Is this discussion absurd, or what?
If you replace [N] with [h] in "sing", you will change the meaning (the word will become unintelligible), therefore [N] and [h] are separate phonemes.
I make phonemic distinctions between [N] and [h].
I distinguish ''ngoo'' (a Welsh word) from ''hoo''. i.e. [Nu:] vs. [hu:]. [N] and [h] are not allophones for me.
I also pronounce ''Ngami'' as [Na:mi].
Bob, ''vain'' and ''vein'' are pronounced the same. What do you mean you pronounce them differently?
Joe writes-''What do you mean you pronounce them differently?''
I mean, In my accent they are pronounced differently. ''vain'' is [v@in] and ''vein'' is [vein]. You might not make the distinction, but I do.
Jim writes-''Indeed, then how can you claim that /h/ and /N/ (as in "hung" = /h^N/) are not allophones?''
Jim also writes-''Give us a minimal pair to show it.''
I've just shewed one above. The one I'm shewing you is ''ngoo'' vs. ''hoo''. I pronounce ''ngoo'' as [Nu:] and ''hoo'' as [hu:].
The discussion has become absurd but the point I had been trying to make is, I think, valid and sane. Certainly, if you take the word "sing" and replace the phone [N] with the phone [h], the word will become unintelligible. But I'm talking not of replacing one phone with another but of replacing one phoneme with another.
Now, if someone were to come up with crazy idea that there was a phoneme, /H/, in English which is pronounced [N] between vowels and at the end syllables and is pronounced [h] at the beginning of syllables, what have we got to proove them wrong? We haven't got any minimal pairs in English.
Yet, we all know that /h/ and /N/ are separate phonemes. How do we know it if there are no minimal pairs (in English, Bob, in English)? I say we know them to be distinct phonemes just as well as we know "pin" and "bin" to be distinct words.
In other words, I don't think we need rely on the existence of minimal pairs to define the phoneme. The, on the other hand, I doubt that I could give an unproblematic defintion without relying on minimal pairs.
Jim, what about if someone came up with this crazy idea that [u] and [ ] were allophones. Couldn't [u] and [ ] be allophones? [u] never occurs at the beginning or end of a word. There are no minimal pairs between [u] and [ ]. Therefore, couldn't we say they are allophones. Thus, we could say that the ''oo'' in ''wood'' and ''book'' is silent.
"Now, if someone were to come up with a crazy idea that there was a phoneme, /H/, in English which is pronounced [N] between vowels and at the end syllables and is pronounced [h] at the beginning of syllables, what have we got to prove them wrong? "
How about this:
First, we notice that the /H/ phoneme has two allophones: [h] and [N].
Then we take the word "sing" [siN] and replace the final phone with [h].
The word becomes unintelligible.
Therefore the two phones [h] and [N] are not elements of the same phoneme. QED :)
Jim, "Now, if someone were to come up with a crazy idea that there was a phoneme, /H/, in English which is pronounced [N] between vowels and at the end syllables and is pronounced [h] at the beginning of syllables, what have we got to prove them wrong? "
Jim, ''Ohio'' is pronounced [OuhaiOu]. So, therefore your idea of [h] and [N] being allophones won't do. ''Ohio'' is an example of a [h] in between two vowel. Other examples are ''behind'' and ''behold'' and ''aha''.
Joe, please don't post messages which don't contain any text of interest to other users (such as the one above), OK?
Jim, I wonder if you're not replying because you agree or because you haven't noticed my post...
What about the Australian ''steak'' vs. ''stake'' distinction? Is that difference only in vowel length. I've heard by Dave (an Australian on the other thread) that it is.
Here's a link to that thread,
It was because I hadn't noticed your post but what you say seems reasonable enough. However, aren't we now extending the definition of the phoneme beyond the simplistic minimal-pair requirement?
All I've been trying to do is bring up a difficulty involved in a definition of the phoneme which relies on minimal pairs. I'd read something about this problem with /h/ and /N/'s having no minimal pairs in English. I'd read that this is a problem for the minimal-pair definition of the phoneme. I'll see whether I can't dig the article up: the author put the point better than I've managed to.
I was just wondering whether it was not possible to come up with a more fundamental definition such that the phonemic distinction between these pairs is a consequence for this definition rather than it's basis.
Mxsmanic, for example, says that vowel length is unimportant in English but aren't /i/ and /i(:)/ distinguished primarily by length? Wouldn't you call /i/ and /i(:)/ different phonemes? I've been trying to find a minimal pair ... does "fifty note" vs. "fifteen oat" count?
"... consequence of ..." not "... for ..."
The 'y' in "fifty" and "ee" in "fifteen" have different vowel positions in my accent, but I agree that vowel length also helps to dinstinguish "fifty note" from "fifteen oat".