LATIN INSCRIPTION AT CIRTA. 379 



dissyllabic prepositions, had an accent on the last syllable. But his ear, he says, could detect 

 no trace of one. Yet many of the later grammarians appear to have held this theory, so fre- 

 quent are their allusions to it. We may safely infer from inscriptions, the oldest manu- 

 scripts, the ancient Grammarians and other sources of information that there were hundreds of 

 cases in which writers felt themselves at liberty to unite two or more words into one or to 

 keep them separate. Qui cumque or quicumque, ubi cumque or ubicumque, magno opere or 

 magnopere, ni mirum or nimirum are a few instances out of many. Some other exceptions 

 to these general laws will be noticed in the course of this paper. 



It appears from what has been said that we English in reading Latin place the accent 

 generally, but by no means always, on the proper syllable. But then we have entirely 

 changed its nature, making it a mere stress, instead of a simple raising of the tone without 

 any lengthening of the quantity. And Praecilius and his contemporaries already did the 

 same. From them and their still more degraded descendants the Italians and other western 

 nations inherited this debased accent which had overthrown and usurped the rights of quantity. 

 In the second line of the Aeneid we read Italiam fato pr6fugus with the accent on the right 

 syllable; but on the same principle we ought to say, and Praecilius indeed and the Romans 

 for centuries after him did say, Lavindque. We flatter ourselves that we thus preserve the 

 quantity ; but that is a mere delusion. It we feel by a mere mental process. Whether we 

 pronounce pr6fu(jus or profugus, quantity is equally violated. In the same way we read 

 Greek with this debased Latin accent, and fancy that we preserve the quantity while sacri- 

 ficing the accent. The modern Greeks read old Greek with the ancient Greek accent debased 

 in the same way into a mere stress. We think them, they think us in the wrong ; and in 

 different ways we are both equally in the wrong. M^riviv ae'iSe Gea in an English or Italian 

 and /jLijvtv aeiSe 9ed in a modern Greek mouth are equally remote from the accent and 

 quantity given to the words by Homer or Demosthenes. 



The thing is so manifest, it would be a waste of words to prove that while Greek was a 

 living tongue, metre was determined by quantity alone, and that accent had no influence on 

 it direct or indirect. In Homer or any other poet verses may be found with identically the 

 same cadence, flow and structure, in one of which the accent shall in every foot agre6 ; in 

 another shall in every foot disagree ; in a third shall sometimes agree, sometimes disagree 

 with the metrical ictus. But in prose as well as verse quantity was of far more importance 

 than accent. This is attested by every technical writer on the subject, from Aristotle down- 

 wards. In the third book of his Rhetoric he gives elaborate directions about the rhythms 

 suitable for the different styles of prose, whether it be an iambic, trochaic, dactylic or 

 paeonic rhythm ; but says not one word of the accent. With Dionysius too accent was quite 

 subordinate. The due proportion and due admixture of iong and short syllables were all- 

 important. 



Nearly the same may be said of Latin. Their poetry from the most ancient recorded times 

 was purely quantitative ; the old Saturnian verses quite as much so as the Aeneid. And in 

 prose too quantity was far the more important element. Cicero and Quintilian attest this as 

 decidedly as Aristotle or Dionysius. The notion of an old lingua rustica in which the people 



