f | share    f | like 

Phonology: Part II

Features & Phonemes

Because our writing system is more or less based on phonemes (represented by letters), we easily mistake phonemes as being the fundamental building blocks of phonology. However, phonemes are made up of features. Therefore it is with reference to distinctive features that we shall showcase a generalization of the Inventory of Phonemes of the English language… this Inventory being what is mistakenly called 'the alphabet'. Be advised that the analysis done here is not necessarily the one and only Ultimate, although it is assuredly a sound general summary based on my impressions and interpretations of others' scholarly analyses.

Traditionally, the great phonemic divide is between consonants and vowels, consonants being articulated with a major obstruction in the oral cavity, vowels not. Another great divide is between those phonemes that are 'singable' or not — all the vowels and some consonants being 'singable', which is to say sonorants; some of the consonants not being 'singable', and those are the obstruents.

The obstruents break down into the groups plosives, fricatives and affricates; and then there is an overlapping group called sibilants, which includes some of the fricatives, and (in English) all the affricates. The plosives are non-continuent because they are articulated with a complete closure that completely halts the aistream before it is released orally (as opposed to nasally): p, t, d, k, g ('g' in 'get' not 'gem'), kw ('qu' in 'quark'), and the glottal stop ʔ (the 'breath-catch' in 'uh-oh!'). ʔ is not so much a phoneme as a semi-phoneme because native English speakers are naturally only semi-cognizant of it. The fricatives are continuents because the airstream is forced through the articulators, creating great friction: f, v, θ ('th' in 'thin' but not 'there'), ð ('th' in 'there' but not 'thin'), s, z, š ('sh' sound), ž ('s' in 'measure', 'z' in 'azure'), and arguably h (as in 'hoe'). The affricates are non-continuent with a continuent release, that is, a plosive held tight to a fricative: t + š => č ('ch' sound), and d + ž => ǰ ('g' in 'gem', 'j' in 'jump'). An overlapping class of obstruent is the sibilants, which are those obstruents articulated with a longitudinal groove along the surface of the tongue: fricatives s, z, š, ž and affricates č, ǰ.

Not only the obstruent consonants can be non-continuent, but also some of the sonorant consonants: the nasals m, n, ŋ ('ng' sound), which are formed with a complete closure in the oral cavity that completely halts the airstream, just like plosives, but, unlike plosives, the airstream is then re-routed through the nose. The non-continuent consonant classes plosives and nasals are known collectively as stops.

The sonorant consonants divide between those that are more consonantal, and the semivowels. The more consonantal sonorants are the nasals m, n, ŋ and the liquids l, r. The semivowels are glides w (as in 'woe') and j ('y' in 'yo'), and arguably the 'aspirate' h (as in 'hoe'). What makes the glides be semivowels is the fact that all that prevents w, j from being vowels ū ('oo' in 'loom'), ī ('ee' in 'feet') respectively is that whereas vowels ū, ī like all vowels are syllabic (they form the nucleus of a syllable), glides are not syllabic. What makes h a semivowel is the fact that (at least in English) it is nothing but a voiceless version of whatever vowel sound comes directly after it: all vowels are voiced (which means the vocal cords vibrate), but h is voiceless (the vocal cords do not vibrate).

The classes we have gone over represent one of the three dimensions of consonants — that of manner of articulation. Another is voicing — whether or not the vocal cords are vibrating (voiced), or not vibrating (voiceless). The voiced consonants are: b, d, g ('g' in 'get' not 'gem'), v, ð ('th' in 'there' but not 'thin'), z, ž ('s' in 'measure', 'z' in 'azure'), ǰ ('g' in 'gem', 'j' in 'jump'), all the nasals m, n, ŋ ('ng' sound), all the liquids l, r, all the glides w, j (like 'y' in 'yo' but not 'busy'). The voiceless consonants are: p, t, k, kw ('qu' in 'quark'), ʔ (the semi-phonemic 'breath-catch' in 'uh-oh!'), f, θ ('th' in 'thin' but not 'there'), s, š ('sh' sound), č ('ch' sound) and h (as in 'hoe').

The other feature-dimension of consonants is place of articulation, which in English are (going from front to back): bilabial (lip against lip) p, b, m, w; labiodental (top teeth into lower lip) f, v; interdental (tip of tongue between teeth) θ (as in 'thin' not 'there'), ð (as in 'there' not 'thin'); alveolar (tongue on alveolar ridge, i.e, the small ridge just behind top front teeth) t, d, s, z, n; alveopalatal (just past alveolar ridge, at beginning of palate or 'dome' of mouth) š ('sh' sound), ž ('s' in 'measure', 'z' in 'azure'), č ('ch' sound), ǰ ('g' in 'gem', 'j' in 'jump'); palatal j ('y' in 'yo'); velar (back of tongue against velum) k, g ('g' in 'get' not 'gem'), ŋ ('ng' sound); labiovelar kw ('qu' in 'quark'), which is velar just like k, g, ŋ, but with the lips rounded like they would be for w; and glottal h (as in 'hoe') and semi-phoneme ʔ (the 'breath-catch' in 'uh-oh!'). The glottis is deep, deep in the throat, the area of the 'voice box'.
NOTE: Because the plosive set t, d are alveolar, their fricative counterparts can be regarded as the alveolar set s, z. However, the alveolar ridge is just one of the two articulators here. All the obstruents have two articulators (except for kw, which has an extra co-articulation). The alveolar ridge and the tongue are the articulators for both sets t, d and s, z; however, the tongue performs differently for each set. With set t, d the tongue is flat and its apex is made use of (apical), whereas for set s, z the tongue is grooved. If you think here of the tongue being the primary articulator, then it is what the tongue is doing that is relevant — which makes the fricative counterparts to set t, d be θ, ð because for both these two sets the tongue is flat (not grooved) and its more or less tip made use of (apical). s, z are sibilant obstruents, wheras t, d, θ, ð are non-sibilant obstruents. You can be more specific with the terms that define the placing of the sets as: t, d apico-alveolar, s, z 'grooved' alveolar, and θ, ð as apico-dental — thus you see what is shared vs. what is not shared, the set t, d being the middle man.
The liquids l, r are difficult to define. I myself would describe liquids in the sense of the tongue being used not so much to obstruct the airstream, as to guide it along, with the vibrations of the airstream sort of 'charging' the tongue with resonance, which can even mean in some cases that the airstream itself takes control, such as in the glorious world-famous Scottish trilled rrrrr. In any case, it is easy to make a definition to distinguish the two liquids from one another: l is lateral because it guides the airstream around its sides, whereas this is not so for r (which varies exceedingly depending on dialect and context).

Vowels are three dimensional, too. There is the vertical axis high (tongue high, jaw not dropped low), low (tongue low, jaw dropped low), and mid (=neither high nor low). The high vowels of Standard English are: ī ('ee' in 'feet'), ɪ ('i' in 'fit'), ū ('oo' in 'loom' but not 'look'), and ʊ ('oo' in 'look' but not 'loom'). And the low vowels are: æ ('a' in 'rat'), ʌ ('u' in 'rut'), and α ('o' in 'rot'). And the mid vowels are: ē ('ai' in 'laid'), ɛ ('e' in 'led'), ō ('oa' in 'load'), and dialectal ɔ ('ough' in 'bought'), and the semi-phoneme schwa ə (like 'o' in 'carrot').

The schwa ə is a semi-phoneme because native English speakers are only semi-cognizant of it, and in any case there is not much to be cognizant of because it is an utterly neutral vowel, present when a vocalic element is necessary, as in the second syllable of 'carrot', just so the word doesn't collapse into one syllable *'carrt', but when a full-fledged vowel phoneme would make the word sound like two words, say, 'care ought' or 'car oat'.

There is the horizontal axis front (tongue at its highest in its front), back (tongue at its highest in its back), and central (=neither front nor back). The front vowels of Standard English are: ī ('ea' in 'lead'), ɪ ('i' in 'lid'), ē ('ai' in 'laid'), ɛ ('e' in 'led'), and æ ('a' in 'lad'). And the back vowels are: ū ('oo' in 'loom' but not 'look'), ʊ ('oo' in 'look' but not 'loom'), ō ('oa' in 'load'), dialectal ɔ ('ough' in 'bought'), and α ('o' in 'lot'). And the central vowels are: ʌ ('u' in 'luck') and semi-phoneme ə ('o' in 'carrot').
ʌ is not necessarily exactly low and central, but basically so, relative to the other vowels of English, and without having to make things messy by bringing up non-distinctive features.
Two features concerning specifically the lips, and that are not distinctive in English because they are automatic, are 1) lips being rounded by default for non-low back vowels ū, ʊ, ō, ɔ; and 2) lips being spread for low front vowel æ.

The third distinctive feature-dimension for vowels is whether the vowel is tense or lax. Tense means effort it put forth to make the opening narrow, and typically the tense vowels last a longer duration; and lax means the opening is not made to be narrow. Here are the tense vs. lax pairs: ī 'feet vs. ɪ 'fit', ē 'laid' vs. ɛ 'led', ū 'loom' vs. ʊ 'look', ō 'boat' vs. dialectal ɔ 'bought', and perhaps α 'rot' vs. æ 'rat', and quite contentiously ʌ 'rut' vs. scemi-phonemic ə 'carrot'.
I submit ʌ is tense (in English!) based on the fact that it is pronounced in a robust, stressed manner, as in the words 'luck, rug, slum'. Schwa sometimes sounds like an unstressed ʌ, hence it's the lax version, all things being relative, and with the assumption all vowels have a tense vs. lax pairing in English (does not Grammar strive for symmetry?). You could say α is the tense, ʌ the lax, but in English ʌ is perceived as a strong independent phoneme definitely not related to α. When an English speaker pronounces α with wider lips (which is the effortless form for English speakers), then out comes æ.
There is one other kind of phonemic vowel in English: the diphthong. A diphthong is composed of a regular vowel (=monophthong) and a glide coming together: aw ('ou' in 'ouch'), aj ('igh' in 'sigh'), and oj (as in the exclamation 'oy!') Even though phonemically (which is to say perceptually and grammatically), those vowels that are unequivocally tense — ī, ē, ū, ō — are regular vowels (monophthongal), the fact remains that, technically and phonetically, they are, respectively, disphthongs ij, ei, uw, ow.


The first principle behind this analysis is binary reductionism. By this term of mine is meant analyzing everything to the extreme — which means cutting everything apart into as many pieces possible — but not doing so at the expense of losing the inherent interconnectedness of the whole. Every split is done in a complimentary fashion, like the heads and tails of a coin which cannot exist independently, such that everything reduces to a “this as opposed to that”, e.g, vowels as opposed to consonants; and the definition of any given “this” is meaningful only in relation to its mutal contrast with a given “that”. Thus the second principle behind this analysis is one of extreme relativity where the identity of any given member of a set is precisely its very relation with the other members of the set — and in this way the underlying implication of the entire analysis is that there really are no individual members or componants; the only reality exists in the whole taken as a whole. However, because the mind can only grasp in snippets (such as in the form of words), there must needs be analysis (the shredding into bits) in order for there to be comprehension (the bringing together of the tattered truths into a meaningful whole). You cannot see the forest for the trees unless you first realize there are trees.
REFERENCES for current page:

Linguistics: An Introduction (2nd Edition) by William O'Grady, Michael Dobrovolsky, Mark Aronoff; St. Martin's Press, New York; 1993.

Phonology in Generative Grammar by Michael Kenstowicz; Blackwell, Cambridge MA & Oxford UK; 1994.

A Dictionary of Phonetics and Phonology by R.L. Trask; Routledge, London & New York; 1996.
« back top home about contact next »

Site & contents © Francis M. Tokarski, Jr. All rights reserved.