Page 20 - Fall 2011

linguistic biography, we might predict her speech would make use of both variants. Ms. Winfrey grew up in rural Mississippi, a region where /ɑ͡􏰀/ monopthongization is common. This is a typical feature of African American English-speaking speech communities as well. Given the racial division in the Southern US during her childhood, we can infer that Ms. Winfrey grew up in a speech community where this monophthongization was common. As an adult, Ms. Winfrey lives and interacts in speech communities where this sort of monophthongization is not common, and where most talkers use a diphthongal pro- nunciation. Hay and colleagues demonstrated that the way Ms. Winfrey pronounced this word varied as a function of both how frequently the word was used and the racial identity of her upcoming guest. Words were defined as high frequency if used five or more times in the corpus under study, and low frequency if used fewer than five times. The researchers found words used more frequently were more likely be pro- duced with the monophthong /tham/: 30% of the frequent words were monophongized, as opposed to only 14% of the infrequent words. The racial identity of the upcoming guest largely influenced Ms. Winfrey’s pronunciation of this vowel as well. Ms. Winfrey was three times more likely to use a monophthongal pronunciation when she was introducing or discussing an upcoming African American guest on her tele- vision program than when she was talking about a non- African American guest. We can interpret her behavior as an accommodation process where she uses a particular variable based on the predicted pronunciation patterns of her guests. This is a process of a sort of global speech style imitation or accommodation. The fact that lexical frequency influences the pattern is also important. The pronunciation variability is not wholly determined by social and interpersonal context; language internal factors, such as lexical frequency, also affect how sounds are produced. A second celebrity example of acoustic imitation takes us to another talk show: Larry King Live. On the Larry King Live show, Mr. King interviewed a range of guests, including celebrities, politicians, and others. Gregory and Webster characteristics of those around us, so let’s consider for a moment regional and dialect variation in speech production. Imagine modeling the pronunciation variants of the vowel /u/, as in the word food or dude, for which there is considerable variation in regional vernacular production. In Minnesota, this vowel is produced with the tongue in a very high and very back position; the lips are also considerably rounded. These articulatory movements cause this vowel to have a very low second resonant frequency. Now, picture to yourself a Californian surfer saying dude. In this stereotypical version, production is quite different from the Minnesotan dude. This type of /u/ pronunciation has an extremely high second reso- nant frequency. Minnesotans and Californians will not global- ly produce these vowels in exactly these caricatured ways; this increase in the second resonant frequency of /u/ is part of a sound-change-in-progress, leaving individuals in all commu- nities at different stages of the change. Talk show hosts imitating their guests The acquisition of regional variation can be simply con- sidered part of what we acquire based on what we hear around us. However, language use does not only vary according to region, and we also find systematic variation based around other macro-sociological categories like class and ethnicity. Our use of language does not reflect a monochromatic mir- roring of what we acquired as children, but rather a flexible matching process of sorts that is largely influenced by who we are speaking with or what we are talking about. Let’s take as an example the speech patterns of two celebrities. Consider first the pronunciation patterns of Ms. Oprah Winfrey from the Oprah Winfrey Show, which were analyzed in Hay et al. (1999). Under analysis was the degree of monophthongization of /ɑ͡􏰀/ in Ms. Winfrey’s speech; that is, did she pronounce a word like time as /thɑ͡􏰀m/ or /tham/? Examples of these two variants are shown in Fig. 3. The primary acoustic differences between these two variants relate to the first and second resonant fre- quencies of the vocal tract. Note, for example, the dynamic tra- jectories in /thɑ͡􏰀m/ on the left compared to the more stable res- onant frequencies in /tham/ on the right. Given Ms. Winfrey’s Fig. 3. Pronunciation variants of the word time. The variant on the left is the diphthongal /thɑ͡􏰀m/ variant and the token on the right is the monophthongal /tham/ variant. Imitation in Speech 19