122 SECTIONAL ADDRESSES 



however, did not fuse with them. Vowel harmony of all the syllables of 

 such a compound often developed as a means of marking off the group. 

 This is Agglutinative speech. It survives in the Bantu tongues {prefixing 

 type) and in the Altaic (e.g. Turkish) as a suffix type. 



The Hamitic-Semitic group carried agglutination so far that the re- 

 lationship-words fused at last with the chief words to which they were 

 affixed or prefixed, and speech became amalgamating. Words were also 

 systematically modified by internal vowel-change to give regular altera- 

 tions of meaning. Dravidian speech is agglutinative — affixing and in- 

 fixing, so that it is rather generalised and may be the ancestor of several 

 of the other main languages. 



The Aryan group developed external suffixes to indicate variations, 

 and so produced infl£ctional language. The three groups Hamitic, 

 Semitic and Aryan also tend to rely on relationship-words and on word- 

 order, and increasingly to neglect the word-forms. Thus they become 

 analytic. There is little doubt in the writer's mind that this sequence 

 (e.g. from Bantu to such an Aryan language as Persian) represents lin- 

 guistic evolution, in much the same sense as the sequence ' amoeba to 

 man ' represents zoological evolution. In both cases many groups 

 branch off from the main stem producing minor independent evolutionary 

 groups. In both cases some descendants stagnate, while others advance 

 rapidly, as stated earlier. 



Before Aryan scholars yield to despair because the foundations of 

 Aryan are ' wrop in mystery,' a promising field would be to explore 

 Dravidian or Altaic for the ancestors of the Aryan. For instance, there 

 are three possible explanations for the accepted resemblances of Finn to 

 Teutonic. The one usually accepted is that Finn has borrowed from 

 Teutonic. But it is also possible that Teutonic has borrowed its ' pecu- 

 liarities ' from Finn. A third view worth considering is that border 

 (e.g. primitive) languages like Keltic or Teutonic, still retain speech 

 characteristics which have been carried over from the more primitive 

 speech (now preserved as a marginal language-zone) from which the Aryan 

 group as a whole evolved. On this view the features common to German 

 and Finn may be an inheritance from the common mother-tongue of 

 Aryan and Altaic. 



The lesson to be derived from the ' Zones and Strata ' technique is that 

 marginal languages should be compared with each other. This means 

 that far-distant speeches may be very well worth comparing. From this 

 point of view, we should actually expect that Basque would resemble 

 some Amerind language ; that Gaelic would resemble Pharaoh's tongue ; 

 and that early Sinitic, early Altaic and early Dravidian (i.e. marginal 

 languages) should be studied to learn something about Proto-Aryan. 

 Thus the writer by no means despairs of the solution of the Aryan 

 problem. 



These ideas have long been engaging the writer's attention. In 1921 

 he published a generalised diagram, which was probably the first utilising 

 the ' Zones and Strata Concept ' as applied to Linguistics. With a few 

 minor alterations it is reproduced as Fig. 12. Here the concept of a 

 central cradleland of culture is adopted. But we must ever bear in mind 

 that we are primarily concerned with events which occurred before 



