Aspects of the Acoustic Analysis of Imitation 381Related ¿ elds, however, have not ignored the utility of the study of imita-
tion. The relevance of imitation to applied linguistics has led to a variety of
studies of its role in second language acquisition, some suggesting that imi-
tation is a plausible learning and modeling behavior (e.g., Markham 1997),
and many focusing on the inability of adult learners to achieve native-like
pro¿ ciency. Research on imitation in speech and hearing science is also well-
known (e.g., Zetterholm 2002). In addition, the study of imitation in forensic
linguistics, such as speaker identi¿ cation, speaker pro¿ ling, voice line-ups
has been carried out for some time (see Hollien 2002). The study of imitative
behavior in child language acquisition (e.g., Richards 1986) and speech pathol-
ogy is also well known (e.g., Dillon et al. 2004; Snow 2001). Many of these
studies of imitation show us the effects of external speech alteration such as
that done by Molina de Figueiredo and de Souza Britto (1996), who showed
that acoustic alterations from disguising the voice by speaking with a pencil
in the mouth can create signi¿ cant differences in the quality of the vowels. In
addition, many of the imitation studies use auditory methods (e.g., Reich 1981;
Markham 1997, 1999; Masthoff 1996; Schlichting and Sullivan 1997).
Arguably, though, a very important (but certainly not the only) justi¿ ca-
tion for the study of imitation is the exploration of the À exibility and limita-
tions of the human language faculty, but the discussion here is limited to a few
of the small collection of acoustic studies of imitation in which internally dis-
guised or imitated speech was examined. For example, Endres, Bambach, and
Flösser (1971) explored the ability to both disguise and imitate. With regard
to disguise, respondents were invited to read a text with a disguise of their
choice three times. The mean formant frequencies of their /a:/, /i:/, and /n/ in
these recordings were compared to those of their normal speech. Results indi-
cated that “there is possibility of considerably changing the formant struc-
ture of vowels and vowel-like sounds as well as the mean pitch frequency by
deliberate disguise of the voice” (1847). In addition, Endres et al. explored
the ability of professional imitators to imitate other speakers by examining
existing tape recordings of imitations and the imitators’ normal speech. In
this case, by comparing the mean formant frequencies of /a/, /e/ and /i:/ they
found that “the imitator can change the formant positions of his voice within
certain limits” but do not reach an exact match in frequency position. Another
study that involved professional imitators was carried out by Zetterholm
(2002). She studied two professional and one semi-professional impersonator
and their ability to imitate famous people (all of whom had different dialects
from the impersonators). Auditory and acoustic methods were used. Results
showed “large” differences in the imitators’ own voices and imitations^1 , and
perceptual tests con¿ rmed the success of the imitators’ efforts. Zetterholm