instance, borrowed some words from Spanish, and Spanish from Portuguese.
It would be conceivable that a few words originating in Spain should pass
into France, and thence into Italy, but it is quite beyond belief that the
large element which the languages from Spain to Roumania have in common
should have passed by borrowing over such a wide territory. It is clear
that this common element is inherited from Latin, out of which all the
Romance languages are derived. Out of the words, endings, idioms, and
constructions which French, Spanish, Italian, and the other tongues of
southern Europe have in common, it would be possible, within certain
limits, to reconstruct the parent speech, but fortunately we are not
limited to this material alone. At this point the Latinist and the Romance
philologist join hands. To take up again the illustration already used,
the student of the Romance languages finds the word for "horse" in Italian
is cavallo, in Spanish caballo, in French cheval, in Roumanian cal, and
so on. Evidently all these forms have come from caballus, which the
Latinist finds belongs to the vocabulary of vulgar, not of formal, Latin.
This one illustration out of many not only discloses the fact that the
Romance languages are to be connected with colloquial rather than with
literary Latin, but it also shows how the line of investigation opened by
Diez, and that followed by Wölfflin and his school, supplement each other.
By the use of the methods which these two scholars introduced, a large
amount of material bearing on the subject under discussion has been
collected and classified, and the characteristic features of the Latin of
the common people have been determined. It has been found that five or six
different and independent kinds of evidence may be used in reconstructing
this form of speech.

We naturally think first of the direct statements made by Latin writers.
These are to be found in the writings of Cicero, Quintilian, Seneca the
Rhetorician, Petronius, Aulus Gellius, Vitruvius, and the Latin