Notes, [MTG] 125-2370

t t l / g»„Beach: The Phonetics ox the Hottentot Language (Cambridge, 1938, W. Ileffer & Son3 Ltd., pp. XV * 329). This book is the work of an extremely fine practical phonetician, a man who has been trained, and who lias further trained himself, to observe and record minute shades of quality, quantity, a>.ui pitch, and a man, moreover, who is not betrayed by the complexity of his material into forgetting the main task of the linguist: to give a clear account of the fundamental structure of the language. The greater part of the book is taken up by a careful &nu detailed description of the phonetics and tonetics of the two :crin dialects of Hottentot, Kama end fora nr including Crique. In addition to this, there are chapters on comparative tonetics, the phonetic history of Hottentot roots, and orthography. 'Finally, three appendices: a study, by Dr* K* tedder, of the phonetic structure c± Hottentot roots; a sample of Griqua-Af rika&ns; end a list of the "personnel of the research''. In his introductory chapters Beach sets out hie technique ant; his method. I he principal means of investigation has been, of course, the "ear-method", writing down from dictation or from phono- graph records; but the findings obtained in this my have b^on supplem ented and controlled by extensive use of palato,graphy and kymography. The general impression left on the reader is that Beach has handled Ms technique competently, and that his phonetic lltffil! . B.H. 2

Ltotf&sCc

li. H. Z and tonetic results are dependable - an Impression which is further strengthened by the fact that the author carefully explains what he considers to be the limitations of his work, and outlines a programme for a continuation; he has not had the time or the op- portunity to finish his analysis of length, stress, and, particular- ly, pitch. It is to be hoped that he will soon be enabled to go back to this important task; the enthusiasm and the capability to which this first instalment bears witness, has linked forever the rane of Beach to the study of Hottentot, and has won him the right to complet,« the task he has set himself. In this day when, more than ever before, the foundations of linguistics and phonetics are in the melting-pot, the description of a language is r peculiarly difficult undertaking. Analysis and description presuppose v. theory, but the practical linguist, who is concerned with the specific problems of e psxticular language or group of languages, is hard put to it to choose among the bewildering multitude of more or loss different theories presented to him by an over increasing number of general linguists in violent disagreement among themselves. It is small wonder that, faced with this situation, Beach has chosen to ignore completely the theoretical discussions of the last ten years, is own theory of language does not emerge clear- ly, perhaps because he has not yet worked out one that satisfies him; but he has succeeded in presenting hie material in such a way that it can be used without difficulty by linguists of other persuasions* And that, after ell, is the best that can be said of any description

Ulde11

B*H. 3

o 1 a language. •The Hottentot speech-chain is divided into the following units; the sentence, the breath-group (both defined as delimited by pauses, of which further information would be desirable), the "root'* (a somewhat surprising term fox what the Prague School call a morpheme, en.L\ wh: t Hielmalev and I have called a oenia ), the syllable , and the phono, ihe “word* is rightly discarded as a unit of no phonetic value for Hottentot, but for the sake of convenience the current division into words is indicated by hyphens between "roots'* which make up one "word"• fhe syllable Is defined, “roughly*, as Be part of the chain of speech which receives a single beat in the rhythmic system", ana we are told that "in Hottentot •• the phonetic and (especially) the tonetic structure of the language reveals a very definite system of syllable division, at any rate for slow and careful speech"* But no such rhythmic principle emerges from what is said later; on the contrary, the syllable is taken for granted arid U3ed to define phon- ernes as syllabic or not. In the opinion of trie reviewer, there are no grounds for the establishment of a syllable other tinn the "root", 'here con be no doubt that Beach is right in fixing on the "root" as the tone unit; it is true that this lias been definitely established only for the "strong roots" (i.e. stems), but the tonetic behaviour of the "weak roots" (formants), as it can be studied in the excellent though not very plentiful texts printed in the book, makes it evident

UlSall

B.H. 4

that they exe to be treated in the same way. As it is also evident that the Hottentot tones are not modulations (i.e. prosodernes capable of characterising a complete utterance) but accents (prosocemes : cle of characterising only a unit small«* than a complete utterance), it follows that the "root” is the Hottentot syllable, the syllable being defined as a unit whose characteristic is one minimal accentual unit' , from this point of view, the structure of the Hottentot "roots’* is a matter of the greatest interest and importance to the science of linguistics, because it affords clear examples, which arc otherwise extremely rare, of syllables comprising two vowels with a consonant in between. ''Roots’* are divided into phones according to the following pi indoles: a consonantal unit occurring either initially or between two vowels in a "root*', is considered a single phone; e vocalic continuum, on the other iumd, is considered two phones ”ii‘ the tongue -position changes materially from beginning to end11 . Vowels and consonants thus receive completely different treatments, and of tliio extraordinary procedure no justification whatever is given either by the author or by the facts as presented to the reader. This must be seen in connexion with the fact that instead of deriving the phone deductively from the phoneme, Beach derives the phoneme in- dnetively from the phone. But any division of the speech-chain that is not based upon significant differences, must be *rbitrsrv. a«

bide li

B.U. 5 -tost phoneticians now agreet because the phonic chain is in itself amorphous. leach’s phoneme definition is f modification of, but no improvement on, that of Daniel Jones* * By phoneme", he says, "is meant a group of rel ted phones (one or more) which c- n be represented unambiguously in a phonetic transcription by a single symbol, "me phone can be represented unambiguously by the seme symbol as another if it never occurs in the same position as the other in the chain of speech". 'his is not only putting the cart before the horse, but it is £■ definition which, if literally applied, would lead to an infinite number of phonemes, because it fails to take into account the overlapping of free variants, as studied particularly by K. p bwirnor and • Swirner4". Fortunately, this oversight has had no in- fluence on the result of the phonemic analysis. because of the admirably clear presentation of the phonetic fact: , it is possible to re-do the linguistic analysis and so to re.uce the inventory of phonemes from 42 (32 consonants and 10 vowels) to 2$ (13 consonants and 12 vowels) for lama. :he 20 clicks can be shown to consist each of a click phoneme (one of the four"types of influx") plus one or more other consonant phonemes (one of the five "types of efflux"), viz. k, h, ?, g , or 'roup kx, all of which, except possibly g , occur also in other combinations, bote that tine triplex roups consisting of a click plus kx conform to Iljelmslov* s lew of complex groups^: a.ll of the four clicks occur in duplex

Uldall

3.11 • 6

groups followed by k, and. k occurs in r duplex group followed by x. The two affricates, kx and to, can be regarded as groups consisting, respectively, of the phonemes k and x and the phonemes t and s, all of which occur separately* finally, m and n must be classified as vowels, not consonants, since, according to beach, they are able to be the only constituents of a "root* characterised by tone, i.e. of a syllable. If it proves possible to identify the nasality of the vowels I, a, o, and u with that which occurs in click groups, which lias been written above with the symbol g , the number of vowel phonemes can be further reduced from 12 to 8, because the nasal vowels can then be regarded as yrsrøps.tasi sequences consisting of the corresponding on 1 vowel phonemes plus the com o • complete inventory of phonemes is then, according to the present analysis: (1) consonants: p,ts,x,h,r,5, (2) vowels: i,o,r,o,u,8,m,n, {i,a,3,u}. Similar reductions are possible in the Horana inventory of phonemes« The Hottentot tones seem to be particularly chaotic and difficult to reduce to order. Beach has wisely chosen to attack the problem by first investigating the tones of the "strong roots” (stem syllables) as pronounced in isolation, and then comparing with those the tones of the same sy 11«-bles in connected speech.} the "weak roots' (formants) cannot be isolated ant; therefore do not

Uldall

B.'I. 7

lemi themselves to the sarae treatment • The result of the analysis is that in Tama all stem syllables in isolation ere pronounced with one of six * inherent tones" (four in Torana), which are considered representatives of as many tonemes. Tut as soon as other "roots", weak or strong, are added, extensive changes of tone set in, so that, for instance, a syllable which in isolation has a low-rising tone, occurs in connected speech both with low-rising, level, and low -falling tones. The rules overnin, ~ these changes have not yet been fully worked out, nor can this be done without a considerable amount of additional material. It is clear, however, that the changes cannot be explained simply as variations within the toneme: each syllable is evidently capable of being characterized by more than one toneme, end there must thus be some kind of tonetic declension, an inflexion of each syllable for tone, parallel to, but not necessarily expressing, the inflexion of nouns for case etc. What is not yet clear is whether this inflexion belongs purely to the plane of expression, i.e. is determined only by influence (direction) from the tones of neighboring syllables, or whet lier it is (also) of a grammatical nature, as seems to he the ease e. in Bfik4. It will have to be investigated also whether all syllables of a given "inherent tone" are subjected to the seme changes under the same tonetic and/or grammatical conditions: it might be that the Hottentot syllables divide into more than the nix, respectively four, tonetic declensions already established. It would seem that the" inherent tones" of both Taiaa and forana could be reduced from respectively six and four simplex tonemes to duplex

uldhil

B.H. '

croups composed of memoers of three tonemes: v high, a mid, t-nc a low; but whether this is feasible can only be known when more material is available and a definitive study of tho tonetic inflexions can be made. In the Introduction Beach expresses a hope that his description of Hottentot may be found useful to comparative philology, the standardisation of Hottentot pronunciation, spelling' reform, and learning to speak Hottentot; no one who reads his book, can fail to realise that it will be of the greatest service not only in the fields envisaged by its author, but also, and particularly, to our knowledge of language in general.

:I.J. did 11.

Ulda II

B.H*

Footnotes. 1. or an. account of the theory of slosae&atics 3oe L. H^olmelev, Ueber die Besiehungen der Phonetik &ur Spracliwiasenacn&t%• rchiv fur vergle ichende Phonetik, Bd. II, 193S* 2* Of. especially Grundfr&gen der Phonometrie* fhonometrische i orsohungen, A.l, Berlin, 19 3&* 3* irecordings of the Fecund International Congress of Phonetic Sciences, Cambridge, 3.936 > p. 53« 4. Of. Ida G. sard: Bhe .Phonetic-and on&l structure of hfik, Gambridge, 3-933 •