The useless and pedantic process of de-assimilation takes other forms, one
of the most common of which is the restoring their foreign plural forms to
words borrowed from Greek, Latin, and Italian. No common noun is genuinely
assimilated into our language and made available for the use of the whole
community until it has an English plural, and thousands of indispensable
words have been thus incorporated. We no longer write of _ideæ_, _chori_,
_asyla_, _musea_, _sphinges_, _specimina_ for _ideas_, _choruses_,
_asylums_, _museums_, _sphinxes_, _specimens_, and the notion of returning
to such plurals would seem barbarous and absurd. And yet this very process
is now going on, and threatens us with deplorable results. _Sanatoria_,
_memoranda_, _gymnasia_ are now replacing _sanatorium_, _memorandums_, and
_gymnasiums_; _automata_, _formulae_, and _lacunae_ are taking the place
of _automatons_, _formulas_, and _lacunas_; _indices_ and _apices_ of
_indexes_ and _apexes_, _miasmata_ of _miasmas_ or _miasms_; and even
forms like _lexica_, _rhododendra_, and _chimeræ_ have been recently noted
in the writings of authors of repute.
Some of these words are no doubt exceptions. _Memoranda_ is preferable
when used collectively, but the English plural is better in such a phrase
as 'two different memorandums'. _Automata_, too, is sometimes collective;
and _lacuna_ always carries the suggestion of its classical meaning, which
makes half the meaning of the word. So again, when the classical form is a
scientific term, it is convenient and well to preserve its differentiation,
e.g. _formulae_ in science, or _foci_ and _indices_ in mathematics; but
such uses create exceptions, and these should be recognized as exceptions,
to a general rule that wherever there is choice then the English form is
to be preferred: we should, for instance, say _bandits_ and not _banditti_.