Is there posttonic stress in English?
Péter Szigetvári
Eötvös Loránd University, Budapest szigetvari@elte.hu
OCP17
2020-02-05
the issue
is the red ostressed?
EPD LPD K&K M-W cf
robot r´@wbot no no no yes Probert pr´@wb@t autopsy ´o:topsij no no no yes leprosy l´epr@sij apricot ´ejpr@kot no no yes yes advocate ´adv@k@t monsoon mons´uwn no yes no no platoon pl@t´uwn October okt´@wb@ no no no no opponent@p´@wn@nt novelettenov@l´et yes yes yes yes
why the confusion?
1. stress is asegmental/melodicphenomenon:
full vs reduced vowel, flapping vs its absence, syncope, epenthesis, loss of j, loss of w, etc
2. but “stress” is also used for a rhythmic/prosodicphenomenon:
involving feet, “stress” clash, “stress” shift, the tonic etc, let’s call this accent(like Vanderslice & Ladefoged 1972, Schane 1979, Gussenhoven 1991)
3. stress is a lexical property of certain vowels
4. the connection between stress and accent is only that accent usually (thought not always) falls on a vowel which is stressed
assumptions about the vowel inventory of BrE
1. “centring diphthongs” have all monophthongized: NEAR i:,
SQUARE e:,CUREo:/u:/@: (Cruttenden 2014)
2. FLEECE andGOOSEare not long monophthongs, but ijand uw (Sweet 1900)
3. “diphthongs” are VC (Trager & Bloch 1941) 4. STRUT is schwa (cf Fabricius 2007)
5. HAPPY-tensing (Wells 1982) has gone through:
*i{#,V}: happy alien h´apij ´ejlij@n
(also *u{#,V}: usual value j´uwZuw@l v´aljuw)
vowel categories in BrE: a phonotactic classification
checked
KITi,DRESSe,TRAP a,STRUT@,LOTo,FOOT u only before a consonant
“nonchecked”
NEARi:,SQUARE e:,START a:,NURSE@:,FORCE o:,CUREu:
before a consonantandword finally note 1
“diphthongs” are VC, vowels can’t occur next to each other at all (neon n´ıjon ≈Nissan n´ısanor vowel v´aw@l≈vassal v´as@l) note 2
unstressed@ is nonchecked: it occurs word finally
full vs reduced vowels
vowels that may occur unstressed: i @ u
KIT(comic k´omik),STRUT (comet k´om@t),FOOT (volume v´oljum)
FLEECE(volley v´olij),GOAT(willow w´ıl@w), GOOSE(value v´aljuw)
vowels that only occur stressed: e a o
DRESSe,TRAP a,LOT o
FACEej,PRICEaj,MOUTHaw,CHOICEoj
long vowels only occur stressed (i: @: u: e: a: o:) idea ajd´ıj@∼ajd´ı:,mature m@tS´uw@∼m@tS´u:,
Europeanj´uw@r@p´ıj@n∼j´u:r@p´ı:n,
butIndia´ındij@∼´ındj@, *´ındi:;usual j´uwZuw@l∼j´uwZw@l, *j´uwZu:l stress cannot be unambiguously guessed from vowel quality ugly ´@glij vsagree @gr´ıj
Kennedy k´en@dijvsKentucky kent´@kij colander k´@l@nd@vsColumbus k@l´@mb@s
vowel(+glide) alternations in English
ablaut sing∼sang,get∼got
umlaut full∼fill,goose guws∼geese gijs,man∼men vowel shift hide hajd ∼hid hid,tone t@wn ∼tonic t´onik vowel reduction ironic ajr´onik ∼irony ´ajr@nij,combine k@mb´ajn∼
combinationk´ombin´ejS@n
these alternations are not phonological, they are historic relics
◮ they are not systematic
idi[´o]tic∼idi[@]t vsrob[´o]tic∼rob[o]t t[@w]ne∼t[o]nic vsph[@w]to∼ph[@w]tic
◮ they are phonetically arbitrary (Kaye 1995)
accent is mobile, stress is stable
accent moves. . .
Taiwán ∼Táiwan míracle thirtéen∼thírteen chápters Piccadílly∼ Píccadilly Círcus . . . but only to a stressed vowel Madríd∼Madríd súmmit
betwéen∼ betwéen chápters cf
t´ajw´on vsm@dr´ıd T´@:t´ıjnvsb@tw´ıjn
stress is segmental/melodic: no flapping before stress
cartoonk`a:t´uwnvs cartoon network k´a:t`uwnn´etw`@:k latexl´ejt`eks vs latest l´ejR@st
levitatel´evit`ejt vs lemmata l´em@R@
Wotanv´@wt`an vs quota kw´@wR@
habitath´abit`atvs palatal p´al@R@l atoll´at`olvs atom ´aR@m
Minotaurm´ajn@t`o: vs monitor m´oniR@
autism´o:t`ız@m vs nautilus n´o:Ril@s
magnetismm´agn@t`ız@m vs magneticm`agn´eRik manateem´an@t`ıj vs vanityv´an@Rij
ketonek´ıjt`@wnvs mosquitom@sk´ıjR@w orotund´or@wt`@nd vs orator´or@R@
stress is segmental/melodic: syncope and epenthesis
no syncope before stress
memorym´em/@rij vs memorize m´em@r`ajz radicallyr´adik/@lij vs radicalize r´adik@l`ajz barbarousb´a:b/@r@s vs barbarism b´a:b@r`ız@m familyf´am/@lij vs jubileedZ´uwb@l`ıj
summarys´@m/@rijvs bummaree b´@m@r`ıj no epenthesis before stress
princepr´ınhtis vs princess pr`ıns´es insolent´ınhtis@l@nt vs insect´ıns`ekt consciousk´onhtiS@s vs insure inS´o:
Gimsong´ımhpis@n vs Gimsoniang`ıms´@wnij@n nymphetn´ımhpif@t∼n´ımhpifitvs n´ımf`et
stress is segmental/melodic: loss of [j] and [w]
[j]/palatality is lost after coronal C before stress volumev´oljuwm vs voluminousv@l´uwmin@s
Danubed´anjuwbvs Danubiand@n(j)´uwbij@n
congratulatek@ngr´atS@l`ejtvs AmE gratuitous gr@t´uwit@s assiduous@s´ıdZuw@svs AmE assiduity`as@d´uw@tij
valuev´aljuwvs absolute´abs@l`uwt venuev´enjuwvs revenue r´ev@n(j)`uw
statutest´atSuwt vs AmE constitutek´onstit`uwt
[w] is optionally retained before stress (and word finally) stimulusst´ımjul@s vs stimulatest´ımju(w)l`ejt
botulinb´otSulinvs botulismb´otSu(w)l`ız@m fortunatef´o:tSun@t vs fortunef´o:tSu(w)n
stress patterns of English: one- and two-vowel words (48%)
legend: w=unstressed,s=stressed,ś=accented source: http://cube.elte.hu
one-vowel words (10%)
w (.64%) and @nd (these are clitics, not words) ś (99%) ant ´ant
two-vowel words (38%) ww (0%)
św (54%): atom ´at@m wś (9%): amok @m´ok śs (30%): atoll ´at`ol sś (7%): antique`ant´ıjk
stress patterns of English: three-vowel words (30%)
www (0%)
1s: śww (28%): abacus´ab@k@s wśw (18%): abandon @b´and@n wwś (0%)
2s: śws (20%): acrobat´akr@b`at swś (6%): arabesque`ar@b´esk śsw (9%): ancestor´ans`est@
sśw (11%): ambition `amb´ıS@n wśs (2%): amortize@m´o:t`ajz wsś (.07%): returnee rit`@:n´ıj
3s: śss (3%): adumbrate´ad`@mbr`ejt sśs (1%): asbestos `azb´est`os ssś (2%): addressee `adr`es´ıj
so what?
◮ stress is binary, a vowel is either stressed or unstressed
◮ stress is a lexically specified property of vowels
◮ stress is stable, a lexically stressed vowel remains stressed, a lexically unstressed vowel remains unstressed
◮ accent is partly lexical (śs vssś,śwsvs swś, etc)
◮ accent is mobile (sś →śs / ś)
thank you for your attention
special thanks to
Karolina, Bartłomiej, & co, NKFI #119863, George Soros slides available at
http://seas3.elte.hu/szigetva/papers.html#ocp17