68 



NA TURE 



[July 15, 1922 



containing clicks, and two of these — isigqoko " hat 

 (adapted from isigcogco " head-ring "), and uTixo 

 " God "• — are not of very recent introduction. As, 

 however, it may be objected that great linguistic 

 changes may take place in seventeen years, we have 

 examined a copy of the native newspaper, " Ilanga 

 las' e Natal," dated February 17, 1922, and find, in 

 two columns, averaging about 175 words each, 31 click- 

 words, excluding repetitions and proper names. Of 

 these, all, with the possible exception of three (two of 

 which may be wrongly printed), are either to be found 

 in Colenso's Dictionary, or are obvious derivatives of 

 words there given. 



It is strange to see -gwidu- " rat " given as peculiar 

 to the Swazi dialect, when igundane is very commonly 

 used in Natal Zulu. Similarly, on p. 86, we have, 

 apropos of the Sesuto -liba " (deep) water," the note, 

 " This is a very interesting penetration far to the south 

 of the Zambezi of a root which is very archaic {-ndiba 

 or -Jiba), and particularly characteristic of the N.W. 

 Bantu." But surely it is the same word as the Zulu 

 isi-ziba, the Ronga tiba, the Swahili ziwa, etc. 



This survey extends over five chapters and is followed 

 by a similar review of the Semi-Bantu languages, after 

 which we have a highly controversial chapter on 

 phonetics and phonology. The note on p. 215 we may 

 leave to be dealt with by scientific phoneticians, but 

 must protest, in passing, against the dictum that 

 " proficiency in speaking an African tongue exactly as 

 it is pronounced ... is only to be acquired by a 

 parrot-like imitation of the natives." While " parrot- 

 like imitation " can only be compassed by those 

 possessed of a really good ear, a faultless pronunciation 

 can often be acquired even by persons of inferior ear- 

 capacity, by attending to the instructions of the 

 phonetician. But this presupposes an analysis of the 

 sounds carried out with that meticulous accuracy for 

 which our author appears to entertain so great a con- 

 tempt. Under the heading " Lingual-palatal-sibilant 

 (p. 217), no notice is taken of the fact that the symbols 

 c, j cover at least two different sounds, one of which, 

 the palatal plosive, is not a compound consonant 

 " composed of a blending of / and sh." The difference 

 is important, because sometimes, as in Chinyanja and 

 Zanzibar Swahili, it serves to discriminate between 

 otherwise similar words. Perhaps, however, this point 

 is covered by what is said on p. 222 as to the palatal- 

 ising of d and /. The final paragraph of this section 

 (p. 219) fails to make clear the distinction between 

 sentence-intonation and significant word-intonation. 



Sir Harry Johnston seems inclined to agree with 

 Prof. Meinhof as to the probable absence of vowel- 

 roots in Proto-Bantu. " A comparison of all the 

 recorded forms often leads to the deduction that the 

 NO. 2750, VOL. I io] 



oldest root of two syllables commenced with a con- 

 sonant, very often a guttural." This term is now 

 disused as not sufficiently precise — it would cover 

 velar, uvular, and faucal consonants. It is not quite 

 accurate, however, to say that Meinhof in all cases 

 " replaced the dubious or missing consonant by a 

 gamma (7)." He sometimes postulates y (bilabial v) 

 and has left the question open for at least sixteen stems, 

 where he was unable to decide what the primitive 

 consonant could have been. It is not quite easy to 

 see what is meant by the next sentence : " My own 

 researches, however, lead me in restoring the missing 

 consonant to greater definiteness ; to a g instead of a 

 7, a k instead of an x, a labial instead of an aspirate." 

 The fact that the form yenda, for instance, is found in 

 a small group of languages (only Shambala and Pare- 

 Gweno, so far as I am aware), while genda, jenda, and 

 enda are common, coupled with the greater difficulty 

 of pronunciation of the voiced velar fricative — a 

 difficulty which seems to be felt very generally in 

 Bantu — seem to indicate that Meinhof may be right 

 here. 



The chapter dealing with " Prefixes, Suffixes and 

 Concords connected with the Noun " is of great interest. 

 Sir Harry appears to show convincing reasons why the 

 fi- or pi- diminutive class (Meinhof's 19th) should be 

 identified with the 8th (vi-) instead of maintaining a 

 separate existence. That it is singular while 8 is 

 plural constitutes no objection, since we find 14 (bit-) 

 fulfilling a similar double function — or rather being 

 treated as plural in some cases (Luganda, Herero, etc.), 

 while in itself it is, strictly speaking, neither singular 

 nor plural. With this example in view, it seems to 

 us that it would have been more logical to place fi-, etc. 

 under 8 without creating for it the special subdivision 

 of 8a. This prefix occurs in Karanga (as noticed on 

 p. 75) in the form swi- (or rather <rz-. with the peculiar 

 " whistling s ")} with vu- corresponding to it as plural. 



We should have thought it probable that the " hono- 

 rific " prefix ka- belonged to a different class (now lost 

 as such, but leaving traces, e.g. in Luganda, in such 

 words as Kabaka, Katonda) from the diminutive (13). 

 There are indications, in Konda, Lamba, and elsewhere, 

 of a class of animals with the prefix ka-. Whether this 

 was originally identical with the last-named, or had 

 any connexion with the Chinyanja words beginning 

 with nanlta- (as nankabai " hawk "), is a problem which 

 remains for solution. 



It seems a pity to confuse the class of infinitives 

 (verbal nouns) with the locatives in ku-, which should 

 properly be Class 17, though no doubt the prefixes had 

 originally the same origin. Words like kuboko " arm," 



1 a- is the International Phonetic Association's symbol for this sound, 

 which (or a similar one) is written by Meinhof s and by Junod s. Sir 

 Harry Johnston has nowhere noticed it. 



