She needn't've played with homophones, which surely would have to've been marked by some sort of pragmatics. (I'm picturing stress on "you" followed by a short pause.) The sentence "I want you more than all the riches of the world" is already ambiguous between "I want you more than (I want) all the R.O.T.W.") and "I want you more than all the R.O.T.W. (want you)." The latter sentence will almost always be true.
There was a similar thing in Dilbert when his girlfriend asks "Do you love me more than that computer?", and Dilbert responds, "No, I love you more than *that* computer". But he thinks, "Don't ask me about the laptop…"
Has there been an LL post on the common 'then'~'than' typo? I can't work out what's causing it.
Are the two actually homophones in any English accent? Those in the far East that have the mat-met merger? Does the Northern Cities Vowel Shift get you anywhere close? And does that even help, since the vowel in 'than' is generally schwa anyway?
As I understand it, it's because treating and as homophones is metanalysis. They're the same word in origin (I'm using etymonline as my source), and my wildly uneducated guess is that they're both subject to a wide variety of pronunciation variations by dialect and stress.
@Pflaumbaum: Native Utah English speaker here. I would only ever pronounce than differently from then in some rare moment when I must emphasize the contrast. In ordinary speech both words have /ɛ/.
But as a spelling stickler, I never ever spell either word worng.
ClockwerkMao: Your example words disappeared, probably because you wanted them in italics but used arrows or other code in a way that made them invisible (other commenters will know). The absence of the words would have been evident in the window that reproduces the text as it will appear. If you are not sure how to fix this, use " " instead of italics and be sure to check the window under "Submit Comment" before you do submit.
@Pflaumbaum. I don't think the "then" for "than" thing is an accent thing. I think most speakers whatever the accent will usually use a reduced form of "than", with a schwa, /ðən/, since the word is usually unstressed. And while some of us will always think of it as /ðæn/ when typing or writing, and thus never spell it as "then" apparently, some folks don't do that and think it in it's unstressed form, thus leaving themselves open to misspelling it as "then", since, unstressed, "then" and "than" are identical or nearly so.
If any speakers have these words both having the same pronunciation in stressed form, I think it's not a matter of accent, but of not having learned the standard stressed pronunciation of "than".
According to the Longman Pronunciation Dictionary, "than" has a strong form with a TRAP vowel and a weak form with a schwa, while "then" only has a strong form with a DRESS vowel. So they are never homophones in this description, in agreement with dw's comment above.
You can read the relevant discussion on John Wells's phonetic blog, where Wells (the editor of LPD) explains in the comments that he doesn't have a weak from for "then", but perhaps ought to include it in future editions as a possibility (though accompanied by some sort of warning for the benefit of EFL transcribers).
The comment thread on John Wells' post has someone claiming that than/then confusion in writing is rare in the UK but common in the US, which is certainly consistent with the notion that the words are homophonous for many US speakers (including the cartoonist whose work is shown above and that cartoonist's assumed audience) even if not in most/all BrEng varieties. I take Wells' point that "than" is pronounced in reduced form the vast majority of the time, but pronouncing unreduced/strong "than" with the TRAP vowel rather than the DRESS vowel sounds like a hypercorrection/pretension/spelling-pronunciation to my AmEng ear. I *think* this means that I conceptualize the words as homophonous in the abstract even if pronounced slightly differently in practice because of the ubiquitous schwaification of "than," which "then" is significantly less prone to for, um, presumably some mix of syntactic and prosodic factors.
In my type of American English, I think the unstressed vowel in than may be [ɛ]-like at times. However, than is unstressed the vast majority of the time and then usually has some degree of stress I think. So the [ɛ]-like phone which occurs in than at times is probably shorter and more centralized than the [ɛ] in then.
Stressed then and than are pretty clearly distinct for me (even though I rarely stress the latter). Although my pre-nasal allophone of /æ/ may be somewhat raised, I still make a distinction between pre-nasal /æ/ and /ɛ/. I do, however, think the capital of Austria is /vi'ænə/. Also mayhem and cayenne are /'meɪhæm/ and /kaɪ'æn/, respectively, for me. I don't know why though. Those might just be idiolectal oddities. Surprisingly though, other people haven't seemed to notice them. Some Americans (but not me) seem to have /kɛn/ for can (as in You should try to do it if you can.). I'm from a part of the Midwest that is supposedly accentless, although we know that that's not possible from a linguistic standpoint.
@Ellen K: I can easily think of situations where "then" is unstressed. In particular, in "If…then…" constructions.
Not for me. I have been racking my brains to try to think of some context in which "then" could be unstressed (to schwa): all I can come up with are meaningless sentence-initial space-fillers such as "Well then" and "Now then". I'm not even sure about them.
I'm skeptical about introspective claims that two vowels are phonetically distinct. Vowels don't occupy discrete points in vowel space; if you plot a bunch of tokens, you see some dispersion around the mean value. See this graph for instance—there's a shit-ton of overlap. It would be interesting to record some spontaneous speech from, say, dw above. I'd bet there's substantially more overlap than his/her intuitions suggest. The stuff about the TRAP vowel and the DRESS vowel represents a slightly idealized view, I think.
My own unreliable intuition is that I don't think I distinguish them consistently. It feels like both have a vowel somewhere in the [ɛ ~ ɨ ~ ɘ] space (note–that's not a schwa). I bet a lot of Northern Cities speakers are similar. I also definitely have an unstressed "then" in utterances like if you don't like it then leave! (I speak a variety that seems to have a lot of reduction to unstressed syllables.)
I'm skeptical about introspective claims that two vowels are phonetically distinct. Vowels don't occupy discrete points in vowel space; if you plot a bunch of tokens, you see some dispersion around the mean value. See this graph for instance—there's a shit-ton of overlap. It would be interesting to record some spontaneous speech from, say, dw above. I'd bet there's substantially more overlap than his/her intuitions suggest.
Fair enough. Put it this way: I have the same degree of certainty that "then" is distinct from "than" as that "end" is distinct from "and".
You're right. There is a lot of overlap there. But do those vowels overlap when they occur in the same phonetic environment? And if they do then is there some other way of distinguishing them, e.g., length (which that graph doesn't show)? I'm only an amateur in linguistics, so please don't be too harsh.
I think in practice than does not have TRAP with any real frequency for most speakers. Generally, the only time we would use it in it's strong form (with TRAP) is when citing the word, or in mathematics, when labeling , as well as, sometimes, when talking about the mathematical operations represented by those symbols.
Otherwise, it's a weak form with a reduced vowel, which very well may have a ɛ-like quality.
Let me clarify. In the above, I'm not assuming that all speakers have TRAP for the strong form of than. Rather, I'm noting that it only has TRAP in it's strong form. And also, I'm talking about speech, not how we hear things in our head when reading or writing.
@GeorgeW, I'm also having trouble with the punctuation in the original strip. Couldn't she just be saying "You're in first place, and all the riches in the world are in second place"? Which ought to be a good enough compliment for anybody.
I'm almost certain there are no American dialects in which /æ/ and /ɛ/ are merged, before nasals or anywhere else. However, there are dialects in which historical /æ/ has been replaced by /ɛ/ in a small set of function words such as "can" and "am", and "than" might as well be in the same class.
I'm almost certain there are no American dialects in which /æ/ and /ɛ/ are merged, before nasals or anywhere else.
Before /ŋ/ perhaps. The vowel in words like "gang" often seems to approach /e/ in closeness. However, there are few words with the DRESS vowel before /ŋ/: one might expect it in "length" or "strength", but many speakers have plain /n/ in these words.
After a quick Google search, I found this paper by Douglas Bigham. It talks about a possible "pen-pan merger" in Southern Illinois English. I haven't read it yet though.
FWIW, the vowels of gang [geɪŋ] and strength [stɹeɪŋkθ] do sound the same to me (AmEng speaker). But length, oddly enough, seems to have a different vowel than strength. It sounds more like [lɪŋkθ]. That's just my impression of my own accent though. And I'm not trying to claim that all Americans talk like me.
AJD's explanation is the best so far. It puts [then for than] in the same area as [of for 've], which is an explicable error because they're usually complete homophones (pace Ellen K., who if I recall right would not agree with this for her own idiolect). And it doesn't require a general merger of /æ/ and /ɛ/ (or /ə/ and /ɛ/) before /n/, or anything like that.
But does it explain [than for then]? There are 7.6 million Google hits for "It was than that". Are we saying that the two are homophonous but people know there are two spellings, so they're just guessing? In which case, why are there hardly any hits for things like "friend've mine"?
You must have misread something I wrote, Pflaumbaum. I certainly did not claim they aren't homophones, and did not make any claims at all about my speech in particular. Certainly my own speech is a factor in any generalization I make, of course, but, still, my comments weren't about my own speech in particular. And, furthermore, the generalization I made is that then and than, when unstressed, are identical, or nearly so.
Pflaumbaum: AJD's explanation is the best so far. It puts [then for than] in the same area as [of for 've], which is an explicable error because they're usually complete homophones
Another similarity is the words "of" and "from". These generally have the STRUT vowel when stressed in North American accents (while in other accents they generally have the LOT vowel). The usual explanation for this is restressing of the unstressed forms with schwa.
For Mount Kosciuszko, Wikipedia says the traditional pronunciation has the LOT vowel in the first syllable of Kosciuszko, but the Macquarie Dictionary says the first vowel is STRUT if I recall correctly.
My theory for the latter pronunciation is that the LOT vowel tends towards the schwa since it doesn't receive primary stress (as is common in Australian pronunciation, e.g. the word 'Australia' itself). But it cannot be completely reduced since the 'Teutonic rule' of English stress would be violated if the first two syllables were both unstressed (the primary stress of 'Kosciuszko' falls on the letter 'u', and 'i' is a separate syllable in the traditional pronunciation). So LOT turns into schwa but is re-stressed as STRUT.