262 ROMANCE LANGUAGES 



t 



municate by word of mouth with any of the permanent neighbors 

 surrounding him. From this point of view there is no delimitation 

 of patois, and accordingly, from this point of view, the only logical 

 method of procedure in determining the characters of a given speech 

 is either to signalize all the phenomena appearing collectively at a 

 certain point of territory, or else to follow out and designate the 

 expansion of a given single phenomenon throughout its territorial 

 extent. But considered in the light of actual conditions the facts are 

 not so certain nor the problem so simple, since, in the domain of 

 dialectology, the question "Who is my neighbor?" exhibits pre- 

 cisely the complexity set forth in the parable, and, as a matter 

 of fact, the existence is discovered of more or less clearly denned 

 speech-barriers, determined by political, social, industrial, and com- 

 mercial conditions, as well as by topographical conditions, which 

 insist that account be taken of them. However, it is rather the 

 nature than the delimitation of speech-phenomena that signifies in 

 philology, and the problems of the present for the student of Ro- 

 mance dialects have to do with the analysis and coordination of such 

 phenomena more urgently than with their distribution. 



In the investigation of the Romance dialects and patois, the 

 present march of progress is led by the gigantic undertaking of 

 Gillie'ron and Edmont, entitled the Atlas linguistique de la France, 

 launched a few years since in response to the programme set forth 

 by Gaston Paris in the following words: 



"II faudroit que chaque commune, d'un c6te", chaque forme, 

 chaque mot, de Pautre, eut sa monographic, purement descriptive, 

 faite de premiere main, et traite"e avec toute la rigueur d 'observa- 

 tion qu'exigent les sciences naturelles." 



To any one unacquainted with the plan of this monumental work 

 a word of explanation here will be welcome. Including in its scope 

 the entire Romance-speaking portion of France, together with the 

 outlying speech-territories properly belonging to it, the promoters 

 of the Atlas linguistique have established at approximately equal 

 distances from each other six hundred and thirty-nine stations, 

 at each one of which M. Edmont has collected and recorded pho- 

 netically, with the utmost possible accuracy, the linguistic facts, 

 that is to say, the patois equivalents of words and phrases, cor- 

 responding to a uniform series of questions prepared by M. Gillie'ron. 

 These results are systematically exhibited in a succession of charts, 

 of which something less than four hundred have already appeared, 

 while it is estimated that the completed work will require a total 

 of some eighteen hundred charts to set forth the material collected 

 by M. Edmont in an itinerary of four years devoted to this work. 

 To indicate by a single illustration the class of material afforded, we 

 may choose the record of the patois equivalents of the word honey- 



