Perhaps typical of their work is an investigation of accent undertaken by Dr. Wilbur L. Schramm, of the University of Iowa in 1935 and 1936.3 He made use of “the microphone, high-quality amplification, the oscillograph, the high-speed output-level recorder and the recording phonograph,”4 and came to the conclusion that accent is a far more complicated phenomenon than the old-time lexicographers ever suspected. “There may be,” he said, “more than one kind of emphasis in speech; a dictionary accent is one thing, an accent which beats the drum for rhythm is another, and a logical emphasis which determines the meaning of the sentence is a third.” One of his associates, Dr. Ruth Ortleb, found that stress itself is by no means an isolated phenomenon, measurable wholly in terms of intensity. It also drags out the duration of a syllable, raises its pitch, and augments its tonal range. And as with syllables, so with words. “In 98% of the cases the emphasized words were of longer duration, in 84% they moved through a wider pattern of inflection, in 75% they reached a higher pitch, and in 71% they reached a lower pitch.” “It is apparent,” concluded Dr. Schramm, “that our old explanations of accent are perhaps too simple,” and that “a complete acoustical description would probably have to take into account at least seven elements: duration of phonation (plus pause, in some cases), magnitude of inflection, highest pitch level, lowest pitch level, average pitch level, average intensity level, and type of inflection.”1 Many other American phonologists now devote themselves to the precise measurements of speech sounds and speech tunes, and the literature of the subject is growing rapidly.2
For many years past philologians have been struggling with the difficulties of representing the gradations of speech in print. No alphabet of any actual language has enough letters to achieve the business, and no artificial alphabet so far contrived has done much more than complicate and obfuscate it. It is, in fact, full of downright impossibilities, as Robert Southey was saying more than a hundred years ago. “Sounds,” he observed, “are to us infinite and variable, and we cannot transmit by one sense the ideas and objects of another. We shall be convinced of this when we recollect the innumerable qualities of tone in human voices, so as to enable us to distinguish all our acquaintances, though the number should amount to many hundreds, or perhaps thousands. With attention we might discover a different quality of tone in every instrument; for all these there never can be a sufficient number of adequate terms in any written language; and when that variety comes to be compounded with a like variety of articulations it becomes infinite to us.”1 Nevertheless, hopeful if imprudent men began to grapple with the problem soon after Southey wrote, and during the 60s of the last century Prince Louis Lucien Bonaparte (1813–1891), a nephew of Napoleon and an amateur philologian of no mean attainments, proposed an alphabet which, at the hands of Alexander J. Ellis, an English phonetician (1814–90), eventually reached 390 characters − 77 for vowels and the rest for consonants and their combinations.2 In 1877 Henry Sweet (1845–1912), another Englishman, reduced the number to 125, but this abbreviated alphabet was quickly found to be inadequate, and improvements upon it were undertaken during the 80s by Paul Passy, a French phonetician. The result was the International Phonetic Alphabet (IPA) of the Association Phonétique Internationale, the latter being Passy’s artifact also.
This alphabet, which includes many new characters hitherto unknown on land or sea, has come into wide use, but its deficiencies are innumerable, and there have been constant changes in it. When, for example, the Practical Phonetics Group of the Modern Language Association adopted it in 1927, it was necessary to add a number of new symbols to indicate peculiarities of American speech,3 and other additions have been since proposed by various other authorities. In 1926 it was given a drastic overhauling by a conference of philologians at Copenhagen, chiefly with the aim of making it more useful for the transcription of non-European languages.4 In 1939 the editors of American Speech petitioned the council of the Association Phonétique Internationale for approval of two new symbols for American sounds devised by John S. Kenyon,1 and at various other times, unless my eyes deceive me, they have slipped in other changes without waiting for a directive from GHQ. Very few practical phonologists have ever attempted to use the IPA without modification. It has become divided into a “broad” form and a “close” or “narrow” form, the former needing fewer symbols than the latter, but suffering a corresponding loss in precision. Daniel Jones used the “broad” form in “An English Pronouncing Dictionary,”2 but he had to borrow four extra vowel symbols from the “narrow” form in order to make his transcriptions intelligible.
In the pamphlets of the British Broadcasting Company he and his colleagues use both the IPA and what they call “modified spelling,” e.g., ǎkwáttic for aquatic, fayt for fête, plaak for plaque, and túrkwoyze for turquoise. No doubt this is a necessary concession to crooners to whom the Greek and Runic letters, the upside-down e’s, c’s and v’s, and the other strange symbols of the IPA would be impenetrable, and perhaps even maddening. Most of the other British phonologists have encountered the same difficulties. Peter A. D. MacCarthy, in his “English Pronunciation,”3 says that the symbols he uses are “in conformity with the phonetic alphabet of the International Phonetic Association,” but proceeds at once to list changes that he has made in it, some of them borrowed from Sweet. In his bibliography,4 listing fifteen works on phonetics by himself and other British authorities, he shows that two use one modification of the IPA, one uses another, six use a third, and five use a fourth. The late H. C. Wyld, in his monumental “History of Modern Colloquial English,” already mentioned, rejected the IPA altogether, and used one of his own invention. “Books about the spoken word,” said A. Lloyd James in “The Broadcast Word,” “all suffer from a serious disadvantage: it is completely and absolutely impossible to represent on paper, by means of conventional print, the simplest facts of speech.… [Its] subtleties are such that no visual symbols can cope with them. The symbol s has to do duty for very many different noises that pass muster, up and down the world, for what is known as ‘the sound of the letter s.’ The ‘sound of the letter l’ has many variants in the English-speaking world, and the l sounds to be heard where English is spoken are legion.”1
With this most American phonologists agree. The late George Philip Krapp made some effort to use the IPA in both “The Pronunciation of Standard English in America,”2 and the second volume (on pronunciation) of “The English Language in America,”3 but in the former volume he sounded a warning to the whole faculty (including himself) that no such artificial alphabet could ever solve the problem of representing all the shades of sound in print. He said:
The professional student of phonetics seems to find it hard to resist the fascination which the game of inventing symbols exerts. The conventional alphabet is obviously inadequate for any scientific purposes, and scores of phonetic alphabets have been invented to take its place. If a phonetic alphabet is an evil, it is a necessary evil. But moderation should be practised in the exercise of this evil, for once started, there is obviously no limit to the number of symbols one may devise as records of his observations. It may be said, moreover, that in the end not even the most elaborate phonetic alphabet can record all the shadings and nuances of speech sounds current daily in good use. For one seeking absolute completeness and precision, some device richer in possibilities than an alphabet must be discovered.4
John S. Kenyon, in his “American Pronunciation,” first published in 1924, used the IPA “in such ways as to adapt it to the peculiarities of American pronunciation,” but apparently found it appreciably short of satisfying, for when he undertook his larger “Pronouncing Dictionary of American English” in collaboration with the late Thomas A. Knott in 19365 he issued a call for suggestions from other phoneticians.6 Whether or not this call brought them anything of value I do not know, but they finally decided to retain the IPA in their dictionary, though with the addition of about twenty-five additional symbols for “less common English sounds and the sounds
of foreign languages,” e.g., y, w and h upside-down, and ? with the dot under the hook omitted. Their explanation of this vitaminized IPA occupied nearly a dozen pages of fine print in the introduction to their dictionary, and must have produced some symptoms of cephalalgia in untutored users of that otherwise able and valuable work. The common, or dirt dictionaries of both England and the United States avoid the IPA with great diligence. Webster 1934, of which Knott was the general editor, offers an explanation of it in the prefatory “Guide to Pronunciation,” written by Kenyon, but in the body of the work it is abandoned for a system going back to Webster himself,1 whereby such homely indicators as ā, ǎ and ä are interpreted by strings of everyday words running along the bottoms of the pages, e.g., āle, ǎdd and ärm.2
There was a time, not remote from today, when the pronunciations ordained by Webster, like his spellings, were accepted as gospel by all right-thinking Americans, and especially by the corps of schoolma’ams, but of late there has been a disposition to question dictionary authority, for the news has got about that language does not follow a rigid pattern but is extraordinarily flexible and changeable. In 1936 Dr. Edward W. Mammen, of the College of the City of New York, printed an analysis of Webster 1934, the Standard, the Century, and various other dictionaries, showing that some of the pronunciations they recommended were not in accord with those then prevailing “in educated colloquial American English.”3 In 1938 Dr. George P. Wilson, of the Woman’s College of the University of North Carolina, followed with another and more devastating paper,4 in 1939 Dr. C. K. Thomas, of Cornell, joined the attack (somewhat mildly) with a third,5 and in 1940 Dr. Karl W. Dykema, of Youngstown College, came on with a fourth, covering both dictionaries and handbooks of “correct” English.1 Even before these attacks, the aforesaid Knott had entertained the savants assembled for an Eastern Public Speaking Conference in New York with a disillusioning account of the way in which dictionary pronunciations are determined, born of his hard experience with Webster 1934.2 The dictionary editor, he said, has to disregard the sounds of words in sentences, and deal with them in isolation, paying no attention to the influence of a given one upon its neighbors and vice versa. Again, he has to frame a pronunciation that will be suitable for the most careful and precise utterance, e.g., that of a public speaker, and pass over the looseness of “ordinary running colloquial.” How, then, are these somewhat pedantic pronunciations arrived at? Knott answered:
We habitually and periodically send around questionnaires containing a very carefully selected list of words with as many as sometimes six different known pronunciations, to just as many people as we dare, people of very widely different classes.… We include a limited number of college and university presidents,… a limited number of linguists and phoneticians,… as large a number as we dare of teachers, heads of departments of speech, [and] … a considerable number of heads of departments of English.… We select some lawyers. We select some of the principal public speakers down in Washington and from various parts of the country. We present them with just as many known pronunciations of these representative words that we have a decent record of, and ask them which one is the one that they prefer.
The answers, said Knott, usually show substantial votes for one or two of the pronunciations submitted. “We find about 125 out of 300 say that there is some one that they use prevailingly. Another hundred habitually use or hear type No. 2, and the other 75 are sprinkled out among the other candidates.”3 Knott said that Webster 1934 did not use the IPA because “hardly anybody knows it yet.” “Take it out,” he went on, “to the first twenty-five teachers that you will meet in an ordinary city public-school system, and if there is one of them who has ever heard of it I should certainly be very, very greatly surprised.”1 And then, in closing:
We probably get five letters every week from school teachers who tell us that such and such a word is not in the dictionary, and why isn’t it? We write back and say: “It is on page 1041, column 2, eight lines from the bottom.” They don’t know the alphabet.
Krapp’s longing for a device to record speech sounds that would be “richer in possibilities than an alphabet,” lately noticed, was already in process of realization when he wrote in 1918. It was the phonograph, which the International Correspondence Schools had been using to teach foreign languages since 1901. But its use for embalming and studying spoken American had to wait until 1924, when Dr. Harry Morgan Ayres, of Columbia University, presented five records of the national speech at the annual meeting of the Modern Language Association. Three years later he and Dr. W. Cabell Greet began a diligent and systematic making of records at Columbia, and this accumulation went on until the beginning of World War II. The materials were obtained from various sources. In some cases radio speeches by eminent public characters, e.g., Roosevelt II and his lady, the Hon. Frances Perkins, General Hugh S. Johnson, Nicholas Murray Butler and Dorothy Thompson, were recorded, and in other cases volunteers from various parts of the country were recruited to provide specimens of their talk. The latter were found mainly on the campus of Columbia, where 14,000 students are gathered annually for the sessions of the Summer School. The director of this Summer School assigns a tree on the campus as a rally-place for the students from each State, so they sort out very conveniently, and finding individuals who spoke the dialect, say, of Columbia, S. C. or Newburyport, Mass., was thus easy. Ayres and Greet were presently joined by Dr. Jane Dorsey Zimmerman, of Columbia, and advanced students were put to work, apparently as kitchen police duty, transcribing the phonograph records in a modified form of the IPA. These transcriptions were printed in American Speech under the editorship of Dr. Zimmerman, beginning in February, 1933, and continuing to the present day. In 1936 she brought out a collection of them running down to June of that year,1 and in 1939 there was a revised edition coming to December, 1939. Most of the student guinea-pigs were set to reading a little fable called “Arthur the Rat,” originally drawn up by the English phonologist, Henry Sweet, to exhibit all of the customary sounds of English, and later improved by Dr. Hans Kurath and Dr. Greet.2 By 1935 records of the speech of no less than 3200 speakers were on file, most of them obtained either from the radio or at Columbia, but also including a number made on trips of exploration by Greet and others.3
Meanwhile, the Linguaphone Institute in New York undertook, on a large scale, the teaching of both foreign languages and English pronunciation by phonograph,4 and in 1943 it added an American English conversation course in which the teachers were Ayres, Greet, Mrs. Zimmerman and various other phonologists.5 It also offered a course “in standard American pronunciation” prepared by Dr. Ray E. Skinner. Finally, it offered two courses in “The Sounds of English,” one based on American pronunciation, prepared by Greet, and one based on British pronunciation, prepared by James. In 1941, cheered by these experiments and doubtful about the usefulness of the IPA, Dr. Bert Emsley, of Ohio State University, proposed boldly that the pronouncing dictionaries of the future be not printed, but recorded on phonograph plates.1 He said:
Pronouncing dictionaries have gone about as far as they can go, in print. Letters cannot duplicate a sound; they can only symbolize it. This situation resembles an older stage, when description of the sounds, which could not symbolize or represent them, had to be replaced, or at least supplemented, by respelling. Just as the Eighteenth Century advanced from description to respelling, it now appears that the Twentieth Century must go ahead from visible to audible devices.… Let us imagine an apparatus which will give (with or without general lexical information) an understandable and reputable pronunciation of any word on demand. Ultimately we should be able to hear the desired word alone, without going through asphasia to get to asthma. A list is not sufficient – there must be some way of sounding and repeating a wanted entry in isolation.
So far this clarion call has not been answered, but no doubt inventive men are busy with it.2
1 See AL4, pp. 3–12, and Supplement I, pp. 1–33.
1 Encyclopedia Americ
ana; New York, 1932, Vol. XXIII, p. 180.
1 Down to 1814 it had sold more than 3,000,000 copies, and down to 1889 62,000,000. Mrs. Emily E. F. Skeel, the lexicographer’s great-granddaughter, tells me that there were 304 editions before 1829. See AL4, p. 385.
American Language Supplement 2 Page 7