STURTEVANT: We all agree. The two points of view that have been expressed, are not in opposition to each other. A complete historical research is not to be attempted before we get to work on the dialects. PROKOSCH: The available historical material is certainly to be utilized. The Managing Committee will surely use it when choosing the communities.

BOLLING: We must communicate to the Historical Societies that assistance is needed now and will be needed later when the speech material has been collected.

STURTEVANT: moved that a secretary be chosen; nominated Graves. Graves was elected by unanimous vote.

HANLEY: moved a substitute for Krapp's motion: That the matter be left to the Managing Committee, or that a special committee be chosen to advise on the choice of localities for study.

MOORE: We are all in agreement, but Professor Krapp's motion is preferable as a definite answer to the question appearing on the docket, and at the same time it allows the use of historical materials in the choice of communities.

BRYAN: moved another substitute: That all readily available information as to the movements of population be utilized when choosing the communities for study.

HANLEY: withdrew his substitute motion.

PROKOSCH: Suggested as an amendment to Bryan's motion that the cooperation of the historical societies be solicited.

BRYAN: That is assumed as a matter of course. But we must avoid delay in carrying out our project.

PROKOSCH: We should emphasize our desire for the cooperation of historians.

KNOTT: seconded Bryan's motion.

KRAPP: Elements of population should not be considered particularly in choosing working centers. They should be chosen for the strikingness of their present dialect. It is better to leave the question open. Naturally we shall use all helpful material. But our historians have not studied the history of the population in sufficient detail to be of much help to the linguists.

LELAND: The Council will endeavor to appoint someone on the Managing Committee who will be competent on the historical side. BRYAN'S substitute motion passed: That all readily available information about sources and movements of population should be utilized in

deciding upon centers to be studied. Main motion not further considered.

II. What action should be taken concerning the use of mechanical
methods of recording speech?

KNOTT: The present conference could not deal with this question. It is one of those which must be left to the Committee. Is this the time

to call for the permanent Committee?

CHAIR: Action of this body is merely tentative in any case, and its function is to express opinions and make recommendations.

KNOTT: moved that decision on Question II be left to a committee composed of Kurath, Greet, Ayres, Russell, and Steinberg. LELAND: This could be only a committee of the Conference, and not a permanent committee.

AYRES: moved a substitute: That the Conference recommends the use of mechanical devices of recording speech and leaves the choice of such devices to the permanent Committee.

KNOTT: withdrew his original motion.

EDGERTON: seconded Ayres' motion.

KENYON: Mechanical devices should not be used exclusively, but should

be supplemented by all the records which can be gotten.

CHAIR: Kenyon's idea can be further considered under the question of the training of field workers.

AYRES' motion, amended to read as follows, was carried: That the Conference recommends, among other means, the employment of mechanical devices of recording materials for the Dialect Atlas. GREET: The Speak-O-Phone is a mechanical device which can be used easily by an inexperienced person. It was used for three months at Columbia University, in making over 400 records, with no mechanical difficulty. This is a better device than the Dictaphone for rough work.

III. What is the relative importance of pronunciation, morphology, syntax, and vocabulary in constructing a dialect atlas? STURTEVANT, followed by CHAIR: The purpose of this Conference is not to decide matters, but to express opinions. You are urged to express your opinions.

BUCK: Do Questions III and IV need discussion?

MOORE: IV does not. As to III, vocabulary is distinctly secondary for our purpose, and the collection of lexical material would take too

much time. Moreover, a Dialect Dictionary is already under consideration by another group.

PROKOSCH: We cannot tell in advance which of these four points will be the most important. In different localities different points may have peculiar interest.

STURTEVANT: Some of the European atlases are based primarily on vocabulary; so there is room for a difference of opinion.

LONG: Vocabulary is confusing rather than helpful in collecting material for speech maps. Pronunciation is the most important element for defining local areas. Words are carried far from their natural habitats in this country.

BOLLING: agreed with Prokosch. All features are of importance. Vocabulary was strongly emphasized in the latest work on Greek Dialects.

KENT: Pronunciation can be recorded by mechanical means. Morphological and syntactical differences are apt to be slighter. Differences of vocabulary are marked, and easily recorded (though not by mechanical means). If we do not consider a point which is the chief feature of other atlases, we destroy a valuable basis for comparison. KURATH: Theoretically, all dialectal features should be included in our study. But if 500 places are to be investigated and each community must be handled, let us say, in a week, for financial reasons or for lack of field workers, then the scope of the material must be reduced. Some phonetic, morphological, syntactical, and lexical matters must be neglected. No definite decision can be reached until after some experience in the field. Perhaps someone will first work up a number of key positions, using a large body of material; then it should be possible to determine which features can be eliminated with the least disadvantage. ARMSTRONG: Question III is one of the most important to discuss, not for purpose of decision, but to let each other know our opinions. If we do not include vocabulary in our Atlas we are going absolutely counter to the method of the French Atlas. But that does not matter if our problem is a different one. We must be practical, and be able to convince the purse holders that our project is not too large to be completed, nor too expensive. If all four of the features mentioned cannot be considered, we want to know which are the more important. ROEDDER: The German Atlas considered only phonology. The present workers on it feel that vocabulary and syntax should have been included. Perhaps 75% of importance should be attached to phonol

ogy, 25% to lexicology, and morphology and syntax might be left to take care of themselves.

LONG: agreed almost entirely with Roedder, but thought that on account of the foreign elements in this country syntax is more important here than elsewhere.

MÜLLER: If we neglect vocabulary, we go counter to the practice of the atlases of France and Italy. 75% of importance should be attached to vocabulary, 25% to other aspects.

KENYON: Our problem is different from that of Europe. We have no old dialects with strongly diversified vocabulary. Therefore vocabulary will be of less importance in our investigation. Pronunciation is more important. Words are borrowed and adapted in phonology to the borrowing dialect. Syntax and morphology are sometimes significant. A phrase like the spellingest child reveals the speaker's dialect.

PARMENTER: It is better to cut down by restricting the territory than by reducing the number of things to study. If a small number of communities are thoroughly covered, a good start will be made. Dissertations and monographs will repeat the process for other places. If we study all phases of speech, we shall enlist the interest of all groups. The Atlas will not be definitive in any case. EDGERTON: endorsed this view, and referring to Kenyon's remarks, said that changes of phonology sometimes result from change of habitat. If we are going to leave out of our investigation the elements which are subject to change, we shall soon have nothing to investigate.

CURME: Syntax is important. Various features of New England syntax may be traced back to usage in southeastern England. Syntax is important because it shows the connection of American dialects with the dialects of definite parts of England.

MOORE: The difference of opinion between Kenyon and Edgerton is fundamental. In his opinion, foreign pronunciations are much more rarely borrowed than foreign words.

CARRUTHERS: supported Edgerton on the basis of his own experience. After three years of residence in southern England he was accepted as a Britisher, but shortly after his return to Canada he dropped all the acquired pronunciations.

HANLEY: agreed with Kenyon. Careful investigation is needed first of all for pronunciation; on the other three points evidence can be

taken down by almost anybody. Moved that the primary emphasis be placed on pronunciation.

KENT and MOORE: called for a vote on the comparative importance of the four features.

BLOOMFIELD: This is an impossible question to vote on. The European atlases are fundamentally unsound, and fail to give a picture of the dialects, because they are made by 'pre-view'. We cannot predict what will be of interest. The ideal is to send out a man for the summer, preferably a native of the region, and have him bring back several hundred pages of text, some phonograph records, and scattered information. From that ideal we must cut down, but not by prejudgment-not by forcing the subject to pronounce a word like wharf which may not exist in his natural dialect. (Bloomfield pronounces it warf, because he knows it only through speakers who use that pronunciation, though in all other wh-words he uses the unvoiced sound.)

BUCK: Though only 10% of the total material gathered may be in syntax, it may be no less important. No definite valuation is possible. JENKINS: agreed with Buck. Prejudgment of the relative importance of the four items is bad. Even the four headings do not cover everything, as for instance word order. It would be bad to neglect vocabulary. In tracing the word abeille in the French Atlas, new and interesting facts were found. EDGERTON, in answer to the objection that vocabulary shifts too much to be studied: The investigator should use his intelligence, and make sure whether the word is native or imported.

KURATH: Some of the most striking differences between class dialects are to be found in morphology. Intonation is of importance regionally. These phases of speech cannot be neglected. It would be premature to pass on the relative importance of the various items at this time.

HANLEY: phrased the motion as follows: It is the sense of this meeting that the primary emphasis in the Dialect Atlas should be on pronunciation, except in so far as discrimination of class dialects involves morphology, syntax, and vocabulary.

RUSSELL: This motion is flexible, and will allow the Committee to vary the emphasis considerably, providing only that pronunciation be the last feature to be trimmed down.

Motion carried by a vote of 28 to 11.

JENKINS: I am not against pronunciation, but against direction of any kind!