Spelling pronunciation
From Wikipedia, the free encyclopedia
- Spelling pronunciation is different from pronunciation spelling
A spelling pronunciation is a pronunciation that, instead of reflecting the way the word was pronounced by previous generations of speakers, is a rendering in sound of the word's spelling. Spelling pronunciations compete, often effectively, with the older traditional pronunciation.
Contents |
[edit] Examples of English words with common spelling pronunciations
- often, pronounced with /t/, though the pronunciation without it is more prevalent. Older dictionaries don't even list the pronunciation with /t/, though the 2nd edition of the OED does (and the first ed. notes the pronunciation, with the comment that it is prevalent in the south of England and "often used in singing" (!); see the Dictionary of American Regional English for contemporaneous citations discussing the status of the competing pronunciations)
- clothes was historically pronounced the same way as the verb close ("Whenas in silks my Julia goes/.../The liquefaction of her clothes" --Herrick), but many speakers now insert a /ð/ sound
- salmon, occasionally pronounced with [l]
- falcon /fælkən/ is standard; cf. the family name Faulkner and a long history of variant spellings without -l-
- comptroller, often pronounced with [mp]; accepted pronunciation is "controller" (and in fact the -mp- spelling is bogus, based on the mistaken idea that the word has something to do with comp(u)tare "count, compute"; in fact it comes from contre-roll "file copy", the verb and its agent noun meaning "compare originals and file copies")
- ye the article, pronounced as if spelled with a Y instead of the printers' mark for the letter thorn
- taking the "insular flat-topped g" of northern scripts as a -z- in names like Mackenzie, Menzies, Dalziel (in the last with the value of /y/ originally)
- tortilla and other words from Spanish with the double-L; similarly maraschino (cherry) with /š/ for proper /sk/
- victuals "vittles" whose -c- (for a consonant lost long before the word was borrowed from French) was reintroduced on etymological grounds, and sometimes pronounced with [kt]
- The pronunciation of waistcoat as spelled is now more common than the previous pronunciation "weskit"
- conduit, historically pronounced "kondit" or "kundit", is now nearly always "conndewit"
- medicine, historically pronounced with two syllables but now quite often with three (some speakers use two when they mean medicaments and three when they mean medical knowledge; three syllables is standard in the USA)
- figure originally rimed with bigger (and still does in the Queen's English); in America the approved pronunciation follows the etymological spelling (copied from Latin figūra)
- trait (traict), has a complicated history: a 15th cent. borrowing from French, it came to be normally pronounced /trey/ in 19th century Britain, by imitation of the current French pronunciation; /treyt/ is gaining in Britain, though, and was always standard in the USA
- Bartholomew formerly pronounced "bartleme" (three syllables, accent on the first) now /barθáləmiw/. (The current standard pronunciation makes hash of the meter of the folk-song Bartholomew Fair.) Similarly Anthony (< Lat. Antonius), now (in USA) /ænθəniy/
- Probably to be included in this general category are the place-names whose traditional ("old fashioned") pronunciations have been displaced by ones influenced by the spelling: St Louis, formerly /sænt luwiy/ now /seynt luwis/, Papillion (Nebraska), formerly /pæpiyow/ now /pəpilyən/, Los Angeles formerly /los æŋgləs/ now /los ænǰələs/, Beatrice (Nebraska, home of Sara Lee) formerly and still somewhat currently /biyǽtrəs/, now /bíyətrəs/
- Personal and as a rule evanescent are such misprisions as taking misled to be the participle of a verb "to misle", construing malefactor as male-factor (a hormonal problem of some sort)
-
- There are also what might be called "misspelling pronunciations", as /sweydow/ for pseudo, lamblast for lambaste, chase lounge for chaise longue "long chair", /eriyeyt/ for aerate, /larniks/ for larynx, possibly /eksetərə/ for etcetera.
[edit] Spelling pronunciations and history
The purpose of writing is to make a permanent record (of whatever duration) of language, whose normal domain is the speech act, which by its nature vanishes instantaneously. Given the complex organization of human language, the symbols of a writing system might refer to various components of language, most commonly words (as in the numerals 1, 2, 3, etc.) or sounds. The latter types are known as phonographic writing systems. Syllabaries write syllables, i.e. the symbols stand for one or more sounds one of which is always a vowel; the symbols of an alphabet stand for individual phonemes. (See syllabary, alphabet.)
In phonographic systems, the degree of regularity with which symbols represent sounds is known as fit. The tightest possible fit would be when a given sound is always represented by the same symbol, and a given symbol always represents the same sound and only that sound. This is rarely found in nature. In many languages written phonographically which have a contrast of vowel length, for example, the contrast may well go unrepresented in the orthography (as in Latin) or partially represented (as in Athenian Greek, whose writing system distinguished between long and short mid vowels, but not high and low vowels). In standard German orthography, vowel length is partly not shown at all in writing (Weg "way, road" has a long vowel; weg "away" has a short one) and partly indicated by inconsistent conventions. In any case, the use of alphabetic diacritics, as in the letter -h- used in English orthography in combination with other letters for specific purposes (th, ch, sh) is only a mild degradation of fit, since ambiguous cases—ambiguity is the inevitable consequence of suboptimal fit—are rare.
The fit of a phonographic writing system is generally best when the language is first "reduced to writing" (as the telling phrase has it), though even in that Edenic state there will almost always be problems if the writing system was originally developed in connection with writing a different language (as is very often the case; signaries are very rarely invented anew). But there is a strong tendency for phonographic written forms, once they become familiar to their users, to remain unchanged even as the language itself changes. This progressively complicates the fit between symbol and sound as phonological splits, mergers, loss, and so on, go unreflected in the orthography.
The fit of written and spoken English is particularly and famously poor. The explanation is of course historical, but is more complicated than the brief statement above. Modern English orthography represents the accumulation of several unrelated writing systems, superimposed one on top of the other. Old English orthography had a very good fit, improved by special symbols to represent sounds not easily symbolized with "Latin values", such as the initial consonant of thin, three, and of wool, wolf (using in those two cases letters from runic signaries), and there were some ingenious adaptations of the Latin signary to the needs of representing Old English, such as using the letter y to represent a high front rounded vowel. These orthographic conventions were in many cases ignored by Anglo-Norman scribes, who introduced such preciosities as qu in place of Old English cw, wrote /ū/ with the letters ou as in French, abandoned the runic letters, and fiddled around with alphabetic diacritics. In at least one case this was an improvement: Old English had a contrast between /k/ and /č/, but both were written with the same letter: so cén "eager" > English keen, but cinn > English chin. On top of this, from Middle English on, English writers adoped the French practice—unique in Romance languages—of writing words borrowed from Latin in an orthography that was, insofar as possible, unchanged from the Latin source, so philosophy, say, cf. Italian filosofia; nation, say, cf. Italian nazione. And of course all of these conventions were largely indifferent to ongoing changes in the language. (When nation was borrowed, it had three syllables, /nāsiūn/ or the like. The rest, as they say, is history.) When our standard orthography started to jell (or gel or gell), there was still a spoken contrast between write, right, rite, and wright, which accounts for their spelling; the vowels of meat and meet were still different; knight and gnat stilll began with a consonant cluster; and so on. These spellings persisted even as changes in the language progressively degraded the fit between the written and spoken forms.
The extraordinarily inconsistent fit between symbol and sound in English has several consequences. Given "silent letters" all over the place, it is easy to introduce letters into words that represent literally nothing at all, not now, not ever, such as the s of island or the w of whole and whore. It is tempting to "restore" (in writing anyhow) letters from earlier states of the language, as when respelling portrait, perfet, det, scism, physionomy as portraict, perfect, debt, schism, physiognomy. You will recognize that the respellings of perfet and physionomy have permanently influenced the pronunciation of the words (though /gn/ in physiognomy is commoner in the US than in Britain); the respelling of scism has given a new pronunciation /skizəm/ a good foothold; the etymologification of debt has been without effect, apart from the headache of remembering the spelling, and portraict failed to make the final cut.
In any case, a literate speaker of English is chiefly bedeviled by the task of having to remember which letters are used to write which sounds in each particular word; there is no method for "doping out" whether a word with the vowel /iy/ should be spelled -ee- or -ea- or even -i- (as in elite and machine) or -e- (as in compete, penal), -ei- (in words of the receive family, and seize, weird) or -ie- (siege, etc.). In most cases our standard spellings are etymological, but by no means all (and that is little or no help anyhow). Whether /ow/ is spelled -oa- or -oCe is largely a matter of chance: road and rode both reflect Old English rád /rād/. But a secondary problem runs the other way: given the general feeling (completely wrong, but general) that writing is primary and language is something evanescent and sloppy that should do its hopeless best to be faithful to the written word, it is unremarkable that from time to time an orthographic form whose fit with the spoken form is poor will be taken seriously by speakers who substitute a kind of guess, based on the letters, for the pronunciation transmitted from speaker to speaker through the generations. The great majority of such guesses are transitory, being soon discarded with reference to the received pronunciation ("mistakes" one might call them). Some however become standard, as in some of the forms mentioned in the section at the beginning of this article.
And while it is true that some such spelling pronunciations have the effect of "reversing history", i.e. restoring (in a sense) sounds lost through regular phonological change, not all do. Many, perhaps most, spelling-based pronunciations of foreign words do no such thing. Pronouncing the -mp- of comptroller, as pointed out above, does no such thing. An interesting example is equip and its derivatives, from French équiper etc. This is a loan word in French, from Low German skippan "to make a ship ready for sea, with provisions, crew, maintenance, etc." The word was borrowed in time for the initial s-cluster in French to become é-; the spelling -qui- is a pure orthographic convention for representing /ki/ (the change of Romance qui, que to /ki, ke/ having provided French scribes—thanks to the persistence of written types in disregard of language change!—with an unambiguous way of writing /k/ before front vowels). That is, there never was a /w/ anywhere in the history of this word, apart from its written representation. A more recent example would be the novelty loverly (from My Fair Lady lyrics: "Ah, wouldn't it be loverly"): the spelling is in an r-less dialect, i.e., British English, and is intended to show the "vulgar" pronunciation of the word in three syllables: love-a-ly, or so. And of course there's no "reversal" in pronouncing /θ/ in thesaurus or Anthony (USA).
All things considered, it's odd that spelling pronunciations aren't commoner in English than they are. Possibly part of the explanation is that literate speakers of English from an early age get so used to the chaotic fit between orthography and pronunciation, the welter of letters that don't stand for anything, and so on, that they're immunized to a degree against changing pronunciations to match the spelling. But not entirely.
[edit] Spelling pronunciation vs. analogical pronunciation
In some cases, we cannot tell if a pronunciation is a true spelling pronunciation. The alternative is that a word is being pronounced analogically, in essence as the "sum of its parts". Thus, forehead is commonly pronounced as a sequence of fore plus head, instead of the historically earlier "forrid"; and waistcoat is commonly pronounced as a sequence of waist and coat, instead of the historically earlier "weskit".
Analogy in this sense (also known as recomposition) can be confused with reanalysis. For example, inmost comes from Old English innemest, which contained the ordinary superlative suffix -est. The later switch to in + most was due to reanalysis of -mest as -most (and led to the creation of a whole family of words of relational meaning: northernmost, outermost, uppermost, etc. Foremost is unusual in this group in having much the same history as inmost, being from OE fyremest, superlative of the word giving modern English former).
[edit] Opinions about spelling pronunciation
Spelling pronunciations give rise to varied opinions. Often those who retain the old pronunciation consider the spelling pronunciation to be a mark of ignorance or insecurity. Those who use a spelling pronunciation may not be aware that it is one, and consider the historically authentic version to be slovenly, since it "slurs over" a letter.
Fowler reports that in his day there was a conscious movement among schoolteachers and others encouraging people to abandon anomalous traditional pronunciations and "speak as you spell".
But it has often been remarked that this trend, though understandable from a socio-psychological point of view, is, from a strictly linguistic perspective, irrational, since writing was invented to represent the sounds of the language – obviously not the opposite: the true, or original, form of the language is the spoken language (cf. the two meanings of the English word tongue, with equivalents in many other languages: Greek glótta, Latin and Italian lingua, French langue, Spanish lengua, etc.; and there are in the world many millions of analphabets, who nevertheless all speak their native languages, sometimes pretty well). Therefore, there is no good reason to "speak as one spells", but there are many good reasons to "spell as one speaks", i.e. to reform the orthography of a language whenever it does not render its pronunciation clearly and unambiguously – which is the task of a writing system. How easy such a reform would be in practice, that is of course quite another matter.
Quite different from spelling pronunciations are phonetic adaptations, i.e. pronunciations of foreign words within the frame of the phonematic system of the language that accepts them: an example of this process is garage ([ga'ʀa:ʒ] in French) sometimes pronounced as ['gæɹɪʤ] in English. Such adaptations are quite natural, and often preferred by speech-conscious and careful speakers.
[edit] Spelling pronunciations in children and foreigners
Children who read a great deal often produce spelling pronunciations, since they have no way of knowing, other than the spelling, how the rare words they encounter are correctly pronounced. Thoughtful parents usually try to correct such children's errors gently. But as this can never extend to every instance, and there are many words which one reads far more often than one hears, what is a spelling pronunciation in one generation often becomes standard in the next.
Well-read second language learners are likewise vulnerable to producing spelling pronunciations.
[edit] In other languages
In French, the first vowel in oignon (onion) is, anomalously, /o/, where general principles would lead one to expect [wa]. The reason is that the spelling of this word is a hangover from the 17th century, when "i" was invariably inserted before "gn": montagne was spelled "montaigne", but pronounced in the same way as today. However, there are provincial school-teachers who insist on pronouncing oignon with a [wa] (personal experience). (The French Academy has recently (1975) decreed an official change in spelling to ognon.)
When English club was first borrowed into French, the approved pronunciation was /klab/, as being a reasonable approximation of the English. Now the standard is /klyb/ (Littré, though Larousse and Oxford prefer /klœb/), on the basis of the spelling. Similarly, shampooing "product for washing the hair" at the time of borrowing was /ʃãpuiŋ/; now it's /ʃãpwæ/
In Hebrew there is a vowel called patach ganuv, consisting of an "a" sign placed underneath a final guttural but pronounced before it: an example is ruach, which looks as if it ought to be *rucha. Where the final consonant is a sounded he (h), many speakers do indeed place the vowel after it, mistakenly pronouncing Eloah (God) as "Eloha" and gavoah (high) as "gavoha". Other examples of spelling pronunciations are the Sephardic "kal" and "tsahorayim": see Sephardic Hebrew language.
In Spanish, the French spelling of "élite" drives speakers to stress the initial "e" although in the original the "i" is stressed (accent marks in French, unlike those in Spanish, don't denote stress).
[edit] Books
- See the index entries under "spelling pronunciation" from Leonard Bloomfield, Language (originally published 1933; current edition 1984, University of Chicago Press, Chicago; ISBN 81-208-1195-X).
- Most of the etymologies and spelling histories above are taken from the Oxford English Dictionary.
