17 



of those accents by the marks employed in other languages. The 

 followino: are the Chinese names of the accents: 



o 



1. pi,g fi,g^, "the even, level, or undisturbed tone." This is di- 

 vided into the upper pi,g fi^g", which they define as " long and clear/' 

 marked by the missionaries with a dash over the vowel (as a), and 

 the lower pijg fi^g, defined as " long and low, or obscure/' marked by 

 a circumflex (as a). 



2. fa.g fi^g, " the ascending tone," marked by the grave accent 

 (as a). 



3. cin fi,g, " the departing tone," marked by the acute accent (a). 



4. ji li.g, " the entering tone," marked by the breve (a). 



It would seem, from the definitions, that the second form might 

 be better represented by the acute, and the third by the grave accent, 

 but with this exception the notation of the missionaries probably 

 represents with sufficient accuracy the tonic distinctions originally 

 made by the Chinese. 



The ji fi^g (called in the Canton dialect yap, or yat-fi,g), repre- 

 sents a short, abrupt tone, which, in the Canton dialect, is always 

 followed by the sound of k, p, or t. " It is as if a man sounding the 

 pi»g ri,g, should be suddenly taken with a hiccup, and stop it half 

 way; if the word lock be sounded, but the last two consonants 

 omitted, it gives the yap fi.g."* 



If it be granted that the ji fi,g furnishes an indication of the 

 probable origin of root forms that terminate in a mute consonant, it 

 would follow that such roots must originally have contained only a 

 short vowel, as is generally the case in the Sanscrit and Teutonic 

 languages, •{* 



The hiatus occasioned by the sudden truncation of sound may also 

 explain the lengthened quantity of a short vowel in Greek and Latin, 

 when followed by two consonants. If the final consonant formed a 

 portion of the primitive root, it is difficult to account for this increase 

 of quantity, but if it was merely a softened close for the ji fi^g, it 

 would naturally retain the temporal increase which attends the at- 

 tempt to pronounce a consonant immediately after the hiatus. 



It is desirable in all etymological comparisons to adopt a notation 

 that shall mark every sound with the greatest possible precision and 



=* "Williams. — Easy Lessons in Chinese, p. 50. 



t Even when a Sanscrit root terminates in a consonant preceded by a long 

 vowel, there is usually another form of the root with the corresponding short 

 vowel, which renders it highly probable that all the consonant-ending roots in that 

 language were originally Ji fi^g. 



VOL. VIII. C 



