On today’s NPR program, “Morning Edition,” the reporter Brendan Byrne kept saying [ashtronaut] instead of [astronaut], exemplifying the change in the speech of some speakers of American English that Y-H-B has written about in American Speech. Here is the article:


MISCELLANY A CASE OF DISTANT ASSIMILATION: /str/ -> /ftr/ lW ATHEN SOUND CHANGE INVOLVES ASSIMILATION, the typical case is one of contact assimilation: the sound that becomes similar to its neighbor is immediately contiguous to the latter. Assimilation at a distance does occur but is a relatively rarer phenomenon when it involves consonants; vowels assimilating to each other in neighboring syllables are quite com- mon, a typical example being that of umlaut (cf. Hock 1986, 64). The recent history of American English includes a sound change that seems to have gone unattested in the scholarly literature.’ This is the change of /s/ to /fl before /tr/ (i.e., a PHONEMIC CHANGE), which involves a palatalization of the initial sound in the cluster /str/, typically in initial position but not exclusively. Thus, for instance, speakers who regularly manifest this pronunciation replace the Standard American English [s] of strong, strategy, strength, Australia(n), restrictive, interest rate, industry, extra, and even history (when pronounced with syncope of the medial vowel) with [I]. The degree of palatalization is not uniform, so that the phonetic realiza- tion can stop short of the full-fledged “phonetic power” of the American [f] found in words like short, shape, ash, etc. (More about the phonetic details later.) This phonemic change seems to be neither dialectal nor regional.2 Over many months of listening to radio and television broadcasts and observing the pronunciation of speakers in the New York area, I have noted it as a regular trait in the speech of the following persons during their television appearances: Richard Nixon (miscellaneous sound bites); Howard P. (“Pete”) Colhoun (panelist on the PBS program Wall Street Week, 9 July 1993); Tracy Austin, Mary Carillo, John McEnroe (USA and CBS broad- casts of the US Open Tennis Championships, Aug.-Sept. 1993); Rick Barry, Hubie Brown (TNT broadcasts of NBA games, 1990-93 seasons); Dick Vitale (ESPN broadcasts of NCAA basketball games); and Cokie Roberts (regular panelist on ABC program This Week with David Brinkley) .3 Based on their overall speech and what can be ascertained about their origins, these speakers are from California (Nixon and Austin), Flushing, Queens, New York (Carillo and McEnroe), New Jersey (Barry, Brown, Calhoun, and Vitale), and Washington, DC, by way of Louisiana (Roberts)-which sug- gests no obvious geographical pattern. Admittedly, this is a highly limited sample, but I have deliberately singled out public figures whose pronuncia- tion is continuously open to observation by others who might wish to confirm for themselves the existence of this trait. I have also registered it among many other (nameless) speakers as a more or less regular phenom- 101

AMERICAN SPEECH 70.1 (1995) enon, and my southern correspondents in places like Birmingham, Ala- bama, have confirmed its incidence in that part of the country. Taking all this into account, I would venture to say that it is a general American innovation, and that it is gaining ground. Such are the facts, to the extent that I am able to present them. What makes this case more interesting than the mere registration of a phonetic peculiarity is its phonological significance. But in order to understand the innovation from the point of view of the sound pattern of American English, we need to back up one step and ask several questions. Is the pronunciation of strong, for example, with [I] instead of [s] properly an assimilation, let alone an assimilation “at a distance”? If so, what is being assimilated to what and in what phonological respect? Finally, in deciding these matters, are there special acoustic data concerning the phonetic realizations of /s/ and /t/ when these phonemes occur before Irl that need to be taken into account? The questions are intertwined; consequently, my discussion will have to do a bit of zigging and zagging between them. First some phonetic details. Judging by the evidence in Olive, Green- wood, and Coleman (1993, 279, 281), the cluster [str] seems to have a peculiar acoustic character. The center frequency of the frication noise of the /s/ moves down rapidly from the high value expected for /s/ (ca. 5 kHz) to a very low value (ca. 2 kHz or even less) right before the onset of voicing for /r/. And what is even more remarkable, there is often no abrupt cessation of the fricative noise (or none long enough to count as a stop), that is, it appears as if there really is no stop. Nevertheless, the uncharacter- istically gradual amplitude change in the noise is apparently enough to cue to the listener the presence of the stop /t/. The spectrogram in Olive, Greenwood, and Coleman (1993, 281) shows this-but without any com- mentary in the accompanying text that recognizes the oddness of the realization of /t/. This acoustic evidence suggests that the initial fricative-phonetically- could be a retroflex [f], just as the voiced fricative noted above (n3) is probably the retroflex [zj. The spectrogram in Olive, Greenwood, and Coleman (1993, 94, fig. 4.8), where the center portions of the voiced fricatives are shown, also makes it clear that retroflex [zj and the sound [3] are practically identical; this would presumably apply to their “voiceless” counterparts. In fact, judging by the spectrograms in Olive, Greenwood, and Coleman (1993, 173, fig. 6.26; 180, fig. 6.31), the retroflex realizations of/s/ are acoustically similar enough to be judged as fronted realizations of I/f. In my own auditory perception of the speakers I heard, I can testify that I consistently heard varieties of [f] and not retroflex [a]. More importantly, none of this disturbs the status of [9] as a realization of /Jf and [zj as a realization of /3/. 102