What counts as probative in establishing genetic relationship

What counts as probative in establishing genetic relationship? The case of the Kx'a languages Bernd Heine and Henry Honken Berlin Workshop on "Genealogical Language Classification in Africa beyond Greenberg", Berlin, February 21 -22, 2010

"Although Greenberg’s work represents considerable progress over that of previous writers, it leaves a number of questions open. His approach is largely inadequate for the PROOF of genetic relationship; it can do little more than offer initial hypotheses, to be substantiated by more reliable techniques like the comparative method. In a number of instances, languages or language groups have been placed in a given family solely on the basis of a handful of ‘look-alikes’, i. e. morphemes of similar sound shape and meaning. The Nilo-Saharan family, in particular, must be regarded as a tentative grouping, the genetic unity of which remains to be established. " (Heine 1992: 32; quoted from Lyle Campbell, p. c. )

Campbell (2003: 281) maintains that principal among the methodological tools for establishing genetic relationship are the following criteria: a Regular sound correspondences in basic vocabulary. b Patterned grammatical evidence, involving e. g. "shared aberrancy" or "submerged features". c Careful attention to eliminating other possible explanations for similarities.

I argue that (a) is necessary and sufficient to establish genetic relationship while (b) has high probative value, that is, it may provide supporting evidence to strengthen a hypothesis but is not necessary, i. e. I do not take it to be a sine qua non for a relationship hypothesis: There is no intrinsic reason to believe that genetically related languages must show "shared aberrancies" and/or "submerged features". In short: Regular sound correspondences in basic vocabulary constitute the only safe criterion for genetic relationship.

Individual-identifying items: In searching for computable evidence based on probability of occurrence, Nichols (1996) proposes an "individual-identifying level" as a generally applicable statistical probative criterion. a Lexical example: The probability that two given languages share the PIE-word *widhewa 'widow', i. e. a word with the consonant sequence *w- *y- *dh- *w is in the range of 0. 000 006 25 (Nichols 1996: 50). b Grammatical example: She observes that the probability that any given language will show the form-meaning pairings characteristic of an "Indo-European gender system" is in the range of less than one in a million (Nichols 1996: 50 -1).

Individual-identifying items 2: Problem: The individual-identifying items technique is useful in providing measurable results, but it is not of help in solving the crucial problem of diachronic reconstruction, namely distinguising between genetic inheritance borrowing. Example: French grammaticalisation : English grammaticalisation have the same meaning and the same consonant sequence g - r - m - t - c -l - z - t - n, i. e. nine consonants, in common. The chances that this similarity is not due to historical relationship is in the range of less than one in a billion. Nevertheless, we know that the term arose in French in 1914 and was borrowed in English in the 1970 s, hence it is useless for establishing genetic relationship.

Individual-identifying items 3: Applying the individual-identifying item technique to the data provided by Mukarovsky (1987) for proving that Mande and Chadic form one genetic stock yields a multitude of similarities that are compellingly suggestive of genetic relationship. Mukarovsky, Hans 1987. Mande-Chadic Common Stock: A Study of Phonological and Lexical Evidence. Vienna: Institute für Afrikanistik und Ägyptologie.

Q: Does Khoisan constitute a genetic (genealogical) unit as hypothesized by Greenberg (1963)? A: We don't know. Q: Are all Khoisan languages historically interrelated? A: Yes, beyond any reasonable doubt. Problem: How to separate the inherited from the areally transferred material?

Genetic classification of South African "Khoisan“ (families). Traditional Contempora Classification ry classifica tion Present-day spoken languages Possible genetic units !Xun Hoan Khoisan Northern Khoisan !Xun (or Ju) !Xun (dialect cluster) Southern Khoisan Hoan !Ui-Taa !Xoon, West- Hoan, N u !Ui-Taa Central Khoisan Khoekhoe: Nama/Damara, Hai om, Aakhoe (dialect cluster). Kalahari Khoe. West: Khwe (Kxoe), Buga, Ani; Naro; G ana, G ui, Haba. East: Shua, Ts ixa, Danisi, Xaise; Kua-Tsua Khoe. Kwadi

Traditional distribution of the !Xun language.

Genetic groupings Cluster Dialect and abbreviated reference form 1 Northwestern (NW) 1. 1 Northern N 1, N 2 1. 2 Western W 1, W 2, W 3 1. 3 Kavango K 2. 1 Gaub C 1 2. 2 Neitsas C 2 3. 1 Ju hoan E 1 3. 2 Dikundu E 2 3. 3 x áó àèn E 3 2 Central (C) 3 Southeastern (SE)

Regular correspondences

Vowels (examples)

*o preceding *m (*o/ m) P-KX *o/ m Hoan oa òàm ‘river bed’ SE-!Xun o do hm n o a n o m m n!o a n!o m m !xo àm !xo m ki o ma o ‘to NW!Xun o Meaning do m ‘throat, bed’ river ‘to make’ n!o ‘to limp’ m ‘to cover’ ‘to taste mouth’

The sequence *a(C)o P-KX Hoan *a(C)o a SE-!Xun NW-!Xun o o Meaning n a n o m m n o m ‘springhare’ n!a m n!o m ‘to crawl’ n!a n!o m m tsxa txo m m ám !o m n!o m ripen’ ‘to ‘ripe, cooked’ ‘to tie together’ o m ‘leg’ ‘thigh’

Pharyngeal vowels (*V ) P-KX *V Hoan V SE-!Xun NW-!Xun V tso a a Meaning V tso a ‘to cut grass’ ‘to pluck’ t a ta m m ‘fall, drop (PL)’ n a n o m m ‘springhare’ n o m

Egressive consonants

Voiced alveolar stop (*d) P-KX Hoan *d SE-!Xun d a h e: òàm ‘river bed’ ó e NW-!Xun d Meaning dàa ‘polecat’ -dí ‘female’ de ‘female’ do hm do m ‘mother’, ‘female’ ‘throat, river bed’ do e ‘smoke’ ki dò àrà o ‘leaf’

Voiceless alveolar stop (*t) P-KX *t Hoan SE-!Xun t t t xa txa NW-!Xun Meaning t txa ‘cut/hit’ t a ta m m ta m ‘fall, drop (PL)’ t àm ta m ‘to taste like’ ta m

Ingressive consonants (click types)

The dental click (* ) P-KX * Hoan SE-!Xun NW-!Xun Meaning o a ‘be absent’ negation marker n a m o n o m n o m u u xo be xòbè ‘springhare’ ‘to enter, insert’ ‘to lend’

The palatal click (* ) P-KX * Hoa n SE-!Xun NW-!Xun (W 2) !! n i Meaning n!! i ‘beast, dangerous -mà animal’ n u hn n!!u hi ‘travel by night’ a n na o n u n !!à nnà ‘to glitter’ ‘to be white, light’ !!u n ‘star’

The alveolar click (*!) P-KX *! Hoan ! !a o !a: , !ao SE-!Xun ! !a o !a u !áú !a o !ó !a o ‘cheetah’ !ù úrú !ó !ú NW-!Xun ! Meaning !a o ‘drop’ ‘to throw down’ !ò ‘leg’ !áó ‘to trek’ ‘small leopard’ !u úrú (W 1) !ú ‘nail’ ‘name’

The lateral click * P-KX * Hoan SE-!Xun NW-!Xun Meaning a o áú ‘well’ (adv) à e à é ‘to cut meat’ o e òè ‘still’ ki u ú i ú ‘to tread’ hai ha i ‘to pull’

The bilabial click (* ) * n o a óa n a n à a ‘sky’ a n g à á u áú dza - úí n u i g à a ‘eye’ a ‘duiker’ o n u i ‘friend, other’

The retroflex click (*!!) P-KX Hoan *!! n a o ám SE-!Xun ! n!a o n!a a n n a ‘be big’ à o m ‘thigh’ ‘leg’ àè 'to ‘to dismiss' depart’ !o m ‘leg’ a e 'to ! àè send message' 'to anounce one's departure to' g u NW-!Xun Meaning g!ò o g ò o ‘bow’ ‘to cough

Click types distinguished in the modern Kx’a languages and in Proto-Kx’a Click type Proto- Hoan Kx'a !Xun dialects W 2 C 1, C 2 All other dialects * + + Alveolar *! + + Palatal * + - + + Retroflex *!! - + + - Lateral * + + Bilabial * + - - - Dental

Click accompaniments

Aspiration (*Ch) P-KX *!h Hoan SE-!Xun !h NW-!Xun !h Meaning !h !ho !hú ‘horn’ !ho n !hu n !hún ‘kill’ (SG) !hu i !húi ‘cord’

Glottalization (*C ) P-KX * Hoan SE-!Xun NW-!Xun Meaning én àè e ‘self’ (reflexive marker) o u u ‘to ún ‘stirring stick’ ‘stick’ ón un ‘hunting bow’ enter, insert’

Glottalized nasalization (*[ n]) Hoa SE-!Xun NW-!Xun n * n. C P-KX n a n Meaning n ( n-á); ‘to sit (SG)’ n (W 1) n a n o m m m ‘springhare’ n!a n!o m ‘to ripen’; ‘ripe, cooked’ m n!o m (W 1) n n u u n u hu (W 1) ‘center’

Structure: Non-probative evidence a b c d e f g Four distinct tones and tone registers Verbal number suppletism Nominal number suppletism Word order AVO, SO Word order head – dependent Possessive order possessor – possessee Serial verb construction of the contiguous type

Structure: Supporting, non-probative evidence h Pre-verbal progressive marker *ai Nominal derivational suffix *-sí j Linker *ke

Example f: Possessive order possessor – possessee àm !o ‘my name’ Hoan !Xun, W 2 dialect m !ú my name

Example k: Linker *ke introducing adjunct participants kì, kè + NP Hoan NW-!Xun SE-!Xun ke kò + NP

Probative evidence: a Verbal number suppletism: "Shared aberrancy" b Nominal number suppletism: "Submerged feature" c Noun-like reflexive construction: "Submerged feature"

a Verbal number suppletism: "Shared aberrancy" Hoan Proto-!Xun Meaning !u i *n!ún ‘stand, SG’ g a n *g a ‘stand, PL’ n a (< * n -a) ki - à *g!ho ‘sit, SG’ q!ha u ‘recline, PL’ *g!ho ‘sit, PL’

b Nominal number suppletism: "Submerged feature" Hoan !Xun (W 2) Singular Plural e a LH Meaning Singular Plural a e LH Meaning óá òén ‘child’ dàbà n á n èé ‘arm’ ám àám ‘thigh’ da hmà da h ‘wife, woman’ me kx àò kx ào ‘owner’ da be ‘child’

c Noun-like (possessive) construction with etymological identity (* aenn) to express reflexivity ya Hoan !Xun, W 2 dialect e n ha e his/ her REFL 's/he himself, he

The “Khoisan” families of Southern Africa Traditional classification Proposed families Sub-families (genetic stocks) Northern Khoisan Possible families !Xun (= Ju) Kx'a Hoan Kx'a Southern Khoisan !Ui-Taa Central Khoisan Khoe Kalahari Khoekhoe Kwadi Khoe-Kwadi

The hypothetical Proto. Kx'a territory.

What is missing: Appropriate evidence to eliminate other possible explanations for some similarities.

Any questions?