Kilian-Hatz 2003: 152. An adverbial derivative from the noun yó 'size, quantity'. Cf. the textual example: kʰóé-à yó-kà-xā "all Bushmen together". Possible synonym: kó(-á)-kà-(x)ā 'all', also derived from a noun: kó 'quantity, multitude' [Kilian-Hatz 2003: 58-59]. However, this word also serves as the derivational basis for 'many' q.v., and it is not clear whether it really emphasizes the semantics of "entirety" rather than "multitude".
Tanaka 1978: 4. Transcribed as wé-ha 'all'; phonetic detail is questionable due to the inaccurateness of the source, but the lexeme itself is most likely accurate.
Cara:
Not attested.
ǀXaise:
Not attested.
Danisi:
Not attested.
Ts'ixa:
Not attested.
Deti:
Not attested.
Kua:
Not attested.
Tsua:
Not attested.
Hiechware:i-he3
Dornan 1917: 98. Meaning glossed as 'together, all'.
Proto-Kalahari Khoe:*we- #
Distribution: Seems to be well preserved in Naro and ǀGwi-ǀǀGana, but otherwise unclear due to lack of attestation. Replacements: Apparently, Kxoe yó-kà-xā cannot be easily reconciled with *we-, since there are no other examples of such a correspondence and since Kxoe yó is identified by Ch. Killian-Hatz as an independent noun meaning 'size, quantity'; we should probably count this as a case of lexical replacement. As to the projection of *we- onto the Proto-Kalahari Khoe level, the only serious supporting argument for this is external comparison: *we- is most probably related to Khoekhoe *hoa 'all'. Additional data from the Eastern branch are sorely necessary to complete the picture.
Visser 2001: 95; Vossen 1997: 417. Quoted as tóú ~ táú (fem.) in [Barnard 1985: 106, 108] (the form ɳǂûm, listed in the same source as a synonym, actually means 'coals, embers').
Distribution: Preserved everywhere except for Naro-ǂHaba. Replacements: (a) In Naro-ǂHaba, replaced by *tʰau, a root that is also attested in ǀǀAni as tʰàú 'flame' [Vossen 1997: 416], in ǀǀGana as tʰàú 'fireplace' [ibid.] and in Kxoe as tʰéú 'spark; tinder' [Kilian-Hatz 2003: 132]. Curiously, the same root is also the main equivalent for 'ashes' in Khoekhoe; however, internal distribution and semantics clearly speak in favor of the semantic shift 'spark, tinder, burning material' > 'embers' > 'ashes' in one small subgroup of Kalahari Khoe, so perhaps it was an areal isogloss that Naro-ǂHaba shared with Khoekhoe; (b) Deti dù 'ashes' = Kxoe dǔ 'place with charcoal; medicinal charcoal' [Kilian-Hatz 2003: 38], implying a trivial semantic shift {'charcoal' > 'ashes'}. Reconstruction shape: Correspondences are largely trivial, with the predictable development of the palatal click into a palatal affricate in East Khoe (*ɡǂoa > *ʓoa).
Kilian-Hatz 2003: 225; Köhler 1981: 508. The meaning in Kilian-Hatz's dictionary is glossed as '(smooth) bark of tree'; cf. also čóró 'shell, bark' [Kilian-Hatz 2003: 127], although the meaning of this secondary synonym is clearly more general, and in [Vossen 1997: 483] the semantics of 'bark' is not even mentioned at all (only 'shell, pod'). Cf. also Buga-Khoe ǀǀxʼũ̌ id. [Vossen 1997: 421].
Visser 2001: 113. Quoted as ǀǀxʼṹˤ in [Vossen 1997: 421] (with unexpected pharyngealization of the vowel); as ǀǀṍ in [Barnard 1985: 46]. Variants from earlier sources, quoted by Barnard [ibid.] as sora ~ cora ~ cera, actually reflect còrō 'peel, shell' [Visser 2001: 103].
Dornan 1917: 98. Meaning glossed as 'the bark of a tree'.
Proto-Kalahari Khoe:*ǀǀxʼũ
Vossen 1997: 421 (*ǀǀxʼũ). Distribution: Preserved everywhere where attested, with the possible exception of ǀGwi-ǀǀGana and Hiechware. Replacements: Although this cannot yet be regarded as definitively proven, it is nevertheless most likely, based on the geographical and topological distribution of involved languages, that at least ǀGwi-ǀǀGana *gure is an areal form, borrowed by these languages from the nearby !Xóõ (cf. !Xóõ gúle 'bark'). Hiechware ǯoreː < *gure with regular palatalization and is most likely also borrowed from a Taa (South Khoisan) source. Reconstruction shape: Correspondences are generally regular, except for the odd pharyngealized vowel in ǂHaba (irregularly replacing the velar affricate) and the same vocalic reflex in at least some dialects of Naro (as witnessed in Vossen's data). The feature of Naro-ǂHaba pharyngealization has not received a completely satisfactory historical explanation so far, and it is not excluded that, based on it, the reconstruction will have to be amended to *ǀǀxʼũˤ.
Visser 2001: 47. Polysemy: 'stomach / womb'. The meanings 'abdomen' and 'belly', according to Visser, are expressed by compound forms: ɳǀã̄ː !áó 'abdomen' (literally 'stomach-around') and ɳǀã̀ː ǂúː 'belly' (literally 'stomach-head'). Quoted as ɳǀnâ in [Vossen 1997: 419]; as ɳǀâ in [Barnard 1985: 64] (along with a variety of synonyms that are not confirmed elsewhere or have different meanings).
Nakagawa 1996: 107. Meaning glossed as 'stomach', but this is most likely the default equivalent for 'belly' as well. Quoted as ɳǀnâ in [Vossen 1997: 419]; as ɳǂâ (with an erroneous transcription of the click influx) in [Tanaka 1978: 12].
Vossen 1997: 419. Transcribed, however, with a different nasal efflux (as ɳǀâ) in the earlier source, [Vossen 1986: 72]. Transcribed as ɳǂâ (with an erroneous transcription of the click influx) in [Tanaka 1978: 12].
Not attested properly. Cf. ǀxore 'the paunch, the stomach' [Dornan 1917: 103]; ǀaː 'a gemsbok's stomach' [Dornan 1917: 99]; ǀxaː 'the paunch of an animal' [Dornan 1917: 102].
Proto-Kalahari Khoe:*ǀa ~ *ɳǀa
Vossen 1997: 419 (*ɳǀna). Distribution: Preserved in all dialects, although the connection between the two basic variants of the root remains unclear. Reconstruction shape: A major irregularity is the conflict between the variant *ɳǀa, with a nasal click, attested in Naro-ǂHaba, ǀGwi-ǀǀGana, and ǀǀAni, and the variant *ǀa, with a zero efflux click, attested in all East Kalahari Khoe languages as well as Kxoe and ǀGanda. The correspondence seems to be unique, but since the rest of the features (dental click, coda in *-a) coincide, and since not a single language that is represented by a dictionary on this list shows any traces of lexical opposition between the two variants, we tentatively consider them as (for the moment) unexplained variants of a single root, without setting up a special proto-phoneme. Perhaps the variation existed already on the proto-level and was conditioned by unknown factors (e. g. the use of the root as an auxiliary word with the meaning 'inside of (smth.)').
Vossen 1988: 72. Quoted as ɡ!uri in [Tanaka 1978: 13]; as synonyms, the same source adds !káo (actually = 'long' q.v.) and ǂdó (probably = ǀǀGana ɡǀǀó q.v.).
ǀǀGana:ɡǀǀó #5
Vossen 1988: 72. Somewhat dubious. In [Tanaka 1978: 13], the same word is listed as ǂdó (with incorrect transcription of the click influx), along with the synonym guri = ǀGwi ɡ!úrí. According to [Vossen 1988: 72], the basic meaning 'big' in ǀGwi and ǀǀGana is thus expressed by two different roots, but Tanaka lists both roots with the same meaning for both languages. Most likely, both words are indeed present in both languages, but their meanings are close; more research is necessary to find out the real state of things.
Dornan 1917: 103, 104. Meaning glossed as 'big, large'. Alternately, cf. also ɡǀowe 'big, large' [Dornan 1917: 96]. Any selection is dubious; we tentatively choose the variant that matches the data of Hiechware's closest relatives (Kua, Tsua).
Proto-Kalahari Khoe:*kai #
Vossen 1997: 445 (*káí 'big'). Distribution: Highly unstable item, preserved only in Naro and a few East Kalahari Khoe languages. Replacements: The only way to reconstruct PKK 'big' is by means of external evidence: Naro and Deti káí are cognate with Proto-Khoekhoe *kai id., although even here one cannot possibly exclude an areal borrowing of the Khoekhoe word into Naro (and from there, into Deti) - this is not highly likely, though, considering that the borrowing would have to have occurred on a paradigmatic level (both in Khoekhoe and in Naro we see tonal derivation of the adjective 'big' from the verbal stem 'to grow, to be(come) big'). We do, however, have arguments for the other equivalents being replacements: (a) ǀǀAni-Kxoe and Danisi-Tsʼixa *ǀau correspond to Nama ɡǀàű- 'manner, way, style, mode; amount', cf. kai ɡǀau-b mari-b 'large amount of money' [Haacke & Eiseb 2002: 185]; the semantic shift {'amount' > 'large'} is quite realistic; (b) ǀGwi-ǀǀGana *ɡ!uri 'big', also found with click loss in Cara, is perhaps related to Nama ɡ!ùri̋-s 'inland, interior (of country); elevated country', adj. ɡ!ùri̋ 'stubborn, obdurate; insubordinate, contumacious' [Haacke & Eiseb 2002: 323] = !Ora !úrí 'proud, stately' [Meinhof 1930: 112], with an original semantics of 'elevated, high-rising'; (c) ǂHaba ǀǀʼũ, ǀXaise ɡǀǀô, and Kua-Tsua ǀǀũ̂ are all phonetically similar, but each comes with its own click efflux, meaning that they are formally irreconcilable with each other, let alone projectible onto the proto-level. At least the Kua-Tsua item is formally comparable to Proto-Khoe *ǀǀũ 'parent, elder' [Vossen 1997: 431]; the other two remain without a good etymology.
Vozzen 1997: 503 (*ʒada). Distribution: Preserved in all daughter languages. Reconstruction shape: Consonantal correspondences are completely regular. Vocalism shows irregular alternations between a and e in the first syllable; of these, a is the more frequent and widely distributed variant, and e may be judged as a potentially regular result of fronting after a palatal affricate (although it should be noted that such cases are only attested after the voiced consonant *ʒ, and never after its voiceless correlates *c and *cʰ).
Dornan 1917: 108. Cf. also !xom 'to bite, divide, break off' in [Dornan 1917: 103].
Proto-Kalahari Khoe:*pa
Vossen 1997: 422 (*pa). Distribution: Preserved everywhere except for Naro-ǂHaba. Replacements: In Proto-Naro-ǂHaba, replaced by *kaˤ or *kʼaˤ, an areal innovation without any known etymology, although it is, perhaps, the same word as Hiechware ka 'to eat' q.v. (also without a proper Khoe etymology). Reconstruction shape: Correspondences are regular and trivial.
Kilian-Hatz 2003: 191; Köhler 1966: 153. Polysemy: 'black / dark'. Cf. Buga-Khoe ɳǂǔ id. [Vossen 1997: 489]. Secondary synonym: dǔ-ʆī 'be black' [Kilian-Hatz 2003: 38], a verbal derivative from the noun dǔ 'charcoal' [ibid.]. It is unclear whether this innovation is in the process of replacing the old word ɳǂú or is just a "stylistic" synonym.
Visser 2001: 57; Barnard 1985: 121. Quoted as ɳǂnǔ in [Vossen 1997: 489]. The "rare" synonym gáí in [Barnard 1985: 121] is not confirmed in more reliable sources.
Dornan 1917: 99. Phonetically could be = ɲʓu-e (cf. the external data), but the suffix -ɲe is very frequently encountered as an adjectival formant on its own.
Proto-Kalahari Khoe:*ɳǂu
Vossen 1997: 488 (*ɳǂu ~ *ɳǂnu). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular. In Proto-East Kalahari, *ɳǂ- > *nʓ- (regularly) with subsequent additional developments in individual daughter languages. In Ts'ixa, the shift is reversed due to its areal connection with Kxoe (West Kalahari).
Vossen 1997: 425 (*ǀʼáò). Distribution: Preserved everywhere except for Tshwa (Kua-Tsua). Replacements: In Proto-Tshwa, replaced with *taka, clearly a non-inherited term that violates the basic phonotactics of the native Khoe root and is most likely borrowed from a Bantu source; however, the exact source of the borrowing remains unclear (no such word for 'blood' is attested in the neighbouring Shona and Tswana languages). Reconstruction shape: Correspondences are regular and trivial.
Visser 2001: 25. Quoted as ɡǀǀǔ in [Vossen 1997: 426] and [Barnard 1985: 65]; the latter source also lists the allegedly synonymous form ǀǀxĩ, not confirmed elsewhere. Distinct from bî 'female breast, milk' [Visser 2001: 3].
#Haba:
Not attested, but cf. pî 'female breast' [Vossen 1988: 74].
ǀGwi:
Not attested reliably, although cf. g̃!u-ša in [Tanaka 1978: 21] (the transcriptional sign g̃ is not explained; -ša is the feminine marker). This word is distinct from pîː 'female breast' [Nakagawa 1996: 115].
Vossen 1997: 426. Quoted as g̃u-ša in [Tanaka 1978: 21] (the transcriptional sign g̃ is not explained; -ša is the feminine marker). Distinct from pî 'female breast' [Vossen 1988: 74].
Dornan 1917: 98. The dental click is probably erroneously transcribed instead of a lateral click, cf. the same situation with 'moon' q.v. Distinct from samʰ 'female breast' [Dornan 1917: 108].
Proto-Kalahari Khoe:*ɡǀǀu
Vossen 1997: 426 (*ɡǀǀu). Distribution: Preserved in all daughter languages where it is attested. Reconstruction shape: Correspondences are generally regular and trivial (with minor unexplained deviations, e.g. click devoicing in Tsua). Semantics and structure: Reconstructible with the semantics of 'male chest' and consequently opposed to *sam 'female breast' [Vossen 1997: 438] and *pi 'milk' [Vossen 1997: 469] (in daughter languages, each of these two roots is sometimes generalized in both meanings, but they must have been more or less semantically distinct on the proto-level).
Kilian-Hatz 2003: 33. May be used both in the transitive and intransitive meanings. Quoted as dǎo in [Köhler 1981: 490]. Cf. Buga-Khoe dáò 'to burn' [Vossen 1997: 500].
Visser 2001: 16. Transitive stem; according to Visser, the corresponding intransitive stem is dàō, with tonal gradation. Quoted as dàó in [Vossen 1997: 500]; as tàú ~ dàú in [Barnard 1985: 131].
Dornan 1917: 92. Cf. also dʰau-o 'to set on fire' [ibid.] (-o is either the directive suffix or the juncture -a; this form proves that the verb may be used in the required transitive sense). Alternately, cf. also ɳǀǀgai 'to burn' [Dornan 1917: 107].
Proto-Kalahari Khoe:*dao
Vossen 1997: 500 (*dao). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular and trivial. Almost completely homonymous with *dao 'road' q.v. (although the original tonal structure may have been different).
Visser 2001: 117. Quoted as ɡǂoro ~ ǀǀairo 'fingernail' or as ǀǀoro 'finger' in [Barnard 1985: 66] (transcription of the first two variants is most likely erroneous).
#Haba:
Not attested.
ǀGwi:
Not attested in reliable sources, but cf. !ore 'nail' in [Tanaka 1978: 64] (we do not include the word on the list since Tanaka's !- may transcribe different clicks, and its correct etymologization is therefore difficult; it is, however, quite likely, that !ore really = ǀǀore).
ǀǀGana:
Not attested in reliable sources, but cf. !ore 'nail' in [Tanaka 1978: 64] (we do not include the word on the list since Tanaka's !- may transcribe different clicks, and its correct etymologization is therefore difficult; it is, however, quite likely, that !ore really = ǀǀore).
Dornan 1917: 94. The dental click is possibly a mistranscription for the lateral click (cf. external data, as well as the possibility of the same mistake in the word for 'moon' q.v., or the reverse mistake in the word for 'snake' q.v.). Also transcribed as ǀdara in [Dornan 1917: 92].
Proto-Kalahari Khoe:*ǀǀoro
Vossen 1997: 436 (*ǀǀado ~ *ǀǀodo). Distribution: Attested in Naro and throughout the Eastern Kalahari subgroup; possibly also in ǀGwi-ǀǀGana, if the somewhat unreliable source on these languages is correct. Reconstruction shape: Correspondences are generally regular, except for some vocalic peculiarities. In Naro, the vowel is pharyngealized, which may point to a former uvular efflux (*ǀǀqoro); however, no traces of uvular effluxes are seen in East Kalahari languages. In Danisi, the first vowel is non-labial (ǀǀáró), which leads R. Vossen to reconstruct *ǀǀado as the primary variant. This, however, would mean independent (and irregular) assimilation in every other language (as well as external cognates in Khoekhoe); it is more reasonable to postulate an irregular dissimilation in one single language.
Vossen 1997: 436. The semantic difference between ǀǀà and ǀǀórò is unclear.
Ts'ixa:
Deti:
Kua:
Tsua:
Hiechware:
Proto-Kalahari Khoe:*ǀǀa
Vossen 1997: 436 (*ǀǀa). Distribution: This is the main equivalent for 'fingernail' in the ǀǀAni-Kxoe group, and its reflexes are occasionally seen in East Kalahari Khoe languages as well. Reconstruction shape: Correspondences are regular and trivial. Semantics and structure: The forms *ǀǀoro and *ǀǀa are in the state of "criss-crossed" complementary distribution in Kalahari Khoe; both are formally reconstructible to top level, but no easy scenarios of semantic shifts can be offered. External parallels (Khoekhoe *ǀǀoro 'fingernail') suggest that *ǀǀoro is more archaic than *ǀǀa; on the other hand, *ǀǀoro is also more widespread in the other Khoisan languages (cf. Proto-Taa *ǀǀqure, etc.), which could hint at its areal provenance. Until more data become available, it is prudent to include both forms on the proto-wordlist as "technical synonyms".
Visser 2001: 59. Formally derived from the adjective ɳǀǀòm 'cloudy' [ibid.]. Idiomatic synonym: túː-ǀʼõ̀ː 'cloud', literally 'rain-hair' [Visser 2001: 105]. In [Barnard 1985: 32], the meaning 'cloud' is rendered either by túː 'rain' q.v. or by ǀóː 'white' q.v.
#Haba:
Not attested.
ǀGwi:
Not attested in reliable sources, but cf. kyu-ko 'cloud' in [Tanaka 1978: 22] (probably the same root as in *tu 'rain' q.v.).
ǀǀGana:
Not attested in reliable sources, but cf. kyu-ko 'cloud' in [Tanaka 1978: 22] (probably the same root as in *tu 'rain' q.v.).
Dornan 1917: 104. Possible mistranscription of the palatal click instead of the required lateral (cf. external data).
Proto-Kalahari Khoe:
Not properly reconstructible due to lack of attestation. The only possible candidate is *ɳǀǀom ~ *ǀǀqom, but the words are only encountered in one branch of Kalahari Khoe (ǀǀAni-Kxoe-Naro), and even there, phonetic correspondences remain unclear (uvular efflux in ǀǀAni-Kxoe irregularly corresponds to nasal efflux in Naro). The word is hardly separable from !Xóõ (Taa) qùm 'to be overcast', but the nature of the connection (genetic? borrowed? which direction?) also remains unclear.
Kilian-Hatz 2003: 173. Meaning glossed as 'be cool, be cold (vb.); coldness; winter, cold season (n.)'. Secondary synonym: ǀxùnú 'be cold; be okay, feel good' (dubious in terms of eligibility for inclusion, since the semantics of the word seems to be "positive" rather than "negative").
Visser 2001: 65. Meaning glossed as 'cold, cool', also as verbal 'be cold' and as nominal 'coldness' (without tonal gradation). Quoted as !ai in [Barnard 1985: 115] (R. Vossen's more correct transcription !xàì is also quoted ibid.). The same source also lists ǂxí ~ ǂxʼé ~ !xí 'cold' as synonyms, but these forms are not confirmed in more reliable sources; they may simply be phonetic corruptions of the correct !xai.
Dornan 1917: 97. Cf. kai caː 'cold water' [Dornan 1917: 100], where kai may or may not be the same word as haiː, heard and transcribed differently (also in kaie ɲaː 'wintry weather'; also in the derivative xai-ɲe 'frost, coldness' [Dornan 1917: 102]). Cf. also a possible synonym: ɡǀuru 'cold', ɡǀuru-wa 'to be cold' [Dornan 1917: 96].
Proto-Kalahari Khoe:*!qʰai
Distribution: Preserved in Naro-ǂHaba, ǀGwi-ǀǀGana, and some of the Eastern languages. Replacements: [1] The most frequent alternative is *ǀxonu, a form that is also found in Naro (as ǀxònō [Visser 2001: 7]) in the meaning 'kind of wind that is cold', and in Kxoe as ǀxùnú 'be cold; be okay, feel good' (see notes on Kxoe). Although the distribution of reflexes is technically sufficient to project it onto the Proto-Kalahari Khoe level in the meaning 'cold', it is suspicious that not a single large dictionary of a Kalahari Khoe language gives it as the primary equivalent for 'cold'; this, coupled with the fact that external Khoekhoe data supports the selection of *!qʰai as the main equivalent, makes us currently judge the term as an innovation (perhaps the original meaning was 'cold wind', as in Naro). [2] Kxoe ǀxʼává has no known parallels in Kalahari Khoe, but is phonetically corresponding to and semantically compatible with Nama ǀʼawa-s 'North'; it is unclear if we should judge this as an areal connection or as a common reflex of a common Proto-Khoe item with the meaning 'North', but in any case, the word is clearly ineligible for the status of Proto-Kalahari Khoe 'cold'. [3] ǀXaise-Deti kárá has no clear etymology; cf., perhaps, Naro !ārà-!ārà 'to be clear (of weather, etc.)' [Visser 2001: 64]? [4] Tsua hùrú is an isolated and unclear form as well. Reconstruction shape: Correspondences for the click efflux in this particular case are rare, and the degree of their regularity is hard to establish without additional data on poorly described languages. On the whole, fluctuation between such reflexes as -qʰ-, -q-, and -x- speaks in favor of selecting the rare uvular aspirated efflux *-qʰ- as the original phonation type (since we accept the general idea of uvular phonemes and click effluxes reconstructible for Proto-Kalahari Khoe); however, this decision may be amended in the future if additional data become available.
Vossen 1997: 458 (*ha). Distribution: Preserved almost everywhere (with the possible exception of ǀGanda). Replacements: ǀGanda ǀǀàḿ = Kxoe ǀǀāḿ 'to run, run into smth.' [Kilian-Hatz 2003: 199]; beyond this, no obvious parallels are found, but it is clear that, provided the semantics of the ǀGanda verb are indicated correctly, we deal with semantic innovation {'to run into smth.' > 'to come'}. Reconstruction shape: Correspondences are regular and trivial (with the exception of *h- > y- in Kxoe, which still seems to be a regular development).
Visser 2001: 119. Quoted as ǀǀʼó in [Vossen 1997: 496]; as ǀǀó ~ ǀǀʼó in [Barnard 1985: 135]. The quasi-synonym kâ, listed by Barnard, is equal to kâ 'absent, finished; not be anymore, be absent, be dead' in [Visser 2001: 33], i.e. most likely represents an occasional euphemism.
Vossen 1988: 78. Curiously, in [Vossen 1997: 496] this form is not listed; instead, we find the form ǀǀʼǒ (with preservation of the click) in the nominal meaning 'death'.
Dornan 1917: 108. Meaning glossed as 'death, to die'; cf. also ohe 'dead', oha-xoː 'a dead body'.
Proto-Kalahari Khoe:*ǀǀʼo
Vossen 1997: 495 (*ǀǀʼo). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular and largely trivial, with the exception of semi-regular (seemingly unconditioned) loss of the click influx in the Tshwa subgroup.
Vossen 1988: 78; Vossen 1997: 453. R. Vossen lists no other equivalent for 'dog'; however, in [Tanaka 1978: 29] two synonyms are given - arugu and aba.
Vossen 1997: 453 (*/h/adi/ku/). Distribution: Preserved in Naro-ǂHaba and many other languages of the West Kalahari Khoe area. Replacements: The situation here is somewhat complex. For Proto-Eastern Kalahari Khoe, the form *ʔaba 'dog' (Vossen 1997: 453) is reconstructible quite unambiguously. Since it is also found in some Western languages (ǀǀGana, most notably), it could be considered the optimal candidate for Proto-Kalahari Khoe in general. However, phonetically similar forms in ǀGanda (ʔápà) and Khoe (ápā) feature a blatant irregularity in the form of an intervocalic voiceless -p-; moreover, intervocalic -p- is highly atypical of inherited Khoe roots in general, violating the normal phonotactics of bisyllabic stems. Considering that all these forms are phonetically similar to Common Bantu *=búà 'dog' (cf. Shona i=mbwa, etc.), it is not unlikely that they were diffused across the languages already after the disintegration of Kalahari Khoe. This leaves the Western forms as more reliable candidates, particularly since they have external parallels in Khoekhoe languages. Reconstruction shape: There are several irregularities connected with *haˤri-gu, namely, the disappearance of h- in ǀǀAni and Kxoe and the presence of pharyngealized vowel articulation in Naro-ǂHaba (these phenomena may be historically connected). These are not crucial enough to deny the word a proper Proto-Kalahari Khoe status, but should be noted all the same.
Dornan 1917: 102. The palatal click symbol here most likely mistranscribes non-click articulation (ejective; cf. the same situation in the word for 'mouth' q.v.). Alternately, cf. tʰa 'to drink' [Dornan 1917: 109] (this could hardly be the same word, although Dornan's transcriptional standards are so low that anything could happen).
Proto-Kalahari Khoe:*xʼa
Vossen 1997: 497 (*kxʼa). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are mostly regular and trivial (although preservation of xʼ- in Danisi is unexpected).
Dornan 1917: 104. Meaning glossed as 'to dry, wither, melt'; cf. the external data in Kua and Tsua that confirm the meaning 'dry', as well as the noun ǀoː 'dryness' [ibid.] (actually the same word).
Proto-Kalahari Khoe:*ǀǀxo
Vossen 1997: 497. Distribution: Preserved everywhere except for Naro and Tshwa. Replacements: Formally, there are two roots competing for the basic slot 'dry' on the PKK level: *ǀǀxo, found in the majority of branches, and *ǀʼo, found in Naro and Tshwa. External comparison even shows that *ǀʼo is better supported, since it is also the default equivalent for 'dry' in !Ora (Khoekhoe). However, more careful analysis of the semantics of *ǀʼo shows that it must have referred to 'excessive dryness': cf. the opposition in Kxoe, where ǀǀxó = 'to be dry', ǀʼóò = 'to dry out, boil over'; externally, note also the meaning 'to be barren (of cow)' for this word in Nama. More data are necessary to confirm this hypothesis; in the meantime, in between this observation and the "majority rule", we postulate *ǀǀxo as the default equivalent for PKK. Reconstruction shape: Correspondences are regular and trivial.
Vossen 1997: 474 (*ǂe). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular, with expected loss of click articulation in East Khoe (where *ǂ- > ɕ- everywhere except for Tsʼixa).
Not attested. Possibly the same word as 'sand' q.v.
Proto-Kalahari Khoe:*xom
Vossen 1997: 425 (*xom). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular and trivial. Semantics and structure: Reconstructible with polysemy: 'earth / sand' (see further notes on 'sand').
Visser 2001: 92. Quoted as ǂʼṹ in [Vossen 1997: 433]; as ǂʼṹ ~ ǂʼṍ in [Barnard 1985: 137]. This is the basic word for 'eat' as applied to various soft substances; it is distinct from xʼōː 'to eat meat' (of humans and animals) [Visser 2001: 39] (see further details under 'meat').
Vossen 1997: 433. Distinct from kʼó 'to eat meat' [ibid.]. In the earlier source [Vossen 1988: 80], both of these meanings are rendered with the verb kʼán; the semantic difference remains unclear.
Dornan 1917: 99. Cf. ɲoː 'to eat grass, to graze; food, meat; to sit down to food, to eat' [Dornan 1917: 108].
Proto-Kalahari Khoe:*ǂʼũ
Vossen 1997: 433 (*ǂʼũ). Distribution: Preserved almost everywhere, with the possible exception of ǀGwi and Hiechware. Replacements: (a) ǀGwi kiam = Naro kāˤm̀ 'sip, keep something in mouth, drink' [Visser 2001: 34]. It is not clear, however, if there was a real replacement in ǀGwi or if we are dealing with just a semantic inaccuracy; (b) Hiechware ka = Naro kǎˤ 'bite' [Visser 2001: 33]. This actually looks like a real replacement, judging by several of the available contexts; however, it should be noted that this is not the original Proto-Khoe equivalent for 'to bite', either, but rather a local isogloss between Naro-ǂHaba and Hiechware. Reconstruction shape: Correspondences are generally regular, with predictable click loss in East Kalahari Khoe.
Vossen 1997: 451. Polysemy: 'egg / testicle / penis'. In order to differentiate between these meanings, the semantics of 'egg / testicle' may be expressed with the compound form ʔíbí-ǀxúrì-!xú, where ǀxúrì = 'seed' q.v. and the third component is not clear.
Vossen 1997: 451. Polysemy: 'egg / testicle / penis'. In order to differentiate between these meanings, the semantics of 'testicle' may be expressed with the compound form ʔíbí-ǀxá, literally 'egg-flesh'.
Vossen 1997: 451 (*ǂʼubi). Distribution: Preserved in all languages where attested. Reconstruction shape: Correspondences are generally regular, with predictable click loss in the Eastern branch (and occasional vocalic assimilation *ǂʼubi > *ǂʼibi).
Vossen 1997: 418. Quoted as ǂxéí in the earlier source [Vossen 1988: 80]; as !kai in [Tanaka 1978: 33] (with incorrect identification of both the click influx and efflux).
Dornan 1917: 90. Transcribed as čai, without the extra length, in idiomatic compounds such as čai ǀʰayo 'eyebrow', etc.
Proto-Kalahari Khoe:*ǂxai
Vossen 1997: 418 (*ǂxai). Distribution: Preserved in all daughter languages. Reconstruction: Correspondences are fully regular, including predictable click palatalization in the Eastern branch.
Kilian-Hatz 2003: 218; Köhler 1981: 501. Polysemy: 'grease / oil / fat'. There does not seem to be any lexical distinction between 'hard fat' and 'liquid fat', but the former meaning may be specified with the idiomatic expression ǀǀʼúū ɳǀǀúì, literally 'frozen fat/oil' [Kilian-Hatz 2003: 209]. Cf. Buga-Khoe ɳǀǀúì id. [Vossen 1997: 435].
Visser 2001: 60; Vossen 1997: 435. Polysemy: 'fat / oil / butter / cheese, cream'. Quoted as ɳǀǀúè in [Barnard 1985: 68]. The secondary synonym ǀuia ~ ǀuiia in the same source is obscure and not confirmed elsewhere.
Dornan 1917: 97. Meaning glossed as 'fat of an animal'. Click transcription is probably erroneous (the articulation is more likely to have been lateral, judging by external data).
Proto-Kalahari Khoe:*ɳǀǀui
Vossen 1997: 435 (*ɳǀǀnui ~ *ɳǀǀui). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are generally regular and trivial. Semantics and structure: It seems that there was no lexical differentiation between 'solid fat' and 'liquid fat, oil' on the proto-level.
Dornan 1917: 95. Same word as 'hair' q.v., despite the difference in transcription.
Proto-Kalahari Khoe:
Not reconstructible due to lack of attestation in most languages. Based on very scarce data (as well as external analogies), it is reasonably safe to assume that '(head) hair' and 'feather' were not differentiated in the proto-language.
Vossen 1997: 435 (*ǀʼe). Distribution: Preserved in the majority of languages. Replacements: Replaced in Deti by the form ɡǀǀǔ = Naro ɡǀǀùú 'make fire' (Visser 2001: 25). Reconstruction shape: Correspondences are regular and trivial.
Dornan 1917: 110. Borrowed from Setswana ƛʰapi id.
Proto-Kalahari Khoe:*ǀǀʼau
Vossen 1997: 436 (*ǀǀʼau). Distribution: Preserved in most languages (except where replaced by Bantuisms). Replacements: The only explicitly attested replacement is Hiechware tʰapi, borrowed from Setswana. Reconstruction shape: Correspondences are generally regular, except for the sporadic (but recurrent) click loss in the Shua branch.
Number:30
Word:fly v.
ǀǀAni:
Not attested.
ǀGanda:
Not attested.
Kxoe:ɡǀǀèvùū1
Kilian-Hatz 2003: 212. Meaning glossed as 'fly (of: birds)'. Alternately, cf. also kʸã̄ã́, glossed as 'run fast; run away, escape, fly; go by car' [Kilian-Hatz 2003: 79]. Dictionary definitions presume that ɡǀǀèvùū is better eligible for inclusion, but additional research on the issue may be required.
Visser 2001: 98. Distinct from ǂéré 'to fly around' [Visser 2001: 83]. Quoted as cã́ĩ̀ˤ by R. Vossen in [Barnard 1985: 139]; Barnard himself gives a completely different word, !xué = !xóé 'to run' [Visser 2001: 66].
#Haba:
Not attested.
ǀGwi:
Not attested in reliable sources. Cf., however, !ʰoe 'to fly' in [Tanaka 1978: 37], most likely the same as hoe 'to run' in [Tanaka 1978: 81]; the lack of click articulation in the latter entry may be erroneous, cf. the complete equivalence of the two meanings in Gǀǀana.
ǀǀGana:
Not attested in reliable sources. Cf., however, hoe 'to fly' in [Tanaka 1978: 37], clearly the same as hoe 'to run' in [Tanaka 1978: 81].
Kilian-Hatz 2003: 78. Polysemy: 'foot / wheel'. Quoted as kʸǎré in [Köhler 1981: 489]. Cf. Buga-Khoe kádè id. [Vossen 1997: 439]. For Buga-Khoe, R. Vossen also lists the form ʒĩ̂ 'foot' as a synonym [Vossen 1997: 440]. However, in Kxoe proper the word ǯĩ̌ means 'bird's foot, talon' [Kilian-Hatz 2003: 42].
Tanaka 1978: 37. Transcribed as ngare; although this is not a reliable source, we tentatively include the word, since it is a natural match with other West Khoe equivalents for 'foot'.
Tanaka 1978: 37. Transcribed as ng!àre; although this is not a reliable source, we tentatively include the word, since it is a natural match with other West Khoe equivalents for 'foot'.
Dornan 1917: 101. Dubious, since the meaning is glossed as 'toes, claws, heels, hoofs; the spoor made by the feet'. However, there is no separate equivalent for 'foot, feet' in Dornan's vocabulary, and the external data clearly shows that 'foot' is the original meaning here; perhaps it was actually surmised under 'toes'.
Proto-Kalahari Khoe:*ɳ!are ~ *!are
Vossen 1997: 440 (*ɳ!adi). Distribution: Preserved in all of West Kalahari Khoe and in the Kua subgroup of East Kalahari Khoe. Replacements: Replaced in the Shua subgroup by *ʒĩ, formerly probably 'toe', cf. ʒê 'toe' in ǀǀAni, as well as the external parallel cĩ-s 'big toe' in Nama; thus, {'toe' > 'foot'}. Reconstruction shape: A big problem with this root is the irregular alternation between the nasalized click ɳ!- and the simple click !-. The former variant is more widespread and better supported by external evidence (Khoekhoe *ɳ!ari 'to drive; wheel'), but the reasons for irregular denasalization in ǀGanda, Kxoe, and Kua would remain unclear. Possible options include an unusually rare click efflux development and contamination between two different roots.
Dornan 1917: 96. The alveolar click is quite likely to have been transcribed in error (instead of the etymologically expected dental click; cf. the same situation with 'near' q.v.). The correct dental click is attested in the verbal form ǀweha 'to be full' [Dornan 1917: 105].
Proto-Kalahari Khoe:*ǀxʼoɛ
Vossen 1997: 503 (*ǀxʼoe). Distribution: Preserved in all daughter languages (unless Hiechware ɡ!wea is a different root rather than a typo, which is not very likely). Reconstruction shape: Correspondences are regular. The original diphthong was probably phonologically -oɛ rather than Vossen's -oe, as seen in the ǀǀAni-Kxoe branch and confirmed by external data (Proto-Khoekhoe *ǀxʼoa rather than *ǀxʼoe). Semantics and structure: Original verbal root ('to fill / be filled').
Visser 2001: 45. Quoted as mã̂ in [Vossen 1997: 441]; as mâ in [Barnard 1985: 140]. Secondary synonym: āū 'to give (polite request)' [Visser 2001: 2], quoted as ou ~ au 'give (polite form)' in [Barnard 1985: 140].
Dornan 1917: 110. This is the only word in Dornan's vocabulary glossed simply as 'to give'. Cf. also maː 'to lend, to give to, to forgive' in [Dornan 1917: 106] (not likely to be the main word for 'give' in this dialect, since it is not attested in the closely related Kua and Tsua in this basic meaning).
Proto-Kalahari Khoe:*ma ~ *mã
Vossen 1997: 440 (Proto-West Khoe *mã, Proto-Khoe *ma). Distribution: Preserved in Naro-ǂHaba and ǀGwi-ǀǀGana, possibly also in some East Kalahari Khoe languages in semantically shifted meanings. Replacements: Replaced (a) in Proto-ǀǀAni-Kxoe by *ǂxa, a root of unclear origin (perhaps related to Proto-Khoekhoe *ǂxa 'to refuse'?); (b) in Hiechware by tʰeː, also of unknown origin. Reconstruction shape: Fluctuation between nasalized and non-nasalized variants of the vowel are unclear, but otherwise correspondences are regular.
Kilian-Hatz 2003: 121. Quoted as tʼôn in [Köhler 1981: 493]. Cf. Buga-Khoe tʼṹĩ̀ id. [Vossen 1997: 445]. The latter source also lists Buga-Khoe qádí as a synonym; in Kxoe proper, the corresponding item qáré is glossed as 'be sweet' in [Kilian-Hatz 2003: 108].
Visser 2001: 62. Quoted as !ã́ĩ̀ in [Vossen 1997: 445]; as !ã́ĩ̀ ~ !kã́ĩ̀ in [Barnard 1985: 116]. The quasi-synonym tʼṍẽ̀ [Visser 2001: 105] = tʼṹẽ́ [Barnard 1985: 116] actually means 'beautiful' (Barnard: "...is usually used in the sense of 'visually beautiful', but like the English word 'beautiful', has a more general sense as well").
Vossen 1988: 83. Cf. also qádè, with polysemy: 'good / pleasant / comfortable / sweet' in [Vossen 1997: 445]. Cf. ɕʼṍẽ̀ 'beautiful' [Nakagawa 1996: 110]; the same word is probably quoted as ekyuen (phonetically = /e/tʼõẽ ?) 'beautiful; good' in [Tanaka 1978: 11, 41].
Vossen 1997: 445. Quoted as kẽ́ĩ̀ in the earlier source [Vossen 1988: 83]. Cf. also qádè, with polysemy: 'good / pleasant / comfortable / sweet' in [Vossen 1997: 445]. Cf. ɕʼṍẽ̀ 'beautiful' [Nakagawa 1996: 110], quoted as tʼṹĩ̀ 'good, beautiful' in [Vossen 1997: 445]; the same word is probably quoted as ekyuen (phonetically = /e/tʼõẽ ?) 'beautiful; good' in [Tanaka 1978: 11, 41].
Vossen 1997: 445 (*tʼon/i/). Distribution: Preserved everywhere, but possibly underwent a slight semantic shift in the Naro-ǀǀGana subbranch. Replacements: Three roots are reconstructible with the overall semantics of 'good' for Proto-Kalahari Khoe. Of these, *tʼoni is the most widespread. Its main competition is *!ãĩ, not found in ǀǀAni-Kxoe, but well attested in all the other subbranches. R. Vossen defines *!ãĩ as 'gut / richtig' and *tʼoni as 'gut / schön'; if this were correct, then *!ãĩ would be a better candidate for the Swadesh slot. However, precise semantic descriptions are not available for these two roots in any of the languages, and reflexes of *!ãĩ in daughter languages are also described by Vossen as 'schön' occasionally. Additionally, there is a strong possibility that the meaning of *!ãĩ, or even its presence in some of the daughter languages (such as Naro), may have been influenced by Khoekhoe languages, where *!ãĩ 'good' is indeed the main (and only) equivalent for the term. The third root, *qari (Vossen: *qadi), in most languages has the specific semantics of 'sweet' (and, figuratively, 'joyful') rather than general 'good'. Reconstruction shape: Correspondences are somewhat vague here, but it makes sense to interpret the fluctuation between coda variants -ũĩ and -on as reflecting an original triphonemic combination *-oni.
Number:35
Word:green
ǀǀAni:
Not attested.
ǀGanda:
Not attested.
Kxoe:ǀxʼáó1
Kilian-Hatz 2003: 173. Verbal stem (meaning glossed as 'be green'). Examples show this wordto be the default equivalent for the required meaning, cf. ǀxʼáó ɡǀã̌ "green leaf". The word should be distinguished from xʼóā 'be green, be unripe, uncooked, raw' [Kilian-Hatz 2003: 74] (see under 'new'), as well as qéú-ʆī 'be bright green, be blue (of: sky)' [Kilian-Hatz 2003: 109].
Visser 2001: 98. Meaning glossed as 'blue/green, light brown'. Quoted as cã̀ ~ cǎˤ in [Barnard 1985: 121]. The same source also lists additional equivalents for 'pale green': còm ~ còˤm = còˤm 'grow, germinate' [Visser 2001: 102] (cf.: coˤma iaː "it is greenish (said when all the trees are green)"), and for 'dark green': !am (not confirmed in other sources; if correct, most likely a borrowing from Nama, cf. Nama !am 'green').
Nakagawa 1996: 109. Meaning glossed as 'blue, green'. Quoted as t ̌sae 'green' in [Tanaka 1978: 42] (it is not quite clear what kind of sound is meant by the transcription t ̌s).
ǀǀGana:
Not attested in reliable sources, but cf. t ̌sae 'green' in [Tanaka 1978: 42] (it is not quite clear what kind of sound is meant by the transcription t ̌s).
Cara:
Not attested.
ǀXaise:
Not attested.
Danisi:
Not attested.
Ts'ixa:
Not attested.
Deti:
Not attested.
Kua:
Not attested.
Tsua:
Not attested.
Hiechware:baratʰi-1
Dornan 1917: 89. The phonetic structure of the word almost certainly excludes a native origin, but the source of borrowing is not clear. Another synonym is čabe 'green' [Dornan 1917: 90], without any known etymology, but more "Khoe" in form.
Proto-Kalahari Khoe:
Not reconstructible due to lack of attestation. Naro cã́ˤ and ǀGwi cxʼāē are phonetically similar, but do not regularly correspond to each other (and even if they did, the distribution of this cognate would not have Proto-Kalahari Khoe status).
Dornan 1917: 111. Polysemy: 'hand / arm / finger'. Dornan also records the meaning 'tail', but external data clearly show that this is a confusion of two near-homonyms.
Proto-Kalahari Khoe:*cʰau
Vossen 1997: 447 (*cʰau). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are mostly regular (examples of Proto-Kalahari Khoe *cʰ- are rare and reflexes are somewhat chaotic, but enough of the languages show an aspirated reflex to make a safe assumption that it goes back to the proto-state).
Visser 2001: 90. Quoted as ǂú in [Vossen 1997: 458]; as !kuː ~ !uː in [Barnard 1985: 63], with erroneous transcription of the click influx articulation.
Vossen 1988: 85. The attested structure of this word is very peculiar and hints at compound origins, but they are not determinable on the synchronic level. In [Tanaka 1978: 45], however, the word for 'head' is quoted simply as mâ.
Vossen 1997: 458 (West Khoe: *ǂu, East Khoe: *ma). Distribution: The two main Kalahari Khoe terms for 'head' have an intersecting distribution: *ǂu is primarily West Kalahari Khoe (but is also found in Tsʼixa, where, however, it could have easily been loaned from West Khoe), *ma is primarily East Kalahari Khoe, but is also found in ǀGwi-ǀǀGana and ǂHaba. Upon first sight, Vossen's *ma has a much wider distribution and has to be regarded as the primary candidate for 'head' in Proto-Kalahari Khoe. However, there are two arguments against that: (a) Since the existence of a separate West Kalahari Khoe branch is dubious, the isogloss between Naro and ǀǀAni-Kxoe is formally just as strong as the isogloss between East Kalahari Khoe and ǀGwi-ǀǀGana. There is also a potential external match for *ǂu 'head' in Nama *ɡǂu-ro 'first, initial' (possibly a suffixal derivate from an older 'head'); (b) The situation with *ma 'head' is also far from clear, since collective evidence shows that its primary shape must have been more complex: Tsua ḿ-ʔà, ǂHaba m̀-kʼá and ǀGwi m̀ʔm̂ point to *m(a)-kʼa or, perhaps, with click loss to *m(a)-!ʼa, where *!ʼa = 'to know, to understand' (cf. in that respect an important external parallel in !Ora: !ʼã 'to hear' > !ʼã-b 'head'). If this is indeed a compound formation, and the widespread variant *mâ is the result of contraction, then the simple root *ǂu is a better candidate for the original Swadesh slot - even despite the fact that the origin of the morpheme *ma remains obscure under this scenario. Reconstruction shape: Correspondences are regular and trivial.
Vossen 1997: 452 (*kúḿ ~ *kóḿ). Distribution: Preserved in the majority of languages. Replacements: Allegedly replaced in ǀXaise with ǀǀám̀ 'to feel', with common semantic shift {'to feel' > 'to hear'}. Reconstruction shape: Correspondences are generally regular. Root vocalism should rather be reconstructed as *u, based both on external correspondences (!Ora kum) and distribution of reflexes between daughter languages.
Kilian-Hatz 2003: 177. Polysemy: 'heart (n.) / be happy, be glad (vb.)'. Cf. Buga-Khoe ǂáò id. [Vossen 1997: 449] (although in the earlier source [Vossen 1988: 85], an entirely different equivalent for Buga is given: cábá).
Visser 2001: 81; Vossen 1997: 449. Polysemy: 'heart / character / mood / want, be about to' (the verbal meaning, according to H. Visser, is restricted to the Xanagas dialect). Quoted as ǂáó ~ ǂáú ~ ɡǀau in [Barnard 1985: 67] (the variant with the dental click is quite dubious).
Vossen 1997: 449 (*ǂao). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular, with predictable click loss (palatalization) in the Eastern subbranch.
Vossen 1997: 452 (*ɳǀǀnâ). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are generally regular. Rainer Vossen's suggested phonological opposition between *ɳǀǀ- and *ɳǀǀn- in modern languages as well as in Proto-Khoe has so far not been properly confirmed.
Visser 2001: 96. Quoted as tí-yá ~ tí-rá in [Vossen 1997: 239]. This is the full accented (emphatic) form. There is also a shortened, non-emphatic, variant in which the first syllable of the form tí-rá is deleted: ra ~ da ~ r [Visser 2001: 75], dà ~ rà [Vossen 1997: 239]. However, the main pronominal root morpheme is unquestionably tí, cf. the 1st sg. object concord marker -tì [Vossen 1997: 239].
Vossen 1997: 369 (*ta). Distribution: Preserved in all daughter languages, albeit with morphological variation. Reconstruction shape: The original Proto-Kalahari Khoe paradigm should probably be reconstructed as *ti (subject form), *ti-a (object form), the way it has been preserved in many Western languages and a few Eastern ones (most notably Kua). This makes it possible to suggest that the variant ta, typical of the Shua subgroup, actually represents a generalization of the old contracted object form (*ti-a > ta). Rainer Vossen's suggestion that *ta is the original root, while ti is actually an old pronominal prefix replacing the true root, is not supported by the distribution of the forms.
Dornan 1917: 96. Meaning glossed as 'to murder, kill, defeat, conquer'; also as nominal 'war, slaughter, conquest, murder'. Cf. also ɡǀohe 'to kill oneself' [Dornan 1917: 95].
Proto-Kalahari Khoe:*ǀxʼũ
Vossen 1997: 497 (*ǀxʼũ). Distribution: Preserved in the majority of languages. Replacements: In Deti the word seems to have been replaced with a reflex of Proto-Kalahari Khoe *!ʰàé 'to stick (into)' [Vossen 1997: 495]. Reconstruction shape: Correspondences are generally regular, except for some unpredictable tonal discrepancies.
Visser 2001: 71. Quoted as !údù in [Vossen 1997: 457]; as !órù ~ !kórù ~ !kúrù in [Barnard 1985: 66]. Cf. also ɳǀǀõãˤ xʼáí 'knee, patella' [Visser 2001: 59], literally 'face (xʼáí) of stone (ɳǀǀõãˤ)' (?) - an unclear formation, unless ɳǀǀõãˤ is not really 'stone', but a homophone with an anatomic meaning, not encountered on its own.
Vossen 1997: 457. According to this source, the variant kúkúrù co-exists with the disyllabic variant kúrù, but in the earlier source [Vossen 1988: 88], only the trisyllabic variant is listed.
Vossen 1997: 457. According to this source, the variant kúkúrù co-exists with the disyllabic variant kúrù, but in the earlier source [Vossen 1988: 88], only the trisyllabic variant is listed.
Vossen 1997: 457 (*ǀǀoe). Distribution: Attested in all major branches except for Nara-ǂHaba, but somewhat unstable. Replacements: There are two main candidates for the status of 'knee' in Proto-Kalahari Khoe, both with comparable distribution: PKK *ǀǀoɛ and PKK *!uru. Of these two, *ǀǀoɛ is better supported by the following two arguments: (a) it is confirmed externally by the presence of Khoekhoe *ǀǀoa as the main equivalent for 'knee', whereas *!uru has no parallels in Khoekhoe; (b) the form *!uru cannot be separated from the trisyllabic variant kukuru, which is attested in some Shua and Tsua dialects and seems to be a probable ancestral stage for the contracted variant *!uru; however, trisyllabic stems like those, with contractions in daughter dialects, are otherwise unknown in Khoe languages, so they are much better explained as the remains of some unclear substrate, perhaps independently borrowed into several languages already after the split of Kalahari Khoe. Since kuru ~ kuTu 'knee' is one of the more widespread "pan-African" roots, its presence in some Kalahari Khoe lineages as an areal feature is understandable, even if not all the details are clear. Reconstruction shape: Reflexes of *ǀǀoɛ are largely trivial. The original open diphthong *-oɛ is preserved in ǀǀAni and regularly corresponds to Khoekhoe *-oa rather than *-oe.
Kilian-Hatz 2003: 203. Polysemy: 'joint / knee / node of reed shaft'. It is not clear whether ǀǀóɛ́ or kûrù is the main equivalent for 'knee' in Kxoe, or if the two words represent dialectal varieties. Since both forms have external correlates with the meaning 'knee', it is recommendable to include them as synonyms.
Vossen 1997: 457. Used either by itself or within the compound form ǀǀóé-ǀʼṍã́, where the second root = 'bone' q.v. The semantic difference between this root and kúkúrù is unclear.
Nakagawa 1996: 107. Quoted as !ʼã́ in [Vossen 1997: 508]; as !an-aː-ha in [Tanaka 1978: 54] (= !ʼã-a-ha with juncture and imperfective verbal suffixes attached).
Vossen 1997: 508. Quoted as ʔã́ã̄ in [Nakagawa 2006: 203]; as an-aː-ha in [Tanaka 1978: 54] (= ʔã-a-ha with juncture and imperfective verbal suffixes attached).
Vossen 1997: 507 (*!ʼã). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are regular; loss of click articulation, however, is unusually widely distributed (it is lost even in languages like ǀGwi and ǀǀGana that typically preserve it), so areal interference is not excluded. Semantics and structure: Should be distinguished from the semantically close root *ǂʼan 'to think' (Vossen 1997: 428), sometimes glossed as 'to know' in sources.
Vossen 1997: 425. In [Vossen 1988: 106], 'leaf' is instead given as ǀxɛ́ù; considering that in the same source, Buga-Khoe 'leaf' is also rendered as ǀxʼɛ́ù = Kxoe ǀxʼáó 'green colour', as in Kxoe ǀxʼáó ɡǀã̌ 'green leaf' [Kilian-Hatz 2003: 173], it may be suspected that Vossen's earlier data incorrectly list the word 'green' instead of the required 'leaf'.
Visser 2001: 96. Quoted as tòːárà ~ dòːˤáràˤ in [Barnard 1985: 47]. The same source lists ɡǀana (fem.) as a synonym in the meanings 'leaf (general)' and 'small leaf' alike. Etymological connections show that this must be the real archaic equivalent for 'leaf (gen.)'; however, R. Vossen does not list it in his comparative series for 'leaf', and in [Visser 2001: 16], ɡǀànā is glossed with the meaning 'little branch which has leaves, part of any tree', which makes it ineligible for inclusion. Most likely, there has been a semantic shift in recent times. Cf. also the phonetically similar ɡǀã̀ã̄ 'grass' [Visser 2001: 96] = ɡǀã ~ ɡǀãː id. [Barnard 1985: 62]. Finally, in [Vossen 1988: 106] the word for 'leaf' is listed as dáˤnàˤ-sá (e. g. the same item as in ǂHaba). More accurate research is needed to disentangle this situation.
#Haba:dáˤnàˤ-sà3
Vossen 1988: 106. The word is listed in its full form (with the feminine marker).
ǀGwi:dàˤnã̀3
Nakagawa 1996: 118. Quoted as dana in [Tanaka 1978: 56]. Different item listed in [Vossen 1988: 106]: ɡǀǀám̀.
Vossen 1988: 106; Vossen 1997: 425. In [Tanaka 1978: 56], however, the equivalent for 'leaf' is listed as dana, i.e. the exact same word as in ǀGwi. In [Nakagawa 2006: 240], two dialectal ǀǀGana forms for 'leaf' are also listed as dáˤnà and ʓáˤnà, respectively.
Dornan 1917: 94. Cf. also ɡǀanaː 'husk, peel of a fruit' [ibid.] - most likely, the same word, with a fictitious distinction "observed" by Dornan (a typical problem with the source).
Proto-Kalahari Khoe:*ɡǀana
Vossen 1997: 425 (*ɡǀana). Distribution: Preserved to a certain extent in all subbranches; reliably reconstructible on the Proto-Kalahari Khoe level. Replacements: Three forms that replace the original 'leaf' in Kalahari Khoe languages (Naro toàˤrà, ǂHaba-ǀGwi *daˤna, Kua-Tsua *tʰana) are all suspiciously similar but do not correspond regularly to each other. Additionally, the Naro form is practically identical with Juǀʼhoan (North Khoisan) dòàˤrà 'leaf', implying some sort of areal isogloss, possibly acquired from a third source that remains unknown (definitely not Bantu). Reconstruction details: The uncontracted form of the root is well preserved in the Eastern languages; in the Western area *ɡǀana > *ɡǀã, undergoing the same process as in Khoekhoe (where the original meaning shifted to 'grass'). This contraction is a special semi-regular development whose conditions are not yet fully understood, but it is recurrent throughout the etymological corpus.
Kilian-Hatz 2003: 203. Quoted as ǀǀoě in [Köhler 1981: 490]. Cf. the dynamic action verb ǀǀóé 'to lie down' [Kilian-Hatz 2003: 203], from which ǀǀòē is derived through a tonal shift (presumably, the result of contraction with the passive suffixal marker -ì). Cf. Buga-Khoe ǀǀóé 'to lie down' [Vossen 1997: 450].
Visser 2001: 115; Vossen 1997: 450. Polysemy: 'lie / go to bed / be in process of / travel, go for a long time'. Two tonal variants are listed: ǀǀóé and ǀǀōē 'to lie; to be'; quite possibly, it is the latter that conveys the static meaning 'to be lying', whereas the former is the dynamic 'to lie down' (cf. a similar situation in Kxoe), but this cannot be deduced with certainty from the available data. A. Barnard quotes the form ɳǂuː in the meaning 'lie' [Barnard 1985: 145], although this is really the verb 'to sit' q.v.
Dornan 1917: 97. Meaning glossed as 'to be asleep', not as 'to lie'. However, since the meaning 'to sleep' already corresponds in Dornan's vocabulary to a different lexical root (ɡǀǀom), it may be cautiously assumed, with the aid of external data, that 'to be asleep' here is actually = 'to lie, (to be asleep while) lying down'.
Proto-Kalahari Khoe:*ǀǀoe
Vossen 1997: 450 (*ǀǀoe). Distribution: Seemingly preserved in all dialects. Reconstruction shape: Correspondences are largely regular and trivial. Semantics and structure: Quite a few languages display the generic polysemy 'to lie / to sleep'; however, since there is an alternate strong candidate for the meaning 'to sleep' (*ǀǀʼom), it is likely that these polysemies represent a more recent development.
Nakagawa 1996: 116. Quoted as kʼã́ĩ́ in [Vossen 1997: 462]; as !kaèn in [Tanaka 1978: 58] (= !ãẽ, with the velar ejective kʼ mistaken for a click).
Vossen 1997: 462 (*kxʼãĩ). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are somewhat unique, particularly for the Kua-Tsua subgroup where we witness a very rare case of palatalization; it should be noted that this very clearly distinguishes the root from *xʼãĩ 'to laugh' (> Kua, Tsua kʼã́ĩ̀), so Vossen's reconstruction *kxʼãĩ for both of these items is not justified (no explanation why -a- is lost in 'liver' and not in 'laugh'). We provisionally amend the reconstruction to *xʼẽĩ, assuming contraction *-ẽĩ > *-ĩ and subsequent palatalization in Kua-Tsua and dissimilation elsewhere, but this is not certified by additional parallels.
Visser 2001: 63; Vossen 1997: 462. Polysemy: 'long / tall'. According to A. Barnard, the Naro word for 'long' is the same as 'big' (káí q.v.), but this is dubious in the light of the rest of the sources [Barnard 1985: 118].
Dornan 1917: 103. Meaning glossed as 'large, long'. The palatal click may be mistranscribing the alveolar click (cf. Kua and Tsua).
Proto-Kalahari Khoe:*!ao
Vossen 1997: 461 (*!ao). Distribution: Preserved everywhere where attested. Reconstruction shape: Correspondences are generally regular, although the alveolar click, as usual, displays a rather chaotic pattern of behavior. In this case, particular suspicion is drawn to the Kua-Tsua forms, since these languages, like all languages of the Shua group, regularly drop the alveolar click (*kao should be expected); this is one of very few examples (see also 'white') where it is preserved, implying that the actual forms might have been reborrowed from a West Khoe source (but which one?).
Visser 2001: 55. Polysemy: 'louse / flea'. Quoted as ɳǂâˤ in [Barnard 1985: 45]; the latter source also lists ɳǀani as a synonym, but this word is not confirmed in Visser's dictionary. Cf. also ǀǀxʼóné 'very small louse; something very small' [Visser 2001: 113].
Vossen 1997: 462 (*kxʼuni). Distribution: Preserved in most languages. Replacements: (a) Naro ɳǂã̀ːˤ, of unknown origin; it is clearly the same word as Juǀʼhoan ɳǂàˤ 'louse', but the latter, likewise, is only restricted within North Khoisan to the Juǀʼhoan dialect, so it seems to belong to the small group of areal Naro-Ju isoglosses (see 'leaf') that may have a substrate origin; (b) Hiechware ɡǀǀam is not supported by any source outside of Dornan's old wordlist; it might be a lexical or semantic mistake, or a genuine replacement of unknown origin. Reconstruction shape: Correspondences are generally regular and trivial.
Vossen 2000: 136. Same stem as 'person', specified with the regular masculine gender morpheme -mà. Should be distinguished from xʼáò 'male, manly' [Vossen 1997: 465].
Visser 2001: 41. Same word as 'person' q.v. with the masculine gender marker. A more specific way to denote 'man' as 'male human being' is xʼáò-kʰōè [Visser 2001: 38], literally: 'male person'. The first component here is xʼáò 'male' [Vossen 1997: 465], the second is kʰoe 'person' q.v.; quoted as kʼáù-kʰùè in [Barnard 1985: 70]. However, text examples show that this compound is rarely, if ever, used in basic contexts; its usage must be restricted to specially marked situations.
Vossen 1997: 465. Meaning glossed as 'male'. In [Tanaka 1978: 60], the meaning 'man (male person)' is rendered as xʼao-kue, literally 'male' + 'person' q.v. It is not clear how basic this compound is, therefore, the whole entry is dubious.
Vossen 1997: 465. Meaning glossed as 'male'. In [Tanaka 1978: 60], the meaning 'man (male person)' is rendered as xʼao-kue, literally 'male' + 'person' q.v. It is not clear how basic this compound is, therefore, the whole entry is dubious.
Dornan 1917: 101. Most likely, a compound of kau 'male' (cf. external data) + čo 'person' q.v.; cf. the similar structure of 'woman' q.v.
Proto-Kalahari Khoe:*xʼao
Vossen 1997: 465 (*kxʼao). Distribution: Preserved in all languages, although see the discussion on semantics below. Reconstruction shape: Correspondences are generally regular and trivial. Semantics and structure: It seems that the original meaning of the morpheme *xʼao is 'male' (applicable to animals and people); the meaning 'man' is perceived as more complex than 'male' and is often expressed by either the masculine gender form of the lexeme *kʰoe 'person' (as in ǀǀAni and Naro) or the compounding of two lexical roots (*xʼao-kʰoe, as in Kxoe and Hiechware). Nevertheless, since 'man'-meaning forms with *xʼao are far more frequent than 'man'-meaning forms without this morpheme, it seems justified to put this morpheme into the main Proto-Kalahari Khoe slot for 'man', leaving *kʰoe in the slot for 'person'. See an absolutely symmetric picture for the item 'woman' q.v.
Number:52
Word:many
ǀǀAni:
Not attested.
ǀGanda:
Not attested.
Kxoe:kó ~ á=kó1
Kilian-Hatz 2003: 16, 58. The forms represent adjectival or adverbial usage of the noun kó 'quanitty, (large) number, multitude'; multiple textual examples confirm that this is indeed the main equivalent of this Swadesh meaning in Khoe. Secondary synonym: tʰíyà, with polysemy: 'be much, be many / be four' [Kilian-Hatz 2003: 133; Köhler 1981: 493] (no textual examples).
Visser 2001: 34. Meaning glossed as 'much, many, more'. Same stem as kāī-ā 'big' q.v., but with different tonal characteristics. A. Barnard does not notice any tonal differences between 'big' and 'many', glossing both meanings as káí [Barnard 1985: 115].
#Haba:
Not attested.
ǀGwi:
Not attested in reliable sources. Cf., however, !tui 'many' in [Tanaka 1978: 60] (which could transcriptionally reflect *ǂui, in which case it would be comparable with Nama ɡǂui 'many').
ǀǀGana:
Not attested in reliable sources. Cf., however, !tui 'many' in [Tanaka 1978: 60] (which could transcriptionally reflect *ǂui, in which case it would be comparable with Nama ɡǂui 'many').
Cara:
Not attested.
ǀXaise:
Not attested.
Danisi:
Not attested.
Ts'ixa:
Not attested.
Deti:
Not attested.
Kua:
Not attested.
Tsua:
Not attested.
Hiechware:cao3
Dornan 1917: 111.
Proto-Kalahari Khoe:
Not reconstructible due to lack of attestation in known sources; the few attested forms are not reconcilable with each other and suggest that the concept is generally quite unstable in Kalahari Khoe.
Vossen 1988: 89. The root here is verbal (xʼò 'to eat meat'); the noun 'meat' is formed with the productive nominalizer -xú (grammaticalized noun 'thing'). More specialized nouns like ǀxá 'muscle; flesh (without bones)' [Vossen 1986: 329] and kání 'cooked meat' [Vossen 1986: 338] are ineligible.
Visser 2001: 39. Meaning glossed as 'meat (with bone) - only used when speaking of animals'. Quoted as xʼó in [Vossen 1997: 433]; as xʼòː ~ kʼòː (com.), also kʼò-xó (com.; cf. the same derivation in Kxoe) in [Barnard 1985: 98] and xʼò-xó in [Vossen 1988: 89]. With tonal gradation, the same root is also used as verbal: xʼōː 'to eat meat'. The meaning 'flesh', 'meat (without bone)', 'body' is said to correspond to the word ǀxàā [Visser 2001: 6].
Dornan 1917: 104. For the morphological structure, see notes on the ǀǀAni entry.
Proto-Kalahari Khoe:*xʼo-xu
Vossen 1997: 433 (*kxʼo 'to eat meat'). Distribution: Preserved in most languages, apparently except for the ǀGwi-ǀǀGana subgroup. Replacements: Merged in ǀGwi-ǀǀGana with *ǀxa 'body; muscle' [Vossen 1997: 468]; note also the distinction in Naro between 'meat (with bone)' and ǀxàā 'meat (boneless)'. Reconstruction shape: Correspondences are regular. Semantics and structure: It seems that the original meaning of the root *kxʼo was verbal ('to eat hard food, eat meat'), since in most cases the nominal meaning 'meat' is expressed by the compound formation *xʼo-xu, lit. 'eat-meat-thing'.
Dornan 1917: 96. The dental click is most likely a transcriptional mistake for the lateral click, as in 'nail' q.v. Cf. also !owe 'a month, the moon' [Dornan 1917: 104] - possibly the same word, mistranscribed in a different way.
Proto-Kalahari Khoe:*ɳǀǀoɛ
Vossen 1997: 468 (*ɳǀǀnoe ~ *ɳǀǀoe). Distribution: Preserved in all languages. Reconstruction shape: The diphthong *-oɛ (rather than *-oe) is preserved as such in numerous languages.
Number:55
Word:mountain
ǀǀAni:
Not attested.
ǀGanda:
Not attested.
Kxoe:ndúndù-1
Kilian-Hatz 2003: 93. The word violates normal Kxoe phonotactics and looks like a Bantu borrowing. Cf. also ɡǀú-ǂʼám̀ ~ ɡǀú-tá-ǂʼám̀ 'slope, hill, mountain' [Kilian-Hatz 2003: 163], where ɡǀú = 'small hill in a plain' and ǂʼám̀ = 'top (side)'; judging by this structure, the meaning 'mountain' is far-fetched, and the expression should rather mean 'mountain top' or 'slope (space from mountain top to bottom)'.
Visser 2001: 106. Polysemy: 'hill / slope / mountain'. Quoted as ǀǀábí (fem.) in [Barnard 1985: 34]; the other synonym listed in the latter source, ǀau ~ cau, is not confirmed in more reliable sources.
#Haba:
Not attested.
ǀGwi:
Not attested in reliable sources. Cf., however, !kabì 'mountain' in [Tanaka 1978: 63].
ǀǀGana:
Not attested in reliable sources. Cf., however, !kabì 'mountain' in [Tanaka 1978: 63].
Cara:
Not attested.
ǀXaise:
Not attested.
Danisi:
Not attested.
Ts'ixa:
Not attested.
Deti:
Not attested.
Kua:
Not attested.
Tsua:
Not attested.
Hiechware:ɡ!oa2
Dornan 1917: 95.
Proto-Kalahari Khoe:
Not reconstructible due to lack of sufficient attestation. The only possible candidate would be Naro ǀǀàbì and its probable cognates in ǀGwi and ǀǀGana (poorly transcribed in Tanaka's dictionary), but more data would be needed to confirm the archaicity of these items.
Vossen 1997: 468. The older source [Vossen 1988: 90] lists an additional synonym as cʼùm 'mouth'. In [Vossen 1997: 464], however, this word is listed as cʼúḿ and more correctly glossed as 'upper lip'.
Vossen 1997: 468. The word cʼóm, listed in the earlier source [Vossen 1988: 90] as a synonym with the same meaning 'mouth', really means 'upper lip' [Vossen 1997: 464].
Dornan 1917: 100. Polysemy: 'beak / mouth'. The palatal click transcription may actually represent a non-click (ejective) consonant in this case. The same word is probably transcribed with a dental click (also mistakenly) as ǀxam in [Dornan 1917: 102].
Proto-Kalahari Khoe:*xʼam
Vossen 1997: 468 (*kxʼam). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are regular and trivial.
Vossen 1997: 471 (*ǀxʼon). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are mostly regular, but some of the languages clearly reflect a bisyllabic variant (*ǀxʼon-i ~ *ǀxʼun-i), usually with vocalic assimilation and contraction (> *ǀxʼũĩ). Since CVCi-type structures do not normally behave this way in Kalahari Khoe, it is possible that this variant goes back to an archaic common gender form (*ǀxʼon-ʔi) with a specific development on the morphemic border.
Nakagawa 2006: Fig. 3.24. Quoted as ǀǀxʼàò-ǀõ̀ã̀ in [Vossen 1988: 91] (the second component is probably a misspelling for ǀʼṍã́ 'bone'); as !xʼao in [Tanaka 1978: 64].
Dornan 1917: 101. Initial ǂ-, as in numerous similar cases, may actually reflect non-click articulation. Cf. also dʰom 'neck' [Dornan 1917: 92] as a potential secondary synonym; however, all the external relatives of this form have the semantically close meaning 'throat' rather than 'neck', so Dornan's translation may be put under doubt. Moreover, the meaning 'throat' is rendered as dom [ibid.] - obviously the same word, with a fictitious consonantal opposition.
Proto-Kalahari Khoe:*!xʼao
Vossen 1997: 446 (*!xʼao). Distribution: Represented in every single subbranch and clearly the optimal candidate for Proto-Kalahari Khoe 'neck'. Replacements: (a) ǀGanda-Kxoe *xɛri, possibly related to Nama xari-s 'gizzard'; (b) Danisi-Tsʼixa *ʓanu < *ɡǂanu, of unknown origin; (c) Tsua kínù, of unknown origin. Reconstruction shape: Correspondences are regular and trivial.
Dornan 1917: 101. The word is easily comparable with its equivalent in Tsua, if the alveolar click symbol here mistranscribes a lateral click (cf. 'fat', 'to lie' with identical situations). Alternatively, cf. also ǀxaie 'new' [Dornan 1917: 102].
Proto-Kalahari Khoe:*xʼoa
Vossen 1992: 471 (*kxʼoa). Distribution: Attested in ǀǀAni-Kxoe and the Shua subgroup. Replacements: (a) Naro + ǀGwi-ǀǀGana have *kaˤba (> ǀGwi-ǀǀGana *qaba) as the main equivalent for 'new'; the word is clearly related to Nama kawa 'new' and could, therefore, by way of external cognate argument be eligible for the main slot; however, *kxʼoa has far stronger internal distribution, implying that the Nama-Naro-ǀǀGana isogloss might really be areal; (b) Tsua ɳǀǀáò, glossed by Vossen with the meaning 'new', is clearly the same word as Proto-Kalahari Khoe *ɳǀǀao 'old' (!) - no hypotheses as to the reasons of such an odd meaning reversal. Reconstruction shape: Correspondences are regular and trivial.
Visser 2001: 57; Barnard 1985: 24. Polysemy: 'night / evening / black'. Same word as 'black' q.v. Quoted as ɳǂnǔ in [Vossen 1997: 470].
#Haba:
Not attested.
ǀGwi:
Not attested in reliable sources. In [Tanaka 1978: 65], the meaning 'night' is rendered as hxae-šika, where the first part = hxaè 'dark' [Tanaka 1978: 26] and the second may constitute a string of derivational suffixes, but whether this is indeed the basic word for 'night' remains to be confirmed.
ǀǀGana:
Not attested in reliable sources. In [Tanaka 1978: 65], the meaning 'night' is rendered as hxae-šika, where the first part = hxaè 'dark' [Tanaka 1978: 26] (= qáè 'dark' [Vossen 1992: 386]) and the second may constitute a string of derivational suffixes, but whether this is indeed the basic word for 'night' remains to be confirmed.
Dornan 1917: 97. The same stem is probably present in xaie ka 'night' [Dornan 1917: 102].
Proto-Kalahari Khoe:*tʰu
Vossen 1997: 470 (*tʰǔ). Distribution: Well represented across all subbranches. Replacements: (a) Naro ɳǂǔ is the same word as 'black', with a secondary polysemous development {'black' > 'night'}; (b) Hiechware haie is etymologically related to ǀǀGana qáè, Nama !xae- 'darkness'; the same root is probably involved in compound formations with the meaning 'night' in ǀGwi-ǀǀGana, but its original meaning was most likely 'darkness' in general rather than 'nighttime' in particular, so the developments in Hiechware and ǀGwi-ǀǀGana are easily interpretable as independent trivial semantic shifts {'dark' > 'night'}.
Vossen 1997: 471 (*ǂui). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are regular (with expected click-to-affricate development in the Eastern branch).
Kilian-Hatz 2003: 19, 138. Quoted as βéː in [Köhler 1981: 535]. Enclitic particle that expresses the basic meaning of negation next to indicative forms of the verb. Cf. Buga-Khoe =bé id. in [Vossen 1997: 203].
Visser 2001: 78, 79; Vossen 1997: 214. The morpheme behaves as a verbal enclitic; distribution between the two variants remains unclear. This negation is employed in present and past tenses; for the future tense, the default negation is títè [Visser 2001: 96; Vossen 1997: 214]. Morphological connections between tá, tāmā, and títè on the synchronic level are essentially of a suppletive nature, but it is almost certain that all three are historically related.
Dornan 1917: 67. Basic verbal enclitic, employed in present tense forms. The corresponding variant for the past tense is =tʰa.
Proto-Kalahari Khoe:*=tama
Vossen 1997: 366 (*tama). Distribution: Found in all the subgroups except for ǀǀAni-Kxoe. Replacements: See 'Semantics and Structure'. Reconstruction shape: We follow Vossen in his conjecture that the variant -ta is historically contracted from *-tama (cf. Tsʼixa -tã where traces of the original nasal consonant are still seen in vowel nasalization), since the bisyllabic variant is confirmed externally (= Nama tama) and there is no explanation of *-ma as an additional suffix with its own meaning and function. Semantics and structure: The morpheme *=ta(ma) is the only one that can be reliably reconstructed for the Proto-Kalahari Khoe level as a basic negation suffix. Most of the languages feature a binary opposition between present-time and past-time negation, but the realization of this opposition differs between languages. One possible scenario is that the original system was similar to the one seen in Danisi, with *=ta(ma) as the present tense negation marker and *=be as the past tense marker; later on, *=be was generalized in the ǀǀAni-Kxoe subgroup, *=ta(ma) was generalized (almost) everywhere else and innovative past tense markers independently sprang up in a few other languages, e.g. Naro and Tsʼixa. This scenario seems somewhat likely since there are no obvious sources for the innovation of *=be in ǀǀAni-Kxoe; conversely, the situation in Naro, for instance, could at least theoretically be explained as an areal innovation under the influence of Khoekhoe. Nevertheless, for the moment we prefer to restrict protolanguage synonymy to *=tama alone, for reasons of distribution.
Vossen 1997: 429 (*ǀúí). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are regular and trivial (with the exception of sporadic click loss in Hiechware, provided the old transcription is accurate).
Visser 2001: 41; Vossen 1997: 409. Quoted as kʰúè ~ kʰóè in [Barnard 1985: 70]. Barnard: "This term is sometimes used specifically for Nharo or Bushman people, in opposition to other ethnic groups".
Dornan 1917: 90, 91. Cf. also čowe 'man' [Dornan 1917: 91] - most likely, the same word.
Proto-Kalahari Khoe:*kʰoe
Vossen 1997: 467 (*kʰoe). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are regular (with expected palatalization in Kua-Tsua, though the distribution of palatalized variants across dialects is somewhat sporadic, as in other similar cases).
Nakagawa 1996: 110. The word is not listed in the comparative monograph [Vossen 1997]; in [Vossen 1986: 93], it is quoted as gyú, and it is assumed by Vossen that this word is not related to tu ~ tyu ~ ɕu in other languages. However, the sequence gy really reflects old dental stops as well, and it is quite unlikely that ǀǀGana has a different root (*du ?) here. Quoted as kyuː in [Tanaka 1978: 77] (note also the use of velar k to transcribe the palatal stop).
Vossen 1997: 477 (*tu). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are regular (palatalized reflexes in ǀGwi-ǀǀGana are not always attested for original *t-, but frequently enough to suggest that the transcription simply reflects idiosyncratic articulations of a single consonant that fluctuates between coronal and palatal variants).
Kilian-Hatz 2003: 219; Köhler 1981: 512. A verbal / adjectival derivative from ǀǀqéú 'red colour, reddish colour / sunset'. Cf. also the secondary synonym: ǀōá 'be clean / be red / be holy'. Since this word does not participate in any derivatives with the semantics of 'red', it is dubious that it continues to serve as the main equivalent for 'red' in Kxoe proper. Cf., however, Buga-Khoe ɳǀóá, glossed as 'red' in [Vossen 1997: 419].
Vossen 1997: 479 (*ɳǀoa). Distribution: Reflexes are found in all subgroups. Replacements: (a) In ǀGanda and Kxoe, the original item has been replaced by the stem *ǀǀqau, originally meaning 'sunset' or 'sunset glow' (Vossen 1992: 385), with the semantic shift {'sunset' > 'red'}; (b) in Tsʼixa-Danisi, replaced by *qa-si of unknown origin; (c) in Kua-Tsua, replaced by *ʔudu (possibly from *!ʼuru or *ǀǀʼuru), also of unclear origin. Reconstruction shape: There are two unusual things with the reflexes of this stem - Naro and ǂHaba show a pharyngealized vowel, and the Eastern subgroup shows no traces of the nasal efflux. There is not enough data to see if these two phenomena are related (e.g. denasalization of the click in East Kalahari Khoe due to the influence of pharyngealized vocalic articulation?), but in any case, both the nasal click and the pharyngealization probably have to be projected onto the proto-level.
Number:66
Word:red
ǀǀAni:
ǀGanda:
Kxoe:
Naro:
#Haba:
ǀGwi:
ǀǀGana:
Cara:
ǀXaise:
Danisi:qâ-sè3
Vossen 1988: 94. This word is listed in the same meaning 'red' as ǀóá, without any notes on possible semantic differentiation. We include both forms as technical synonyms.
Nakagawa 1996: 110. Quoted as díó in [Vossen 1997: 494]; as gyô in [Tanaka 1978: 80] (both transcriptions are imperfect attempts to represent a voiced palatal stop).
Vossen 1997: 494. In the meaning 'path', quoted as a compound form: dáò-ǀõ̀ã̀ in the earlier source [Vossen 1988: 92] (ǀõ̀ã̀ = lit. 'child', i. e. 'little').
Vossen 1997: 493 (*dao). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are regular and generally trivial (except for some scattered palatalized reflexes /*d > ʓ/ in some languages).
Visser 2001: 96. Polysemy: 'root / quiver (which is made from a root)'. Quoted as tʼǒbé ~ tʼǒˤbéˤ́ ~ tʼõ in [Barnard 1985: 47]; distinct from ǂkabi 'terminal root' [ibid.].
Tanaka 1978: 80. Although this is not a reliable source, we cautiously include the word, since it is clearly the same as in Naro and other West Khoe languages, although the exact phonetic transcription may be seriously off. Cf. also !xʼáí 'tree root' in [Nakagawa 1996: 140].
Tanaka 1978: 80. Although this is not a reliable source, we cautiously include the word, since it is clearly the same as in Naro and other West Khoe languages, although the exact phonetic transcription may be seriously off.
Cara:
Not attested.
ǀXaise:
Not attested.
Danisi:
Not attested.
Ts'ixa:
Not attested.
Deti:
Not attested.
Kua:
Not attested.
Tsua:
Not attested.
Hiechware:papasi #-1
Dornan 1917: 108. Meaning glossed as 'a root, a branch'. Semantics is dubious; in any case, the phonetic structure of the word clearly betrays a non-Khoe origin.
Proto-Kalahari Khoe:*toˤbe #
Distribution: Only attested in Western languages; information on most Eastern languages is lacking (Hiechware papasi is a highly dubious entry). Nevertheless, since there are no other candidates and no clear indication of the secondary origin of this root, it may be tentatively projected onto the top level. Reconstruction shape: Correspondences are not fully regular; standard *tobe can hardly develop into *toɛ in Kxoe, but one could suggest a "leniting influence" if the original vowel was pharyngealized, as in Naro (*toˤbe). Even that does not explain root vowel -a- in attested ǀGwi-ǀǀGana forms, but these come from a phonetically unreliable source (Tanaka's dictionary) and need to be double-checked.
Number:69
Word:round
ǀǀAni:
Not attested.
ǀGanda:
Not attested.
Kxoe:mbírì-1
Kilian-Hatz 2003: 87. Probably of Bantu origin (initial mb- is not allowed in "native" Kxoe words). Cf. also tʰínì 'to roll (smth.), turn inside off', tʰínì-ʆān 'to turn; to be round' [Kilian-Hatz 2003: 132].
Kilian-Hatz 2003: 108. Polysemy: 'sand / be sandy (vb.)'; cf. also qáń ~ qã́ĩ́ 'to bury' [ibid.], possibly derived from the same stem. Quoted as qǎn 'sand' in [Köhler 1981: 501]. The word xóḿ 'earth' q.v. may also be used in the meaning 'sand', but this is not the primary meaning.
Vossen 1997: 425 (*xom). Distribution: Attested in all daughter languages, usually with the meaning 'earth', more rarely with polysemy 'earth / sand'. Replacements: Kxoe qàn̄ goes back to the verbal root *qan 'to cover with (ashes / sand)' and seems to be a recent innovation, introducing a lexical differentiation between 'earth' and 'sand'. Reconstruction shape: Correspondences are regular and trivial. Semantics and structure: Reconstructible with polysemy: 'earth / sand'.
Vossen 1997: 481. A textual example in [Vossen 2000: 139] suggests ɳǀǀnâ as a synonym: cá náɳǀǀnâ-tà 'what did you say?' However, [Vossen 1997] only quotes the meaning 'to tell, narrate' for this latter verb, which is also its unquestionable etymological meaning.
Kilian-Hatz 2003: 84; Köhler 1981: 498. Polysemy: 'make a sound, make noise / say'. In the English-Khwe index, Kilian-Hatz also attaches the meaning 'say' to such verbs as ǀʰĩ̂ [Kilian-Hatz 2003: 164] and ɳǀǀâ [Kilian-Hatz 2003: 215], but these are really transitive predicates meaning 'to say, tell (smth.), inform of (smth.)', less eligible for inclusion.
Visser 2001: 46. Quoted as mĩ́ in [Vossen 1997: 481]; as míː ~ mĩ́ː in [Barnard 1985: 152]. Two other synonyms listed by Barnard, kira and tame, are not confirmed in other sources.
Tanaka 1978: 82. Although this is not a reliable source, the entry is very likely to be correct, due to external matches; phonetic transcription, however, may be off (e. g. the actual pronunciation may be mi, as in the other languages).
Tanaka 1978: 82. Although this is not a reliable source, the entry is very likely to be correct, due to external matches; phonetic transcription, however, may be off (e. g. the actual pronunciation may be mi, as in the other languages).
Dornan 1917: 106. Meaning glossed as 'to say to'; cf. also me-he 'to say to oneself'. Alternately, cf. kwa 'to say' [Dornan 1917: 105] or a tʰeː tʰama 'to say to' [Dornan 1917: 89]. We concentrate on me (tentatively) because of external data.
Proto-Kalahari Khoe:*mĩ
Vossen 1997: 480 (*mĩ). Distribution: Preserved everywhere where attested (for many languages, the exact equivalent for 'to say' is lacking or unclear). Reconstruction shape: Articulation of the vowel fluctuates between simple *i and nasal *ĩ, sometimes even within the same language (cf. Naro); the nasal variant seems to be confirmed by external data (Khoekhoe), but influence of the initial nasal consonant cannot be excluded as a factor either.
Visser 2001: 3. Quoted as mô in [Vossen 1997: 490]; as m ~ mû in [Barnard 1985: 152]. Of all the sources, only Visser's dictionary lists the irregularly denasalized variant with b-; it also gives a special variant mâ [Visser 2001: 45], said to be used exclusively by older people - unclear if it is etymologically related to mô.
Dornan 1917: 106. Meaning glossed as 'to see, appear, sight'.
Proto-Kalahari Khoe:*mũ
Vossen 1997: 490 (*mũ). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are generally regular, except for the lack of nasalization and contraction (*mũ > mu > m) in several Western languages (possibly triggered by the rare phonotactic structure of the item).
Kilian-Hatz 2003: 171. Meaning glossed as 'seed (corn); pip (e. g. of lemon)'. Secondary synonym: bótʰò ~ bòtʰó 'seed' (a recent Bantu borrowing). Cf. Buga-Khoe ǀxùí id. [Vossen 1997: 480].
Dornan 1917: 90. An obvious compound, where the last morpheme quite likely transcribes a misheard *ǀxuri, i.e. reflects the proper Khoe root for 'seed'. The first morpheme, čai = 'eye' q.v.; the second remains unidentified, but may be the same as in ča-kai 'eyes (pl.)' (where kai is, however, not a productive plural morpheme). The whole expression is thus, literally, 'eye-smth.-seed' or 'eyes-seed'. In fact, the "proper" word ǀxuri is also attested, under the meaning 'a grain of seed', in [Dornan 1917: 103].
Proto-Kalahari Khoe:*ǀxuri
Vossen 1997: 480 (*ǀxudi). Distribution: Preserved in most languages. Replacements: In ǀXaise, allegedly replaced by a reflex of Proto-Khoe *coro 'shell, pod' [Vossen 1997: 483]; provided Vossen's semantic glossing is accurate, we should postulate the semantic shift {'shell' > 'seed'}. Reconstruction shape: Correspondences are mostly regular and trivial, except for the irregular elision of *-r- in Kxoe (unexplained).
Kilian-Hatz 2003: 190. Derived from the dynamic action verb ɳǂṹ 'to sit down' by means of the passive/intransitive suffix -i. Quoted as ɳǂnũ̀ĩ́ in [Köhler 1981: 496]. Cf. Buga-Khoe ɳǂnú 'to sit down' [Vossen 1997: 491].
Visser 2001: 56. This is the static verb ('to be sitting'), regularly derived from the dynamic action verb ɳǂṍː 'to sit down' [ibid.]. The dynamic verb is quoted as ɳǂnú ~ ɳǂnṹ in [Vossen 1997: 491] and as ɳǂǔ ~ ɳǂú in [Barnard 1985: 154].
Vossen 1997: 491. The meaning is given as 'to sit down' (there is hardly any doubt that the static verb 'to sit' is formed from the same root). Quoted as ɳ!(u)o in [Tanaka 1978: 87] (with incorrect identification of the click influx).
Vossen 1997: 491. The meaning is given as 'to sit down' (there is hardly any doubt that the static verb 'to sit' is formed from the same root). Quoted as ɳ!(u)o in [Tanaka 1978: 87] (with incorrect identification of the click influx).
Vossen 1997: 491 (*ɳǂ/n/u ~ *ɳǂ/n/ũ). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are generally regular (including click loss in Eastern languages), although fluctuations between nasal and non-nasal articulation are unpredictable. External parallels in Khoekhoe suggest that the nasal vowel is primary. Semantics and structure: The simple morpheme *ɳǂũ clearly had a dynamic meaning ('to sit down'); the static meaning was most likely denoted by combining the root morpheme with a vocalic suffixal extension, such as in Kxoe or Naro. However, since the suffixes in Kxoe and Naro are actually different, and since the static form itself is very rarely attested in wordlists for the other languages, it is not easy to determine the exact shape of the stem in Proto-Kalahari Kxoe.
Kilian-Hatz 2003: 65. Polysemy: 'skin / hide / leather / blanket made from hide'. Quoted as kʰò in [Köhler 1981: 554]. Secondary synonym: ǀǀèí 'skin; blanket' [Kilian-Hatz 2003: 201] (semantic difference unclear, the word seems to be in less frequent usage, judging by the contexts). Cf. Buga-Khoe kʰǒ id. [Vossen 1997: 448].
Vossen 1997: 448. The same word is listed as cʰǒ in the earlier source [Vossen 1988: 95]; this palatalized variant seems to be the regular expected development, so the non-palatalized variant in [Vossen 1997: 448] may be a mistake.
Dornan 1917: 91. Cf. also ǯoruː 'the skin' [Dornan 1917: 98], a form that could possibly be connected to ǯoreː 'bark' q.v., although the vocalic alternation remains unexplained.
Proto-Kalahari Khoe:*kʰo
Vossen 1997: 448 (*kʰo). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are regular and mostly trivial (Hiechware čo reflects sporadic palatalization that is frequently observed in various Kua-Tsua dialects).
Vossen 2000: 141, 142. The basic (etymological) meaning of the word is 'to lie down', but R. Vossen also gives the translation 'sleep' on p. 142; also the same meaning in [Vossen 1988: 96].
Dornan 1917: 96. Alternatively, cf. ǀoe 'to sleep' [Dornan 1917: 104], which is probably the same word as 'to lie' q.v.; !we 'to sleep' [Dornan 1917: 105] may be yet another mistranscribed variant of the same word.
Proto-Kalahari Khoe:*ǀǀʼom
Vossen 1997: 484 (*ǀǀʼom). Distribution: An unstable item, but reflexes are found in all subgroups (Western and Eastern), ensuring that it is reconstructible in the required meaning for Proto-Kalahari Khoe. Replacements: (a) Many languages seem to merge the meanings 'to lie' and 'to sleep' within a single root *ǀǀoe (see under 'to lie'); given its wide spread, this synonymy may have already been partially active in Proto-Kalahari Khoe, but the separate root for 'to sleep' was clearly still present at that stage; (b) An exclusive situation is found in Kxoe, where it seems that the original equivalent for 'sleep' (> Kxoe ǀǀʼóm̀) has become specialized in the meaning 'to perch at dusk, roost (of bird)' [Kilian-Hatz 2003: 208]. The replacement ǂʼóm, despite phonetic similarity, represents an entirely different root, cognate with Naro ǂʼōm 'to be sleepy, drowsy' [Visser 2001: 92] and, further on, Nama ǂʼòm̋-s 'sleep (n.)'; based on the semantic distinction in Naro, we may suggest a semantic shift {'to be sleepy, drowsy' > 'to sleep'} in Kxoe. Reconstruction shape: Correspondences are generally regular and trivial between all languages in which the original root has been preserved.
Visser 2001: 9. Quoted as ǀxʼáré in [Vossen 1997: 457]; as ǀxáré ~ ǀxáːré in [Barnard 1985: 118]. The probable suffixal origins of -ré are betrayed by such semantically close entries as ǀxʼā-nà 'useless, old, without value; small thing (something despisable)' and ǀxʼāī-ǀxʼáì 'very small' [Visser 2001: 8].
Vossen 1988: 97. Somewhat dubious, since this is the same word as 'short' q.v.; furthermore, in [Tanaka 1978: 88], 'small' is rendered as ǀare, e. g. the same word as in ǀGwi. This entry may require amendment.
Cara:ɕṹĩ́4
Vossen 1988: 97.
ǀXaise:ɕṹĩ́4
Vossen 1988: 97.
Danisi:ʔòré5
Vossen 1988: 97. The later source [Vossen 1997: 457] lists Danisi ǀʼáré in the meaning 'small', but the same source also indicates that in all of its close relatives, the word means 'thin' or 'narrow'; it is highly probable that the Danisi semantics of this word also deviates from the basic meaning 'small'.
Dornan 1917: 97. Cf. also kare 'a little' (adverbial form?) in [Dornan 1917: 101].
Proto-Kalahari Khoe:*ǀxʼare
Vossen 1997: 457 (*ǀxʼade). Distribution: This root is the basic equivalent for 'small' in Naro, ǂHaba, and ǀGwi, as well as in Kua-Tsua, which makes it the only word for 'small' to be found in both Western and Eastern languages and, consequently, the optimal candidate for 'small' on the Proto-Kalahari Khoe level. In several Eastern languages the word is only attested in related meanings such as 'thin' or 'short' [Vossen 1997: 457], but these are likely secondary (or, in the case of 'thin', may reflect a proto-language level polysemy). Replacements: At present, none of the equivalents for 'small' within separate dialectal clusters have any reliable external etymologies: ǀǀAni-Kxoe *kũ (possibly < *!ũ), Shua *ɕũĩ (possibly < *ǂũĩ or *ǂuni), and Tsʼixa-Danisi *ʔore (possibly < *!ʼore) remain unexplained. Reconstruction shape: Correspondences are regular.
Vossen 1997: 476. Polysemy: 'smoke / tobacco'. (Actually, only quoted by R. Vossen in the meaning 'tobacco', but no other word for 'smoke' has surfaced so far).
Visser 2001: 104. Used both as a noun and a verb ('to smoke /of fire/'). Quoted as cʼínì in [Vossen 1997: 476]; as cénè 'to smoke (of fire)' in [Barnard 1985: 109]. In the substantive meaning 'smoke', however, the latter source only lists the noun ǀxoru [Barnard 1985: 33, 109] - a highly dubious inclusion, probably the same as ǀʰúrú 'be in need of tobacco, long to smoke' [Visser 2001: 11], clearly ineligible for inclusion.
Dornan 1917: 111. Meaning glossed as 'flame, smoke'.
Proto-Kalahari Khoe:*cʼani
Vossen 1997: 476 (*cʼani). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are generally regular, except for some unpredictable fluctuations between vocalic reflexes in the first syllable (because of the assimilative influence of *-i in the second one).
Visser 2001: 93. Static verb ('to be standing'), distinct from the dynamic action verb tẽ̀ẽ̄ 'to stand up' [ibid.]. Only the latter is listed in [Barnard 1985: 155] as tẽ́ ~ tĩ̌ 'stand, stand up'.
#Haba:
Not attested; cf. the dynamic equivalent tìń 'to stand up' in [Vossen 1997: 418].
Nakagawa 1996: 115. Quoted as té in [Vossen 1997: 495]. Cf. ɕěn 'to stand' [Nakagawa 1996: 108] (the actual meaning is most likely dynamic: 'stand up') = tàń 'stand up' [Vossen 1997: 418] = kíyan 'stand' [Tanaka 1978: 91] (kiy- here is an attempt to transcribe a palatal stop).
Vossen 1997: 495 (*tai). Distribution: Preserved everywhere where attested. Reconstruction shape: Correspondences are generally regular and indicate *ɛ rather than *e as the original vowel (see notes on structure). Semantics and structure: Vossen's reconstruction *tai presumes a morphemic border between the two vowels (*ta-i), confirmed by the corresponding dynamic action stem *ta-n as well as the symmetric paradigm for the verb 'to sit' (*ɳǂũ 'to sit down' vs. *ɳǂũ-i 'to be sitting'). Combination of root morpheme *ta- + intransitive suffix *-i results in morphemic contraction *ta-i > *tɛ already on the Proto-Kalahari Khoe level (so Vossen's reconstruction is actually "internal").
Visser 2001: 89. Quoted as ǂóˤnù ~ ǀōnù in [Barnard 1985: 33] (the click efflux is identified incorrectly, and the second variant with the dental click is also probably erroneous).
Dornan 1917: 102. The entry ǀǀala, glossed as 'the heavens, the stars' in [Dornan 1917: 100], is too semantically ambiguous (and completely unconfirmed by external data) to be included on the main list.
Proto-Kalahari Khoe:*ǀxani
Vossen 1997: 496 (*ǀxani). Distribution: Preserved everywhere except for Naro-ǂHaba and ǀGwi-ǀǀGana. Replacements: In a large part of the Western Kalahari Khoe area, the original root has been replaced by an innovation of unclear origin that has at least three phonetically close, but irregular variants (ǂoˤnu ~ ǂxonu ~ ǂonu); this circumstance implies that it may have a substrate origin. It is also very similar to the Ju root *ǂũʰ 'star', yet another of several interesting isoglosses between North Khoisan and the Naro area of Kalahari Khoe. Reconstruction shape: Correspondences are regular; the only frequent development is a partial ore full assimilation of the first vowel under the influence of the second one.
Dornan 1917: 96. Cf. ɡǀǀwaː 'a grave' - most likely, the same word, with a fictitious distinction in vowel length. Alternately, cf. also karo, glossed as 'a large hail stone; a stone' [Dornan 1917: 101].
Proto-Kalahari Khoe:*ɳǀǀoaˤ
Vossen 1997: 495 (*ɳǀǀoa ~ *ɳǀǀnoa). Distribution: Well represented in most of the subgroups. Replacements: (a) Deti-Cara ǀʼui 'stone' is clearly related to Khoekhoe *ǀʼui 'stone, mountain' and is subsequently reconstructed in those meanings as *ǀʼui on both the Proto-Kalahari and Proto-Khoe levels in [Vossen 1997: 495]. However, the complete absence of this root from any other Kalahari branch or language other than Deti and Cara raises doubts about its Proto-Kalahari status. One might consider the possibility of a relatively recent borrowing from Khoekhoe (though this is also dubious because of the lack of direct contact and because no other similar cases have been detected so far), or some sort of specialized original semantics that prompted the root to become extinct in all other languages, but widen its meaning to 'stone' in Deti-Cara (also an odd scenario), but in any case, the widespread distribution of *ɳǀǀoa efficiently blocks *ǀʼui from aspiring to Proto-Kalahari Khoe status in the basic function of 'stone'; (b) ǀXaise káró, Tsua qárò is a recent innovation that still shows a specialized meaning in Dornan's Hiechware records (karo 'large hail stone'). Reconstruction shape: Correspondences are mostly regular. Pharyngealization has to be set up based on Naro-ǂHaba data (Naro also has nasalization on the diphthong, which is hard to explain). Vossen's opposition between *ɳǀǀ- and *ɳǀǀn- remains unverified. Semantics and structure: It is worth noticing that Proto-Kalahari Khoe *ɳǀǀoa does not seem to display the frequent polysemy 'stone / mountain' (equivalents for 'mountain' either remain unattested in Kalahari Khoe languages or represent completely different roots).
Kilian-Hatz 2003: 153. Polysemy: 'sun / time / turn (n.)'. Quoted as ǀáḿ in [Köhler 1973: 186]. Cf. also čã̂, glossed as 'sun, sunlight, heat of the sun; be light, be hot' in [Kilian-Hatz 2003: 123]. Cf. Buga-Khoe ǀáḿ id. [Vossen 1997: 492].
Dornan 1917: 100. Also transcribed as ɡǀǀxam in [Dornan 1917: 95], with polysemy: 'sun / light / day' (the two forms are most likely the same word, with an incorrectly transcribed lateral, rather than dental, click in the latter).
Proto-Kalahari Khoe:*ǀam
Vossen 1997: 492 (*ǀáḿ). Distribution: Preserved in most languages. Replacements: The only known replacement is in Deti (and possibly Cara, with two alleged synonyms), where the original word has been ushered out by a reflex of Proto-Kalahari Khoe *kʰobo '(to be) warm; to sweat' [Vossen 1997: 505], thus, with the semantic shift {'warm' > 'sun'}. Reconstruction shape: Correspondences are regular and trivial.
Vossen 1988: 99. Semantic difference from ǀáḿ remains unclear; based on general typological considerations and the etymology of both forms, it may be that ǀáḿ is 'sun' (physical body), whereas kʰóbó is 'sunlight; heat of the sun' - but cf. the situation in Deti, where R. Vossen records only kʰóbó in the meaning 'sun'. We may be dealing with a case of "transit synonymy".
Visser 2001: 2. Somewhat dubious. In Visser's dictionary, this verb is glossed as 'go through'; the idea of 'swimming' only springs up for the derivative baˤra-xʼao 'one who goes through a difficult situation, one who suffers' (literally 'go-through-male-person'), with one further example: cʰaː baˤra-xʼao-ba 'swimmer' (literally 'water-go-through-male-person') [Visser 2001: 3]. However, in [Vossen 1997: 490] the verb bàdá is expressly glossed as 'swim', and cf. also the external data from other Non-Khoekhoe languages. Meanwhile, in Visser's dictionary the meaning 'swim' is rendered as gã̀ã̄-tòè [Visser 2001: 28], literally 'duck-move'; and in [Barnard 1985: 156], the same meaning is rendered as ǀǀábú, literally 'to push'. Clearly, the concept of 'swimming' is quite unstable in Naro (for understandable reasons); we prefer to make an exception here and go along with Vossen's choice, considering the external parallels and the lack of consistency in between the "competing" sources.
Vossen 1997: 490 (*bada). Distribution: Preserved almost everywhere where attested, though there are many gaps in the wordlists. Replacements: Possibly replaced in Hiechware, but the source is old and unreliable, and the alleged replacement has no etymology. Reconstruction shape: Correspondences are regular and trivial, except for pharyngealization in Naro (which may possibly be archaic). Semantics and structure: This is one of several Khoe roots beginning in *b-, and as such, its origins are suspicious, as well as its original semantics in light of the general meaning 'go through', recorded by Visser for Naro. Nevertheless, there is no conclusive evidence that speakers of Proto-Kalahari Khoe were unfamiliar with the concept of 'swimming', and *bara is clearly the optimal (and only) candidate for this meaning in the protolanguage.
Dornan 1917: 111. Confused by Dornan with a near-homonymous (but not completely homonymous, as seen in external data) cau 'hand' q.v.
Proto-Kalahari Khoe:*cao
Vossen 1997: 488 (*cáó). Distribution: Preserved in all languages. Reconstruction shape: Correspondences are regular and trivial.
Number:85
Word:that
ǀǀAni:ɳǀné=tè1
Vossen 2000: 137. A compound form, consisting of ɳǀné 'this' q. v. followed by the far deixis particle tè. Cf. the same use of this particle in ʔá 'that one mentioned before' : ʔá-tè 'that yonder, very far away' (also ʔá-àkà-tè id., where the middle element = ʔákà 'far'). The principal meaning of 'far deixis' is transparently expressed by the morpheme tè.
Kilian-Hatz 2003: 166. Used both in the adjectival and substantive functions. In [Köhler 1981: 528], quoted as ɳǀĩ́; the same source also gives an additional variant ɳǀnĩ́-ʔē, with the same additional emphatic (or directive) particle as in ɳǀné-ʔē 'this one there, that one' (see under 'this'). According to Kilian-Hatz, may also be encountered in abbreviated variants: ń ~ ŋ́ [Kilian-Hatz 2003: 89, 99], but these look more like neutralizations of the opposition between 'this' and 'that' q.v., i. e. general deictic stems without specific reference to the location of the speaker.
Visser 2001: 1. Quoted as ãː in [Barnard 1985: 127]. It is not quite clear from existing descriptions to which degree of "far deixis" this particular morpheme is referring (unlike ẽ̄ː, which gets a more precise definition in Visser's dictionary).
Vossen 1997: 254. This is clearly a secondary derivative formation from ɳǀnẽ́ 'this' q.v., expanded with the aid of the particle chain =si-ha that seems to take upon itself the expression of the actual meaning of "far deixis". However, cf. also ʔáā 'that', clearly related to the main morphemes with the meaning of "far deixis" in Naro, ǂHaba, and ǀǀGana, in [Nakagawa 2006: 214]; in [Vossen 1997], the existence of this pronoun in ǀGwi is not mentioned. In order to avoid a lexicostatistical error, we include both forms as synonyms.
The systems of demonstrative pronouns in Kalahari Khoe languages differ rather widely from each other, so reconstruction here is complicated and needs to be taken on a step-by-step basis. (1) For ǀǀAni-Kxoe, it makes sense to assume the originality of the Kxoe system: *ɳǀe 'this', *ɳǀĩ 'that (interm.)', *ɳǀǀa 'that (distant)'. In ǀǀAni, the latter two pronouns fell out of use; instead, *ɳǀe became the general deictic stem and a new distal deixis pronoun was formed from it (with the addition of =te, a morpheme of unclear origin). (2) In Naro-ǀǀGana, we generally observe a binary system: *ɳǀe 'this' vs. *ʔa- 'that' (occasionally, both stems are extended by means of additional suffixes). (3) In East Kalahari Khoe, the most common system is also binary: *u 'that' vs. *i 'this' (Tsʼixa restores *ɳǀe, probably under the influence of the Western group). If we also bring into consideration the external data of Khoekhoe (where we reconstruct *ɳǀǀa- 'that' vs. *ne 'this'; the latter form may be thought of as a sporadic simplification from an earlier *ɳǀe due to the auxiliary / grammatical status of the word), this makes it likely that the click-containing pronominal forms are more archaic, while the monovocalic forms are secondary. These may have two possible sources of origin: (a) external (cf. Bantu *o 'that', which may have been the source of at least *u in East Kalahari) and (b) extended from original referential pronouns, cf. the existence of i as a referential 'that' and a as a referential 'this' in Kxoe. Consequently, we propose to provisionally reconstruct something like *ɳǀe- for the proximal 'this' and *ɳǀǀa- for the distal 'that' on the Proto-Kalahari Kxoe level (even if the latter is retained, of all living languages, only in Kxoe).
Kilian-Hatz 2003: 215. Quoted as ɳǀǀnǎ in [Köhler 1981: 528]. There are two degrees of "distal deixis" in Kxoe, but the descriptions are contradictory: Kilian-Hatz defines ɳǀǀnàā as "far dist.", whereas Köhler defines it as 'celui là', opposed to ɳǀnĩ́ 'celui qui est là-bas, de l'autre côté'. Technically, we treat both stems as synonyms.
Visser 2001: 25. Meaning glossed as 'that (thing) there (near addressee)'. Cf. also the complex derivative ẽ̄ː-sí-hã̄ː 'that there (far away)' [Visser 2001: 26]; the same form is quoted as ʔì-sí-hà in [Vossen 1997: 254].
Kilian-Hatz 2003: 165; Köhler 1981: 528. Adjectival stem, cf. ɳǀné ŋgú-hɛ̀ "this house", etc. It is also used within more complex forms, such as: ɳǀné-ʔē 'this one there, that one', ɳǀné-ʔé 'this one here' (specifying different "subtle" degrees of deixis). According to Kilian-Hatz, may also be encountered in abbreviated variants: ń ~ ŋ́ [Kilian-Hatz 2003: 89, 99], but these look more like neutralizations of the opposition between 'this' and 'that' q.v., i. e. general deictic stems without specific reference to the location of the speaker.
Visser 2001: 48. Quoted as ɳǀné in [Vossen 1997: 254]; as ɳǀí ~ ɳǀe in [Barnard 1985: 128]. As a possible synonym, R. Vossen also lists ɳǀǀà in the meaning 'this' [Vossen 1997: 254], but in [Visser 2001: 57], this pronoun is explained as 'something or someone already referred to', i. e. a special anaphoric word that is not directly involved in the basic deictic opposition 'this / that'.
Nakagawa 1996: 114. Meaning glossed as 'this (deictic)'. Quoted as ɳǀné in [Vossen 1997: 254]. Cf. ɳǀǀnã́ 'this (anaphoric)' [Nakagawa 1996: 114]. In [Tanaka 1978: 97], 'this' is listed as ɳǀui(-ša) ~ ɳǀu-ša; it is not clear if this is the same word, since the vocalism is very different.
Vossen 1997: 254. In [Tanaka 1978: 97], 'this' is listed as ɳǀui(-ši) ~ ɳǀu-ša; it is not clear if this is the same word, since the vocalism is very different.
Vossen 1997: 255. Possible secondary synonym: ɳǀní (not included here since the semantics is not quite clear; this could in reality be an "intermediate 'that'" rather than a true synonym to ǐ 'this').
Vossen 1997: 255. Possible secondary synonym: ɳǀní (not included here since the semantics is not quite clear; this could in reality be an "intermediate 'that'" rather than a true synonym to ǐ 'this').
Visser 2001: 75. 97. Masculine cáː vs. feminine sáː. Quoted as cá ~ cá-cí (masc.), sá ~ sá-sí (fem.) in [Vossen 1997: 239]. These are the full accented (emphatic) forms. The shortened, non-emphatic, variants are ci (masc.), si (fem.) [Visser 2001: 76, 102], cì (masc.), sì (fem.) [Vossen 1997: 239]. The same variants with high instead of low tone function as 2nd sg. object concord markers: -cí, -sí [Vossen 1997: 239].
Vossen 1997: 370. Distribution: Preserved in all daughter languages. Replacements: No true replacements, but the change to cí / sí in ǀGwi may have taken place under the influence of the 1st person pronoun forms. Reconstruction shape: The opposition of masculine *ca vs. feminine *sa is well supported by data from all subgroups. Semantics and structure: Comparison with the Proto-Khoe feminine marker *-sa suggests that the opposition *ca (m.) / *sa (f.) may go back to the combination of the pronominal root *=a with an archaic masculine marker *c= (only preserved in this pronoun) and the usual feminine marker *s=. However, such segmentation is formally impossible, only hinted at, on the Khoe level of analysis.
Vossen 1997: 510 (*dam). Distribution: Preserved in all daughter languages. Reconstruction shape: The variant tàˤm̄, attested in Visser's dictionary of Naro, speaks in favor of amending the reconstruction to *daˤm; however, pharyngealized articulation is not attested in any other languages or even in the two alternate sources on Naro (Vossen's and Barnard's records), so perhaps we are dealing with some idiosyncratic peculiarity here.
Number:89
Word:tooth
ǀǀAni:ságárì-1
Vossen 1988: 100. The word has a distinctly non-Khoe shape, but the source of borrowing is unknown.
Vossen 1997: 509. In [Tanaka 1978: 99], 'tooth' is glossed as !toʓi - obviously a different word, but not identifiable with anything in more reliable sources so far.
Vossen 1997: 509. In [Tanaka 1978: 99], 'tooth' is glossed as !toʓi - obviously a different word, but not identifiable with anything in more reliable sources so far.
Dornan 1917: 96. Probably the same word is also (mis)transcribed as !xo in [Dornan 1917: 103].
Proto-Kalahari Khoe:*ǀǀũˤ
Vossen 1997: 509 (*ǀǀũ). Distribution: Preserved in all daughter languages except for ǀǀAni, where it has been replaced with a strange trisyllabic word of unclear provenance. Reconstruction shape: Correspondences are generally regular and trivial. Pharyngealized articulation of the vowel is pointed at by the reflexes in Naro and ǂHaba, as in several other cases. (Note that Vossen and Barnard do not attest pharyngealization in Naro, but Visser's dictionary clearly marks it).
Dornan 1917: 97, 112. Polysemy: 'tree / medicine'. Cf. the variant without marked vowel length in hi-ǀwa 'a bush or shrub' [ibid.].
Proto-Kalahari Khoe:*yi
Vossen 1997: 419 (*yi). Distribution: Preserved in all daughter languages. Reconstruction shape: This is a rare case of initial *y-, and a unique one of initial *yi-; however, correspondences are quite straightforward in all cases.
Dornan 1917: 64, 100. Polysemy: 'two / both / twice'. Final -ɲe is probably an adjectival or verbal suffix, but it is not clear to what degree of accuracy it has been transcribed.
Proto-Kalahari Khoe:*ǀam
Vossen 1997: 510 (*ǀam). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular and trivial.
Dornan 1917: 103. Meaning glossed as 'to go, to come from'.
Proto-Kalahari Khoe:*!ũ
Vossen 1997: 441 (*!ũ). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are generally regular; loss of *!- is fully regular for all East Kalahari Khoe and many West Kalahari Khoe languages, but the click is expectedly preserved in Naro and ǀGwi and is reliably reconstructible on the basis of both internal and external data.
Number:93
Word:warm (hot)
ǀǀAni:
Not attested. In the meaning 'warm' (not 'hot') the form kʰóbò is quoted [Vossen 1997: 505].
ǀGanda:
Not attested. In the meaning 'warm' (not 'hot') the form kʰóbò is quoted [Vossen 1997: 505].
Kxoe:čã̂1
Kilian-Hatz 2003: 123. Meaning glossed as 'be light, be hot'; also as the noun 'sun, sunlight, heat of the sun'. Distinct from kʰóvò [Kilian-Hatz 2003: 67] = kʰóβò [Köhler 1981: 493] 'warm' and from ǀǀʼóō 'be warm, be gentle (of air)' [Kilian-Hatz 2003: 208].
Visser 2001: 45. Polysemy: 'hot / feverish'. The form is regularly derived from the verb kùrū, glossed in [Visser 2001: 44] as 'get warm, be warm'. Distinct from ǀǀʼōò 'warm, not hot, lukewarm' [Visser 2001: 119] (R. Vossen quotes this as ǀǀʼô and also indicates the semantics of 'warm' rather than 'hot' [Vossen 1997: 504]. The semantics 'hot' for the main entry is also confirmed in [Barnard 1985: 117]: kùrú 'hot', distinct from câ ~ cã̂ 'boiling' [Barnard 1985: 115].
Visser 2001: 77; Vossen 1997: 240. Morpheme corresponds to the dual stem. Specific forms are: sì=cá=ḿ (masculine), sì=sá=ḿ (feminine), sì=kʰá=ḿ (common). These are full (emphatic) forms; short variants without the prefixal element sì= are also in use: cà=m̀, sà=m̀, kʰà=m̀. Corresponding object forms all reduplicate the main pronominal morpheme: cá=ḿ-ʔm̀, sá=ḿ-ʔm̀, kʰá=ḿ-ʔm̀ [Vossen 1997: 240].
Vossen 1997: 245. Morpheme corresponds to the dual stem. Specific forms are: cá=m̀ (masculine), sá=m̀ (feminine), kʰá=m̀ (common). Corresponding object forms are: cà=m̀, sà=m̀, kʰà=m̀.
Vossen 1997: 244. Morpheme corresponds to the dual stem. Specific forms are: hí=cè=bè (masculine), hí=sè=bè (feminine), hí=kʰè=bè (common). Corresponding object forms are: cè=mà, sè=mà, kʰè=mà.
Vossen 1997: 241. Morpheme corresponds to the dual stem. Specific forms are: í=cè=bè (masculine), í=sè=bè (feminine), í=kʰè=bè (common). Corresponding object forms are: í=cè=mà, í=sè=mà, í=kʰà=mà.
Dornan 1917: 65. Dornan only lists the form c=e 'we', corresponding to the 1st p. plural, common gender form in the closely related Kua and Tsua languages. There is little doubt that Hiechware possessed a paradigmatic system that was quite close to Kua and Tsua, but Dornan did not record it.
Proto-Kalahari Khoe:*=m
Vossen 1997: 370 (m. *cabe, f. *sabe, c. *kʰabe). Distribution: Preserved in all daughter languages, though the actual form is amended in some of them. Reconstruction shape: Vossen reconstructs the main root morpheme for this dual pronoun as *=be, based on the isogloss between two subgroups: ǀGwi-ǀǀGana and Kua-Tsua. The more common variant *=m is tentatively explained by him as the result of secondary nasalization, similar to the development of the nominal masculine gender *-ba > -ma in many Kalahari Khoe languages. However, unlike the nominal masculine gender, in this particular case not only is the variant with -b- much less widespread in Kalahari Khoe, it also goes against the external evidence of Khoekhoe (Nama and !Ora also have -m as the main root morpheme). It seems more likely that the change *-m > -be is an areal process, triggered not by phonetic but by analogical development (a possible scenario: *-m > -me by analogy with the plural forms in -e, followed by denasalization > -be because of the phonotactic peculiarity of the syllable me). The opposite scenario would have required us to assume an independent and phonetically unmotivated development *-be > -m in several branches. Semantics and structure: The complete paradigm is reconstructible as m. *ca=m, f. *sa=m, c. *kʰa=m.
Vossen 1997: 238. Morpheme corresponds to the plural stem. Specific forms are: ǀǀ=é (masculine), y=é (feminine), t=é (common). Corresponding object concord marker is attested only for the masculine form: -ǀǀè [ibid.].
Visser 2001: 77; Vossen 1997: 240. Morpheme corresponds to the plural stem. Specific forms are: sì=ǀǀ=áé (masculine), sì=s=é (feminine), sì=t=á (common). Variation between the allomorphs =é and =á is not explainable synchronically. These are full (emphatic) forms; short variants without the prefixal element sì= are also in use: ǀǀ=àè, s=è, t=à. Corresponding object forms all reduplicate the main pronominal morpheme: ǀǀ=áé-ʔà, s=é-ʔá, t=á-ʔá [Vossen 1997: 240].
Vossen 1997: 245. Morpheme corresponds to the plural stem. Specific forms are: ǀǀ=â (masculine), s=ê (feminine), t=â (common). Corresponding object forms are: ǀǀ=à, s=è, t=à.
Vossen 1997: 244. Morpheme corresponds to the plural stem. Specific forms are: há=ǀǀ=à (masculine), há=s=è (feminine), há=t=à (common). Corresponding object forms are: ǀǀ=à, s=è, t=à.
Vossen 1997: 241. Morpheme corresponds to the plural stem. Specific forms are: í=ǀǀ=áè (masculine), í=s=ê (feminine), í=t=ê (common). Corresponding object forms are: í=ǀǀ=à, í=s=à, í=t=à.
Vossen 1997: 250. Morpheme corresponds to the plural stem. Specific forms are: k=á (masculine), s=íè (feminine), c=íè (common).
Hiechware:
Proto-Kalahari Khoe:
Vossen 1997: 373 (m. *ǀǀae, f. *se, c. *tae). Distribution: Preserved in all daughter languages, though the actual form is amended in some of them. Reconstruction shape: The original form of the 1st p. pl. root morpheme is unquestionably *-e. Variation between -e and -a in some of the actual forms in some languages may be due to contraction with the gender prefixes, the details of which vary between specific idioms. Semantics and structure: The complete paradigm should probably be reconstructed as m. *ǀǀa=e, f. *sa=e, c. *ta=e (although the Eastern branch rather points to *ca=e, which cannot be a normal phonetic development from *ta=e).
Vossen 2000: 139. The same source also mentions ná 'what, which?' as a synonymous form, with one example where the word functions as syntactic object. It is, therefore, probably a contraction: né + ʼà (object marker), cf. má-ʼà 'whom?' from ma 'who?'.
Kilian-Hatz 2003: 82. Literally 'which thing?' (see 'who' for further notes on the interrogative morpheme). Substantive form, distinct from the adjectival formation ndéú < *ndé-xú 'what kind of, which' [Kilian-Hatz 2003: 91].
Visser 2001: 23; Vossen 1997: 263. Quoted as du ~ duː in [Barnard 1985: 164]. The morpheme m, whose meaning R. Vossen gives as 'what?' in [Vossen 1997: 263], is in fact a general interrogative morpheme, usually encountered in adverbials (m̀-ǀámá 'when?', m̄-dà 'where?', etc.).
#Haba:
Not attested.
ǀGwi:yǐ #4
Vossen 1997: 264. Cf.: ɳǀé yǐ xô-ʒè ʔè "what is this thing?" [ibid.]. This complex form ('what-thing') is listed in [Tanaka 1978: 106] as î-ho-ši. On the other hand, the same source, in the same meaning 'what?', also lists forms that are clearly the same as in ǀǀGana - nuː ~ nuː-še ~ nuː-e; they are, however, not mentioned in R. Vossen's grammatical notes. The entire situation really requires additional investigation.
Vossen 1997: 264. Cf.: ɳǀé nǔ-ʒè ʔè "what is this?" [ibid.]. The exact form is not clear, since the grammatical notes on the same page list the actual form as dǔ rather than nǔ; but cf. also the forms for 'what?' in [Tanaka 1978: 106]: nuː ~ nuː-se ~ nuː-e ~ nuː-ma (with various gender and number suffixes). It is possible that dǔ in [Vossen 1997: 264] is a local misprint, although external comparison clearly shows that nǔ does indeed correlate with dǔ in other languages (e. g. Naro).
Dornan 1917: 72, 106. This form is elicited through the example: nao e are heaha "what have they done?" (other contexts listed by Dornan are not diagnostic).
Proto-Kalahari Khoe:*n-du
Vossen 1997: 381 (*(n)du). Distribution: Well attested in the Eastern branch and in several Western languages as well. Reconstruction shape: Khoe languages do not normally allow for word-initial consonant clusters, but initial nd- for this stem is well attested in several languages, signifying a certain degree of uniqueness. In this situation, it is quite likely that Naro dǔ, Deti dú etc. are phonotactic simplifications of the original *ndu. Semantics and structure: The existence of a simple root *ndu is highly unlikely for phonotactic reasons; the form may be analyzed as an old contraction of a bisyllabic compound (e.g. *na-du) where the first morpheme is the same general interrogative stem that functions by itself in several Khoe languages with the adjectival meaning 'which?' (e.g. Kxoe, Cara ná, Kua nǎ). In this case, the second morpheme should probably go back to some old noun meaning smth. like 'thing', though no isolated word like *du or *tu is attested with anything close to that meaning. Nevertheless, already on the Proto-Khoe level *ndu must have been understood as an inseparable entity, given its firmly monosyllabic behavior in all daughter languages. Replacements: It looks as if the original compound *n-du is prone to being replaced by other forms containing the general adjectival interrogative morpheme *na-: (a) ǀǀAni né, Tsʼixa né is an areal isogloss, probably a contraction from *na-e or *na-i; (b) Kua ná-ũ̀, Hiechware na-o is another obvious compound, restricted to the Kua-Tsua subgroup. In both cases, the second morpheme may historically go back to one of the deictic stems (*i 'this', *u 'that'). Another case is (c) Kxoe mà-xú, composed of *ma 'who?' + *xu 'thing'.
Kilian-Hatz 2003: 175. Verbal stem; meaning glossed as 'be white, be beige, be whitish'. The suffix -ʆi is used to derive verbs from nouns, and the corresponding nominal stem !ʼé 'something white' is attested in [Köhler 1981: 500], but not in Kilian-Hatz's dictionary. Antiquity of this stem is dubious, since it contains the alveolar click !-, which is generally lost in Kxoe in verbs of Proto-West Khoe origin. However, no identifiable source of borrowing has been detected so far. Cf. also the secondary synonym xùɛ́-ʆi 'be bright white, be very white' [Kilian-Hatz 2003: 147].
Vossen 1997: 506 (*xóé). Distribution: Preserved in several East Kalahari Khoe languages (Cara, ǀXaise, Deti; also in Hiechware as a representative of "old Tshwa" speech patterns), as well as in ǀǀAni, at the other end of the Kalahari Khoe continuum. Replacements: Upon first sight, the most obvious candidate for Proto-Kalahari Khoe 'white' would be *!ʼu, whose internal distribution is at least as strong as that of *xoɛ and which is also supported by external data (Nama *!ʼu-ri 'white'). However, a serious problem with this solution is that the reflexes of this root in Kua-Tsua (Kua !ʼǔ, Tsua !ʼú) show preservation of the alveolar click, a near-unique situation, since the East Kalahari Khoe branch regularly drops this click. This implies that the Kua-Tsua forms were probably borrowed - perhaps from Naro or some other source. In the end, the only language whose reflex cannot at all be placed under doubt for phonetic reasons is Naro itself; and it cannot be excluded, either, that the Naro form, too, is ultimately a borrowing from a Khoekhoe source (cf. particularly the preservation of the root *xoɛ in Naro in the emphatic form xõẽ-xõẽ-se 'very white'). On the other hand, it must be noted that the form *xoɛ (a) is not particularly stable in itself, as we see it being easily replaced by other items, such as *!ʼe for ǀGanda-ǀǀAni and *!o- for Danisi and Tsʼixa, both of unclear origin; (b) may be nominal in origin, cf. its meaning in Kxoe - xóɛ́ 'foam, froth', as well as external cognates (most likely related to Nama goa, !Ora gòà-b 'froth, foam'. All this means that the reconstruction is highly tentative. Reconstruction shape: Correspondences for those languages in which the item is attested are generally regular; ǀǀAni data speak in favor of reconstructing the word as *xoɛ rather than *xoe (this also agrees with the reflexes in Khoekhoe, where *oɛ > oa).
Visser 2001: 18. Quoted as di in [Vossen 1997: 263] and in [Barnard 1985: 165]; also used with gender markers (dí-bá masc., dí-sá fem., dí-ná general).
Vossen 1997: 264. The quasi-synonymous mã̂, judging by the examples, is used rather as an adjectival question word referring to already identified people, e. g.: ɳǀé-ǀǀù xʼáó-kʰòè-ǀǀù mã́-ǀǀù ʔè "who (i. e. what) are these men?" [ibid.].
Vossen 1997: 264. The quasi-synonymous mâ, judging by the examples, is used rather as an adjectival question word referring to already identified people, e. g.: ɳǀé-s kʰóè-sà má-s ʔè "who (i. e. what) is this woman?" [ibid.].
Vossen 1997: 265. According to R. Vossen, this interrogative pronoun is in free variation with a different stem, nǎ.
Tsua:
Not attested.
Hiechware:na-re3
Dornan 1917: 66. Cf.: kau nare "who are you?", nare ča kwa moː "whom do you see?".
Proto-Kalahari Khoe:*ma
Vossen 1997: 379 (*mã). Distribution: Very well attested throughout both West and East Kalahari Khoe languages. Replacements: (a) Naro dǐ, ǀGwi-ǀǀGana dí. Origin is unclear, but in the light of Naro + ǀǀGana du 'what?' < Kalahari Khoe *n-du (see 'what' for details), it may be suggested that di also < *n-di, where *n- is the general adjectival interrogative morpheme and *di is some sort of nominalizer (perhaps the same as the relative particle di?). If this analysis is correct, then these forms may actually go back to the exact same source as (b) Hiechware na-re 'who?' < *na-di. However, we still mark these replacements with different numbers because even if the analysis is correct, these two replacements must have been independent of each other in two different branches. Reconstruction shape: The original animate interrogative morpheme is unequivocally reconstructible as *ma (Vossen has *mã with nasalization, but the nasalized variant is more rare and might actually be secondary).
Vossen 2000: 136. Same stem as 'person', specified with the regular feminine gender morpheme -hɛ̀. Should be distinguished from ɡǀǀɛ- 'female, feminine' [Vossen 1997: 506].
Kilian-Hatz 2003: 211. Also with the fem. gender marker: ɡǀǀɛ̀ɛ̄-kʰòè-hɛ̀. A compound formation, consisting of ɡǀǀɛ̀ɛ̄ 'woman, female' (a frequent compound participant) and kʰòè 'person' q.v. The first morpheme is quoted as ǀǀɛ̌ in [Köhler 1981: 491]; judging by the examples on the same page, simply referring to 'woman / wife' as ǀǀɛ̌ is "vulgar", and the neutral expression is the compound ǀǀɛ̌-xʼōē. Cf. Buga-Khoe ɡǀǀɛ- 'female' [Vossen 1997: 506].
Visser 2001: 41. Same word as 'person' q.v. with the feminine gender marker. A more specific equivalent for 'woman' as 'female human being' is ɡǀǀàē-kʰòè [Visser 2001: 24], with polysemy: '(young) woman / wife'. Literally: 'female person', from ɡǀǀàē 'female' and kʰoe 'person' q.v.; quoted as ɡǀǀáí-s-kʰùè in [Barnard 1985: 71] (with an additional female gender suffix after the first morpheme). However, most text examples show that the simple form kʰóè-sa is used in basic situations.
#Haba:
Not attested properly, but cf. ɡǀǀae- 'female' in [Vossen 1997: 506].
Vossen 1997: 506. This source only lists the morpheme ɡǀǀae- in the meaning 'female', but cf. ɡǀǀaí-kue ~ ɡǀǀaí-ko-ši 'woman' (= ɡǀǀae- 'female' + kʰoe 'person' q.v.) in the phonetically unreliable, but lexically informative dictionary [Tanaka 1978: 108].
Vossen 1997: 506. This source only lists the morpheme ɡǀǀɛ- in the meaning 'female', but cf. ɡǀǀaí-kue ~ ɡǀǀaí-ko-ši 'woman' (= ɡǀǀɛ- 'female' + kʰoe 'person' q.v.) in the phonetically unreliable, but lexically informative dictionary [Tanaka 1978: 108].
Vossen 1997: 506. Meaning glossed as 'female'; the word 'woman' is probably formed from this root and 'person' q.v., although this has not been explicitly stated in the source.
Vossen 1997: 506. Meaning glossed as 'female'; the word 'woman' is probably formed from this root and 'person' q.v., although this has not been explicitly stated in the source.
Vossen 1997: 506. Meaning glossed as 'female'; the word 'woman' is probably formed from this root and 'person' q.v., although this has not been explicitly stated in the source.
Vossen 1997: 506. Meaning glossed as 'female'; the word 'woman' is probably formed from this root and 'person' q.v., although this has not been explicitly stated in the source.
Vossen 1997: 506. Meaning glossed as 'female'; the word 'woman' is probably formed from this root and 'person' q.v., although this has not been explicitly stated in the source.
Vossen 1997: 506. Meaning glossed as 'female'; the word 'woman' is probably formed from this root and 'person' q.v., although this has not been explicitly stated in the source.
Vossen 1997: 506. Meaning glossed as 'female'; the word 'woman' is probably formed from this root and 'person' q.v., although this has not been explicitly stated in the source.
Vossen 1997: 506 (*ɡǀǀae). Distribution: Preserved in all languages, although see the discussion on semantics below. Reconstruction shape: Correspondences are generally regular and trivial. Semantics and structure: It seems that the original meaning of the morpheme *ɡǀǀae is 'female' (applicable to animals and people); the meaning 'woman' is perceived as more complex than 'female' and is often expressed by either the feminine gender form of the lexeme *kʰoe 'person' (as in ǀǀAni and Naro) or the compounding of two lexical roots (*ɡǀǀae-kʰoe, as in Kxoe and Hiechware). Nevertheless, since 'woman'-meaning forms with *ɡǀǀae are far more frequent than 'man'-meaning forms without this morpheme, it seems justified to put this morpheme into the main Proto-Kalahari Khoe slot for 'woman', leaving *kʰoe in the slot for 'person'. See an absolutely symmetric picture for the item 'man' q.v.
Number:100
Word:yellow
ǀǀAni:
Not attested.
ǀGanda:
Not attested.
Kxoe:čèrè-tòè-ǂʼú1
Kilian-Hatz 2003: 124. Literally: 'silver terminalia (Terminalia sericea)' (čèrèé) + 'root' (tòɛ̄, q.v.) + 'color' (ǂʼú), with slight morphophonological changes of the individual stems.
Visser 2001: 15. Literally: 'tortoise-egg' (see further on 'egg'). The idiomatic expression is confirmed in [Barnard 1985: 121], where it is quoted as dam ǂobi. The latter source also lists the synonym ǀkore, not confirmed elsewhere.
#Haba:
Not attested.
ǀGwi:
Not attested.
ǀǀGana:
Not attested in reliable sources, but cf. dam-!ubi 'yellow' in [Tanaka 1978: 110] (= Naro dàḿ ǂʼūbī q.v.).
Dornan 1917: 96. Phonetically probably = ŋgu-/o/, cf. the entry in the closely related Kua.
Proto-Kalahari Khoe:*ɳ!u
Vossen 1997: 507 (*ɳ!u). Distribution: Preserved in most languages. Replacements: Replaced in the ǀGwi-ǀǀGana subgroup with the word xʼai of unclear origin. Reconstruction shape: Correspondences are regular and trivial, with predictable click loss in most languages.
Visser 2001: 70. Polysemy: 'heavy / difficult / pregnant'. Quoted as !úḿ in [Vossen 1997: 490]. Completely different word quoted in [Barnard 1985: 117]: tʰu (not confirmed in the rest of the sources).
Vossen 1997: 490 (*!óḿ ~ *!úḿ). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are generally regular, with expected click loss in most of the languages. Vocalic reflexes (as well as external data) indicate that *-om is preferable as the original coda to *-um.
Dornan 1917: 104. Possibly the same word is mistranscribed as ɡ!uː 'near' in [Dornan 1917: 96]; the alveolar click is quite likely to have been transcribed in error (instead of the etymologically expected dental click; cf. the same situation with 'full' q.v.).
Proto-Kalahari Khoe:*ǀu
Vossen 1997: 470 (*ǀû). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular and trivial.
Visser 2001: 78. Quoted as dàˤbéˤ in [Vossen 1997: 481]; as dǎˤbé (com.) in [Barnard 1985: 98] (the female gender form of the same word is said to mean 'salt pan').
Dornan 1917: 92. Cf. also debeː 'a salt pan' (either the same word or a contraction with an older suffix).
Proto-Kalahari Khoe:*doˤbe
Vossen 1997: 481 (*dobe). Distribution: Preserved in all languages where attested. Reconstruction shape: The only problem is with the first root vowel. Pharyngealization, attested in Naro and ǀGwi-ǀǀGana, may be archaic. Quality should be reconstructed as labial, since the coda *-abe is typically preserved without labialization in all Kalahari Khoe languages and occasional dissimilation *-obe > -abe is more probable.
Visser 2001: 116. Quoted as ǀǀúm̀ in [Vossen 1997: 460]; as ǀǀkum ~ ɡ!um in [Barnard 1985: 119] (the variant with the alveolar click is probably erroneously transcribed).
Dornan 1917: 104. Cf. also !xo kadi 'short' [Dornan 1917: 103] - a compound formation whose individual parts are not identifiable. Dubious, especially in the light of external data.
Proto-Kalahari Khoe:*ǀǀom
Vossen 1997: 460 (*ǀǀom). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular and trivial.
Kilian-Hatz 2003: 184; Köhler 1981: 489. Secondary synonym: ǂqòyō 'snake' [Kilian-Hatz 2003: 194]. Semantic difference is not explained; the latter form is not found in Köhler's materials. It should be noted that the phonotactic structure of both forms is somewhat atypical for Kxoe; the second syllable -yō suggests either some obscure suffixal formation or a borrowing (from an unknown source).
Dornan 1917: 94, 95. The variant with the dental click is in better agreement with external data; the variant with the lateral click is either corrupt or represents a different word.
Proto-Kalahari Khoe:*ǀxʼao
Vossen 1997: 485 (*ǀxʼao). Distribution: Well attested in all major branches of the family. Replacements: (a) Kxoe ǂʼíyō, of unclear origin (and with a rather weird phonetic shape for a Khoe root); (b) Tsʼixa-Danisi ǀǎ-kà-kũ̀, also of unclear origin and looking like a euphemistic replacement for the original word. (One possible way of analysis is smth. like 'belly-walker', cf. Tsʼixa ǀâ 'belly' and kũ 'to go'). Reconstruction shape: Correspondences are generally regular, with the expected phonetic shift *ǀxʼ- > *ǀʼ- in East Kalahari Khoe.
Number:106
Word:snake
ǀǀAni:
ǀGanda:
Kxoe:
Naro:
#Haba:
ǀGwi:
ǀǀGana:
Cara:
ǀXaise:
Danisi:ǀǎ-kà-kũ̂3
Vossen 1988: 97. This word (see notes on Ts'ixa) is listed as the primary equivalent for 'snake' in this source; the difference between it and the older equivalent ǀʼàó remains unclear.
Ts'ixa:
Deti:
Kua:
Tsua:
Hiechware:
Proto-Kalahari Khoe:
Number:107
Word:thin
ǀǀAni:
Not attested.
ǀGanda:
Not attested.
Kxoe:ǀíní1
Kilian-Hatz 2003: 155. Cf. also ǀǀʰã̀ĩ́ 'be meagre, be thin; be weak, be feeble' [Kilian-Hatz 2003: 214] (only ǀíní is clearly applicable to inanimate objects).
Naro:ɳǀõ̀ẽ̄ˤ2
Visser 2001: 49. Quoted as ɳǀùíˤ by R. Vossen in [Barnard 1985: 120]. A. Barnard himself lists the equivalent ɡǀǀaba, not confirmed in other sources [Barnard 1985: 120]. Cf. also càīˤ 'thin (persons, animals)' in [Visser 2001: 98].
Visser 2001: 90. Polysemy: 'wind / air'. Quoted as ǂʼã́ in [Vossen 1997: 507]; as ǂãː ~ tʼãː (fem.) in [Barnard 1985: 33] (the second variant is a curious case of idiosyncratic secondary "de-clickification", at least as heard by A. Barnard).
Vossen 1997: 507 (*ǂʼã́). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular (predictable click palatalization and loss in the Eastern branch).
Visser 2001: 54. Meaning glossed as 'worm in manure, beetle larva, caterpillar'; this is the closest word in the dictionary to the required 'earthworm', distinct from ǀxʼã̄õ̀ 'ringworm (in the body)' [Visser 2001: 8] and several even less eligible synonyms.
#Haba:
Not attested.
ǀGwi:
Not attested.
ǀǀGana:
Not attested.
Cara:
Not attested.
ǀXaise:
Not attested.
Danisi:
Not attested.
Ts'ixa:
Not attested.
Deti:
Not attested.
Kua:
Not attested.
Tsua:
Not attested.
Hiechware:ɡǀǀxaie1
Dornan 1917: 95. Other possible synonyms include ǀǀau [Dornan 1917: 101] and bokoː [Dornan 1917: 89], borrowed from Setswana (cf. se=boko 'worm').
Visser 2001: 44. Polysemy: 'year / age / time / birthday / all the time, always'. Quoted as kúdí in [Vossen 1997: 454]; as kúrí (fem., com.) in [Barnard 1985: 113, 125].
Vossen 1997: 454 (*kúdí). Distribution: Preserved in all daughter languages. Reconstruction shape: Correspondences are regular and mostly trivial (with expected palatalization in the Kua-Tsua subgroup).