Upload
others
View
10
Download
0
Embed Size (px)
Citation preview
BBN–ANG–243 Phonological analysis
3–4. Contrast in English consonants
Zoltán G. Kiss, Péter Szigetvári, Miklós Törkenczy
Dept. of English Linguistics, Eötvös Loránd University
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 1 / 98
aims
aims for today and next time
◮ contrast among English consonants
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 2 / 98
aims
aims for today and next time
◮ contrast among English consonants
◮ di=erent models/analyses of contrasts
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 2 / 98
aims
aims for today and next time
◮ contrast among English consonants
◮ di=erent models/analyses of contrasts
◮ laryngeal contrast: phonetics and distribution
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 2 / 98
aims
aims for today and next time
◮ contrast among English consonants
◮ di=erent models/analyses of contrasts
◮ laryngeal contrast: phonetics and distribution
◮ when contrast disappears: neutralization
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 2 / 98
aims
aims for today and next time
◮ contrast among English consonants
◮ di=erent models/analyses of contrasts
◮ laryngeal contrast: phonetics and distribution
◮ when contrast disappears: neutralization
◮ voicing assimilation
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 2 / 98
models
steps of scientific approaches
1. observe something interesting in the real world that needs explanation
2. set up a hypothesis = create a model that represents but is not equalto reality, use this model to predict future events in the real world
3. test the hypothesis: collect data, see if the model predicts what we seein the real world
4. if the data. . .
◮ . . . support hypothesis: retain it◮ . . . do not support hypothesis: reject it, look for new explanations
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 3 / 98
models
science: fitting models to reality
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 4 / 98
models
modelling language
modelling language is di;cult: we don’t have access to ‘reality’ = the brain
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 5 / 98
models
language as a black box = generating “machine”
the model of language needs to be built based on the input and the output
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 6 / 98
models
modelling sound contrast
our focus today and next time: to set up a model of sound contrast
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 7 / 98
contrast
speech = bundle of phonetic features
◮ we can think about speech sounds as a bundle of phonetic features
◮ these features can be related to articulation: articulatory gestures inthe vocal tract
◮ lips, teeth, tongue position, position of the velum, manner ofarticulation, vocal fold vibration, etc.
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 8 / 98
contrast
example
What phonetic features are needed to produce the word base?
Features b eI s
V-height mid
V-clipping clipped
Voicing voiced voiced voiceless
Nasality oral oral oral
C-place labial alveolar
C-manner stop fricative
. . . . . . . . . . . .
– are all these necessary to express contrast?
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 9 / 98
contrast
two approaches to features
classical phonemic & generative approach
◮ only those features are part of language that are unpredictable andcontrastive
◮ predictable features are introduced by phonological rules
◮ simpler model
phonetically-grounded approach
◮ even predictable (“redundant”) features are part of language (not justspeech)
◮ phonetically “richer” model
◮ one given contrast can be aided (“cued”) by several features, not justone
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 10 / 98
contrast
F. de Saussure: language = a system of contrasting signs
For language, only those phonetic features are “relevant” which can beused to express contrast between words and which are unpredictable.
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 11 / 98
contrast
phonetic di=erence 1: vowel length and nasality
Long Clipped Nasalized
save [seIv] ∼ safe [seIf ] ∼ sane [seIn]made [meId] ∼ mate [meIt] ∼ main [meIn]maze [meIz] ∼ mace [meIs] ∼ main [meIn]
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 12 / 98
contrast
phonetic di=erence 2: vowel height
Mid Low
pale [peIl] ∼ pile [paIl]
Dave [deIv] ∼ dive [daIv]
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 13 / 98
contrast
phonetic di=erences
1. [eI] ∼ [eI] ∼ [eI]: long ∼ clipped ∼ nasalized
2. [eI] ∼ [aI]: mid ∼ low
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 14 / 98
contrast
not all phonetic di=erences are equal
◮ vowel height causes contrast: pale 6= pile
◮ vowel clipping and nasality never cause contrast, no minimal pairs
◮ vowel height: part of language (part of the inventory), contrastive,phonemes
◮ vowel clipping and nasality: not part of language (not part of theinventory), redundant, allophones
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 15 / 98
contrast
predictability
s v – [eI]
s f – [eI]
s n – [eI]
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 16 / 98
contrast
unpredictable: [eI] or [aI]?
p l – both [eI] or [aI]
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 17 / 98
contrast
two levels of analysis = two approaches to sounds
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 18 / 98
contrast
two levels of analysis = two approaches to sounds
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 19 / 98
contrast
two levels of analysis = two approaches to sounds
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 20 / 98
contrast
two levels of analysis = two approaches to sounds
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 21 / 98
contrast
phonology: mapping phonemes onto allophones
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 22 / 98
contrast
generative model of phonology
◮ underlying/lexical representation (input): only those sound propertiesthat are contrastive (= phonemes)
◮ phonological rules that derive the surface representation from theunderlying representation
◮ surface representation (output = speech): contrastive + predictablesound properties
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 23 / 98
contrast
underlying representation: only contrastive properties
(underspecification)
Properties b eI s
V-height mid
V-clipping !
Voicing voiced ! voiceless
Nasality oral ! oral
C-place labial alveolar
C-manner stop fricative
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 24 / 98
consonant contrasts
consonant contrasts in English
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 25 / 98
consonant contrasts
the consonant inventory (contrastive consonants)
◮ sonorants◮ glides: /j w/◮ liquids: /l r/◮ nasals: /m n N/
◮ obstruents◮ stops: /p b t d k g/◮ fricatives: /f v T D s z S Z h/◮ a=ricates: /Ù Ã/
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 26 / 98
consonant contrasts
the contrast of obstruents in English
what’s the contrastive feature in these?:
tie – diewriter – riderbeat – bead
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 27 / 98
consonant contrasts
main questions
◮ what features make the contrast possible?
◮ do these features “survive” in all position?
◮ are they the same in other languages, like HUN & GER?
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 28 / 98
consonant contrasts
obstruent groups
OBSTRUENTSStops & A=ricates Fricatives
/p/ /b/ /f/ /v//t/ /d/ /T/ /D//k/ /g/ /s/ /z//Ù/ /Ã/ /S/ /Z/
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 29 / 98
laryngeal contrast
the traditional view
tie ↔ die/t/ ↔ /d/
file ↔ vile/f/ ↔ /v/
“voiceless” ↔ “voiced”
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 30 / 98
laryngeal contrast
laryngeal contrast in obstruents
◮ the phonological contrast of “voicing” is signalled (“cued”)by a complex of features: there are several correlates of this contrast
◮ vocal fold vibration is only one of them
◮ let’s refer to the phonological contrast as laryngeal contrast
◮ voicing is a narrowly used phonetic term: vocal fold vibration (alsocalled: phonation)
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 31 / 98
laryngeal contrast
some phonetic correlates of laryngeal contrast in
obstruents
◮ voicing/phonation: vocal fold vibration
◮ Voice Onset Time (VOT)
◮ relative length of preceding vowel
◮ glottalization
◮ release noise/burst: intensity & length
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 32 / 98
laryngeal contrast voicing
framework of the larynx
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 33 / 98
laryngeal contrast voicing
the vocal folds + glottis in the larynx
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 34 / 98
laryngeal contrast voicing
states of the vocal folds
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 35 / 98
laryngeal contrast voicing
cross-section of the larynx
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 36 / 98
laryngeal contrast voicing
vocal fold vibration: the Bernoulli e=ect
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 37 / 98
laryngeal contrast voicing
steps of vocal fold vibration
vocal fold vibration happens because of air pressure changes (aerodynamicreasons):
1. vocal folds loosely close
2. air pressure increases below vocal folds
3. air pressure blows vocal folds apart (glottis opens)
4. speed of air particles increases through narrow glottis
5. air pressure decreases below/within vocal folds ⇒ vocal folds suckedtogether (Bernoulli e=ect)
6. vocal folds are closed again, a cycle like this repeats itself approx.100–300 times/second
7. the cycles last until the state of glottis changes (e.g., opens to producea voiceless sound)
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 38 / 98
laryngeal contrast voicing
types of voicing
1. passive/modal/spontaneous voicing: open oral cavity – this helps tostart and maintain voicing because air pressure will be low in themouth but high below vocal folds; sonorants
2. passive devoicing: closure/constriction in mouth – this creates highair pressure above vocal folds, which inhibits vocal fold vibration;obstruents
3. active voicing: extra articulatory e=ort is needed to maintain voicing inobstruents, e.g., lower larynx to enlarge oral cavity
4. active devoicing: extra articulatory e=ort to maintain voicelessness
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 39 / 98
laryngeal contrast voicing
phonetic di;culty of voicing in obstruents
◮ voicing is di;cult to maintain in obstruents: they prefer to be devoiced
(every language has voiceless stops, there are languages that onlyhave voiceless stops, but none that have only voiced stops)
◮ languages use 2 strategies:◮ additional articulatory gestures (active voicing): Hungarian, Spanish,
Polish, Dutch. . .◮ partial or full devoicing (passive devoicing): English, German, Swedish,
Norwegian, Danish. . .
◮ lenis stops: ‘weakly voiced’, ‘not voiced in all positions’
◮ in English: ‘voiced’ obstruents are typically devoiced, except between 2sonorants: ready, bandit. . .
◮ in Hungarian: “voiced” obstruents are typically voiced in all positions(initially and word-finally, too): bab, babos. . . ; cf. voicing assimilationthough
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 40 / 98
laryngeal contrast voicing
classic English literature: Jones
Jones (1918: 154)
In voiced plosive consonants the amount of voice heard during the stopmay vary. [. . . ] When a voiced plosive [. . . ] occurs between two vowels (asin about), voice sounds throughout the whole of the stop.
In English when /b d/ and /g/ occur initially [. . . ], they are partially devoiced[. . . ] i.e. voice is not heard during the whole of the stop but only during partof it, generally the latter part. With some speakers the voice disappearsaltogether [. . . ].
With many speakers [. . . ] final voiced plosives [are] partially or evencompletely devoiced [. . . ]. [These] consonants are very weak voicelessplosives consonants.
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 41 / 98
laryngeal contrast voicing
classic English literature: Gimson
Gimson (1962: 32; 152)
A voiceless/voiced pair such as [s, z] are distinguished not only by thepresence or absence of voice but also by the degree of breath and musculare=ort involved in the articulation. Indeed, [. . . ] in certain situations, thevoice opposition may be lost, so that the energy of articulation becomes asignificant factor. Those English consonants which are usually voiced tendto be articulated with relatively weak energy, whereas those which arealways voiceless are relatively strong. Thus, it may be important to define[s] as strong or fortis and [z] as weak or lenis.
The lenis series /b d g/ may have full voicing [. . . ] when they occur [. . . ]between voiced sounds e.g. labour, leader, eager [. . . ]. In initial andespecially in final positions, [. . . ] while remaining lenis, may be partiallyvoiced or completely voiceless.
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 42 / 98
laryngeal contrast VOT
laryngeal contrast in stops/plosives: Voice Onset Time
definition
the time between the release of the stop and the start of vocal fold vibrationof the vowel or sonorant
direction of VOT
◮ positive (aspiration, fortis stops = voiceless aspirated)
◮ zero (neutral/lenis = voiceless unaspirated)
◮ negative (truly voiced stops (unaspirated))
length of VOT
– in the case of positive and negative VOT: short or long
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 43 / 98
laryngeal contrast VOT
positive, long-lag VOT: aspiration (#paces)
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 44 / 98
laryngeal contrast VOT
aspiration in closeup
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 45 / 98
laryngeal contrast VOT
negative VOT: voiced stop (Spanish)
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 46 / 98
laryngeal contrast VOT
negative VOT: voiced stop (Spanish)
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 47 / 98
laryngeal contrast VOT
positive VOT: aspirated, voiceless stop = fortis
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 48 / 98
laryngeal contrast VOT
positive VOT: aspirated, voiceless stop = fortis
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 49 / 98
laryngeal contrast VOT
zero VOT: unaspirated, voiceless stop = neutral
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 50 / 98
laryngeal contrast VOT
zero VOT: unaspirated, voiceless stop = neutral
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 51 / 98
laryngeal contrast VOT
zero VOT: unaspirated, voiceless stop = neutral
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 52 / 98
laryngeal contrast VOT summary
VOT typology: summary
– contrast based on VOT: 2 main types of languages
1. voicing languages:zero VOT ⇔ − VOT= voiceless-unaspirated ⇔ voiced-unaspiratede.g., Spanish, Hungarian, French, Dutch. . .
2. aspirating languages: zero VOT ⇔ + VOT= voiceless-unaspirated (lenis) ⇔ voiceless-aspirated (fortis)e.g., English, German, Danish. . .
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 53 / 98
laryngeal contrast VOT summary
comparing English & Hungarian
ENG HUN
pig big pig(ment) big(ott)
aspirated? + − − −
voiced? − − − +
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 54 / 98
laryngeal contrast VOT summary
warning!
◮ be careful: don’t be mislead by spelling
◮ spelling uses arbitrary symbols to indicate contrast
◮ if two languages use the same letter, it does not necessarily mean that
the contrast is based on the same phonetic feature!
◮ the p letters are the same, but the phonetic content is very di=erent:
E. pig 6= H. pigment; E. big 6= H. bigott
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 55 / 98
laryngeal contrast VOT summary
contrast in English vs. Hungarian
E pig—big: /ph/—/p/
H pig(ment)—big(ott): /p/—/b/
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 56 / 98
laryngeal contrast positions
laryngeal contrast of stops/plosives vs. positions
◮ so far two phonetic features have been used for the laryngeal contrast:voicing & aspiration
◮ they are not equally active in all phonetic positions
1. between sonorants, before a stressed vowel: repél – rebél2. word-initial, before a stressed or unstressed vowel: tíe – díe3. between sonorants, before an unstressed vowel: wríter – ríder
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 57 / 98
laryngeal contrast positions
1. between sonorants, before a stressed vowel
repél rebélvoiced? − +aspirated? + −
– both features are active in this position for the contrast
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 58 / 98
laryngeal contrast positions
2. word-initial, before a stressed or unstressed vowel
tíe díevoiced? − −
aspirated? + −
– only aspiration is active in this position for the contrast
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 59 / 98
laryngeal contrast positions
3. between sonorants, before an unstressed vowel
wríter rídervoiced? − +aspirated? − −
◮ only voicing is active in this position for the contrast
◮ note: length of stops is relatively short here, and voicing may continuethroughout the stop: /t/ may become a flap [R] in American English,but not /p/ or /k/: rápid – rábid still contrast
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 60 / 98
laryngeal contrast word-final
absolute word final position: beat – bead
◮ in this position, voicing is di;cult to maintain
◮ since nothing follows the stop, aspiration is also impossible
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 61 / 98
laryngeal contrast word-final
no contrast in beat – bead?
beat beadvoiced? − −
aspirated? − −
◮ has English given up contrast in word-final position?
◮ or maybe there are features other than voicing that get activated hereto maintain the contrast. . .
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 62 / 98
laryngeal contrast word-final
halfway summary
◮ overall topic: how to model phonological contrast
◮ focus: laryngeal contrast in consonants (obstruents)
◮ 2 models: classical generative vs. phonetically-grounded
◮ laryngeal contrast is signalled/“cued” by several phonetic features
◮ features so far: voicing/phonation, VOT
◮ these features are not equally active in all positions
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 63 / 98
laryngeal contrast word-final
ranking of positions based on laryngeal contrast
preservation in stops
1. medial, between sonorants, before a stressed V (repél – rebél) >2. word-initial, before a stressed/unstr. V (tíe – díe) >
3. medial, between sonorants, before an unstressed V (wríter – ríder)
X > Y = ‘X has more/better features to maintain the contrast than Y’(where X and Y are positions)
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 64 / 98
laryngeal contrast word-final
no contrast in beat – bead?
beat beadvoiced? − −
aspirated? − −
◮ has English given up contrast in word-final position?
◮ or maybe there are features other than voicing that get activated hereto maintain the contrast. . .
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 65 / 98
laryngeal contrast neutralization
neutralization: the beer goggle e=ect
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 66 / 98
laryngeal contrast neutralization
neutralization
The disappearance of contrast under a given condition.(= The local suspension of a phonological oppositionbetween two or more contrastive sound segments.)
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 67 / 98
laryngeal contrast neutralization
neutralization examples: vowel reduction
◮ a wide range of vowels can appear in a stressed syllable but inunstressed syllables, vowel contrast is reduced to a handful of vowels(primarily the schwa)
◮ senténtial ∼ séntence e ∼ @
systémic ∼ sýstem e ∼ @
morálity ∼móral æ ∼ @
symbólic ∼ sýmbol 6 ∼ @
atómic ∼ átom 6 ∼ @
harmónious∼ hármony oU ∼ @
mystérious ∼mýstery I@ ∼ @
dráma ∼ dramátic A: ∼ @
sulphúrious ∼ súlphur jU@∼ @
◮ opposition: full vowels, condition: unstressed syllable, output: /@/
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 68 / 98
laryngeal contrast neutralization
neutralization examples: /s/ and /S/
◮ /s/ is in contrast with /S/
◮ so – show, mass – mash, parcel – partial, universal – controversial, etc.
◮ word-inital, pre-consonantal position: /S/ only before /r/, and /s/ isbefore any other consonant
◮ /Sr/: shrub, shrivel, shrink, shrug. . . but never */St/, */Sp/, */Sk/, etc.
◮ /s/ + C: steam, sport, sky, etc.
◮ opposition: /s/ – /S/, condition: word-inital, pre-consonantal position,output: either /Sr/ or /s/ + C (where C 6= /r/)
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 69 / 98
laryngeal contrast neutralization
neutralization examples: nasal + stop clusters
◮ nasals contrast with respect to place of articulation
◮ sin – SIM – sing: /n/ – /m/ – /N/
◮ before a stop: only one can occur, whose place depends on followingstop
◮ e.g., print /nt/, but no /m/ or /N/ before /t/
StopsNasal /p/ /t/ /k/
/m/ limp — —/n/ — tent —/N/ — — link /Nk/
◮ opposition: /n/ – /m/ – /N/, condition: before a stop, output: onlyone nasal can occur, the contrast is suspended before a given stop
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 70 / 98
laryngeal contrast neutralization
neutralization?
◮ laryngeal contrast in word-final position
◮ beat – bead, back – bag, loose – lose, leaf – leave, etc.
◮ opposition: obstruents, condition: word-final position, output: onlyvoiceless-unaspirated obstruents
◮ based on this, beat and bead are supposed to be pronounced the sameway: beat [bi:t] = bead [bi:t] (homophones)
◮ this does not seem to be the case!
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 71 / 98
laryngeal contrast neutralization
“redundant” features to the help
◮ correlates of laryngeal contrast so far: voicing and aspiration – butthey are not active in word-final position
◮ there are other correlates of the laryngeal contrast
◮ they seem to emerge more saliently when contrast is in danger (as inword-final position):
◮ relative length of preceding vowel◮ glottalization◮ other features: release noise, articulatory strength/e=ort/force
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 72 / 98
laryngeal contrast vowel length
relative length of preceding vowel
◮ vowels are shorter (clipped) before fortis obstruents than before lenisobstruents: Pre-Fortis Clipping
◮ speak – speed,mate – made,rope – robe,write – ride,root – rude,cap – cabloose – lose,leaf – leave
◮ clipping is redundant/predictable but it cues the contrast here
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 73 / 98
laryngeal contrast glottalization
pre-glottalization/glottal reinforcement
◮ glottal closure quickly closes down the voicing of the vowel, followedby the oral closure of the fortis stops & a=ricate
◮ happens word-finally or when they are followed by another consonant
◮ right [raIPt], shop [S6Pp], shot [S6Pt], shock [S6Pk], April ["eIPprl],fatness ["fæPtn@s], football ["fUPtbO:l], reach [ri:PÙ], etc.
◮ it only happens for the fortis consonants:mate [meIPt] – made [meIt],seat [siPt]– seed [si:t]
◮ it is another indicator of the fortis – lenis contrast!
◮ it happens where the contrast between fortis & lenis stops couldpotentially disappear
◮ note: glottalization may well be just a more salient/forceful versionof pre-fortis clipping: the vowel is cut by glottal closure
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 74 / 98
laryngeal contrast glottalization
contrast is salvaged in beat – bead
beat beadvoiced? − −
aspirated? − −
preceding V shorter? + −
glottalization? + −
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 75 / 98
laryngeal contrast glottalization
redundancy is actually important
◮ remember: classical phonemic/generative model claims that whateveris predictable has no “information value”, it’s noncontrastive, hencenot part of language (only part of speech)
◮ contrary to the classical phonemic/generative model,predictable-redundant features may actually be important to maintaincontrast in certain situations
◮ vowel-clipping is predictable, yet it is crucial in preserving contrastword-finally (beat – bead, bit – bid, rope – robe, etc.)
◮ it is only sporadically present in other contexts like repél – rebél, whereother features safely signal the contrast
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 76 / 98
laryngeal contrast other features
other features?
◮ research is ongoing whether other features play a role in laryngealcontrast preservation or not
◮ release noise: length and intensity seem to be only present in fortisstops
◮ articulatory strength: lenis stops are claimed to be articulated with lesse=ort
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 77 / 98
laryngeal contrast other features
release noise
◮ can be a cue of laryngeal contrast for stops in English
◮ the tongue is more saliently released after the fortis stops than afterthe lenis stops
◮ /t/ seems to have the noisiest release, it is often a=ricate-like /ts/
◮ “voicing” languages often use this cue for the contrast: voiced stopsmay have a voiced release, which may evolve into a schwa-like vowelsound (as in French, bag [bag@], buzz [b2z@])
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 78 / 98
laryngeal contrast other features
articulatory e=ort?
◮ traditional literature (Gimson) often cite this as a cue for laryngealcontrast
◮ fortis obstruents: more energy, articulatory e=ort, stronger contact ofthe articulatorslenis obstruents: relatively weak energy, less articulatory e=ort
◮ this e=ort di=erence is supposed to remain active in all positions,including word-finally
◮ problem: no reliable phonetic definition of ‘energy’, ‘e=ort’, ‘strength’exists, and experiments have failed to show its relevance in laryngealcontrast
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 79 / 98
laryngeal contrast other features
Ladefoged & Maddieson (1996) on articulatory strength
The Sounds of the World’s languages, pp. 96, 98:
“Measures of the force of contact between the articulators [. . .] generallyfailed to show that pairs such as /p/ and /b/ di=ered in the expected way,and the idea of articulatory strength was widely considered amongphoneticians to be discredited. [. . .] There does not seem to be anindependent use of articulatory strength as a contrastive parameter.”
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 80 / 98
laryngeal contrast other features
articulatory e=ort?
◮ traditional fortis may simply mean ‘voiceless strongly aspirated stop’,lenis: ‘a voiceless stop without aspiration’, and so fortisness/lenisnessis simply a synonym for + vs. zero VOT (aspiration vs. no aspiration)
◮ articulatory strength is what may cause the length di=erence betweenclosure: fortis stops are longer (and the vowel before them is shorter)than lenis stops
◮ it may only be a relevant factor in the laryngeal contrast of fricatives:feel – veal, file – vile, sip – zip, leaf – leave, bus – buzz, etc.
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 81 / 98
laryngeal contrast fricatives
correlates of laryngeal contrast for fricatives
◮ /T/ – /D/, /f/ – /v/, /s/ – /z/, /S/ – /Z/
◮ voicing/phonation, preceding vowel length and intensity signal thecontrast
◮ aspiration, glottalization, release do not seem to play a role
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 82 / 98
laryngeal contrast fricatives
1. medial, between sonorants, before a stressed V
◮ conféction – convéction, defíed – divíde
◮ voicing/phonation is the primary cue, no danger for contrast
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 83 / 98
laryngeal contrast fricatives
2. word-initial, before a V
◮ sip – zip, cellar – Zellar, fain – vein, fault – vault, feel – veal,sheet /Si:t/ – gite /Zi:t/, thigh /TaI/ – thy /DaI/
◮ voicing/phonation is the primary cue: initial fricatives seem to beactively voiced in English (unlike the stops), no danger for contrast
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 84 / 98
laryngeal contrast fricatives
3. medial, between sonorants, after a stressed V
◮ míssle – mízzle, grístle – grízzle, rífle – ríval, Óphir /"@Uf@/ – óver/"@Uv@/, Áisha – Ásia, Ásher – ázure, tréssure – tréasure, Confúcian –confúsion
◮ voicing/phonation is the primary cue, no danger for contrast
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 85 / 98
laryngeal contrast fricatives
4. absolute word-final position
◮ leaf – leave, brief – breve, calf – calve, safe – save, bus – buzz, race – raise,hiss – his, ruche /ru:S/ – rouge /ru:Z/, teeth /ti:T/ – teethe /ti:D/, loath
– loathe
◮ for similar reasons as for stops, vocal fold vibration in this position isdi;cult to maintain
◮ relative vowel and consonant length emerge to maintain the contrast
◮ /T f s S/: have a shorter vowel before them and they are articulatedlonger than
◮ /D v z Z/: preceding vowel is relatively longer and they are articulatedrelatively shorter
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 86 / 98
laryngeal contrast fric + stop
fortis fricative + stop clusters
◮ so far we have not seen neutralization of the laryngeal contrast
◮ fortis fricative + stop clusters:◮ /s/ + C: speak, sport, spring, stéreo, stúpid, string, school, scheme, sketch,
discóver, displáy, expláin. . .◮ /f/ + C: caftán, fiftéen
◮ the laryngeal contrast is completely neutralized in this position: onlyan unvoiced-unaspirated stop (= lenis) may occur here (especially after/s/)
◮ this stop “sounds” like a usual lenis stop for native speakers when weget rid of the fricative – DEMO
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 87 / 98
laryngeal contrast ranking
ranking of positions based on laryngeal contrast
preservation in stops
1. medial, between sonorants, before a stressed V (repél – rebél) >2. word-initial, before a stressed/unstr. V (tíe – díe) >
3. medial, between sonorants, before an unstressed V (wríter – ríder) >4. absolute word-final (beat – bead) >
5. after fortis fricatives (sport/stop/school)
X > Y = ‘X has more/better features to maintain the contrast than Y’(where X and Y are positions)
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 88 / 98
laryngeal contrast voicing assimilation
regressive voicing assimilation (RVA)
◮ when two (or more) obstruents with di=erent laryngeal specificationsstand next to each other: C1C2
◮ within the same word (C1C2) or across a word boundary (C1# C2)
◮ laryngeal specification of C2 influences/spreads to/is assimilated by
C1: he was sent /z/ + /s/ → [ss], good time /d/ + /t/ → [tt]
◮ if either only fortis+fortis or lenis+lenis clusters are possible, then thiswould be laryngeal neutralization of obstruents before anotherobstruent
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 89 / 98
laryngeal contrast voicing assimilation
RVA in Hungarian is neutralizing
◮ háztól ‘from house’ /z/+/t/ → [st]
◮ népzene ‘into the flour’ /p/+/z/ → [bz]
◮ Koszos lettem a mé[st]ol.‘I became dirty from the ?lime / ?honey’mész /me:s/ ‘line’ ↔ méz /me:z/ ‘honey’mésztol = méztol: complete neutralization
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 90 / 98
laryngeal contrast voicing assimilation
sonorants do not cause RVA
◮ mésznek ‘for the lime’ /s/ + /n/ → *[zn]
◮ mésznek does not become méznek: no laryngeal neutralization
◮ képnek ‘for picture’ [pn] *[bn], töröknél [kn] *[gn] ‘at Turk(ish)’,zokni ‘socks’ [kn] *[gn]
◮ reason: sonorants are passively voiced, passive voicing cannot spreadto other sounds
◮ only actively voiced and actively devoiced/fortis sounds can spreadtheir voicing and devoicing/fortis feature to other sounds – as inHungarian for example
◮ in English too: batman [tm] *[dm], putney [tn] *[dn], replay [pl] *[bl]
◮ Slovak is exceptional: Krásny kvet má pät’ malých lupienkov. [dm]‘The beautiful flower has 5 petals.’Položili kvet a . . . . [da]‘They laid the flower and. . . ’
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 91 / 98
laryngeal contrast voicing assimilation
laryngeal properties of word-initial obstruents in English
1. /p t k T f s S/: contain active devoicing/fortisness, prediction: theycan cause devoicing in preceding obstruents
2. /D v z Z/: seem to contain active voicing, prediction: they may causevoicing in preceding obstruents
3. /b d g/: do not contain a voicing feature, prediction: they do notcause voicing in preceding obstruents
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 92 / 98
laryngeal contrast voicing assimilation
fortisness of /p t k T f s S/ can spread in English
◮ is Pete going? [sp], live show [fS], grade four [tf ], bead show [ts]
◮ devoicing can spread (C1 becomes voiceless)
◮ but the contrast is not neutralized as other correlates of the contrastdo not change: the vowel will not become shorter (and if C1 is a stop,there is no glottalization either)!
◮ thus: bead show will not become beat show even though both finalconsonants are voiceless
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 93 / 98
laryngeal contrast voicing assimilation
contrast in beat show vs. bead show
beat show bead showvoiced? − −
aspirated? − −
preceding V shorter? + −
glottalization? + −
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 94 / 98
laryngeal contrast voicing assimilation
/D v z Z/ may spread (some) voicing in English
◮ work zebra ?[gz], what’s this? ?[zD], beat Zoë ?[dz]
◮ voicing from the lenis fricatives may spread to C1
◮ even though voicing may spread, the contrast is not neutralized asother correlates of the contrast do not change: the vowel will stayshorter and there may be glottalization, too!
◮ thus: beat Zoë will not become bead Zoë even if both final consonantsare (partially) voiced
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 95 / 98
laryngeal contrast voicing assimilation
contrast in beat Zoë vs. bead Zoë
beat Zoë bead Zoëvoiced? (+) (+)
aspirated? − −
preceding V shorter? + −
glottalization? + −
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 96 / 98
laryngeal contrast voicing assimilation
/b d g/ do not spread voicing in English
◮ pop group [pg], beat band [tb], black dress [kd], birthday [Td],wishbone [Sb]
◮ there is no laryngeal neutralization in this context
◮ thus: beat band will not become bead band – C2 does not influence C1
at all: all laryngeal features will remain in beat, including itsvoicelessness
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 97 / 98
laryngeal contrast voicing assimilation
contrast in beat band vs. bead band
beat band bead bandvoiced? − −
aspirated? − −
preceding V shorter? + −
glottalization? + −
z. g. kiss (elte|delg) analysis 3–4 | consonant contrast 98 / 98