MÜLLER: cited the groups of 'essential words' studied in the Atlas of Italy: names of countries, proper names, words for child, adult, sickness and health, sorrow and joy, and so on. We might follow their classification.

BOLLING: Marcel Cohen, in his vocabularies for the investigation of aboriginal languages, gives the principal ideas to be looked for in a language.

KNOTT: Depot and station are given as variants in the Word List. These do not vary according to locality, but according to the speaker's education.

KURATH: They should be studied none the less, as indicative of class dialects. It is almost impossible to approach the problem from the general point of view mentioned by Professor Blake. We must start from the details which we already know. Our intention is not to make a semantic study, but to mark dialectic differences in the use of words.

KNOTT: Various Anglicisms such as lift for elevator, luggage for baggage might be studied.

BOLLING: Also phrases such as nothing in it or nothing to it, and idiomatic expressions like a quarter to, or a quarter of.

C. Morphology and Syntax.

KURATH: The material in the Circular is very sketchy. Curme's and Leonard's lists are intended to suggest the sort of things that should be investigated.

MOORE: A suggestion for procedure in distinguishing vulgarisms from colloquialisms: A list could be made of expressions which are generally considered 'vulgar,' as Him and me was there, and of border cases, such as Who did you see, It don't seem that way. The occurrence of such 'vulgarisms' and 'border-line cases' should be the object of a thorough investigation. Expressed the opinion that forms such as don't for doesn't and who for whom cannot be called 'vulgar' if they are used by people who do not use recognized 'vulgarisms.' CURME: The list labeled Some Syntactical Features is incomplete. The English present perfect tense should be added. It is unusual. In other languages the perfect has become a simple past. This process, in which the past idea overshadows the present, can be seen in speakers who say I done it. Another point is the tendency of Americans to refer to inanimate things as she, whereas the English use the masculine. KURATH: We can do little here in this matter. The literature on the

subject must be carefully sifted, and exhaustive lists of morphological and syntactical peculiarities must be prepared, from which certain features will be selected for study.

D. Specimen Text.

KURATH: explained that the Specimen Text in the Circular stood as he had constructed it in 1924. Ayres and Greet had altered it somewhat for their phonographic records of American speech (Victor Talking Machine Co.). Admitted that some sentences were stilted and that some literary words were to be eliminated: certain undesirable features, he said, had been caused by having to insert three test words per line. Queried whether this were a desirable text. Thought that the text should contain much conversation and dialogue. Asked whether the subject matter was to be changed. Pointed out the advantages of a connected text, especially for the matter of intonation; for by the time the subject reaches the second paragraph, he is more normal than he would be if individual questions were put to him. Recognized that the text was to be supplemented by answers to a number of questions by way of check, since the reading pronunciation is not always the same as the speaking pronunciation. Among the disadvantages of a connected text is the subject's tendency to drop into a sing-song.

BLOOMFIELD: A better subject for a text might be worked up-something which a person could tell as though it happened yesterday. EDGERTON: Literary and cultivated words should be avoided as far as possible.

RUSSELL: The intonation of a subject is not natural when he reads. The method of the French Atlas was to submit questions to the subject, and the answer was a natural rebound to the question asked. Do not psychologists say that the written symbol inevitably arouses a control?

METFESSEL: Not if the text is familiar and the reader has dramatic ability.

RUSSELL: would suggest using Gilliéron's procedure (cited above) in addition to the Specimen Text.

ALEXANDER: suggested that a dialogue between two unsuspecting subjects would be valuable.

GREET: It is my experience that the microphone 'scares the subject into being natural.' Impromptu inquiries as to where the subject was born, where he went to college, his present interests, etc., have been

found useful. The Specimen Text in the Circular is 'abominable.' It has been used at Columbia, with a few changes, simply because they started with it and the element of comparison was valuable. Phrases were cited as bad, i.e., unnatural: shirk making a choice, as usually, could not (instead of couldn't), dreary old loft, beams and rafters, joists, fear and horror, hurried out of their holes, the idea of it (those who do not insert an r do not use the phrase), withoug varying his voice, quite dead.

KURATH: These phrases are exactly the ones which I inserted as tests

of one thing or another. I had to distribute the test-words at the time, because I had to jot down the phonetic symbols while the subject read at his usual rate.

RUSSELL: The use of phonograph records will obviate the necessity of having three test-words in a line.

HANLEY: How about using test passages and then something familiar like the Mother Goose Rimes?

KURATH: That would be all right in addition to the Specimen Text. METFESSEL: A specimen text, the same for each subject, is valuable for purposes of comparison. But more natural and less conscious speech should also be recorded.

AYRES: explained that passages were omitted at Columbia simply to cut down the time. A continuous narrative is desirable to get people off their guard.

GREET: suggested phrasing the story, and putting one phrase on each line.

KURATH: Phrasing varies from person to person; forcing it would make the reading unnatural.

GREET: Then at least the passage should be more rhythmical.

KURATH: The passage is very much more unwieldy than it will need to be with mechanical means of recording. It was 'doctored up' with additional test phrases.

RUSSELL: moved that Questions IX, X, XI of the unfinished morning docket be considered together. Carried.

IX. Should the foreign language areas be studied immediately?
X. Should the Atlas include the American Indian languages?
XI. How much attention should be paid to the English of the for-
eign born?

CHAIR: asked for information from an expert on the survey of American
Indian languages, as to its bearing on the present project.
BLOOMFIELD: It has no bearing on it.

KRAPP: Would the fact that early settlers in America spoke Indian languages to some extent have a bearing on the present pronunciation of certain American words, for instance, place names like Chicago? BLOOMFIELD: Loan-words from the Indians present no different problem from that of any other loan-words. No consideration of them is needed in this project (of course, in a complete history of the words in question, the speech of early settlers would have to be considered). RUSSELL: IX is probably included under Krapp's resolution of this morning: that the work shall start in areas which have relatively clearly defined speech characteristics, but any limitation shall be in the number of places studied, not in the whole area to be considered. The Spanish Southwest will, then, be studied eventually. The interplay of the three elements, Indian, Spanish, and English, makes it an interesting and important section to study. In large areas English is not spoken at all.

EDGERTON: Do you suggest that the Spanish of the section be studied? RUSSELL: Yes, eventually. It may disappear soon. It would be less

of a task to include the Spanish than to study the English separately. One individual could handle both.

KURATH: The resolution mentioned said specifically English-speaking areas. It would be hard to get workers who are competent to make out suitable vocabulary lists, etc., for all the foreign language groups in the country. This undertaking should be omitted, for the present at least. The English of these areas, however, must be included. RUSSELL: knew someone who would be competent to do the work for the Spanish section. Estimated that if $1000000 were necessary to get the English data, $1000 would be enough for the Spanish. KRAPP: This task is sure to be done anyway, and should be done much more thoroughly than we are prepared to do it. It is as important for the Spanish as for the English.

HANLEY: agreed with Krapp and Kurath. Cited Mencken's appendix to show that the importance is in the influence of English on the foreign languages, not vice versa. Therefore the interest in a study of foreign language groups is chiefly that of students of these languages.

KENYON: What term are we to use in speaking of this project? STURTEVANT: moved that this project be tentatively called A Linguistic Atlas of the United States and Canada.

ROEDDER: That would include the foreign languages.