3. The phoneme-grapheme correspondences of English, 1: Consonants
© 2015 Greg Brooks, CC BY http://dx.doi.org/10.11647/OBP.0053.03
3.1 The general picture: the regular spellings of English consonant phonemes
This chapter can be summed up by saying that 13 of the 24 consonant phonemes of RP have highly regular spellings (though for two of these, /w, ŋ/, positional constraints have to be stated), while the other 11 have to be analysed according to position in the word.
So, the 11 consonant phonemes /b d g h m n p r t θ ð/ are regularly spelt <b d g h m n p r t th th> respectively; /w/ (which occurs only before vowel phonemes and therefore does not occur word-finally) is regularly spelt <w> initially, <u> medially (but see the note below Table 3.1); and /ŋ/ (which occurs only after short vowel phonemes and therefore does not occur initially) is regularly spelt <n> before /k, g/, however spelt, otherwise <ng>.
The main regularities for the other 11 consonant phonemes are summarised in Table 3.1, by position in the word. For seven phonemes final position has to be subdivided, and final /s, k/ have a further sub-subdivision. The entries for /ʤ, s, k/ blur the distinction between phonemes and graphemes in defining word positions – for more detail on these phonemes’ complicated correspondences, and for the 2-phoneme grapheme <x>, see sections 3.7.1, 3.7.4 and 3.7.6 and Tables 3.3 and 3.4.
Position in word |
||||
Phoneme |
Initial |
Medial |
Final |
|
/∫/ |
sh |
ti |
sh |
|
/f/ |
f |
f |
ff |
|
/v/ |
v |
v |
ve |
|
/j/ within /juː/ |
See Table 5.1 |
|||
/j/ elsewhere |
y |
i * |
(does not occur) |
|
in monosyllables after a short vowel spelt with one letter |
otherwise |
|||
/l/ |
l |
l |
ll |
L |
/ʧ/ |
ch but <t> before /uː/ |
t |
tch |
ch |
/ʒ/ |
(rare) |
si |
(does not occur) |
ge |
/z/ |
z |
s |
zz |
S |
/ʤ/ |
j |
<g> before <e, i, y>, otherwise <j> |
dge |
ge |
/s/ |
s |
s |
ss |
in other monosyllables <ce>; in polysyllables <s> |
/k/ |
c |
c |
ck |
in other monosyllables <k>; in polysyllables <c> |
but <k> before <e, i, y> |
* N.B. Many occurrences of medial /j/ (and some of medial /w/) are actually not represented in the spelling at all – see sections 3.8.7-8.
In sections 3.5, 3.7 and 3.8 I set out the consonantal phoneme-grapheme correspondences of English, under the consonant phonemes listed in the order in which they appear in Table 2.1 in chapter 2.
The consonant phonemes fall into two main categories, those which have a doubled spelling, such as <bb> for /b/ in rabbit, and those which do not. Within those which do have a doubled spelling there is a further important distinction, between those whose doubled spellings are rare in one-syllable words and those whose doubled spellings are regular at the end of one-syllable words after a short vowel phoneme spelt with one letter (Crystal, 2012, especially chapters 7 and 8, explains how this division goes back many centuries). Sections 3.5 and 3.7 cover these two categories of consonant phonemes with doubled spellings, and section 3.8 those which do not have a doubled spelling.
This trichotomy (the Greek etymology of this word officially means ‘cutting into three’, but unofficially could also mean ‘haircut’ – how neat is it that the word meaning ‘cutting into three’ could also mean ‘splitting hairs’?) does not quite accommodate /r/. It does have a doubled spelling (<rr>) and therefore does not belong in section 3.8 (phonemes without a doubled spelling). But /r/ does not occur word-finally in RP, so is not even a candidate for section 3.7 (phonemes whose doubled spellings are regular at the end of one-syllable words after a short vowel phoneme spelt with one letter). Yet /r/ spelt <rr> is not just rare in one-syllable words – it is non-existent – so it might seem not to fit into section 3.5 (phonemes whose doubled spellings are rare in one-syllable words) either. However, section 3.5 is where I have put it, on the grounds that (a) there are some medial examples of /r/ spelt <rr>, e.g. error; (b) many other examples of /r/ spelt <rr> arise from suffixation, e.g. preferring, referral; (c) in these respects /r/ is similar to the other phonemes in section 3.5.
Within each group I list the phonemes in alphabetical order of the letter(s) comprising their basic spellings, except that in section 3.5 /r/ is dealt with after /t/; /r/ is dealt with last because that leads on naturally to the treatment in section 3.6 of a special process involving /r/, namely /r/-linking, hence the interruption in the order of sections.
Under each consonant phoneme I deal with the spellings in this order:
1) The basic grapheme. In my opinion, each of the 24 consonant phonemes of English has a basic grapheme, the one that seems most natural as its spelling. The identification of <si> as the basic grapheme for /ʒ/ may seem curious – but this is the least frequent phoneme in English speech and <si> is its most frequent spelling. As you will see from the percentages at the beginning of each section, the basic grapheme is also, in 20 cases, the most frequent spelling of that phoneme – the exceptions are /z, ʤ, ∫, j/.
2) Other graphemes which are used for the phoneme with reasonable frequency. By reasonable frequency I mean at least 5 per cent of the occurrences of the phoneme in running text.
3) The doubled spelling, if the phoneme has one – 16 of the 24 consonant phonemes do (indeed, a few have more than one). Most doubled consonant spellings consist of the basic single-letter grapheme written twice, but some have a different pattern. Most of the doubled spellings are quite rare in stem words. For some guidance on when to spell a consonant double see chapter 4. None of the doubled spellings of English consonant phonemes ever occur in word-initial position (with the two exceptions noted under /l/ in sections 3.7.5 and 4.1), so word-initial position is not mentioned in the entries about doubled spellings in this chapter (except under /l/).
4) The doubled spelling plus final <e>, if the phoneme has such a spelling.
5) Oddities, graphemes which are used to spell that phoneme only rarely.
6) Any 2-phoneme graphemes in which the phoneme is represented. Almost all the 2-phoneme graphemes are also Oddities, but a few belong to the main system (see section 3.4) and are included there.
7) Any 3-phoneme grapheme in which the phoneme is represented. Both 3-phoneme graphemes are definitely Oddities.
Some entries end with Notes, and a few have Tables.
Under most phonemes I give the frequency of occurrence of each major grapheme as a spelling of the phoneme, using the information in Edward Carney’s massive study A Survey of English Spelling (1994). He gives two frequencies for most phoneme-grapheme correspondences:
- text frequency, that is, the frequency with which the correspondence occurs when you count all the correspondences in a large set of pieces of continuous prose, but discounting derived forms of stem words, e.g. past tenses, and all function words, e.g. of, is, there, where. Because Carney lemmatised his corpus (that is, reduced all the words to stem forms), his text frequencies for doubled consonants are probably systematically underestimated, since large numbers of occurrences of doubled consonant letters arise from suffixation – see sections 4.2 and 4.3.1;
- lexical frequency, that is, the frequency with which the correspondence occurs when you count all and only the correspondences in a dictionary.
Usually the two frequencies are similar, but where a particular correspondence occurs in only a few words but those words are very common, the text frequency will be high and the lexical frequency low (and vice versa where a correspondence occurs in many words but those words are rare). For this chapter and chapter 5 I’ve used only Carney’s text frequencies since those (mainly) represent what readers encounter. However, my lists of examples range far and wide within English vocabulary, and take in words which are so rare that they certainly did not contribute to Carney’s text frequencies. An odd category here is words in which /ɪə/ is spelt <ier> - this category is never mentioned by Carney; presumably no such words turned up in the corpus he compiled and analysed.
I give no frequencies for doubled spellings plus final <e> since these are all rare, and in most cases the frequencies for the Oddities are lumped together.
3.4 The main system and the rest
Under each phoneme I separate the correspondences with graphemes into what I consider to be the main system and the rest (this distinction is very similar to that between major and minor units postulated by Venezky, 1970: 52-55). The correspondences which I include in the main system are those which seem to me to operate as part of larger regularities, even though pretty rarely as absolute rules. For the consonant phonemes the larger regularities comprise the basic correspondences, the correspondences which have reasonable frequency as I’ve defined it above, and the doubled spellings, but not the doubled spellings plus <e>, the 2-and 3-phoneme graphemes (except a few 2-phoneme graphemes which are of reasonably high frequency), or the Oddities. In this chapter (and in chapters 5, 9 and 10) correspondences which have reasonable frequency are shown in 9-point type, the rest in smaller 7.5-point type.
Three quite rare correspondences are, however, included in the main system – /k/ spelt <q>, /ʒ/ spelt <ge>, and /uː/ spelt <ue>. For /k/ spelt <q> this is because <q> would otherwise not appear in the main system at all, but <q> is a grapheme of written English and therefore has to be included; also, the 2-phoneme sequence /kw/ is mainly spelt <qu>. /ʒ/ spelt <ge> is needed to complete the pattern of correspondences in word-final position – see Table 3.1. And although /uː/ spelt <ue> is very rare, I found it necessary to include it in the main system because the mirror-image correspondence (<ue> pronounced /uː/) is one of only two frequent correspondence of <ue> - see section 10.37.
3.5 Consonants with doubled spellings which are rare in one-syllable words: /b d g m n p t/, plus /r/
For the incluson of /r/ in this section see section 3.2.
Despite their rarity in stem words, the doubled spellings of these consonant phonemes arise very frequently from suffixation, e.g. rubbed, budding, begged, skimmed, skinned, hopped, pitted, preferring (see sections 4.2 and 4.3.1).
The main system
Basic grapheme |
<b> |
98% |
e.g. rabid |
Other frequent graphemes |
(none) |
||
Doubled spelling |
<bb> |
<1% |
medially, regular before final /əl/ spelt <-le> after a short vowel spelt with a single letter, e.g. babble – see section 4.3.3; there are also independent medial examples, e.g. abbey, abbot, bobbin, cabbage, dibber, hobbit, hobby, hubbub, rabbi, rabbit, ribbon, rubber, rubbish, Sabbath, shibboleth, stubborn – see sections 4.3.4 and 4.4.5-6; word-finally, only in ebb – see section 4.3.2 |
The rest
Doubled spelling + <e> |
(does not occur) |
|
Oddities |
1% in total |
|
<bh> |
only in abhor and its derivatives abhorred, abhorrent, plus bhaji, bhang(ra), bhindi, Bhutan and a few other rare words from the Indian sub-continent. <b, h> are usually separate graphemes at a morpheme boundary, as in clubhouse, subheading |
|
<bu> |
only in build, buoy, buy. See Notes |
|
<pb> |
only in the compound words cupboard, raspberry, plus Campbell |
|
2-phoneme graphemes |
(none) |
Notes
For the compound words gooseberry /ˈgʊzbriː/), raspberry /ˈrɑːzbriː/), strawberry /ˈstrɔːbriː/) see section 6.10.
I analyse <bu> in build, buoy, buy as a grapheme spelling /b/ because this is more economical than adding /ɪ/ spelt <ui>, /ɔɪ/ spelt <uoy> and /aɪ/ spelt <uy> to the list of graphemes; cf. <gu> under /g/, section 3.5.3, and <cu> under /k/, section 3.7.1.
The main system
Basic grapheme |
<d> |
98% |
e.g. bud |
Other frequent grapheme |
<ed> |
(not counted in percentages) |
|
Doubled spelling |
<dd> |
2% |
medially, regular before final /əl/ spelt<-le> after a short vowel spelt with one letter, e.g. griddle – see section 4.3.3; other medial examples include addictive, additive, adduce, bladder, buddy, cheddar, fodder, judder, ladder, midden, rudder, ruddy, shoddy, sodden, sudden, teddy, toddy, widdershins – see sections 4.3.4 and 4.4.5–6; perhaps also the compound word granddad, but see section 4.4.7; word-finally, only in add, odd, rudd, Sudd – on add, odd see section 4.3.2 |
The rest
Doubled spelling + <e> |
(does not occur) |
|
Oddities |
<1% in total |
|
<bd> |
only in bdellium |
|
<ddh> |
only in Buddha and derivatives, saddhu |
|
<de> |
only in aide, blende, blonde, horde and (for)bade pronounced /(fəˈ)bæd/ (also pronounced /(fəˈ)beɪd/ with <d> alone spelling /d/ and <a.e> spelling /eɪ/). The <e> in blonde marks it French-style as feminine (masculine: blond) |
|
<dh> |
only in a few loanwords and names from the Indian subcontinent, e.g. dhobi, dhoti, dhow, Gandhi, jodhpurs, sandhi, Sindh |
|
2-phoneme graphemes |
(none) |
Note
/d/ is almost always spelt <ed> in past forms of regular verbs ending in a voiced consonant other than /d/ or a vowel, e.g. ebbed, flowed. The only exceptions are laid, paid which would (if they were spelt regularly) be *layed, *payed – cf. delayed, played and sections 5.7.1 and 6.5. See also the entry for <ed> in chapter 10, section 10.15.
The main system
Basic grapheme |
<g> |
92% |
e.g. beg |
Other frequent graphemes |
(none) |
||
Doubled spelling |
<gg> |
2% |
medially, regular before final /əl/ spelt <-le> after a short vowel spelt with one letter, e.g. muggle – see section 4.3.3; other medial examples include aggressive, beggar, dagger, doggerel, haggis, jagged, maggot, nugget, ragged, rugged, rugger, sluggish, trigger – see sections 4.3.4 and 4.4.5-6; word-finally, only in egg – see section 4.3.2 |
The rest
Doubled spelling + <e> |
(does not occur) |
|
Oddities |
2% in total |
|
|
<ckgu> |
only in blackguard /ˈblægəd, ˈblægɑːd/ |
|
<gh> |
word-final only in ugh; otherwise only in afghan, aghast, burgher, ghastly, ghat, ghee, gherkin, ghetto, ghillie (also spelt gillie), ghost, ghoul, ogham, sorghum and a few more rare words |
|
<gu> |
word-initially, only in guarantee, guard, guerrilla, guess, guest, guide, guild, guilder, guile, guillemot, guillotine, guilt, guinea, guise, guitar, guy and a few more rare words; |
medially, only in baguette, beguine, dengue, disguise, languor (the <u> surfaces as /w/ in languid, languish – see section 7.2) and suffixed forms of a few words in next category, e.g. cataloguing; phonemically word-final only in brogue, drogue, fatigue, fugue, intrigue, plague, rogue, vague, vogue and a few more rare words where the final written <e> is part of a split digraph with the vowel letter preceding the <g> – see also next paragraph, and Notes |
||
|
<gue> |
only word-final and only in analogue, catalogue, colleague, decalogue, demagogue, dialogue, eclogue, epilogue, ideologue, league, monologue, morgue, pedagogue, prologue, prorogue, synagogue. In some of the words ending <-ogue> US spelling has <-og>, which is simpler in the stem forms but means that in, e.g., cataloging the first <g> (less regularly) spells /g/ before <i>, a problem which the spelling with <u> avoids. The only word in which final <g, u, e> are all separate graphemes is segue /ˈsegweɪ/ |
2-phoneme graphemes |
For all of these see Notes |
|
|
/gz/ |
4% |
|
(1) spelt <x> |
only in some polysyllabic words of Latin origin, namely anxiety pronounced /æŋˈgzaɪjɪtiː/ (also pronounced /æŋˈzaɪjɪtiː/), auxiliary, exact, exaggerate, exalt, exam(ine), example, exasperate, executive, executor, exemplar, exemplify, exempt, exert, exigency, exiguous, exile, exist, exonerate, exorbitant, exordium, exuberant, exude, exult, plus exotic from Greek and a few more rare words; also in Alexandra, Alexander and becoming frequent in exit pronounced /ˈegzɪt/ (also pronounced /ˈeksɪt/). For anxiety see also under /ŋ/ in section 3.8.2 |
|
(2) spelt <xh> |
only in about 7 polysyllabic words of Latin origin: exhaust(ion), exhibit, exhilarat-e/ion, exhort, exhume – but in some derivatives <xh> spells /eks/, e.g. exhibition, exhortation, exhumation |
|
/gʒ/ spelt <x> |
only in luxuriance, luxuriant, luxuriate, luxurious |
Notes
In blackguard (also spelt blaggard), guarantee, guard the <u> is technically redundant because <ckg, g> would spell (and be pronounced) /g/ without it. But in all the other words with <gu> the <u> has to be there in order to prevent the <g> appearing to spell (and be pronounced) /ʤ/. It’s because guild, guy must be analysed this way that I analyse build, buy (and by extension buoy) as having /b/ spelt <bu> (see section 3.5.1, and cf. <cu> under /k/, section 3.7.1).
The regular 2-grapheme spelling of /gz/ is <gs>, e.g. dogs. The sequence <gz> seems to occur only in zigzag.
The 2-phoneme sequence /gʒ/ seems to occur only in the four words listed above and to have no 2-grapheme spelling.
The 2-phoneme sequence /gw/ is almost always spelt <gu>, e.g. in anguish, distinguish, extinguish, guacamole, guano, guava, iguana, language, languish, linguist, penguin, sanguine, segue, unguent. Exception: wigwam. The converse does not hold – most occurrences of <gu> are pronounced either as /g/ or as 2 phonemes (/g/ plus a vocalic pronunciation of <u>) – see section 9.15.
For <go> in allegory, category see section 6.10.
The main system
Basic grapheme |
<m> |
96% |
e.g. sum |
Other frequent graphemes |
(none) |
||
Doubled spelling |
<mm> |
3% |
medially, does NOT occur before final /əl/ spelt <-le>; medial examples include comma, commune, cummerbund, hammock, hummock, immense, plummet, rummage, slummock, summit and some derived forms, e.g. dia/pro-grammatic, immodest – see sections 4.3.4 and 4.4.5-6; never word-final |
The rest
Doubled spelling + <e> |
<mme> |
now only in oriflamme and (non-computer) programme since gram and its derivatives are no longer spelt *gramme, etc. |
Oddities (all word-final only) |
<1% in total |
|
|
<gm> |
only in apophthegm, diaphragm, epiphragm, paradigm, phlegm, syntagm. /g/ surfaces in some derivatives: paradigmatic, phlegmatic, syntagma(tic) – see section 7.2 |
|
<mb> |
only in dithyramb, lamb; climb, limb; aplomb, bomb, catacomb, comb, coomb, coxcomb, coulomb, hecatomb, rhomb, tomb, womb; crumb, dumb, numb, plumb, rhumb, succumb, thumb and a few more very rare words. /b/ surfaces in some derivatives: dithyrambic, bombard(ier), bombast(ic), |
|
<mbe> |
only in buncombe (‘nonsense’; also spelt bunkum), co(o)mbe (‘short valley’; also spelt coomb); and contrast flambe /ˈflɒmbeɪ/, where |
|
<me> |
never initial; mainly word-final and there only in become, come, some, welcome and the adjectival suffix /səm/ spelt <-some>, e.g. handsome (contrast hansom); medially only in camera, emerald, omelette, ramekin pronounced /ˈræmkɪn/ (also pronounced /ˈræmɪkɪn/) – see section 6.10 – and Thames |
|
<mn> |
only in autumn, column, condemn, contemn, damn, hymn, limn, solemn. /n/ surfaces in some derivatives: autumnal, columnar, columnist, condemnation, contemner, damnable, damnation, hymnal, solemnity – see section 7.2 |
|
<nd> |
only in sandwich pronounced /ˈsæmwɪʤ/ (also has a ‘regular’ spelling pronunciation /ˈsændwɪʧ/) |
2-phoneme grapheme |
/əm/ spelt <m> |
only word-final, e.g. chasm, enthusiasm, orgasm, phantasm, pleonasm, sarcasm, spasm, several words ending in –plasm (e.g. ectoplasm), chrism, prism, schism and all the many derived forms ending in –ism, macrocosm, microcosm, abysm, aneurysm (also spelt aneurism), cataclysm, paroxysm, algorithm, rhythm and a few other very rare words; also in film pronounced /ˈfɪləm/ in some Irish accents. See Note |
Note
In all but the last three of the words just listed with word-final /əm/ spelt <-m> the preceding phoneme is /z/ spelt <s>, so the regular spelling of word-final /zəm/ is <-sm> (only exception: bosom). This is one of only a handful of cases where the spelling of a final syllable is more predictable as a whole than from its separate phonemes, which here would predict (for example) *chasam, *prisom, etc. However, word-final /əm/ with other preceding phonemes has various 2-grapheme spellings in, e.g., alyssum, balsam, besom, fathom (but contrast the 1-grapheme spelling in algorithm, rhythm), gypsum, hansom, lissom, opossum, ransom, transom and all the adjectives ending <-some>.
The main system
Basic grapheme |
<n> |
97% |
e.g. tin |
Other frequent graphemes |
(none) |
||
Doubled spelling |
<nn> |
<1% |
medially, does NOT occur before final /əl/ spelt <-le>; medial examples include anneal, annual, annul, biennale, binnacle, Britannic, cannibal, chardonnay, cinnabar, cinnamon, ennui, innocent, punnet, tannic, tinnitus, tintinnabulation, zinnia – see sections 4.3.4 and 4.4.5-6; word-finally, only in Ann, djinn, Finn, inn – on Ann, inn see section 4.3.2 |
The rest
Doubled spelling + <e> |
<nne> |
only word-final and only in Anne, cayenne, comedienne, cretonne, doyenne, tonne and a few other rare words |
Oddities |
3% in total |
|
|
<dne> |
only in Wednesday |
|
<gn> |
word-initially, only in gnarl, gnash, gnat, gnaw, gneiss, gnome, gnosis, Gnostic and gnu analysed as /n/ spelt <gn> plus /juː/ spelt <u>; medially, only in cognisance (also pronounced with /gn/), physiognomy, recognise pronounced /ˈrekənaɪz/ (usually pronounced /ˈrekəgnaɪz/); word-finally, only in align, arraign, assign, benign, campaign, coign, condign, consign, deign, design, ensign, feign, foreign, impugn and a few other very rare words in –pugn, malign, reign, resign, sign, sovereign, thegn; also phonemically word-final in champagne, cologne where the final written <e> is part of a split digraph with the letter before the <g> spelling a diphthong. /g/ surfaces in some derivatives: agnostic, diagnosis, prognosis, malignant, pugnacious, repugnant, assignation, designation, resignation, signal, signature – see section 7.2 |
|
<gne> |
only word-final and only in cockaigne, epergne, frankalmoigne /kəˈkeɪn, ɪˈpɜːn, ˈfræŋkælmɔɪn/ |
<kn> |
1% never word-final; medially, only in acknowledge, knick-knack; otherwise only word-initial and only in knack(er(s)), knap, knave, knead, knee, knell, knew, knick(er(s)), knickerbocker, knick-knack, knife, knight, knit, knob, knobbly, knock, knoll, knot, know(ledge), knuckle and a few more very rare words |
|
|
<mn> |
only word-initial and only in mnemonic, mnemonist. /m/ surfaces in amnesia, amnesty – see section 7.2 |
|
<nd> |
only in grandfather, Grandma (hence the frequent misspelling *Granma – cf. section 4.4.7 on |
|
<ne> |
non-finally, only in vineyard (and even there it’s stem-final), vulnerable pronounced /ˈvʌlnrəbəl/ - see also Notes and section 6.10 (I refuse to analyse the alternative pronunciation /ˈvʌnrəbəl/ with loss of the first /l/ because it would add an otherwise not-needed grapheme <lne> to the inventory); otherwise only word-final and only in about 35 words, namely borne, bourne, bowline, Catherine, clandestine pronounced /klænˈdestɪn/ (also pronounced /ˈklændəstaɪn/), cocaine, compline, crinoline, demesne, (pre)destine, determine, discipline, engine, ermine, examine, famine, feminine, genuine, gone, groyne, heroine, hurricane pronounced /ˈhʌrɪkən/ (also pronounced /ˈhʌrɪkeɪn/), illumine, intestine, jasmine, marline, masculine, medicine, migraine, moraine, none, peregrine, ptomaine, saccharine, sanguine, scone pronounced /skɒn/ (also pronounced /skəʊn/), shone, urine, vaseline, wolverine. In all these words the <e> is phonographically redundant, in that its removal would not affect the pronunciation. However, without <e> done, none would become don and the prefix non- (and changing their spellings to dun, nun would cause other confusions). Also, the <e> keeps borne, heroine visually distinct from born, heroin |
<ng> |
only in length, lengthen, strength, strengthen pronounced /lenθ, ˈlenθən, strenθ, ˈstrenθən/. See also under /k, ŋ/, sections 3.7.1, 3.8.2 |
|
|
<nt> |
only in croissant, denouement, rapprochement |
|
<nw> |
only in gunwale |
|
<pn> |
only word-initial and only in words derived from Greek πνευ̑μα pneuma (‘breath’) or πνεύμων pneumon (‘lung’), e.g. pneumatic, pneumonia |
2-phoneme graphemes |
/ən/ |
only in Haydn (I mention him in memory of Chris Upward of the Simplified Spelling Society) and most contractions of not with auxiliary verbs, i.e. isn’t, wasn’t, haven’t, hasn’t, hadn’t, doesn’t, didn’t, mayn’t, mightn’t, mustn’t, couldn’t, shouldn’t, wouldn’t, oughtn’t, usedn’t, some of which are rare to the point of disuse, plus durstn’t, which is regional/comic; in all of these except mayn’t the preceding phoneme is a consonant. Other contractions of not with auxiliary verbs (ain’t, aren’t, can’t, daren’t, don’t, shan’t, weren’t, won’t), i.e. all those with a preceding vowel phoneme (except mayn’t) are monosyllabic (though some Scots say /ˈdeərənt/ with a preceding consonant and linking /r/ and therefore two syllables). Curiously, innit, being a contraction of isn’t it, reduces isn’t to a single syllable. See Notes |
|
/nj/ |
see under /j/, section 3.8.8 |
Notes
/ən/ has several 2-grapheme spellings, e.g. in cotton, ruffian, written.
For <ne> in confectionery, generative, stationery, vulnerable see section 6.10.
The main system
Basic grapheme |
<p> |
95% |
e.g. apt |
Other frequent graphemes |
(none) |
||
Doubled spelling |
<pp> |
5% |
medially, regular before final /əl/ spelt <-le> after a short vowel spelt with one letter, e.g. apple: other medial examples |
include apply, apprehend, cappuccino, dapper, frippery, hippodrome, hippopotamus, guppy, opponent, oppose, opposite, scupper, supper, supply, support – see sections 4.3.4 and 4.4.5-6; word-finally, only in Lapp |
The rest
Doubled spelling + <e> |
<ppe> |
only in grippe, steppe |
Oddities |
<1% in total |
|
|
<b> |
only in presbyterian pronounced /prespɪˈtɪəriːjən/ (also pronounced /prezbɪˈtɪəriːjən/) |
|
<bp> |
only in subpoena /səˈpiːnə/ |
|
<gh> |
only in misspelling of hiccup as *hiccough |
|
<pe> |
only in canteloupe, troupe, plus opera in rapid speech – for <pe> in opera see section 6.10 |
|
<ph> |
only in diphtheria, diphthong, naphtha, ophthalmic and shepherd. The first four also have pronunciations with /f/ – e.g. /ˈdɪfθɒŋ/ versus /ˈdɪpθɒŋ/ |
2-phoneme graphemes |
(none) |
The main system
Basic grapheme |
<t> |
96% |
e.g. rat |
Other frequent grapheme |
<ed> |
(not counted in percentages) See Notes |
|
Doubled spelling |
<tt> |
3% |
medially, regular before final /əl/ spelt <-le> after a short vowel spelt with one letter, e.g. rattle; other medial examples include attention, attract, attribute, |
battalion, battery, butter, button, buttress, chitterlings, falsetto, glutton, jitter(s), mattress, rattan, smattering, tattoo, tittup – see sections 4.3.4 and 4.4.5-6; word-finally, only in bott, boycott, butt, matt, mitt, mutt, nett, putt, watt. See also Notes |
The rest
Doubled spelling + <e> |
<tte> |
only word-final and only in about 23 stem words, namely baguette, brunette, cassette, coquette, corvette, croquette, epaulette, etiquette, garrotte, gavotte, gazette, maisonette, omelette, oubliette, palette, pipette, pirouette, roulette, serviette, silhouette, toilette, vignette, vinaigrette, and a few derived forms, e.g. cigarette, launderette, rosette, statuette, suffragette, and some other rare words. In latte <tt, e> represent separate phonemes, as do <u.e, tt> in butte |
Oddities |
1% in total |
|
|
<bt> |
only in debt, doubt, subtle. /b/ surfaces in debit, indubitable, subtility – see section 7.2 |
|
<ct> |
only in Connecticut, indict, victualler, victuals. /k/ surfaces in indiction – see section 7.2 |
|
<dt> |
only in veldt |
|
<phth> |
only in phthisic, phthisis pronounced /ˈtaɪsɪk, ˈtaɪsɪs/ |
|
<pt> |
only in Deptford, ptarmigan, pterodactyl (Greek, = ‘wing finger’), pterosaur (Greek, = ‘wing lizard’), Ptolem-y/aic, ptomaine, receipt and a few more very rare words. /p/ surfaces in archaeopteryx, helicopter (Greek, = ‘ancient wing, spiral wing’), reception, receptive – see section 7.2 |
|
<te> |
mainly word-final and in that position in at least 120 words, namely - ate pronounced /et/ (also pronounced /eɪt/, which requires a different analysis: /t/ spelt <t> and /eɪ/ spelt <a.e>), Bacchante, composite, compote, confidante, debutante, definite, detente, dirigiste, enceinte, entente, entracte, exquisite, favourite, granite, hypocrite, infinite, minute |
(‘sixtieth of an hour’), opposite, perquisite, plebiscite, pointe, requisite, riposte, route, svelte - about 30 nouns/adjectives in /ət/ spelt <-ate> where the verbs with the same spelling are pronounced with /eɪt/, e.g. advocate, affiliate, aggregate, alternate (here with also a difference in stress and vowel pattern: noun/ adjective pronounced /ɔːlˈtɜːnət/, verb pronounced /ˈɔːltəneɪt/), animate, appropriate, approximate, articulate, associate, certificate, consummate (here with also a difference in stress and vowel pattern: adjective pronounced /kənˈsʌmət/, verb pronounced /ˈkɒnsjəmeɪt/), coordinate, curate (here with also a difference in meaning and stress: noun (‘junior cleric’) pronounced /ˈkjʊərət/, verb (‘mount an exhibition’) pronounced /kjʊəˈreɪt/), degenerate, delegate, deliberate (here with also a difference in syllable structure: adjective /dɪˈlɪbrət/ with three syllables and an elided vowel – see section 6.10; verb /dɪˈlɪbəreɪt/ with four syllables), designate, desolate, duplicate, elaborate, estimate, expatriate, graduate, initiate, intimate, legitimate, moderate, pontificate (here with unrelated (?) meanings: noun (‘pope’s reign’) pronounced /pɒnˈtɪfɪkət/, verb (‘speak pompously’) pronounced /pɒnˈtɪfɪkeɪt/), precipitate (but here only the adjective has /ət/; the noun as well as the verb has /eɪt/), predicate, separate (here too with a difference in syllable structure: adjective /ˈseprət/ with two syllables and an elided vowel – see section 6.10; verb /ˈsepəreɪt/ with three syllables), subordinate, syndicate, triplicate. In the verbs and the many other nouns and adjectives with this ending pronounced /eɪt/, the <e> is part of the split digraph <a.e> spelling /eɪ/ and the /t/ is spelt solely by the <t> - a further set of at least 60 nouns/adjectives (some of which are derived forms) in /ət/ spelt <-ate> with no identically-spelt verb, e.g. accurate, adequate, agate, appellate, celibate, chocolate, climate, collegiate, conglomerate, |
||
inanimate, in(sub)ordinate, inspectorate, - possibly just one word where both noun and verb have <-ate> pronounced /ət/: pirate - <te> spelling /t/ also occurs medially in a few words in rapid speech, e.g. interest, literacy, literal, literary, literature, sweetener, veterinary – see section 6.10 In all cases where /ət/ is spelt <-ate> the <e> is phonographically redundant (that is, it does not indicate a ‘long’ pronunciation of the preceding vowel letter and could therefore be omitted from the spelling without altering the pronunciation; hence I have not analysed such words as having /ə/ spelt <a.e> and /t/ spelt <t>), but in two cases it makes the words visually distinct from words without the <e> and with an unrelated meaning: point, rout. Carney does not recognise <te> as a spelling of /t/ and this probably means that percentages for my analysis would be slightly different from his |
||
|
<th> |
only in Thai, thali, Thame, Thames, Therese, Thomas, thyme, Wrotham /ˈruːtəm/ |
|
<tw> |
only in two and derivatives, e.g. twopence, twopenny. /w/ surfaces in between, betwixt, twain, twelfth, twelve, twenty, twice, twilight, twilit, twin – see section 7.2 |
2-phoneme graphemes |
/tθ/ |
only in eighth. See section 4.4.7 |
|
/ts/ |
|
|
(1) |
only in Alzheimer’s, bilharzia, Nazi (but Churchill said /ˈnɑːziː/), scherzo, schizo(-) |
|
(2) |
only in intermezzo, paparazzi, pizza, pizzicato |
Notes
/t/ is always spelt <ed> in past forms of regular verbs ending in a voiceless consonant other than /t/, e.g. walked. See also the entry for <ed>, section 10.15.
/ts/ also has 2-grapheme spellings, the regular one being <ts>, plus the Oddity <tz> – for the latter see under /s/, section 3.7.6.
Occurs only before a vowel phoneme (in RP).
The main system
Basic grapheme |
<r> |
94% |
e.g. very |
Other frequent graphemes |
(none) |
||
Doubled spelling |
<rr> |
4% |
medially, does NOT occur before final /əl/ spelt <-le> and arises mainly from suffixation (see Notes), but there are some independent examples, e.g. arroyo, barrow, berry, borrow, burrow, carrot, derrick, garrotte, guerrilla, herring, horrid, hurry, lorry, mirror, (to)morrow, parrot, porridge, scurrilous, serrate, sorry, squirrel, stirrup, terrine, warrant, wherry, worrit, worry – see sections 4.3.4 and 4.4.5-6; never word-final as a separate grapheme – see Notes |
The rest
Doubled spelling + <e> |
(does not occur as a spelling of /r/ – but see Notes) |
|
Oddities |
2% in total |
|
|
<re> |
only in forehead pronounced /ˈfɒrɪd/ |
|
<rh> |
only word-initial and only in a few words mainly of Greek origin, namely rhapsody and several other words beginning rhapsod-, rhea, rheme, rhesus, rhetor(ic), rheum(ati-c/sm), rhinestone, rhinoceros and several other words beginning rhin(o)-, rhizome and several other words beginning rhizo-, rhododendron and several other words beginning rhodo-, rhodium, rhomb-ic/us, the Greek letter name rho, rhotic, rhubarb, rhyme, rhythm and a few other rare words |
<rrh> |
only medial and only in a few words of Greek origin, namely amenorrhoea, arrhythmia, cirrhosis, diarrhoea, gonorrhoea, haemorrhage, haemorrhoid, lactorrhoea, logorrhoea, pyorrhoea, pyrrhic. N.B. In catarrh, myrrh the <rrh> is not a separate grapheme – see Notes (but in catarrhal /r/-linking occurs – see section 3.6) |
|
|
<wr> |
except in awry, only in initial position and only in wrap, wrasse, wreck, wren, wrench, wrest(le), wretch(ed), wriggle, wring, wrinkle, wrist, write, wrong, Wrotham /ˈruːtəm/, wrought, wry and a few more rare words |
2-phoneme graphemes |
(none) |
Notes
The only stem words in which final <-rr, -rre, -rrh> occur are carr, charr, parr, err, chirr, shirr, skirr, whirr, burr, purr; barre, bizarre, parterre; catarrh, myrrh. Because there is no /r/ phoneme in these words (in RP), these letters do not form separate graphemes but are part of the trigraphs or four-letter graphemes <arr, err, irr, urr, arre, erre, arrh, yrrh> spelling variously /ɑː, ɜː, eə/ – see the entries for those phonemes in sections 5.5.1, 5.5.2 and 5.6.3 and, for some suffixed forms, the next section. For err see also section 4.3.2.
In words like preferring, referral, the <rr> is due purely to a spelling rule involving the suffix – see the next section and section 4.2. In such words the letters <err> spell the vowel /ɜː/ and the <rr> also spells the linking /r/ consonant – for /r/-linking see section 3.6, and for dual-functioning section 7.1. But in berry, errant, guerrilla, herring, wherry, abhorrent, demurral, garrotte, <e, o, u, a> spell /e, ɒ, ʌ, ə/ and the <rr> simply spells /r/ without influencing the pronunciation of the vowel; similarly in the other words listed above as having independently-occurring medial <rr>.
Although word-final /r/ does not occur in RP when words are pronounced in isolation, words which end in letter <r> after a vowel letter retain the possibility of a /r/ phoneme surfacing when a suffix or the next word begins with a vowel phoneme. For example, I pronounce the phrase dearer and dearer with three /r/ sounds, corresponding to the first three occurrences of the letter <r>: /ˈdɪərərənˈdɪərə/. For more phonological detail see Cruttenden (2014: 224, 315-7).
Many people call this phenomenon ‘<r>-linking’, using the name of the letter <r>. I prefer to call it ‘/r/-linking’, using (in speech) the sound of, or (in writing) the symbol for, the phoneme /r/ because that is what the link consists of in speech. Moreover, various other graphemes which can spell /r/ allow /r/-linking – see, for example, <rrh> in catarrhal in the entry for /r/ just above. /r/-linking is one of four special processes which I have identified as operating in English spelling (for the others see section 6.10 and chapter 7).
In Table 3.2 I have assembled all the examples of /r/-linking mentioned in this book.
In some cases the pre-linking ‘phoneme’ is actually a 2-phoneme sequence.
In a few categories where, before linking, the last phoneme of the stem is /ə/ spelt <er, or>, /ə/ is deleted in speech and <e, o> in writing, and the <r> is left to spell /r/. This process needs to be distinguished from vowel elision (see section 6.10), where a vowel letter is written even though there is no vowel phoneme at that point in the spoken word.
Where <e>-deletion (see section 6.4) occurs, I analyse the phoneme before the linking /r/ (provided it has not been deleted or elided) as spelt by the pre-linking grapheme minus <e>, even when that phoneme has changed.
Except where stated:
1) stress placement and the phoneme before the linking /r/ remain unchanged;
2) the /r/-linking grapheme continues to function as part of the spelling of the preceding phoneme (dual-functioning – see section 7.1), even when that phoneme has changed and/or <e>-deletion has occurred. This principle is adopted in order to avoid introducing some correspondences for which there is no other warrant in my analysis, e.g. <a> alone spelling /eə/ in vicarious. For more detail see section A.8 in Appendix A.
Phoneme before /r/-linking |
Grapheme spelling that phoneme |
/r/-linking grapheme |
Example(s) |
Notes |
/ə/ |
<ar> |
<r> |
polarise |
|
familiarity, hilarity, peculiarity, polarity, vulgarity |
Stress shifts to last syllable of stem, the vowel there shifts to /æ/ and is spelt only by <a>, and <r> spells only /r/ |
|||
vicarious |
Stress shifts to last syllable of stem, and the vowel there shifts to /eə/ |
|||
<er> |
ethereal, managerial |
Stress shifts to last syllable of stem, and the vowel there shifts to /ɪə/ |
||
hyperintelligent, interagency, leverage, offering, sufferance |
/ə/ may be elided – see section 6.10 |
|||
<eur> |
amateurish |
|||
<er> |
foundress, hindrance, laundress, ogress, temptress, tigress, waitress, wardress, wintry |
/ə/ is deleted, as shown by the disappearance of the penultimate <e> of the stem, and <r> spells only /r/ |
||
<or> |
actress, ambassadress, conductress, dominatrix, executrix |
/ə/ is deleted, as shown by the disappearance of the penultimate <o> of the stem, and <r> spells only /r/ |
||
for instance, prioress, terrorist |
||||
authority |
Stress shifts to last syllable of stem, the vowel there shifts to /ɒ/ and is spelt only by <o>, and <r> spells only /r/ |
|||
/ə/ |
<or> |
<r> |
authorial, dictatorial |
Stress shifts to last syllable of stem, and the vowel there shifts to /ɔː/ |
<our> |
favourite |
/ə/ may be elided – see section 6.10 |
||
<r>, plus deletion of <u> from final syllable of stem |
glamorise, rigorous, vigorous |
|||
<re> |
<r> following <e>-deletion |
central, fibrous, lustr-al/ous, metrical, spectral |
/ə/ is deleted (as shown by disappearance of <e>), and <r> spells only /r/ |
|
mediocrity, sepulchral, theatrical |
/ə/ is deleted (as shown by disappearance of <e>), stress shifts to syllable before suffix if not already there, vowel there shifts to /ɒ, ʌ, æ/, and <r> spells only /r/ |
|||
calibration |
/ə/ is deleted (as shown by disappearance of <e>), stress shifts to first syllable of suffix, and <r> spells only /r/ |
|||
acreage, massacreing, ochreous, ogreish /ˈeɪkərɪʤ, ˈmæsəkərɪŋ, ˈəʊkərəs, ˈəʊgərɪʃ/ |
/ə/ is not deleted (as shown by retention of <e>), and <r> spells only /r/, but the schwa and /r/ seem to be spelt by the <e> and <r> in reverse order |
|||
<ure> |
injurious |
Stress shifts to 2nd syll-able of stem, and vowel there shifts to /jʊə/ |
||
adventurous, natural, naturist, procedural, treasury |
/ə/ is spelt only by the <u> and may be elided, especially in derived adverbs – see section 6.10 – and <r> spells only /r/ |
Phoneme before /r/-linking |
Grapheme spelling that phoneme |
/r/-linking grapheme |
Example(s) |
Notes |
/ə/ |
<ur> |
<r> |
murmuring |
|
sulphuric |
Stress shifts to last syllable of stem, and the vowel there shifts to /jʊə/ |
|||
/ɜː/ |
<er> |
conference, deference, preference |
Stress shifts to first syllable, and last vowel phoneme of stem shifts to /ə/ (or may be elided – see section 6.10) |
|
<err> |
<rr> |
errant |
Preceding vowel shifts to /e/, and <rr> spells only /r/ |
|
<irr> |
whirring |
|||
<urr> |
purring |
|||
<er> |
<rr> arising from consonant letter doubling (see section 4.2) |
conferring, deferring, preferring, referral |
||
<ur> |
furry, occurring |
|||
demurral |
Preceding vowel shifts to /ʌ/, and <rr> spells only /r/ |
|||
/ɑː/ |
<ar> |
sparring |
||
<arre> |
<rr> following <e>-deletion |
bizarrery |
||
<arrh> |
<rrh> |
catarrhal |
||
<ar> |
<r> |
cigarette, czarina |
Stress shifts to suffix, and in cigarette vowel phoneme preceding /r/ shifts to /ə/ |
|
/wɑː/ |
<oir> |
memoirist |
||
/eə/ |
<heir> |
inherit |
Too complicated to analyse |
|
<air> |
repairing |
|||
<aire, heir> |
millionairess, heiress |
Stress shifts to final syllable |
||
<ayor> |
mayoral, mayoress |
In mayoress, stress shifts to final syllable |
||
<ear> |
wearing |
|||
/eə/ |
<ere> |
<re> |
thereupon |
|
<r> following <e>-deletion |
wherever, compering |
|||
<are> |
staring |
|||
preparedness |
||||
/aɪə/ |
<ire> |
<r> following <e>-deletion |
entirety |
2nd <e> surfaces as /ɪ/ – see section 7.2 |
wiring |
||||
inspiration |
Stress shifts to first syllable of suffix, vowel in last syllable of stem shifts to /ɪ/ or /ə/ and is spelt only by the <i>, and <r> spells only /r/ |
|||
satirical |
Stress shifts to last syllable of stem, the vowel there shifts to /ɪ/ and is spelt only by the <i>, and <r> spells only /r/ |
|||
<yre> |
lyrical |
Vowel in stem shifts to /ɪ/ and is spelt only by the <i>, and <r> spells only /r/ |
||
pyromaniac |
||||
/ʊə/ |
<ure> |
enduring, surety |
||
<oor, our> |
<r> |
boorish, touring |
||
/ɔː/ |
<ar> |
<rr> arising from consonant letter doubling (see section 4.2) |
warring |
|
<or> |
abhorrent |
Preceding vowel shifts to /ɒ/ and is spelt only by <o>, and <rr> spells only /r/ |
||
<or, oar, oor, our> |
<r> |
mentoring, hoary, flooring, pouring |
||
/ɔː/ |
<ore> |
<r> following <e>-deletion |
boring |
|
/ɪə/ |
<ere> |
interfering |
||
sincerity |
Preceding vowel shifts to /e/ and is spelt only by <e>, and <r> spells only /r/ |
|||
<ear, eer, ier> |
<r> |
dearer, hearing, cheering, tiering |
||
/aʊə/ |
<our, ower> |
devouring, towering |
An even fuller analysis would also mention cases of /r/-linking occurring where an intervening consonant phoneme has been dropped, as in the place- and surname Wareham /ˈweərəm/ (where the /h/ of the Anglo-Saxon placename element ham was dropped many centuries ago) and the British Tommy’s adage about medals: ‘Win ’em and wear ’em’ – here the end of the sentence is also pronounced /ˈweərəm/, the /ð/ phoneme of RP /ˈweə ðəm/ (with no /r/) having been elided. But this book is not about placenames, surnames or accents other than RP.
Sometimes /r/-linking is overgeneralised to words which do not have a letter <r> in the written form (and never had, and still do not have, a /r/ phoneme in any accent of English when pronounced in isolation): the best-known example is law and order pronounced /ˈlɔːrəˈnɔːdə/ (‘Laura Norder’) with ‘intrusive /r/’, rather than /ˈlɔːwəˈnɔːdə/. (But this phrase never seems to be pronounced /ˈlɔːrəˈndɔːdə/ (‘Lauren Dawder’), with the <d> of and made explicit.) An example that occurs in children’s speech is drawing pronounced /ˈdrɔːrɪŋ/ rather than /ˈdrɔːɪŋ/. Cruttenden (2014: 316) provides several more examples.
On the other hand, /r/-linking is sometimes avoided where the spelling suggests it would be natural. For example, the recorded announcers at Sheffield railway station say /ˈplætfɔːm fɔː ˈeɪ, ˈmænʧɪstə ˈeəpɔːt, ˈmænʧɪstə ˈɒksfəd ˈrəʊd, ˈʃaɪə əʊks/ rather than /ˈplætfɔːm fɔːˈreɪ, ˈmænʧɪstəˈreəpɔːt, ˈmænʧɪstəˈrɒksfəd ˈrəʊd, ˈʃaɪərəʊks/ for ‘Platform 4A’, ‘Manchester Airport’, ‘Manchester Oxford Road’, ‘Shireoaks’.
Almost all instances of /r/-linking are also examples of what I call dual-functioning. That is, after linking, the <r>, etc., continues to function as part of the grapheme spelling the pre-linking phoneme while also spelling /r/ in its own right. Exceptions shown in Table 3.2 where an <r> ceases to function as part of the grapheme spelling the pre-suffixation phoneme and therefore only spells /r/ after suffixation are: familiarity, hilarity, peculiarity, polarity, vulgarity, foundress, laundress, ogress, temptress, tigress, waitress, wardress, actress, ambassadress, conductress, dominatrix, executrix, protrectress, authority, mediocrity, sepulchral, theatrical, central, fibrous, lustr-al/ous, metrical, spectral, calibration, demurral, inspiration, satirical, lyrical, abhorrent, sincerity.
For other categories of dual-functioning see section 7.1.
For cases in which /ə/ may be elided after /r/-linking see also section.
For ‘linking /w/’ and ‘linking /j/’ see sections 3.8.7-8. Like /r/-linking, both occur frequently between a stem word and a suffix or a following word beginning with a vowel phoneme. However, there are two key differences: (1) in /w/- and /j/-linking, the quality of the glide between stem and suffix or next word is entirely predictable from the stem-final phoneme, whereas /r/-linking never is (in RP), and can be explained only historically – it occurs where once there was a postvocalic /r/; (2) similar /w/- and /j/-glides occur within many stem words where there is no indication of them in the spelling – /r/-linking never occurs within stem words.
3.7 Consonants with doubled spellings which are regular at the end of one-syllable words after a short vowel spelt with one letter: /k ʧ f ʤ l s v z/
In addition to their frequency in stem words, the doubled spellings of /k f l s/ occasionally arise from suffixation, e.g. picnicking, iffy, modelling, gassing (see sections 4.2 and 4.3.1).
The main system
For all these categories see Notes and Table 3.3.
Basic grapheme |
<c> |
59% |
e.g. cat Regular in all positions except (1) before <e, i, y>, where the regular spelling is <k> (2) before final /əl/ spelt <-le> after a short vowel spelt with one letter, where the regular spelling is <ck> (3) word-finally in one-syllable words, where the regular spelling is <ck> after a short vowel spelt with one letter, otherwise <k> For other exceptions see below |
Other frequent grapheme |
<k> |
21% |
regular before <e, i, y>, e.g. kelp, kit, sky, including word-finally within split digraphs, e.g. like, make; also word-finally in one-syllable words except those where <ck> is regular. Only exceptions: ache, Celt, Celtic, sceptic and one pronunciation of words beginning encephal-; arc, chic, disc, and a few more words |
Doubled spelling |
<ck> |
6% |
regular in word-final position in one-syllable words after a short vowel spelt with one letter, e.g. crack; also before final /əl/ spelt <-le> after a short vowel spelt with one letter, e.g. heckle – see section 4.3.3; for other occurrences medially in stem words see Table 3.3; there are several word-final occurrences in polysyllables, e.g. derrick, dunnock, haddock, hammock, hummock, slummock |
Frequent 2-phoneme grapheme |
/ks/ spelt <x> |
5% |
word-initially, only in the Greek letter- name xi pronounced /ksaɪ/; regular medially, e.g. buxom, maxim, next (for exceptions see below); also finally where the /s/ is part of the stem, e.g. box (only exception: aurochs) |
Rare grapheme |
<q> |
3% |
e.g. quick See <cq, cqu, qu, que> within the Oddities, below, and Notes |
The rest
Doubled spelling + <e> |
(does not occur) |
|
Oddities |
6% in total |
|
|
<cc> spells /k/ |
- before <e, i, y>: only in baccy, biccy, recce /ˈrekiː/ (short for reconnoitre), soccer, speccy, streptococci - where the next letter is not <e, i, y>: in about 45 words mainly of Latin origin, namely acclaim, acclimatise, accolade, accommodate, accompany, accomplice, accomplish, accord, accost, account, |
accoutrement, accredit, accrete, accrue, acculturate, accumulate, accurate, accursed, accuse, accustom, desiccate, occasion, occlude, occult, occupy, occur, succour, succubus, succulent, succumb. Words of non-Latin origin in this group are broccoli, buccaneer, ecclesiastic, felucca, hiccup, mecca, moccasin, peccadillo, peccary, piccolo, raccoon, scirocco, staccato, stucco, tobacco, toccata, Wicca, yucca See Notes for the complementary value of <cc> before <e, i, y>, and on why <cc> is not the doubled spelling of /k/ |
||
|
<cch> |
only in bacchanal, Bacchante, bacchic, ecchymosis, gnocchi, saccharide, saccharine, zucchini. In bacchanal, Bacchante, ecchymosis, saccharide, saccharine, the <h> could be deleted without altering the English pronunciation – see just above; but in bacchic, gnocchi, zucchini this change might make them look as if they were pronounced with /ks/ |
|
<ch> |
mainly in words of Greek origin, e.g. amphibrach, anarchy, anchor, archaic and every other word beginning /ɑːk/ (except arc, ark), brachial, brachycephalic, bronchi(-al/tis), catechis-e/m, |
synchronise, synecdoche, technical, technique, trachea, triptych, trochee. Words of non-Greek origin in this group are ache, aurochs, baldachin, chianti, chiaroscuro, cromlech, Czech, lachrymose, masochist, Michael, mocha, oche, pinochle, pulchritude, scherzo, schooner, sepulchre; also broch, loch, pibroch, Sassenach when pronounced with /k/ rather than Scots /x/ (for this symbol see section 2.3) |
||
|
<cq> |
only in acquaint, acquiesce, acquire, acquisitive, acquit |
|
<cqu> |
spells only /k/ (not /kw/) only in lacquer, picquet, racquet |
|
<cu> |
only in biscuit, circuit (contrast ‘circuitous’ where the <u> ‘surfaces’ – see section 7.2.2); cf. <bu> under /b/, section 3.5.1, and <gu> under /g/, section 3.5.3 |
|
<g> |
only in length, lengthen, strength, strengthen pronounced /leŋkθ, ˈleŋkθən, streŋkθ, ˈstreŋkθən/ (for their alternative pronunciations see under /n/, section 3.5.5) - for the rationale of this analysis see Notes under /ŋ/, section 3.8.2 – and in angst /æŋkst/, disguise /dɪsˈkaɪz/, disgust pronounced /dɪsˈkʌst/, i.e. identically to discussed; disguise, disgust are also pronounced /dɪzˈgaɪz, dɪzˈgʌst/, i.e. with both medial consonants voiced rather than voiceless |
|
<gh> |
only in hough |
|
<ke> |
only in Berkeley, burke |
|
<kh> |
only in astrakhan, gurkha, gymkhana, khaki, khan, khazi, khedive, sheikh, Sikh |
|
<kk> |
only in chukker, dekko, pukka and inflected forms of trek, e.g. trekkie |
|
<qu> |
as a digraph spelling only /k/ (not /kw/) occurs initially or medially (never finally – cf. next paragraph) in about 50 words mainly of French origin, namely bouquet, conquer (/w/ surfaces in conquest – see section 7.2), coquette, croquet, croquette, etiquette, exchequer, liqueur, liquor, liquorice, maquis, mannequin, marquee, marquetry, masquerade, mosquito, parquet, piquant, quatrefoil, quay, quenelle, quiche, so(u)briquet, tourniquet, and, in conservative RP-speakers’ accents, questionnaire, quoits; also medially in applique, communique, |
manque, risque – see next paragraph; also phonemically but not orthographically word-final in opaque; claque, plaque; basque, casque, masque; antique, bezique, boutique, clique, critique, mystique, oblique, physique, pique, technique, unique; bisque, odalisque; toque; peruque; brusque pronounced /bruːsk/, and a few more rare words where the final written <e> is part of a split digraph with a preceding vowel letter spelling variously /eɪ, ɑː, iː, əʊ, uː/. The words basque, casque, masque, bisque, odalisque and brusque pronounced /bruːsk/, where there is also an <s> before the <qu>, cause a special extension to the definition of a split digraph – see section A.6 in Appendix A and the Notes under <a.e, i.e, u.e>, sections 10.4/24/38 |
||
|
<que> as a trigraph spelling only /k/ (not /kw/ plus vowel) |
occurs word-initially only in queue and medially only in milquetoast (where it is nevertheless stem-final in a compound word); otherwise only word-finally and only in about 18 words mainly of French origin, namely: (1) with a preceding consonant letter such that <que> could be replaced by <k> without changing the pronunciation: arabesque, barque, basque, brusque pronounced /brʌsk/ (also pronounced /bruːsk/), burlesque, casque, catafalque, grotesque, marque, masque, mosque, torque and the derived forms picturesque, romanesque, statuesque. However, in this group barque, basque, casque, marque, masque, torque are kept visually distinct from bark, bask, cask, mark, mask, torc (2) with a preceding vowel letter such that <que> could be replaced by <ck> without changing the pronunciation: baroque, cheque (cf. US check), monocoque, plaque pronounced /plæk/ (also pronounced /plɑːk/) |
|
<x> spells /k/ (not /ks/, etc.) |
only in coxswain and before /s/ spelt <c> in a small group of words of Latin origin, namely exceed, excel(lent), except, excerpt, excess, excise, excite |
Other 2-phoneme graphemes |
See also Notes |
|
|
/k∫/ |
|
|
(1) spelt <x> |
only in flexure, luxury, sexual /ˈflekʃə, ˈlʌkʃəriː, ˈsekʃ(uːw)əl/ |
|
(2) spelt <xi> |
only in anxious, complexion, connexion (also spelt connection), crucifixion, fluxion, (ob)noxious |
|
/ks/ |
|
|
(1) spelt <xe> |
only in annexe, axe, deluxe, (River) Exe. The <e> in axe is redundant, as the US spelling ax shows (but cf. the ‘Three-Letter Rule’, section 4.3.2). The <e> in annexe /ˈæneks/ (‘addition to building or document’) is also phonologically redundant (and mainly omitted in US spelling) but, where used, differentiates this word visually from annex /əˈneks/ (‘take over territory’). Similarly, deleting the final <e> from the French spelling of de luxe would get too close to soap and washing powder |
|
(2) spelt <xh> |
only in exhibition, exhortation, exhumation – for exhibit, exhort, exhume see under /g/, section 3.5.3 |
3-phoneme grapheme |
/eks/ spelt <x> |
only in X-ray, etc. One of only two 3-phoneme graphemes in the whole language |
Notes
For adverbs with the unstressed ending /ɪkliː/ spelt <-ically> see section 6.10.
It is unphonological but true that it is easier to state the main correspondences of /k/ in terms of following letters rather than following phonemes. (For an attempt to do it phonologically see Carney, 1994: 217).
<k> is used to spell /k/ mainly before the letters <e, i, y>, that is, just where <c> would usually spell /s/ – see below. There are very few exceptions:
1) where /k/ is spelt <c> despite being before <e, i>: Celt, Celtic, sceptic, all of which have alternative spellings with <k> (and the Glasgow football club is in any case /ˈseltɪk/), arced, arcing, synced, syncing (which means that the spelling synch for this verb is better); also several words beginning encephal-, all of which have two pronunciations, with /s/ (where the spelling with <c> is regular) or /k/ (where it is irregular), e.g. encephalitis /enˈsefəlaɪtəs, eŋˈkefəlaɪtəs/ - note too the alternation between /n, ŋ/ in the first syllable. Also, in July 2006 the derived form chicest /ˈʃiːkɪst/ appeared on a magazine cover, and in May 2010 ad hocery appeared in the Guardian. There seem to be no exceptions with /k/spelt <c> before <y>
2) where /k/ is spelt <k> despite not being before <e, i, y>: alkali, askance, blitzkrieg, hokum, kale, kangaroo, kaolin, kappa, kapok, kaput, klaxon, kleptomaniac, koala, kohl, kopek, Koran, kosher, krypton, kwashiorkor (twice), leukaemia, mazurka, oakum, okay, okra, paprika, polka, sauerkraut, shako, skate, skulk, skull, skunk, sudoku, tektite, ukulele.
<k> is also the regular spelling of /k/ at the end of one-syllable words where the preceding phoneme is NOT a short vowel, e.g. bark (only exception: burke); also in two-syllable words after a consonant letter and before final /əl/ spelt <-le> – but there are very few words in this set, namely ankle, crinkle, rankle, sparkle, sprinkle, tinkle, twinkle, winkle, wrinkle (exceptions: circle, uncle).
<q> is almost always used to spell /k/ when followed by /w/, which in this context is always spelt <u>. For exceptional spellings of /kw/, see /w/, section 3.8.7.
In addition to its single-grapheme spelling, <x>, /ks/ has several 2-grapheme spellings. There are none in initial position (where /ks/ perhaps occurs only in xi anyway). Word-finally, where the /s/ is not part of the stem the regular spelling in one-syllable words after a short vowel is <cks>, and <ks> in other one-syllable words; <cs> is regular in polysyllables – for exceptions to all of these see Table 3.3. Medially, there are three 2-grapheme spellings of /ks/: <xc> is rare – see the last entry among the Oddities above; <cs> is even rarer – it seems to occur only in ecstasy, ecstatic, facsimile, frolicsome, tocsin (contrast toxin); <cc> occurs in a few words mainly of Latin or French origin where the following letter is <e, i, y>, namely accede, accelerate, accent, accept, access, accident, accidie, coccyx, eccentric, flaccid pronounced /ˈflæksɪd/ (also pronounced /ˈflæsɪd/), occident, occiput, Occitan(e), succeed, success, succinct pronounced /səkˈsɪŋkt/ (also pronounced /səˈsɪŋkt/), vaccine.
It is because <c> before <e, i, y> almost always spells /s/ that <cc> can not function as the doubled spelling of /k/ – before a suffix beginning with <e, i, y> the second <c> would represent /s/ (as in the group of words just listed). So when a suffix beginning with a vowel letter is added to words ending in /k/ spelt <c>, the <c> is usually doubled to <ck>, e.g. bivouacked, picnicking, trafficked – but the principle of avoiding <c> spelling /k/ before <e, i, y> is not applied to arced, arcing, chicest (cf. above and section 4.2). <cc> also has the very rare pronunciation /ʧ/ only in bocce, cappuccino (see next section).
/k∫/ has scarcely any 2-grapheme spellings, but cf. baksheesh.
The word ache is one of the few where the split grapheme <a.e> has two consonant letters in its midst (it could equally well be spelt *ake).
As Carney says (p.216), ‘/k/ is the most divergent of the consonants’. For this reason, a further analysis of the major spellings of /k/ is given in Table 3.3.
In each main box below, the regular spelling is stated at the top or above the relevant set of words. (For exceptions besides those in the Table, see the 2- and 3-phoneme graphemes and Oddities above).
initial |
medial |
final |
|
/k/ before <e, i, y> |
<k>, e.g. kelp, kit, kyle Exceptions: Celt, Celtic |
<k>, e.g. sketch, skit, sky Exceptions: sceptic; lichen pronounced /ˈlaɪkən/ (also pro-nounced /ˈlɪʧən/); chicken, cricket, jacket, mackerel, pernickety, pickerel, pocket, rocket, sprocket; mackintosh; finicky Words ending in <-c> usually add <k> when a suffix beginning with a vowel letter is added, e.g. bivouacked, picnicking, trafficked (but arced, arcing, chicest do not) |
<k> Occurs only within split digraphs, e.g. cake, eke, bike, poke, rebuke, fluke, tyke Exception: ache |
/k/ not before <e, i, y> |
<c>, e.g. cake, close, coal, cross, cute Exceptions: kagoul (also spelt cagoule), kale, kangaroo, kaolin, kappa, kapok, kaput, kayak, klaxon, klepto-maniac, koala, kohl, kopek, Koran, kosher, krypton |
<c>, e.g. scale, eclogue, scorch, across, acute Exceptions: With <ck>: between short vowel spelt with one letter and final /əl/ spelt <-le>, e.g. crackle (see section 4.3.3), plus beckon, buckaroo, buckshee, chickadee, cockatoo, cockaigne, cockatrice, cockney, gecko, hackney, hickory-dickory, huckster, jackanapes, lackadaisical, reckon, rucksack, sackbut With <k>: in two-syllable words after a consonant letter and before final /əl/ spelt <-le>, namely ankle, crinkle, rankle, sparkle, sprinkle, tinkle, twinkle, winkle, wrinkle (exceptions: circle, uncle), plus alkali, askance, blitzkrieg, hokum, leukaemia, mazurka, oakum, okay, okra, paprika, periwinkle, polka, sauerkraut, shako, skate, skulk, skull, skunk, sudoku, |
In one-syllable words after a short vowel spelt with one letter: <ck>, e.g. back, beck, trick, clock, duck Exceptions: bloc, choc, doc, hic, mac, roc, sac, sic, spec, tec, tic; flak, suk, trek, yak In other one-syllable words: <k>, e.g. ark, bank, brook, freak Exceptions: arc, chic, disc, franc, orc, talc, torc, torque, zinc In polysyllables: <c>, e.g. politic Exceptions: With <ck>: alack, attack, bailiwick, bannock, barrack, bollock, bullock, burdock, buttock, cassock, Cossack, derrick, dunnock, fetlock, fossick, gimcrack, gimmick, haddock, hammock, hassock, haversack, hemlock, hillock, hollyhock, hummock, limerick, |
tektite, ukase, ukulele With <x> spelling only /k/ because the following <c, sw> spells /s/: exceed, excel, excellent, except, excerpt, excess, excise, excite; coxswain |
mattock, maverick, niblick, paddock, pollack, ransack, rollick, rollock, rowlock, rucksack, shamrock, slummock, tussock, warlock With <k>: aardvark, asterisk, basilisk, batik, bergomask, berserk, Bolshevik, bulwark, damask, kopek, mountebank, muzak, obelisk, Slovak, sputnik, springbok, tamarisk, tomahawk, yashmak |
||
/ks/ |
Very rare – perhaps occurs only in Greek letter name xi pro-nounced /ksaɪ/ |
<x>, e.g. maxim, next, toxin Exceptions: - where <x> occurs but spells only /k/ (see box above): exceed, excel, excellent, except, excerpt, excess, excise, excite; coxswain; - others: accede, accelerate, accent, accept, access, accident, accidie, coccyx, eccentric, flaccid, occident, occiput, Occitan, succeed, success, succinct, vaccine; ecstasy, ecstatic, facsimile, frolicsome, tocsin |
Where /s/ is part of stem: <x>, e.g. fax, perplex, six, box, influx, pyx Only exception: aurochs Where /s/ is not part of stem, = is a suffix: all the non-suffixed forms of such words belong in the two boxes above, and in all these cases, when suffixed, /ks/ is spelt with the non-suffixed grapheme plus <s> |
The main system
For all these categories see Notes
Basic grapheme |
<ch> |
65% |
regular initially except before /uː/, e.g. chin, church; also finally (except in one-syllable words after a short vowel spelt with one letter), e.g. church (exceptions: despatch, dispatch, eldritch); rare medially, but cf. bachelor, duchess |
Other frequent grapheme |
<t> |
25% |
regular medially, e.g. actual, intuition; also initially before /uː/ spelt <u, u.e>, e.g. tulip, tune; never word-final |
Doubled spelling |
<tch> |
10% |
word-initially, only in Tchaikovsky; rare medially – does NOT occur before final /əl/ spelt <-le> – but there are a few examples, e.g. butcher, crotchet, hatchet, ketchup, kitchen, patchouli, pitcher, ratchet, satchel, (e)scutcheon, tetchy, wretched; regular in word-final position in one-syllable words after a short vowel spelt with one letter, e.g. match; exceptions: much, rich, such, which, niche pronounced /nɪʧ/, kitsch, putsch; also (irregularly after a diphthong/long vowel) in aitch, retch |
The rest
Doubled spelling + <e> |
(does not occur) |
|
Oddities |
<1% in total |
|
|
<c> |
only in cellist, cello, cicerone (twice), concerto (second <c>) |
|
<cc> |
only in bocce, cappuccino |
|
<che> |
only in niche pronounced /nɪʧ/, which could be spelt *nitch; niche is also pronounced /niːʃ/ |
|
<ci> |
only in ancient, ciabatta |
|
<cz> |
only in czardas /ˈʧɑːdæʃ/, Czech /ʧek/ |
|
<te> |
only in righteous |
|
<th> |
only in posthumous |
|
<ti> |
only in con/di/indi/in/sug-gestion, question, rumbustious and the derived forms combustion, exhaustion. In words like nation, lotion, equation I count the <i> as part of a digraph with the preceding consonant letter – see /∫, ʒ/, sections 3.8.3-4 |
|
<tsch> |
only in kitsch, putsch |
2-phoneme graphemes |
(none) |
Notes
Because /ʧ/ is a sibilant consonant, adding any of the suffixes regular noun plural and third person singular person tense verb (spelt <es> - for an exception see next sentence) and regular singular and irregular plural possessive (spelt <’s>) to a stem ending in /ʧ/ adds a syllable /ɪz/ as well as a morpheme: matches, detaches, (the) Church’s (mission). The only word ending in /ʧ/ where the stem already ends in <e> appears to be niche pronounced /nɪʧ/; in this case the ending is just <s>. See also /z/, section 3.7.8, and /ɪ/, section 5.4.3.
The regular spellings of /ʧ/ are:
- initially, <t> before /uː/, otherwise <ch>
- medially, <t>
- finally, <tch> in one-syllable words after a short vowel spelt with one letter, otherwise <ch>.
Examples:
- initial <t> before /uː/: tuba, tube, tuber, Tuesday /ˈʧuːzdiː/, tuition /ʧuːˈwɪ∫ən/, tulip, tumour, tumult(uous), tumulus, tuna, tune pronounced /ˈʧuːn/, tunic, tureen, tutor
- initial <ch> otherwise: chin, church
- medial <t>:
- a small set of words ending in /ʧən, ʧəs/ spelt <-tian, -tion, -tious>: Christian, combustion, con/di/indi/in/sug-gestion, exhaustion, question, rumbustious
- many nouns ending in /ʧə/, which is mostly spelt <-ture>, e.g. adventure, capture, creature, culture, picture
- a set of adjectives ending in /ʧuːwəs/, which are all spelt <-tuous>, e.g. tortuous, virtuous
- a small set of nouns in <-tuary>: actuary, estuary, mortuary, obituary, sanctuary, statuary, voluptuary, whether pronounced with /ʧuːwəriː/ or /ʧəriː/. For the elision of the <u> see section 6.10
- a larger set of adjectives in <-tual>: accentual, actual, conceptual, contractual, effectual, eventual, factual, habitual, intellectual, mutual, perpetual, punctual, ritual, spiritual, textual, virtual, etc., whether pronounced with /ʧuːwəl/ or /ʧəl/. For the elision of the <u> see again section 6.10. The elision seems even more prevalent in the derived adverbs in <-tually>
- a small set of words where /ʧ/ is spelt <t> and a following medial /ə/ (always in the penultimate syllable, with the stress on the antepenultimate syllable) is spelt <u>: century, congratulate, fistula, flatulen-ce/t, fortunate, petulan-t/ce, postulant, postulate, saturate, spatula, titular
- a ragbag of words with /ʧ/ spelt <t> followed by /uː/ spelt <u, u.e> (occasionally in word-final position as <ue>), e.g. impromptu; gargantuan, perpetuate; attitude, multitude, solitude; habitué; statue, virtue; intuition, pituitary; costume; fortune, importune, opportune; virtuoso; obtuse; de/in/pro/sub-stitute, de/in/pro/re/sub-stitution
- just two words where the stress falls on the syllable following /ʧ/ spelt <t> and that syllable contains <ur(e)> spelling /ʊə/: centurion, mature. In all the other words with medial /t/ spelt /ʧ/ listed above the stress falls on an earlier syllable.
- final <tch> after a short vowel spelt with one letter in monosyllables: match, sketch, pitch, botch, hutch, butch
- final <ch> otherwise: attach, arch, church.
Exceptions (other than those listed under Oddities):
- initial <t> other than before /uː/: none
- initial <ch> before /uː/: only chew, choose
- initial spellings other than <t, ch>: <tch> only in Tchaikovsky
- medial spellings other than <t>: only archer, bachelor, cochineal, duchess, duchy, lecher, lichen pronounced /ˈlɪʧən/ (also pronounced /ˈlaɪkən/), macho, treacher-y/ ous; butcher, crotchet, hatchet, ketchup, kitchen, patchouli, pitcher, ratchet, satchel, (e)scutcheon, tetchy, wretched as stem words, but there are also many derived forms, e.g. lurcher, marcher, matching, preacher, righteous, (re)searcher, teacher; also the words in <ti> listed in the Oddities
- final <tch> in monosyllables after a diphthong or long vowel: only aitch, retch pronounced /riːʧ/ (also pronounced /reʧ/, where <tch> is regular)
- final <tch> in polysyllables: only despatch, dispatch, eldritch
- final <ch> in monosyllables after a short vowel: only much, rich, such, which
- final spellings other than <tch, ch>: see Oddities.
As a spelling of /ʧ/, <ti> is rare and occurs only at the beginning of the final syllable of a stem word and immediately after a stressed syllable ending in /s/ spelt <s>, e.g. question.
All the words in which /ʧ/ is spelt <t> were formerly pronounced with the sequence /tj/, and conservative RP-speakers may still pronounce them that way (or imagine that they do). Pronunciations with /tj/ would require an analysis with the /t/ spelt <t> and the /j/-glide either spelt <i> (where that is the next letter) or subsumed into the spelling of a 2-phoneme sequence with the following vowel. However, I think that in current RP the process of affricating /tj/ to /ʧ/ is virtually complete (as Cruttenden, 2014: 83 says), and has eliminated pronunciations with /tj/, which I have therefore ignored. An English friend who once did a year’s teaching exchange in a primary school in the United States would ask the pupils every day, ‘What day is it?’ and on Mondays, Wednesdays, Thursdays and Fridays they would answer. But on Tuesdays they would insist she give the name of the day, and would then delightedly point out, ‘You say Choozdee /ˈʧuːzdiː/!’ (as opposed to their /ˈtuːzdiː/ ‘Toozdee’, where the /j/-glide has been dropped without affricating the /t/, or perhaps was never present).
For the parallel affrication of /dj/to /ʤ/ see section 3.7.4, and see also section 5.4.7.
The main system
Basic grapheme |
<f> |
84% |
e.g. fish |
Other frequent grapheme |
<ph> |
11% |
in many words of Greek origin, e.g. philosophy. See Notes |
Doubled spelling |
<ff> |
4% |
regular in word-final position in one-syllable words after /ɑː/ spelt <a>, e.g. staff, and after a short vowel spelt with one letter, e.g. gaff, cliff, off, gruff; cf. section 4.3.5 and /ɑː/, section 5.5.1; for off see also section 4.3.2 (exceptions: graph, gaffe, chef, clef, if; strafe is not an exception because it has /ɑː/ spelt <a.e>); also regular medially before final /əl/ spelt <-le> after a short vowel spelt with one letter, e.g. duffle – see section 4.3.3; there are also some independent medial examples, e.g. affray, buffet (both pronunciations and meanings), chiffon, offer, proffer, ruffian, soffit, suffer – see section 4.3.5. Also see Notes |
The rest
Doubled spelling + <e> |
<ffe> |
only in gaffe, giraffe, pouffe; also in usual pronunciation of difference, different, sufferance – for the elided vowel see section 6.10 (but not in afferent, efferent) |
Oddities |
1% in total |
|
|
<fe> |
only in carafe, housewife (‘sewing kit’ pronounced /ˈhʌzɪf/) and, in rapid speech, conference, deference, preference, reference – for the elided vowel see section 6.10 |
|
<ft> |
only in often, soften pronounced /ˈɒfən, ˈsɒfən/ |
|
<gh> |
medially, only in draught and derived forms of the following words; otherwise only word-final and only in chough, cough, enough, laugh, rough, slough (‘shed skin’), sough, tough, trough |
|
<pph> |
only in sapphic, sapphire, Sappho /ˈsæfɪk, ˈsæfaɪə, ˈsæfəʊ/ |
|
<v> |
only in kvetch, svelte, svengali, veldt |
2-phoneme graphemes |
(none) |
Notes
In monosyllables, the default spelling is <f>, except where <ff> is regular as defined above. There are a few exceptions with the Greek <ph> spelling: graph, lymph, morph, phase, phone, phrase.
In polysyllabic stem words, there are the following tendencies in the distribution of the three main spellings of /f/:
- <ph> occurs almost solely in words of Greek origin, and <f, ff> in other words – but how (unless you have studied Classical Greek, as I did, or know modern Greek) can you tell which words are of Greek origin? Though few people could answer this explicitly, many internalise the word-elements which have that origin and require the <ph> spelling, e.g. graph, lymph, morph, phag, phall, pharmac, pharyng, phase, pheno, phleb, phil(e/o), phob(e), phon(e/ic/o-), phor(e), phosph, photo, phrase, phren, phyll, phys, phyt(e/o), soph, spher, taph. As Carney (1994: 229) points out, there is further guidance towards <ph> in polysyllabic words if other Greek word-elements are present, e.g. anthrop, apo, chloro, chron, crypto, dia, dys, epi, eu, geo, hiero, hydro, hyper, hypo, lexi, macro, meta, micro, oid, ology, ortho, peri, scop, syn, tele, thus aiding correct spelling of, e.g., apocrypha(l), chlorophyll, cryptographer, diaphragm, euphemism, euphoria, hieroglyphics, metaphor, psephology.
Words with the element /fænt/ were all once spelt with <ph> and most still are (e.g. phantasm, phantom) but, awkwardly, three of the commonest are now spelt with <f>: fantasise, fantastic, fantasy.
The Greek elements neo, para may be misleading; they have taken on lives of their own in modern English (e.g. neocon, paramedic, where the second element in each word has a Latin origin) and might therefore mislead writers into, e.g., *neofite, *paraprophessional.
Beyond this, one can only list some of the commoner of the other words containing <ph>: alpha, aphid, aphrodisiac, asphyxiate, blaspheme, cenotaph, dolphin, elephant, hyphen, lymph, nymph, orphan, phalanx, pheasant, phenol, phial, philistine, phloem, phoenix, siphon/syphon, sphinx, sylph, trophy, zephyr, and four words where the pronunciation varies between /f/ and /p/: diphtheria, diphthong, naphtha, ophthalmic. Words of non-Greek origin in this set are: caliph, cipher/cypher, nephew (also pronounced with /v/), pamphlet, pharaoh, Pharisee, phwoar!, samphire, seraph, triumph.
Nothing but a source of confusion would be lost if all these words with /f/ spelt <ph> were instead spelt with <f>, as the cognate words are in Italian and Spanish.
- <ff>
1) For 2-syllable words ending in /əl/ spelt <-le> and with a short vowel spelt with one letter in the first syllable, see above.
2) There is a strong tendency for /f/ to be written <ff> in the middle of two-syllable words where the immediately preceding vowel phoneme is short and written with a single letter, e.g. offer. For examples and exceptions, see section 4.3.5.
3) There is also a strong tendency for /f/ to be written <ff> rather than <f> at the end of the third from last syllable of a word /f/ is an exception to a wider rule in this respect. For examples and exceptions, see section 4.4.5.
4) Word-finally in polysyllables not of Greek origin <ff> predominates: bailiff, caitiff, chiffchaff, dandruff, distaff, handcuff, mastiff, midriff, plaintiff, pontiff, rebuff, riffraff, sheriff, tariff, plus fisticuffs; contrast belief, (hand)kerchief, mischief, relief; caliph, seraph, triumph.
5) In the few remaining words not of Greek origin, <ff> again predominates: affidavit, affiliate, affinity, effeminate, efficacious, effrontery, paraffin, ragamuffin; contrast cafeteria, defeasible, deferential, defibrillate, defoliant, nefarious.
- <f> is the default spelling. The only generalisation, however, is that it is regular in consonant clusters in words not of Greek origin, e.g. afraid, after, deflate, deflect, defray, defrock, kaftan; exceptions include affray, effrontery.
The main system
For all these categories see also Notes.
Basic grapheme |
<j> |
29% |
never word-final; regular initially (e.g. jet), and medially when not followed by <e, i, y>, e.g. ajar, banjo, cajole, conjugal, enjoy, juju, major, (maha)rajah, sojourn. On this basis, the initial <j> in jujitsu is regular, but the medial one is not |
Other frequent graphemes |
<g> |
51% |
never word-final (except Reg, veg); regular medially before <e, i, y> |
|
<ge> |
10% |
word-initial only in geograph-er/y, geomet-er/ry, Geordie, George, Georgia(n), georgic; rare medially, but cf. burgeon, dungeon, gorgeous, hydrangea, pageant, pigeon, sergeant, sturgeon, surgeon, vengeance where the following /ə/ (or /ɪ/ in pigeon) is spelt <a, o, ou>; also dangerous, vegetable if <e> is elided - see section 6.10; also in the derived forms singeing, swingeing to prevent confusion with singing, swinging, and bingeing, spongeing, whingeing to avoid the misapprehension that there might be verbs to *bing, to *spong, to *whing (but fringing, impinging never retain the <e>); mostly word-final, e.g. binge, blancmange, change, disparage, flange, fringe, garage |
pronounced /ˈgærɪʤ/, haemorrhage, hinge, image, language, lounge, mortgage, orange, impinge, scavenge, singe, sponge, village, whinge and hundreds of words ending in <-age> where the <e> is also part of the split digraph <a.e>, e.g. age, rage, stage. See section 7.1 for dual-functioning, section 10.4 for<a.e>, and section A.6 in Appendix A for the rarity of other split digraphs with included <g> |
|||
Doubled spellings |
<dge> |
5% (with <dg>) |
never word-initial or medial; regular in word-final position in one-syllable words after a short vowel spelt with one letter, e.g. bridge, judge. See next paragraph, and section 6.4 on when <e>-deletion does and does not occur before suffixes beginning with a vowel letter |
|
<dg> |
never word-initial or -final; medially, does NOT occur before final /əl/ spelt <-le>, and most occurrences arise from deleting <e> from <dge> before suffixes beginning with a vowel letter, e.g. bridging. However, there are a few words with independent medial /ʤ/ spelt <dg>: badger, budgerigar, budget, budgie, codger, cudgel, didgeridoo, dodgem, fidget, gadget, kedgeree, ledger, midget, podgy, smidgen, smidgin, todger, widget; also, as more obviously belonging to this set than to a set with medial <dge>, provided <eo> is recognised as a spelling of /ə/ (see p.155), bludgeon, curmudgeon, dudgeon, gudgeon, smidgeon, widgeon |
The rest
Doubled spelling + <e> |
(cannot occur because unsuffixed word-final doubled spelling already ends in <e>) |
|
Oddities |
5% in total |
|
|
<ch> |
only in ostrich, sandwich, spinach pronounced /ˈɒstrɪʤ, ˈsæmwɪʤ, ˈspɪnɪʤ/ |
|
<d> |
never word-final; frequent initially and medially before /uː, ə, ʊə/ spelt with various graphemes involving letter <u>, namely <eu, eur, ew, u, ua, ue, u.e, ur, ure>, e.g. (initially) deuce (cf. the homophone juice), various words beginning with (Greek) deuter-; dew, due (which are homophones, and cf. the further homophone Jew); dual/ duel |
|
<di> |
only in cordial pronounced /ˈkɔːʤəl/ (also pronounced /ˈkɔːdiːjəl/), incendiary, intermediary, stipendiary, subsidiary pronounced with /-ʤəriː/, soldier |
|
<dj> |
only in about 10 words of Latin origin: adjacent, adjective, adjoin, adjourn, adjudge, adjudicate, adjunct, adjure, adjust, adjutant, plus djinn |
|
<gg> |
only in arpeggio, exaggerate, loggia, suggest and the derived forms Reggie, veggie, vegging. The last three words appear to be the only examples of <gg> spelling /ʤ/ arising from consonant-doubling before a suffix – see section 4.2 |
|
<gi> |
only in allegiance, contagio-n/us, egregious, legion, litigious, plagiaris-e/m, region, religio-n/us and the derived forms collegial, prestigious, vestigial |
|
<jj> |
only in hajj |
2-phoneme graphemes |
(none) |
Notes
The words geograph-er/y, geomet-er/ry could alternatively be analysed as having initial /ʤ/ spelt <g> and the following /ə/ spelt <eo>, but this would entail a counter-intuitive analysis of Geordie, George, Georgia(n), georgic as having /ɔː/ spelt <eor>, so I have retained the analysis of geograph-er/y, geomet-er/ry as having initial /ʤ/ spelt <ge> and the following /ə/ spelt <o>.
Because /ʤ/ is a sibilant consonant, addition of any of the suffixes regular noun plural and third person singular person tense verb (both spelt <s> where the stem ends in <e>, otherwise <es>) and regular singular and irregular plural possessive (spelt <’s>) to a stem ending in /ʤ/ adds a syllable /ɪz/ as well as a morpheme: languages, sandwiches, (the) bridge’s (collapse). See also /z/, section 3.7.8, and /ɪ/, section 5.4.3.
To summarise, the regular spellings of /ʤ/ are:
- in word-initial position: <j> (73% of spellings in that position)
- in medial position before <e, i, y>: <g>
- in medial position otherwise: <j>
- in stem-final position in unsuffixed one-syllable words after a short vowel spelt with one letter: <dge>
- in stem-final position when <dge> loses the <e> before a suffix beginning with a vowel letter (see section 6.4): <dg>
- otherwise in word-final position (including dual-functioning of the <g> within split digraphs): <ge>.
Exceptions (in addition to the Oddities):
- initial <g> (27% of spellings in that position): gaol, gee, gel (/ʤel/ ‘viscous liquid’; contrast gel pronounced /gel/, ‘posh’ version of girl), gelatin, gelignite, gem, geminate, Gemini, Gemma, gen, gender, gene, general, generate, generic, generous, genial, genie, genital, genitive, genius, gent, gentle, genuflect, genuine, genus, most words beginning with (Greek) geo ‘earth’, e.g. geographic (but not those listed above as having initial /ʤ/ spelt <ge>), Geoff(rey), geranium, gerbil, geriatric, germ, German, gerrymander (also spelt with <j>), gerund, gestate, gesture, giant, gibber, gibbet, gibe, giblets, gigantic, gigolo, gill (/ʤɪl/ ‘quarter of a pint’; contrast gill /gɪl/ ‘lung of fish’), Gillingham /ˈʤɪlɪŋəm/ in Kent (but /ˈgɪlɪŋəm/ in Dorset and Norfolk), gimbal(s) (also pronounced with /g/), gimcrack, gin, ginger(ly), ginseng, gipsy, giraffe, giro, gist, gym(nas-t/ium), gyp, gypsum, gypsy, gyrate, various words beginning with (Greek) gyro-, gyves
- medial <j> before <e, i>: only in jujitsu, majest-y/ic and words with the Latin element <ject> (‘throw’), namely ab/de/e/in(ter)/ob/pro/re/sub-ject, conjecture, trajectory (no exceptions before <y>)
- medial <g> not before <e, i, y>: only in margarine (also pronounced with /g/), second <g> in mortgagor
- medial <dg>: see list above
- medial <ge>: see list above
- final <g>: only in Reg, veg
- final <dge> in words of more than one syllable: only in abridge, cartridge, (ac)knowledge, partridge, porridge.
All the words in which /ʤ/ is spelt <d> were formerly pronounced with the sequence /dj/, and conservative RP-speakers may still pronounce them that way (or imagine that they do). Pronunciations with /dj/ would require an analysis with the /d/ spelt <d> and the /j/-glide subsumed into the spelling of a 2-phoneme sequence with the following vowel. However, I think that in current RP the process of affricating /dj/ to /ʤ/ is virtually complete (as Cruttenden, 2014: 83 says) and has eliminated pronunciations with /dj/, which I have therefore ignored.
For the parallel affrication of /tj/to /ʧ/ see section 3.7.2, and see also section 5.4.7. In the case of /ʧ/ there are very few spellings competing with <t> in initial and medial positions before /uː/, etc., and <t> is therefore the regular spelling. However, for initial and medial /ʤ/ there are many words spelt with <j> before /uː/, etc., so that <d> cannot be considered the regular spelling in these circumstances, or ‘promoted’ to the main system. This also means that words in which /ʤ/ is spelt <d> cannot be predicted and just have to be learnt.
For the ending /ɪʤ/ see also under /ɪ/, section 5.4.3.
The main system
Basic grapheme |
<l> |
75% |
e.g. lift |
Doubled spelling |
<ll> |
18% |
e.g. fill. See Notes |
Frequent 2-phoneme grapheme |
/əl/ |
8% |
e.g. dazzle, debacle, table, visible. See Notes |
The rest
Doubled spelling + <e> |
<lle> |
medially, only in decollet-age/e; otherwise only final. Regular in the ending -ville, e.g. vaudeville; also in bagatelle, belle, braille, chanterelle, espadrille, fontanelle, gazelle, grille, pastille, nacelle, quadrille (but not reveille, tagliatelle where the <e> spells /iː/). In chenille, tulle I analyse /l/ as spelt <ll> and <i.e, u.e> as split digraphs spelling /iː, uː/ - see sections 5.7.2, 5.7.6, A.6 – and medially in guillemot <lle> spells /liː/ |
Oddities |
<1% |
|
|
<gl> |
only in a few Italian loanwords, e.g. imbroglio, intaglio, seraglio, tagliatelle |
|
<le> |
except in Charles, only word-final and only in aisle, cagoule, clientele, gargoyle, joule, isle, lisle, voile. On isle, lisle see also /aɪ/ spelt <is>, section 5.7.3 |
|
<lh> |
only in philharmonic, silhouette |
Other 2-phoneme graphemes |
/əl/ |
only word-final and only in axolotl, dirndl, shtetl |
|
/lj/ |
only in carillon |
Notes
At 18%, <ll> has the highest frequency of all the doubled consonant spellings (at least in stem words, i.e. discounting consonant-doubling before suffixes):
- It occurs in the two exceptional word-initial doubled consonant spellings llama, llano.
- It is regular in word-final position in one-syllable words after /ɔː/ spelt <a>, e.g. all, and after a short vowel spelt with one letter, e.g. shall, ell, ill, moll, cull, bull. Exceptions: col, gal, gel (both /gel/, posh pronunciation of girl, and /ʤel/ ‘lotion’), mil (abbreviation of millimetre), nil, pal; gal, pal would otherwise look identical to gall, pall and be pronounced with /ɔː/. For all, ell, ill see also section 4.3.2.
- There is a strong tendency for /l/ to be spelt <ll> in the middle of two-syllable words where the immediately preceding vowel phoneme is short and written with a single letter. For examples and exceptions, see sections 4.3.4 and 4.4.6.
- There is also a preference for /l/ to be spelt <ll> rather than <l> at the end of the third from last syllable of a word. /l/ is an exception to a wider rule in this respect. For examples and exceptions, see section 4.4.5.
- There appear to be only a few polysyllabic non-compound words ending in <ll>: chlorophyll and some rare words in -phyll, idyll, plimsoll (also spelt plimsole). All other non-compound polysyllables end in <l>, except those listed above under <lle>.
- Similarly, there appear to be only three three-syllable words with <ll> at the end of the second syllable/beginning of the last syllable: embellish, intellect, parallel.
- In words of more than three syllables, no generalisation seems possible, so here is a list of such words which have <ll>: allegory (3 syllables if the <o> is elided), alleviate, alligator, alliterate, various words beginning with (Greek) allo-, ballerina, calligraphy, camellia, collaborate, collateral, fallopian, hallelujah, hallucinate, hullabaloo, illegible, illegitimate, illiberal, illimitable, illiterate, illuminate, mellifluous. Alleviate, alliterate and all those just listed beginning <coll-> and <ill-> belong to the (to most people, meaningless) category of words with Latin roots and assimilated Latin prefixes – see section 4.3.1.
On reducing <ll> to <l> in compound words see section 4.4.7.
In non-final positions, the 2-phoneme sequence /əl/ only has straight (non-reversed) 2-grapheme spellings, e.g. allowed, aloud. But in final position, although the reversed spelling <-le> predominates, there is considerable variation between that and several non-reversed spellings. The situation is too complex to summarise at this point; see sections 4.3.3 and 4.4.2-3.
For 2-grapheme spellings of /lj/ see under /j/, section 3.8.8.
For <lle> in chancellery, jewellery see section 6.10.
The main system
For all these categories see Notes and Tables 3.4-5.
Basic grapheme |
<s> |
79% |
e.g. sat, persuade, bias Regular (1) initially and medially where the next letter is not <e, i, y>; (2) in most unstressed final syllables of polysyllables; (3) in various suffixes and contracted forms after a non-sibilant voiceless consonant; (4) within split digraph <o.e> |
Other frequent graphemes |
<c> |
15% |
e.g. city, decide Regular initially and medially where the next letter IS <e, i, y>. Never word-final |
|
<ce> |
e.g. fence, mice Except in a few suffixed forms (see section 6.4), only word-final, where it is regular after /n/ and when the <e> is also part of split digraphs <a.e, i.e, u.e, y.e>, e.g. ace, ice, puce, syce (for dual-functioning see section 7.1, and for split digraphs section A.6 in Appendix A), but is otherwise unpredictable |
|
|
<se> |
only word-final, where it is regular after /l, p/, after <r> forming part of a vowel di-/tri-graph, and after most vowel digraphs |
|
Doubled |
<ss> |
regular word-finally in one-syllable words after /ɑː/ spelt <a> and after a short vowel spelt with one letter, e.g. grass, fuss; also in suffixes <-ess, -less, -ness> and in stressed final syllables of polysyllables |
|
Rare 2-phoneme grapheme |
/ks/ |
see under /k/, section 3.7.1. Though rare as a correspondence for /s/, this counts as part of the main system because of its higher frequency as a correspondence for /k/ |
The rest
Doubled |
<sse> |
except in divertissement, only word-final, e.g. bouillabaisse, crevasse, duchesse, finesse, fosse, impasse, lacrosse, largesse (also spelt largess), mousse, noblesse, palliasse, wrasse and a few more rare words |
Oddities |
6% in total |
|
|
<cc> |
only in flaccid, succinct pronounced /ˈflæsɪd, səˈsiŋkt/ (also pronounced with /ks/) |
|
<ps> |
only word-initial and only in some words of mainly Greek origin, e.g. psalm, psalter, psephology, pseud(o) and many compounds, psionic, psittacosis, psoriasis, psych(e/o) and many compounds, and a few more very rare words. /p/ surfaces in metempsychosis – see section 7.2 |
|
<sc> |
only in abscess, abscissa, adolescen-t/ce, ascend, ascertain, ascetic, corpuscle, crescent (also pronounced with /z/), descend, discern, disciple, fascicle, fascinate, isosceles, lascivious, miscellany, muscle, (re)nascent, obscene, omniscient, oscillate, plebiscite, prescient, proboscis, proscenium, rescind, resuscitate, scenario, scene, scent, sceptre, sciatic(a), science, scimitar, scintilla, scion, scissors, |
scythe, susceptible, transcend, viscera(l), viscid and a few more rare words, plus suffixed derivatives of next group. /k/ surfaces in corpuscular, muscular – see section 7.2 |
||
|
<sce> |
only in stressed final syllables of verbs ending /es/ spelt <-esce>, e.g. acquiesce, coalesce, convalesce, deliquesce, effervesce, evanesce and some other very rare words, plus reminisce |
|
<sch> |
only in schism pronounced /ˈsɪzəm/ |
|
<st> |
only medial. Between a short vowel spelt with one letter and final /əl/ spelt <-le>, <st> is the regular spelling of /s/, but this is a small set: pestle, trestle, bristle, Entwistle, epistle, gristle, thistle, Thistlethwaite, Twistleton, whistle, apostle, jostle, Postlethwaite, throstle, bustle, hustle, rustle and the derived forms nestle, wrestle (exceptions: hassle (but I once received an email with this word spelt *hastle, showing the power of this sub-rule), tassel, corpuscle, muscle, tussle) (/t/ surfaces in apostolic, castellan, castellated, epistolary – see section 7.2); also, with a preceding long vowel (in RP), castle; also occurs before final /ən/ spelt <-en>, but this is an even smaller set: glisten, listen and the derived form christen with preceding short vowels, fasten with a preceding long vowel (in RP), and the derived forms chasten, hasten, moisten with preceding long vowels or diphthongs. The only other examples of medial /s/ spelt <st> within stem words appear to be forecastle in either of its pronunciations /ˈfəʊksəl, ˈfɔːkaːsəl/, mistletoe, ostler. In nestle, wrestle, christen, chasten, hasten, moisten, fasten, /t/ has been lost at a morpheme boundary. Other examples of compounds with lost /t/ so that <st> spells /s/ are chestnut, Christmas, durstn’t, dustbin, dustman, mustn’t, waistcoat /ˈweɪskəʊt/ and sometimes ghastly. |
This loss of /t/ at a morpheme boundary is one small aspect of a very frequent process which is too widespread and complicated to tackle in this book, focused as it mainly is on citation forms of stem words – see Appendix A, section A.1 |
||
|
<sth> |
only in asthma, isthmus if pronounced without /θ/ |
|
<sw> |
only in answer, coxswain, sword /ˈɑːnsə, ˈkɒksən, sɔːd/ and boatswain pronounced /ˈbəʊsən/ (also pronounced /ˈbəʊtsweɪn/) |
|
<t> |
only the penultimate <t> in about 10 words ending in <-tiation>, e.g. differentiation, initiation, negotiation, propitiation, transubstantiation, and only for RP-speakers who avoid having two occurrences of medial /∫/ in such words (see Notes under /∫/, section 3.8.3). In French, on the other hand, <t> is one of the most frequent correspondences for /s/ |
|
<z> |
only in blitz(krieg), chintz, ersatz, glitz, howitzer, kibbutz, kibitz, klutz, lutz, pretzel, quartz, ritz, schmaltz, schnitzel, seltzer, spritz(er), Switzerland, waltz, wurlitzer |
Other 2-phoneme graphemes |
/ts/ spelt <z, zz> |
see under /t/, section 3.5.7 |
|
/ks/ spelt <xe> |
only in annexe, axe. See comments under /k/, section 3.7.1 |
|
/ks/ spelt <xh> |
only in exhibition, exhortation, exhumation – for exhibit, exhort, exhume see under /g/, section 3.5.3 |
3-phoneme grapheme |
/eks/ spelt <x> |
see under /k/, section 3.7.1 |
Notes
For the very few occasions when <ss> is reduced to <s> in compound words see section 4.4.7.
/s/ is the phonological realisation of various grammatical suffixes (regular noun plural and third person singular person tense verb (both spelt <s> where the stem ends in <e>, otherwise <es>), regular singular and irregular plural possessive (spelt <’s>)), and of is, has when contracted (also spelt <’s>), after any voiceless non-sibilant consonant (/p t k f θ/). As just shown, in all these cases the spelling contains <s>.
However, because /s/ itself is a sibilant consonant, adding any of the suffixes just listed to a stem ending in /s/ adds a syllable /ɪz/ as well as a morpheme: horses, fusses, Brooks’s. On this and the topic of the previous paragraph see also /z/, section 3.7.8, and /ɪ/, section 5.4.3.
Because /s/ is almost as divergent as /k/, a further analysis of the major spellings of /s/ is given in Tables 3.4-5. As with /k/, it is unphonological but true that it’s easier to state all the initial and medial correspondences of /s/, and some of the stem-final ones, in terms of following letters rather than following phonemes. (For an attempt to do it phonologically see Carney, 1994: 234-6).
There are several words in <-se> in which the <e> appears redundant since the <s> alone would spell /s/ and the <e> is not part of a split digraph, namely carcase (also spelt carcass), purchase; mortise (also spelt mortice), practise, premise (also spelt premiss), promise, treatise (cf. thesis); purpose; porpoise, tortoise; apocalypse, apse, collapse, eclipse, elapse, ellipse, glimpse, prolapse, relapse, traipse. In copse, corpse, lapse the <e> is equally redundant phonographically but serves to differentiate these words visually from cops, corps, laps.
The virtual non-existence of <-oce> and the rarity of <-oss> as spellings for word-final /əʊs/ mean that <-ose> is almost entirely predictable as the spelling of this stem-word ending. However, this seems to be one of very few examples where a pattern of this sort is more reliable than predictions from the separate phonemes. On this see also section A.7 in Appendix A.
For the medial /s/ in eczema see section 6.10.
In each main box below, the regular spelling is stated at the top in bold.
For other exceptions see the 2- and 3-phoneme graphemes and Oddities above.
Initial |
medial |
|
/s/ not before <e>, <i>, <y> |
<s>, e.g. sale, scale, skull, slime, smooth, snake, soap, spill, still, suave, swede Exceptions (all with <c>): caecum, caesium, caesura, coelacanth, coelenterate, coeliac, coelom, coenobite, coenocyte (there are none where the next phoneme/ letter is a consonant) |
<s>, e.g. descant, askance, asleep, dismiss, consonant, dyspnoea, disaster (second <s>), persuade, aswill Exceptions (there are none where the next letter is a consonant, but those in capitals in this list are exceptions where the next phoneme is a consonant): apercu, facade (lacking French cedillas); ambassador, assail, assassin (first <ss>), assault, assay, cassava, commissar, dissatisfy, essay, massacre, pessary, reconnaissance, renaissance; hassle, tussle; associate, assonance, assorted, bassoon, blossom, caisson, dissociate, dissolute, dissonant, lasso, lesson, lissom, voussoir; alyssum, ASSUAGE, assume pronounced /əsˈjuːm/, DISSUADE; also EMISSARY, NECESSARY, PROMISSORY with elided vowels (see section 6.10) |
/s/ before <e>, <i>, <y> |
<c>, e.g. ceiling, city, cyclic Exceptions (all with <s>): sea, seal, seam, seance, sear, search, season, seat, sebaceous, sebum, secant, sec-ateurs, secede, seclude, second, secret(e), secretary, sect(ion), secular, secure, sedan, sedate, sedentary, sedge, sediment, sedition, seduce, see, seed, seek, seem, seep, seethe, segment, segregate, segue, self, sell, semantic, semaphore, semblance, semen, semi, seminar, semiotics, semite, semolina, senate, send, senile, senior, sense, sensual, sent-ence, sentient, sentiment, sentinel, sentry, |
<c>, e.g. accept, decide, bicycle Exceptions: Words ending in /sɪs/ are spelt <-sis>, e.g. Sis (abbreviation of sister), thesis (only exception: diocese) Words ending in /sɪtiː/ preceded by a consonant letter or by /ɒ/ spelt <o> are spelt <-sity>, e.g. adversity, density, diversity, falsity, immensity, intensity, pervers-ity, propensity, sparsity, university, varsity; animosity, curiosity, generosity, impetuosity, verbosity, virtuosity, viscosity (exceptions: scarcity; atrocity, ferocity, precocity, reciprocity, velocity; and cf. city itself) |
sepal, separate, sept, septic, sepulchre, sequel, sequence, seques-ter, sequin, seraph, serenade, serene, serf, serge, sergeant, serial, series, serious, sermon, serpent, serrated, serum, serve, sesame, session, set, settle, seven, sever, severe, sew, sewer (in both pronun-ciations and meanings), sex; sibilant, sibling, sibyl, sick, side, sidereal, sidle, siege, siesta, sieve, sift, sigh, sight, sign, Sikh, silage, silent, silica, silk, sill, silly, silo, silt, silver, simian, similar, simmer, simper, simple, simultan-eous, sin, since, sincere, sine, sinew, sing, singe, single, sinister, sink, sinuous, sip, siphon, sir, sire, siren, sisal, sister, sit, sitar, site, situation, six, size, sizzle; sybarite, sycamore, syce, sycophant, syllable, syllogism, sylph, symbiosis, symbol, symmetry, sympathy, symphony, symposium, symptom, synaesthesia, synagogue, synapse, synch(ro-), synergy, synod, synonym, synopsis, syntax, synthesis, syphilis, syphon, syringe, syrup, system, syzygy |
Words ending in /sɪv/: if preceded by a short vowel spelt with one letter they are spelt <-ssive> (e.g. massive), otherwise <-sive> (e.g. adhesive) (but N.B. sieve itself) Other exceptions: with <s>: abseil, absent, arsenal, arsenic, beseech, consecutive, consensus, consent, consequen-ce/t, corset, counsel, desecrate, disembark, dysentery, insect, morsel, prosecute, transept; basin, consist, disinfect, disinherit, misinform, misinterpret, parasite, transit; asylum; apostasy, argosy,controversy, courtesy, ecstasy, fantasy, greasy, heresy, hypocrisy, idiosyncrasy, jealousy, leprosy, minstrelsy, pleurisy, prophesy (verb, pronounced /ˈprɒfɪsaɪ/ - the noun prophecy, pronounced /ˈprɒfɪsiː/, has regular <c>); autopsy, biopsy, catalepsy, curtsy, dropsy, epilepsy, gipsy, narcolepsy, necropsy, tipsy with <ss>: antimacassar, assegai, assemble, assent, assert, assess, asset, casserole, cassette, connoisseur, cussed (‘stubborn’), delicatessen, dissect, dissemble, disseminate, dissension, dissent, dissertation, disservice, essence, essential, fricassee, lessen, masseu-r/se, mussel, necessity, trousseau; admissible, assassin (second <ss>), assiduous, assign, assimilate, assist, assize, brassica, bassinet, chassis, classic, classify, dissident, dissimilar, dissipate, fossil, gossip, jurassic, lassitude, messiah, permissible, (im)possible, potassium, prussic, triassic; embassy, hussy |
For /ks/ see above and under /k/, section 3.7.1. For /s/ as a grammatical suffix, see above.
Categories listed in the left-hand column below apply to both monosyllables and polysyllables except where stated.
For exceptions besides those in the Table, see the 2- and 3-phoneme graphemes and Oddities above.
In mono-syllables after /ɑː/ spelt <a> and after a short vowel spelt with one letter: <ss> |
Examples: brass, class, glass, grass, pass; ass, bass (/bæs/ ‘fish’), crass, lass, mass; bless, cess, chess, cress, dress, guess, less, mess, ness, press, stress, tress; bliss, hiss, kiss, miss, p*ss; boss, cross, doss, dross, floss, loss, moss; buss, cuss, fuss, muss, truss; puss. Exceptions: gas, yes, Sis (abbreviation of sister), this, bus, plus, pus, thus, us Extension: There appears to be only one other one-syllable stem word in which a long vowel/diphthong is spelt with a single letter before word-final /s/: bass (/beɪs/ ‘(player of) large stringed instrument’/‘(singer with) low-pitched voice’) |
After /n/: <ce> |
Examples: (monosyllables) dance, chance, glance, lance, prance, stance, trance; fence, hence, pence, thence, whence; mince, prince, quince, since, wince; nonce, once, sconce; bounce, flounce, ounce, pounce, trounce; dunce; (polysyllables) abundance, evidence and hundreds of other words ending <-ance/-ence>, convince, evince, province, ensconce, an/de/(mis)pro-nounce Exceptions: (monosyllables) manse; dense, sense, tense; rinse; (polysyllables) expanse; condense, dispense, expense, immense, incense (noun and verb), intense, license, nonsense, recompense, suspense; response |
After /l, p/: <se> |
Examples: (monosyllables) dulse, else, false, pulse; apse, copse, corpse, glimpse, lapse, traipse; (polysyllables) convulse, impulse, repulse; apocalypse, eclipse, ellipse, col/e/pro/re-lapse Exceptions: (none) |
After <r> forming part of a vowel di-/tri-graph: <se> |
Examples: (monosyllables) arse, coarse, course, curse, Erse, gorse, hearse, hoarse, horse, morse, Norse, nurse, purse, sparse, terse, verse, worse; (polysyllables) adverse, averse, concourse, converse, discourse, disburse, disperse, diverse, endorse, immerse, intercourse, intersperse, inverse, obverse, recourse, rehearse, reimburse, remorse, reverse, traverse, transverse, universe Exceptions: (monosyllables) farce, fierce, force, pierce, scarce, source, tierce; (polysyllables) commerce, divorce, enforce, reinforce, perforce, resource |
After other vowel digraphs: <se> |
Examples: (monosyllables) cease, crease, grease, lease, geese, goose, loose, moose, noose, douse, grouse, house (noun), louse, mouse; (polysyllables) decease, de/in-crease, release, porpoise, tortoise, caboose, papoose, vamoose Exceptions: (monosyllables) sauce, peace, fleece, deuce, niece, piece, choice, voice, juice, sluice; gneiss; (polysyllables) invoice, rejoice |
In words ending /əʊs/: <s> within split digraph <o.e> spelling /əʊ/ (see also Notes above Table) |
Examples: (monosyllables) close (adjective/noun), dose; (polysyllables) the ‘sugar’ words dextrose, glucose, lactose, sucrose (all of which have alternative pronunciations in /əʊz/), and adjectives comatose, lachrymose, morose, verbose and dozens of others (see also section A.7 in Appendix A) Only exceptions: Groce (rare surname); gross, engross |
Where any other long vowel, diphthong or /juː/ is spelt with a split digraph: <ce>, such that the <e> functions as part of both graphemes (for dual-functioning see section 7.1) |
Examples: (monosyllables) ace, brace, dace, face, lace, grace, mace, pace, place, race, space, trace; dice, ice, lice, mice, nice, price, rice, slice, spice, splice, thrice, trice, twice, vice; puce, spruce, truce, syce; (polysyllables) apace, de/ef/out/re-face, disgrace, embrace, en/inter-lace, dis/mis/re-place, retrace; advice, caprice, device, entice, police, sacrifice, suffice; Exceptions: (monosyllables) base, case, chase; use (noun); (polysyllables) a/de-base, encase; obese; concise, paradise, precise; abstruse, obtuse, recluse; also merchandise, abuse, excuse, refuse as nouns and diffuse as an adjective |
In stressed final syllables of polysyllables: <ss> |
Examples: abyss, address, amiss, assess, caress, confess, discuss, dismiss, distress, duress, excess, express, impress, morass, possess, process (verb), profess, progress (verb), prowess, recess, redress, repress, remiss, success Exceptions: none (?) |
In polysyllables ending in unstressed /sɪs/: <s> |
Examples (from many that could be given): Only exception: diocese |
In other cases of final unstressed /ɪs/ in polysyllables: <ce> |
Examples (from many that could be given): apprentice, auspice, chalice, justice, practice Exceptions: axis, cannabis, marquis, metropolis, pelvis; practise, premise, promise, treatise; premiss N.B.: mortice, mortise has both spellings, and cf. Latin (rigor) mortis (‘(stiffness) of death’) |
In other unstressed final syllables of polysyllables: <s> |
Examples (from many that could be given): bias, canvas, corpus, cosmos, fabulous, horrendous, rickets, syllabus, tonsilitis, virus Exceptions: <se> only in carcase, purchase, purpose <ss> in abscess, access, albatross, blunderbuss, buttress, canvass, carcass, compass, congress, cutlass, egress, embarrass, empress, harass, ingress, isinglass, mattress, all the compounds of press, process (noun), progress (noun), trespass, windlass |
The main system
Basic grapheme |
<v> |
98% |
e.g. oven |
Other frequent grapheme |
<f> |
only in of, and roofs pronounced /ruːvz/ (neither counted in percentages) |
|
Doubled spelling |
<ve> |
2% |
regular in word-final position, e.g. give, have, positive. Exceptions: bruv, chav, derv, gov, guv, lav, leitmotiv, of, rev, satnav, shiv, Slav, sov, spiv; <ve> also spells /v/ in average, deliverable, evening (noun, ‘late part of day’, pronounced /ˈiːvnɪŋ/, as distinct from the verb of the same spelling, ‘levelling’, pronounced /ˈiːvənɪŋ/), every, leverage, several, sovereign – cf. section 6.10 – but is very rare medially in stem words (but see Notes) and never occurs initially |
The rest
Doubled spelling + <e> |
(cannot occur because doubled spelling already ends in <e>) |
|
Oddities |
<1% in total |
|
|
<bv> |
only in obvious pronounced /ˈɒviːjəs/ |
|
<ph> |
only in nephew pronounced /ˈnevjuː/ (also pronounced /ˈnefjuː/), Stephen (also spelt Steven) |
|
<vv> |
only medial and only in bevvy, bovver, chavvy, chivvy, civvy, divvy, flivver, lavvy, luvv-ie/y, navvy, revving, savvy, skivvy, spivvery, spivvy |
2-phoneme graphemes |
(none) |
Notes
<vv> is very rare (it occurs only in the words just listed, most of which are slang), and in word-final position <ve> functions in its place as the doubled spelling of /v/ (see next paragraph).
English spelling has a tacit rule that words must not end in <v>, except for a few slang and foreign words and modern abbreviations (see above under ‘Doubled spelling’). Word-finally, therefore, the regular spelling of /v/ is <ve>, which occurs in at least 1,000 words. At least 700 of these are polysyllabic adjectives and nouns ending in unstressed /ɪv/ spelt <-ive>, e.g. adjective, endive, expletive, gerundive, massive, narrative, olive, relative. (A century ago Dewey tried, but failed, to have the phonographically redundant final <e> removed from the US spelling of these words; it is equally redundant in all the non-split digraph categories mentioned below, but is probably even more resistant to change there. He might have had more success if he had advocated removal of the redundant final <e> in words ending in <-ate, -ite> where the <e> is also not part of a split digraph – for these words see sections 3.5.7 and 9.34 – or any of the other graphemes with redundant final <e>, of which there are many).
Among the small number of remaining polysyllables are the preposition above, the noun octave, and groups of words with:
1) /l/ spelt <l> preceding <ve>, e.g. dissolve, evolve
2) vowel digraphs preceding <ve>, e.g. bereave; receive and several other words in <-ceive>; deserve; achieve, believe, relieve, reprieve, retrieve
3) split digraphs where the <e> is part of both the digraph and <ve>, e.g. behave, conclave, forgave; alive, archive, arrive, deprive, naive, ogive, recitative, revive, survive; alcove, mangrove.
Among monosyllables with final /v/ spelt <ve> are those with
1) a short vowel phoneme immediately preceding <ve>, namely have; give, live (verb, /lɪv/); dove, glove, love, shove; and – with its unusual digraph spelling of a short vowel – sieve
2) /l/ spelt <l> preceding <ve>, e.g. salve pronounced /sælv/), valve; delve, shelve, twelve; solve
3) vowel digraphs preceding <ve>, e.g. waive; calve, halve, salve pronounced /sɑːv/; carve, starve; mauve; greave, heave, leave; sleeve; nerve, serve; grieve, thieve; groove; curve
4) split digraphs where the <e> is part of both the digraph and <ve>, e.g. gave, shave, suave, wave; breve, eve; drive, five, hive, jive, live (adjective, /laɪv/), swive, wive; cove, drove, move, prove; gyve.
The point of this long analysis has been to show in how few words ending in a single vowel letter plus <ve> the <e> forms digraphs both with the <v> and with the vowel letter (for dual-functioning see section 7.1). In this respect <ve> is very unlike the other principal word-final consonant digraphs formed with <e>, namely <ce, ge, se>.
Medially, <ve> occurs in:
- the few words mentioned above under ‘Doubled spelling’
- a large number of regular plural nouns and singular verbs, e.g. haves (vs have-nots), gives, grieves, initiatives, dissolves, lives (verb), loves, improves, preserves, mauves
- a small number of irregular plural nouns ending in /vz/ spelt <-ves> where the singular forms have /f/ spelt <f>, namely calves, dwarves, elves, halves, hooves, leaves, loaves, scarves, (our/your/them-)selves, sheaves, shelves, thieves, turves, wharves, (were)wolves. On 4/1/15 the form behalves appeared in the Observer; various websites decry this form as obsolete or unnecessary
- a very few similar words where the <f> in the singular is within the split digraph <i.e>: knives, lives (/laɪvz/; the singular verb of the same spelling is pronounced /lɪvz/), (ale/good/house/mid-)wives (but if housewife ‘sewing kit’ pronounced /ˈhʌzɪf/ has a plural it is presumably pronounced /ˈhʌzɪfs/).
Why the irregular words in the third category in this list with no <e> in the singular have an <e> in the plural is unclear (i.e. they could be spelt *elvs, *leavs, etc.), unless verbs like calve, halve, leave, salve, shelve, thieve and their singular forms calves, halves, leaves, salves, shelves, thieves influence the plural nouns, or the strong prohibition on word-final <v> (see above) extends to stem-final position in plurals.
A couple of the nouns just listed have alternative, regular plurals: dwarfs, turfs, and the only plural of lowlife is said to be lowlifes (though I have seen the form lowlives in print).
Conversely, roofs, which (officially) has only that regular plural spelling, has both the regular pronunciation /ruːfs/ and the irregular pronunciation /ruːvz/, but the latter pronunciation is hardly ever recognised in the spelling as *rooves (though this form has been printed twice in The Guardian: (1) 18 July 2009, main section, p.39 (in a puzzle); (2) 26 September 2009, Review section, p.7 (in a poem); internet exploration revealed various people wondering if *rooves or roofs was the plural spelling because they say /ruːvz/). This pronunciation and the spelling roofs provide the only other example, besides of, of /v/ spelt <f>.
The main system
For all these categories see Notes.
Basic grapheme |
<z> |
5% (with <ze> - |
regular in word-initial position, e.g. zoo (only exceptions: sorbet pronounced /ˈzɔːbeɪ/ (also pronounced /ˈsɔːbeɪ/), sauerkraut if pronounced with German /z/, and the words in <x-> listed below); medially, only in amazon, azyme, bazooka, bedizen, benz-ene/ol, bezel, bezique, blazer, bombazine, bonanza, brazen, cadenza, chimpanzee, coryza, crazy, denizen, enzyme, extravaganza, frozen, gazebo, gazump, gizmo, (hap)hazard, hetero-/mono-zygous, influenza, lazy, lizard, magazine, mazurka, muzak, ozone, phizog, plaza, protozoa, razor, samizdat, schnauzer, spermatozoon, stanza, teazel/teazle (also spelt teasel), trapez-ium/oid, vizard, vizier, vizor (also spelt with <s>), wizard, wizen(ed), zigzag; |
word-finally without a following <e> only in fez, phiz, quiz, topaz, whiz, the abbreviated compound word showbiz and a few other very rare words; phonemically word-final within split digraphs, only in amaze, blaze, craze, daze, gaze, glaze, graze, haze, laze, maze, raze; trapeze; cloze, doze, froze, plus the four nouns ending in /aɪz/ always spelt <-ize> (assize, capsize, prize, size) and the large number of verbs ending in /aɪz/ spelt <-ize> (almost all of which have alternative spellings in <-ise>) |
|||
Other frequent graphemes |
<s> |
93% (with <se>) |
word-initial only in sorbet if pronounced /ˈzɔːbeɪ/and sauerkraut if pronounced with German /z/ Regular (1) in medial position, e.g. chisel, preside, seismic, talisman; (2) word-finally (see above for exceptions with <z>, and below for all other exceptions); (3) in various suffixes and contracted forms after a vowel or non-sibilant voiced consonant – see Notes |
<se> |
never word-initial or -medial (except medially in compound words, e.g. gooseberry /ˈgʊzbriː/), housewife ‘sewing kit’, pronounced /ˈhʌzɪf/; also in miserable if pronounced /ˈmɪzrəbəl/ - see section 6.10); regular word-finally in content words after a long vowel or diphthong spelt with a digraph, e.g. blouse, bruise, cause, chauffeuse, cheese, choose, drowse, noise, parse, please, raise |
||
Doubled spelling |
<zz> |
<1% |
regular at the end of one-syllable words after a short vowel spelt with one letter – but see Notes; also regular before word-final /əl/ spelt <-le> after a short vowel spelt with one letter, e.g. dazzle (see section 4.3.3); otherwise only in blizzard, buzzard, dizzy, fizzog, gizzard, grizzly, mizzen, muezzin, muzzy, pizzazz, razzmatazz, scuzzy, snazzy, tizzy, wazzock. For <zz> arising from consonant-doubling before a suffix see section 4.2 |
The rest
Doubled spelling + <e> |
(does not occur) |
|
Oddities |
2% in total |
|
|
<cz> |
only in czar(ina) |
|
<sc> |
only in crescent pronounced /ˈkrezənt/ (also pronounced /ˈkresənt/) |
|
<ss> |
only medially and only in Aussie, brassiere, dessert, dissolve, hussar, Missouri, possess, scissors |
<ts> |
only in tsar |
|
|
<x> |
word-initially, only in some words of Greek origin, namely xanthine, xanthoma, xanthophyll, xenon, xenophobia and several other words beginning xeno-, Xerox and several other words beginning xero-, xylem, xylene, xylophone and several other words beginning xylo-. Medially, only in anxiety pronounced /æŋˈzaɪjɪtiː/ (also pronounced /æŋˈgzaɪjɪtiː/). French loanwords ending in <-eau> are sometimes in the plural written French-style with /z/ spelt <x> rather than <s>, e.g. |
indeed, my dictionary gives only the <x> form in bandeaux, chateaux, rondeaux, tableaux. In my opinion the <x> form is outmoded and unnecessary |
||
|
<ze> |
only word-final and only in adze, bronze after consonant phonemes, plus baize, booze, breeze, freeze, frieze, furze, gauze, maize, ooze, schmooze, seize, sleaze, sneeze, snooze, squeeze, wheeze after long vowels or diphthongs spelt by digraphs - in all these words the <e> is phonographically redundant but spellings without it would look odd. In the hundreds of verbs ending /aɪz/ which may or must be spelt with <-ize>, plus the few other stem words where <z> appears within split digraphs (see above), it is unnecessary to analyse the <z> as also being part of digraph <ze> because the <z> spells /z/ without the <e> - but see Notes |
2-phoneme graphemes |
/gz/ |
<1% see under /g/, section 3.5.3 |
|
/ɪz/ |
only, following an apostrophe, in regular singular and irregular plural possessive forms ending in a sibilant consonant (/s, z, ∫, ʒ, ʧ, ʤ/), e.g. Brooks’s (book), jazz’s (appeal), Bush’s (government), (the) mirage’s (appearance), (the) Church’s (mission), (the) village’s (centre), (the) geese’s (cackling). See Notes |
Notes
/z/ is the phonological realisation of various grammatical suffixes (regular noun plural and third person singular person tense verb (both spelt <s> where the stem ends in <e>, otherwise <es>), regular singular and irregular plural possessive (spelt <’s>)), and of is, has when contracted (also spelt <’s>), after any vowel or voiced non-sibilant consonant (/b d g l m n ŋ v ð/). As just shown, in all these cases the spelling contains <s>.
However, because /z/itself is a sibilant consonant, adding any of the suffixes just listed to a stem ending in /z/ adds a syllable /ɪz/ as well as a morpheme: fuses, quizzes, jazz’s. On this and the topic of the previous paragraph see also /s/, section 3.7.6, and /ɪ/, section 5.4.3.
<zz> is regular word-finally in one-syllable words after a short vowel spelt with one letter, but there are only seven words in this set (buzz, fizz, frizz, fuzz, jazz, tizz, whizz, plus two polysyllables: pizzazz, razzmatazz) and there are 10 counter-examples (as, fez, has, his, is, phiz, quiz, was, whiz and cos, the abbreviation of because; cos, the lettuce and the abbreviation of cosine, vary in pronunciation between /kɒz/ and /kɒs/) – but three of the counter-examples (has, is, was) are verbs following the rule that grammatical endings in /z/ are spelt <s>.
<zz> is also regular before word-final /əl/ spelt <-le> after a short vowel spelt with one letter (see section 4.3.3), and very rare other than in this and the context mentioned in the preceding paragraph.
<z> is regular in word-initial position, and very rare elsewhere, except in one class of verb stem endings. In British English, almost all the verbs whose stems end in /aɪz/ can be spelt with either <-ise> or <-ize>, e.g. atomise/atomize. There are a few which can only be spelt with <-ize>, namely capsize, prize, size, and a larger group which can only be spelt with <-ise>, namely advertise, advise, apprise, chastise, circumcise, comprise, compromise, despise, devise, enterprise, excise, exercise, franchise, improvise, incise, merchandise, premise (/prɪˈmaɪz/ ‘base argument upon’), prise, realise, revise, rise, supervise, surmise, surprise, televise.
Exceptions to the <-ise/-ize> choice are a few verbs which are mainly spelt with <-yse> in British English (though US spellings in <-yze> are becoming commoner): analyse, breathalyse, catalyse, dialyse, electrolyse, paralyse. And plural nouns and singular verbs whose stems end in /aɪ/ and which therefore end in /aɪz/ when suffixed (e.g. dies, dyes, lies, vies) follow different rules.
Of the 15 words ending /eɪz/ spelt with a split digraph, 11 are spelt with <-aze> (amaze, blaze, craze, daze, gaze, glaze, graze, haze, laze, maze, raze), and 4 with <-ase> (erase, phase, phrase, ukase). Vase appears to be the only word ending in /ɑːz/ (in RP).
The only word in which word-final /iːz/ is spelt <-eze> is trapeze, and cerise, chemise, expertise, reprise, valise appear to be the only ones in which it is spelt <-ise>. Besides these spellings in which /iː/ is spelt by a split digraph (and in theory /z/ is spelt only by the <z>, though it is immaterial whether one instead recognises the <e> as part of digraph <ze> and therefore as dual-functioning), word-final /iːz/ has four further spellings in which /iː/ is spelt by a non-split digraph, and /z/ definitely by either <ze> (breeze, freeze, sneeze, squeeze, wheeze; frieze) or <se> (appease, (dis)ease, (dis)please, heartsease, pease, tease; cheese). In all other cases, of which there are dozens, especially nationality/language words, plus these, word-final /iːz/ is spelt <-ese>. This may seem to be one of the few cases where the spelling of a final /VC/ pattern is more predictable as a unit than from correspondences for its two phonemes, but actually it is predictable from them: the regular spelling of /iː/ in closed final syllables of polysyllables is <e.e> (see section 5.7.2), and the regular spelling of phonemically word-final /z/ is <s>. See also section A.7 in Appendix A.
There are only 2 words ending /əʊz/ spelt <-oze>: doze, froze; the rest are spelt with <-ose>, e.g. chose, close (verb), hose, pose, prose, rose, those.
There are no words ending /(j)uːz/ spelt <-uze>, and only 1 spelt <-ose>: lose; the rest are spelt with <-use>, e.g. abuse (verb), accuse, amuse, bemuse, excuse (verb), enthuse, (con/dif/ef/in/suf-)fuse, hypotenuse, muse, peruse, refuse (verb), ruse, use (verb).
In all cases other than those listed, the regular spelling of medial and final /z/ is <s>, including the grammatical suffixes mentioned above.
3.8 Consonants without doubled spellings: /h ŋ ∫ ʒ θ ð w j/
Occurs only before a vowel phoneme, therefore never word-finally.
The main system
Basic grapheme |
<h> |
97% |
e.g. behave, have |
Other frequent graphemes |
(none) |
The rest
Oddities |
<j> |
only in fajita, jojoba (twice), marijuana, mojito, Navajo |
|
<wh> |
3% - only in who, whom, whose, whole, whoop(er/ing), whore |
2-phoneme graphemes |
(none) |
Notes
/h/ is rare medially, but cf. adhere, behave, behind, bohemian, cahoots, clerihew, cohere, cohort, enhance, inhabit, inherit, mayhem, perhaps pronounced /pəˈhæps/ rather than /præps/, prehensile, shanghai, and compound words such as anyhow, meathook, mishap, mishit, peahen, poorhouse, prehistoric, sawhorse, sunhat, warhead, warhorse.
Because Carney (1994) did not include function words such as who, whom, whose in his frequency counts, his percentage for /h/ spelt <wh> is distinctly lower than if he had included them.
Occurs only post-vocalically, and therefore never word-initially (in English). Also, except in very rare cases such as spraing, occurs only after short vowel phonemes (and even then never after /ʊ/ (in RP); also very rare after /e, ə/ - see Notes)
The main system
Basic grapheme |
<ng> |
75% |
e.g. bang, sing, zing, long, lung |
Other frequent grapheme |
<n> |
25% |
before /k, g/, however spelt, e.g. sink, zinc, anxious, conquer, ankle, uncle, length; longer, kangaroo, anxiety. See Notes |
The rest
Oddities |
<nc> |
only in charabanc /ˈʃærəbæŋ/ |
|
<nd> |
only in handcuffs, handkerchief /ˈhæŋkʌfs, ˈhæŋkəʧɪf/ |
|
<ngh> |
only in dinghy, gingham, Singhalese /ˈdɪŋiː, ˈgɪŋəm, sɪŋəˈliːz/ (contrast shanghai /ʃæŋˈhaɪ/) |
|
<ngu> |
only in a very few suffixed forms of words in next category, namely haranguing, tonguing. See also end of section 6.4 |
|
<ngue> |
only in harangue, meringue, tongue /həˈræŋ, məˈræŋ, tʌŋ/ (contrast dengue /ˈdeŋgeɪ/) |
2-phoneme graphemes |
(none) |
Notes
The conclusion that /ŋ/ before /k/ is spelt <n> is based on words like ankle /ˈæŋkəl/, carbuncle, crinkle, peduncle, periwinkle, rankle, sprinkle, tinkle, twinkle, uncle, winkle, wrinkle, where /k/ is clearly spelt <k, c>, so that the preceding /ŋ/ must be represented by the <n>. Then the same analysis must apply to angle /ˈæŋgəl/, even though this means that here the letters <n, g> do not form the grapheme <ng> and do not jointly spell /ŋ/. The same applies to finger /ˈfɪŋgə/, but in singer /ˈsɪŋə/ there is no /g/ (in RP, though there is in Lancashire), so that here the letters <ng> do form a single grapheme representing /ŋ/.
The words length, lengthen, strength, strengthen pronounced /leŋkθ, ˈleŋkθən, streŋkθ, ˈstreŋkθən/ (for their alternative pronunciations see under /n/, section 3.5.5) and angst /æŋkst/ appear to need an analysis in which /ŋ/ is spelt <n> and /k/ is spelt <g> - if so, these words and disguise /dɪsˈkaɪz/, disgust pronounced /dɪsˈkʌst/, i.e. identically to discussed (disguise, disgust are also pronounced /dɪzˈgaɪz, dɪzˈgʌst/, i.e. with both medial consonants voiced rather than voiceless) would be the only occurrences of /k/ spelt <g>, though the spelling of /ŋ/ with <n> in angst, length, lengthen, strength, strengthen conforms to the analysis of the many words with this correspondence just given (see also under /k/ in section 3.7.1).
The words length, lengthen, strength, strengthen are also among the very few in which /ŋ/ occurs after /e/. The only other examples seem to be dengue, dreng ‘free tenant in ancient Northumbria’, enchiridion, enclave pronounced /ˈeŋkleɪv/ (also pronounced /ˈɒŋkleɪv/), enkephalin, ginseng, nomenclature, the abbreviation SENCo (‘Special Educational Needs Coordinator’) pronounced /ˈseŋkəʊ/, and words beginning encephal- when pronounced /eŋˈkefəl-/ (also pronounced /enˈsefəl-/). The only cases in which /ŋ/ follows /ə/ may be words like concur(rent) if pronounced /kəŋˈkɜː, kəŋˈkʌrənt/.
Although long, strong, young end in /ŋ/ (in RP) and are therefore to be analysed as containing /ŋ/ spelt <ng>, the comparative and superlative forms longer, longest, stronger, strongest, younger, youngest and the verb elongate all have medial /ŋg/, so here the /g/ has ‘surfaced’ and is represented by the <g> (see section 7.2), and the /ŋ/ is spelt <n>. Similarly with diphthongise, prolongation, which gain a /g/ relative to unsuffixed diphthong, prolong. By contrast, in longevity /lɒŋʴʤevɪtiː/ the surfacing phoneme is /ʤ/.
The word anxiety has two pronunciations: /æŋˈgzaɪɪtiː, æŋˈzaɪɪtiː/, where the second lacks /g/. Both are rarities: the first is the only instance of /ŋg/ where the /g/ is not spelt <g>; the second is the only case where /ŋ/ is spelt <n> without a following /k, g/.
The main system
Basic grapheme |
<sh> |
37% |
e.g. ship, fish; regular in initial and final positions; rare medially, but cf. ashet, baksheesh, banshee, bishop, buckshee, Bolshevik, bolshie, bushel, cashier, cashmere, cushion, dasheen, dishevel, fashion, geisha, kosher, kwashiorkor, marshal, pasha, pashmina, ramshackle, sashay, worship, yashmak and words with the suffix -ship. For exceptions in initial and final positions see the Oddities, below. Also see Notes |
Other frequent graphemes |
All these graphemes occur only medially |
For all these categories see Notes and Table 3.6 |
|
|
<ti> |
55% (with <ci, |
regular medially, e.g. nation, but there are many exceptions |
|
<ci> |
e.g. commercial, crucial, delicious, judicial, logician, magician, official, racial, social, special |
|
|
<si> |
e.g. aversion, emulsion, pension, repulsion, reversion, tension, torsion, version |
|
|
<ssi> |
e.g. accession, admission, discussion, emission, intercession, obsession, passion, percussion, permission, recession, remission, session |
|
Rare grapheme |
<ce> |
regular medially in /eɪ∫əs/ spelt <-aceous>, e.g. cretaceous, curvaceous, herbaceous, sebaceous – see Notes and Table 3.6; otherwise only in cetacean, crustacea(n), Echinacea, ocean, siliceous |
The rest
Oddities |
8% in total |
|
- in initial position |
||
|
<ch> |
only in 30+ words of mainly French origin, namely chagrin, chaise, chalet, chamfer pronounced /ˈʃæmfə/ (also pronounced /ˈʧæmfə/), chamois (whether pronounced /ˈʃæmiː/ or /ˈʃæmwɑː/), champagne, chancre, chandelier, chaperone, charabanc, charade, chardonnay, charlatan, Charlotte, chassis, chateau, chauffeu-r/se, chauvinis-m/t, chef, chemise, chenille, cheroot, chevalier, chevron, chi-chi (twice), chic, chicane(ry), chiffon, chignon, chivalr-ic/ous/y, chute |
|
<s> |
only in sugar, sure and (German pronunciations of) spiel, stein, strafe, stumm |
|
<sch> |
only in schedule (also pronounced with /sk/), schemozzle, schist, schistosomiasis, schlemiel, schlep, schlock, schmaltz, schmo(e), schmooze, schnapps, schnauzer, schnitzel, schnozzle, schuss, schwa |
|
<sj> |
only in sjambok |
- in medial position |
||
|
<c> |
e.g. officiate, speciality, specie(s), superficiality and sometimes ap/de-preciate, associate. See Notes |
|
<ch> |
only in about 20 words of mainly French origin, namely attaché, brochure, cachet, cachou, cliché, crochet, duchesse, echelon, embouchure, Eustachian, machete, machicolation, machine, marchioness, nonchalant, parachute, pistachio, recherché (twice), ricochet, ruching, sachet; also sometimes in (Greek) chiropody (hence the punning shop name Shuropody) |
|
<che> |
only in rapprochement |
|
<chs> |
only in fuchsia |
|
<s> |
only in asphalt pronounced /ˈæ∫felt/ (also pronounced /ˈæsfælt/), censure, commensurate, ensure, insure, tonsure |
|
<sc> |
only in conscie, conscientious, crescendo, fascis-m/t. See Notes |
|
<sch> |
only in maraschino, meerschaum, seneschal |
|
<sci> |
only in conscience, conscious, fascia, luscious. See Notes |
|
<se> |
only in gaseous pronounced /ˈgeɪ∫əs/ (also pronounced /ˈgæsiːjəs/). See Notes |
|
<ss> |
only in assure, fissure, issue, pressure, tissue |
|
<t> |
mainly before <-iate> with the <i> spelling /iː/ (and with ‘invisible’ /j/-glide), e.g. differentiate, expatiate, ingratiate, initiate, negotiate, propitiate, satiate, substantiate, vitiate, plus minutiae, otiose pronounced /ˈəʊʃiːjəʊs, ˈəʊʃiːjəʊz/ (also pronounced /ˈəʊtiːjəʊs, ˈəʊtiːjəʊz/), partiality, ratio; also novitiate pronounced /nəˈvɪʃiːjət/ (also pronounced /nəˈvɪʃət/). See Notes |
- in final position |
||
|
<ce> |
only in liquorice pronounced /ˈlɪkərɪ∫/ |
|
<ch> |
only in Welch and, in phonemically word-final position, fiche, gouache, moustache, niche pronounced /niː∫/, pastiche, quiche, ruche, where the <e> is part of the split digraphs <a.e, i.e, u.e> spelling /ɑː, iː, uː/ |
|
<che> |
only in about 12 words of mainly French origin, namely avalanche, barouche, brioche, cache, cartouche, cloche, creche, douche, farouche, gauche, louche, panache |
2-phoneme graphemes |
/k∫/ |
|
|
(1) spelt <x> |
only in flexure, luxury, sexual /ˈflekʃə, ˈlʌkʃəriː, ˈsekʃ(uːw)əl/ |
|
(2) spelt <xi> |
e.g. anxious: see under /k/, section 3.7.1 |
Notes
Because /∫/ is a sibilant consonant, adding any of the suffixes regular noun plural and third person singular person tense verb (both spelt <s> where the stem ends in <e>, otherwise <es>) and regular singular and irregular plural possessive (spelt <’s>) to a stem ending in /∫/ adds a syllable /ɪz/ as well as a morpheme: quiches, fishes, Bush’s. See also /z/, section 3.7.8, and /ɪ/, section 5.4.3.
Some rules could probably be given for when <sh> is regular medially, but these would be more complicated than giving the list of examples, above.
As spellings of medial /∫/ in stem words, <ti, ce, ci, sci, se, si, ssi> occur only at the beginning of the final syllable of a word and immediately after the stressed syllable, and the final syllable is always one of /əl, ən, əs/ or (very rarely) /əm/ (consortium pronounced /kənˈsɔːʃəm/ (usually pronounced /kənˈsɔːtiːjəm/), nasturtium), /əns/ (conscience) or just /ə/ (e.g. consortia pronounced /kənˈsɔːʃə/ (usually pronounced /kənˈsɔːtiːjə/), fascia, militia); and <si> is always preceded by a consonant letter (except in Asian). Exceptions with these features but other spellings of medial /∫/: bushel, marshal, seneschal; cushion, Eustachian, fashion; fissure, fuchsia, geisha, pressure.
The default spelling of medial /∫/ is <ti>; for example, it is regular in words ending in /eɪ∫ən, eɪ∫əl, iː∫ən, əʊ∫ən, (j)uː∫ən/, e.g. nation, spatial, accretion, lotion, evolution, pollution (exceptions: Asian, racial, cetacean, crustacean, Grecian, ocean, Confucian, Rosicrucian). However, because medial /∫/ has so many other spellings, the major patterns are set out in Table 3.6. What does not come over particularly clearly even then is that the only case where there is substantial three-way confusion is over final /ɪ∫ən/: the majority spelling is <-ition>, e.g. volition, but there is competition (!) from <-ician, -ission> – see the top right-hand and bottom left-hand boxes of the Table (and beware of Titian).
Default spelling: <ti> |
Exceptions (in addition to those in <sh> listed under the basic grapheme and those listed among Oddities above or under Subpatterns and Sub-exceptions below): <ci>: facial, glacial pronounced /ˈgleɪ∫əl/ (also pronounced /ˈgleɪsiːjəl/), racial, (e)special, financial, provincial, social, commercial, crucial; Grecian; academician, electrician, logician, magician, mathematician, mortician, musician, obstetrician, optician, patrician, phonetician, physician, politician, statistician, tactician (N.B. most words in <-ician> are occupations); suspicion; Confucian, Rosicrucian; precious, specious; siliceous; auspicious, avaricious, capricious, delicious, judicious, malicious, meretricious, officious, suspicious, vicious; atrocious, ferocious, precocious and various other rare words <si>: controversial, torsion <ssi>: fission |
|
Each of the subpatterns below is an exception to the rule that the default spelling is <ti>, and each subpattern has its own sub-exceptions (some of which revert to <ti>) |
||
Subpattern |
Sub-exceptions |
|
For /eɪ∫əs/ the regular spelling is <-aceous>, |
audacious, capacious, contumacious, efficacious, fallacious, gracious, loquacious, mendacious, perspicacious, pertinacious, pugnacious, rapacious, sagacious, tenacious, vivacious, voracious; gaseous, spatious, Ignatius |
|
For /ɪ∫əl/ the regular spelling is <-icial>, e.g. artificial, beneficial, (pre)judicial, official, sacrificial, superficial (but this is a very small set) |
initial plus 4 other rare words in <-itial> |
|
For /∫ən/ preceded by /ɜː/ spelt <er, ur> or by /l, n/ spelt <l, n>, the regular spelling is <-sion>, e.g. (a/re-)version, excursion, emulsion, expulsion, pension, tension |
Cistercian, coercion (also pronounced with /ʒ/), exertion, Persian (also pronounced with /ʒ/), tertian; gentian |
|
For /∫ən/ preceded by /æ, e, mɪ, ʌ/ spelt <a, e, mi, u>, the regular spelling is <-ssion>, e.g. (com)passion; (ac/con/inter/pro/re/se/suc)cession, con/pro-fession, ag/di/e/in/pro/re/retro/trans-gression, (ad/com/e/inter/intro/manu/per/re/sub/trans)mission; con/dis/(re)per-cussion |
national, ration, (ir)rational; Prussian, Russian |
Spellings of medial /∫/ with <c, sc, t> always have a following <i(e)>, but the <i(e)> is a separate grapheme spelling /iː/. Some of the relevant words have alternative pronunciations with /s/, e.g. appreciate /əˈpriː∫iːjeɪt, əˈpriːsiːjeɪt/, negotiate /nɪˈgəʊ∫iːjeɪt, nɪˈgəʊsiːjeɪt/, species /ˈspiːʃiːz, ˈspiːsiːz/.
And then there seems to be a phonological constraint in many RP-speakers’ accents against medial /∫/ occurring twice in words ending in /iːjeɪt/ which already have one medial /∫/ and then would acquire another if suffixed to end in /iːˈjeɪ∫ən/. For example appreciation and negotiation are mainly pronounced /əpriːsiːˈjeɪ∫ən, nɪgəʊsiːˈjeɪ∫ən/, not /əpriː∫iːˈjeɪ∫ən, nɪgəʊ∫iːˈjeɪ∫ən/. But this in turn does not apply in words which do not end in /eɪ∫ən/ spelt <-ation>: conscientious is always pronounced with two occurrences of medial /∫/: /kɒn∫iːˈjen∫əs/, and recherché obviously has two: /rəˈʃeəʃeɪ/. The constraint also clearly does not apply to words with the –ship suffix, e.g. relationship /rɪˈleɪʃənʃɪp/.
The least frequent phoneme in spoken English.
The main system
For all three categories see also Notes.
Basic grapheme |
<si> |
91% (with <s>) |
e.g. freesia, vision. Only medial |
Rare graphemes |
<s> |
only medial before <u> and only in casual, usual, visual; (dis/en/fore-)closure, composure, embrasure, erasure, exposure, leisure, measure, pleasure, treasure, treasury, usur-er/y/ious |
|
|
<ge> |
4% |
never initial; medially, only in bourgeois(ie), mange-tout; regular in word-final position, where it occurs only in about 25 words of mainly French origin, namely beige, cortege, concierge, liege, melange, |
rouge and, with the <e> also forming part of the split digraphs <a.e, i.e, u.e> (for dual-functioning see section 7.1), in badinage, barrage, camouflage, collage, corsage, decalage, décolletage, dressage, entourage, espionage, fuselage, garage pronounced /ˈgærɑːʒ/, massage, mirage, montage, triage, sabotage; prestige; luge; only exception in word-final position is raj /rɑːʒ/ |
The rest
Oddities |
5% in total |
|
|
<ci> |
only, exceptionally but increasingly, in coercion pronounced /kəʊˈwɜːʒən/ (usually pronounced /kəʊˈwɜːʃən/) |
|
<g> |
initially, only in genre, gilet; medially, only in aubergine, conge, dirigiste, largesse, negligee, protege, regime, tagine and lingerie pronounced /ˈlænʒəriː/ (also pronounced /ˈlɒnʤəreɪ/); never word-final |
|
<j> |
only in jihad, raj and some rare French loanwords, e.g. bijou, goujon, jabot, jalousie, jupe |
|
<se> |
only in nausea, nauseous pronounced /ˈnɔːʒə(s)/ (also pronounced /ˈnɔziːjə(s)/) |
|
<ti> |
only in equation /ɪˈkweɪʒən/ |
|
<z> |
only in azure pronounced /ˈæʒə, ˈeɪʒə/ (also pronounced /ˈæzj(ʊ)ə, ˈeɪzj(ʊ)ə/), seizure /ˈsiːʒə/ |
|
<zi> |
only in brazier, crozier, glazier pronounced /ˈbreɪʒə, ˈkrəʊʒə, ˈgleɪʒə/ (also pronounced /ˈbreɪziːjə, ˈkrəʊziːjə, ˈgleɪziːjə/) |
2-phoneme grapheme |
/gʒ/ spelt <x> |
see under /g/, section 3.5.3 |
Notes
Because /ʒ/ is a sibilant consonant, adding any of the suffixes regular noun plural and third person singular person tense verb (both spelt <s> where the stem ends in <e>, otherwise <es>) and regular singular and irregular plural possessive (spelt <’s>) to a stem ending in /ʒ/ adds a syllable /ɪz/ as well as a morpheme: massages, (the) Raj’s (collapse). See also /z/, section 3.7.8, and /ɪ/, section 5.4.3.
As spellings of /ʒ/, <si, s> occur only medially and immediately after the stressed syllable, and are preceded by a vowel, and <s> is always followed by <u>. Almost all spellings with <si> are followed by <-on>, e.g. vision, but there are a few others, namely crosier, hosier(y), osier.
Although Carney gives 91% for <si, s> combined, it is clear that the great majority of these must be <si> spellings, since there are rather few words with /ʒ/ spelt <s> and a large number with /ʒ/ spelt <si>. This is why I have classified /ʒ/ spelt <s> as a rare grapheme.
Treating <ge> as the regular spelling of word-final /ʒ/ is justified by the first six words of French origin listed above: here the preceding vowel phonemes (plus the /n/ in melange) are represented without the aid of the word-final <e>. In the other 19 words <ge> is clearly still spelling /ʒ/, but it is necessary (and parallels other parts of the analysis) to analyse the <e> as also forming part of the split digraphs <a.e, i.e, u.e> spelling /ɑː, iː, uː/ (even though the last two correspondences have only one instance with included <g> each) – for dual-functioning see section 7.1. Then I analyse the /e/ in cortege as spelt only by the first <e> because it is a short vowel and no short vowels (in my analysis) are spelt by split digraphs – see section A.6 in Appendix A. Then <g> has to be recognised as a grapheme spelling /ʒ/ separate from <ge> because of the few words listed where this correspondence occurs initially and medially and the following vowel letters are obviously (involved in) separate graphemes.
The spelling <zh> is also used to represent /ʒ/, but because this occurs only in transcriptions of Russian names, e.g. Zhivago, Zhores, I have not added it to the inventory of graphemes.
The main system
Basic grapheme |
<th> |
100% |
Other frequent graphemes |
(none) |
The rest
Oddities |
<phth> |
only in apophthegm /ˈæpəθem/, phthalate /ˈθæleɪt/ |
|
<the> |
only in Catherine with first <e> elided (see section 6.10), saithe (/seɪθ/, ‘fish of cod family’) |
2-phoneme grapheme |
/tθ/ |
see under /t/, section 3.5.7 |
Notes
In the rare word saithe, the only function of the <e> seems to be to keep this word visually distinct from saith (/seθ/, archaic form of says) with the rare spelling <ai> for /e/.
See also the Notes under /ð/, next.
The main system
Basic grapheme |
<th> |
100% |
Other frequent graphemes |
(none) |
The rest
Oddity |
<the> |
<1% |
only word-final and only in breathe, loathe, seethe, sheathe, soothe, staithe, teethe, wreathe. See Notes |
2-phoneme graphemes |
(none) |
Notes
In all the words listed under Oddity, the vowel digraphs preceding <-the> spell a long vowel or diphthong, and there is therefore no need to analyse the final <e> as part of complex split graphemes <ea.e>, etc. However, in bathe, lathe, unscathed (the free form scathe meaning ‘to harm’ does not occur, but underlies both unscathed and scathing), swathe, blithe, lithe, tithe, writhe, clothe, hythe, scythe, the final <e> is part of the split digraphs <a.e, i.e, o.e, y.e> spelling /eɪ, aɪ, əʊ, aɪ/, so here /ð/ is again spelt <th>. The final <e> keeps breathe, loathe, seethe, sheathe, soothe, teethe, wreathe, bathe, lathe, swathe, clothe visually distinct from breath, loath (also spelt loth), seeth (/ˈsiːjɪθ/, archaic third person singular present tense form of see), sheath, sooth, teeth, wreath, bath, lath, swath, cloth. For the few minimal pairs differing only in having /θ/ or /ð/ see section 9.36.
The fact that both /θ/ and /ð/ are spelt <th> is useful in writing: people whose accents have different distributions of the two phonemes nevertheless spell the relevant words identically. This is particularly the case with some plural nouns, e.g. baths pronounced /bɑːðz/ in RP but /bæθs/ by many people from the North of England; the singular has /θ/ in both cases. But this does not help people trying to read unfamiliar words containing <th> - though again see section 9.36.
Occurs only before a vowel phoneme, and therefore never word-finally
The main system
For all these categories, and /w/ not represented at all, see Notes.
Basic grapheme |
<w> |
64% |
e.g. word. regular initially, rare medially |
Other frequent graphemes |
<u> |
31%, |
e.g. quick, language. Never initial, regular medially |
|
<wh> |
5% |
medially, only in erstwhile, meanwhile, narwhal, over-/under-whelm; otherwise only initial and only in whack, whale, wham(my), whang, wharf, what, wheat, wheedle, wheel, wheeze, whelk, whelp, when, whence, where, wherry, whet, |
whether, whey, whiff, whiffle, Whig, which, while, whim, whimper, whimsical, whimsy, whin, whine, whinge, whinny, whip, whippersnapper, whippet, whippoorwill, whirl, whir(r), whisk, whisker, whisk(e)y, whisper, whist, whistle, whit, white, whither, whitlow, Whitsun, whittle, Whitworth, whiz(z), whoa, whomp, whoopee, whoops, whoosh, whop, whump, whup, why, whydah and a very few other rare words (e.g. whilom) |
The rest
Oddities |
<1% in total |
|
|
<hu> |
only in chihuahua (twice) |
|
<ou> |
only in Ouija |
|
<ww> |
only in bowwow, glowworm, powwow, skew(-)whiff, slowworm |
2-phoneme graphemes |
/wʌ/ spelt <o> |
only in once, one – unless you prefer to consider the /w/ as not being represented in the spelling at all – see Notes and section 9.0 |
|
/wɑː/ |
See also Notes |
|
(1) spelt <oi> |
only in a few words more recently borrowed from French, e.g. bourgeoisie, coiffeu-r/se, coiffure, pointe, soiree, toilette |
|
(2) spelt <oir> |
mainly word-final and only in a very few words more recently borrowed from French, namely abattoir, avoirdupois, boudoir, memoir, noir, reservoir, soiree, voussoir. /r/-linking occurs in memoirist, noirish – see section 3.6 |
|
(3) spelt <oire> |
only word-final and only in a very few words more recently borrowed from French, namely aide-memoire, conservatoire, escritoire, grimoire, repertoire |
|
(4) spelt <ois> |
only word-final and only in a very few words more recently borrowed from French, namely avoirdupois, bourgeois, chamois (the animal, pronounced /ˈ∫æmwɑː/, as opposed to the leather made from its skin, pronounced /ˈ∫æmiː/, the latter also being spelt shammy), patois (contrast fatwa). /z/ surfaces in bourgeoisie – see section 7.2 |
|
/waɪ/ spelt <oy> |
only in foyer pronounced /ˈfwaɪjeɪ/, voyeur. Here the <y> is both part of the digraph <oy> spelling /waɪ/ and also a single-letter grapheme spelling /j/. For dual-functioning see section 7.1 |
3-phoneme grapheme |
/ˈwaɪə/ |
spelt with a single grapheme <oir> only in choir – one of only two 3-phoneme graphemes in the entire language |
Notes
If we follow Crystal (2012: 131-2), ‘more recently’ in terms of loanwords from French means after the Great Vowel Shift, which ended about AD 1600.
Although phoneme /w/ never occurs at the end of an English stem word, letter <w> is very frequent in word-final position, where it always follows a vowel letter with which it forms a digraph spelling a vocalic sound. See also ‘linking /w/’ later in these Notes.
The 2-phoneme sequences /kw, gw/ are almost always spelt <qu, gu> respectively.
<u> spelling /w/ occurs not only in the familiar /kw/ spelt <qu>, e.g. quick, squash, but also in a few words after /k, g, s, z/ spelt <c, g, s, ss, z>, e.g. cuirass, cuisine, cuissse; anguish, distinguish, extinguish, guacamole, guano, guava, iguana pronounced /ɪˈgwɑːnə/, language, languish, linguist, penguin, sanguine, segue, unguent; persuade, pueblo, puissan-ce/t, pursuivant, suave, suede, suite; assuage, dissuade; Venezuela (usually pronounced with /z/) and some very rare words; otherwise perhaps only in ennui, etui /ɒnˈwiː, eˈtwiː/. In these contexts <u> is clearly a consonant letter, though it is very rarely taught as having (like <y>) both vowel and consonant functions.
In linguistic terms it is unnecessary to analyse /kw/ as a single phonetic unit since in all words containing /kw/ (those where it is spelt <qu>, plus the Oddities acquaint, acquiesce, acquire, acquisitive, acquit (in these five words /k/ is spelt <cq>), awkward, coiffeur, coiffeuse, coiffure, cuisine, kwashiorkor and even choir) the /k/ is spelt separately, as it is also in compounds like backward. However, for teaching purposes many authors treat /kw/ and <qu> as units in close correspondence. Even though this ignores not only the (admittedly few) words just listed where /kw/ is not spelt <qu> (including acquaint, acquiesce, acquire, acquisitive, acquit) but also the 60+ words where <cqu, qu, que> are not pronounced /kw/ (see under /k/, section 3.7.1), this pragmatic approach to /kw/ spelt <qu> seems justified by the high frequency of this correspondence.
The frequency of <wh> is actually higher than 5% because Carney (1994: 253) did not count the function words what, when, whence, where, whether, which, while, whither, why. Scots and others who have the phoneme /ʍ/ (think ‘hw’) in their accent presumably have little difficulty in knowing which words to spell with <wh>, but the rest of us just have to learn them.
/wʌ/ also has 2-grapheme spellings, e.g. <wo> in wonder. /wɑː/ has the 2-grapheme spellings <oi> in coiffeur, coiffeuse, coiffure, <ua> in guacamole, guano, guava, iguana, suave, and <wa> in kwashiorkor. /waɪə/ has the 2-grapheme spellings <wire> in wire and <uire> in acquire, quire, require.
Medial /w/ spelt <w> is quite rare in stem words. The only words in which a medial <w> has only the function of spelling /w/ seem to be awkward (second <w>), fatwa, kiwi, kwashiorkor. There are rather more examples in which the <w> is both a grapheme in its own right spelling /w/ and part of one of the digraphs <ew> spelling /(j)uː/, e.g. in ewer, jewel, newel, sewer (/ˈsuːwə/, ‘foul drain’), skewer, steward, or <ow> spelling /əʊ/ - in stem words this occurs only in bowie, rowan (in its English pronunciation /ˈrəʊwən/) - or /aʊ/ in bowel, dowel, rowel, towel, trowel, vowel; bower, cower, dower, flower, glower, power, shower, tower; coward, dowager, howitzer, prowess, plus the Scottish pronunciation of rowan /ˈraʊwən/. For words containing /əʊ, aʊ/ see also sections 5.7.4, 5.6.2, and for dual-functioning section 7.1.
There are also various stem words where medial /w/ occurs in the pronunciation as a glide between /(j)uː, əʊ, aʊ/ and a following vowel phoneme but has no representation in the spelling, for example:
- after /(j)uː/: altruism, bruin, casual, cruel, cruet, doing, dual, duel, duet, duo, fluid, genuine, gruel(ling), iguana pronounced /ɪgjuːˈwɑːnə/, pirouette, ruin, silhouette, suet, toing, truant, (in)tuition, usual, zoology, zoomorphic
- after /əʊ/: coerce, coincide, coition, coitus, co-op, cooperate, co-opt, coordinate, co-own, egoism, froing, going, heroic, heroin(e), jingoism, Noel, no-one, oboist, phloem, poem, soloist, spermatozoon, stoic(al)
- after /aʊ/: devour, flour, hour, lour, our, scour, sour and dour pronounced /ˈdaʊwə/, all of which end in /ˈaʊwə/, plus the stray medial example of sauerkraut.
This pattern of representing or not representing medial /w/ after /(j)uː, əʊ, aʊ/ before a following vowel phoneme is paralleled when stem words ending in those phonemes have a suffix beginning with a vowel phoneme added. In all such cases there is a ‘linking /w/’ glide between the stem and the suffix (see Cruttenden, 2014: 152), but this is represented in the spelling only if the stem word already ends in <w>, otherwise not. If there is a <w> it continues to function as part of the spelling of the stem-final vowel while also now spelling /w/ - the familiar dual-functioning (see section 7.1). If the stem does not end in <w> the spelling simply ignores the /w/-glide. For examples of both categories see Table 3.7.
Preceding phoneme(s) |
/w/ represented in spelling by <w> |
/w/ not represented in spelling |
/(j)uː/ |
brew-er/ing, chewing, few-er/est, hewer, Jewish, leeward pronounced /ˈluːwəd/ (also pronounced /ˈliːwəd/), mewing, new-er/est, renewer, sinew-ous/y, stewing, (inter/re-)view-er/ing |
do-er/ing, lassoing, toing (and froing); canoeing, shoeing; mooing, shampooing, shooing, tattoo-ing/ist; rendezvousing; accru-al/ing, argu-able/ing, continu-al/ance/ing/ous, gluing, issu-able/ing, pursu-ance/ing, rescuable, subduable, statuesque, su-able/ing, virtuous |
/aʊ/ |
allow-ance/ing, avowal, bowing (‘inclining the head’), miaowing |
Maoist; plough-er/ing |
/əʊ/ |
sewer (/ˈsəʊwə/, ‘one who sews’), sewing; blowing, bowing (‘playing stringed instrument’), follow-er/ing, N.B. Where the <ow> is not stressed, not only does linking /w/ occur but the vowel is often reduced to /ə/ and is therefore spelt simply by the <o>, e.g. widower. |
plateauing; hoeing, toeing; (toing and) froing, going |
And this pattern also occurs when stem words ending in /(j)uː, əʊ, aʊ/ are followed by a word beginning with a vowel phoneme. In all such cases there is a ‘linking /w/’ glide between the two words, but again this is represented in the spelling only if the first word already ends in <w>, otherwise not. Since this book is mainly concerned with stem words and their derivatives, only a few examples of inter-word /w/-glides not represented in the spelling will be given: to estimate /tuˈwestɪmeɪt/, go away /ˈgəʊwəˈweɪ/, slough of despond /ˈslaʊwəvdɪsˈpɒnd/.
See also the parallel phenomenon of ‘linking /j/’ (next section), and the subtly different one of /r/-linking (section 3.6).
Occurs only before a vowel phoneme, and therefore never word-finally.
For all these categories, the percentages, and /j/ not represented at all, see Notes.
The main system
Basic word-initial grapheme |
<y> |
19% |
e.g. yellow. Regular (very few exceptions) initially except before /uː, ʊə/; rare medially |
Other frequent grapheme |
<i> |
c.5% |
e.g. onion. Only medial |
Frequent 2-phoneme graphemes |
/juː/ |
See also the main entry for /juː/, section 5.7.5 |
|
|
(1) |
62% |
e.g. union, illusion, cute. Regular as spellings of /juː/ in non-final and closed final syllables respectively; <u> word-final only in coypu, menu, ormolu |
|
(2) spelt <ew> |
11% |
e.g. few. Regular as spelling of /juː/ word-finally in monosyllables |
Rare 2-phoneme grapheme |
/juː/ spelt <ue> |
for percentage |
regular as spelling of /juː/ word-finally in polysyllables, e.g. pursue |
The rest
Oddities |
All medial |
<1% in total |
|
<h> |
only in a very few words between 2 vowels, namely annihilate, vehement, vehicle, vehicular |
|
<j> |
only in hallelujah and majolica pronounced /maɪˈjɒlɪkə/ (also pronounced /məˈʤɒlɪkə/) |
|
<ll> |
only in French-like pronunciations of bouillabaisse /buːjɑːˈbes/ and marseillaise /mɑːseɪˈjez/ and Latin American Spanish-like pronunciation of tortilla /tɔːˈtiːjɑː/ |
Other 2-phoneme graphemes |
3% in total |
|
|
/juː/ spelt <eau, eu, ewe, ui, ut, uu> |
see under /juː/, section 5.7.5 |
|
/jʊə/ |
see Notes below and under /ʊə/, section 5.6.5 |
|
/jə/ |
see under /ə/, section 5.4.7 |
|
/lj/ spelt <ll> |
only in carillon. Most occurrences of /lj/ (all medial) have one of the 2-grapheme spellings <li, lli> - see the groups of words in the Notes beginning battalion, civilian. The sequence /lj/ also occurs in halyard, failure but in both the /l/ is spelt <l>; in halyard the /j/ is spelt separately as <y>, but in failure is subsumed with the final /ə/ in <ure> |
|
/nj/ spelt <gn> |
only in chignon, cognac, gnocchi, lasagne, lorgnette, mignonette, monsignor, poignant, seigneur, soigné, vignette |
Notes
Although /j/ never occurs at the end of an English stem word when pronounced alone or before a word beginning with a consonant phoneme, letter <y> is very frequent in word-final position, where it always spells a vocalic sound, either singly or as part of a digraph. See also ‘linking /j/’ later in these Notes, and section 11.5.
The correspondences for initial /j/ are fairly straightforward:
- if the word is a monosyllable /j/ is spelt <y>, e.g. yacht, yolk. Only exceptions: ewe, Ewell (contrast yew, you, Yule), uke, Ure, use (both the noun /juːs/ and the verb /juːz/ and their derived forms)
- if the word is a polysyllable and the next phoneme is /uː/, the /j/ is usually subsumed into <u> spelling the 2-phoneme sequence /juː/, e.g. union, university (exceptions: eucalyptus, eucharist, euchre, eugenic, eulogy, eunuch, euphemism, euphorbia, euphoria, Eustachian, euthanasia, ewer with /j/ subsumed instead into <eu, ew>; Yucatan, Yugoslav, Yupik with /j/ explicitly represented as <y>)
- if the word is a polysyllable and the next phoneme is /ʊə/, the /j/ is usually subsumed into <ur> spelling the 2-phoneme sequence /jʊə/ - but this category includes mainly urea and various words derived from it, e.g. urethra, urine, urology (exceptions: eureka, eurhythmic)
- otherwise, that is in polysyllables where the next phoneme is not /uː, ʊə/, initial /j/ is almost always spelt <y>, e.g. yellow, etc.
In medial positions, large numbers of instances of /j/ are subsumed into 2-phoneme spellings of /jə, jʊə, juː/, and lists of these can be found in sections 5.4.7, 5.6.5 and 5.7.5 respectively. There are also a very few instances of other 2-phoneme spellings – see above. Otherwise, where a consonant phoneme precedes /j/ the predominant spelling appears to be <i>, e.g. in:
- a group of words ending /jən/ spelt <-ion> after a stressed syllable: battalion, billion, bunion, champion, companion, dominion, million, minion, onion, opinion, pavilion, pinion, union, vermilion
- a small group of words ending /jəriː/ spelt <-iary> after a stressed syllable: apiary, auxiliary, aviary, breviary, domiciliary, pecuniary, topiary (contrast January with final /jəriː/ spelt <-uary> and requiring /je/ to be analysed as spelt <ua>) (but for incendiary, intermediary, stipendiary, subsidiary pronounced with /ʤəriː/, i.e. with /dj/ affricated to /ʤ/, see section 3.7.4)
- a further ragbag: (with /lj/ spelt <(l)li>) civilian (3rd <i>), colliery, milieu, plus, in rapid speech, brilliant (2nd <i>); (others) behaviour, envious, piano, pronunciation (1st <i>), spaniel, (inter-/re-)view, etc.
In these cases after a consonant <i> is clearly a consonant letter, though even more rarely than <u> is it taught as having (like <y>) both vowel and consonant functions. There are a few exceptions with <y>: banyan /ˈbænjæn/, biryani, canyon /ˈkænjən/, halyard, lanyard, vineyard.
Otherwise, i.e. in situations not yet covered and where /j/ is both preceded and followed by vowel phonemes, the predominant spelling of medial /j/ is probably zero, i.e. it is not represented at all. There are a few exceptions among the Oddities above, plus just three with <i> (alleluia, onomatopoeia, pharmacopoeia, where the <i> can be taken to represent /j/ since the preceding <u, oe> spell /uː, iː/), and rather more with <y>:
- beyond, bowyer /ˈbəʊjə/, lawyer, sawyer, yoyo, where <y> spells only /j/, plus bayonet, cayenne, crayon, mayonnaise, rayon, which can be economically analysed as having /eɪ/ in their non-final syllables spelt (regularly – see sections 5.7.1, 6.3) just by the <a>, so that /j/ is spelt just by the <y>
- abeyance, where /j/ is spelt <y> but the <y> is also part of <ey> spelling /eɪ/
- arroyo, doyenne pronounced /dɔɪˈjen/, foyer pronounced /ˈfɔɪjə/, loyal /ˈlɔɪjəl/, Oyer (and Terminer), royal /ˈrɔɪjəl/, soya, where /j/ is spelt <y> but the <y> is also part of <oy> spelling /ɔɪ/
- coyote /kaɪˈjəʊtiː/, doyen and doyenne both pronounced /dwaɪˈjen/, foyer pronounced /ˈfwaɪjeɪ/, kayak /ˈkaɪjæk/, papaya, voyeur /vwaɪˈjɜː/, where /j/ is spelt <y> but the <y> is also part of <ay, oy> spelling /aɪ/.
(For dual-functioning see section 7.1).
A list of words in which /j/ is not represented at all in medial positions could go on for pages. A few examples (illustrating the range of spellings within which /j/ is invisible) are: (with final /ə/) bacteria, cochlear, idea, linear, meteor, senior (all these words and many others would be analysed by Carney as ending in /ɪə/, whereas I assign them to /iːjə/); (others) aphrodisiac, appreciate, archaic, audience, axiom, caviar, chaos, chariot, create, creole, dais, deify, diary, dossier, foliage, genius, hilarious, jovial, lenient, mosaic, museum, odious, pantheon, period, radii, radio, reinforce, ruffian, serviette, simultaneous, society, soviet, spontaneity, stallion, tedium, triangle, triage, video.
This pattern of representing or not representing medial /j/ between two vowel phonemes is paralleled when stem words ending in /aɪ, eɪ, ɔɪ, iː/ have a suffix beginning with a vowel phoneme added. In all such cases there is a linking /j/-glide between the stem and the suffix, but this is represented in the spelling only if the stem word ends in a digraph ending in <y>, otherwise not. If there is such a digraph the <y> continues to function as part of the spelling of the stem-final vowel while now also spelling /j/ - the familiar dual-functioning (see section 7.1). If the stem does not end in a digraph ending in <y> the spelling simply ignores the /j/-glide. For examples of both categories see Table 3.8.
Preceding phoneme |
/j/ represented in spelling by a digraph ending in <y> |
/j/ not represented in spelling |
/aɪ/ |
shanghaiing; higher, highest, sighing; beautifying, defying, dryer, dyer, flyer, fryer, supplying, plus: - some words obeying the <y>-replacement rule (section 6.5), e.g. alliance, amplifier, defiance, drier, flier; - the four words which obey the - two Oddities: dyeing, eyeing. See also paragraph below Table |
|
/eɪ/ |
betrayal, conveyance, layer, layette, obeying, playing, prayer pronounced /ˈpreɪjə/ (‘one who prays’), preying, purveying, surveying |
crocheting, inveighing, laity, neighing, ricocheting, segueing, weighing |
/ɔɪ/ |
annoyance, boyish, buoyant, cloying, destroyer, enjoying, joyous, toying |
|
/iː/ |
jockeying, moneyer, volleying |
absenteeism, agreeing, beauteous, fleeing, fre-er/est, liar, nauseous pronounced /ˈnɔːziːjəs/, orgiastic, precising, seeing, leylandii, fasciitis, skiing, taxiing, plus many words obeying the <y>-replacement rule (section 6.5), e.g. acrimonious, bolshi-er/est, calumniate, carrier, centurion, comedian, dalliance, dutiable, enviable, historian, industrial, luxuriance, melodious, memorial, remedial, studious, twentieth, thirtieth, etc., variable, variance |
The words with stems ending in <y> which I’ve placed under ‘/j/ not represented in spelling’ opposite /aɪ/ in Table 3.8 may seem not to belong there but in the other column. The <y> here could be analysed as representing both the /aɪ/ and the /j/, but this would be the only example in my entire analysis of a letter having two single-letter functions – in all other cases the dual-functioning letter’s first function is as part of a di- or trigraph (see section 7.1). Also, none of the other words in this box would support such an analysis, and as it happens none of the digraphs ending in <y> which spell /aɪ/ occur word-finally (see section 5.7.3); hence there are no words to put in the left-hand box opposite /aɪ/.
Some further instances of ‘invisible /j/’ occur within suffixes, e.g. <-ial, -ian> /iːjəl, iːjən/ (though again Carney would assign these to /ɪə/).
And this pattern also occurs when stem words ending in /aɪ, eɪ, ɔɪ, iː/ are followed in running speech by a word beginning with a vowel phoneme. In all such cases there is a ‘linking /j/’ glide between the two words (see Cruttenden, 2014: 152), but again this is represented in the spelling only if the first word already ends in a digraph ending in <y>, otherwise not. Since this book is mainly concerned with stem words and their derivatives, only a few examples of inter-word /j/-glides not represented in the spelling will be given: I understand /aɪjʌndəˈstænd/, inveigh against /ɪnˈveɪjəˈgeɪnst/, ‘hoi polloi’ is Greek /ˈhɔɪpəˈlɔɪjɪzˈgriːk/, free offer /friːˈjɒfə/.
See also the parallel phenomenon of ‘linking /w/’ (previous section), and the subtly different one of /r/-linking (section 3.6).
The only percentage stated by Carney (p.256) is 19% for /j/ spelt <y>. In order to work out other percentages I have ignored non-represented /j/. For /j/ spelt <i> I deduced a percentage as follows: Mines et al. (1978, Table A-1, p.236) show that the ratio of initial to medial /j/ is about 4:1. Since /j/ spelt <y> is very rare medially, and the Oddities and the 2-phoneme graphemes spelling sequences other than /juː/ are negligible, it is safe to take the ‘junior partner’ in this ratio as medial /j/ spelt <i>. Hence the figure of c.5% as the percentage for this spelling of /j/.
It follows that 2-phoneme spellings of /juː/ constitute most of the remaining 76%. Carney (p.201) states that <u, u.e> are 82% of spellings of /juː/, hence {82% x 76%} = 62% of the spellings of /j/, and that <ew> is 15% of spellings of /juː/, hence {15% x 76%} = 11% of the spellings of /j/. The remaining {100%-(19%+5%+62%+15%)} = 3% of spellings of /j/ are mainly the minority 2-phoneme spellings of /juː/, plus spellings of /jə, jʊə/.