Spelling pronunciation

Spelling pronunciation occurs when people display their literacy by using a pronunciation that more closely reflects the spelling, under the impression that to do so must be more correct.

In English

 * The áccents show stress and pronunciation: see English phonemes

Spelling pronounciation is an ongoing phenomenon in English and can be attributed to the increase in literacy over the last half-century. Broadcasters such as the BBC and Al Jazeera, especially in their news bulletins, are particularly prone to it. A typical example is the sounding of the h before r in Bahrâin, Tehràn and Tahrìr. Indeed it is often the sounding of a particular consonant, which may have traditionally long been silent, that is at issue. Further examples are:
 * the t in óften
 * the l in sálmon
 * both rs in lîbrary or Fébruary (should it end exactly like Jánuary?)

In the case of sálmon, sounding the l is an occasional American phenomenon, and indeed there are other cases of spelling pronunciation that constitute recognised differences between British and American English, most notably the suffix -ary (unstressed and with a silent a in British)/-ãry (like -érry in American).

Depalatalisation
Much spelling pronunciation concerns depalatalisation, usually that of s, c or t when followed by i or u as an 's' sound where a 'sh' is normal. A classic example is a word beloved of broadcasters, íssue: should it be pronounced *íssyûe (IPA /'ɪsju/) or palatalised as *íshûe (IPA /'ɪʃu/)? Clearly, the unpalatalised version *íssyûe reflects the spelling and is of greater antiquity, but is has long been the norm to say *íshûe. Even so, *íssyûe has never quite gone away, as it is perceived that resisting the urge to palatalise to the 'sh' sound is somehow more correct. The same can be heard with tíssue, and of course there is no palatalisation in assûme and consûme: both always have -syûme: a -shûme pronunciation would sound comically uneducated. But foreigners should learn to palatalise words that normally have this pronunciation, if they are not to sound mannered. Nobody says *insyûrence for insûrance; indeed it most often sounds like *inshŏrance (British) or *inshûrance (American).

And sûre itself retains its palatalisation: AmE *shûre, BrE sůre certain = shŏre sea: it does not sound like seŵer. spêcies however has been showing signs of returning from *spêeshíz to *spêessíz or *spêessêez. Meanwhile, Cambôdia, ‘Cambodge’ in French, was for a time in the 1990s Kampuchêa in English: we can see that the -dia represents a palatalisation to a 'j' sound that has been lost in English, and that the K- spelling was a forlorn attempt to restore it in the form of a 'ch'.

Other words not normally palatalised are: assûme, dûe, ensûe, euthanâsia (-zìə), Galícìa (see below), Parísìan (-zì-), presûme (-zyûme), pursûe, redûce and other words in -dûce, sûicide, sûit (cf. shoôt), synaesthêsia (though anasthaêsia might be), Tunísìa (BrE *Cheŵ-níziə, AmE *Tunìzhə) and Valéncia. Other normally palatalised words are: amnêsia (-zhə), apprêciâte, Âsia, assôciate, atrôcious, caucâsian and other words in -sian (-zhn), Chrístian (*Kríshchən) and other words in -tian (-shn), Indonêsia, milítia, negôtiate, Patrícia, Përsian, pléasure (-zh-), précious and other words in -cious, préssure (-sh-) and other words in -sure, sôciology, sôldier, substántial, Croâtia (Crô-, and other words in -tia), vísion and other words in -sion (-zhn), vítiate, volítion and other words in -tion (-shn).

Ending -ies lengthening from -íz to -êez
From around the beginning of the twenty-first century, another very strong tendency, perhaps influenced by the spelling pronunciation of some foreign learners, is to lengthen the vowel sound in the ending -ies in nouns, but, interestingly, not in verbs (cárries is still *cárríz). Apart from monosyllables (the bêes' knêes) the -êez sound has traditionally been used only for the plural of words from Greek ending in -is: crîsis, plural crîsês (*crîsêez), but it can now be heard in the plurals of nouns in -y, such as pàrties, in the speech of younger people (and Tony Blair):

This appears to be the continuation of a change, as we can hear in films from the mid-twentieth century how the -y ending too was once pronounced í. But this new change can be considered unfortunate for listeners, especially learners, as it reduces the distinction between spoken singular and plural to just the 'z' sound (written s).

Sometimes Dâvies can be heard with this pronunciation (often in Ameican English), instead of being homophonous with Dâvis. The tendency, often heard on BBC News bulletins, may also be spreading to the past tense of verbs ending in -y: accòmpaníed becoming accòmpanìêd; míschíef can also be heard with the spelling pronunciation, and one can also hear intéstíne pronounced like Chrístìne.

Other English spelling pronunciations
These are arranged alphabetically:
 * The strong form of the indefinite article a (in normal speech always the weak form, schwa, [ə]), pronounced like the letter's name, *ây, can be heard in self-conscious speech: to withstánd â fûture dròught
 * A regular pronunciation of âi in agâin and agâinst can sometimes be heard instead of *agén and *agénst. This is perhaps unfortunate, as it can be confused with a gâin: 'celebrating a gain' or 'celebrating again'?
 * While it is normal to pronounce the n in autúmnal, because of the change of stress from åutumn (*åwtəm), one now sometimes hears also the n of cólumn (*cóllum) in cólumnist
 * An aspirated h in Bahrâin: it represents an Arabic sound, but traditionally just gives the àh sound in English, without being pronounced separately, as h is not normally sounded before r; see also Tehràn, below
 * For Cátalan many people say 'Cátalán', presumably under the impression it is more correct, but they would never do this with Itálian or indeed Chílean
 * A Spanish-style Colómbia instead of Colòmbia (= Colúmbia British, trademark) as if there were any danger of confusion (also Colómbo for Colòmbo). Similarly, some affect a Spanish pronunciation of Chílê, which traditionally sounds exactly like chílly. There has also been a slight tendency to pronounce Chílean ‘Chilèan’, perhaps inspired by the Spanish version ‘Chileno’
 * There are various pronunciations of Fébruary: in AmE *Fébûãry, like Jánûãry, and in BrE *Fébry or *Fébyəry, like *Jányəry. The spelling pronunciation pronounces both rs: AmE Fébrûãry, BrE *Fébrəry
 * If Galícian were as common as patrícian (*pətríshən), one would not hear *Gəlíssiən, let alone 'Gəlíthìən' (the latter after the European Spanish 'th' sound of the 'c'). Similarly, if Ossétia were not relatively new to most broadcasters, one would expect it to rhyme with Lucrêtia *Osêsha; as it is, one is tempted to misspell it 'Osettia'; the same applies to Ingushétia
 * A regularised Kÿrgyz can be heard replacing the spelling pronunciation Kỳr- in that and Kyrgyzstān
 * In óften, *óffen, some people pronounce the traditionally silent t
 * Monosyllables with -oor and -our: toûr alternates with tŏur; BrE pŏor increasingly sounds like poôr, although dŏor remains unaffected
 * The pronunciation of prãyer traditionally rhymes with lãir (of an animal); the spelling pronunciation has two syllables and rhymes with lâyer
 * The traditional *crézzent may be giving way to a hissing -ss- sound in créscent; similarly, péssimism and péssimist can be heard with a hissing sound instead of the usual péz- pronunciation
 * Some speakers sây sâys and sâid, but most still sây *séz and *séd
 * Shreŵsbury is the spelling pronunciation of the English town long known as *Shrôzebury
 * Sûéz instead of *Sûíz
 * Tehràn is traditionally *Té-ràn: the Persian pronunciation, pronouncing the h after a closed [e] sound, is reflected in the spelling, but since the sequence -ehr- does not come naturally in English (hence an alternative, but little used, spelling, Teheràn) there has developed a tendency to lengthen the first vowel to the sound in èh?, making the syllable sound like Tây; cf. '''Bàhrain
 * Weak pronunciations of syllables traditionally with the schwa sound sometimes give way to strong ones: an example is Thaîlánd /taɪlænd/ (sounding like two words, 'tie land') instead of Thaîland /taɪlənd/: one would not expect such a pronunciation for England, Scotland, Ireland, Iceland, Greenland or New Zealand, and it recalls the days of the British Empire: Basûtolánd, Swàzilánd  and Matabêlelánd do indeed usually retain such a pronunciation. And similarly, there is a tendency for people called Jûliet and Márian, with schwa as the final vowel, to pronounce their names more like Juliétte and Mariánne
 * In the suffix -wíck the w is usually silent, but in Lërwick, it is sounded – compare Ẁarwick, Bérwick, Késwick and Sméthwick (*Wórrick, *Bérrick, *Kézzíck, *Smédhíck)
 * The suffix -wích: Woòlwich, Gréenwich and Nórwich, which traditionally echo pórridge, can now be heard with final consonant unvoiced, as always in Ípswich

Glottal starts
In the opposite direction to spelling pronunciation is a trend to replace the traditional word-linking sounds r and w with glottal stops (\) even when this means leaving silent a written letter. This may also be the result of a foreign learners’ habit, that of pronouncing each word separately, without liaisons. These could be called 'glottal starts', as they attach themselves emphatically to the vowel beginning the second word: with the loss of a linking r sound: and even where written as one word:

But this does not happen with the consonantal 'y' liaison: hîgh úp is still normally pronounced *hiyúp. And conversely, people still tend to join words where the first ends in a with an imaginary, invisible 'r', by analogy with the -er ending: Chîna-América relâtions (*Chinərəmericərəlâshənz).

Other converse trends in English
One often hears Beijíng pronounced with a 'zh' sound instead of a normal j.

Sometimes the names of places in the Spanish-speaking world that have c before e or i, such as Valéncia, are pronounced with the unvoiced 'th' sound, 'Valénthia', echoing the pronunciation of Madrid (not all Spanish has this sound). Also, Nicarágûa is giving way to a more Spanish-sounding *Nicarágwa/-àgwa.

Some English speakers pay homage to a local pronunciation by calling Kósovo 'Kosova' (the fact is that Kosovo is the Serbian name and Kosova is the Albanian one).

Tbilìsi can sometimes be heard with an extra, invisible 'i', 'Tíb-', replacing the usual schwa sound before the b (*Təblêessy).

In Catalan
In Catalan, x was traditionally pronounced [s] before a consonant. It is now more and more pronounced [ks] in this position:
 * text “text”: traditionally [ˈtest], now usually [ˈtekst].
 * explicar “explain”: traditionally [əspliˈka], now usually [əkspliˈka].

In French
Several inconsistencies of French spelling lead to spelling pronunciations.

Double consonants
The double consonants ll, mm and nn, usually pronounced as single consonants ([l], [m], [n]), may be phonetically doubled—or more exactly lengthened—in some words. This infringes the traditional pronunciation rules (explicitly, phonotactics) of French, where doubled or lengthened consonants are usually impossible inside a word:
 * illusion “illusion”: normally [ilyzjɔ̃], becoming [illyzjɔ̃].
 * sommet “summit”: normally [sɔme], becoming [sɔmme].
 * grammaire “grammar”: normally [gʀamɛʀ], becoming [gʀammɛʀ] (note that the pronunciation [gʀɑ̃mɛʀ] may be also heard: it is archaic or regional).
 * inné “innate”: normally [ine], becoming [inne].

Silent or pronounced letters
Silent letters (whether they are archaic or artificial) may be pronounced:
 * legs “legacy”: traditionally [lɛ], now usually [lɛg].
 * Aulnay-sous-Bois (a town in Ile-de-France): normally [onɛ su bwa], becoming [olnɛ su bwa].
 * Lanoux, Giraudoux, Amouroux, Roux (surnames): normally [lanu, ʒiʀodu, amuʀu, ʀu], becoming [lanuks, ʒiʀoduks, amuʀuks, ʀuks].
 * sioux “Sioux”: normally [sju], becoming [sjuks].
 * août or aout “August”: traditionally [au], now [u] and more and more [ut].
 * persil “parsley”, cil “eyelash”, sourcil “eyebrow”, chenil “kennel”, gril “grill/broiler”: traditionally [pɛʀsi, si, suʀsi, ʃəni, gʀi], now usually [pɛʀsil, sil, suʀsil, ʃənil, gʀil] (note that other words with a similar -il-ending keep a silent l: gentil “nice”, outil “tool” are always pronounced [ʒɑ̃ti, uti]).

Difficult reading interpretation
In some words, graphemes with an irregular pronunciation (or at least with a difficult reading interpretation), may be pronounced according to the general value of the graphemes:
 * gageure or gageüre “challenge”: normally [gaʒyʀ], becoming [gaʒœʀ] (the sequence ge should be pronounced [ʒ] before u, but speakers, instead of analyzing ge+u, see the more usual sequence g+eu where eu is normally pronounced [œ]).
 * geôle “jail”: normally [ʒol], becoming [ʒeol] (the sequence ge should be pronounced [ʒ] before ô, but speakers, instead of analyzing ge+ô, see the more usual sequence g+é+o).
 * Bruxelles “Brussels”: normally [bʀysɛl], becoming [bʀyksɛl].
 * Auxerre (a town in Burgundy): normally [osɛʀ], becoming [oksɛʀ].
 * Metz (a town in Lorraine): normally [mɛs], becoming [mɛts].
 * Montpellier (a town in Languedoc): usually [mɔ̃pəlje], but also [mɔ̃pelje] (theoretically, e is pronounced [e] before two consonants; the native, Occitan name Montpelhièr [mumpeˈʎɛ] could also explain [mɔ̃pelje] in French).

Œ
The grapheme œ, usually pronounced [œ] in the common word œil [œj] “eye”, should theoretically be pronounced differently, [e/ɛ], in learned words of Latin or Greek origin. However, some speakers prefer to use the [œ]-like pronunciation (which may be [œ] or [ø], depending on adjacent sounds):
 * œsophage “(o)esophagus”: normally [ezɔfaʒ], but usually [øzɔfaʒ].
 * œdème “(o)edema”: normally [edɛm], but usually [ødɛm].
 * œstrogène “(o)estrogen”: normally [ɛstʀɔʒɛn], becoming [œstʀɔʒɛn]. Interestingly, this word is often re-spelled estrogène by doctors and biologists in order to promote the pronunciation [ɛstʀɔʒɛn].

Ch
The grapheme ch is usually pronounced: However, since the grapheme ch is ambiguous, several words originating from Greek have adopted for ch an inconsistent [ʃ]-sound instead of a consistent [k]-sound: Sometimes, this inconsistency causes contradictory pronunciations within a same word family:
 * 1) [ʃ] (the English sh-sound) in words from non-Greek origin: chien [ʃjɛ̃] "dog", riche [ʀiʃ] "rich", méchant [meʃɑ̃] "nasty".
 * 2) [k] in words from Greek origin (where ch comes from the Greek letter χ): archaïque [aʀkaik] "archaic", chaos [kao] "chaos", tachycardie [takikaʀdi] "tachycardia".
 * machine [maʃin], not [makin]* "machine"
 * chimie [ʃimi], not [kimi]* "chemistry"
 * chimère [ʃimɛʀ], not [kimɛʀ]* "chim(a)era"
 * psyché [psiʃe] "psyche" — psychédélique [psikedelik] "psychedelic", psychologue [psikɔlɔg] "psychologist"
 * bronche [bʀɔ̃ʃ] "bronchus", bronchite [bʀɔ̃ʃit] "bronchitis" — bronchiolite [bʀɔ̃kjɔlit/bʀɔ̃ʃjɔlit] "bronchiolitis" — bronchoscopie [bʀɔ̃kɔskɔpi]] "bronchoscopy"
 * chirurgie [ʃiʀyʀʒi] "surgery" — chiroptère [kiʀɔptɛʀ] "chiropter"
 * monarchie [mɔnaʀʃi] "monarchy" — monarque [mɔnaʀk] "monarch"
 * monachisme [mɔnakizm/mɔnaʃizm] "monachism" — monacal [mɔnakal] "monachal"
 * pachyderme [pakidɛʀm/paʃidɛʀm] "pachyderm"
 * Michel [miʃɛl] "Michael" — Michel-Ange [mikɛlɑ̃ʒ] "Michelangelo", Michaël [mikaɛl] "Michael"

In Italian
In Italian, there is a tendency, when speaking slowly, to sound the 'i' in words where it was originally introduced merely as a spelling device to soften the preceding 'c' or 'g', as for example in the names Luciano, Giovanni and Giuseppe (compare Luke, John and Joseph).

In Portuguese
Normally, Portuguese final unstressed -o is pronounced as a very short /ʊ/, much like the 'u' in English 'put'. But the name of the euro, the Europe-wide currency that in 2002 replaced the Portuguese escudo, is pronounced by most speakers most of the time without this characteristic. Only when speaking rapidly do the Portuguese say 'yuru'. The reason is no doubt awareness that the word is derived from 'Europa', where the 'o' is very much an 'o' and stressed.