| It has been widely shown that infants and adults are capable of using only prosodic information to discriminate between languages. However, it remains unclear which aspects of prosody, either rhythm or intonation, listeners attend to for language discrimination. Previous researchers have suggested that rhythm, the duration and timing of speech segments, plays an important role in linguistic processing and drives language discrimination, especially for infants during the first year of life. However, the experiments in this thesis were conducted in an attempt to show that infants and adults strongly attend to intonation, or pitch cues, and, in some cases, weigh this information more heavily than rhythmic information.;The experiments described in chapter 2 test American English-listening adults on their ability to discriminate their native language from a prosodically similar non-native language, German, and from a non-native dialect, Australian English. An acoustic analysis shows there are subtle rhythmic and intonational differences that listeners could potentially use to discriminate between both pairs. Then, American English listeners' ability to discriminate prosodically-similar languages is examined using (1) low-pass filtered and (2) re-synthesized speech, which preserves rhythmic timing and pitch, but degrades or eliminates segmental information, (3) monotone re-synthesized speech, which contains only rhythmic timing information, and (4) scrambled speech, which distorts both rhythmic timing and pitch information. Results show that listeners are capable of using both rhythmic timing and pitch to discriminate between the languages, but weight pitch cues over rhythmic timing cues, even when pitch is uninformative and hinders discrimination.;American English-learning infants' discrimination ability is tested in chapters 3 and 4, using the Headturn Preference Paradigm. In Chapter 3 a series of experiments on infants' discrimination of American English and German are presented. The ability to discriminate between these two languages is found to develop between 5- and 7-months of age. Like adults, 7-month-olds can discriminate between these two languages using only prosodic cues, as evidenced by their ability to discriminate between low-pass filtered speech stimuli. Finally, using stimuli with re-synthesized intonational contours, evidence is provided suggesting that, like adults, infants also weight intonation over other cues such as rhythm.;In Chapter 4 data on American English-learning infants' ability to discriminate of a number of additional language pairs, including American & Australian English, American English & Dutch, British English & Dutch, and American English & Japanese are presented. Infants failed to discriminate these pairs in all but two of the tested cases---five-month-olds were able to discriminate between British English & Dutch and American English & Japanese. However, the behavior observed in these two cases does not fit with previous reports on infant discrimination. The ability to discriminate British English & Dutch was retested using a different methodology, visual habituation, and more typical behavior was observed. The implications for infant testing methodologies are discussed. Finally, in chapter 5, the findings from the current study and concluding remarks are summarized. |
