Consonant clusters

Posting about Sprite recently, specifically mentioning consonant clusters, reminded me that I have not fully explored these in this blog. Some time ago I posted a list of common words with the possible clusters of English, but didn’t analyse the clusters themselves.

All of the two-phoneme initial consonant clusters in English start with /s/ and/or finish with /w/, /l/, /r/ or /j/ (the ‘y’ sound in you, not the ‘j’ sound in jewel). Starting with: 
/sw/ switch, suede
/sl/ sleep, slow

The other ones starting with /s/ are:
/sm/ small, smile
/sn/ snow, snap
/sp/ speak, spend
/st/ state, still
/sk/ school, scale, skin
the rare, Greek /sf/ (sphere, sphinx)
and the very rare, Greek /sθ/ (sthenic (I don’t know what this means, but it appears on lists of consonant clusters)

The ones ending in /w/ are:
/tw/ twenty, twelve
/dw/ dwell, dwarf
/kw/ question, quite
/gw/ guava, Guam
/θw/ thwart

The ones ending in  /l/ are:
/pl/ place, play
/bl/ black, blood
/kl/ close, clear and scientific words starting with chl
/gl/ glass, glance
/fl/ floor, flat

The ones ending in /r/ are:
/pr/ problem, provide
/br/ bring, break/brake
/tr/ try, train
/dr/ draw, drive
/kr/ create, cross and many Christian and scientific words (chrome, chrono-)
/gr/ great, group
/fr/ from, phrase
/θr/ three, through
/ʃr/ shrug, shrink

Finally, the ones ending in /j/ are:
/mj/ music, museum
/nj/ new/knew, nuisance
/pj/ pupil, pure
/bj/ bureau
/tj/ tune, tube
/dj/ during, duty
/kj/ queue/cue, curious
/fj/ future, fuel
/vj/ view
/hj/ huge, humour

In table form:

switch, suede sleep, slow

m small, smile

music, museum
nsnow, snap

new/knew, nuinsance
pspeak, spend
place, playproblem, providepupil, pure

black, bloodbring, break/brakebureau
tstate, stilltwenty, twelve
try, traintune, tube
dwell, dwarf
draw, driveduring, duty
kschool, scale, skinquestion, quiteclose, clearcreate, crossqueue/cue, curious
guava, Guamglass, glancegreat, group
fsphere, sphinx
floor, flatfrom, phrasefuture, fuel

three, through

shrug, shrink

huge, humour

(I tried to make a table unambiguously showing which combinations are possible and which aren’t, but couldn’t make one that is both complete and clear.)

There are several things to notice from this list and table. The first is that all the clusters with /j/ arise from the interplay of the first consonant with the following /u:/. Some varieties of English, especially in the USA, do not do this, instead pronouncing new as /nu:/, tune as /tu:n/ and dew as /du:/. Others, especially in the British Isles, turn huge and humour into yooj and yumour.

The second is that in order to fit into a cluster together, the two sounds need to be similar enough that our mouths can move seamlessly from the first to the second, but not either too similar or too different. Examples of ‘too similar’ are /tl/ and /dl/. /mw/ is used in mwah or muah, verbalising a kiss, while /pw/ and /bw/ are regarded as childish or a speech impediment for /pr/and /br/ respectively. Examples of ‘too different’ are /sb/ (compare /sm/ and /sp/), /sd/ (compare /sn/ and /st/) and /sg/ (compare /sk/). 

The third is the difference between pronunciation and spelling. The pronunciation /kw/ is usually spelled <qu> (and vice versa), but note the informal ‘Kwik-e-mart’-type spelling. On the other hand, the spelling digraphs <ch>, <ph>, <th> and <sh> represent the single phonemes /tʃ/, /f/, /θ/ and /ð/, and /ʃ/, so by themselves do not appear on these lists. Thus <chr>, <phr>, <thr> and <shr> are two-phoneme clusters, not three. 

The fourth is that although all the examples I have given are of these clusters at the start of words, these clusters can occur at the start of any syllable within a word. It’s just easier to search for them at the start of words. These, other and longer clusters can also occur between syllables (that is, at the the end of one syllable and the start of the next), but in those cases can always be broken down into single phonemes or the clusters listed here.

It should come as no surprise that the three-phoneme initial consonant clusters of English follow the same patterns. Indeed, they are, with one rare exception, made from (s) + (p, t, k) + (w, l, r, /j/). They are:

/spl/ split, splash
/spr/ spring, spread
/spj/ spurious, spew
/str/ street, strong     
/stj/ student, studio
/skw/  square, squad
/skl/ sclerosis
/skr/ screen, scream
/skj/ skew 

The rare exception is smew, a small water bird from northern Europe. This still fits the general pattern, though, and like the other examples with /j/, it happens because of the interplay between the preceding consonant and the following /u:/.

All of this presents medium-sized problems for English-speaking children and large problems for English language learners. Most languages have no, far fewer or far simpler initial consonant clusters. (Some even have no clusters between syllables.) English-speaking children often drop one of the sounds completely, pronouncing swim as wim, or change one or more sounds, pronouncing three as free or fwee. English language learners often insert a neutral vowel (usually a schwa or similar) somewhere, often influenced by their own language. I’ve heard Spanish-speaking students pronounce Spain and states as Espain (cf España) and estates (cf estados) (note that estates is another, related word). And, which is what got me thinking about this in the first place, Koreans say 스프라이트 (seu-peu-ra-i-teu). English language learners also drop one or more sounds, especially in the middle or at the ends of of words: one student momentarily confused me by saying he’d eaten Madonna for breakfast. Which brings me to another, probably even bigger topic (to be continued, sometime), that of final consonant clusters, of which there are more, and more variations.

Finally, not arising from any of the above, and almost worth a separate post, English speakers with some knowledge of other languages and cultures have been exposed to different initial clusters, whether they can pronounce them accurately or not. Off the top of my head I can think of bwana, Dmitri (originally a follower of Demeter), fjord, Gstaad, Gyeongbokgung, Kjeller, Mstislav (msti – “vengeance” and slav – “glory, fame”), Mraz, Rwanda, schmuck, Sjöberg, Sri Lanka, Vltava, zloty. But most of these follow the same patterns of (s or z) and/or (m, n, p, b, t, d, /k/, g, /f/, v etc) and/or (w, l, r, /j/). There are various exceptions to that generalisation, which I won’t analyse, but probably have explanations within their own languages. Note that some of these represent different sounds or combinations than in English; for example, in Korean, ‘y’ is officially classed as a vowel. 

(More information about consonant clusters in general, and in English.)


One thought on “Consonant clusters

  1. Pingback: motovlog | Never Pure and Rarely Simple

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s