462 ON LATIN PRONUNCIATION. 



letter, and, oa the other hand, some of the sounds have redundant or, 

 at any rate, equivalent representatives. It would be beside the purpose 

 of the present paper, even if it were possible, to attempt an exposition 

 of the nature and range of the sounds employed by the Romans. The 

 vowels, their most important letters, we know had, each of them, a 

 variety of sounds. These variations, however, did not arise from any 

 connection between sound and quantity. They were regulated rather 

 by the manner in which the letters were combined, and further by the 

 custom of the time. We may reasonably infer the latter to have been 

 the case from the fact that similar changes have been undergone, and 

 are even now in progress, in modern languages — a conclusion which is 

 confirmed by the early Grammarians. In treating of the sounds of 

 letters these writers are invariably silent as to any distinction between 

 long and short vowels, so far as sound is concerned. Indeed there 

 are passages in which it is espressly stated that there is no such 

 distinction. Probus, for example, says, " A autem et ^ naturam sua3 

 vocalitatis, sive correpta3, sive productje, custodiunt." Moreover many 

 instances are cited of words changing their accent or their quantity, 

 but no mention is made of any corresponding alteration in the vocal 

 character of the syllable. Does it not, in fact, seem extremely improba- 

 ble that all words which include syllables of doubtful quantity, or 

 syllables whose quantity depends on position, should have two pronun- 

 ciations almost totally distinct ? There are, it is true, some statements 

 in the writings of the Grammarians which, at first sight, appear to con- 

 flict with this conclusion ; but the apparent discrepancy will, I think, 

 disappear upon a closer esaminatioa of the various passages. One 

 example may suffice. Capella says : " B vocalis duarum Graecarum 

 vim possidet. Nam cum corripitur e Graecum est ut hoste ; cum pro- 

 ducitur 7] est ut ab hac die." The meaning of this quotation depends 

 entirely upon the significance of the word viin ; that it refers here only 

 to quantity may be gathered from the use which Victorinus makes of 

 the corresponding word potestas : — " Potestas est quae in ratione me- 

 trica valet cum aut producta aut correpta est." In fact Capella could 

 not have meant by vim any distinction in sound, for in his time -q and 

 e had become identical in that respect, as Sextus Empiricus informs us : 

 — " Correptum Eta (ait) fieri Epsilon, productum contra Epsilon fieri 

 Eta." On the whole it is fair to conclude that sound and quantity have 

 no necessary connection whatever; but are in their nature distinct, the 

 one from the other ; and further that the vowel sounds depended upon 



