79 



xli., 12) I transcribed it fkaSla]. In the same way the word 

 for "house" is heard as [werli], [warli], or [wadli], and is 

 spelt "wodli, wurli," by othor observers, and the word for 

 "man" sounds as [jerli], [jadli], or [jardli]. Such varied 

 transcriptions seem to indicate, in these and many other 

 words, that there is something wrong about the hearing and 

 the method of representation, and this belief was strengthened 

 when I found that Mrs. Newchurch rejected such pronunci- 

 ations as [ka/'la, kadla] or [kaSla] as incorrect. The sound 

 in question appears to be a simple one — an I formed by press- 

 ing the slightly reflexed point of the tongue against the front 

 part of the palate. The tongue is not so much reflexed as 

 for [?•], nor is it pushed so far back on the palate, but in both 

 cases the preceding vowel is indistinct in quality, owing to the 

 peculiarly elevated position of the tongue, so t'lat one feels 

 in doubt whether it should be represented by [a], \q\ [o], 

 or even [e]. I therefore write these words [ka/a, waZi]. In 

 Wirrung I heard the word for "fire" as [kala] or [kolaj, with a 

 very open vowel, but the I appeared fairly normal or only 

 slightly reflexed. Whether [l] was originally [d] is an open 

 question, but very likely it was. The sound is quite distinct 

 from Spanish //, Welsh voiceless II, or the thick, backward 

 Russian /, although nearest to the last-named. 



Then tliere are many words in which [r] is more or less 

 heard before a palatal [«], \f\ or [cf], such as [maniguri] or 

 [ma^fguri, mar^guri], three; [marna] or [ma^^a], big; [kartu] 

 or [ka^u], wife. It is quite possible that here 'also we are 

 dealing with simple reflexed-palatal sounds bearing the same 

 relation to [n] and [t] that [/] does to [1] and [r] to [r]. There 

 is this difference, however, that this strong, reflexed-palatal 

 [^] or [^], or a very similar sound, occurs at the beginning of 

 some words, a position in which I have never heard [r] or \l\ 

 Of course a reflexed-palatal sound such as [/'] or [/] affects 

 a following [n], [t], or [d] so that they become reflexed-palatal 

 also. The same changes occur in other languages which 

 possess the [r] sound. Paul Passy, in his "Changements 

 phonetiques," p. 179, when discussing assimilation of con- 

 sonants, writes [ba?i] for the Swedish [barn]. Barn, "child." 

 In the Swedish text given in the "Principles of the 

 International Phonetic Association," p. 25, Nordan, "the 

 north wind," is transcribed [no:r/an], with the statement 

 that [c?] is an inverted (reflexed) [d]. In the "Expose des 

 principes" of the same association, p. 19, is a Norwegian 

 text, where it is stated that [^i, rd, rn, ri, rs] may be 

 reduced to \t, d, n, I, s], and in the text occur the words 

 [trae:rna], traeerne, "the trees"; ['vakkart], vakkert, "beau- 

 tiful" ; [ko :rn9], korne, ''corn." These words might therefore. 



