English phonology

Two types of l

Many varieties of English, including SBE, have two types of l: the first sound of lick lik sounds differently from the last sound of kill kil. In fact, the l in kill and killing kiliŋ sound different. Phonetically speaking, the l in kill is velarized, which means that the back of the tongue is raised against the velum, giving the l some u colouring. The l’s of lick and killing lack this u colouring. The u-coloured consonant is called dark L, the other one clear L, or sometimes light L. Dark — velarized — L is represented as ɫ, clear L is simply l.

Accents of English show great variation with respect to the distribution of clear and dark L. General American and Scottish English lacks this distinction and has velarized, ie dark, L in all positions, while Irish English has clear L only. Both variants occur in Standard British English. The two types of l are in complementary distribution, hence they are allophones of a single phoneme l. We are going to examine their distribution presently.

As already noted, the l in lick is clear (lik), while that in kill is dark (kiɫ). It would be very bold to draw conclusions of so little data, so let’s see some more examples. They have been grouped according to the type of l that occurs in them.

word finallyWe can see that clear L is followed by a vowel in each of the above examples (eg lick), while dark L is either followed by a consonant (eg kilt) or is word final (eg kill). Claiming that word-final l is always dark is wrong. Consider the word killing kiliŋ. Since the stem of this word is kill, which itself is a word, the l in this word is word final. Being word final does not necessarily mean being followed by a space in writing, it simply means being at the end of a word, and the part kill in killing is a word. This is marked as kill#ing by phonologists.^»The symbol # is called the hash mark and in phonology it represents a word boundary. Note that the part shill- in shilling is not a word, hence the l in shilling is not word final. But this makes no difference, the l is clear in both cases.

The l is also word final in killed, but in this case it is followed by a consonant, the d of the past tense suffix. In fact, the word-final allophone of l is selected by the segment that follows the word boundary, ie the first segment of the suffix or of the next word. Thus killing kiliŋ has clear L just like kill Inge kil iŋgə or shilling ʃiliŋ, and killed kiɫd has dark L just like kill Dan kiɫ dan or build biɫd. That is, the word boundary does not influence whether a l is clear or dark in any way.

To summarize, clear L occurs before a vowel and dark L occurs elsewhere, ie before a consonant or at the end of an utterance (before a pause) in Standard British English. This can be given by the following formulas:

l: __(#)V
ɫ: __(#)C and __##^»The two hash marks mean a pause.

The parenthesized word boundary symbols indicate that it is indifferent from the point of view of this regularity if the vowel or consonant that follows the l is in the same word or at the beginning of a suffix or the next word.

yodThere are two modifications that have to be added to the above formulation of the distribution of clear and dark L. On the one hand, the l is clear in the following examples.

value valjuw, million miljən, schoolyard sguwljaːd

In each of these cases we find j (called yod) after the l, which is clear. This is unexpected, since j is a consonant,^»We know that j is a consonant, because we find the indefinite article a (not an) and the definite article ðə (not ðij) before words that begin with it. and before consonants l is supposed to be dark. Here we are not going to investigate possible reasons for this peculiarity, we only record it. The formulas accordingly modified as given below:

l: __(#){V j}
ɫ: __(#){p t k b d g ʧ ʤ f v θ ð s z ʃ ʒ m n l r w h} and __##

The trouble with these formulas is that neither “V or j”, nor “any consonant but not j” is a natural class, ie there is no standard set of sounds that includes exactly these segments. Hence we either list all the relevant segments (as above) or create arbitrary sets that have to be defined.

syllabic lFurthermore the distribution of clear and dark L given above has to be restricted to nonsyllabic occurrences of l. Consonants, among them l, typically occur before or after a vowel. In some languages, among them English, some consonants may occupy the middle of a syllable too, which is more common for vowels to occupy. The word little litl contains two syllables (li and tl), but only one vowel, in the first syllable. The second syllable contains only two consonants, of which l occupies the “vowel” position: it is syllabic.^»Hungarian generally has no syllabic consonants, but many speakers do pronounce one in the second syllable of Lidl. Other speakers pronounce this name as lidli, with a vowel in the second syllable, too.

A syllabic l, marked as l̩, is dark irrespective of what follows it, vowel or consonant. So the second, syllabic l̩ is dark in each of the following examples:

little Ann litl̩ an, little Eunice litl̩ juwnis, little Thelma litl̩ θelmə