Summary e form of the indefinite article before h is not an easy matter, as mentioned among others, by Jespersen, in A Modern English Grammar. e existence today of written corpora and other online resources allows for a wider and, hopefully, a more reliable examination of variants. e present paper presents the results obtained from an analysis of the forms of the indefinite article before words beginning with h found in the Gutenberg corpus, therefore, in literature, and in e British National Corpus (BNC). Quantitative data are presented and accounted for in a synchronic and diachronic perspective. Key words: variants, loss of [h], synchrony, diachrony Povzetek Katero obliko nedoločnega člena uporabiti pred h ni preprosta odločitev, kar poleg drugih avtorjev omenja že Jespersen v A Modern English Grammar. Danes številni korpusi pisnega jezika in ostali elektronski viri dopuščajo zelo široko in verjetno tudi bolj zanesljivo analizo variant. V tem članku so predstavljeni rezultati analize rabe oblik nedoločnega člena pred besedami, ki se začenjajo s h in se nahajajo v Gutenbergovem korpusu, torej v književnosti, in v britanskem narodnem korpusu (e British National Corpus). Kvantitativni podatki so predstavljeni in utemeljeni tako s sinhronega kot diahronega vidika. Ključne besede: variante, izguba [h], sinhronija, diahronija DOI: 10.4312/elope.2.1-2.11-24 For an English teacher in France, the problem of the form of the indefinite article before words beginning with h is not an unimportant one. Otto Jespersen in A Modern English Grammar( part VII, p 407) qualifies it as “not quite simple”. e indefinite article, he adds, has two forms: “an” before a vowel and “a” before a consonant. But from Chaucer to Kipling, the form “an” has been used not only before vowels but also before h. And then Jespersen gives an impressive list of authors (unfortunately without the exact indication of the particular passage and the work) who have used “an” before about 33 English words beginning with h in stressed syllables and 19 before h in unstressed syllables. As he confesses “for years I took the harmless trouble of noting down all the examples I came across of an before h but probably left out many instances I found in which a was used before h.” T o my knowledge the first person interested in the form of the indefinite article before h and before initial [ju] spelt u or eu was Louis Feipel. In 1929 he undertook the counting of the occurrences of an and a before h in about 300 recently published books or translations (most of them published between 1921 and 1927) by authors of repute, equally divided between American and British ones, and presented the results in an article entitled “A” and “AN” before “H” and Certain Vowels. e paper was written in response to a statement made by Professor J.T . Hillhouse which led one to believe that there was a high degree of uniformity in the use of a before words beginning with an aspirated h, before words beginning with the long sound of , and before words beginning with the sound of and . Professor Hillhouse quotes Fowler’s Dictionary of Modern English Usage (1926) which purports to represent the prevailing usage among present-day writers, to this effect: “An was formally used before the unaccented syllable beginning with ‘h’, but now that the ‘h’ in such words is pronounced, the distinction has become pedantic, and a historical should be said and written; similarly, an humble is now meaningless and undesirable. A is now usual also before vowels preceded in fact, though not in appearance, by the sound of ‘y’ or ‘w’ (a unit, a eulogy, a one)”. First, Feipel analyses words beginning with h (the form of the indefinite article in front of words beginning with initial initial [ju] spelt u or eu will be dealt with in another paper) according to whether the syllable containing it is accented or not. us, he divides the words into two classes. He mentions that there is unanimity in the use of the form an before the four words containing a silent h: (borrowed in ME from French with no h in spelling) an heir, an hour, an honour and an honest. eir pronunciation without [h] was long prevalent. And he concludes that the occurrence of an a form is practically unthinkable. In the first class containing words with a stressed first syllable he investigates the words humble and hundred/hundredfold : Word Article British English American English humble a 2 0 an 0 4 You can see the results of the investigation of the word humble in the table above. He considered that the regular form a before h was only found in British English while in American English the “abnormal an form” was found. But as in the rest of his article he never asks himself questions as to why this is so. We are actually told by Wells (1986) that a h-less form of the word humble is still found in the American south so that this is exactly what one would expect to find here. Word Article British English American English hundred/hundredfold a 0 0 an 5 2 In the case of hundred/hundredfold the prevailing form an may have the explanation given by Jespersen who admits to it being “especially tenacious”. In the second class the syllable containing h is not accented. And the first word he analyses is hotel, which if we are to be accurate has the first syllable stressed with secondary stress. e investigation of this word revealed another “anomalous” situation, to use Feipel’s words. It is mainly the British authors who use “an hotel” – only one example was found in an American author. He says: “In the 300 volumes examined ‘an hotel’ preponderated markedly over ‘a hotel’.” But no figures are given. is is, in a way, what one would expect, given that according to D. Jones’s Dictionary “some British speakers use the form without h always, others when it is preceded by an”. e form in American English reflects probably the preference of the Americans for spelling pronunciation. In the same class he analyses historic(al) and historian. Word Article British English American English Remarks historic(al) a 7 4 2 authors use both an 4 4 historian a 0 2 an 1 0 e use of the two forms of the indefinite article has an explanation in the pronunciation of these words. In British English the word historic(al) is sometimes pronounced without h , though only after an according to D. Jones and Wells. We are told that some RP speakers treat h in unstressed syllables as in historical as if it belonged to the group hour, heir, honour, eg an historical novel . But Gimson mentions that such pronunciations, as well as humour as /ju:m«/, are used by a minority only. In its turn the word historian is sometimes without h when after the indefinite article an according to Wells. In the following 16 examples the an form seems to be “preponderant”. Word Article British English American English hallucination An 6 2 A 3 0 hysterical An 5 1 A 1 3 horizon 1 An 4 1 A 2 0 hypothesis An 1 2 A 1 0 habitué An 1 1 A 1 0 hereditary An 2 0 A 1 0 hermaphrodite An 2 0 A 0 0 hermetical(ly) An 2 0 A 0 0 herbaceous 2 An 1 0 A 0 0 hieratic An 1 0 A 0 0 hydraulic An 1 0 A 0 0 Hellenic An 1 0 A 0 0 hypertrophy An 1 0 A 0 0 Hidalgo An 0 1 A 0 0 Hungarian An 0 1 A 0 0 histrionic An 0 1 A 0 0 We find it difficult to follow his analysis for some of the words as to my knowledge “preponderant” means “larger in number or more important than other people or things in a group” and these words occur in only one form (from hieratic to the end). In the table below the a form is preponderant: Word Article British English American English harmonious a 5 0 an 1 0 hypnotic a 3 1 an 3 0 harangue a 1 0 an 0 0 hyena a 1 0 an 0 0 Herculean a 0 1 an 0 0 Havana a 2 1 an 0 0 Here again some of the words are found only preceded by a such as: a harangue, a hyena, a Herculean, a Havana. ere are also cases of even break: Word Article British English American English habitual a 2 0 an 2 0 heraldic a 1 0 an 0 1 hiatus a 1 0 an 0 1 hilarious a 0 1 an 1 0 hypocrisy a 1 0 an 1 0 In my turn I investigated the usage of the indefinite article before words beginning with h, both synchronically and diachronically. Nowadays, such an investigation is facilitated by the existence of a considerable number of computer corpora. I began by investigating the electronic texts available on the internet (part of the Project Gutenberg). I, like Feipen, focused my investigation on texts of literature, but unlike him I did not analyse translations for the obvious reason that the author’s choice of one or the other form may be intentional, reflecting a different dialect be it regional or social. I found a number 3 of works of literature in the corpus where the form an of the indefinite article in front of words beginning with h was not used except with the 4 words containing the silent h mentioned above. e number of authors whose work I have investigated and who include variation in the use of the indefinite article is observed is 15: Author Work an+h a+h Remarks D. Defoe 1660 - 1731 Robinson Crusoe 1719 an humble British English Moll Flanders 1722 a hundred From London to Land’s End 1724 an hundred (3) 4 a hundred both forms occur S. Johnson 1709 - 1784 Journey to the Western Isles of Scotland 1773 an hereditary, an humble a habitation, a historian, a hundred British English J. Austin 1775 -1817 Love and Friendship 1790 an happiness, an historian, an horrible, an horrid an humble British English Sense and sensibility 1811 an habitual, an heavy, an heightened an hysterical a happy(10), a happier, a horrid, a hurried, a husband, a hurry, a happiness Pride and Prejudice 1813 an hopeless an hurried , an husband Emma 1815 an hesitation an hundred (2) Ch. Darwin 1809 - 1882 On the Origin of Species 1859 an hermaphrodite, an hexagonal, an herbaceous, an hereditary, an homopterous British English e Voyage of the Beagle, 1909 an herbarium, an horizon, an humble(2) E.A.Poe 1809 - 1849 e Fall of the House of Usher 1839 an habitual American English R. H. Dana 1815 - 1882 T wo Years before the Mast 1840 an hermaphrodite American English Ch. Dickens 1812 -1870 A Tale of T wo Cities 1859 an hotel a honest (3) (Mr. Cruncher ), a honouring (Mr. Cruncher) British English Great Expectations 1860-1 An hotel a honour ( Joe) Oliver T wist 1837-8 an habitual, an habitual, an hysterical a honour (Charley), a honour (Mr. Sikes) H. Melville 1819 - 1891 Moby Dick 1850 an hypothesis American English G. B. Shaw 1856 - 1950 An Unsocial Socialist 1883 a honest (2) British English K. Grahame 1859- 1932 e Wind in the Willows 1908 an hysterical British English J. K. Jerome 1859-1927 ree Men in a Boat 1889 an hotel (2) a hotel(1) British English Both forms A. C. Doyle 1859 - 1930 e Memoirs of Sherlock Holmes 1894 an historical, an Havana a history British English E. F . Benson 1867 - 1940 Michael 1916 an hallucination a happy a hurry a hundred (12) British English B. Russell 1872 - 1970 Proposed Roads to Freedom 1918 an historical a history British English J. Buchan 1875 - 1940 e irty – Nine Steps 1915 a history, a hotel a hundred (4) British English Greenmantle 1916 an hotel (2), an hydraulic, an hysterical A hotel, a hideous life, a hundred Both forms e first conclusion to draw is that in the works of authors such as Defoe or Jane Austin there is vacillation in the form of the indefinite article before words beginning with h in stressed syllables (e.g. hundred, happy, horrid, husband, hurry). ey seem to be in free variation. But later, with Darwin, the form an is used in front of words beginning with h if the syllable is unstressed. Words with a stressed syllable preceded by an (besides an a form) are reduced to humble and hotel both with an earlier h-less form. Beginning with the works of Dickens we find the form a in front of the words containing a silent h. Here we are, fittingly, in the 19 th century, when concern about a “correct” form of speech became more marked than before. People became increasingly aware of forms of speech different from those of their immediate circle. It is in the 19 th century that London became a favourite setting for fiction and with Dickens the Cockney speaker entered English literature. He was the most extensive and successful depicter of Cockney speech. Convinced that people are their behaviour, are their words, are their gestures, Dickens uses visual suggestions of non-standard speech rather than phonemic transcription of speech. Of the 6 salient phonemic features which characterise Cockney speech, h- insertion and h- dropping are the most constant signals. Dickens makes use of the apostrophe for slurred speech (e.g. unnat’ral), and because it is difficult to signal h insertion he uses the form a in front of the words with silent h such as honest and honour. He does not exploit the Cockney for comic purposes. He saw the Londoners as individuals with their own culture. He uses the a form of the indefinite article to represent the speech of Mr. Cruncher the uneducated and superstitious odd-job- man in the Tale of Two Cities. “It’s enough for you,” retorted Mr. Cruncher, “to be the wife of a honest tradesman, and not to occupy your female mind with calculations when he took to his trade or when he didn’t. A honouring and obeying wife would let his trade alone altogether.” e form a instead of an is also found in the speech of Joe the uneducated blacksmith in Great Expectations, and of the two thieves Mr. Sikes and Charley Bates in Oliver Twist: “Which you have that growed,” said Joe, “and that swelled, and that gentle-folked;” Joe considered a little before he discovered this word; “as to be sure you are a honour to your king and country.” ‘She’s a honour to her sex,’ said Mr. Sikes, filling his glass, and smiting the table with his enormous fist. ‘Here’s her health, and wishing they was all like her!’ ‘Well, it is a honour that is!’ said Charley, a little consoled.” It is extremely difficult to represent in writing the addition of [h] before a word spelt with a mute h (the consequence of the unphonetic character of the English spelling); this is why ackery used “Hhonour”. Elizabethan, and even 18 th century, authors, who represented vulgarisms so frequently, do not seem to use omissions and misplacing of h’s as characteristic of low class speech. After Dickens, G. B. Shaw also uses the form a instead of an before the word honest as a means of denoting non-standard English (in the put-on accent of Jeff Smilash alias Sidney T refusis in An Unsocial Socialist). As a matter of fact Shaw had some knowledge of phonetics (as seen from the preface to Pygmalion), so his Cockney reflects better the contemporary reality than Dickens’s. “#e young lady’s hi,” he said suddenly, holding out the umbrella, “is fixed on this here. I am well aware that it is not for the lowest of the low to carry a gentleman’s brolly, and I ask your ladyship’s pardon for the liberty. I come by it accidental-like, and should be glad of a reasonable offer from any gentleman in want of a honest article.” I continued my research investigating another corpus – this time a synchronic one, namely the BNC. e British National Corpus (BNC) is a very large corpus (100 million word collection of samples) of written (90%) and spoken (10%) language from a wide range of sources, designed to represent a wide cross-section of current British English, both spoken and written. I checked all English words beginning with h investigated by Feipel and this is what I obtained. I did not include the words which were preceded by the a form of the article only. an+Word Number Domain 5 a+Word Number Domain habitual 12 6 /11 7 I2, BT2, WA3, L3, SS2 habitual 17 / 15 A,L3, BT2, SS5, I2, WA3, Sk habitué 8 3 / 3 A2, C habitué 1 I hallucination 9 / 9 SS, I5, A, Sk2 hallucination 9 / 9 A4, I 5 harangue 1 I harangue 4 / 4 I3, BT Hellenic 3 / 2 A1, Sk 2 Hellenic 1 BT herald 1 Sk herald 8 I5, BT3, WA3 heraldic 5 /3 L, WA3, BT heraldic 3 /3 L2, WA1 herbaceous 1 L herbaceous 5 / 3 I, L3, NS hereditary 18 / 12 WA4, NS, SS2,L hereditary 30 / 28 L3, C2, NS15, WA15, I, AS2, Sk 1 hermaphrodite 2 / 2 A, I hermaphrodite 7 / 7 heuristic 3 / 2 AS3 heuristic 11 / 8 NS6, A,SS2, C, AS hiatus 1 A hiatus 27 / 27 hierarchical 7 / 7 AS2, SS3, WA2 hierarchical 102 / 69 hilarious 6 / 5 L, A3, SS2 hilarious 29 / 28 historian 42 / 33 historian 87 / 65 historic 159/ 130 historic 129 / 95 historical 191 / 135 historical 293 / 161 historically 13 /3 historically 18 / 15 history 2 / 2 BT , I history 820 / 481 histrionic 1 A histrionic 4 /4 I3, WA holistic 5 / 5 BT , L, SS3 holistic 52 / 43 holy 1 I holy 100 / 75 homogeneous 4 / 4 AS, C, SS, NS, homogeneous 100 / 72 Horatian 1 A Horatian 2 / 2 A2 horizon 2 / 1 A horizon 40 / 24 hospital 3 / 3 WA, SS, Sk hospital 823 / 435 hotel 80 / 61 I2,WA12, SS12; L7, A5, AS, BT3, C4. Sk8 hotel 764 / 423 human 1 C human 1242 humble 5 / 2 WA humble 85 / 72 hundred 10 / 10 A,WA2, SS2, Sk4, I hundred 4721 / 1375 Hungarian 1 A Hungarian 64 / 55 hydraulic 2 / 2 L, AS hydraulic 42 / 30 hypnotic 6 / 5 A, BT , SS, I, WA hypnotic 15 / 12 hypothesis 27 / 12 BT7, AS3, SS9, NS9 hypothesis 75 / 50 hysterical 17 / 16 SS6, A2, L1, I6, WA2 hysterical 27 / 25 A, I, SS, L, WA 5 6 7 8 If we compare the data obtained from the BNC and the data found by Feipel, we realise the differences are slight even if 100 years have passed between them. On one hand, three words (hidalo, Havana and hermetic ) were not found in the BNC; on the other, there is a tendency to use a preponderant a instead of a preponderant an in the words beginning with h in unaccented syllables: hereditary, herbaceous, hermaphrodite, histrionic, hydraulic, hypothesis, hysterical. So there are 35 words in Present-day British English beginning with h which are used with the two forms of the indefinite article. In the majority of cases the form a is, however, preponderant. e form an is preponderant in habitué, Hellenic and historic. ere is an even split in hallucination and heraldic. Feipel considered that historic and historical were evenly divided between an and a; According to the BNC the word historic seems to take more an and historical more a. A more thorough investigation is probably needed in this case. I also investigated the 4 words beginning with a mute h in the corpus and collected the following data: an+word Number Domain a+word Number Domain heir 83 / 83 Heir 0 honest 312 / 242 Honest 2 / 2 I, Sk 1 honour 122 / 122 Honour 0 hour 5967 / 1648 Hour 21 / 21 A3, WA2, SS2,I3, L2,Sk 8,C1 As can be seen from the table above, two of the words show variation in the form of the indefinite article: honest and hour. Honest preceded by a was found in the spoken data and once in the Imaginative domain; there were 8 occurrences of a in front of hour in the spoken data too. Unlike the examples found in the Gutenberg corpus the examples from the BNC are not examples of Cockney speech. So there seems to be great variation in the form of the indefinite article before h in present- day English, as there was also variation in the past. It is true that terms like “variability”, “variation” and “variety” are terms which have always held negative connotations. According to a large number of dictionary definitions they have been associated with some degree of unreliability, lowering of standards, or as falling wide of the accepted norm in many walks of life. We must admit that human nature has a strong tendency to favour conformity to standards, uniformity and conservatism, and to disfavour non-conformity, diversity, change. But as linguists we accept the widely-held view that language variation and change are natural processes rather than symptoms of degeneration and decay. In order to account for the variation of the form of the indefinite article in front of words beginning with h we investigated the history of /h/ in English. We analysed first the etymology of the words examined (found in OED on CDROM), and put down the date of their entry into the language, as can be seen in the following table: Noun Etymology Date of entry Pronunciation habitual Ad.med.L habitual-is, from OF habit, abit 16 th h↔∪bΙτΣΥ↔λ habitué F habitué,L habituat 19 th h↔∪bΙτϕueΙ 9 hallucination late L alucination-em, F hallucination 17th h↔,lu :sΙ∪νeΙΣ↔ ν harangue OF arenge , harangue,OHG hring 17th h↔∪ρΘΝ Hellenic L Hellenicus 17 th he∪λι:nΙκ herald OF herault but Germanic origin,OHG haren 14 th ∪heρ↔λδ heraldic F héraldique 18 th h↔∪ρΘλδΙκ herbaceous L herbace-us ,F herbacé 17 th h↔∪beΙΣ↔σ hereditary L hereditari-us 17 th h↔∪ρεδ↔τρΙ hermaphrodite Ad L hermaphroditus, a Gr 14 th h∈:∪μΘφρ↔δαΙ τ heuristic Gr, cf.Ger.heuristik 19 th hjΥ↔∪ρΙστΙκ hiatus L hiatus 16 th hαΙ∪εΙτ↔σ hierarchical Gr, cf F hierarchique 17th ,ηαΙ↔∪ρΑ:κΙκλ hilarious L hilari 19 th hΙ∪λε↔ρΙ↔σ historian F historien, L historia 16 th ηΙ∪στÿ:ρΙ↔ν 10 historic Ad L historic-us, a Gr 17 th hΙ∪στÿρΙκ 11 historical Ad L histori-cus, a Gr 16 th hΙ∪στÿρΙκ↔λ 12 historically Ad L histori-cus, a Gr 16 th hΙ∪στÿρΙκ↔λi 13 history L historia 14th ∪hΙστρι histrionic L histrionic-us, cf.F 18th ,hΙστρΙ∪ÿνΙκ holistic Gr. 20 th h↔Υ∪λΙστΙκ holy OE, OHG 11th ∪η↔Υλι homogeneous L homogene-us 17 th ,hÿ∪δΖι:νι↔σ Horatian L Horatian+us 18 th h↔∪ρεΙΣ↔ν horizon OF orizonte, L horizont-em 14 th h↔∪ραΙζ↔ν 14 hospital OF hospital, L hospitale 14 th ∪ηÿσπΙτ↔λ hotel F hôtel 17 th η↔Υ∪τελ 15 human F humain,L humanus 14 th ∪hju:m↔ν humble OF umble, humble, L humil-em 13 th ∪η℘μβλ hundred OE 10th ∪η℘ndrΙδ Hungarian L Hungaria 17 th η℘n∪γε↔ρΙ↔ν hydraulic Ad.L hydraulic-us, a Gr. 17 th hαΙ∪δρÿ:λΙκ hypnotic F hynotique, L hypnoticus 17 th hΙπ∪νÿτΙκ hypothesis Gr, cf.F hypothèse 16 th hαΙ∪πÿΤ↔σΙσ hysterical L hysteri-cus fromGr, cf F hysthérique 17 th ηΙ∪στερΙκ An investigation of their etymology shows that most of them are words borrowed from French or via French from Latin. e only 2 exceptions are holy and hundred which are OE words. All these words begin with what is conventionally referred to as a voiceless glottal fricative. ese are the two sources of the Present day English /h/in initial position. First, it is the OE sound h which in initial position was similar to the Mod. E. sound – e.g. habban (have) – and it is represented by the letter h. en, a new initial /h/ was introduced in English in words borrowed from French having an initial letter h. But the letter h in French has two values: a) a mute h in words of Latin origin : l’homme, un me , un ηομμε ανδ, b) a , b) a so called aspirated value which hinders elision and liaison: le homard, un homard. e “h aspirée”comes from words of Germanic origin and is now mute too. In Latin the letter h represented an aspirated sound in word-initial position. But apparently in the classical period a weakening of the aspirated sound took place in non-onomatopoeic words, which explains the recommendation of Quintilian in the 1 st century AD to keep h both in spelling and pronunciation. But the tendency among the “pseudo-cultivated” to insert h before words beginning with a vowel 16 was known to arouse the irony of contemporary writers in the 1 st century BC. e spelling of French words of Latin origin beginning with h was erratic in OF and ME, probably representing an h-less pronunciation. AN scribes whose command of English was incomplete often vary in their treatment of initial h – both omitting and inserting it incorrectly. It was common for different areas to have even different spelling conventions for the same sounds and these conventions had to be reconciled when a standardised spelling arose. In the 4 words already mentioned (heir, honour, honest, hour) the h-less pronunciation had been retained in spite of the spelling (the explanation lies in their frequent use); but in most cases the letter h began to be pronounced in late ME and early Modern English – a phenomenon that continued well into the 19 th century. e words: herb, humble, hospital, and humour are examples of late sounding of h. So writing affected speech. V . Fromkin (1993 ) considers that spelling influences the pronunciation of infrequently used words and gives the example of not pronounced in hour and honest but pronounced elsewhere. For example herb is undergoing sounding of h now. It is pronounced in Standard British English but not in Standard American English. e use of the form an of the indefinite article in front of a word beginning with h may reflect the loss of [h] in pronunciation. Lack of [h] in words adopted through French from Latin does not depend on English sound change, but on French pronunciation. But we can talk about pre-vocalic h-loss 17 as an English sound change. Pre-vocalic [h] loss refers actually to two 18 different phenomena: one that belongs to vulgar or dialectal speech confined to stressed syllables, and one a normal process that belongs to educated speech and is confined to unstressed syllables. Milroy, who studied the history of h loss in stressed syllables, reached the conclusion that the phenomenon started in Early ME about the 14 th century and had a certain amount of prestige before then spreading to the lower orders of society, and from urban to rural regions. But there seems to have been a period in the 16 th century when h-dropping must have been associated with ignorance and lack of education, and therefore avoided (mainly by the middle class). It seems that the upper and middle classes would treat h as a stylistic variable, omitting the etymological h in informal speech and restoring it in more careful and public speech. By the 18 th century /h/ dropping was certainly recognised as vulgarism. en there was the second [h] loss phenomenon in unstressed syllables, which is viewed by linguists as a “more normal process” but, as mentioned by Dobson, there is very little evidence of it. is phenomenon accounts for the variant pronunciation of 7 words in our corpus in the two up-to-date pronunciation dictionaries used and for the variation in the form of the indefinite article. But one thing is normally neglected in the presentation of the history of h, namely that it is a weak segment in English (it is regarded as a strong, voiceless onset of the vowel that follows), and therefore it is a candidate for spontaneous loss. Linguists have looked for evidence of h- loss in OE but have not found any, although there may have been some loss in OE according to Milroy. H-loss in Germanic languages does not seem natural – probably because of the heavy stress on initial syllable. Anyway, what is sure is that the French-English contact was the single most important influence on the rapid progress of h-loss in ME. As far as the history of polite “Received” English is concerned it seems to have been one of slow restoration of a segment that had been lost in many dialects and was variable in some others. is explains why words beginning with h before a weakly stressed vowels, as in historic, Hungarian, hysteria, etc., are pronounced now with a restored [h], while the literary convention persists of writing an, and even pronouncing it, before the following h. So we can view h as a variable with a long and fluctuating history characterised by a general tendency to progressive and conscious stigmatisation of h-loss. And in conclusion, as an explanation for the two existing forms a historical or an historical (novel) has been given, it seems that these two variant forms can be tolerated for a long period without discernible movement toward reduction of variants. is phenomenon is called in the literature “personal - pattern variation” as opposed to geographical variation, stylistic variation, etc. As for the variant forms of the indefinite article in front of hour and honest found in the BNC I wonder if this is not to be considered as hypercorrection on the part of some middle-class speakers who are afraid of dropping the [h] and insist on pronouncing it even when it should not be pronounced just because it exists in spelling.