220 AHsti)i Morris Harmon, 



ation that we have been inchned to overlook, and because it shows 

 that this accentuation was not confined to vulgar speech, but was 

 in vogue in the best circles as early as the fourth century. 



It would be beyond the scope of this treatise to consider how 

 much older than Ammianus the custom of Greek accentuation may 

 be. In any case such an investigation would bear little fruit at 

 present, owing to the paucity of data. Further study of Latin prose- 

 rh3^thm may eventually help to solve the problem.^ 



We come now to the matter of loan-words derived from languages 

 other than the Greek. In the lists of Greek words given above a 

 number of words like Calycadims are cited, which were borrowed 

 from some other language through the medium of Greek : their 

 treatment is of course in no wise different from that of pure Greek 

 words. So in Ldranda (I. 11. 10) and in Artogc'rassa CAQTCcyfjQai'. 

 II. 123. 18) Ammianus' accent agrees with the Greek. The curious 

 name Caenos Gallicanos (I. 38. 1) evidently is influenced by the 

 Greek : it may be corrupt, for Gallicanus, with which it appears to 

 be connected, uniformly has an accented penult in Latin (I. 201. 21 : 

 II. 60. 7 ; 228. 15). Bardxmalcha (II. 5. 24) and Nadrmalcha (II. 6. 24; 

 21. 29) are not to my knowledge mentioned elsewhere, so that we 

 are ignorant of the accent which they bore in Greek. How Am- 

 mianus pronounced Amida is uncertain, for we find both Amida 

 (I. 168. 2) and Amida (222. 6). Perhaps he was uncertain himself 

 in the matter. In Amano (I. 28. 17) and Naessum (II. 71. 13; I. 246. 

 12 ; 247. 30) Ammianus accents the antepenult, whereas in Greek 

 the accent is certainly on the ultima in the one {rb 'Af/avor) and 

 probably in the other {iVicior>6<^). The pronunciation of Abarne 

 (I. 167. 10) with antepenultimate accent is paralleled by the Greek 

 form "i(jf«(>ro^, but the word itself is a transliteration of 'A^uitvii. A 

 similar thing occurs in Bczabde (I. 253. 18; 222. 16), a transliteration 

 of B%iCi^<Stj. Finally, I may cite Gdlatae (I. 68. 18 ; II. 60-61) Sdr- 

 matae (I. 136. 18; 21 ; 32; 137. 11 ; IL 198. 7), and sdtrapa (I. 181. 2; 

 II. 51. 19; 36.32; 121. 12). In how far any of these words may 

 derive its accentuation from the language in which the word itsell 

 originated I cannot attempt to sa}'.^ 



' It may be noted here that in Sedulius the words Cycladas (Huemer 

 p. 171. 11) and (in the oblique cases) tyrannis (p. 210. 11 : 222. 19 : 

 253. 10) have the Greek accent (cf. Candel p. 87). No other loan-words 

 occur in the clausula with him, except perhaps Esaias (p. 274. 4). 



* On satrapa^ however, see Seelman, p. 49. On a previous occasion (i*?//."/;/. 

 Mils. 1907, p. 159) I have hinted tliat the accent of rumpia in Latin may 

 possibly be Tkracian. 



