ANALYTIC ORTHOGRAPHY. 361 
SANSCRIT CEREBRALS. 
491. These are thus described by Wilkins (Gramm. 1808,)—“ This series of consonants 
is produced by turning and applying the tip of the tongue far back against the palate; 
producing a hollow sound, as if proceeding from the head.’ Lepsius and Ellis add the 
common Sanscrit 7 of other authors. Wilkins says that in Bengal the d is “pronounced 
like a very obtuse rv.” See §199. We will assume that the Sanscrit + is a common 
trilled 7, and that the Bengali sound is a trilled cerebral (not d but) r, and to be so. 
written. Hichhoff (p. 80) excludes the / as fictitious; Wilkins makes it the Welsh Ji. 
EKichhoff uses his dotted i for a Birman sound, which others consider Polish J. 
492. Lepsius, t d n 8 z r l t a 
Ellis, te de ne fc je re Ic teH dcuH 
Miller, t d n sh r l th dh 
Bopp, t ee i es r Lo 
Kichhoff, 1 D N S R L DE GaeeD Hl 
493. Most of the Sanscrit forms have a horizontal curve below (_) by which we propose 
to denote them, placing it below or after the base letter, as in— 
RoC thy wo) te oan 
494. Another mode is to use the small italic capitals, 7, »,s, &c.,—and Bengali x, which 
would represent the point of contact as removed towards the throat. §471. In Hllis’s 
notation, ‘c’ is a diacrit, ‘k’ being used for the cay power. 
SIGMALS. 
@ 
- Oo a 
8, 3, VY, VY, v%) v, xv, > Ty Ye. 
I 2 8 A 6 8 7 B8 oO WD 
495. The affinity of the aspirates of seize is rather with the dentals than the palatals, 
with which (§ 158) they are often classed. Most authors represent them by s, z, the latter 
being a perversion to which we are exceedingly averse. Bopp uses ¢ for English dz, 
which is also wrong. Although we do not approve of any double letters, z may be wanted 
for Russian zd, a power which is constantly before us in Greek, and perhaps in Hebrew, 
(Hwald.) Compare 6f0¢ and Aeolic Jed0¢; Hormuz and Hormuzd; Oromazes or Oromasdes ; 
Ezra-s or Esdra-s. The use of English z would tend to destroy the etymologic value of 
every word (like ‘ozone’) transliterated with it from Greek, whilst a new character would 
indicate a new sound (‘o30n’) in a corrupted or naturalised form. This objection would 
be weakened if Greek z were transliterated with zd, but we fear that very few would use 
forms like ozdon, or horizdon. 
496. The English or French z is unrecognised in Greek, Latin, and Spanish, and it 
