The Alignment of IPA in Learner's Dictionarys

Vowel Length & Vowel Quality

Several hundred years ago, the English language had a distinction based on vowel length as in German(ex. Stadt and Staat). Each short vowel and its corresponding long vowel had almost the same vowel quality. But short vowels and long vowels have gradually turned in different directions. They are mutually distant today. Consequently, former vowel length difference has been transformed into current vowel quality difference. Because of such history, however, we still have common terms such as Short-U and Long-U.

This is my understanding. Words such as "short" and "long" are not about vowel length. But in phonetics they are used for length or duration. Verry puzzling. So we need a clear explanation.

Nevertheless, it is better to learn short-A,E,I,O,U and long-A,E,I,O,U in pairs. This Short-Long relationship should be considered when dictionarys show us IPA list.

<<My teacher of (American English) says SUCKS has a short a: [saks] while SOCKS has a long a: [sa:ks]. Is it true? Many times I hear this.>>

No. The two vowels are very different. They're produced in different parts of the mouth and neither is longer than the other, at least in General American English.

<<I would say sucks has a short U and socks has a short O, and neither have an A of any kind. But apparently I would be wrong.>>

No, you're just thinking of spelling, which is completely removed from what we're talking about. "Short u" implies XSAMPA [U], which is the sound in "book." "Short o" at least in British English implies [Q], which is their sound for words in "sock" (and Lazar's dialect too). However, General American has [V] for "suck" (like BrE) and [A] for "sock."

General American English is broad, but aren't contrastive lengths, such as the traditional one between "sucks" and "socks", still maintained in most American dialects? Because [V] and [A] move close together in quite a number of dialects.

<<I would say sucks has a short U and socks has a short O, and neither have an A of any kind. But apparently I would be wrong.>>

No, you're just thinking of spelling, which is completely removed from what we're talking about. "Short u" implies XSAMPA [U], which is the sound in "book." "Short o" at least in British English implies [Q], which is their sound for words in "sock" (and Lazar's dialect too). However, General American has [V] for "suck" (like BrE) and [A] for "sock."


Yeah, I kind of figured I'd get crap for that.

There was no XSAMPA when I was in grade school, and short U was "uh" as in "cut" and long U was "oo" as in "cute".

Short O was "ah" as in "cot", and long O was "oh" as in "coat".

We didn't worry too much about international dialectical variation when I was seven or eight.

There are no minimal pairs in Mxsmanic's dialect of English that are distinguished solely by vowel length. In mine there are plenty however, according to Mxsmanic, my accent would be some substandard abberation.

I will agree, though, that the terms "long" and "short" are misused in English-speaking schools. Uriel writes "those terms describe the sound of the vowel, not the length of time you hold it." Yes, that's exactly how we're taught but in phonetics it's the other way around.

Minimal pairs are not black and white. Some minimal pairs are more significant than others, because they occur more plausibly and frequently in speech. Minimal pairs involving vowel length, even in pronunciations where they exist, are typically near the bottom of the list. One reason why vowel length is no longer phonemic in standard English is that there are so few useful occasions for it to form minimal pairs. Since it provides a distinction that nobody really needs, it isn't used anymore.

For example, some people distinguish Mary, merry, and marry. For them, there are phonemic distinctions among these words. But in reality, these words rarely occur in similar contexts, and indeed they rarely are used at all, so phonemic distinctions between them aren't very important. As a result, most people pronounce them all the same way today.

Some dialects of English may develop phonemic vowel length to some length to compensate for their non-rhotic character and the loss of diphthongs that mark the 'r' in words. However, standard pronunciations don't do this. Standard non-rhotic pronunciations generally replace the 'r' with a vowel, forming a distinctive diphthong; and rhotic pronunciations simply pronounce the 'r'. Thus, neither requires phonemic vowel length, and neither uses it.

Well, Mxsmanic, excuse me for speaking this aberant sub-standard form of the language. Please do accept this apology I make on behalf of myself and my entire country. It's so good to know that there are people like you out there determining what is standard English and what is not.

When you refer to the significance of a minimal pair isn't best to specify the context? Those minimal pairs which occur more plausibly and frequently in speech certainly are more significant in the context of ESL, yes. However even those other rarer ones are significant with respect to establishing what your phoneme set consists of.

"Minimal pairs involving vowel length, even in pronunciations where they exist, are typically near the bottom of the list." you write. What I'm saying is that you're leaving somethings out. Minimal pairs involving vowel length only are typically near the bottom of this important-for-ESL list.

Nor are minimal pairs the be all and end all. There may be millions of people who merge "merry" and "marry" this doesn't mean that we should ignore the distinction between /e/* and /{/ (X-SAMPA). There aren't too many minimal pairs for /S/ and /Z/ however it's still important to pronounce them distinctly.

Monophthongisation of the centring diphthongs (e.g. "air" & "ear" going from [e@] & [I@] to [e:] and [I:]) is one way that phonemic vowel length has crept into Australian English. Another way is the merging of the positions of the vowels in "some" and "psalm" leaving only the length distinction (which we all still do make) to distinguish the phonemes.

But, of course, us Aussies can go and shove our stinking substandard aberant dialect where the Sun doesn't shine.

* I use /e/ for the vowel in "get".

It isn't necessary to specify the context when speaking of the significance of minimal pairs, as the notion of relativity and context are implied by the mere use of the term significance.

There is no fixed set of phonemes. Some sounds are more phonemic than others, in that they more often serve to distinguish meaning. Many sounds never distinguish meaning, and are thus not phonemes at all; but among those that do, their phonemic quality varies with the frequency and importance of the situations in which they perform this function.

Inevitably, strong phonemes are more universal and durable than weak phonemes. Weak phonemes may disappear completely from one pronunciation to another; whereas strong phonemes may shift the sets of allophones to which they correspond when one moves from one pronunciation to another, but they do not disappear.

Thus, if the vowel in "man" changes in a dialect to sound like the vowel in "moon" in standard English, then the vowel in "moon" will change also to sound like something else and distinctive, in order to preserve the important phonemic distinctions marked by these vowels. In contrast, a weakly phonemic difference, like that in the minimal set Mary, merry, marry that exists in some dialects, may simply be obliterated in other dialects, since it just isn't important enough to preserve. Many weak phonemic distinctions are serendipitous and ephemeral.

When it comes to american English, your teacher is right as socks would be long(sa:ks) and sucks would be the cut sound according to the IPA where you can look up when you say it(try to roll ur eyeballs upwards)..you would not be able t do it with the shwa sound.thats enhancing the skill where I teach.

Socks and sucks both have a short vowel in my American English, although it's not the same vowel for both words.

Things get more complicated here with allophonic vowel length, as not only does such not correlate to vowel quality in many dialects, in particular North American English dialects, but furthermore the allophonic vowel length can be critical for word recognition *completely separately* from vowel quality alone. For example, take the aforementioned examples of [bit] and [bi:t]; I myself would normally perceive them as the words "beet" and "bead" respectively, *not* as "bit" and "beet", as in my dialect for word-final consonants the length of preceding vowels is actually more important than the voicing or even length of the word-final consonants themselves. Even though I would in semi-formal speech pronounce those two words more like [bit] and [bi:t_X] (you could also transcribe such as [bi:d_0]), the actual length of the word-final consonant is secondary to the vowel length despite the vowel length being allophonic from a generative standpoint. Of course, my dialect is somewhat of an extreme case in this regard, but even still, allophonic vowel length does form a significant part of word recognition in many North American English dialects nonetheless.

In our native tongue here in India, The rate of speech becomes a big issue when it comes to speaking in english as our dilect is a 4 beat dilect, a person would either speak too fast or too slow; more experienced is the former than the latter.The vowel length therefore helps us help them bring thier rate of speech(ROS) down when they would in a high rate of speech manner say "dIp lIk" (both short sounds as they tend to get clipped with the high (ROS) instead of Di:p LeIk (IPA).

So the vowel length is more important in the relavance bit.