[Sidenote: A Question of Accent.]
I suppose there is no gainsaying the authority of "general usage" in the
matter of English pronunciation—even when that usage is etymologically
wrong. If there is one instinct in the Anglo-Saxon race which is at
once widespread and admirable, it is surely our instinct to avoid even
the semblance of preciosity; the Prig is justly our pet abhorrence.
Maybe some of us incline to carry this instinct a thought too far; as,
for instance, the educated English lady who, when taken to task by an
American for saying _sónorous_, replied: "We always say _sónorous_; of
course we know well enough that it really is _sonórous_, but it would
sound awfully priggish to say so in every-day talk!" But she was an
extreme example, and, though I still persist in saying _sonórous_,
I am far from wishing to undo the long-done work of that "general
usage" which has given us _bálcony_ (for _balcóny_) and _anémone_ (for
_anemóne_). About _paresis_ I may be in some doubt, for the word is so
young in general use that there may still be time to check the spread
of the illiterate _parésis_. The latter pronunciation does not seem
to me to have been consecrated by sufficiently long usage to have won
indisputable authority; there may be a chance for _páresis_ yet!
There are, however, many words in our language, derived from the Latin,
on the accentuation of which both authority and usage are still divided;
and I cannot think the time past for etymology fairly having something
to say about these. Yet it seems to me that the etymological rule for
accenting such words, as it is commonly set down, leaves a good deal to
be desired in point of logic. It is that syllables which are long by
derivation should be accented, that those which are short should not;
and by it we get _compénsate_, _contémplate_, etc.; but a large number
of recognizedly educated people say _cómpensate_ and _cóntemplate_, and
also have the authority of some excellent lexicographers therefore. What
authority there may be for throwing the accent upon the penult in these
words cannot yet be considered as final.
A word which leads me to an explanation of my idea is _elegiac_—which
the Standard Dictionary now gives as _elégiac_ only, but which used
to be pronounced _elegíac_ by most cultivated English speakers. It
is rather a scholarly word, and I fancy most scholars to-day still
pronounce it _elegíac_; it seems to me that there still hangs about
_elégiac_, as Walker said in his day, a "suspicion of illiteracy."
But, if _elegíac_ is right, why is it right? The rule for accenting
syllables that are long by etymology does not hold good here, for the
_i_ in _elegiācus_ is short, as it is also in the Greek _elegiakós_.
It seems to me so highly probable as to amount almost to a certainty,
that scholarly Englishmen fell into the habit of saying _elegíac_ simply
because they had already formed the habit of saying _elegiācus_. They
accented the _i_ in English because it was accented in Latin; and in
Latin it is accented, not because it is long (which it is not), but
because the _a_ which follows it is short. And, if English scholars
said _elegíac_ from habit, may not the results of a similar Latin habit
be found in our pronunciation of hosts of other English words of Latin
origin?
The rule for accentuation I would propose is this: "If the syllable
which is penultimate in the English word is accented in the Latin, it
should be accented in the English word also; if, however, this syllable
is unaccented in Latin, the accent in the English word should fall
back upon the antepenult." Thus the penultimate _i_ in _elegiac_ is
accented because the corresponding _i_ is accented in _elegíacus_. An
old school-master of mine used to insist upon our saying _Quirínal_,
because the _i_ was long; I maintain that _Quírinal_ is right, because
the second _i_ in _Quirinālis_ is unaccented. This rule would give us
_cóntemplate_ and _cómpensate_ because the syllables _tem_ and _pen_
are unaccented in _contemplātus_ and _compensātus_ respectively. (It
is of no avail to argue in favor of _contémplate_ that the _tem_ is
long, and accented in _contémplo_; our English word is derived from the
Latin participle, not from the first person singular of the present
indicative.) _Désiccate_ would be right on the same principle, and
_desíccate_, wrong.
By this rule of mine we can preserve an English pronunciation as nearly
like the original Latin as it is in the spirit of our language to do;
and, where authority and usage are wellnigh equally divided, this seems
to me worth while.