ANALYTIC ORTHOGRAPHY. 377 
theory of the catenary curve drawn from the consideration of a single detached link hung 
upon two nails,—such factitious forms being less like words than a link is like a chain. 
583. Every English word of three or more syllables requires the vowel x, a, 1, or a syl- 
lable without a vowel, when the structure of the word does not interfere with it, as gra- 
duate, self-sameness, portmanteau, and the difficulty is to decide upon the proper vowel, as 
in candidate, agitate, elevate, expedite, avenue, maladiction,—for vernacular practice can- 
not be controlled by the consideration that the original was an adverb rather than an 
adjective, unless it can be shown that the adverbial form has been preserved in speech, 
and we think it is not. With the spelling we have nothing to do. 
584. Phonetic readings of ‘usual’ and ‘feature, with zh-y-oo, tsh-y-oo, we do not consider 
English, because y of w or yoo has become zh, sh, leaving a vowel without a preceding y. 
(§ 311.) Nor is dif-thong correct, having been dissimilated, (§292—3,) and none but a 
scholar, a greekist, (not Greeceist,) or foreigner, could sanction an English form like 
‘dif-thon-dzhise’ (for dip-thong-ise,) a form which would allow but one word for singing 
and singeing. Farther, a cramberry is no more a cran berry than ‘amber’ is anber, 
‘imbue’ inbue, or ‘aunt’ amt, Latin A*MiTa.* 
585. It is supposed by some that English spelling “corresponded at some time or other 
to the sound of words.” (Miller, p. xviii.) We think not. English never had signs 
for its commonest vowels, and as it is difficult to determine where they occur, the 
orthoepists find it easier to follow the accidents of a spelling which at no time represented 
the language, than to enter seriously upon an inquiry into the laws of English speech. 
We admit that words like action once had o, and ended with own, as in Spanish and 
German, but we doubt whether the on of honest ever formed a part of them, and we know 
that it does not occur in the English of 1858. The vowel of ebb is common enough in 
English, but hardly so common as to occur thirteen times in fourteen consecutive words 
containing sixty elements, thus constituting one-fifth of the whole. Yet it has been 
indicated as occurring thus in the fragment—“ several passages were then inserted, and in 
* We do not recommend our own pronunciation,—forms like tra-vlr, difrns, instnsz, genrl, temprus, being too 
condensed—too Attic, for ordinary use, besides being more influenced by the spelling than the genius of the 
language allows. In looking through the Phonetic periodicals, whilst preparing this essay, we find that we have 
been ignorant of the name of many public characters. To us there was a fictitious Clanricard within two weeks, 
and whilst we know that our two friends ‘“‘Mackay” are respectively Mecé and Mecdj, we do not know the name 
of the poet Charles Mackay, though we have heard him named Mxct. We mispronounced the proper names 
Tyrwhitt, Napier, Hereford, Bowring (a gentleman we have more recently met,) Keightley (which we had classed 
with Weightman,) Howick, Moore, Mavor, Latham, Youatt, Lowth, Houghton (Hoton, which we classed with 
Hough, or Huf,) ‘Aurora Leigh,’ leg? lay? lee? lie? Once, when in Boston, Massachusetts, with a fellow 
traveller, we wished to see a public building of which we had read, named Faneuil Hall, and after discussing what 
we should ask for, we wisely concluded that the natives would not understand us, or would laugh at our 
pronunciation—so we neither saw the building nor learnt its name. 
