157

Why are certain words pronounced with diphthongs on their own but with monophthongs in compounds? For example:

Words pronounced with diphthongs on their own: Michael, Christ, wise, drive

Their pronunciations with monophthongs in compounds: Michaelmas, Christmas, wizard, driven

All of these changes occur in stressed syllables. Compare that of cycle and bicycle, tricycle which occurs in unstressed syllables.

P/s: Just to be clear, I added the "historical-change" tag because I suspected this phenomenon had something to do with the Great Vowel Shift.

Vun-Hugh Vaw
  • 5,400
  • 21
    For what it's worth, I generally hear the Y in cycle and bicycle pronounced differently, in basically the same way is Christ/Christmas. – aslum Jun 01 '16 at 16:55
  • 1
    'Michaelmas'? Sounds like a Michael Scott-ism... – TylerH Jun 02 '16 at 18:33
  • 7
    @aslum: The OP is not saying that the is pronounced the same way in bicycle as in cycle; rather, (s)he's saying that that pronunciation difference is to be expected, since the is unstressed in bicycle (whereas in the pairs (s)he's asking about, the corresponding syllable is stressed in both words). – ruakh Jun 03 '16 at 06:05
  • 1
    @ruakh Indeed. I checked and a question regarding "cycle/bicycle" had been asked. I just wanted to make clear that I didn't ask that same question again. – Vun-Hugh Vaw Jun 04 '16 at 13:34
  • As long as your at it, what about cyclical being pronounced like it is "SICK-lick-all" ? Is that a dialect issue or something else? – O.M.Y. Jun 04 '16 at 22:22
  • 1
    @O.M.Y. it's the same thing, other than the fact that the /aɪ/ that becomes a /ɪ/ is spelled with a "y" rather than an "i". – hobbs Jun 04 '16 at 22:31
  • @hobbs I hear what you are saying but I have also heard cyclical pronounced "SIKE-lick-all". Which is correct? – O.M.Y. Jun 04 '16 at 22:43
  • @O.M.Y., for what it's worth, 'cyclic', as in 'cyclic group', is (I think) universally pronounced 'SICK-lick' by mathematicians, and I think 'cyclically', as in 'cyclically permuted', nearly as often pronounced 'SICK-lick-lee'; so I'd imagine that 'cyclical' would almost always be 'SICK-lick-all' amongst us. (Of course that sheds no light on why 'cycle', which we pronounce as all other humans, shifts pronunciation when it becomes 'cyclic', only lends some weight to the claim that it does so shift.) – LSpice Jun 05 '16 at 17:49
  • 1
    @LSpice: FWIW In my maths department it was always "SIGH-click" and "SIGH-click-lee"... – psmears Jun 05 '16 at 22:18
  • @psmears, so much for my universality hypothesis! Is it fair to guess that your maths department is English? Maybe "universally … by American mathematicians" would have been less blatantly false. – LSpice Jun 05 '16 at 23:19
  • See also http://english.stackexchange.com/q/336090. – tchrist Sep 06 '16 at 00:55

2 Answers2

291

Short answer

The PRICE vowel that we hear in the word wise, /waɪz/, has a systematic relationship with the KIT vowel which we hear in the word wizard, /'wɪzəd/. As we add syllables to the base of a word in English, we tend to reduce the length of the vowel in the base. This is so that we can accommodate the new syllables and still preserve the perceived stress timed rhythm of English. When we add syllables to a base word containing the PRICE vowel, the vowel is very likely to change to a KIT. Other pairs of long and short vowels have this same type of relationship in English.

Full answer

What the Original Poster is observing here is a side-effect of the way English organises stressed syllables in connected speech.

English is a 'stress-timed' language. What this means is that the syllables in English utterances do not come at regular intervals in the way that they do in Japanese for example, or in Spanish. Instead English utterances give the impression that the stressed syllables come at regular intervals. In actual fact this is not strictly what's happening, in the sense that although they give this impression, the stressed syllables do not occur at strictly regular intervals at all.

The upshot of this is that the following utterances, for example, will take roughly the same amount of time to pronounce in English:

  • fish, peas, beans, rice.
  • fish and then some peas and a banana and a casserole.

The first utterance there has only four syllables, but the second has fifteen. However, in normal relaxed speech, these sentences will have approximately the same duration. This is because they each have four stressed syllables:

  • fish, peas, beans, rice.
  • fish and then some peas and a banana and a casserole.

Now, if you're a native speaker, if you tap out the stressed syllables as you say the two utterances, you'll realise that the stressed syllables seem to come at regular intervals, regardless of what other material comes in between them.

English has a system for enabling this to happen. One part of this system is that many grammatical words have two forms. They have a strong form which we use if these words are stressed (or stranded). They have a weak form which we use if they are not stressed. So for example, the strong form of can is /kæn/ and the weak form is /kn/ or /kən/. You will be able to hear the difference in the following examples:

  • I can dance.
  • I can dance.

What you will notice is that these weak forms have a reduced vowel, most usually a schwa, /ə/. This vowel is easy to pronounce very quickly because it does not require any large movements of the articulators.

Another effect that we find is that a given stressed syllable will become shorter when more unstressed syllables are added. So the if we compare the /mæn/ ('man') in the word man and the /mæn/ in the word manager, we will find that /mæn/ is much longer in the first word than in the second. Compare:

  • He's a man.
  • He's a manager.

The reason for this is that with the word manager, we need to be able to squash the rest of the word in before the next stressed syllable arrives. Reducing the length of the vowel in /mæn/ enables us to squish in the next two syllables without radically increasing the length of time it takes to say the word. This is known as rhythmic clipping.

Within words themselves we can see processes similar to the ones above at work. English has families of vowels. In British English we can notionally divide these families according to the perceived length of the vowel. So in one group we have the long vowels, /i:, u:, ɑ:, ɔ:, ɜ:/ (as in bead, booed, bard, board, bird), and the diphthongs. In the other we have the 'short' vowels, which include, amongst others /ɪ, e, æ, ɒ, ʌ/ (as in bid, bed, bad, bod, butt).

Now, there seems to be a systematic relationship between many long vowels in English and other specific short vowels. If you speak English, these qualities will seem to be logical. In fact, they aren't. In terms of the actual sound there is almost no phonetic relationship between the long vowels and their short vowel counterparts in modern English. For example, the vowels in the words weight and bad are not very similar. We represent these sounds by the symbols: /eɪ/ and /æ/ respectively. However, these vowels have a very close relationship in the language. For example, the following words have the /eɪ/ vowel:

  • grateful
  • sane
  • inflame

The following words, on the other hand, even though they have the same root, usually have the vowel /æ/:

  • gratitude
  • insanity
  • inflammatory

This change from long to short vowels usually happens when there are extra syllables added to the base or root of the word. The more syllables there are in a word the more short vowels and the fewer long vowels we are likely to see. We can think of this as a kind of phonological rhythmic clipping. It helps to reduce the amount of time needed to say that syllable in order to accommodate the other unstressed syllables. The following long and short vowels have this same relationship:

  • eɪ / æ
  • i: / e
  • aɪ / ɪ
  • əʊ / ɒ
  • aʊ / ʌ

These relationships can be seen in the following pairs of words respectively:

  • chaste / chastity
  • penal / penitentiary
  • wise / wisdom
  • joke / jocular
  • south / southern

The Original Poster's Question

The /aɪ/ diphthong that we find in words such as Michael, Christ and wise has a phonological relationship with the short vowel /ɪ/ in English. Because English is stress timed, it has systematic methods for reducing vowel lengths in order to accommodate unstressed syllables without radically disrupting the natural rhythm of the language. When we add syllables to base words containing the PRICE vowel, /aɪ/, it is likely to change to a KIT vowel /ɪ/. This is what we see in the Original Poster's examples:

  • 'maɪkl /'mɪklməs
  • 'kraɪst / 'krɪsməs
  • 'waɪz / 'wɪzəd
  • 'saɪkl / baɪsɪkl

Transcription note

I have used a standard British English transcription, as used by John Wells in Longman Pronunciation Dictionary

Araucaria - Him
  • 47,146
  • 74
    This is absolutely fascinating. I'd never heard of, or noticed, the regular timing of the stressed syllables but it's absolutely there once you start listening. It's most obvious in emphatic speeches, i think, of politicans for example: it even enables them to bang their fist along with the stresses. – Max Williams Jun 01 '16 at 12:47
  • 1
    @MaxWilliams Yes, English is really good for that! – Araucaria - Him Jun 01 '16 at 12:49
  • 26
    Absolutely fantastic answer. Also, for someone wanting to hear the difference between syllable and stress timing within a language, compare Brazilian Portuguese (strongly stress timed) and European Portuguese (strongly syllable timed). – user0721090601 Jun 01 '16 at 13:19
  • 2
    @Araucaria By "phonological relationship", did you mean they've developed from the same older form (e.g. Old English /i/). As OE /i/ developed into Modern English, after the Great Sound Shift, it turned into /aɪ/ in standalone words, but /ɪ/ in compounds due to "rhythmic clipping"? – Vun-Hugh Vaw Jun 01 '16 at 13:52
  • 5
    @Vun-HughVaw I really meant synchronically in modern English. However, you're right that these long and short vowel pairs were originally long and short alternations of the same vowel in Middle English. During the Great Vowel Shift the long versions changed quite radically and the short versions didn't. I don't know very much about Middle English at all though! – Araucaria - Him Jun 01 '16 at 14:21
  • I'm guessing stress timing has a very important relationship with iambic pentameter? – pjc50 Jun 01 '16 at 15:18
  • @pjc50 I assume it must. I doubt that iambic pentameter would be effective and suspect may not even be possible with a very strictly syllable timed language. But I'm not sure! – Araucaria - Him Jun 01 '16 at 15:50
  • 12
    I like that you have the disclaimer about the "regular" intervals not being regular. But then you go ahead and say that "fish, peas, beans, rice" and "fish and then some peas and a banana and a casserole" have "the same duration" in "normal relaxed speech"; I'm very dubious of this. The second one definitely seems longer when I say it to myself. Isochrony is a complicated topic; there is an interesting answer on Linguistics SE by Robert Fuchs that suggests that it mainly applies to vowel durations, not to the duration of whole utterances. – herisson Jun 01 '16 at 16:06
  • 5
    @sumelic Tap it out as you go! The point here is that perceptually English appears to be strictly syllable timed. But if you measure it, it isn't. Having said that, if you were to stick English on Fuch's notional continuum it would be fairly near the end of the stress timed pole. Certainly, I've never met a non-native speaker for example, whose speech was too stress timed! But it's the general effect that counts, not the numbers on the gadgets! If you measure those two sentences they'll be much closer together than if you add another single syllable to the first list. – Araucaria - Him Jun 01 '16 at 16:16
  • 1
    Of course they have the same overall rhythm (except "casserole" has secondary stress on the last syllable for me), so the number of taps will be the same. But as you've acknowledged, the number of taps is not actually directly proportional to the physical duration. To make them the same duration, I have to say "fish, peas, beans, rice" more slowly than my normal tempo, and "fish and then some peas and a banana and a casserole" faster than my normal tempo. This is possible, but I don't think it's particularly natural. – herisson Jun 01 '16 at 16:21
  • 1
    @sumelic It's fairly well related to the duration, I reckon. What you'll find though is that they are not as regular as they appear to be to the naked ear. Thing to do is just to get someone to say them and time them, I suppose. I don't agree even a tiny bit though with your say them at a slower tempo. Why? Because that list appears before the other. Why would you say that list unnaturally slowly in anticipation of the next. That won't stack up at all! – Araucaria - Him Jun 01 '16 at 16:26
  • That's exactly my point. I disagree that people would say these two phrases with the same duration. If I say "fish, peas, beans, rice" normally, it definitely takes less time than it does to say "fish and then some peas and a banana and a casserole" normally. – herisson Jun 01 '16 at 16:28
  • 7
    @sumelic Try fish, peas and a banana and a casserole. That's 50% of one list and 50% of the other. Try and do that so that the third and fourth stresses don't come out perpetually at the same intervals as the first two syllable whilst saying it naturally. It's well nigh impossible. Try it! – Araucaria - Him Jun 01 '16 at 16:33
  • 2
    Do you mean fewer long vowels rather than less long vowels? – cfr Jun 02 '16 at 02:42
  • 1
    @MaxWilliams For interest - Isochrony – J... Jun 02 '16 at 10:32
  • 5
    I recorded myself saying fish, peas, beans, rice. and fish and then some peas and a banana and a casserole on my iPhone 3, the latter is approximately a second longer (4 secs.) Hardly any difference considering the length of the sentence. What I did notice is that I really stressed on the first word, in both instances, and then the rest flowed. Perhaps if the sentence was in the middle of a dialogue the emphatic stress would be less noticeable. – Mari-Lou A Jun 02 '16 at 14:43
  • I like this answer as much as I like the one about different stresses of vowel sounds based on whether it's the noun or verb form (record vs record, etc) – ebwb Jun 02 '16 at 21:50
  • 10
    "If you speak English, these qualities will seem to be logical. In fact, they aren't" - I feel like every answer on this site should include this somewhere! – user56reinstatemonica8 Jun 02 '16 at 23:41
  • hi @sumelic. If you're dubious - you're wrong. ask any English songwriter. – Fattie Jun 03 '16 at 17:23
  • 1
    indeed, competent English readers/speakers, will take longer to say "Fish, peas, beans, rice." than the other phrase if presented that way on a card in the studio, since any sophisticated reader will interpret it to be a somewhat contextually emphatic list (and hence punch it even more than you structurally have to in English); the same reader ... – Fattie Jun 03 '16 at 17:32
  • 5
    ....when it comes to "Fish and then some peas and a banana and a casserole." will indeed even further try to force it into the inevitable stress syllables - because English speakers can and do absolutely rip through the "filler". People think of Spanish, etc, as very fast; English rips but only through the interstitials. The "filler" in "Fish and then some peas and a banana and a casserole." becomes a blur, a hint, just a hustle of air, barely pronounced. (A central difficulty for speech recognition software.) – Fattie Jun 03 '16 at 17:34
  • 1
    Great answer. Particularly useful for me, when my four year old asks me why a word is spelled/pronounced in an apparently illogical manner! – Joe Jun 03 '16 at 17:56
  • @MaxWilliams Funny how these things are language dependent; if a Czech politician tried to do the banging thingy, he'd certainly both look and sound pretty silly :) – yo' Jun 03 '16 at 21:02
  • 3
    This is possibly the best answer I've read on any SE site. Thank you so much for writing it all out! – Ross Presser Jun 04 '16 at 05:20
  • Is this by any chance why children's books (think The Cat in the Hat) have the "sing-song" rhythms that they do? – O.M.Y. Jun 04 '16 at 22:28
  • 3
    I remember being at a lecture on Shakespeare where the professor provided convincing evidence that people in those times spoke MUCH faster than we do today. Speeches on stage were extremely fluid and subject to little or no dramatic pauses at all, so obviously the adjustment of the syllabic emphasis would have been critical. Amazing I never heard of this concept before. Thank you @Araucaria – O.M.Y. Jun 04 '16 at 22:41
  • @O.M.Y. That's very interesting! – Araucaria - Him Jun 04 '16 at 23:10
  • 1
    @Araucaria -- I have no direct contact information (this was years ago) but the instructor was Dr. Colin Cox if you want to follow up on this. He had some rather interesting ideas on the Bard himself as well. – O.M.Y. Jun 05 '16 at 02:47
  • 1
    @Araucaria The unvoiced consonants are actually fairly irrelevant here: the voicedness of the plosives is only a factor in homorganic lengthening, but in Christ, wise, and Michael, the vowel is inherently long, and the short vowel is the result of pre-cluster shortening (I believe Michaelmas, a contraction of Micheles mæsse, reflects an encroachment on pre-cluster shortening to contexts with no cluster). Drive/driven, on the other hand, is regular ablaut for class I strong verbs in OE. – Janus Bahs Jacquet Jun 05 '16 at 17:05
  • 3
    The crux here is (which you don’t really bring up in your answer—a worthy addendum, I feel) that there is an intricate and far from regular correspondence between the various different (historically) long/short vowel pairs in Modern English, and that this correspondence is the result both of (mostly) regular historical lengthenings and shortenings that hinge on phonetically predictable circumstances and of verbal ablaut patterns that go back even further and were then themselves subject to these same lenghtenings and shortenings later on. – Janus Bahs Jacquet Jun 05 '16 at 17:10
  • 1
    (I’ve added a substantial bit of detail to the section on pre-cluster/polysyllabic shortening in my answer on the ‘milk/find’ question that also makes it encompass the words in this question.) – Janus Bahs Jacquet Jun 05 '16 at 17:34
  • Could you add links for the first use of PRICE/KIT for those of us unfamiliar with those terms? – JDługosz Jun 05 '16 at 20:13
  • 3
    @JDługosz Let me hunt around for a good link (might take a day or two as I'm mad busy right now). In the meantime, PRICE, KIT and so forth are the names of "lexical sets" an invention of Professor John Wells. He chose one word (which it was felt was unlikely to be confused with other words), and used this word as a title for a) the vowel that it contained and b) the set of all the words in the language that use that vowel. So essentially PRICE isusually just the name for the /aɪ/ vowel that you hear in the word price. ... – Araucaria - Him Jun 05 '16 at 20:54
  • 3
    @JDługosz This means that you don't need any training in phonetics or reading transcriptions to understand which vowel is being talked about! It also gets round the problem of different writers using different symbols to represent the same vowel - and also makes it easier to talk about them. Instead of saying "The iiiii vowel", we can say "the FLEECE vowel", and so forth. Until I get round to finding a good link, here's one to John Wells' phonetics blog on the subject: Lexical sets – Araucaria - Him Jun 05 '16 at 20:59
  • Comments are not for extended discussion; this conversation has been moved to chat. – Matt E. Эллен Jun 07 '16 at 09:28
  • You correctly indicate the vowels are changed to make words easier to pronounce, and you give an excellent explanation of what a stress-timed language is. But similar vowel changes occur in strongly syllable-timed languages too, such as Spanish, so this is not exclusively a stress-timed thing. I wrote ananswer about this but it's been deleted. See discussion at http://meta.english.stackexchange.com/q/8065/69422 I think your answer is very well written, though I disagree with a small part of it, so +1. I fail to see how deleting my answer improved the site, however, hence the meta question. – Level River St Jun 07 '16 at 20:36
  • @LevelRiverSt I was very sad to see your answer deleted. I think your answer should have been taken as putting the English phenomenon in context. (I don't think that the reasons the vowels change in Spanish are even vaguely the same as they are in English at all though!) – Araucaria - Him Jun 07 '16 at 21:50
  • If "English has systematic methods for reducing vowel lengths in order to accommodate unstressed syllables", "obtainable" should be pronounced /əbtænəbl/, not /əbteɪnəbl/. The same rule should apply to "maintainable". Of course, "maintenance" has a different spelling and pronunciation. How can we explain the difference in o and e sound between "obey" and "obedience". I think "penalty" is a better example than "penitentiary" as the accent is not on the first syllable. –  Jun 08 '16 at 17:58
  • 2
    @Rathony I don't think these systematic methods apply to words like "obtainable" or "maintainable", which were probably coined in English rather than borrowed as a whole from French or Latin. Highly active morphemes such as -able, -ing, -ful, etc. usually don't affect the pronunciations of the main vowels of the stems. Not to mention, based on the spellings alone, /eɪ/ as in "obtain" may not be the same as /eɪ/ as in "chaste", therefore it might not have any relationship with /æ/. – Vun-Hugh Vaw Jun 18 '16 at 02:20
  • I suppose this is grist for the mill!! – Hot Licks Jul 09 '20 at 13:01
-2

Christmas was said as krisstmahs while Christ was said as Kreest,the ee sound was the long version of ih so ''I'' in english was once said the way Dutch,Danish etc say it as modern english ''ee''.A word that has one syllable's easier to be given vowel length then ones that have more than one,also y and w sounds tend to fall/merge with the following/previous sound when the the semivowel is followed by a sound that is very created closely in the mouth to where they are eg: y sounds and s/sh/ch and w sounds and r(WRite/sORe/cROissant) also consonants next to each other have a tendency for one of them to fall,since it's easier to pronounce only one of them(french orthography is full of this trend since the orthography was never changed and also cause it depends on the position of the word), now when the great vowel shift happened(that took ages) krisstmas only lost it's t and mahs became mss/muhs Christ went from kreest became krihyst to kruhyst(many irish and canadians still say it like this) to kraahyst(many americans say it like this), some people have it gone further to something along to Krahyst or even Kroyst(many southern brits)

Artsaregreat
  • 17