the west boundary of northern France. Mr Wicksteed suggests that the words ‘et mentibus Aragonie terminati’ (‘and bounded by the mountains of Aragon’) (lines (61, 62) have become displaced, and ought to be inserted immediately after the clause (lines 52, 53) ‘a Januennium finibus incipientes’ (‘beginning from the frontier of the Genoese’). In that case the term ‘Mountains of Aragon’ does not mean the Pyrenees, but the mountains south of that range which enclose Aragon on the west, south, and east. By adopting this suggestion the difficulty about the west boundary of oïl is removed, and the statement as to the limits of oc is made more precise; for Genoa and the mountains of Aragon (in the sense just explained) give approximately the east and west limits of the territory of oc, including its Spanish portion. It may be mentioned that Raja makes similar transpositions in the text in two other places. (See his larger edition, p. 168, note 2; p. 138, note 5.)
63. 'The term Apennines appears to be here equivalent to the Pennine Alps, which may fairly be taken as forming part of the south boundary of the oïl territory.
CHAPTER IX
[Dante further restricts his survey to the threefold language of the South, and proves by the fact that many words arc common to all its forms tj the language was at first uniform. This leads to a discussion of the cause of the variation in languages, and an explanation of the institution of a conventional invariable literary language (grammar) to obviate the inconveniences due to the mutability of natural speech.]
We must now put whatever reason we possess to the proof, since it is our purpose to investigate matters in which we are supported by the authority of none, namely, the change which has passed over a language which was originally of
