pality
into contact with Englishmen. Nor did the Romans ever establish their
language (I know not whether they wished to do so) in this island, as we
perceive by that stubborn British tongue which has survived two
conquests.[482]

In Gaul and in Spain, however, they did succeed, as the present state of
the French and peninsular languages renders undeniable, though by
gradual changes, and not, as the Benedictine authors of the Histoire
Litteraire de la France seem to imagine, by a sudden and arbitrary
innovation.[483] This is neither possible in itself, nor agreeable to
the testimony of Irenaeus, bishop of Lyons at the end of the second
century, who laments the necessity of learning Celtic.[484] But although
the inhabitants of these provinces came at length to make use of Latin
so completely as their mother tongue that few vestiges of their original
Celtic could perhaps be discovered in their common speech, it does not
follow that they spoke with the pure pronunciation of Italians, far less
with that conformity to the written sounds which we assume to be
essential to the expression of Latin words.

[Sidenote: Ancient Latin pronunciation.]

It appears to be taken for granted that the Romans pronounced their
language as we do at present, so far at least as the enunciation of all
the consonants, however we may admit our deviations from the classical
standard in propriety of sounds and in measure of time. Yet the example
of our own language, and of French, might show us that orthography may
become a very inadequate representative of pronunciation. It is indeed
capable of proof that in the purest ages of Latinity some variation
existed between these two. Those numerous changes in spelling which
distinguish the same words in the poetry of Ennius and of Virgil are
best explained by the supposition of their being accommodated, to the
current pronunciation. Harsh combinations of letters, softened down
through delicacy of ear or rapidity of utterance, gradually lost their
place in the written language. Thus _exfregit_ and _adrogavit_ assumed a
form representing their more liquid sound; and _auctor_ was latterly
spelled _autor_, which has been followed in French and Italian. _Autor_
was probably so pronounced at all times; and the orthography was
afterwards corrected or corrupted, whichever we please to say, according
to the sound. We have the best authority to assert that the final _m_
was very faintly pronounced, rather it see