The Pitch Range of Italians and Americans. A Comparative Study

Linguistic experiments have investigated the nature of F0 span and level in cross-linguistic comparisons. However, only few studies have focused on the elaboration of a general-agreed methodology that may provide a unifying approach to the analysis of pitch range (Ladd, 1996; Patterson and Ladd, 1999; Daly and Warren, 2001; Bishop and Keating, 2010; Mennen et al. 2012). Pitch variation is used in different languages to convey different linguistic and paralinguistic meanings that may range from the expression of sentence modality to the marking of emotional and attitudinal nuances (Grice and Baumann, 2007). A number of factors have to be taken into consideration when determining the existence of measurable and reliable differences in pitch values. Daly and Warren (2001) demonstrated the importance of some independent variables such as language, age, body size, speaker sex (female vs. male), socio-cultural background, regional accents, speech task (read sentences vs. spontaneous dialogues), sentence type (questions vs. statements) and measure scales (Hertz, semitones, ERB etc.). Coherently with the model proposed by Mennen et al. (2012), my analysis of pitch range is based on the investigation of LTD (long-term distributional) and linguistic measures. LTD measures deal with the F0 distribution within a speaker’s contour (e.g. F0 minimum, F0 maximum, F0 mean, F0 median, standard deviation, F0 span) while linguistic measures are linked to specific targets within the contour, such as peaks and valleys (e.g. high and low landmarks) and preserve the temporal sequences of pitch contours. This investigation analyzed the characteristics of pitch range production and perception in English sentences uttered by Americans and Italians. Four experiments were conducted to examine different phenomena: i) the contrast between measures of F0 level and span in utterances produced by Americans and Italians (experiments 1-2); ii) the contrast between the pitch range produced by males and females in L1 and L2 (experiment 1); iii) the F0 patterns in different sentence types, that is, yes-no questions, wh-questions, and exclamations (experiment 2); iv) listeners’ evaluations of pitch span in terms of ±interesting, ±excited, ±credible, ±friendly ratings of different sentence types (experiments 3-4); v) the correlation between pitch span of the sentences and the evaluations given by American and Italian listeners (experiment 3); vi) the listeners’ evaluations of pitch span values in manipulated stimuli, whose F0 span was re-synthesized under three conditions: narrow span, original span, and wide span (experiment 4); vii) the different evaluations given to the sentences by male and female listeners. The results of this investigation supported the following generalizations. First, pitch span more than level was found to be a cue for non-nativeness, because L2 speakers of English used a narrower span, compared to the native norm. What is more, the experimental data in the production studies indicated that the mode of sentences was better captured by F0 span than level. Second, the Italian learners of English were influenced by their L1 and transferred L1 pitch range variation into their L2. The English sentences produced by the Italians had overall higher pitch levels and narrower pitch span than those produced by the Americans. In addition, the Italians used overall higher pitch levels when speaking Italian and lower levels when speaking English. Conversely, their pitch span was generally higher in English and lower in Italian. When comparing productions in English, the Italian females used higher F0 levels than the American females; vice versa, the Italian males showed slightly lower F0 levels than the American males. Third, there was a systematic relation between pitch span values and the listeners’ evaluations of the sentences. The two groups of listeners (the Americans and the Italians) rated the stimuli with larger pitch span as more interesting, exciting and credible than the stimuli with narrower pitch span. Thus, the listeners relied on the perceived pitch span to differentiate among the stimuli. Fourth, both the American and the Italian speakers were considered more friendly when the pitch span of their sentences was widened (wide span manipulation) and less friendly when the pitch span was narrowed (narrow span manipulation). This happened in all the stimuli regardless of the native language of the speakers (American vs. Italian).