s which produce the growth, or, according to
others, constitute the history of language, is under the control of man.
The phonetic decay of language is not the result of mere accident; it is
governed by definite laws, as we shall see when we come to consider the
principles of comparative grammar. But these laws were not made by man; on
the contrary, man had to obey them without knowing of their existence.

In the growth of the modern Romance languages out of Latin, we can
perceive not only a general tendency to simplification, not only a natural
disposition to avoid the exertion which the pronunciation of certain
consonants, and still more, of groups of consonants, entails on the
speaker: but we can see distinct laws for each of the Romance dialects,
which enable us to say, that in French the Latin _patrem_ would naturally
grow into the modern _pere_. The final _m_ is always dropped in the
Romance dialects, and it was dropped even in Latin. Thus we get _patre_
instead of _patrem_. Now, a Latin _t_ between two vowels in such words as
_pater_ is invariably suppressed in French. This is a law, and by means of
it we can discover at once that _catena_ must become _chaine_; _fata_, a
later feminine representation of the old neuter _fatum_, _fee_; _pratum_ a
meadow, _pre_. From _pratum_ we derive _prataria_, which in French becomes
_prairie_; from _fatum_, _fataria_, the English _fairy_. Thus every Latin
participle in _atus_, like _amatus_, loved, must end in French in _e_. The
same law then changed _patre_(pronounced _pa-tere_) into _paere_, or
_pere_; it changed _matrem_ into _mere_, _fratrem_ into _frere_. These
changes take place gradually but irresistibly, and, what is most
important, they are completely beyond the reach or control of the free
will of man.

Dialectical growth again is still more beyond the control of individuals.
For although a poet may knowingly and intentionally invent a new word, its
acceptance depends on circumstances which defy individual interference.
There are some changes in the grammar which at first sight might seem to
be mainly attributable to the caprice of the speaker. Granted, for
instance, that the loss of the Latin terminations was the natural result
of a more careless pronunciation; granted that the modern sign of the
French genitive _du_ is a natural corruption of the Latin _de illo_,--yet
the choice of _de_, instead of any other word, to express the genitive,
the choice of _illo_, inste