Early Modern Phonology: Some not-so-great vowel shifts


Short vowels


Variation between /ć/ and /α/

-/ć/ became /α/

-in early English, before /r/: harm, hard

                   -a preceding /w/ rounds it: warm, reward

-not yet evident in earlier rhymes like arm and warm, regard and reward

          -in standard British English

-sporadically before /ns/ and /nt/

                             -dance, plant

-before voiceless fricatives (except /š/)

                             -class, pass, path

                                      -but not if another vowel follows:

                                                -classical, passage


-read more about this: Crystal 93, 207; Algeo and Pyles, 165, 213 


ME /U/ “centered and unrounded in most environments” to /^/ (but)

-in Shakespeare’s English, run, mud, cut would have rhymed with put, bush

-love and come also in this category: they’re really “U” words that happen to be spelled with an <o>

-change sometimes blocked by

          -preceding labial consonants (esp. followed by /l/, /š/, /č/)

                   -full, pull, wool, wolf

                   -bush, push


-read more: Crystal 69; Origins 163; Problems 7.4.3 (p. 152) 


Following /r/ tends to do odd things to vowels

§       e.g. the “NURSE” shift, centres

o      girl /I/ ->

o      fern /ε/ ->

o      hurt /U/ ->

§       we know that this sound change happened before /r/ was lost in some dialects!


§       e.g. can lower vowels, esp. /εr/ to /ar/

o      permanently with native words like OE steorre -> star

·        temporarily with loanwords like servant, mercy

o      sometimes variant forms coexist

§       person

§       clerk

§       vermin


-read more: Problems 7.4.7 (153)




Long vowels


Exceptions to the Great Vowel Shift


Effect of following /r/ (can lower preceding vowel): Problems 7.4.8

-/o/ didn’t rise to /u/: floor

-/ε:/ didn’t rise to /i/: wear, bear, tear, pear

          -train wreck: pair, pare, pear

                   -but: ear, hear, etc.


-/ε:/ sometimes shortened before the GVS happened: Problems 7.4.4

          -death, deaf

          -bread, head


-/u/ sometimes laxed/shortened to /U/ after the GVS happened

          -words where it shortened early joined the unrounding of /U/->/^/ (cut)

                   -flood, blood: /o/ -> /u/ -> /U/ -> /^/

-words where it shortened later missed the boat (or the but) and stayed at /U/

                   -good and book: /o/ -> /u/ -> /U/


A diphthong or two (Origins 164)


Modern relevance: do you say news /njuwz/ or /nuwz/?


[-EmodE /yu/ came from a variety of OE/ME sources

          -/εu/: dew, beauty

          -/eu/: due, new, fruit

          -/iu/: spew, rule]


-EmodE /yu/

          -usually disappeared after /r/, e.g. rude

          -/y/ retained in some contexts, e.g.

                   -initially: use, Europe

-after labials: pure, beauty, music

-after velars: cute, argue, hew

          -/y/ variably deleted in other contexts

-for some older Canadians, particularly stigmatized after /t/, /d/, /n/,

e.g. tune, duke, news

          -after /s/ and /z/, /y/ involved in another sound change called assibilation

                   -sugar, sure /sy/ -> /š/


-read more: Crystal 341; Problems 7.4.6 


For further reading: Jack Chambers’ research on ‘yod-dropping’ in Canadian English




Early Modern English Phonology: Consonants


Main source:

Lass, Roger. “Phonology and morphology.” Volume III 1476-1776

of the Cambridge History of the English Language. Cambridge: CUP, 1999.


The phoneme /h/: its postvocalic allophones ([ç] and [x])

§       either change to /f/

o      mostly word-finally: tough, laugh

o      but draft/draught (perhaps the <f> because there are so few /_ft/ words)

§       or disappear

o      word-finally: sigh, through

o      before word-final /t/: right, thought

§       if the vowel was short, the process lengthened it

§       <night> /nIçt/ -> /nit/

·       (now eligible for the GVS -> /nait/)

§       Old French delit -> ME delit now spelled delight

o      “reverse spelling”: ass. that <_ight> /_it/ -> /_ait/

§       so if <night> can be /nit/ then /delit/ can be <delight>


Now /h/ is left only word-initially: not a very stable phoneme

o      in OE, words could begin

o      with /h/: hors, habban

o      or with a cluster: hlúd ‘loud’, hwíl ‘while’

§       word-initial [h] lost in cluster: loud

§       in some ME mss, sometimes also word-initially, e.g. adde ‘had’

·       used to be blamed on Norman scribes, but it’s a native process


o      late Latin and Old French had lost word-initial /h-/, so loans into ME reflect that

-not spelled in ostler (‘hosteler’, somebody at an inn who looks after horses)

                   -spelled in heir, honest

                   -spelled and reintroduced in history, human, humour

-prestige of Latin, written language/spelling pronunciation

                             -“h-dropping” a social issue from C18th onwards

                   -regional variation: herb, human...


Read more: Origins 166, Problems 7.5.1


Loss of postvocalic /r/ in context

-very early, lost first before /s/:

-OE baers -> bass (fish)

          -change often reversed

          -but there are some interesting doublets

-arse, horse, curse, burst ->

-think about the /ć/ vs /a/ pronunciation of ass vs arse!


-then, in some dialects (ancestors of RP, rural East Anglia), /r/ lost much more widely

          -postvocalically (car or park)

                   -sporadically from C15th; widespread by C18th


-not initially approved of (r-dropping as bad as h-dropping)

-Walker: /r/ is “sometimes entirely sunk” in London


-do non-rhotic American accents (e.g. eastern seaboard, South)

-descend from these nonrhotic English dialects?

-e.g. East Anglian Puritans settling in the eastern seaboard

-or imitate it because of sustained contact?

          -stay tuned


read more: Crystal 69, 93, 307, 312; Origins 167, 213; Problems 7.5.5



Loss of /l/ in some contexts (/l/ and /r/ similar)

          -lost after low back vowels before certain consonants

                   -velar: talk, walk; folk, yolk, Holmes

                             -but not before most dentals: salt, bolt

-labial: half, palm

-in some dialects, restored (spelling pronunciation)

-calm, palm, balm, alms


          -more recent: vocalization of /l/ in Cockney English (M375)

                   -pill [p`io], bottle [o] or [u]


Read more: Origins 166, Problems 7.5.4



Loss of final stops in some word-final consonant clusters, e.g.

          -loss of final alveolars in certain clusters

          -when /s/ is involved

                             -handsome, landscape

                             -castle, hasten, wrestle, ostler, Christmas

                             -my electricians talk about joyces in the walls:

joist -> jois, plural joises


-loss/glottalization of [t] after voiceless consonants

                             -exploited by makers of “Duc(t/k) tape

                                      -not reflected in standard spelling


          -loss of final stop in other clusters

                   -/g/ after /ŋ/: sing /sIŋg/ -> /sIŋ/


Two new phonemes


Two new phonemes arise that filled gaps in the existing system


          -in OE, [ŋ] only occurred before /k, g/: sing /sIŋg/

-[ŋ] used to be the allophone of /n/ before /g/

-but in lexical words, word-final stops often lost in clusters

          -sing /sIŋg/ -> /sIŋ/ with loss of word-final stop

-now sing is a minimal pair with sin, so the /ŋ/ is officially a phoneme

-/ŋ/: a velar nasal to go with the velar stops


Not the same as ‘g-dropping’ in present participles, e.g. running /-in/


Read: Crystal 69, Problems 7.5.3 (155)


/ž/:    -result of

(a)             borrowing from French (beige) and

(b)            a set of sound changes called palatalization or assibilation

-measure: from /zy/ to /ž/; now minimal pair with mesher

-now there’s a voiced alveopalatal fricative to go with the voiceless one /š/

-remember that English consonants tend to come in voice/less pairs


Assibilation in context

          /zy/ -> /ž/: seizure

/sy/ -> /š/: nation, ocean

          -explains onset of sugar, sure

          -Love’s Labours Lost: puns on shooters and suitors

                   -both kinds of hunters

                   -sometimes reversed: suitors not /š/

-if /sj/ before /u/ didn’t become /š/, /ĵ/ could be dropped: Susan, supreme

-still regional variation in words like tissue, sexual


/dy/ -> /ĵ/: soldier

          -where does Cajun come from?

          -sometimes reversed: Indian


/ty/ -> /č/: creature

          -what’s another form of this word?


Read more: Crystal 69, Origins 164, Problems 7.5.2 (155)


Many sound changes not reflected in spelling

          -Cajun an exception

          -some only in informal/regional varieties? ass, hoss, varmint     


Spelling pronunciation has reversed some of them

-native processes: calm

-etymological respellings:

-vulgar Latin fallita -> OF, ME faute -> EmodE fault

                   -sometimes changes pronunciation: fault, throne, habit

                   -sometimes doesn’t: dette -> debt