[Although a generally accepted solution of the problem of Hittite h has not yet been found, the discussion of various theories adduced has brought out a number of facts and probabilities which must be recognized in further attempts, and which have narrowed down the problem considerably. It will consequently be advantageous to examine such of these hypotheses as have not been definitely abandoned, and to determine, if possible, what features of these must be incorporated in a theory that may hold promise of satisfying all conditions.]

There are three distinct types of French spoken in Southern Louisiana. The first, that with which this paper deals primarily, I have chosen to call Standard Louisiana-French. The term may or may not be appropriate, but it is at least convenient. The correct name would be Creole-French, but unfortunately that name has been applied too often to the Negro-French dialects of Louisiana, the French Antilles, and the Mascarenes for its use to be unambiguous, especially since not all are familiar with the application of the name ‘Creole’ in Louisiana only to the white descendants of European (French, Spanish) settlers. Standard (Louisiana-)French is still spoken by many of the more cultivated whites—particularly by the urban population—of Southern Louisiana, and more especially of what is known as the ‘Southwest Louisiana Parishes’.

In Language 10. 43ff. (1934), R. G. Kent, in his review of L. Bloomfield's Language (New York, 1933), mentions Bloomfield's transcription of the Russian word for city, gorod. Bloomfield writes ['gorot], while Kent would prefer ['gort]. The justification given to Kent by Bloomfield for his transcription is: 'Weakening of unstressed syllables in Russian is sufficiently indicated when the place of accent is given, i.e., ['gorot] but plural [goro'da] tells as much as ['gorǝt, gǝra'da]: in fact, to the persons who know the rule which you cite, it tells more, since each of these transcriptions indicates both the accented and the unstressed forms of each vowel-phoneme, whereas ['gort], for instance, fails to tell whether the second syllable has [o] or [a] or [i].'

As is shown in my paper, ‘The Phonemic Principle’, the only completely satisfactory method of treating the sounds of a language is to classify them according to the elemental types, or phonemes, recognized by the given language and to describe the types in terms of norms and ranges of deviation. The present paper presents such a description. The phonetic techniques employed in the course of the study were aural observation and reproductive experiment (attempting to pronounce words after the informant and getting his reaction to the reproduction). Now these are the techniques employed successfully by children in acquiring their native languages, but also employed without complete success by abnormal children and by most adult foreigners. However, a trained phonetician may ordinarily be expected to succeed if he takes sufficient pains. The only real deficiency in these techniques is that they do not provide a means of describing sounds with final accuracy and objectivity (such as is possible when laboratory instruments are used). In the present case, I am limited to a rough indication of the norms of the different phonemic types and sub-types; the normal deviation is practically nowhere sufficiently great to be measurable by my ear.