Two people (nomina sunt odiosa)––one the female host of a morning news program, the other the network’s Pentagon correspondent––are talking to each other on the radio about the war in Iraq, and each consistently pronounces the name of the country differently: one says Ir[á]q with what is called a “broad A” (rhymes with rock) in the linguistic literature, the other says Ir[ǽ]q with what is called a “flat A” (rhymes with rack). Then the topic switches to Iran, and the same difference in their rendering of the stressed vowels perdures. But they both say Pakistan and Afghanistan with flat vowels in the appropriate syllables throughout. When the male co-host jumps into the conversation, the same distribution of variants applies to his speech: he has the broad vowel in both Iran and Iraq but the narrow vowel in the two ‘Stans.
Then a clip is played of a recorded interview with an Army captain in Iraq, who consistently uses the flat vowel [æ] in his pronunciation of the two countries’ names. This interview is followed by one with an enlisted man, whose stressed vowel in the two nomina propria (proper nouns) conforms to that of the officer but differs in the value of the initial (unstressed) vowel, which he pronounces with the diphthong [ai] (rhymes with eye) and secondary stress, viz. [àirǽk]––just as he might the first vowel of Italian, both blatantly down-market, non-standard pronunciations. The interviewer in both cases is the network’s (female) Baghdad correspondent. She consistently—whether interviewing or just reporting from Iraq—maintains the pronunciation with a flat vowel, i.e. Ir[ǽ]q.
Then an excerpt from the governor of Alaska’s speech to the 2008 Republican National Convention is broadcast, and she too says what the enlisted man said, namely [àirǽk], with a diphthong in the first syllable and a flat A in the second syllable.
What’s going on? Why this variation among native speakers of American English in the rendering of the (stressed) sound A?
Before essaying an answer, one needs to keep in mind the following salient external facts about the dramatis personae. (1) The two co-hosts of the program (one based in Washington, the other in Los Angeles) are not in regular contact with military personnel––unlike both the Pentagon and the Baghdad correspondents; (2) the Army personnel are members of that social group by definition, differing only in rank and (probably) education; (3) the Alaskan governor is also the commander of that state’s National Guard and even has a son who is a member of that unit.
Now for some general data about this variation.
Vacillation between [a] (“broad” A)’ and [æ] (“flat” A) is a persistent feature of American speech, particularly in loan words or nomina propria (proper nouns), as in the twofold pronunciation of the stressed vowel of Colorado, Nevada, Iran, Iraq, Milan, and so on. Whereas no true Westerner would be caught dead saying Color[á]do or Nev[á]da, many of them, along with other Americans, do habitually say Ir[á]n, Ir[á]q, and Mil[á]n, instead of the long- standing and traditional Ir[ǽ]n, Ir[ǽ]q, and Mil[ǽ]n. In the case of loan words, including designations of foreign places or things, even where initially there is vacillation between [ɑ] and [æ], as in Viet Nam (cf. the preference for [nǽm] over [nám] to render the slangy [originally military!] abbreviation ‘Nam), American speech in modern times seems to favor pronunciations that speakers likely construe as approximating the donor/original language’s sounds, especially in the case of a smattering of knowledge of foreign, mostly European, languages. In this respect, American speech has tended to diverge from traditional British English––and the older American tradition––where anglicization has long been the norm (cf., for instance, the different rendering of names like Kant or Dante; or of words like pasta and mafia). Viewed from this perspective, pronunciations like Ir[á]n simply conform to a current tendency.
Recently, however, there has been a marked augmentation of the domain affected by the tendency––specifically, to include unfamiliar words, whether or not a particular word is ascertainably foreign and “known” to a speaker as such. In this new situation, the emphasis falls on unfamiliarity: the word in question is either not part of a speaker’s active vocabulary or is used sporadically. It may have been acquired from other speakers who are equally unfamiliar with it. In such cases, the pronunciation is likely to be at variance with the common or traditional pronunciation. Take the recently manifested vacillation in the stressed vowel of the journalistic buzz word (a Sanskrit borrowing), mantra. The foreign provenience of this word is clearly irrelevant as far as these speakers are concerned. Its new transferred meaning––that is, anything repeated as a set piece, especially a political slogan, the dictionary meaning being a type of prayer––is the sense journalists who have the broad vowel have evidently assimilated and foregrounded. But the traditional pronunciation m[ǽ]ntra is either unknown or eschewed. I propose to explain this appearance of [ɑ] for [æ] as deriving from insecure knowledge of the word as such, not its meaning.
This analysis is confirmed indirectly by cases where unfamiliarity cannot be invoked as the reason for [ɑ], but markedness could be. In a broadcast some years ago of his commentary, “The Nature of Things” (Vermont Public Radio), the naturalist Will Curtis several times pronounced the word habitat with [ɑ] for both of the relevant (stressed and unstressed) vowels. This untraditional pronunciation of a word in common use can be chalked up to its valorization as marked in the sense of “special” or “restricted.” When a speaker accords salience or special status to a word that contains a vowel that can be rendered [ɑ] or [æ], [ɑ] may be utilized as a means of mirroring the marked value of the word in context. Curtis (whose topic was the disappearance of habitat for certain flora and fauna) evidently––and unconsciously––did this with habitat.
This analysis joins hands with the earlier one, in that “unfamiliarity” is one of the concrete meanings of the abstract designation “marked.” The foreignness of words lends itself typically to subsumption under the category of marked value, hence the special or restricted phonetic features commonly found in the pronunciation of foreign words unless and until they are nativized (if ever). This is especially true of names. Thus Yasser Arafat, while he was alive and his name constantly being gibbered in the media, was pronounced with some combination of [ɑ]’s and [æ]’s, although the thoroughly anglicized version––all [æ]’s––is also extant. I once heard a speaker wishing to dignify his ownership of the very expensive car called a Lamborghini pronouncing the first vowel [ɑ] instead of [æ]. The vowel [ɑ], through its occurrence in what is perceived as American “educated” speech in words like rather, as well as in British English (tomato, banana), has become associated with marked (= foreign, formal, “high” style) pronunciation, whence its natural utilization as a phonetic mark of special status.
Imitation of prestige dialects is likely to account for examples like the garden-variety word pistachio or the name Andrea being pronounced with [ɑ] rather than the plebeian [æ]. (The recent appearance of the spelling Ondrea to render the name bears this out. )
Now––finally!––we come to an explanation of the strange distribution of broad and flat A that gave rise to this discussion.
The persons whose speech on the radio served as the source of data about the variation in the stressed vowel of Iraq and Iran break up into: (A) those who are familiar with the traditional (i.e. local, in situ) American English pronunciation by virtue of their contact with military personnel and those close to that speech community; and (B) those who (unconsciously?) think that the correct pronunciation should approximate what they take to be the vowel of the source language––here Arabic and Farsi, respectively. The first group follows the older norm, the second the emerging one. The same would apply mutatis mutandis to speakers who have the broad vowel in the relevant syllables of Afghanistan and Pakistan.
All these data speak in favor of the idea that the historically older urge of Americans to render foreign (European) words “correctly” at the expense of native phonetic norms has been subsumed, as but one specific manifestation, under the newer and more general drive for “authenticity.” Truth is (mis-?)identified with the authentic. Thus, K[á]nt and D[ɑ]nte persist as the only pronunciations in American speech (where the British norm has K[æ]nt and D[æ]nte) not because of a desire to acknowledge the foreignness of the names but because nativizing their pronunciation might run the risk of making one’s acquaintance with them seem less than authentic. Hence it is the avoidance of anything that, through speech, might be taken as a sign of inauthentic knowledge that seems to explain not only the proliferation of Ir[á]q and Ir[á]n but pronunciations like m[á]ntra, pist[á]chio and even h[á]bit[à]t as well.