I was wondering. I was reading in one forum about the dark L someone said about how you can also have a dark L in an unstressed syllable so a word like only (ˈoʊnli) will actually have dark L. Can someone enlighten me a little.bit.more about this please? However, a word like maleficent wouldn't have a Dark L, it'll have a Light L, since the Letter L Is the stressed in the word. I need a little help about unstressed L's being a Dark L.
To start with, I would say not to worry about this too much, as I think it would be hard to perceive the difference in this position. For me, "dark l" (i.e. /l/ with a back articulation, phonetically transcribed as [ɫ]) is most noticeable in its effect on the pronunciation of a preceding vowel, so when I try to pronounce the /l/ in a word like "only" as dark vs. light, I can't hear much difference.
It seems somewhat plausible to me that /l/ in a word like "only" might be dark, although it certainly is light in many accents. If you're getting into the phonetic details, you might want to know that according to the phonetician Jack Windsor Lewis, "only" is often pronounced without any [l] sound at all (see my answer to the question Is it common to pronounce 'only' as 'own-knee'?).
Different accents are known to follow different rules for the distribution of light and dark /l/. As Janus Bahs Jacquet said, there are supposed to be accents that take one of the two extreme options of "all /l/s are light" or "all /l/s are dark". And there are a number of different types of in-between rules:
For "RP" English, the rule is typically described as using light /l/ before a vowel (or before /j/), and dark /l/ elsewhere. Word boundaries are supposed to be irrelevant, so we would hear light /l/ in a context like "full of sand".
For "SBE" (which is supposed to be a modern British accent, unlike "RP") the rule may be somewhat different. One exception to the "preceding a vowel" rule that I've seen given is syllabic l: according to this web page, that would be dark in "SBE" even when a vowel follows.
Bermúdez-Otero (2013) cites Sproat and Fujimura (1993) as describing a different rule for the distribution of dark /l/ in a certain accent of American English. In this accent, word-final /l/ is always dark, regardless of whether a vowel or a consonant follows, but word-internal /l/ is light when a vowel follows. The examples given to illustrate this are "Beelik" (light l) vs. "Beel equates" (dark l).
There are accents of American English and of British English that show l-darkening in more contexts (yellow and village are the examples given). Bermúdez-Otero analyzes these accents as having a "foot-based" rule, like the rule for t-flapping. For this rule, stress is relevant: /l/ cannot be dark at the start of a stressed syllable, but it can be dark at the start of an unstressed syllable (assuming that we syllabify words like "holy" as "ho.ly").
I didn't see a detailed analysis of how this rule was supposed to apply. For t-flapping, Bermúdez-Otero says there are actually two rules involved: at the word level, /t/ in non-foot-initial position is supposed to be lenited, and then at the phrase level, lenited /t/ is supposed to be flapped when it is followed by a vowel and preceded by /r/ or a vowel. This allows us to explain why we don't see flapped /t/ in words like "guilty". If a similar two-part process is supposed to account for l-darkening, then we would only expect to see "dark" l when the preceding sound is a vowel or /r/. But it seems plausible to me that some accents might have a simple rule of darkening non-foot-initial /l/ regardless of the identity of the preceding sound.
My accent has what Bermúdez-Otero would call a "foot-based" type rule, and I use dark /l/ in angrily and daily. Based on the pronunciation of the vowel, I believe that I use light l in lowly, slowly and grayly, and sometimes in gaily, presumably because the morpheme boundary is more salient than the one in daily. I always use dark l in "-Vly" words when there is no morpheme boundary before the /l/, e.g. holy, mealy, steely, wooly.
I use dark l in words like wholly, totally, finally.
I think that I use light l in words like invisibly, unbearably.
I think that I use light l in only, although as mentioned in the first two paragraphs, it doesn't seem very noticeable.
I don't think the "l" in "only" would be dark, because the change to dark "l" depends on the syllable position of "l"; "l" is dark in the offset of a syllable, by which I mean the part of the syllable after the vowel. "Only" is divided into two syllables "on-ly", and "l" is at the beginning of the second syllable, even though the vowel of the second syllable is not stressed, so the "l" is probably not dark.
By the same reasoning, the "l" of your other example, "curly", should also not be dark, but this is less clear than the first example, because the syllable division could be "curl-y", in which case the "l" is at the end of a syllable and should be dark. In my own speech, the syllable division is "cur-ly", and the "l" is not dark, just as for your first example, but this might not be true for other people. There are two complicating factors here that might give a different result:
(1) If you take away the last vowel in the first example, you are left with "onl", which is not a possible word of English. But in the second example, you are left with "curl", which is a word, and in fact is the word from which "curly" is derived. Since syllable boundaries sometimes tend to follow the divisions between meaningful parts of words, this might cause the syllable division "curl-y".
(2) Sometimes a single consonant in English between a stressed vowel and an unstressed vowel is shifted into the preceding syllable with the stressed vowel. Since "ur" is a single stressed vowel (even though it's spelled with two letters), and "-y" is an unstressed vowel, the "l" might shift into the first syllable for this reason. That should make the "l" dark. (But that doesn't happen in my own pronunciation.)