Language desk
< June 29	<< May \| June \| Jul >>	July 1 >

Welcome to the Wikipedia Language Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

June 30

Illegal, irregular

As is well known, the English prefix "in-" can assimilate with what follows it

intolerable, inapplicable: no assimilation
imperfect, imbalance: assimilates, /m/
ingratitude, inclement: assimilates, /ŋ/

This isn't complete. But let's turn to spelling for a moment:

illegal, illogical: "n turns into l"
irrelevant, irregular: "n turns into r"

I'm now going through Rochelle Lieber, Introducing Morphology, 2nd ed, and am puzzled by what I read on pp 177-178. Lieber says about pronunciations that in illegal etc. we have /ɪl/ and in "irregular" etc. we have [ɪɹ]. This is no mere typo: she also says a nasal consonant assimilates [...] to the point and manner of the consonant if it is a liquid.

I'd say instead that the "n" of "in-" simply disappears in front of a liquid. For me, the words are /ɪˈligəl/ and /ɪˈɹɛgjʊlə/, not /ɪlˈligəl/ and /ɪɹˈɹɛgjʊlə/. Do I perhaps misperceive my own pronunciation or speak an unusual lect of English? Or is Lieber perhaps wrong? -- Lips McGee (talk) 07:05, 30 June 2018 (UTC)[reply]

I agree with you, and I think most British dictionaries represent the words that way, but some American dictionaries use the Lieber pronunciation. Perhaps there are regional differences? Dbfirs 07:22, 30 June 2018 (UTC)[reply]

Thank you for the sanity check, Dbfirs. Lieber is American and she primarily has US English in mind, so fair enough -- if it is indeed true for US English. I wonder what Americans will say about this. In my own (non-US) English, /ll/ can occur (e.g. in "well-loved"), but I don't think that /ɹɹ/ is even possible, anywhere. Lips McGee (talk) 07:36, 30 June 2018 (UTC)[reply]

I would say that in General American and Standard Canadian English, there are essentially two L's (or R's) being pronounced distinctly. In both irregular and illegal, the first syllable is stressed, which I think (?) is not the case in most British accents. However, I still think your interpretation is correct and that the doubling of the consonants is perhaps something that got added in afterwards. We like stretching initial "I" words here in NA (compare Eye-rack and Eye-ran). Matt Deres (talk) 13:12, 30 June 2018 (UTC)[reply]

I'm surprised by this. I would expect two distinct L sounds for the double L of "illegal" only as a strong form of emphasis ("I said ILlegal"), and similarly for "irregular". I live in Canada but my English is a mixture of different sources. --76.69.47.228 (talk) 05:45, 1 July 2018 (UTC)[reply]

This has nothing to do with English. It's a Latin prefix and like some other prefixes it assimilates with the first consonant of the root. In Latin those were literally long consonants [lː] and [rː], and this still exists in Italian (irregolare, illegale). English has no long consonants as such, so these ll and rr are purely orthographic.--Lüboslóv Yęzýkin (talk) 15:50, 30 June 2018 (UTC)[reply]

But Lüboslóv Yęzýkin, I believe that in my own lect of English "well-loved", for example, has an /l/ coda and an /l/ onset. I can feel my tongue moving between the first /l/ and the second (going from a "dark 'l'" to a regular one, I suppose). Therefore I can easily believe that "illegal" could be pronounced similarly. Indeed, if by magic I woke up half a century ago and found myself dictating material to a secretary and thus wanting to be very clear, I might well pronounce it this way. But I'm confident that in normal conversation I don't. As for "irregular", maybe I'm crippled by my non-rhotic English, but I can't put an /ɹ/ coda in front of an /ɹ/ onset even if I want to. I'd guess that in L1 General American (and other rhotic lects) it is possible. I'm wondering whether Americans and others do actually do this for "illogical", "irrelevant" and the rest. I'm also willing to believe that the doubling of "l" and "r" is (at least in today's English) merely an orthographic convention. But if so, here's what's strange: Lieber isn't a phonologist, and her areas of particular interest within morphology, as far as I understand them, are not particularly related to phonology. But she is a very eminent morphologist. This is an introductory morphology text in its second edition. It's hard to believe that a mistake such as this would get through. Additionally, the pronunciation guides in respected American English dictionaries may be flawed, but I think they're worth taking seriously. Lips McGee (talk) 10:04, 1 July 2018 (UTC)[reply]

@Lips McGee: She is simply wrong (I have just read those pages). The suffix in- does not become [ɪl] and [ɪr] in English. This was in Latin. And these exact Latin forms were borrowed directly (or via French, though it does not really matter) into English "as is" with the doubled consonants. But in English these consonants are not doubled in pronunciation, only in spelling. You must not confuse this case with the natural doubling at the word boundaries (well-loved), the latter is an entirely different synchronic process, whereas the double letters in the borrowing adopted long ago is diachronic. If anything, [l] and [r] of [ɪl] and [ɪr] belong to the root, so practically it is fair to say that before liquids the suffix in- becomes [ɪ]. So there are four allomorphs of the suffix in English: [ɪn], [ɪm], [ɪŋ], and [ɪ]. In Latin, as I said, the latter was different. It would be more correct - if she presented English examples all borrowed from Latin - she were speaking about the assimilation processes in Latin directly. Particularly, when she is not afraid to "shock" or "burden" the reader with exotic languages and gives such examples quite often (the next page she presents Zoque - even I never heard of such a language; at least the next pages she gives examples from Turkish - definitely more known and less exotic; so it would not be much difficult to the reader - if speaking about the allomorphy in Latin borrowings - she turned to Latin itself, a very well-known language).--Lüboslóv Yęzýkin (talk) 08:53, 2 July 2018 (UTC)[reply]

Thank you, Lüboslóv Yęzýkin. If it's a mistake -- and yes, I think it is -- then it's pretty surprising, considering the publisher, series, author, and the fact that it's a second edition. Now I'm particularly interested in Matt Deres's comment above: I would say that in General American and Standard Canadian English, there are essentially two L's (or R's) being pronounced distinctly. You're a Canadian; even in fairly careful speech, do you actually pronounce them in this way? Lips McGee (talk) 12:53, 2 July 2018 (UTC)[reply]

@Lips McGee: I've just re-read the whole topic. And yes, Merriam-Webster gives an alternative pronunciation with the doubled consonants. I even did not believe them, not the least because their transcription is always somewhat sloppy and confusing, so I deliberately rechecked in Kenyon & Knot's A Pronouncing Dictionary of American English. And yes, they clearly listed the variants with the double consonants. And the Cambridge English Pronouncing Dictionary gives the same, though less clear, but in some words such as illegible it is definitely clear. Though they do not believe that in irr- double [rr] is possible, while Kenyon & Knot thought it is. The latter also thought that single [r] is an Eastern/Southern feature. So I must have been wrong at least in some way. Or rather RP/British-centric. Lieber is American, so probably her pronunciation is this and she must know better for herself. But I always considered such a pronunciation affected or a spelling pronunciation which is more typical for non-native speakers. But Americans are well-known for spelling pronunciations, so I am not surprised very much. Probably, I must thank you that you have made me learn something new I was not aware of. English never ceases to present something new.--Lüboslóv Yęzýkin (talk) 13:37, 2 July 2018 (UTC)[reply]

See this page as an illustration to my point that this is often a foreign accent. See also this. And this.--Lüboslóv Yęzýkin (talk) 14:02, 2 July 2018 (UTC)[reply]

Thank you, Lüboslóv Yęzýkin. It's interesting that the Cambridge English Pronouncing Dictionary says that /ll/ is possible for "illegitimate", etc., but still doesn't recognize /rr/ in "irrational", etc. (I suppose the only way to check the latter is via oscilloscope.) English is indeed fascinating. (I only recently realized that adverbs can modify nouns: "the shortage internationally of XYZ" = "the international shortage of XYZ"; I mean, it was there in my L1 English competence, but missing from my conscious knowledge. But there must be thousands of languages that are just as fascinating as English but that lack the vast literature that has built up around the syntax, morphology, phonology, etc. of English Lips McGee (talk) 13:45, 3 July 2018 (UTC)[reply]

Lips McGee I think so (though it's hard to not over enunciate when considering stuff like this). If they're not two distinct sounds (one phoneme to each syllable), they're definitely elongated. I'm struggling to find a minimal pair, but compare eligible/elongated/erase to illegal/irregular, for example. I hear those initial consonants as different for the two groups - in the first group it's clearly only one sound, but in the other it sounds stretched and/or split. Your mileage may vary, of course. Matt Deres (talk) 15:52, 4 July 2018 (UTC)[reply]