Global synopsis: phonetic and phonological variation in English world-wide

Conclusion

The variability of the pronunciation(s) of English(es) around the globe tends to be highly complex, multi-faceted and multidimensional, and strongly conditioned by regionality and sociolinguistic factors. Hence, generalizations are problematic, and unavoidably face the difficulty of simplifying an apparently chaotic reality too radically – in this context the emphasis needs to be on details, individual distributions, and local or regional patterns. Nevertheless, I conclude by suggesting a few general patterns and observations that the data summarized above seem to imply and that should deserve more thorough investigation and possibly substantiation.

– The amount of variability found in a given area seems to correlate with the “historical depth” of the independent evolution of the respective variety of English. A very large amount of minute detail characterizes the dialectal landscape of England, and the British Isles in general. In North America, there is still a fairly wide range of pronunciation details to be observed. In contrast, the pronunciations of AusE and NZE are relatively homogeneous (though marked by substantial social class distinctions), and differences within SAfE are primarily socially and ethnically conditioned. African and Asian varieties seem to be relatively more homogeneous – although the amount of variability to be found should not be underestimated.

– While British English varieties are characterized by an elaborate system of diphthongs, and long vowels have commonly tended to develop glides, this does not hold for many of the "younger" varieties. African and Asian varieties functioning as or derived from L2's tend to be marked by more uses of monophthongs; they have considerably fewer vowels with offglides than, say, British, American, or "Broad" Australian dialects.

– Some varieties display a tendency to reduce the fairly rich system of RP lax ("short") vowels. One cause of this may be to avoidance of schwa (possibly due to a tendency toward syllable-timing rather than stress-timing) and hence the replacement of schwa by some other, fully stressed vowel. Another reason may be the fact that the system of short front vowels of RP and StAmE is relatively crowded, with the vocalic space being divided by as many as four vowels ([ɪ, ε, æ​, a]) . In West African varieties, for instance, the tendency is to have a five-vowel system. The exact arrangement patterns are regional: For example, in CamE and NigP STRUT merges with LOT; elsewhere STRUT tends to join the TRAP class.

– The importance of length in distinguishing phonemes is definitely waning. Even in RP corresponding "long" and "short" vowels (like KIT and FLEECE) are actually distinguished not primarily by their respective duration but rather by slight qualitative differences. In American English, and even more so in African and many Asian varieties, the importance of length is considerably reduced; in some varieties it is simply not contrastive at all. KIT and FLEECE, FOOT and GOOSE, LOT and THOUGHT tend to merge in quite a number of varieties – unconditionally in some, in specific environments (e.g. quite commonly before /l/) in others.

– A few ongoing sound changes have been described, but there is no globally uniform process to be observed. The closest thing to a supra-regional sound change seems the tendency to front back vowels as part of the so-called "Southern Shift", to be observed both in southern hemisphere countries and in the Southern US. Conversely, however, in the northern US "short" (or "checked") vowels show a tendency toward an up- and backwards rotation.

Basically, it should be possible to categorize sound distributions into some with a practically global outreach and others with rather strictly local extensions. A preliminary classification along these lines yields the following listings (by necessity, the listings are tentative and cannot claim to be exhaustive):

(1) Globally predominant sound realizations:

– KIT as [ɪ];

– DRESS as [ε];

– FOOT as [ʊ];

– FLEECE as [i:];

– GOOSE as [u:];

– THOUGHT, NORTH as [ɔ:];

– CHOICE as [ɔɪ];

– MOUTH as [aʊ];

– happY as [i];

– [d] for word-initial /ð/; also [t] for initial /θ/;

– alveolar for velar nasals in–ing endings.

(2) Distinctive sound realizations that may serve to characterize specific regions::

– [ʊ] in STRUT;

– [a] in LOT, CLOTH;

– [ɔ], [œ], [a] or [ε] in NURSE;

– [ɪ, i] in FLEECE;

– [ɪə] in FACE and [ʊə] in GOAT;

– [ʉ] or [ʊu] in GOOSE;

– [əɪ], [ɔɪ] , [ae] or monophthongal [a:] in PRICE;

– [əʊ] or [æʊ] in MOUTH;

– [ia] in NEAR;

– [o: , ɔ:] in CURE;

– homophony of KIT and FLEECE;

– homophony of FOOT and GOOSE;

– homophony of LOT and THOUGHT;

– homophony of LOT and STRUT;

– homophony of NEAR and SQUARE;

– the "Scottish Vowel Length Rule";

– unaspirated word-initial /p/, /t/, /k/;

– glottal stop for word-fi nal /t/;

– word-initial [kj-], [gj-], [bw-] for k-, g-, b-;

– velar fricative onset in wh- words;

– [v] for /w/;

– existence of velar fricative;

– word-initial /h/-deletion and /h/-insertion;

– /r/ realized as apical or uvular trill;

– velarization of word-final nasals (e.g. [-ŋ] in down);

– existence of tonal distinctions.

While global patterns do not serve the purpose of accent discrimination well, local pronunciations are more useful for determining a speaker's place of origin. For such a purpose, the features of the third list seem most recommendable:

(3) Features which seem particularly useful for the identification of regional accents in a global perspective:

– [æ] vs. [a] in TRAP;

– [ɒ] vs. [ɑ] in LOT;

– [ɑ:] vs. [a] vs. [æ] in BATH;

– [eɪ]  vs. [e:] in FACE, [əʊ] vs. [oʊ] vs. [o:] in GOAT;

– [ə] vs. [a, Λ] in lettER and commA;

– nasalized vowels before nasals;

– lenisation / voicing / flapping of intervocalic /t/ (writer = rider);

– jod-dropping;

– (frequency and conditions of) word-final consonant cluster deletion;

– rhoticity;

– existence of intrusive /r/;

– stress shift;

– tendency towards syllable-timing;

– high-rising terminal contour.

In general, however, no single pronunciation detail will suffice to serve such a purpose. In other words, distinctive accents are never distinctive because of any specific feature found there, but always because of the unique mix of pronunciation choices in a given region.