134 
On the Letter *r.* 
and from which we have in Enghsh the compound participle forlorn^ 
utterly lost, Germ, verloren ; in the verb is (where the s is not a ter- 
mination as in he loves, but a radical), plural arne and are; in was, 
O. E. ives, plural ware(n) or more commonly weren or were, though 
we not unfrequently, in provincial dialects, have the r even in the 
singular, as always in the subjunctive. Nor is it only in English 
(and German) that s often stands for the sound of ; it is always so 
in French, and most frequently in Italian, if the s comes between 
two vowels ; and in all probability this was also the case in Latin 
and Greek. Hence in those early Latin forms in which an s stands 
for a later r, the single s — pronounced as a sonant — seems to have 
been used almost exclusively. Such forms are Fusius = Ftirius, 
Maso — Maro (~— mason ?), Icebesum = liberum, plorasit plorarit 
= ploraverit, ocisit -= occirit — - occiderit, verhesit = verberit — verbe- 
rarit, plusimus = plurimus, &c.=" 
Now any one who chooses to sound the vibrant r and then the z, 
or vice versa, will readily perceive that the transition involves no very 
considerable change in the position of the vocal organs. But a 
curious illustration of the affinity of these two sounds has come 
under my notice in the case of a former pupil of mine, who, when 
a little boy, was quite incapable of distinguishing them in his pro- 
nunciation. I have frequently, when he was about eleven years old, 
made him stand by my side and pronounce words containing these 
sounds ; and as he tried word after word, I have found myself quite 
unable to determine whether the singular sound he produced was 
more like rose or roar, Ccesar or Ccerar. 
The surd sibilant s is one step more remote from the r than the z 
is, inasmuch as in the s the vocal chords of the larynx apparently 
cease to vibrate. But as it is impossible always to decide in ancient 
languages whether the sibilant with which we may be dealing was 
pronounced as a surd or as a sonant, we may disregard the distinc- 
tion in quoting a few other instances. Such are the a in aiavix-^a as 
representing! the r of cerumna; the corruption of the classical Busta 
Gallic.a into Portogallo ; the use of r as a nominative termination in 
Old Norse, the corresponding termination in Latin and Greek being 
s (e. g. blindr = caecws) ; the use of r in Irish as the termination of 
* A very different mode from the one here suggested, of accounting for the 
consonantal change now under notice is implied in the supposition that "r was 
perhaps originall}' like English th in those " (see Elementary Lat. Grammar, by 
Mr. Roby, of St. John's, Cambridge, p. 2). The point is not argued; but to 
establish so remarkable an hypothesis, it would seem to be necessary to show 
at least one nation employing this same Hebrew-derived symbol R to stand for 
the sound of th in those, and equally necessary to find numerous cases in which 
r has been changed into this th, or vice versa. Such evidence will be very 
hard to adduce. 
+ So Mr. Paley conjectures : see his note on Eur. Med. 19. 
