Do /ɪə/, /eə/, /ʊə/ diphthongs actually exist in General American as phonemes?

Question

The Handbook of English Pronunciation. (Marnie Reed, John Levis referring to J.C. Wells)

Аs the pronunciation of most speakers is rhotic, there are no centring diphthongs, because the vowels /ɪə, eə, ʊə/ in words such as peer, pair, and poor are a sequence of a monophthong followed by /r/ so the rhyme of these words is /ɪr/, /er/, and /ʊr/ respectively.

Acoustics of American English Speech: A Dynamic Approach (Joseph P. Olive, Alice Greenwood, John Coleman)

There are five diphtongs in American English, /eʲ/ as in bait, /aʲ/ as in bite, /ɔʲ/ as in boy, /aʷ/ as in cow and /oʷ/ as in boat.

Oxford Dictionary marks /ɪə/, /eə/, /ʊə/ as strictly British.

Wikipedia quoting Wells in "General American English" article.

When prosodically salient, the lax vowels /ɪ, ʊ, ɛ, ʌ, æ/ tend to be realized as centering diphthongs [ɪə, ʊə, ɛə, ʌə, æə] instead of the more usual long monophthongs [ɪˑ, ʊˑ, ɛˑ, ʌˑ, æˑ] when they precede a word-final voiced consonant, so that the word good in the sentence that's very good! tends to be pronounced [ɡʊəd] instead of [ɡʊˑd].

That is why fear [fɪr] may become [fɪər] or poor [pʊr] may become [pʊər] when emphasized just as well. In my humble non-native speaker opinion.

Why, despite all that, do I keep seeing guides like this that insist on existence of these diphthongs in AmE? Is it a sort of alternative theory I'm not aware of, or do these authors interpret diaphonemic notation literally?

Answer 1

No, they do not.

The diphthongs /ɪə, eə, ʊə/ found in Received Pronunciation stem historically from the sequences of /iː/, /eː/ (now /eɪ/), or /uː/ + /r/. In about the 16th century, /iːr/ etc. started to become realized more like [iːər] etc. Later they became more like [ɪər] etc. At this point it would be unreasonable to posit the diphthongs as phonemes because /iː/ etc. were consistently realized as diphthongs before /r/ (so the diphthongs were merely allophones). But then it became fashionable to drop /r/ after vowels in Britain. This gave rise to the diphthongs /ɪə, eə, ʊə/ as distinct phonemes in British English.

General American didn't go through the r-dropping, so even if the diphthongal realizations were retained (which they kind of are—see below), it would be more theoretically sound to analyze the diphthongs as allophones of the long vowels which they originate from. But in GA, /iː, eɪ, uː/ and /ɪ, e, ʊ/ are neutralized before /r/, meaning there are no minimal pairs of /iːr/ etc. and /ɪr/ etc., in that nearer and mirror rhyme and so do Mary and merry (which they do not in RP). This is why some phonemicize near as /niːr/ and others as /nɪr/. (Often the length mark <ː> is omitted and <ɛ> substitutes for <e> in analyses of AmE, but here they are retained for comparison with BrE.)

/niːr/ and /nɪr/ are both defensible, but the latter is more common because key-ring and hearing typically don't rhyme (some pronounce zero, hero, etc. with /iː/, but these are rare exceptions and in free variation with /ɪ/). And although the vowel is short and monophthongal when followed by a vowel, as in caring or nearer, when preceded by a pause or consonant, as in cared or near, the vowel is often diphthongal and resembles the RP counterpart (except followed by /r/). But that doesn't mean they are phonemes in GA, which requires minimal pairs. (Also note marry /ær/ also merges with Marry, merry /er/.)

See Wells's Accents of English (1982), vol. 1, sections 3.2.1 and 3.3.1 for more.

Those who insist on the existence of centering diphthongs as phonemes in AmE do so either because they think that allows for better comparison with RP, because their model accent is different from what linguists have referred to as "General American", "Network English", etc., or because they don't know what they're talking about.

Answer 2

I'm not going to answer the question as to whether these are separate phonemes. That depends on the definition of phoneme, and which native speaker you ask. For the standard linguistic definition of phoneme (where you require minimal pairs), there doesn't seem to be much justification for treating these as separate phonemes.

I will try to answer the question as to why some dictionaries treat these as separate phonemes.

Besides the standard linguistic definition, there are other definitions of phonemes, for example.

The central concept in phonology is the phoneme, which is a distinctive category of sounds that all the native speakers of a language or dialect perceive as more or less the same.
(Thomas E. Murray, The Structure of English: Phonetics, Phonology, Morphology. Allyn and Bacon, 1995)

Many dictionaries seem to use a definition closer to this one, where they treat two things are different phonemes if they are perceived to be different sounds by speakers of the language.

This is presumably why some dictionaries use different phonemes in chintz (/tʃɪnts/) and wince (/wɪns/); even though nearly all English speakers pronounce these so as to rhyme, not all English speakers perceive these words as rhymes.

Many Americans perceive the r-colored vowels to be a completely different set of vowels from the regular vowels. In fact, for some vowels (and some Americans), this is completely justified. The vowel I use in cord does not sound very much like the vowel in either cawed or code. Similarly, the vowel I used for card is somewhere in between cad and cod. So if the dictionaries use this definition of phoneme, there is justification for having a different phoneme for /ɪə/ than either /ɪ/ or /i/.

To make things more complicated, some Americans perceive the vowel of beer to be the same as in bid (/ɪ/), while some perceive it as the vowel of bee (/i/), and some perceive it as a different vowel altogether. So if you chose either /i/ or /ɪ/ for this vowel, you would disagree with some Americans' perceptions of the phoneme.