Why do so many languages have both an alveolar "light L" [l] and a velarized "dark L" [ɫ] allophone?

Question

Various dialects in all three of English, Portuguese, and Dutch have accents that contain velarized L allophones, which are sometimes known as “dark L’s”, at the syllable coda.

Why is this?

There are two different L allophones: a “light L” [l] is pronounced more at the front of the mouth while a “dark L” [ɫ] is pronounced more at the back of the mouth. For example:

• In some English dialects, the words light [lʌɪt] and lit [lɪt] have “light L’s” at the start, while the words occult [əˈkʰʌɫt] and full [fʊɫ] have “dark L’s” at the end of the syllable.

• In some Portuguese dialects, the words ler [leɾ] ‘read’ and libra [ˈliβɾɐ] ‘pound’ both start with light L’s, but the words fácil [ˈfasiɫ] ‘easy’ and mal [maɫ] ‘bad’ both end with a dark L in those Portuguese dialects that do this, although those same words in Spanish always have only a light L.

Wikipedia mentions:

In a number of languages, including most varieties of English, the phoneme /l/ becomes velarized in certain contexts, a sound often called "dark l". Some languages, like many North American dialects of English, may not have a "clear" /l/ at all, or use it only before front vowels (especially [i]).

The velarized alveolar lateral approximant (dark l) is a type of consonantal sound used in some languages. It is an alveolar, denti-alveolar, or dental lateral approximant, with a secondary articulation of velarization or pharyngealization. The regular symbols in the International Phonetic Alphabet that represent this sound are ⟨lˠ⟩ (for a velarized lateral) and ⟨lˤ⟩ (for a pharyngealized lateral), though the dedicated letter ⟨ɫ⟩, which covers both velarization and pharyngealization, is perhaps more common.

How does this phenomenon arise phonologically, and why do native speakers of languages in which this occurs not consider those distinct phonemes? Is this a property that the physical and acoustic characteristics of the buccal cavity lend themselves to naturally, or is it instead some sort of mental adaptation that occurs only in certain cultures?

Are there any languages that have both kinds of L — meaning alveolar or “light” or “clear” on the one hand versus velarized or “dark” on the other hand — as distinct phonemes with minimal pairs rather than as mere allophones?

Answer 1

I've been looking for a functional explanation in the literature, and this apparently isn't a question that has been explored: why is the change from clear to dark l so common? Dark l has a very low F2, and clear l has a higher F2, so it is unlikely that this is acoustically driven. I suspect that it is related to things necessary to articulate a lateral.

[l] would differ from [t] in that [l] has only a little bit of contact in the front, whereas [l] has contact going back to the molars. It has been observed that in the articulation of [l] (this is from Ladefoged & Maddieson's The sounds of the world's languages), the jaw is lower and the tongue is lower, making the lateral opening more possible. (There are other ways, and thus we have palatal [ʎ] as well, where the side cavity is in the front). Lowering the tongue encourages backing (there an empty space in the back where the tongue can easily go, and not much room in the front). So some degree of pharyngealization is probably a natural but not irresistible physiological consequence of what you have to do to achieve a side cavity while maintaining a short alveolar closure.

Unfortunately, it is very difficult to model the degree of physiological necessity of this tongue lowering and backing: there are certainly languages with clear l. What is empirically unclear, at the moment, is how laterals and plain alveolars compare, articulatorily, in languages which absolutely do not have l-velarization. A physiological comparison tongue retraction and backing in Spanish l vs. English clear l in maximally non-darkening environments would be informative.

One further factor that is surely relevant is that the velarization effect is most prominent in the syllable coda. It may be that in the onset, duration of l is shorter and there is a rapid transition to a more prominent vowel formant, whereas in the coda, what comes after l is probably much less prominent acoustically, and the duration of l is longer therefore potentially more noticeable.