A Khoekhoegowab tumblr

What follows are the independent and inclusive (relative to the speaker) personal pronouns of Khoekhoegowab in table form. The comparison between different authors elucidates some of the differences between the authors’ approaches. A few remarks up-front: I have chosen four sources (references here): Olpp, Hagman, Böhm, and Haackes dictionary. The orthography in Olpp corresponds to modern convention throughout, so I won’t specify it separately. Save some cosmetic adaptations (as using the correct grapheme for the lateral click instead of two slashes, as in Böhm), the orthography is directly adopted from the sources. That is, with the exception of the entries from Haackes dictionary that are written between slashes: these are based on his pronunciation outline (according to which entries are ordered), but I have omitted diacritics for nasals. I mostly included them as sort of a guide for the tone pattern, so they should not be read as a broad IPA transcription, as might be assumed:

1) MASCULINUM¹

	1st person	2nd person	3rd person
Singular	tita tiíta tita /tìi̋tà/	sats saáts sa.ts /sàa̋ts/	ǁîb ǁ'ı͠ıp ǁ'aĩ.p /ǁììb/
Dual	sakhom saákxm̀ sa.kḫɇm /sàa̋khȍm/	sakho saákxò sa.kḫo /sàa̋khȍm/	ǁîkha ǁ'ı͠ıkxà ǁ'aĩ.kḫa /ǁììkhȁ/
Plural	sage saáke sa.ke /sàa̋gè/	sago saáko sa.ko /sàa̋gò/	ǁîgu ǁ'ı͠ıku ǁ'aĩ.ku /ǁììgù/

2) FEMININUM²

	1st person	2nd person	3rd person
Singular	tita tiíta tita /tìi̋tà/	sas saás sa.s /sàa̋s/	ǁîs ǁ'ı͠ıs ǁ'aĩ.s /ǁììs/
Dual	sam saám̀ sa.m /sàa̋m̏/	saro saárò sa.ro /sàa̋rȍ/	ǁîra ǁ'ı͠ırà ǁ'aĩ.ra /ǁììrȁ/
Plural	sase saáse sa.se /sàa̋sè/	saso saáso sa.so /sàa̋sò/	ǁîdi ǁ'ı͠ıtì ǁ'aĩ.ti /ǁììdȉ/

3) NEUTRUM³

	1st person	2nd person	3rd person
Singular	– – – –	– – – –	ǁî-i ǁ'ı͠ı'ì ǁ'aĩ.‘i /ǁìì-ȉ/
Dual	sam saám̀ sa.m /sàa̋m̏/	saro saárò sa.ro /sàa̋rȍ/	ǁîra ǁ'ı͠ırà ǁ'aĩ.ra /ǁììrȁ/
Plural	sada saátà sa.ta /sàa̋dȁ/	sadu saátù sa.tu /sàa̋dȕ/	ǁîn ǁ'ı͠ıǹ ǁ'aĩ.n /ǁììn̏/

A few remarks:

First, one might infer that there are differences in interpretation concerning the voicing of the plosives. In fact, Khoekhoegowab does not have phonemic constrasts between voiced and voiceless plosives. As a result, it is somewhat arbitrary if one chooses to write b or p, d or t, g or k. Beach identifies them as rather voiceless consonants, but not strictly so. Modern orthography has rules when to write one or the other, but these are determined by tone pattern and do not refer to actual differences in pronunciation. Don’t forget that the entries from Haacke are not meant as broad or phonemic transcriptions.

Then, it is obvious that Hagman also indicates tone, contrary to the Olpp, Böhm, or indeed modern orthography. Hagman distiguishes three register tones: high (acute accent), middle (no diacritic), and low (grave accent). Tone bearing units (TBUs) are vowels or morpheme-final nasals (that is, consonants). Haacke, on the other side, distinguishes four register tones: very high (double acute accent), high (acute accent), low (grave accent), and very low (double grave accent). The lateral /l/ and the rhotic /r/ can be TBUs, too, as in 'skoli’ /skól̀ȉ/ (school). If this is the case only in foreign words I do not know (the same segments also trigger a morphophonological rule that changes the gender/number marker of masculine singular words if they occur in stem-final position, e.g. /xam+b/ gives /xammi/, not */xamb/).

In Hagman and Böhm, nasal vowels are indicated by a tilde above the corresponding vowels, which corresponds to the IPA indication to the same effect. Modern orthography (as in Olpp’s examples) uses a circumflex. The latter is also used in Haacke’s dictionary, though I did not replicate them here.

Hagman uses double vowels to mark long vowels. Olpp (and modern orthography) uses a macron above the vowel, save the vowel is a nasal, in which case it is long by default. Böhm does not differenciate between long and short vowels. Thus, in modern orthography 'V̅’ and 'V̂’ correspond to the segments [Vː] and [Ṽː], respectively. It should be noted that the two vowels involved in the formation of one long vowel segment may still be TBUs for different tones, as can be seen in several examples in the tables above. As a consequence, the representation of double vowels is indispenable if tones are to be marked. Indeed, Haacke’s dictionary is organized according to the this principle.

In the pronoun 'sa.kḫɇm’ (modern orthography: 'sakhom’), Böhm uses the grapheme ɇ. He adopts this from Beach’s phonetics book of Khoekhoegowab, though beach transcribes it as an ə. Beach calls it a “neutral oral vowel phoneme,” that is very rare and “may be considered an irregular part of the Hottentot vowel system.” (He also notes that in orthography at that time it was designated by any of the graphemes ḁ, e̥, i̥, o̥, u̥. It is clear from his transcription as well as from his description that 'ə" represents a schwa. Hagman does not mention this phoneme; he mentions schwa, but only in another context, namely as an allophonic variation that /a/ undergoes “before the high vowels /i/ and /u/,” so that “e.g., /kai/ "big” is realized as [kəi], and /nau/ “the other” is realized as [nəu].“ Beach notes a similar variation of /a/, but includes /m/ as a segment that triggers it and describes the assimilation as "a tendency to pronounce a slightly closer.” On the contrary, in Hagman’s rendering of the pronoun, 'saákxm̀’, the vowel dissappears altogether, and /m/ becomes the (syllabic) TBU. This, of course, suggests that the /o/ in this pronoun (and also indicated as TBU in Hackes transcription) is very reduced in pronunciation, if present at all.

As a final point, the signification of the apostrophe might cause some confusion. Earlier, it was used to indicate the glottal stop /ʔ/, but this is today written as a hyphen. So, the word 'khoe-i’ (human being) is pronounced as [kʰoeʔi]. However, modern orthography has another convention concerning the glottal stop after the click consonants, that deviates both from earlier convention and the convention of other Khoisan languages. Thus, /ǁV/ (V = vowel) is pronounced [ǁʔV], whereas [ǁV] is actually written as /ǁgV/. While not problematic as a simple convention, it is somewhat confusing when comparing different orthographies. With regard to the personal pronouns, this convention explains a) why all the apostrophies following clicks in Hagman and Böhm simply disappear in Olpp, and b) why in other positions it is replaced by a hyphen. And so, Böhm’s ǁ’aĩ.’i is changed beyond recognition to ǁî-i [ǁʔiːʔi] in modern orthography. The convention to add a /g/ to the click where non is actually produced is due to an articulatory analysis of click consonants. Already Meinhof (as quoted by Böhm) noted:“As a closure at the soft palate is necessary for the formation of the clicks, a g or k is quasi prepared. The listener believes to perceive a g or k, indeed…, while the native disputes, that such a sound is actually pronounced. Upon closer study one has to agree,… What is relevant is the quiet vowel onset.”⁴

But why does Böhm identify an /a/ in “ǁî-i”, anyway, when he writes it as “ǁ’aĩ.’i” (see also all the other pronouns, where /a/ follows the click+glottal stop combination)? To be honest, I am not really sure. Böhm - following Meinhof - distiguishes a “firm” and an “aspirated” vowel onset (spiritus lenis vs. spiritus asper). Thus, the glottal stop and aspiration after clicks belong to the vowel onset. So, I might speculate that directly after the release of the (click) closure and the subsequent production of the glottal stop, the articulators are in a position favorable to produce rather something like a schwa. However, neither does Böhm write a schwa, nor do are all his examples of words with /click+glottal stop/ followed by an /a/.

---

^{1. 'manelik’ in Olpp, 'masculine’ in Hagman, 'Genus masculinum’ in Böhm↩}

^{2. 'vroulik’ in Olpp, 'feminine" in Hagman, 'Genus femininum’ in Böhm↩}

^{3. 'gemeenslagtig’ in Olpp, 'indefinit-common’ in Hagman, 'Genus comune’ in Böhm↩}

^{4.“Da für die Bildung der Schnalze ein Verschluß am weichen Gaumen notwendig ist, wir gewissermaßen ein g oder k vorbereitet. Der Hörer glaubt auch g oder k zu vernehmen…, während der Eigeborene bestreitet, daß solche Laute wirklich gesprochen werden. Bei näherer Untersuchung muß man ihm recht geben,… Das, worauf es ankommt, ist der leise Vokaleinsatz.”↩}

In my latest post, I reproduced Gerhard Böhm’s spelling of the Khoekhoegowab word khoeb “man”: kḫoe.p. Now, there are several differences from modern orthography. There is the p instead of b, and the point separating the word root from the p enclitic. But there is also the ḫ instead of h. Surely, if Böhm simply wanted to represent an [h] he would have written one. As I suggested in the same post, Böhm means to represent an /x/, be it as an indication of an actual speech sound or of a phonological fact, or both.

Now, why does he not simply write kxoe.p, then, as most other scholars do? This would also be consistent with the IPA alphabet, of course, which uses [x] to represent the voiceless velar fricative.

I can only speculate about the reasons. There is a similarity between the graphemes ḫ and h that comes in handy in the context of a language where the phonemes they represent are also tightly related. Writing x instead of ḫ might somehow muddy this relationship. Of course, one has no such choice if orthography is standardized - but Khoekhoegowab orthography wasn’t when Böhm wrote his book.

But where does that ḫ come from, anyway? Did Böhm just invent a new grapheme by adding a breve to an h, and decided that it represent a velar fricative?

As it turns out, he did not. According to Wikipedia, “Ḫ, ḫ (h-breve below) is an extended Latin letter to transliterate.” The languages (or scripts) to transliterate are, according to the same entry, Arabian, Akkadian, Hittite, and Egyptian. A closer look reveals that it is to represent a dorsal fricative in all cases (though the case seems less clear for Hittite).

Following the Wikipedia article, the ḫ transliterates a certain Egyptian hieroglyph, which does not tell us a lot. Specifically, it transliterates the hieroglyph Aa1 in Alan Gardiner’s sign list - an “unclassified” hieroglyph, though Gardiner speculates it to represent a human placenta. The speech sound it represents is, according to Gardiner, “like ch in Scotch loch.” That is, it is the voiceless velar fricative, [x]. I should note that Gardiner contrasts this in Egyptian to the (voiceless) palatal fricative, [ç], which is represented by the hieroglyph F32 (“animal’s belly with teats”), and transliterated as ẖ (“perhaps like ch in German ich”). However, this account is somewhat incomplete: the hieroglyph Aa1 is a uniliteral (or alphabetical) sign. There are also biliteral (as the hieroglyph D43 ḫw, “forearm with hand holding a flagellum”) and triliteral signs (as the hieroglyph S34 ˁnḫ, “tie or strap, especially sandal-strap (as symbol of life known as ‘the ankh’)”) containing ḫ.

Similarly, in the transliteration of Akkadian (cuneiforms), “[t]he phoneme ḫ should be pronounced like ch in German ach and Scottish loch”, according to Huehnergard’s “A Grammar of Akkadian.” Again, this refers clearly to a velar fricative (contrary to Wikipedia that has it as an uvular fricative). (Note that German distinguishes two sounds, both written as ch: [ç] as in ich, and [x] as in ach, though the contrast is not phonemic.) Contrary to Egyptian, which contrasts - according to Gardiner - [h], [ç], [x], and [h’] (the latter being an emphatic consonant), Akkadian’s only corresponding (central, non-sibilant) fricative seems to be exactly [x], as in ḫurāṣum “gold”.

OK, but all this is not as exotic as you’d wish? There must be more? Well, Christmas is just over, but here you go anyway. There is an ancient, extinct, unclassified¹, yet untranslated language, Meroitic, that is also thought to have had a velar fricative. Meroitic had an alphabetical script the graphemes of which were based on Egyptian hieroglyphs (though not entirely, and some of them modified beyond recognition; also, the script represents vowels). This script uses a strongly modified version of the Egyptian hieroglyph Aa1 to represent the velar fricative, and this is - inter alia - found to be transliterated as ḫ ², as in the word dḫe “child” (the meanings of a few words are known).

As it turns out, Böhm has published on several Afroasiatic languages. He has also written a book on the Meroitic language, “Die Sprache der Aithiopen im Lande Kusch”, so there. Of course, I cannot know if any of this is why he chose to use to use ḫ for the /x/ phoneme.

Anyway, speculating a bit is fun.

Pertaining to the contrast between [k] and [k͡x] in Khoekhoegowab, Beach writes:“kx differs from k in two respects: in the first place it is pronounced with stronger aspiration, and in the second place the tongue is removed from the velum somewhat more slowly so that the cognate fricative x is heard while the tongue is being lowered.” That is, [k] is followed immediately by the homorgan [x] without intermediate plosion, thus it is a genuine fricative, [k͡x].

However, Beach also observes that a “variant pronunciation of kx used by some speakers in strong roots and by nearly all speakers at times in weak roots is a strong plosive kh”. Although Beach notes that “current orthography” uses kh for the kx phoneme" he refers to Schulze¹ who “uses the more correct spelling kχ.” That is, Beach has no doubts that /k͡x/ is the real thing - he uses the spelling kx throughout.

By contrast, Johanna Brugman notes in her doctoral thesis,“For my speakers, frication in the affricate can be quite weak and acoustically similar to strong aspiration. This is not the case with the velar fricative, in which production can range from [x] to [χ], but which is always clearly distinguished from [h].” Hence, Khoekhoegowab has a dorsal fricative, the realization of which ranges from velar [x] to uvular [χ], but that is clearly distinguished from the glottal fricative [h]. In contrast, the affricate /kx/ might actually be realized as an aspirated plosive [kʰ]. Hence, while [x] and [h] are clearly contrasting segments, there is no such contrast between [k͡x] and [kʰ].

Hagman’s short chapter concerning phonology just notes that tongue articulation of the latter consonant in the affricates /ts/ and /kx/ is “rather lax, and their release is often followed by slight glotal friction, e.g. [tsh] and [kxh].” In this, he agrees with Beach, who also suggests kxh as narrow transcription for his kx phoneme. Nonetheless, as Beach notes, Nama does not have “two velar affricates distinguished by strength of aspiration” (i.e. there is no contrast /x/:/xʰ/), and so he does not indicate this difference in transcription.

Böhm notes that “affricates in Nama correspond to aspirates in !Gora”, suggesting a dialectal variation. For the word khoeb (in modern orthography) “man”, he writes khoe.p /kʰoeb/ in !Gora, and kḫoe.p /kxoeb/² in Nama. More interestingly, and contrary to Beach, he contrasts ḫ to kḫ. That is, where Beach is concerned with the contrast between the velar plosive [k] from the velar africate [k͡x], Böhm contrasts the latter with the velar fricative [x]. He actually sees a more general pattern here:“Of course, this is not a coincidence, but we will see that the dichotomie friction : plosion is predetermined in the system of click sounds (ǀ and ǁ as opposed to ǂ and !): the "click-like” sounds are paralleled by clicks.“³ In his account, !Gora preserves properties of an older state of the language. Thus, Nama - having an affricate while the [k]:[kx] contrast is neutralized in !Gora - gave in to a need ”(…) to reinstate the vanishing, but psychologically, systemically required affricate"⁴ as a partner of [x]. I am not sure what forces, exactly, Böhm sees at work here, but his argumentation reminded me strongly of Sapir’s paper The Psychological Reality of Phonemes and what he calls “phonetic illusions.” Giving several examples, Sapir argues that differences that are present on a phonological level, but not in the physical production of the corresponding speech sounds, might still be so “real” to the speaker that he nonetheless “perceives” them.⁵ Thus, I read Böhm as saying that the phonetic contrast [x]:[k͡x], all but neutralized in !Gora, is represented on a phonological level (as the contrast /x/:/k͡x/), and found its way back to the surface in Nama.

To conclude, the status of the velar fricative in Khoekhoegowab is somewhat elusive. Standard orthography represents the segment in question as kh, as in khoes “woman.” But as this was already true when Beach wrote his book, orthography is a bad guide (as always). Following the literature I used for this post, it seems that the velar fricative was once of high importance, but has been transformed to a aspirated velar plosive by now. Another interpretation is that the realization of a /k͡x/ phoneme as [kʰ] instead of [k͡x] was limited to certain morphemes in Beach’s time, but that today there is more or less free variation between the two. Nevertheless, if I read Böhm correctly, the analysis of either [kʰ] or [k͡x] as variants of an underlying phoneme /k͡x/ should be kept in mind, phonetic realization regardles.

As I mentioned in my first post, there are only few textbooks on Khoekhoegowab. I won’t deal with literature predating the 20th century. Most of what I’ll be writing about I draw from four books:

• Roy S. Hagman, Nama Hottentot Grammar, a doctoral thesis turned textbook from the 70s.
• Gerhard Böhm, Kḫoe-kowap: Einführung in die Sprache der Hottentotten, a grammar in German language relaying lots of the literature preceding it.
• J. Olpp, Nama-Gramatika, a textbook written in Afrikaans, learner-friendly in that it apparently targets language learners rather than people interested in the more academic issues (i.e. phonology, morphology, syntax, etc.), also contains exercises.
• D. Beach, The Phonetics of the Hottentot Language, an older book from the 30s, and limited to Khoekhoegowab phonetics (as the title indicates), but still a “classic” on the topic.

As you can guess from the titles, the term Khoekhoegowab was not used as recently as a couple of decades ago. Reasons include that it is the language of various ethnic groups and, correspondingly, that there are various dialects. On the other side, the term Hottentot is elusive and considered derogatory today. I won’t use it other than in quotations. Also, you will notice that second comes the name Nama, while the name Damara - one of the major ethnic groups speaking Khoekhoegowab - is absent in the titles. This is interesting, because even though Hagman’s Nama Hottentot Grammar is exclusively based on his work with Damara speakers, it is still the term Nama that made it into the title.

As concerns this puzzle and the “neutral” term Khoekhoegowab, that thusly covers the whole dialect continuum, I’ll quote from an Interview with Prof. Wilfrid Haacke, one of the most eminent scholars on the language¹:

First of all, the word “Khoekhoegowab” is not a neologism like “Khoisan”, but a revitalization of the original name of the language. Only a few months after he arrived at the Cape, in January 1653, Jan van Riebeeck talked about the Khoekhoe in his journal for the first time. The so-called “Hottentots”, as they were called back then - but today, the word is not used any more - called themselves Khoekhoe. “Khoe” means human being, and Khoekhoe means the typical, the real humans, so to speak. The word “gowab” means language, Khoekhoegowab thusly Khoekhoe language. In Namibia, too, “Khoekhoegowab” was still used when the missionaries arrived in the early 19th century. They came from the South, from Cape Town. First, they worked almost exclusively among the Nama in the South, not among the Damara. Krönlein worked in Berseba for 26 years. He was the man who did most for the language during the 19th century. And in 1889, he published his famous dictionary under the title “Wortschatz der Khoi-Khoin (Namaqua-Hottentotten).” In doing so, he indicated that Nama is a special dialect of Khoekhoegowab, but that there is also something distinct from it. This he knew all too well. Friedrich Rust revised this book. It was published in 1969 and is now titled “Nama Wörterbuch.” The word Khoekhoegowab disappeared at the turn of the century, that is at the beginning of the 20th century, from the literature. This traces back to the fact that the missionaries focused on the Nama, because they settled in the homesteads of Nama chiefs, where they got into contact with more people.

The Damara political system was not as structured. They lived in small groups throughout the country. Only Vedder studied the Damara language and culture by about 1930 for record. But by then, the name “Khoekhoegowab” had disappeared and the language was called Nama. The Damara were very upset about this and claimed that they speak “Damara.” Anyway, the question if the latter is a language or a dialect was a point of contention among linguists. This is because there are no clear boundaries and criteria in this respect, if a dialect is to be regarded as a language. There is a very polemic debate between Nama and Damara that is still carried out in the media. Public authorities suggested to call the language “Nama/Damara”, as a compromise. The former Southwest African Broadcasting Corporation made that “Damara/Nama.” From this alone, one can spot a skirmish. Pastor Eiseb, with whom I have written the dictionary, has long been advocating to dispose of this. It was as nonsensical as “South-West Africa/Namibia.” He could remember that his grandmother talked about Khoekhoegowab, and thus, he suggested this name once again. I added the evidence from the archives. Before independence, this debate was brought to the so-called Nama/Damara Language Committee of the Department of Education, and back then, the majority of the opinions indicated that they’d accept that. As an outsider, I kept out of the debate entirely during the assembly. But at the end, a very prominent Nama member gave his opinion: history had taught them that they’d only ever lost out through compromises. “I speak Nama. Irrespective of whether you speak Khoekhoegowab or whatever, I speak Nama.” Thus, the debate was off the stove for ten years. And after independence, the debate was revived by Pastor Eiseb, and the name Khoekhoegowab was accepted, because said gentlemen was not a member anymore. That is, the name “Khoekhoegowab” has only been revived, after all. It was a historical mistake - that came about only through circumstances involving the work of the missionaries - that the language was called “Nama.” It is a dialect name that was elevated to a language name.

Obviously, I am not in a position to make any pronouncements concerning this dispute. I’ll use the term Khoekhoegowab througout, for the simple reason that is seems to be the current scholarly term.

So…

I have to study some literature on Khoekhoegowab, the most widespread of the so-called Khoisan languages. Unfortunately, its orthography (and, for that matter, its name) had not been standardized until recently. This makes reading older textbooks on the language tough, at times. Also, as research on the language is rather scarce - as compared to, say, French or Japanese - there are not that many textbooks out there; thus, those older textbooks are sort of inevitable.

As a consequence, I have to deal with some non-standardized issues. In doing so, it might be necessary to find out why one author chose to put something - e.g. a phonetic transcription - in one specific way, while another settled for a different option.

Writing down the information I gather to disentangle some of the resulting puzzles is a useful way to clarify and organize my thoughts. In terms of workload, it makes no difference if I put it in a computer file only, or in a blog, too. Also, making it available online might force me, to some extent, to be more thorough in putting down my thoughts. Which is why this tumblr exists, even though it is probably of very limited interest (to say the least).