May 20, 1080] 



NATURE 



51 



discredited the labours of better men. There is much 

 in ctj'mology which must always defy analysis, there is 

 much which will have to be corrected hereafter, but this 

 will matter little if we have once learnt the lesson that 

 change of sound and meaning can only take place in 

 accordance with fixed and invariable law. Etymology 

 is but a means to an end, and that end is partly the 

 history of the development of thought and civilisation as 

 reflected in the fossil records of speech, partly the dis- 

 covery and illustration of the laws which govern the 

 shifting and decay of sounds and the modifications of 

 sense." 



The whole subject of phonetics is of course treated in 

 a masterly manner, and well illustrated with diagrams 

 and useful tables of Lautverschiebung as applicable to 

 the Semitic, Bantu, Finno-Tataric, and Aryan families. 

 The last, especially, is very full, including the Oscan and 

 Umbrian, the Old Welsh and Gaulish, besides those 

 usually given. It need scarcely be added that this, like 

 all other branches, is brought well up to date, a good 

 instance of which is afforded by the reference to the use 

 already made of the phonograph in the scientific treat- 

 ment of phonetics. Most readers will here learn, probably 

 for the first time, the curious fact that " all sounds may 

 be reproduced backwards by simply beginning with the 

 last forms indented on the tin-foil : sociabilily, for example, 

 becoming ytibilaishos. Diphthongs and double con- 

 sonants may be reversed with equal clearness and pre- 

 cision, so that bite, which the phonograph pronounces 

 ba-cct, becomes tee-db. In this way we have learnt that 

 the ch of cheque is really a double letter, the reversed 

 pronunciation of the word being keshl" (i. 335). 



The question of mixed languages, that is, mixed in 

 their structure, claims a good deal of attention, and is 

 handled with considerable reserve. But the important 

 truth is loudly proclaimed that the "physiological races 

 of the modern world are far more mixed than the lan- 

 guages they speak ; the physiologist has much more 

 difficulty in distinguishing his races than has the glotto- 

 logist in distinguishing his families of speech" (i. 366). 

 This is perhaps as far as it is safe to go at present, and 

 is sufficient for practical purposes. It points out that it 

 is in the nature of ethnical groups to mix, and of linguistic 

 groups [to keep aloof, thus vindicating for language its 

 rightful position in anthropological studies. It is not 

 always or necessarily a test of race, but it is often an 

 indispensable collateral agent of research, becomes under 

 special circumstances, and with all due precaution, a 

 final court of appeal, and in many cases bears witness to 

 the presence of racial elements which would not other- 

 wise be suspected. Its development also is extremely 

 slow, slower even under certain conditions than that of 

 physical types themselves, as shown, for instance, in the 

 case of the Osmanli, Magyars, and many Finnish and 

 Turkoman tribes, all of whom continue to speak purely 

 agglutinating Finno-Tataric tongues, although through 

 intermixture they have been largely assimilated to the 

 Caucasian ethnical type. 



The chapter on Roots (vol. ii.) is accompanied by a 

 table of all known languages, for the classification of 

 which Fr. Miiller seems mainly responsible. The ap- 

 pended references to authorities will be found extremely 

 useful, but the [classification itself is defective in many 



respects, and calls for revision in future editions. Sonrhay 

 and Ilaussa, for instance, ought not to be grouped 

 together, nor have Wakuafi (read Ki-Kwafi) and Masai 

 anything in common with the Nuba and Fulah groups. 

 It is not clear why Berber any more than Egyptian (both 

 Hamitic) should be described as sub-Semitic ; but it is 

 still more startling to find Brahui amongst the neo- 

 Sanskritic tongues in company with Siah-Posh, which 

 latter would appear to belong ratbcr to the Galcha or 

 pre-Sanskritic of the Eastern Turkestan Highlands, and 

 which is unaccountably excluded altogether from the 

 table. Etruscan, in spite of Corssen, is grouped apart as 

 agglutinating, though there are many good authorities for 

 this view. But Horpa is not a Tibcto-Burman isolating 

 tongue, nor are Lolo and Mautse properly linguistic 

 terms, but rather collective Chinese names of hill-tribes, 

 mostly probably of Caucasian stock and untoned speech. 

 The " Mon-.Annam " family has no existence, the Mon or 

 Talain having little to do with the Annam, and nothing 

 at all with Kambojan, which belongs to a totally different 

 connection. The Miztec, Matlalzinca, Totonac, and other 

 Mexican tongues are described as isolating, all being 

 polysynthetic, some, such as the Miztec, in the very 

 highest degree with "bunch-words" of fifteen and even 

 seventeen syllables. 



The second volume is largely occupied with some of the 

 principal linguistic families typical of the several orders 

 of speech, followed by concluding chapters on Compara- 

 tive Mythology and the Origin of Language, all handled 

 in a masterly manner, extremely suggestive even when 

 somewhat heterodox, and accompanied by much inci- 

 dental matter of great value and interest. The statement 

 (p. 324) that " the characteristics of race were fixed before 

 the invention of speech " is one of those astonishing 

 paradoxes which seem inseparable from original thought^ 

 but which remain none the less paradoxes. It is scarcely 

 conceivable that the yellow, black, fair, and other funda- 

 mental types of mankind should have become slowly 

 differentiated before man had acquired the faculty of 

 speech, that is, the very faculty by which the human is 

 distinguished from all other species, and that the art was 

 then "invented" in various independent centres. But 

 though it cannot be argued on this ground that "the idioms 

 of mankind have had many independent starting-points"' 

 (p. 323), few will probably question the conclusion that 

 linguistic science "can throw no light on the ethnological 

 problem of the original unity or diversity of the human 

 race" (p. 324). Such questions arc truly "the task of 

 the ethnologist, not of the student of language" {ib.) 

 And even should the hope have to be finally abandoned 

 of ultimately establishing the original un"ify of human 

 speech, no argument could thence be deduced in favour of 

 the original diversity of the human species. Dispersions 

 of babbling tribes, whether originally one or not, probably 

 took place at various stages in the evolution of human 

 speech, or at times while it was still in process of forma- 

 tion, or when little more than the faculty existed, so that 

 it must needs have afterwards developed into types no 

 longer reducible to one hypothetical primeval type. This 

 hypothetical type becomes daily more shadowy, continually 

 retreating to the background of an inconceivably remote 

 past, according as the astonishing complexity and diversity 

 of articulate speech is revealed to the earnest student of 



