311 



moulded into their old Celtic shape, by removing the middle aspirations 

 and vowel infections, and otherwise applying the laws developed by 

 Zeuss. And not only the body of the words and roots has to be recon- 

 structed, before it can be useful in any way, but the much harder task 

 has to be attempted of restoring the terminations. As the Celtic languages 

 are members of the Indo-Germanic family of languages, which origi- 

 nally possessed a very rich system of inflections, it follows of necessity 

 that the worn out terminations of the Irish and Welsh must have been 

 preceded by fuller forms analogous to those of the Sanskrit, Greek, and 

 Latin. This is further borne out by the testimony of the Gaulish in- 

 scriptions already deciphered. The a-bases of the old Irish decline : 

 hall, haill, laull, hall [n]. Corresponding forms of the Gaulish inscrip- 

 tions are : -os, -i, -u, -on. The dative plural in Irish ends in a mere h : 

 the inscription of Msraes has matre-ho Nemausica-ho, with a termination 

 ho, only one step removed from the Latin hus. Even where as yet we 

 have not actual forms of Gaulish inscriptioDS to guide us, we must, by 

 the laws of comparative philology, try to gain some idea what they may 

 have been in the Gaulish stage. To do otherwise — to interpret Gaulish 

 inscriptions through the assumption of Irish or "Welsh inflections — 

 would just be as ridiculous as to expect Swedish grammatical forms on 

 a runic stone, or Italian want of inflection in an inscription of Caesar's 

 time. 



Likewise, where the vocabulary of the modern Celtic fails us, we 

 must have recurrence to the other and chiefly the older branches of the 

 Indo-Germanic languages, as the Celtic may have lost, and has actually 

 lost, old roots in use in Gaulish times. Thus dede, ''he gave," from the 

 well-known Indo-Germanic root dd, is on the inscription of Msmes, but 

 such a root is entirel}^ unheard of in the later Celtic. 



The first question which presents itself is the purport of the name 

 Dontaurion. It is clear that this is either a nominative neuter, or ac- 

 cusative neuter, or accusative masculine. Considering the great proba- 

 bility of its being the name of a genius, good or evil, we shall choose 

 the third supposition. The base of it is clearly Bontamio. Since dont 

 would be as odd a form for a root as aurio for a suffix, we are driven 

 to the conclusion that the word is a compound of don + taurio. At first 

 sight there is a slight difiiculty in this assumption, since the Gaulish 

 compounds generally show a vowel at the end of the first word ; how- 

 ever, in itjugdunum, another form of Lugudunum, we have an example not 

 only of the first part ending in a consonant, but of that ending being 

 brought about through the loss of the original vowel u. We are there- 

 fore at liberty to treat the don either as the true form of the base of the 

 first word, or else as a shortening of a base dono, donu, doni, according 

 as the case may require. Assuming dono as the original form, the word 

 bears a strong resemblance to Ir. duine, a man, which points back to 

 donio, the vowel being altered as in Gaulish mori -sea = Ir. muir. Simi- 

 lar alterations of the o by the influence of a following i, we have in Ir. 

 slond, significatio, sluindid, significat ; londas, indignatio, coUuindi, cum 

 amaritudine, etc. {vid. Zeuss, 16, 18), 



