1076 Wisconsin Academy of Sciences, Arts, and Letters. 
satisfactory summary, was impossible for two or three reasons,— 
these dialects do not pertain directly to the subject and the ma¬ 
terial is too uncertain. What material is one to select to be sure 
that it is trustworthy ? Even if we can feel reasonably certain of 
having a text as the writer originally wrote it, all the old poets 
have many literary or dialectal forms,—Rustico di Filippo has 
such words as merze, auciso, sengnore, caunosciuto, etc.*—and, 
as for the prose-writers, the most unlearned, as soon as he took 
a pen in his hand, seemed to feel within him the spirit of the 
Roman of old and Latinized his words, if he was not already 
confused by dialectal forms. What little material of my own 
I have used for examining the dialects has been largely gath¬ 
ered, on the basis of reliability, from Monaci’s Crestomazia and 
from editions of documents in the various journals,—-the 
Archivio Glottologico, the Propugnatore, the Archivio Storico, 
the Zeitschrift fur romanische Philologie, and so on. Many 
of the documents which were examined are not mentioned in 
the following pages, because they added nothing new or ap¬ 
peared of dubious reliability. So the treatment of the dialects, 
such as it is, is most discouraging and probably there will never 
be an entirely satisfactory treatment. However, whatever 
may be the objections in detail to the conclusions as to the 
dialects and also as to the development in Italian, there seems 
to be no sufficient cause to doubt the general laws of develop¬ 
ment. 
6. The objection may be made that chronology has not been 
strictly followed. My excuse, insufficient as it may seem to 
some, was the fear of further complicating an intricate sub¬ 
ject by robbing it of the advantage of an occasional alphabetic 
arrangement. It is the writer’s impression—an impression 
not capable of indubitable demonstration—that the early docu¬ 
ments represent the Florentine dialect in the actual process of 
certain phonetic changes, which only later completed their de¬ 
velopment,—this appears to be the case in certain instances of 
vacillation between E and I. 1 In general, however, the state 
§6. Note 1. Parodi seems to hold the same view. In a discussion 
of the language of the Frammenti di un lihro di hanchieri fiorentini 
