Commenters on a recent post ("Australian hypocoristics") discussed the vowel quality of the first syllable of McDonald's in detail and at length. The issues involved are interesting enough to deserve a post of their own.
One central question is whether there is more than one qualitatively distinct type of schwa (reduced unstressed vowel) in English, either within or across varieties.
The null hypothesis here, in my opinion, is that there's a single category, representing a phonologically unspecified neutral vowel, which takes on various qualities depending on phonetic environment, speaking rate, and various paralinguistic factors. This null hypothesis might be false, but I've never seen any clear evidence to refute it.
Many people, including some famous phoneticians, have advanced a contrary view, namely that (at least in some varieties of English) there are two distinct schwas, one higher and fronter, the other lower and backer. A supposed near-minimal-pair, for some of the schwa-splitters of my acquaintance, is Alice vs. Dallas. I've never seen any evidence from natural speech that supports this view — performance of (near-) minimal pairs is always suspect in such cases, since it's subject to facultative disambiguation.
Another position, not inconsistent with this first one, is that there are characteristic differences in the pronunciation of such vowels among dialects or varieties of English, with some varieties having higher-fronter schwas, and others having lower-backer ones. This certainly might be true, but again, I haven't seen any real evidence.
Why, besides a general scientific show-me attitude, am I skeptical of these opinions? One reason is that I've learned not to trust my own intuitions in similar cases.
For example, I have the very strong intuition that I pronounce ladder and latter differently. And when I produce these words as a decontextualized minimal pair, there are small but significant average differences between the two words, in the mean values of various phonetic measurements such as the relative duration and amplitude of the medial voiced tap — although the distributions for the two categories are highly overlapped. But in normal usage, these differences evaporate, or are overwhelmed by other sources of variation, or both. And when I try to identify randomly-selected examples of my own pronunciations of latter or ladder, even if the exemplars are taken from decontextualized minimal-pair performances, my performance is indistinguishable from random guessing, .
I reckon that my intuition in this case comes not only from the fact that the words are spelled differently, but also from the fact that latter is a word that I've mostly read rather than heard, while ladder is a word that I've mostly heard rather than read. In any case, I've learned from experiences of this kind that my own intuitions about how I pronounce things are not always reliable.
What about the argument from variable nickname formation, advanced by more than one commenter? Thus
I was surprised to learn recently that Americans sound the first syllable of 'McDonald's' as 'Mick', whereas we Brits pronounce it with a schwa or as 'Mac'. Hence the American hypocorism 'Mickey D's', and the British one 'Maccy D's'.
Other commenters (correctly) denied that any such general trans-Atlantic difference exists, but some still felt that finer-grained dialect differences in the pronunciation of initial unstressed M(a)c might underlie the nickname differentiation.
But entirely independent of how speakers may pronounce schwas, there's quite a bit of variation in how they behave in promoting particular schwas to the status of full vowels. An interesting example emerged from some fascinating (but unfortunately unpublished) work by Marian Macchi, done long ago at Bell Labs. (What follows is based on decades-old memories, and should be evaluated as such…)
Marian had observed that American English speakers generally exhibit anticipatory coarticulation between schwas and the vowels of following syllables: thus in adept the initial vowel is somewhat [ɛ]-like; in adapt the initial vowel is somewhat [æ]-like; and in adopt the initial vowel somewhat [ɒ]-like. She hypothesized that in more carefully-articulated speech, this coarticulatory effect would be reduced or eliminated. So she set up an experiment based on a dictation paradigm, where local high-school students were asked to read lists of words to an experimenter, who occasionally pretended not to have heard them clearly, and interrupted with something like "Excuse me?" or "Could you repeat that?"
What she found was that in the (definitely hyperarticulated) repetitions, some of the subjects indeed produced un-coarticulated (or at least less coarticulated) schwas, [əˈdɛpt] [əˈdæpt] [əˈdɒpt] — but others went in the opposite direction, producing clearly-articulated forms with complete assimilation of the schwa as a full copy of the following vowel, [ɛˈdɛpt] [æˈdæpt] [ɒˈdɒpt] . . .
These were all teen-aged speakers from the same high school in a suburban area of northern New Jersey, apparently all speakers of the same variety of English. Marian's experiment suggests that some speakers apparently are prone to promote predictable context-conditioned phonetic variation to phonemic or even lexical status, while others aren't.
But there's another source of potential variation, which is specifically relevant in clipping McDonalds: speakers are certainly aware that M(a)c is a morpheme, and that there are quite a few names where it ends up with main word stress, e.g.
In those cases, as the spelling "Mac" suggests it should be, the vowel is [æ].
So even someone who perceives the unstressed vowel in McDonald's as closer to [ɪ] than to [æ] might still choose [æ] as the stressed version, based on how the M(a)c morpheme is pronounced when stressed. [And, as a commenter observed, there's the "Big Mac" to provide morpho-phonemic guidance!] Alternatively, someone who is aware of the morpho-phonology might still decide to go with (their perception of) the phonetics.
(There's some relevant additional discussion in "Symbols and signals in g-dropping", 3/23/2011.)