Kora: A Lost Khoisan Language of the early Cape and the Gariep by Menán du Plessis - 1. The linguistic classification of Kora

Introduction.

This chapter [1] begins within an overview of the Khoisan languages in general, and briefly describes the three or four main families that make up the collectivity of the Southern African Khoisan languages. The chapter concludes –in the electronic version– with short accounts of various conjectured relations between languages of the KHOE family and certain languages of northern or eastern Africa, as well as with with other languages of the southern region, including with other Khoisan languages.

The chapter is organised as follows:

            Divisions and distributions of the Khoisan languages: a general overview.

General characteristics of the JU and TUU families.

            General characteristics of the KHOE family.

            The Kalahari and Khoekhoe branches of KHOE.

The Khoekhoe branch of KHOE:

(i)         West Coast varieties of Khoekhoe (Giri, Nama)

(ii)               Central and south-eastern varieties (Cape Khoekhoe > Kora, eastern Khoekhoe); dialects of Kora.

 Additional discussion appears in the electronic version under the following headings:

Mooted relations between the KHOE languages and languages of northern or eastern Africa.

Relations between the KHOE languages and other Khoisan languages.

Relations between KHOE languages and local languages of the BANTU family.

Relations between the KHOE languages and varieties of Afrikaans.

Divisions and distributions of the Khoisan languages: a general overview.

As we have already briefly noted, the term Khoisan [2] is used by linguists today purely as a blanket term for the non-BANTU (and non-CUSHITIC) ‘click languages’ of Africa, and does not imply the existence of any familial relationships between the member groups. Some scholars include two isolate click languages of Tanzania (Hadza and Sandawe) within the scope of a so-called Macro-Khoisan, although there is little evidence to suggest that these two languages are related even to each other, let alone to any of the southern African languages.

The Khoisan languages of southern Africa are divided by most linguists into three or four separate families. [3] The largest of these is the KHOE family, which includes not only the various Khoekhoe KHOE varieties of Namibia and South Africa, but also the Kalahari KHOE languages of Angola, Namibia, Botswana and Zimbabwe, as shown in Figure 1.1. Divisions shown for the Kalahari branch are based partly on those set out in Rainer Vossen’s comprehensive study [4] of the family, and partly on recent work by Tom Güldemann [5]. They may not be complete or entirely correct, and it is not certain that they are as absolute as the diagram may suggest, particularly in the case of the still under-studied eastern languages. Ts’ixa, for example, has certain characteristics that make it seem intermediate between western and eastern Kalahari. [6] The placement of Kwadi, an extinct Angolan variety, is discussed elsewhere. Some varieties, including the latter, are now extinct.

The unity of the KHOE family is universally accepted by linguists, and barring a few minor sub-divisions that still need to be more firmly established, the information set out in the diagram can be treated as fact. This unity is manifested not simply in the use of typologically similar grammatical strategies by the various members of the family throughout its branches. More specifically, this relatedness is reflected in the use of cognate morphemes for the purpose. Examples include the verb extensions, the particles used to express tense, aspect and modality, and the postpositions. It has also been possible to project systematic reconstructions, on the basis of regular patterns of cross-varietal phonetic alternations, of more than a hundred lexical roots that would have been present in the vocabulary of the hypothetical ancestral language, ‘Proto-Khoe’. (The final section of this chapter will discuss various theories that propose more controversial relationships between the KHOE family and other languages on one hand, or possible substrate influences on early KHOE on the other.)

The map (Figure 1.2) shows past and present distributions of the KHOE languages throughout southern Africa during the modern period of recorded history.

Fig. 1.2. Distribution of languages belonging to the KHOE family, with their range indicated by the black line. Note that the distributions shown here conflate older and present-day data of the modern period. It should be kept in mind also that the early clans, which were probably always fairly limited in size, were highly mobile, and seem to have covered a large range of territory over the course of different seasons. Within South Africa, the seasonal movements of the Khoi clans were facilitated by their possession of pack-oxen and the portability of their dwelling structures. The migrating bands of Khoi were frequently accompanied by small communities of !Ui-speaking San people. (Satellite image from Google Earth.)

 

The term Khoekhoe is simply an abstract label used to identify a sub-set of the KHOE languages that have certain linguistic characteristics in common. It is not assigned to this branch on the basis of any perceived cultural identity of their speakers. It was at one time supposed that the languages that make up the Kalahari branch [7] of the KHOE family on the other hand must be a separate class of ‘Bushman’ languages, on no other grounds than that their speakers were mainly hunter-gatherers! This was the reason for their unfortunate identification by Dorothea Bleek as ‘Central Bushman’, [8] even though the probable relationship between these languages and Khoekhoe languages such as Nama had been pointed out on the basis of comparative data by Samuel Dornan as early as 1917. [9] Further comparative evidence in support of the obvious connection was later put forward by Maingard, [10] and with Rainer Vossen’s presentation [11] of a set of reconstructed roots for Proto-Khoe –following a previous set of postulations from Kenneth Baucom– [12] the matter was settled conclusively.

Since the KHOE family shows the greatest overall dialectal proliferation and diversity, it is probable that it is the oldest of the Khoisan language groups. While there is no reliable way of dating a language family in the absence of written records, we can nevertheless form a (very) rough approximation using the evidence we have for other well-documented language families of a known age from other parts of the world, such as the Germanic family. [13] On this basis, the KHOE family looks as though it may be around 2000 years old, although this is probably a generous estimate.

The other Khoisan language groups of southern Africa are shown in Figure 1.3. (Note that linguists disagree on the question of relationships between them.) [14] The list of dialects shown in the diagram is not exhaustive, and internal divisions of some groups are still debated. Several names are not strictly glossonyms (names for languages) but merely endogenous (own) or exogenous (others’) names for speaker communities. Most of the !Ui languages and several belonging to the Taa group are now extinct. ?Khomani, which Westphal identified as N|huki, appears to have been closely related to if not identical with the varieties of N|uu (or N?ng) that have been documented by scholars working over the past decade and a half, following the re-discovery of a few elderly speakers in the 1990s. [15]

The two groups identified as JU and TUU are in many ways quite different in terms of their morphology and syntax – yet they nevertheless have a few broad typological features in common, as will be outlined later below. The JU languages, which were given this name by Ernst Westphal, [16] were originally labelled ‘Northern Bushman’ by Dorothea Bleek, as a consequence of which some linguists still refer to the group as ‘Northern Khoisan’. The !Ui (or ‘!Kwi’ in the original spelling) and Taa language groups were also assigned their names by Westphal, having been previously grouped together by Dorothea Bleek as ‘Southern Bushman’. The alternative label TUU for the !Ui-Taa or ‘Southern Khoisan’ group has been suggested by Tom Güldemann. [17] The map (Figure 1.4) gives an approximate indication of the rather limited distributions of the two families.

Fig. 1.4. Distribution of languages belonging to the JU (Northern Khoisan) and Taa (Southern Khoisan) groups. Within these only very approximately delimited ranges, speakers tend to be found in small and dispersed villages, often close to water, and in among the settlements of other communities, so that the boundary lines should not be taken to indicate either exclusive or widespread distribution. The location of the few villages where speakers of ?’Amkoe may still be found is adjacent to and just slightly north-east of the Taa range. For various reasons, as discussed in the text, it is difficult to give an accurate picture of the former distribution of the !Ui languages (also Southern Khoisan) of South Africa. (The dispersed places within South Africa where brief records were made during the twentieth century probably reflected the last stages of movement on the part of very small surviving groups or even individuals.) (Satellite image from Google Earth.)

It is not easy to give an accurate picture of the former range of the South African !Ui languages. The so-called ‘Bushmen’ of the Cape tended to live in small but mobile bands. They may have covered an extensive territory as they moved after the changing vegetation of the different seasons, and the associated migrations of game animals. Their small groups were originally often closely associated with individual clans of the Khoi, whom they may have had no choice but to accompany on their own migrations in quest of grazing and water for their herds in addition to veldkos and game. [18] They would have done so in much the same way as some of the serf classes in the Kalahari, including speakers of Kalahari KHOE varieties, Taa dialects, and the Sotho-Tswana-related language Kgalagadi, were formerly compelled to accompany their Tswana (and Kalanga) masters wherever they travelled with their livestock. [19] This social factor may have played a significant role in certain aspects of the region’s dynamics, and was almost certainly more pertinent from a linguistic point of view than economic culture. [20] Historical evidence concerning other societies and social strata throughout much of older Africa tells us that such groups were sometimes required to modify aspects of their speech in order to show deference.

It is also a matter of stark historical record that speakers of the !Ui dialects were subjected to a ruthless targeting throughout much of the 18th and early 19th centuries, being hunted down and massacred in their hundreds by commando groups, particularly on the margins of the outlying districts of the slowly but steadily expanding settlements. This makes it likely that at least some varieties of !Ui –which famously included the |Xam language documented by Wilhelm Bleek and his sister-in-law Lucy Lloyd– would have perished along with their speakers, without ever having been documented. Nevertheless, the women and children were usually spared in these raids and taken instead to live on farms, where subsequent marriages to Khoekhoe-speaking (and by now also Dutch-speaking) farmworkers were not uncommon, and where !Ui varieties were occasionally preserved in the context of the home. (Two of the speakers who worked with Bleek and Lloyd in the later part of the 19th century spoke Kora in addition to |Xam.)[21]

A few surviving communities seem to have formed alliances with some of the remaining independent Khoi clans, while speakers of the N|uu dialects once found north of the Gariep appear to have moved still further north, into the Kalahari. Some of the latter people, then identified as ?Khomani, were among the groups who were assembled at Bain’s Camp in 1936 to coincide with the Empire Exhibition in Johannesburg, where various linguists had the opportunity to work with them and obtain recordings of their language.[22] (When human rights advocates and linguists began working with the surviving ?Khomani San in the late 1990s and early 2000s, they discovered that some of the people in the community were descendants of people who had attended this camp.)[23] The accumulated evidence from all of the varieties that survived long enough to be documented (even if only fragmentarily) suggests that the dialectal differences across the !Ui group were not extensive, while it is known that even the differences between the varieties spoken north and south of the Gariep were not so great as to prevent mutual intelligibility.[24]

Across the JU spectrum,[25] the various dialects may be grouped into three or four regional clusters,[26] with Ju|’hoan being the southernmost member, and the least differentiated.[27] While the dialects are all fairly closely related, this does not automatically mean that speakers of Ju|’hoan, for example, would readily or even at all understand someone speaking one of the dialects from northern Namibian (and north-western Botswana) or the southern Angolan dialects of !Xun.[28]

A few linguists[29] believe that the previously unplaced language ?’Amkoe, formerly known by the name of one of its dialects as Eastern ?Hoan,[30] is related to JU, and propose to call the unified grouping KX’A (or Ju-?Hoan). The linkage is still open to debate, however, given the great typological disparity between ?’Amkoe and the JU languages, where the extent of these differences makes it puzzling that a sub-component of the lexis should at the same time be so systematically and self-evidently similar to equivalent words in JU languages, while another sub-component should so closely resemble equivalent words in Taa varieties such as !Xoon, even to the extent of featuring bilabial clicks.

The JU and TUU families are each far less internally branched than the KHOE family, and so are in all probability considerably younger than the latter, at least if they are treated as two independent and unrelated groups – which is the approach currently taken by most Khoisanists.[31] It is a widespread popular and sometimes even academic assumption that the JU and TUU (or !Ui-Taa) groups must be ‘ancient’. However, this is not greatly plausible from a linguistic point of view, as just noted, and the idea (which is in any case remarkably vague) seems to have its roots in little more than a belief in absolute cultural chronology, superimposed on a colonial fallacy of socio-economic or even biological essentialism.[32]

As far as the modern surviving Khoisan languages are concerned, it is difficult to form a reliable estimate of speaker numbers. Many speakers live in outlying rural villages far from any major city centre.  Furthermore, many of these communities have a cross-border distribution, spanning two or more countries, where approaches to census-taking are not always uniform. The only relatively robust language is found in Namibia, where varieties of Namibian Khoekhoe (which includes dialects of Nama, Dama and Hai?Om) are estimated to be spoken today by as many as 200,000 people[33] (out of a total population of approximately 2.1 million)[34]. Namibian Khoekhoe, which may be taken as a subject both at school and university level, is spoken across all generations, and is used in a wide variety of situations and contexts. Most speakers are bilingual, typically speaking also English or Afrikaans, if not both.[35] Other Khoisan languages still spoken in Namibia (but sometimes also in neighbouring Botswana) include dialects of !Xun and Ju|’hoan (JU family), and dialects of the Taa language !Xoon (TUU family). It has recently been estimated[36] that the total number of people speaking JU dialects across Angola, Namibia and Botswana is around 16,000. Remaining speakers of Taa are believed to be about 2,600.[37]

To the north of Namibia in south-eastern Angola, Khoisan languages once found included various !Xun dialects belonging to the JU family, as well as Khwe dialects belonging to the western division of the Kalahari KHOE languages (and found also in the Caprivi strip). Just under 4,000 speakers of !Xun were relocated to South Africa following the end of the liberation struggles in the region, along with speakers of Khwe dialects.

Botswana is home to the majority of the Kalahari KHOE languages, in addition to some of the JU and TUU languages already mentioned, as well as Eastern ?Hoan (now more often referred to as ?’Amkoe), which has perhaps fewer than 30 speakers left.[38] The total number of people still speaking languages belonging to the Kalahari branch of KHOE is around 25,000, based on estimates given by Brenzinger[39] for Naro (10,000), ?Gana-|Gui (2,500), Shua and Tshwa (4,100) ), and Khwe-?Ani (8,000).

In the south-western part of Zimbabwe, perhaps a dozen elders from a small cluster of village-based communities at Tsholotsho still speak varieties of eastern Kalahari KHOE such as Ts’ixa and Tjwao[40]. Most of the Tjwa people now speak Ndebele or Kalanga.

Of the Khoisan languages once spoken in South Africa, the only one still moderately viable (though fragile) is Nama. Brenzinger reports[41] that the number of speakers remaining may now be less than 2000, where most are older than fifty. Most members of the South African Nama communities in the Northern Cape now speak Afrikaans. As for the Khoekhoe language we are concerned with in this book, namely Kora, there were at the time of writing only two known speakers left, and virtually all the present day descendants of the Korana people now speak Afrikaans, English, Sotho or Tswana. The last remaining member of the !Ui group is N|uu, as noted earlier, where the number of elders who still speak it is dwindling with each year that passes, so that at the time of writing there were only three left. Members of the ?Khomani San community for the most part now speak Nama and Afrikaans or Tswana. Although some speakers of Khwe (Kalahari KHOE) and !Xun (JU) dialects have recently come to be resident in South Africa, as noted earlier, the languages in question are not indigenous to the country.

General characteristics of the JU and TUU languages.

As we have already mentioned, the Khoisan language families of southern Africa other than KHOE have a number of broad typological properties in common. Most notably, both the JU and TUU (!Ui-Taa) groups have systems of noun-classification that are reflected in a set of multiple grammatical genders, where these are similar to those of most Niger-Congo languages. They do not make reference to any semantic category of natural gender such as masculine or feminine, but are based rather on properties such as animacy, edibility, utility, or shape. The !Ui languages differ from those of the Taa and JU groups in having only two genders, as was first noted by Wilhelm Bleek.[42] (The !Ui genders seem to be based on features primarily of animacy and inanimacy.) The more complex gender system of Ju|’hoan was established only in the late 1960s by Ernst Westphal and Jan Snyman,[43] and it was no doubt an awareness of the situation in the JU languages that motivated Tony Traill to search for and establish the equally complex gender system of the !Xoon language (Taa group). The isolate Eastern ?Hoan (now often referred to as ?’Amkoe) is reported to have no genders.

The genders in both families are not indexed by any overt morphology, but are mostly visible only in the selection of pronouns. Even so, languages of the !Ui and Taa groups nevertheless make some limited use of gender-indexical noun suffixes, while both JU and TUU groups have varieties where a few nouns still carry prefixes.[44] Although some authors prefer to assign them numbers, the genders can be referred to most simply in terms of their paired singular-plural pronouns. It is notable –given that the JU and TUU groups are not generally thought to be related– that nouns assigned to the pronominal gender ka/ka in JU varieties frequently have counterparts (with an overall similarity in form and meaning) in the ha/ha gender of !Xoon (which is a TUU language).

Languages of both the JU and TUU families also favour a verb-second sentence pattern, which places the verb after the subject and before any object, to give the pattern Subject-Verb-Object or SVO, although different orderings may occur in subordinate clauses. Languages belonging to both families use a few basic morphemes plus a range of verbal auxiliaries to express some types of negation; as well as tense, aspect and modality; and to impart directional implications. Verbal compounds (or ‘complex predicates’) and verbs with grammaticalised extensions are found in both JU and TUU groups (and also in KHOE languages). While a few basic (or ‘true’) adjectives are found, most descriptive specifiers are either derived from nouns or else are verbal forms used in association with a relative construction, much as is the case also in the BANTU languages. 

Another feature common to the languages of both groups is the use of suppletion. This means that an entirely different morpheme –not merely an inflected form– is introduced to complete certain parts of a paradigm. In English, for example, we use ‘went’ to complete the past tense of ‘go’. This feature is perhaps most widely seen in languages of the JU family, where it may be manifested in the use of two different words to form the singular and plural forms of a given referring expression, as well as the use of different words to express the same predicate, depending on whether the subject is singular or plural, and whether the verb in question is used transitively.

The syntax of the various JU languages has been fairly thoroughly described in a number of works over the past few decades. The syntactic structures of the !Ui and Taa languages, on the other hand, are only just beginning to be described in detail, and it is possible that further commonalities will be discovered as this work proceeds.

With regard to their phonetic inventories, languages of both the JU and TUU families reflect a greater range of contrastive vowel colourations than the KHOE languages – that is, in addition to the use of semantically significant nasalisation, which is a feature common to all Khoisan languages. The additional vowel qualities, which may also combine with one another, include pharyngealisation, breathy-voicing and glottalisation, although the !Ui languages seem to have featured only pharyngealisation.[45] One feature that distinguishes the TUU languages (and ?’Amkoe) from JU is the use of the bilabial click (?), which typically occurs, however, in only a small set of words.

General characteristics of the KHOE family.

The KHOE family is set sharply and quite unmistakeably apart from the JU and TUU families by its gender system. Much like the systems of Afroasiatic and Indo-European languages, the KHOE system divides nouns into categories that line up with the distinction between masculine and feminine in the case of animate referents. A third category is available for neutral or indeterminate reference. In languages of the Khoekhoe branch, these grammatical genders are overtly indexed by means of suffixes that mark the nouns as masculine or feminine.

It is often noted that the KHOE languages are also distinguished typologically by a general preference for a verb-final order (SOV) in the sentence, where the verb is placed after the subject (S) and any object (O). While this is true in principle, overall ordering of constituents in the KHOE languages is in reality highly flexible, and seems to be driven primarily by pragmatic considerations of focus and topic. As we will see later, it may even be possible to analyse Khoekhoe languages using such typological terms as ‘topic prominent’, much like Mandarin, Cantonese, Japanese and Malay. As is typically the case in languages that place the verb after the subject and any object, the adpositions in KHOE languages pattern in a parallel way, and are placed after the noun. For this reason, they are frequently referred to as postpositions (rather than prepositions).

The Kalahari and Khoekhoe branches of KHOE.

The differences between the languages belonging to the Kalahari and Khoekhoe branches of the KHOE family are not entirely well-defined.[46] One rather broad difference, though, is the greater overall diversity seen in the Kalahari grouping. The western sub-groups constituted by varieties of Khwe[47], Naro and ?Gana[48] may differ from one another in various aspects of their morphology and syntax, particularly in the expression of tense and aspect. These western varieties differ in turn from eastern subgroups such as varieties of Shua[49] and Tshwa in a number of respects, with the latter being distinguished amongst other things by the reduced number of clicks in their consonant inventories –and in particular the rarity of (post)alveolar (!) and palato-alveolar (?) clicks in eastern varieties– and some differences in their morphology.

The following are some of the specific respects in which the Kalahari varieties differ from Khoekhoe.

,

·         Most Kalahari languages do not make a distinction between inclusive and exclusive reference when using pronouns of the first and second persons (for example, to express ‘we-all’ versus ‘just us, not you’). Such a distinction is, however, made in Khoekhoe languages.

·         The Kalahari varieties tend to build their absolute pronouns on a base such as xa or ha in western dialects, and e in eastern dialects , whereas the Khoekhoe varieties use  ?’ãi (~ ?’i). 

·         Kalahari varieties frequently omit the suffixes that express grammatical gender, and reflect gender only indirectly through the dependent pronominal forms.

·         It is more common and sometimes the norm in Kalahari varieties for the masculine singular suffix to be expressed as –m(i) rather than the characteristic –b(i) of Khoekhoe languages.

·         Some Kalahari varieties express the 3rd person masculine plural (‘they [male]’ by means of a sufffix such as –?u, or in some varieties –dzi or –re, whereas Khoekhoe languages use –ku.

·         Kalahari languages typically use a linker (sometimes termed a ‘juncture’) between the verb and certain other grammatical elements.[50] Khoekhoe languages, on the other hand, do not have this feature.

Interestingly, Kora has preserved a number of features that are absent from Nama, yet which occur in the Kalahari languages. Examples include the occurrence in Kora of an ejective affricate both as a phoneme and (in some dialects) as a click accompaniment,[51] a few aspects of its morphology, such as the use of an accompanitive verb extension -|xoa, and various items of vocabulary. There is a sense in which Kora almost appears to be the ‘missing link’ between Kalahari and Khoekhoe varieties of KHOE.

The Khoekhoe branch of KHOE.

The Namibian varieties of Khoekhoe include Nama, which is spoken in the south of Namibia, and various dialects spoken in the north of the country by the Damara people[52] and the Hai?Om. The differences between the varieties are mainly phonetic in character, although some minor differences in morphology and vocabulary are also found.[53]

The original South African varieties of Khoekhoe, insofar as we have records of them, can be divided very broadly into two groups, consisting of:

(i)                 a small or at least not greatly differentiated set of dialects with a distribution along the West Coast, starting perhaps near the Olifants River, and extending as far north as Namaqualand; and

(ii)               a more extensive set of dialects with a central and south-eastern distribution extending from Table Bay and Saldanha Bay into the interior of the country up to and along the Gariep as well as beyond the Vaal River; and also in a more easterly direction along the south coast, as far as the Kei River (if not beyond).

(i) West Coast varieties of Khoekhoe (Giri, Nama)

The early West Coast varieties were spoken by communities such as the Chariguriqua (which may have meant the Little Guriqua), the Grigriqua (perhaps Garigurikua or Gurigurikua, later Griqua or Griekwa), and the Amaqua (!Amaqua). These dialects seem to have had close affinities with the varieties of Nama spoken in the northern reaches of the West Coast (or Little Namaqualand), and in the southern parts of the country known today as Namibia (formerly Great Namaqualand). As the Dutch settlers at the Cape over the course of a century and a half claimed more and more of the Khoi herders’ traditional grazing veld, some of the clans from the West Coast began to change their migration patterns and moved further north, or even crossed the Gariep, while others moved inland. Various suggested migration routes are set out in map form in Nienaber’s comprehensive study of the Khoi clan names.[54]

While occasional deadly outbreaks of smallpox at the Cape are known to have had a particularly devastating impact on the vulnerable local populations, the Khoi were certainly not entirely wiped out by the disease. In some cases, small groups accepted employment on their farms of the slowly advancing settlers, for example as herders of livestock and wagon drivers, where they rapidly began to acquire Cape Dutch. Many others moved away from the shifting frontiers of the Cape, while from the early1800s onwards, some made the choice to settle permanently in the vicinity of mission stations, both in Namaqualand (in the far north-western sector of the Cape) and in the interior of South Africa. Here they typically became bilingual – learning to speak, read and write Dutch (and in some cases, English) in addition to sustaining their own Khoekhoe variety, even if the latter was perhaps increasingly used only in the private setting of the home. Those Khoi of the West Coast who moved inland to mission stations such as Klaarwater (originally !Ari?amma,[55] but subsequently renamed Griquatown) co-existed with some of the Korana clans as well as people from diverse other communities.[56] Other mission stations of the interior included those of the Berlin Missionary Society at Bethany and Pniel.

In the remote and arid region of Namaqualand, the mission stations eventually became the centres around which a number of ‘reserves’ were formed (Richtersveld, Steinkopf, Leliefontein, Komaggas and Concordia). Although traditional matjieshuis structures could still occasionally be seen among more conventional modern buildings as recently as the early 1990s, the Nama language by this time was in decline, having been widely replaced by Afrikaans.[57] Since the ending of apartheid, attempts have been made to regenerate the language by introducing it to children in pilot programmes at a few of the local schools, and at the time of writing, plans were being announced to introduce the language more formally into the curriculum of schools in the region.

We have very few records for the older varieties of the West Coast, but it turns out that some members of the Links family interviewed by Lucy Lloyd in 1879 were Griqua rather than Korana. In particular, the small amount of material obtained from Siela (Cela) is recognisably different from Kora, and seems to represent a variety of Giri. While the speech of Piet Links himself featured a number of unmistakeable Kora characteristics, including the presence of the ejective affricate, there are various instances in the narratives he dictated where a more western and Giri-like influence occasionally manifests itself, not only in the morphology and lexis, but also in the syntax. (It is conceivable that the members of his family would have considered him to speak Giri with a strong ‘Korana accent’.)

By the 1920s and 30s, there were very few speakers of Giri left. The phonetician Douglas Beach who worked in the field at this time was able to provide only a short paragraph of general observations concerning phonetic characteristics of the Griqua variety,[58] although Meinhof contributed a short illustrative vocabulary,[59] having obtained some limited information from two or three speakers who visited the mission station at Pniel where he was staying in 1928. 

Perhaps the most lasting record of the dialects of the early clans of the West Coast and Northern Cape is to be found in local place names, such as Garies, Komaggas and Nababeep, to mention only a few. A number of sources have been suggested for the name Garies,[60] including !arib ‘ridge, long low mountain’ (which has also been proposed as the origin of the name Gariep, since the river flows through steep rocky gorges at certain points along its course). It may simply arise, however, from !harib ‘town’. (Some of the elderly N|uu speakers among the ?Khomani San, who have a high proportion of Khoekhoe words in their speech, initially gave the word g!ari as the specific name for Upington, but later revised its meaning to the more general ‘town, city’.)[61] Various meanings have likewise been proposed for Komaggas, including ‘olive-rich’ (!’ummaxas), while the name Nababeep is open to a number of interpretations, including one that involves an incorporation of !nabab ‘rhinoceros’.

Many more place names of the present day West Coast and Northern Cape are recognisably Khoekhoe, even though it is often difficult to work out what the exact forms of the originals were, or what their meanings would have been. It is known, at least, that numerous old names incorporate references to water, as Nienaber and Raper have pointed out,[62] where these may take the form of words meaning ‘artery’, ‘stream’, ‘river’, ‘ford’, ‘bend’, ‘waterfall’, ‘inlet’, ‘confluence’, ‘mouth’, ‘pan’, ‘vlei’, ‘spring’ or ‘well’. The nomadic Khoi also distinguished and named various types of terrain, climatic region and geological substrate –probably because of the different vegetation types and animals associated with them– while many more place names directly incorporate the names of plants or animals.[63] The Kareeberg Mountains may take their name, for example, from local taaibos or Rhus species (Nama !areb) –that is, if the name is not based on the word for ‘ridge’– while it is likely that original Khoekhoe names were the source of river names such as the Olifants and the Buffels.

(ii) Central and south-eastern varieties (Cape Khoekhoe > Kora, eastern Khoekhoe)

There is general consensus[64] that the Khoi clans once based in the vicinity of Table Bay and, surrounding areas, such as the Goringhaiqua, the Goringhaikona, the Gorachouqua and possibly also the Cochoqua (or Saldanhars) of Van Riebeeck’s journals were the forebears of clans such as the Hoogstanders or High-standers (the !Urimã?’ais or !Uri?’aikua), the Korana (!Ora?’ais), and the Smalwange or Narrow Cheeks people (?’Oxokua). These are among the clans who were met with again in the 1770s, by observers such as Hendrik Wikar (who mentions the Kouringas or ‘High Kraal’)[65] and Robert Gordon. Travellers like these last two encountered the dispersed clans not only in the interior and along the middle and upper stretches of the Gariep, but even beyond the Gariep and the Vaal.[66]

It is difficult to form an accurate estimate of the original numbers of the Khoi at the Cape, particularly since most of the clans seem to have visited Table Bay and the surrounding areas only at certain times of the year. In 1655, however, a party of the Company’s hunters returned to the Dutch Fort with a report that they had come across a place where ‘there are fully 1000 huts put up.’[67] Van Riebeeck himself rode out to see this massed assembly of the Saldanhars (Cochoquas), and the whole incident was subsequently recorded as follows:

Coming near and beyond the Redoubt Duynhoop, we found the country everywhere so full of cattle and sheep, as far as the wood, where our people lie, fully 3 mylen from this, and fully ½ myl broad, that we could hardly get along the road, and the cattle required to be constantly driven out of our way by the Hottentoos, otherwise it seemed impossible to get through; not only were the numbers of cattle impossible to be counted, but the same might be said of the number of herds of cattle; and it was just the same with the people, of whom we could see at one look around us, probably 5000 or 6000, young and old, for their curiosity to see was such that we were so enclosed by them, that we could scarcely see over them from horseback; there were also about 4 to 500 houses, rather large, and pitched in circles close to each other, within which the cattle are kept at night, the circles could scarcely be walked round in a half hour, and looked like regular camps. The Caepmans [Goringhaiquas] were also there with their houses and cattle.[68] 

The direct link between the Cape Khoi who regularly visited Table Bay and the Korana is attested in the first place by historical records, but is also confirmed by linguistic evidence –fragmentary and inevitably imprecise as this may be. A comprehensive list of the early records of Cape Khoekhoe has been compiled by Gideon Nienaber,[69] whose indispensable reference work also contains a near exhaustive collation of comparative sources for each instance of an old Khoekhoe word encountered in the early documents, indexed by its Afrikaans translation equivalent. (Some of these sources are described in more detail in the following chapter.)

For the eastern varieties of Cape Khoekhoe, only a few brief records have come down to us from people who travelled during the late 18th century to the outer regions of the slowly expanding settlement, along both coasts and as far afield as the Gariep in the north and the Great Fish River in the east. These travellers include two Swedish naturalists – Anders Sparrman, who travelled in the Cape between 1772 and 1776, and Carl Peter Thunberg, who travelled independently of Sparrman, between 1772 and 1775. Both contributed valuable if fragmentary examples of the local languages,[70] while Sparrman’s data included a list of words and phrases illustrating not only the variety spoken by the Houteniqua Khoi, but also an early example of Xhosa.[71] The German explorer Franz von Winkelmann managed to record a list of eastern Khoekhoe (and Xhosa) words in 1788–1789.[72] Sadly, though, no copy has ever been found of the catechism reported by Lichtenstein[73] to have been written in Gona and ‘printed with his own hands, in the year 1806’ by Johannes Theodorus van der Kemp of the London Missionary Society’s station at Bethelsdorp.)

Ultimately, and much as in the case of the western clans, perhaps the most enduring aspect of the eastern Khoi legacy is to be found in local place names. The name for the Sneeuberg mountains, for example, translates the original Khoekhoe name recorded by Gordon on one of his maps as ‘Noa Gore’[74] (perhaps n!hoa xodi, although it must be said that the Kora word for ‘snow’ was recorded only as either !xoab or !oab, while Xhosa has iliqhwa ‘sleet’, and Zulu has iqhwa ‘ice, frost’). Many rivers, mountains and passes of the southern and south-eastern Cape still preserve their original Khoekhoe names, at least in adapted versions such as Kariega, Gourits, Gamtoos, Kei, Keiskamma, Tsitsikamma and Kareedouw while many more bear the translated Dutch equivalents of their older names, such as the Vogel River, for which Gordon recorded the original name Kaniga (kx’anis ‘bird’), and the Buffalo River (which on early maps often bore the alternative name Kouka or Kaugga, possibly from gaob ‘wildebeest’ rather than |aob ‘buffalo’).

When all of this early lexical evidence is collated and compared, it is clear that Kora was close to Cape Khoekhoe, and that it was far more so than Giri or Nama. As Maingard once put it, ‘the similarity of the Korana and the Cape dialects […] in possessing together certain phonetic peculiarities and common words […] definitely constitute these two as a homogeneous linguistic group as against the Nama dialect.’[75] In addition, as Nienaber has noted, it is apparent that ‘the eastern dialects have a close affinity with the extinct Cape and Kora forms, in contrast with Nama’.[76] Examples of the kinds of evidence that show the close relationship between Cape Khoekhoe, Kora and eastern Khoekhoe are provided in Figure 1.5. 

Dialects of Kora.

Just as in the case of any other language, the entity we are referring to as Kora consisted of a number of different dialects.[77] Maingard[78] studied a few dialects still spoken in the 1930s that he believed were plausibly representative of the groupings identified by Wikar[79] in the late 18th century,[80] and found that various linguistic features seemed to corroborate the latter’s early division of the Korana, with certain characteristics of the more western groups making them appear closer to Nama. (Another way of interpreting this data might be in terms of the dispersed groups of the former West Coast clans on one hand, and the clans of Table Bay and the interior on the other.) Engelbrecht[81] in turn compared aspects of the Khoekhoe dialects spoken by the Lukas people and the Karoshebbers (or Karosdraers) on one hand, as against varieties spoken on the other by the Links, Kats[82] and Kraalshoek people. He similarly concluded that the varieties spoken by the first group were closer to Nama.

The existence of what may have been a further minor dialect within Kora, not previously recognised as such but suggested by records made independently by Lichtenstein,[83] Burchell[84] and Wuras,[85] has come to light during the course of the present study. The most salient feature of this variety was a more frequent use of –m as opposed to –b for the masculine singular suffix. These cases seem to have occurred in words that contained a nasalised vowel, and probably developed out of an assimilation involving the intrusive nasal segment that could appear after such a vowel in certain varieties of Kora, and the masculine suffix. (The same process was probably responsible for the variant form Tsuni?goam sometimes seen for the name of the mythological being, Tsui?goab.) The original nasalisation of the vowel occasionally seems to have disappeared subsequently, as seen in some of the examples below.

                                                Meinhof          Wuras              Lichtenstein

            eye                               mub                 mumb              muhm

            thigh                            tib                    -                       tiim

wild cat                       |hõab               |hoãm              -

calf, steer                     |nõab               |noãm              t’1nom

tooth                            ?ub                  ?um                 t’1kuhm

ear, leaf                       ?nãub              ?naum             t’1naum

moon, month               ?xãb                 ?xam                t’2kaam

hand, arm                    ?’õab               -                       t’2koam

hunger                         !’ãb                  !am                  -

brother                         !’ãb                  !kãm                t’2kaam

The speech of our consultant Ouma Jacoba Maclear reflects a similar tendency, in her pronunciation, for example, of |ãb ‘grass’ as |ãm.

The remaining sections of this chapter, which are found in the electronic version, will provide brief discussion of a range of theories and conjectures about relationships between the KHOE languages and various other African languages, beginning with an account of longstanding proposals for a connection between the KHOE languages (or in some cases just the Khoekhoe branch of the family) and one or another language or language family from further north or in the east of Africa.

Hypotheses concerning relationships between languages of the KHOE family and various other languages of Africa.

(i)                 Mooted relations between the KHOE languages and languages of northern or eastern Africa.

Before he came to South Africa in 1855, Wilhelm Bleek had already presented his doctoral thesis[86] on ‘sex-denoting languages’, where he proposed a division of languages based on their use (or not) of gender systems where the noun classifications are aligned with the categories of natural gender. In terms of this framework, and drawing on the very scant sources of information then available for Nama, he proposed a connection between the Khoekhoe language Nama, and not only various languages such as Ancient Egyptian and Galla that would be classified today as part of Afroasiatic, but also Indo-European languages, which at the time were referred to as Indo-Germanic. This idea, insofar as it suggested a connection between Nama (as the stand-in for Khoekhoe languages) and other languages of north-eastern Africa, was further developed by later scholars, and finally found its way (in a modified form, and minus the proposed link with Indo-European) into the Hamitic hypothesis of Meinhof, which appeared in 1912.[87]

When Otto Dempwolff a few years later in 1916 published his extensive study[88] (which includes texts) of a newly-found click language of East Africa, Sandawe, he suggested that it too belonged to the supposed Hamitic group,[89] and even offered a short comparative list of words for Nama and Sandawe, where he claimed various vague and semantically only tenuously connected similarities. (There should hardly be any need to add that the mere presence of clicks in languages that are otherwise utterly different, not to mention separated by a vast geographic distance, is not enough to ‘prove’ a relationship between those languages. It has recently been proposed by two linguists working independently of one another that, given the right combination of co-articulatory events, clicks have the potential to emerge,[90] while a recent case of click emergence in exactly the predicted environment has even been documented.[91] In principle then, it is possible that clicks may have arisen separately on more than one occasion, in different parts of the continent.)

There is another click language spoken in the country now known as Tanzania, namely Hadza.[92] This isolate language does not appear, however, to have any connection either with the nearby Sandawe[93] or with any of the Khoisan languages of southern Africa. (The location of Sandawe and Hadza is shown in Figure 1.6.) As hardly needs stating, the mere fact that the language contains clicks is not enough to point to an actual relationship with any other languages that make use of similar sounds; while the fact that its speakers have (no doubt through force of social circumstance) largely preserved a form of the hunter-gatherer lifestyle once common to all humanity –not only throughout Africa but the entire world– is entirely irrelevant. Dorothea Bleek nevertheless included her own data for this language in her Bushman Dictionary, labelling it CIII – where the implied linkage with so-called Central Bushman (in fact Kalahari KHOE) languages, such as Hie Tshware (CI) and Naro (CII) could not have been based on much more than the fact that all three languages have a gender system that includes categories of ‘masculine’ and ‘feminine’.

The idea of a Tanzanian connection for languages belonging to the KHOE family of southern Africa remains alive, and has recently been revisited by Tom Güldemann, who has proposed in a succession of papers that Sandawe might be connected to the southern languages[94] via an intermediate higher-order grouping based on KHOE and the extinct Angolan variety Kwadi.[95] Güldemann suggests that the speakers of his mooted ancestral Khoe-Kwadi language were relatively recent immigrants, whose arrival nevertheless pre-dated that of any groups of people speaking BANTU languages,[96] and that it was this hypothetical group of people[97] –rather than merely the Khoi (or ‘Khoekhoen’), as previously suggested by others– who were responsible for bringing the first sheep into the region.

Is the scenario of a north-eastern connection for the KHOE languages plausible? We will leave it to readers to make up their own minds, but would point out that the linguistic evidence for a higher-level common ancestor (Khoe-Kwadi) is not compelling, since no-one has been able to present a systematic and comprehensive set of arrays showing regularly repeated phonetic correspondences across shared words with plausible semantic linkages, as opposed to merely a few isolated instances of similar-looking words. It is possible that Kwadi was simply a kind of auxiliary code or inner language, given that it was spoken only among themselves by a closed and very small circle of older men within a community whose members otherwise spoke an ordinary Kwanyama-like BANTU language known as Kwanyoka. There is similarly no conclusive linguistic evidence to support the idea of a familial connection between Sandawe and the KHOE languages.

Fig. 1.6. The location of the Hadza and Sandawe languages in Tanzania, and of Kwadi in the south-west of Angola. (Satellite image from Google Earth.)

Finally, Güldemann hypothesises that contact with supposed substrates of pre-existing Khoisan languages belonging to the TUU and JU families, in the area of the Cape on one hand,[98] and the Kalahari on the other,[99] may have contributed to certain specific properties of the modern Khoekhoe and Kalahari varieties of KHOE.[100] Most recently he has begun collaborating with biologists in an attempt to confirm these theories of contact-induced cross-influence (diffusion) through genetic studies.

Is the scenario of local areal influence plausible? It is undeniably true that there must have been significant and sustained contact between speakers of KHOE languages, and speakers of other Khoisan languages in southern Africa. We have already noted, however, that the linguistic evidence does not support the idea of any great age for either of the JU or TUU language families, at least when they are considered separately as two distinct and unrelated entities, while the social circumstances mentioned earlier in this chapter would probably not have been conducive to a shift in the proposed direction. In the overwhelming majority of cases –where the few exceptions are languages of the Kalahari region– it is clear that the direction of any influence has been from KHOE sources into the other languages.[101] Our knowledge of the eastern Kalahari KHOE languages has expanded steadily over the past few years, while the present work in turn contributes to a better knowledge of the Khoekhoe KHOE spectrum. In light of what we know now, it has become increasingly doubtful that there are any properties of either Kalahari or Khoekhoe KHOE languages so strikingly anomalous as to warrant special explanation in terms of diffusion.

(ii)               Relations between the KHOE languages and other Khoisan languages.

As just noted, there has been an extensive influence of KHOE languages on other Khoisan languages of southern Africa. In addition to lexical borrowings, there has also in some cases been an uptake of morphology, and it is rather astonishing to find that the Ju|’hoan paradigms of tense and aspect share almost all their morphology with Khoekhoe. The JU languages also use some of the verb extensions that have been reconstructed for Proto-Khoe, such as the accompanitive *-|xoa. While this kind of structural borrowing is by no means unheard of, it is certainly less common than lexical borrowing.

In the case of the !Ui-Taa (or TUU) languages, borrowings are sometimes localised and present in only one or two varieties, so that they probably reflect the relatively recent kinds of contact that occurred, for example, when surviving members of San groups were taken in by Khoi communities. In other cases, though, and more significantly, loanwords from a Khoekhoe source are so widely and systematically present throughout all the known varieties that the borrowing must have occurred at an early stage, prior to any dispersal of the speakers and the dialectal diversification of the family.

While the presence of Khoekhoe loanwords in |Xam has been noted in the past,[102] the pervasive presence of such loanwords throughout the !Ui languages becomes even more apparent when the vocabulary of Kora and Cape Khoekhoe is taken into account. Indeed, most of the Khoekhoe loanwords in both |Xam and N|uu varieties appear to have come quite specifically from Kora rather than Giri or Nama, as is apparent from certain distinctive phonetic properties of the words in question. There is something about this overall picture of wholesale borrowing that seems a little unusual, and the topic might be a fruitful area of research for future scholars willing to investigate the subject from a fresh perspective of relative social status and power relations, rather than race.[103]

(iii)             Relations between KHOE languages and local languages of the BANTU family.

The apparent influence of Khoisan languages on the Nguni languages of South Africa has long been the subject of discussion, and was the subject of a detailed study by Meinhof,[104] who attributed not only the clicks but also certain other sounds in Xhosa (such as the ejective affricate) to a Khoekhoe source, since they were not the expected reflexes for Xhosa of the sounds reconstructed for Proto-BANTU.[105] Other authors have since explored the topic in detail, and it is generally now believed that the small set of post-alveolar clicks found in Sotho were probably derived secondarily from one of the Nguni languages, rather than directly from a Khoisan source. The dental clicks found in the tekela Nguni language Swati may also have been obtained indirectly.[106] An additional feature of the Nguni languages nowadays often attributed to the influence of Khoisan languages is the phenomenon of tonal depression, where an expected High tone surfaces as Low when it occurs after a depressor consonant, which is typically breathy-voiced. (The details and mechanisms of this borrowing are not generally spelled out, while the specifically responsible Khoisan languages are never identified.)[107]

There are many respects in which the general scenario of Khoisan influence on the Nguni languages is ultimately unsatisfactory. It is problematic, for one thing, that words featuring the initial ejective affricate kx’ in Xhosa and Zulu –where it is usually represented by the letters ‘kr’ (Xhosa) or ‘kl’ (Zulu)– are only rarely found to have plausible sources in semantically equivalent lexis of any Khoisan language, whereas they often have clear affines in languages of the Sotho-Tswana group. (The sound itself certainly occurs in many Khoisan languages, including Kora.)

Apart from the lack of strong evidence for any widespread and sustained borrowing from Khoisan languages into the Nguni languages, there are also indications that there must have been some influence in the reverse direction. There are numerous click words in the Nguni languages, for example, that have long been known to be intrinsically BANTU, in the sense that they have obvious non-click cognates in other related languages, and can even be mapped from Proto-Bantu. What is surprising is that a number of these essentially ‘BANTU words with clicks’ are also found in Khoisan languages. An example is provided by a word meaning ‘shake, agitate’, which comes to be used for ‘churn (butter)’, and which is reconstructed for Proto-Bantu as *-k?´p(?d)- ‘shake off, spill’. While this root has regular reflexes in several Bantu languages, a click form of it turns up in Zulu as –g?u?uza.[108] What is more, this click form of the word also occurs throughout the KHOE languages – so widely and systematically that it was possible for Rainer Vossen to reconstruct it for the ancestral language, as *?~ubu for Proto-Khoe itself and the Khoekhoe branch, and as *?~nubu (*?~?ubu) for the Kalahari branch. (The word has also been borrowed into other Khoisan languages.)

Another topic that might repay further investigation is a set of ambivalences associated with the palato-alveolar clicks represented by the symbol ?. For one thing, it has long been noted that these clicks with their diverse accompaniments alternate to varying degrees with affricated non-click equivalents in various languages of the eastern sub-groups of Kalahari KHOE.[109] The prevailing view is that these patterns reflect a process of gradual click loss in the eastern varieties[110] and this is one of the assumptions underpinning the reconstructions proposed for the KHOE family by Vossen. It is less often appreciated that the non-click forms of these words in many cases bear a striking resemblance to semantically linked counterparts in various BANTU languages, particularly those belonging to the Sotho-Tswana group, where the Sotho-Tswana equivalents can be mapped unproblematically from Proto-Bantu, and where the various affrications (the result of palatalising and alveolarising processes) are associated with the well-known influence of the Class 5 prefix.[111]

On a different note, it was observed by Walther Bourquin[112] that, where it is possible to find click words with a shared occurrence in both a Khoekhoe language and one of the Nguni languages, the palato-alveolar clicks (?) of the Khoekhoe languages are in some instances represented in the Nguni language by a dental click (|), but in other instances by a (post)alveolar click (!). An initial consideration of the Kora data suggests that this phenomenon might not be entirely random, since there are some cases within Kora itself where dental and palato-alveolar clicks occur as variants of one another (| ~ ?), while there are other cases where (post)alveolar and palato-alveolar clicks (! ~ ?) appear as alternates. It would be worth investigating whether there is any kind of patterned correlation between these two internal sets and the distributions noted by Bourquin.[113]

A number of BANTU languages spoken in the Okavango region also contain click words, although (except in the case of Yeyi) this is generally on a very much smaller scale than occurs in the Nguni languages, most often involving fewer than a hundred words, where only one click is used, typically with only a limited range of elaborations. These cases will not be discussed here, partly because they do not directly involve Kora, and partly because the local dynamics of their emergence may have been slightly different.[114]

(iv)             Relations between the KHOE languages and varieties of Afrikaans.

In recent years there has been a small surge of academic interest in the possible impact of Khoisan languages on the emergence of varieties of Afrikaans from Cape Dutch, beyond the obvious contribution of a few loanwords and expressions, such as abba ‘carry baby on back’ and kamma kamma ‘seemingly, in a make-believe manner’ (which are of the same order as various loanwords from Malay, such as piesang ‘banana’ and baadjie ‘jacket’). While the pioneering Dutch scholar Hans den Besten considered it likely that some process of early pidginisation and creolisation at the Cape would have played a part,[115] and that local Khoekhoe dialects may have contributed an actual structural influence, other scholars have doubted that there is strong (if any) evidence for either the creolisation or any associated structural influences.[116] Current research accordingly focuses more on identifying features that may have arisen in some varieties from ordinary scenarios of contact or second language influence. This is an intriguing area of ongoing investigation, and it may benefit researchers to include a consideration of the Kora material in future studies, given the continuity between Cape Khoekhoe and Kora, and the once widespread distribution of the language throughout much of central South Africa.

Endnotes