in Naro (Visser 2013: 200), Nama (Hagman 1977: 89–91; Haacke 2013: 338), Haiǁom (Widlok 2013: 163) and !Ora (Meinhof 1930: 55). If one is inclined to speculate, the element may even be related to the Kwadi negation particle tʃi ~ ci (or tʃe ~ ce) which is found combining with different morphological elements in declarative and interrogative clauses (Westphal n.d. c). While no further reconstruction will be attempted here, it is still tempting to consider the possibility of Kalahari Khoe imperatives preserving archaic negation morphology.
47Due to a lack of comprehensive data, it is difficult to comment on the history of other morphological elements linked to the imperative clause type. Nevertheless, some very preliminary observations may be added in order to showcase opportunities for future research:
48The imperative conjunction na appears in Khwe, Ts’ixa, Shua, Tshwa and Gǀui-Gǁana (19), but seems to be absent in Naro and Khoekhoe. Its distribution may therefore merit a reconstruction to the stage of proto-Kalahari Khoe. It is unclear whether na is related to other grammatical markers of the same form which occur in the coordinating and subordinating morphology across Khoe, e.g. Naro nā(kā) (cf. Visser 2013: 390–391).
49The dedicated clause-final particle ke adding background information to commands is only attested in Ts’ixa and Naro, but seems especially interesting from a historical viewpoint. In their comparative survey of clause-second structures in southern African “Khoisan” languages, Güldemann et al. (2019) show a link between the overuse of particles marking thetic, i.e., background information, and the grammaticalization of declaratives. The Nama declarative marker ke, albeit a clause-second morpheme delimiting a pragmatically marked prefield, is formally identical with the particle ke flagging background information in Naro and Ts’ixa. One may therefore speculate that the Nama declarative and the imperative subordinator ke are historically related and go back to a proto-Khoe morpheme marking theticity (which, on its part, may trace back to a copula, cf. Güldemann et al. 2019: 59). However, given the strikingly different syntactical frameworks, in combination with the fact that clause-second markers of the shape kV are widespread in the Kalahari Basin (Güldemann et al. 2019), compare, e.g., Nǀuu ke (cf. Collins & Namaseb 2011: 8–9), any further attempt to pursue a historical relationship between different particles ke in Khoe should proceed with care.
50In this article, I have attempted to highlight possible areas of interest for the study of imperatives in the Khoe-Kwadi language family and beyond. As the data from Ts’ixa shows, the expression of commands can be described in terms of morphosyntax, but also needs to take into account the conversational and ethnolinguistic context in which a directive statement is made. By further adopting a comparative perspective, I have stressed the need for a comprehensive description of imperative clauses across the family. The limited data available shows a close association between imperative morphology and other addressee-oriented utterances, such as vocatives and interrogatives. While the expression of imperatives by means of unmarked verb stems could be shown to be a cross-family feature, the morphology of imperatives displayed a considerable degree of cross-linguistic variation, which may in part have been triggered by language contact. Both historical and areal perspectives on imperatives in Khoe-Kwadi therefore seem to provide promising fields for future research.
51Data on Ts’ixa was collected under a research permit issued by the Ministry of Youth, Sports and Culture (2010–2013) of the Government of Botswana and funded by the a.r.t.e.s. Graduate School of the University of Cologne, the German Academic Exchange Service (Deutscher Akademischer Austauschdienst) DAAD, the project “The Kalahari Basin area: a ‘Sprachbund’ on the verge of extinction” within the European Science Foundation EUROCORES programme EUROBABEL, and by the Humboldt University of Berlin. The author is currently funded through contract CEECIND/02765/2017 financed by the Foundation for Science and Technology (FCT, Portugal), and this paper was written as part of the FCT-funded projects PTDC/BIA-GEN/29273/2017 and 2022.06259.PTDC (http://www.doi.org/10.54499/2022.06259.PTDC).
52I would like to thank Bonny Sands for comments on a previous version of this paper, and Chris Collins for generously sharing his (then) unpublished manuscript on Kuasi with me.
| 1/2/3 |
1st / 2nd / 3rd person |
| acc |
accusative |
| advz |
adverbializer |
| attr |
attributor |
| ben |
benefactive |
| c |
common gender |
| caus |
causative |
| coll |
collective |
| com |
comitative |
| compl |
completive |
| conj |
conjunction |
| cont |
continuous |
| cop |
copula |
| decl |
declarative |
| dem |
demonstrative |
| dim |
diminutive |
| dir |
directive |
| dist |
distal |
| drm |
discourse reference marker |
| du |
dual |
| dur |
durative |
| emph |
emphasis |
| f |
feminine |
| fut |
future |
| imp |
imperative |
| inst |
instrument |
| inter |
interrogative |
| ipfv |
imperfective |
| j |
juncture |
| loc |
locative |
| m |
masculine |
| mpo |
multi-purpose oblique |
| neg |
negation |
| part |
particle |
| pl |
plural |
| prf |
perfect |
| poss |
possessive |
| prog |
progressive |
| proh |
prohibitive |
| prox |
proximal |
| pst |
past |
| q |
question particle |
| quot |
quotative |
| ref |
referential |
| refl |
reflexive |
| sbj |
subject |
| sbjv |
subjunctive |
| sfx |
suffix |
| sg |
singular |
| sub |
subordination |
| tam |
tense-aspect-mood |
| voc |
vocative |
| vol |
volition |