¹ The linguistic atlas project was initiated in 1930, under the sponsorship of the American Council of Learned Societies and the direction of Hans Kurath. So far only the first section, Linguistic atlas of New England (3 vols. of maps in 6) and Handbook of the linguistic geography of New England, has been published (1939–43). Field work for the linguistic atlas of the Middle and South Atlantic States was completed in 1949; the records are on file at the University of Michigan, and were consulted for this article with Kurath's permission. The records from southwestern Ontario, essentially contiguous to the Canadian communities investigated in connection with the work in New York State, will be included in the linguistic atlas of the North Central States; field work for this project is nearing completion under the direction of A. H. Marckwardt, with whose permission this material was used by the present writers.

The authors appreciate the assistance of J. W. Downer in checking data in the files of the Linguistic atlas.

Citations from previous publications have been changed only in styling, so as to conform to the practices of Language: (1) linguistic forms are italicized; (2) phonetic notations are enclosed in square brackets; (3) transcriptions within slant lines are phonemic.

² In general the intrusive [h] before a vowel seems to be most common before stressed syllables, especially before extra-loud stress. In such dialects [h] is not phonemic; nonsignificant preaspiration of initial vowels is a common function of loud stress. See Richard A. Proctor, The misused h of England, Atlantic monthly 55.593–601 (1885).

³ Joseph Wright, English dialect grammar §240 (Oxford, 1905). An important exception is the pronoun it (which usually occurs with weak stress), where the historical form [hit] is associated with lack of social prestige; compare stressed [him] but weak-stressed [im]. Recent investigations have found exceptions to the traditional picture of preaspiration as a non-distinctive feature of initial vowels in southern British dialects. See Helge Kökeritz, The phonology of the Suffolk dialect: Descriptive and historical §221 (Uppsala, 1932); Bertil Widén, Studies on the Dorset dialect §83 (Lund studies in English, No. 11; 1949). This modification of Wright's statement is less probably due to recent influence of the standard language than to the fact that recent investigations have been more systematic than those on which Wright based his observations. Thus the 1938 survey of southern England by Guy S. Lowman Jr. disclosed that dialects west of London largely retain postvocalic /-r/ in first, barn, beard, etc., although it has been traditional to state that this sound was lost everywhere in southern England. See also Widén §7.

⁴ Otto Jespersen, Modern English grammar 1.2.942 (Copenhagen and London, 1949).

⁵ Dickens' representation of Uriah Heep as saying 'umble indicates that by 1850, the date of David Copperfield, the pronunciation of initial /h-/ in such words had become for some people a badge of social respectability. Conversely, the practice of representing the pronunciation without /h-/ as 'dialectal' and typical of an unsavory character such as Uriah Heep may have increased a tendency to look upon the pronunciation of these words with /h-/ as socially preferred. See Sumner Ives, A theory of literary dialect, Tulane studies in English 2.137–82 (1950). The mixture of attitudes toward these pronunciations is shown in a series of letters to Notes and queries, prompted by the appearance of David Copperfield and Dickens' representation of 'umble as a substandard form. In a debate among amateurs over the propriety of pronouncing initial /h-/ in English words of Romance derivation, pronunciations without /h-/ were variously described as a mark of culture, a vulgarism, and an affectation assumed by younger members of the clergy. Humor was frequently said to be one of the words where the pronunciation without /h-/ was acceptable in cultured speech. See Notes and queries, 1st series 8.54, 229, 298, 393–5, 551 (1853).

⁶ Except for a few dialects, chiefly in the southeastern United States, where here and hear have /hj-/, this cluster occurs only before /u/, mostly in words of Romance derivation.

⁷ Thomas Pyles, Linguistics and pedagogy: The need for conciliation, College English 10.393 (1948–9): ‘To this day “the baby whales” and “the baby wails” sound exactly alike in my pronunciation, as they do in the pronunciation of many speakers in all parts of the country, cultured and uncultured alike.‘

⁸ Mabel Y. Spears (of Utica, N. Y.), in an article called W'ere, w'en, w'y, w'ich, w'at, Col. Eng. 11.38–9 (1949–50). Yet the cultured informant interviewed for the linguistic atlas in the Utica area has no /h-/ in whip and humor.

⁹ T. R. Lounsbury, The standard of pronunciation in English 23, 194, 197 (New York and London, 1904), insists that the pronunciation of /hw-/ rather than /w-/ in words like whip is overwhelmingly predominant ‘in polite society’ in the United States, but concedes that there is ‘wavering’ in the usage for humor.

W. D. Whitney, The elements of English pronunciation, Oriental and linguistic studies (2d series) 268–70 (New York, 1874), is even more prescriptive in his attitude. ‘That those who say hwen and hyu have preserved an earlier and fuller sound, which has suffered corruption and abbreviation in the mouths of the other party, admits no serious question. ... With a great part of the vulgar speakers of English, the tendency is toward eliminating the surd instead of the sonant of the combination which ought to contain both, converting when into wen.‘

A. J. Ellis, Early English pronunciation 4.1125 (London, 1875), suggests that Whitney's attitude was perhaps influenced by the way he had heard people pronounce his own name: 'Mr. Whitney is an incontrovertible authority as to the way in which he pronounces, and wishes others to pronounce, the initial sounds of his own name, but ... he must have met with many who disputed it. Possibly he is often called ... [i.e. a pronunciation with /w-/ for /hw-/], as he certainly would be generally in London.' Whitney's conclusions were also disputed by E. S. Sheldon, secretary of the American Dialect Society, and a native of Maine, in A New Englander's English, Dialect notes 1.42 (1890–6): 'Humor I think I used to pronounce [yû] without the [h]. The wh question I cannot here discuss fully, but I should regularly write [hw] in the words concerned, and not [w]. But some Americans do say [w], especially in the word wharf, common in New England at least, as [wɔəf] or [wɔf]. Compare the dialect of the Biglow Papers. I remember a college classmate of mine, some twenty years ago, who came from Cincinnati, and who regularly said [w] for the written wh as in where, when, etc. Perhaps the [h] now common was introduced or at least assisted by the schoolmasters.'

¹⁰ C. K. Thomas, Introduction to the phonetics of American English 102 (New York, 1947): ‘In all English dialects the substitution of [j] for [hj] ... is considered substandard, though common in most large cities. Similarly, the substitution of [w] for [hw] ... is usually considered substandard in America, though this also is common in most large cities. In the South of England, on the other hand, the substitution of [w] for [hw] is normal.‘

A. G. Kennedy, Current English 209 (Boston, 1935): ‘Some people regularly simplify this combination to [w] alone, saying wen, were, wy, and so on, and everyone does it to some extent in hasty and careless speech. Some speakers, it is true, intensify the diphthong [iu] after initial [h] so as to produce a combination [hj] in such words as Hewlett, huge, human, humor, and then, just as in the case of [w] above, the [h] is lost, and such pronunciations as [jumr] and [jumən] result.‘

¹¹ H. L. Mencken, The American language, Supplement 2.95 (New York, 1948). In The American language⁴ 350 (1937), Mencken asserted that the attempt to enforce the pronunciation of /h-/ in such words was an affected imitation of the British: ‘The majority of Americans seem to have early abandoned all effort to sound the h in such words as when and where. It is still supposed to be sounded in England, and its absence is often denounced as an American barbarism, but as a matter of fact few Englishmen actually sound it, even in their most formal discourse. ... The Americans do not sound the h in heir, honest, honor, hour, and humor and their derivatives, and frequently omit it in herb, humble, and humility.‘ It is difficult to understand how Mencken could have got the picture of British and American usage reversed.

¹² S.v. humor: 'The h, formerly silent, is now generally pronounced, both in England and the United States, although many good speakers, following the older orthoëpists, as Smart, omit it, esp. in the senses referring to mental states.' In the synopsis of opinion and disputed points of pronunciation—a feature carried over intact from earlier Merriam-Webster dictionaries—all dictionaries except that of the Rev. James Stormonth are indicated as recognizing both /hj-/ and /j-/ pronunciations of humor; see Introduction §277. In another passage (Introduction §258) we read: 'wh as when, what, which, is, in America, usually pronounced as h+ a voiceless w, no voice being heard until the beginning of the following vowel. In England, and by many American speakers, it is usually pronounced simply as a voiceless w, although still often, especially in the South of England, as an ordinary w.'

¹³ Three pronunciations are given: with [hju-, him-, ju-]. The [ju-] type is indicated (Introduction §210) as ‘most likely to occur in “sense of humor,” “mood,” and the verb.‘ Words of the whip and whoa types are cited only with [hw-], but with the notation that [w-] pronunciations are common.

¹⁴ S.v. humour: ‘The pronunciation of the initial h is only of recent date, and many still omit it, esp. in the senses under II’, which deal with feelings. — S.v. wh-: ‘/w-/ is now universal in English dialect speech, except in the four northernmost counties and north Yorkshire, and is the prevailing pronunciation among educated speakers. /hw-/ is general in Scotland, Ireland, and America, and is used by a large proportion of educated speakers in England, either from social or educational tradition, or from a preference for what is considered a careful or correct pronunciation.‘

Cf. Jespersen, Mod. Eng. grammar 1.13.51: ‘In order to indicate the retention of the old sound as an Irishism B[ernard] Shaw has recourse to the spellings hwat and hwy. (John B[ull's Other Island] 9,77).‘

¹⁵ Hermann Michaelis and Daniel Jones, A phonetic dictionary of the English language [with entries in phonetic transcription] (Hanover and Berlin, 1913); Daniel Jones, An English pronouncing dictionary (London and Toronto, 1917).

¹⁶ The /ju-/ form of humor is labeled ‘dialectal’ by Michaelis-Jones, ‘old-fashioned’ by Jones. Michaelis-Jones has no entries under [hw-]; Jones gives [hw-] as a secondary (i.e. less frequent) pronunciation for words of the whip group, except for whoa, which is given only with [w-].

¹⁷ H. W. Fowler, A dictionary of modern English usage 240 (Oxford, 1926): 'Humour is still often or usually pronounced without the h sound; the derivatives now being rarely without it, humour itself will probably follow suit.' Ibid. 466–7: 'The broad principles [of socially acceptable pronunciation] are: pronounce as your neighbors do, not better; for words in general use, your neighbor is the general public.'

¹⁸ George Hempl, MLN 6.310–1 (1891); cf. Jespersen, Mod. Eng. grammar 1.13.51. Hempl suggested that the large proportion of Scotch and Irish settlers may have helped American pronunciation to retain the /hw-/. He noted that /w-/ regularly occurred ‘only in Maryland and in certain circles in New York City.‘

¹⁹ Grandgent's observations were based on answers to a mailed questionnaire. ‘My correspondents are nearly unanimous in favor of hw in all the examples except wharf, whoa! and the interjection why.‘ For full particulars, see C. H. Grandgent, American pronunciation again, MLN 8.277–8 (1893).

²⁰ G. P. Krapp, The English language in America 2.245–6 (New York, 1925): ‘The pronunciation ... varies ... both pronunciations occurring in all regions and at all levels of speech, though the former [i.e. the one with /hw-/] is the more common. Especially when relatively lightly stressed, initial wh is likely to be pronounced as w.‘ Krapp observes that the pronunciation with /w-/ has been opposed by early American grammarians and other authorities. It was, however, regarded as standard by the Royal standard English dictionary (1778) of the Londoner Perry and was taken over in the early American editions of that work. In his Travels (1.468), Timothy Dwight reported, without disapproval, that Bostonians often omit /h-/ in words written with wh.

In his Dissertations on the English language (Boston, 1789), Noah Webster spoke of /w-/ in whip and the like as a fault. He asserted that the ‘pure English stock’ in the United States keeps /h-/ in these words, and that its loss is a foreign corruption. See Ellis, EEP 4.1068 (1875).

²¹ J. S. Kenyon, American Pronunciation⁶ §227 (Ann Arbor, 1935): ‘The distinction between [hw] and [w] by which, e.g., whether is distinguished from weather is still standard usage in America, though there are a great many speakers who do not make the distinction. Reliable statistics are lacking as to whether the substitution of [w] for [hw] is increasing here. It has probably been frequent for many generations.‘ Kenyon cites (§348.2) several pronunciations of humor, suggesting a semantic difference between the forms with and without /h-/, but does not indicate their relative distribution.

²² Hans Kurath, American pronunciation, SPE Tract 30.284 (1928): 'In white, wheat, whale, etc., the [w] is always preceded by a voiceless fricative; only in the exclamatory why! and in weakly stressed position, as in somewhat, whatever, whenever, and unstressed when, what, where, this voiceless sound is slighted.' Id., The origin of the dialectal differences in spoken American English, Mod. phil. 25.386–7 (1927–8): 'Among the retarded sounds one might mention ... the first consonant in wheat [ʍi: t] which with us is still generally pronounced as a voiceless labial fricative followed by a voiced glide (except in parts of the South; elsewhere rather commonly in unstressed whatever, whenever, wherever, and the exclamatory why!), while it became fully voiced in Southern English.'

Stuart Robertson, Development of modern English³ 227 (New York, 1941), observes that, contrary to Kurath's statement in the SPE tract, initial /w-/ in words like whip does occur in New York City and Philadelphia.

²³ A. H. Marckwardt, Introduction to the English language 46 (New York, 1942): ‘Like [h] the sound [ʍ] tends not to appear in unstressed situations, where [w] is frequently substituted for it. ... There is also some difference between British and American practice in respect to [ʍ] in such words as whine, whether, whale, Southern British tending to employ [w] to a greater extent.‘

²⁴ Sylvester Primer, Charleston provincialisms, PMLA 3.91 (1888): 'In the combination wh the h is always silent.' Id., The pronunciation of Fredericksburg, Virginia, PMLA 5.199 (1890): '[In Fredericksburg] the h never disappears in the combination wh as in Charleston, S. C.'

²⁵ O. F. Emerson, The Ithaca dialect, Dialect notes 1.168: ‘In [the Ithaca dialect] the sound [i.e. /hw-/] is regularly preserved, though /w-/ for /hw-/ is occasionally heard as an individual peculiarity. In unstressed syllables /hw-/ is sometimes reduced to /w-/, sometimes lost.‘

²⁶ William A. Read, Variant pronunciations in the New South, Dial. notes 3.530 (1905–12). Using a written check list, he found, for informants in the Gulf States, 69 examples of /hj-/ in humor, 172 of /j-/. In his article Southern American dialect, Am. speech 8:2.39 (April 1933), Read states that in the South /hw-/ is usual in white, etc., but that /w-/ sometimes occurs in New Orleans.

²⁷ E. F. Shewmake, English pronunciation in Virginia 31 (University of Virginia diss. 1920; Davidson, N. C., 1927): ‘The omission of the [h] sound in words like white and where is not characteristic of the Virginia dialect except in one word why, and not in this one except when it is used with exclamatory or expletive force.‘

²⁸ W. Cabell Greet, A phonographic expedition to Williamsburg, Virginia, Am. speech 6.170 (1930–1): '[ʍ] stressed is stable on the coast as well as elsewhere in the South.' Id., A record from Lubeck, Maine, and remarks on the coastal type, Am. speech 6.401 (1930–1): 'There is some uncertainty in the employment of [ʍ]. I believe that normally [w] for [ʍ] is more common than the retention of [ʍ]. This is not true, I think, of polite speech in Boston and Cambridge.' Id., Delmarva speech, Am. speech 8:4.62 (Dec. 1933): '[ʍ] is commonly replaced by [w] except in the word wharf, which is very common ... [w] for [ʍ] in stressed syllables occurs rather often in the speech of coastal Virginia and South Carolina, and I am told that it is current in Philadelphia, as it is in New York City, particularly in Brooklyn and elsewhere on Long Island. Generally in the United States [ʍ] is stable in stressed syllables, though usually modified to [w] in unstressed syllables.' Greet does not cite the evidence that led him to revise his judgment of two years before.

²⁹ Harry M. Ayres, Bermuda English, Am. speech 8:1.9 (Feb. 1933), reports that /hw-/ is normal in Bermuda in which and while, though whaling has initial [v].

³⁰ Cleanth Brooks, The relationship of the Alabama-Georgia dialect to the provincial dialects of Great Britain 41 (Baton Rouge, 1935), is in doubt about the interpretation of spellings like w'ich, w'en, w'y, w'at in the Uncle Remus stories of Joel Chandler Harris—since he had himself done no field work in the dialect area which he attempts to compare with British speech.

³¹ Oma Stanley, The speech of East Texas 56 (Am. speech reprints, No. 2, 1937): ‘The initial [h] has not been restored in herb, humble, humor ... in East Texas speech; that is, they have not acquired the spelling pronunciations which they have in some sections.‘ Stanley reports (ibid. 71) that in East Texas /hw-/ is normal in words like whip and wheelbarrow.

³² C. K. Thomas, Pronunciation in Upstate New York VII, Am. speech 12.122–3 (1937); Pronunciation in Downstate New York II, Am. speech 17.154–6 (1942). Thomas obtained most of his data from readings of a prepared text, which would probably yield more occurrences of /hw-, hj-/ than would be found in normal speech. Most of his informants are college students.

For Upstate New York he reports 27 occurrences of humor with /hj-/, 5 with /j-/. The proportions of /hw-/ to /w-/ vary from 4 to 3 for wheeze, to 28 to 1 for when. For the three words on which he has the most information, the figures are 321 to 84 (wheelbarrow), 293 to 58 (while), 224 to 58 (what). For subdivisions within the Upstate area, the percentages of occurrences of wheelbarrow with /w-/ are: ‘Southern’ (lower Hudson) 57; ‘Eastern’ (Mohawk and Upper Hudson) 35; ‘Northern’ (Adirondacks and St. Lawrence) 3; ‘Central’ (Finger Lakes) 16; ‘Western’ (Genesee to Lake Erie) 15. He concludes that there is a trend toward /w-/ in the southern and eastern sections, the forms lacking /h-/ being most common in the neighborhood of large cities.

For the Downstate New York area—New York City, Long Island, and suburban Rockland and Westchester Counties—Thomas finds a relatively higher proportion of forms without /h-/. For humor he has 10 examples with /hj-/, 7 with /j-/. Only one word (where, with a non-significant proportion of /hw-/ in 6 records to /w-/ in 5) has /hw-/ more often than /w-/. The ratio of /w-/ to /hw-/ is generally about 3 to 2: wheelbarrow has 52 /w-/ to 46 /hw-/; white 159 to 113; while 112 to 104; whale 89 to 25; whistle 63 to 15.

³³ A. Tressider, Sounds of Virginia speech, Am. speech 18.271 (1943). Tressider's informants were girls, students at Madison College, a state teacher's college in Harrisonburg, Va. He reports /j-/ more common than /hj-/ in humor, though occurring only twice in huge. Most informants have /hw-/ in where, etc., but forms with /w-/ appear occasionally, especially among students from the counties on the eastern shore of Chesapeake Bay; out of six informants from this area, three have /w-/.

³⁴ Mencken, The American language, Supplement 2.203, 209. The pronunciation of words like whip with /w-/ is twice noted as characteristic of the Philadelphia area.

³⁵ Hubbell's dissertation, The pronunciation of English in New York City (New York, 1950), is the most detailed study of the speech of a single American community now in print. (Using the Atlas records, Yakira H. Frank in 1948 completed a dissertation on the same subject at the University of Michigan; but her study is at present available only on microfilm.) Hubbell bases his conclusions on phonographic recordings of both prepared texts and free conversation. He distinguishes both age levels and cultural levels among his informants, and recognizes the limitations of read material.

Of /h-/ Hubbell writes (42): 'So far as its phonetic distribution is concerned, metropolitan speech differs from that of much of America in that the initial cluster /hw-/ (in words like when and white) does not occur in the pronunciation of most New Yorkers ... Furthermore, the initial cluster /hj-/ (in words like huge and humorous) is very frequently lacking on the uncultivated and intermediate levels ...' Again (54): 'On the uncultivated and intermediate levels of New York speech the cluster /hj-/ does not occur, all words like huge being pronounced with /j-/ alone. (Such pronunciations are very common in the speech of Columbia undergraduates who come from the city.) The distinction recorded in our dictionaries between /j-/ as a variant in humor and its derivatives, and /hj-/ alone in the other words of this group appears in the speech of [five informants]. But more commonly New Yorkers are consistent in pronouncing all these words with /hj-/ or /j-/.' And again (52): 'In words of the type of whale, wheel, when, the cluster /hw-/ may occur, but the most common pronunciations on all levels of metropolitan speech are [those without /h-/]. Speakers who employ /hw-/ consistently in all the words of this historical class are rather rare. In uncultivated speech /w-/ is universal; on the intermediate and cultivated levels, /hw-/ may not infrequently be heard, but its use more often than not is sporadic—the speaker will employ it only in a restricted number of words or will pronounce the same words now in one way and now in the other ... There can be little doubt that the /hw-/ pronunciations are for the most part consciously adopted ones, adopted because of the widespread notion that pronouncing whale and wail, whet and wet as homonyms is “incorrect.” New Yorkers of Irish birth usually employ /hw-/ or some variant of it in these words, but the pronunciation rarely survives in the speech of the second generation.'

³⁶ Handbook of the linguistic geography of New England 9 (Providence, 1939): ‘The loss of [h] in wheelbarrow, whetstone, etc., is common on the coast from Marblehead to Machias, and occurs in scattered points elsewhere ... In wharf, a seashore term, the loss of [h] is much more widespread. It seems that the conservative coast towns from Long Island Sound to New Brunswick have preserved in this word a type of pronunciation that was widely current in New England in Colonial times and that this coastal pronunciation has been retained in the upland through contact with the seashore.‘ Cf. ibid. 23 and Chart 8.

³⁷ Fieldworkers were encouraged to record conversational responses, i.e. forms which occurred spontaneously in the informants' conversation while discussing a particular topic suggested by the fieldworker. These are especially significant when they differ from answers to direct questioning.

³⁸ When field work is completed for Marckwardt's Linguistic atlas of the North-Central States and for H. B. Allen's Linguistic atlas of the Upper Midwest, comparable data will be available for the northern half of the country as far west as the Rocky Mountains.

³⁹ The distribution of /h-/ before vowels does not appear to be significant for American English. For words of Old English or Scandinavian derivation, the absence of /h-/ seems to be confined to sporadic foreign-language settlements. For words of Romance derivation, where there seems to be some variation in British usage, hotel (which is stressed on the first syllable by perhaps a majority of American speakers) always has /h-/, and hostler is only a book word; humble and the proper name Humphrey or Humphries seem to have geographical variants, the forms without /h-/ being apparently more common in the South Atlantic States than elsewhere, but are very difficult to elicit in a conversational interview.

⁴⁰ Most of the other words in which some speakers have /hj-/ and others have /j-/, such as huge and human, are difficult to elicit in a conversational situation. Moreover, of this group humor seems to be the word occurring most frequently with /j]-/; see Thomas' studies reported in fn. 32.

⁴¹ While was systematically recorded only in New England: stressed what was not investigated in New England. Wheel (the baby) was also not investigated in New England; in the Midland and the South, the verb for this action is generally roll or ride. What time is it? was recorded throughout the Atlantic seaboard; but in this phrase, what has reduced or weak stress, and is thus not comparable with the other words. Cf. fnn. 22 and 23.

⁴² This procedure is not strictly phonemic, since conceivably some speakers may have weak murmur-onset before semivowels or vowels, and strong aspiration for /h-/- However, in the absence of a phonemic analysis for each informant the assumption that murmur is an allophone of /h-/ is at least a working hypothesis by which the data can be interpreted.

⁴³ No attempt has been made to identify the communities in which the variants whetrock, whetseed, whetter, and whet occur, since all these have the same initial.

⁴⁴ Explanations of the facts are risky, of course. Nevertheless, whip does often occur with secondary or tertiary stress in conversation and in compounds—situations where the /h-/ might be weakened; and the common practice of moistening a whetstone with water or spittle may occasionally tend to an association with wet.

⁴⁵ In two communities in southern Ontario, the fieldworkers recorded wheel in conversation with /w-/, though wheelbarrow was offered as a response to direct questioning with /hw-/.

⁴⁶ Kurath, A word geography of the eastern United States, Fig. 3 (Ann Arbor, 1949).

In his radio program Where are you from?, Henry Lee Smith Jr. set up an undelimited dialect type which he called Central Atlantic Seaboard; it probably coincides with the area where /h-/ is lost in these words. As a derivative of Smith's area we may include the St. John's Valley in southern New Brunswick, which was settled at the end of the 18th century by Loyalists from the New York City area. See Handbook of the linguistic geography of New England 17, 23, 33, 238–40.

⁴⁷ Kurath, Word geography 62–3.

⁴⁸ Possibly the loss of /h-/ in wharf began earlier in England, since two of the four earliest examples cited in the Oxford dictionary (including the earliest) lack the /h-/. On the basis of the New England evidence alone, Kurath labels wharf a ‘seashore term’ (Handbook 9); see fn. 36. The contrast between the distributions of /w-/ in wharf and in whip is, however, much more spectacular in the South than in New England.

⁴⁹ See Kurath, Word geography 16, 66, Fig. 108.

⁵⁰ Instances in which the fieldworkers recorded a falling diphthong with an unrounded first element are included with /hju-/.

⁵¹ In two Vermont communities, the fieldworker recorded music, beautiful, and bureau with /mu-, bu-/, with a diphthongal but fully rounded syllabic.

⁵² Lorenzo D. Turner, Africanisms in the Gullah dialect 27, 243 (Chicago, 1949).

⁵³ Under direct questioning, one informant used /hw-/ in wheelbarrow and another did so in whetstone. Each of these, however, also offered conversational instances of the same word with /w-/.

⁵⁴ Wright, Eng. dial. gram. §240, reports that /hw-/ is preserved in the Shetlands and Orkneys, most of Scotland, Ireland, Durham, Northumberland, Cumberland, Westmoreland, northern Yorkshire, and the Isle of Man; elsewhere in England it has become /w-/. See also Wright, An elementary historical New English grammar §284 (Oxford, 1924). Neither Wright's grammar nor the English dialect dictionary (London, 1898–1905) includes whoa, whose pattern of distribution in the United States is markedly different from that of whip, wharf or whinny.

E. Kruisinga, A grammar of the dialect of West Somerset §246, §315 (Bonn, 1905), finds humor with /j-/ in Somerset, whip and the like with /w-/. Occasionally what and when appear with /h-/ and no /w/.

Kökeritz, The phonology of the Suffolk dialect §223, reports that in Suffolk the phoneme /w/ has a voiceless allophone in the neighborhood of voiceless sounds; but there is no evidence of a phonemic contrast between the two kinds of [w].

⁵⁵ The peculiarities of Charleston pronunciation have been attributed to the Huguenot settlement in the early 18th century. See Primer, The Huguenot element in Charleston's pronunciation, PMLA 4.214–44 (1889). Untrained observers have frequently asserted Negro influence to be responsible for those features of Southern white speech which are not found in their own dialects. For an evaluation see McDavid, The relationship of the speech of American Negroes to the speech of Whites, Am. speech 26.3–17 (1951); review of Turner, Africanisms in the Gullah dialect, Lg. 26.323–33 (1950). The subject of borrowing has been treated by Einar Haugen, The analysis of linguistic borrowing, Lg. 26.210–31 (1950); Problems of bilingualism, Lingua 2.271–90 (1950).

⁵⁶ See R. L. Meriwether, The expansion of South Carolina 1729–1765 (Kinsport, Tenn., 1940).

⁵⁷ Compare Jespersen's statements: 'Humour and hotel are now pronounced with [h] by some educated speakers, without [h] by others.' (Mod. Eng. grammar 1.2.943.) /w-/ for historical /hw-/ 'is not ... nowadays regarded as nearly so “bad” or “vulgar” as the omission of [h], and is, indeed, scarcely noticed by most people. In fact, a great many “good speakers” always pronounce [w] and look upon [hw] as harsh or dialectal. In some schools, however, especially girls' schools, [hw] is latterly insisted on.' (Ibid. 1.13.51.) Cf. Robertson, Development of modern English 242.

⁵⁸ R. E. Zachrisson, The English pronunciation at Shakespeare's time as taught by William Bullokar, Skrifter utgivna av K. Humanistiska Vetenskaps Samfundet i Uppsala 22.6.107. Zachrisson's word-lists do not suggest any loss of /h-/ in whip and the like.

⁵⁹ Jespersen, Mod. Eng. grammar 1.2.943.

⁶⁰ Ellis, EEP 4.1145 (1875).

⁶¹ Karl Luick, Historische Grammatik der englischen Spräche 1.704, 753 (Leipzig 1921).

⁶² Kökeritz, The phonology of the Suffolk dialect §370; especially significant are such hyperforms as where for were. See also Harold Whitehall, Some fifteenth-century spellings from the Nottingham Records, Essays and studies in English and comparative literature 13.70 (Ann Arbor, 1935); H. C. Wyld, A history of modern colloquial English 311–2 (New York, 1920).

⁶³ Wyld, op.cit. 138–9, 180; A short history of English §286 (London, 1927).

⁶⁴ Ellis, EEP 4.1050–5 (1875).

⁶⁵ Krapp, The English language in America 2.245–6.

⁶⁶ Jespersen, Mod. Eng. grammar 1.13.51.

⁶⁷ Ellis, EEP 1.188 (1869), 2.573,605 (1869), 4.1145 (1875).

⁶⁸ Henry Sweet, The history of English sounds, Trans. Phil. Soc. 525–6 (1873–4): The loss of /h-/ in the Old English clusters /hr-, hl-, hn-, hw-/ ‘is at the present moment being carried out with the only remaining member of the group‘—i.e. /hw-/.

⁶⁹ History of English sounds 268 = §917 (Oxford, 1888): ‘Toward the close of [the 18th century] ... [hw] began to be levelled under [w], and in the present cent. the change was carried out universally, even among those who still retained [h] as a mark of gentility. But of late years it has begun to be restored in Southern [British] educated speech, partly by the influence of the spelling, partly by that of Scotch and Irish pronunciation, so that in another generation it will probably be completely restored. It is now pronounced in unstrest words, where it was probably weakened into [w] in the period when it was a natural sound.‘

⁷⁰ W. W. Skeat, Influence of Anglo-French pronunciation upon modern English, Trans. Phil. Soc. 452 (1899–1902): ‘It is in the South that [hw] has become a mere [w], whilst in the Northumbrian district it is still fairly maintained.‘

⁷¹ Thus, in the South Carolina piedmont, urban informants—like the mercantile and plantation caste of the coastal plain, especially Charleston—generally lack the constriction of postvocalic /-r/ in barn and beard, though rural speakers often have it. See McDavid, Post-vocalic /-r/ in South Carolina: A social analysis, Am. speech 23.194–203 (1948).

^71a See Kurath, Word geography 16, 41, 66, and Fig. 108.

⁷² A. L. Davis and R. I. McDavid Jr., Northwestern Ohio: A transition area, Lg. 26.264–73 (1950).

⁷³ Anders Orbeck, Early New England pronunciation 9 (Ann Arbor, 1927).

⁷⁴ Kurath, Word geography 12, 13, 48, 56, Fig. 67.

⁷⁵ When we try to plot on a map of the British Isles what we learn from the English dialect dictionary about items that have a clearly defined regional distribution in the United States, we discover how exceedingly fragmentary our knowledge of British dialects actually is. Thus, of the many names current in this country for the earthworm, only eaceworm and angledog are recorded in the EDD. It is hoped that these gaps in our knowledge—of British phonology no less than of British vocabulary—will be filled by the dialect surveys now begun in England under the direction of Harold Orton of Leeds, and in Scotland under that of Angus McIntosh of Edinburgh.