Parametric description and long-domain features

1. Segmental vs. parametric views of speech

In studying the IPA alphabet, we may become so focussed on the linear sequence of consonants and vowels that we forget the fundamental fact that speech is produced by six independently-controlled organs (lungs, larynx, jaw, velum, tongue and lips) all moving simultaneously and continuously.

Multiple parallel articulatory parameters
The figures above show the up-and-down movements of small metal pellets placed on several articulators (two on the tongue) during two utterances of the phrase "perfect memory". The raising of the tongue dorsum is labelled κ, of the tongue tip τ and of the lower lip β. It can be seen that in the upper version, in which the two words are clearly separated, movements κ and τ are strongly overlapping, whereas β comes later. In the lower version, in which the two words are run together, movements κ, τ and β are all rather strongly overlapping.

2. Pitch, tone and intonation

Pitch refers to the perception of relative frequency (e.g. perceptually high-pitched or low-pitched) of the vibrating vocal cords.

Intonation refers to the rise and fall of voice pitch over entire phrases and sentences, as in English:


3. Terminology. For this reason, intonation is sometimes called a supra-segmental feature. This terminology implies that phonetic and phonological features can be separated into two subsets: segmental and suprasegmental features. This is not really correct, as we shall soon see. For this reason, I prefer to avoid the term "suprasegmental". Instead, I shall talk about the prosodic function of features. Almost any feature can be used prosodically or segmentally, depending on the language.

In English, we also use pitch (in conjunction with other features) to distinguish a few pairs of words; e.g. DEfect (n.) vs. deFECT.

It is quite common for pitch and stress to be closely related. In some languages, in fact, pitch seems to play a role that is rather like stress in English: these are often referred to as pitch-accent languages. E.g. Croatian:


Long rising

Short falling
Listen to Croatian vǐle 'fairies' ʋǐːle vile ʻfairiesʼ
Listen to Croatian vîle 'hayfork' ʋîle vile ʻhayforkʼ
Listen to Croatian dǔga 'rainbow' dǔːɡa duga ʻrainbowʼ
Listen to Croatian dûga 'stave' dûga duga ʻstaveʼ

Tone refers to significant (i.e. meaningful, constrastive, phonemic) constrasts between words signalled by pitch differences. Tone may be lexical, as in Mandarin Chinese:
 
Tone number Description
IPA transcription example Meaning
1 high level [má] ʻmotherʼ
2 high rising [māá] ʻhempʼ
3 low (falling+)rising [màá] ʻhorseʼ
4 high fall [mâ] ʻscoldʼ
"no tone/neutral tone" (depends on preceding syllable)
[ma] (question marker)

Or grammatical tone, as in many African languages, e.g. Edo:
 
Tense Monosyllabic verbs Disyllabic verbs
Timeless [ì mà] ʻI showʼ [ì hrùlè] ʻI runʼ
Continuous [í mà] ʻI am showingʼ [í hrùlé] ʻI am runningʼ
Past [ì má] ʻI showedʼ [ì hrúlè] ʻI ranʼ

However, there may also be non-pitch aspects of tone. Lexical tones are often related to durational, phonatory and vowel quality distinctions as well as frequency distinctions. For example, Mandarin Chinese tone 3 (low rise) is long with creaky voice, Hunanese tone 2 has breathy or chesty voice. Tibetan tone 1 words have voiceless initial consonants whereas tone 2 words have voiced beginnings. Long vowels in tone 4 or 5 open syllables in Thai are checked by a final glottal stop.

All of these variations in pitch require the tension of the vocal cords to be adjusted up and down at the same time as the other articulators are moving, as in Browman and Goldstein's diagram in section 1 above.

4. What else can be a prosody? Virtually anything!

4.1. Laryngeal features

In addition to tone and intonation, several other laryngeal features are prosodic in some languages. In Cœur d'Alène, a dying Salishan language spoken in northern Idaho by few speakers ("less than a handful" in 1966), glottalization (probably creaky voice, glottal stops and glottalic consonants) is used as a diminutive morpheme: 

mar-marím-EntEm-ilts vs. m̰-m̰ar̰-m̰ar̰ím-En̰tem̰-il̰ts
\___________________/
\__________________________/
not glottalized
glottalized
ʻthey were treated one by oneʼ
ʻthey little ones were treated one by oneʼ

A complex of creaky voice, pitch and loudness in Danish (stød) distinguishes words in a way that parallels the use of contrastive tones in Swedish or Norwegian, e.g. [bœn] ʻbeansʼ vs. [bœnʔ], or [thɒŋgn̩] ʻthe thoughtʼ vs. [thɒŋʔgn̩]. (See Laver 1994 and Grønnum 1998 for further examples.) Although I have transcribed these using a glottal stop symbol, the glottalization is more dispersed, and may involve creaky voiced vowel, creaky voiced sonorant consonants, falling pitch, and glottal closure. It is therefore usually analysed as a prosodic feature, rather than a segmental feature. Fischer-Jørgensen (1987) is a very thorough study (part of it conducted here in the Oxford Phonetics Laboratory in 1981! A copy is in the lab library.)

Voicing assimilation is another (perhaps less spectacular) example of a feature not staying nicely confined within a segment. For example, in Russian, there is a maximum of 1 voicing distinction in consonant clusters. A cluster takes on the voicing of the last consonant in the sequence, without regard to boundaries:

Morphemic boundary: /gorod+k+a/  
[gorotka] ʻlittle townʼ
Clitic boundary: /mtsensk# bi/  
[mtsenzgbɨ] ʻif Mcenskʼ
Word boundary: /mtsensk## bil/  
[mtsenzgbɨl] ʻit was Mcenskʼ

Sonorants e.g. [m] are not distinctively voiced, and they are transparent to voicing assimilation:

iz # mtsensk+a  
[is mtsenska] ʻfrom Mcenskʼ
ot # mzd+i
[od mzdɨ] ʻfrom the bribeʼ

Some French examples:

anticipation of [+voice]:    
anticipation of [-voice]:
je passe vite [paz vit]

chemin de fer [t fɛːʁ]
la tête droit [tɛd dʁwat]

coup de pied [ku t pje]
avec vous [avɛg vu]

esprit de corps [ɛspʁi t kɔːʁ]
place d'armes [plaz daʁm]

tout de suite [tu t sɥit]


rez-de-chaussée [ʁe t ʃose]


4.2. Velum movement
 

Terena/Tereno (southwestern Mato Grosso, Brazil) (Bendor-Samuel 1960).

3rd person possessives
(non-nasal)

1st person possessives
(nasal)
eˈmoʔu ʻhis wordʼ
ẽˈmõʔũ ʻmy wordʼ
ˈajo ʻhis brotherʼ
ˈãj̃õ ʻmy brotherʼ
ˈowoku ʻhis houseʼ
ˈõw̃õŋgu ʻmy houseʼ

Sundanese (Austronesian, W. Java) (Robins 1957):

ɲaian [ɲãĩãn], ʻto wetʼ
miasih [mĩʔãsih], ʻto loveʼ
kumaha [kumãhã], ʻhow?ʼ

4.3. Tongue root prosodies

4.3.1. Spread of velo-pharyngealization in Tashlhiyt Berber. This language (like various other Afroasiatic languages, such as Arabic) has a contrast between pharyngealized and non-pharyngealized coronal consonants, as well as some pharyngeal consonants. In words containing these consonants, entire syllables or even the whole word becomes pharyngealized, as you can hear from these audio clips:

lYwr!d  lˤwrˤd ʻhillʼ
Listen to Berber lYwr! 'to grease'  zʁwrˤ ʻto greaseʼ
Listen to Berber zr! 'to watch'  zrˤ ʻto watchʼ
Listen to Berber t!n!gd!t! 'you drowned'  tˤnˤgdˤtˤ ʻyou drownedʼ
Listen to Berber nnqqwr!t! 'silver'  nnqqwrˤtˤ ʻsilverʼ

4.3.2. [ATR] harmony

In a number of languages of West Africa, a particular pattern of vowel harmony is found in which the vowels fall into two sets: more advanced and higher vowels {i, e, o, u} and correspondingly less advanced and lower {ɪ, ɛ, ɔ, ʊ}. [a] is typically "neutral", and can co-occur in words with vowels of either set. In the more advanced and higher vowels, the tongue root has been found to be more advanced, and hence they are given the feature [+Advanced Tongue Root] or [+ATR]: the non-advanced vowels are consequently labelled [-ATR]. See Laver (1994: 289-291) and references therein, and Tiede (1996) for MRI images of Akan [i] vs. [I]. The vowel system of Ega  (Kwa, Côte d'Ivoire) (Connell, Ahoua and Gibbon 2002) is quite typical:
 

High, [+ATR]

i

u

High, [-ATR]

I

ʊ
Mid, [+ATR]
e

o

Mid, [-ATR]

ɛ

ɔ
Low



a


Examples of words containing vowels in the two sets (note that Ega also has tones, not transcribed here):


[+ATR]



[-ATR]

Listen to Ega efi 'eye' efi
ʻeyeʼ

ɔvɛ ʻsilk cotton treeʼ

oji
ʻcoldʼ
Listen to Ega UvE 'dog' U ʻdogʼ

efe
ʻtaking flightʼ

ɛɲɛ ʻarrivalʼ

uje
ʻbundle of woodʼ
Listen to Ega OsI 'woman' ɔsI ʻwomanʼ

emo
ʻsmellingʼ

ɛzɔ ʻquarrelʼ

ize
ʻwoodʼ

ɛcI ʻlaughingʼ

4.4. Tongue body prosodies; lip prosodies

Vowel harmony systems involving frontness vs. backness, openness vs. closeness or lip-rounding vs. lip-spreading are widely documented in the phonetics and phonological literature, and exemplify the prosodic use of tongue body and lips.

4.5. Tongue tip prosodies

Finally, a number of languages have harmony systems involving tongue-tip consonants: all the coronal consonants of a word are either more advanced. [+anterior], or less advanced, [-anterior]. In some languages this distinction is dental vs. alveolar, in others alveolar vs. post-alveolar. In Coleman (2003), I present evidence that English also exhibits this kind of [anterior] harmony to some extent. For example:

Chumash (Chumashan, California (extinct))
k-sunon-us
ʻI obey himʼ

k-ʃunon-ʃ ʻI am obedientʼ
Tahltan (Athabaskan, NW Canada)
ɛsdɑn
ʻI'm drinkingʼ

dɛθkwʊθ ʻI coughʼ
Zayse (Omotic, Ethiopia)
zatst
ʻleadʼ

ʔiʃitʃ ʻfiveʼ
English
ʌthəsɑɪ
ʻutter sighʼ

ʌthəʃɑɪ
ʻutter shyʼ


4.6. Manner of articulation (degree of stricture)

5. Prosody reconsidered

The preceding examples have shown that many more features than duration and pitch are (or can be) prosodic. What, then, do we mean by ʻprosodyʼ if not just tone, stress, intonation, and loudness, the traditional ʻsuprasegmentalʼ features?

a) Features (or groups of features) that not located at a single place in the sequence of consonants and vowels.

b) For example, (groups of) features associated with a whole syllable, word or phrase.

c) Also, features of the boundaries of syllables and words (e.g. assimilation, liaison, absence vs. presence of sounds in particular syllable or word positions). ʻGrenzsignaleʼ (Trubetzkoy 1969: 273-297). Recognition of the non-segmental behaviour of features, and the close relationship between features and the specific places in syllable or word position in which they occur, led to the origin in the mid 1930ʼs to the London school of "prosodic phonology", under J. R. Firth, which presented a thorough critique of and offered a theoretical alternative to phonemic phonology. Some of the ideas of the prosodic school have influenced the contemporary mainstream of phonological theory, especially via the framework of Autosegmental Phonology.

References

Bendor-Samuel, John T. (1960) Some problems of segmentation in the phonological analysis of Tereno. Word 16, 348-355.

Coleman, John (2003) Discovering the acoustic correlates of phonological contrasts. Journal of Phonetics 31, 351-372.

Connell, Bruce, Firmin Ahoua and Dafydd Gibbon (2002) Ega. Journal of the International Phonetic Association 32, 99-104.

Fischer-Jørgensen, Eli (1987) A phonetic study of the stød in Standard Danish. Annual Report of the Institute of Phonetics, University of Copenhagen vol. 21. 55-265.

Grønnum, Nina (1998) Danish. Journal of the International Phonetic Association 29, 99-105.

Laver, John (1994) Principles of Phonetics. Cambridge University Press.

Robins, Robert H. (1957) Vowel nasality in Sundanese: a phonological and grammatical study. Studies in Linguistic Analysis. The Philological Society. 87-103.

Tiede, Mark (1996) An MRI-based study of pharyngeal volume contrasts in Akan and English. Journal of Phonetics 24, 399-421.

Trubetzkoy, N. S. (1969) Principles of Phonology. (English translation of Grundzüge der Phonologie.) University of California Press.