Apart, however, from the matter of context, one may make a rough classification of the romances on the ground of these references. Leaving aside the few narratives (e.g. _Sir Percival of Galles_, _King Horn_) which contain no suggestion that they are of secondary origin, one may distinguish two groups. There is, in the first place, a large body of romances which refer in general terms to their originals, but do not profess any responsibility for faithful reproduction; in the second place, there are some romances whose authors do recognize the claims of the original, which is in such cases nearly always definitely described, and frequently go so far as to discuss its style or the style to be adopted in the English rendering. The first group, which includes considerably more than half the romances at present accessible in print, affords a confused mass of references. As regards the least definite of these, one finds phrases so vague as to suggest that the author himself might have had difficulty in identifying his source, phrases where the omission of the article ("in rhyme," "in romance," "in story") or the use of the plural ("as books say," "as clerks tell," "as men us told," "in stories thus as we read") deprives the words of most of their significance. Other references are more definite; the writer mentions "this book," "mine author," "the Latin book," "the French book." If these phrases are to be trusted, we may conclude that the English translator has his text before him; they aid little, however, in identification of that text. The fifty-six references in Malory's _Morte d'Arthur_ to "the French book" give no particular clue to discovery of his sources. The common formula, "as the French book says," marks the highest degree of definiteness to which most of these romances attain.