Download Speech Perception Development in Infants: Phoneme and Prosodic Discrimination - Prof. Diog and more Study notes Linguistics in PDF only on Docsity! Nadim Van De Fliert - Gabriel Modderman 499A - Psycholinguistics - Diogo Jusczyk 1997 Chapter 4: How speech perception develops during the first year September 27, 2007 PHONE DISCRIMINATION - generalization - vowel sensitivity tends to occur before consonant sensitivity. - hypothesis - because vowels occur more frequently than consonants - hypothesis - because vowels carry more phonetic and prosodic information than consonants - observation - sucking rate of infants increased when listening to their native language (French), versus listening to a non-native language (Russian). - interpretation - infants are capable of recognizing phone and prosodic patterns matching their language, enabling them to ignore information they don't need - interpretation - infants show a preference for sound patterns heard in a prenatal environment, regardless of linguistic significance - secondary testing - infants whose native language (after birth) was French, but whose prenatal environment was not one surrounded by native French still showed more interest in passages of the French language than Russian, English, or Italian. - Perceptual Assimilation Model - (1) non-native phones that map to a single native phoneme are harder to distinguish, (2) non-native phones that map to different native phonemes are easier to distinguish, (3) non-native phones that do not map to any native phoneme are easier to distinguish - generally right, but Best (who proposed model) has found discrepancies, meaning more is at play. LANGUAGE DISCRIMINATION - Tools infants may use to identify native vs. non-native speech passages: - If it contains phones never found in the native language - If it contains phone clusters never found in the native language - NOTE: experiments have shown markedness is quantitativly determined by how frequent a native phenomenon occurs - If the prosodic or suprasegmental patterns do not match those of the native language . . . then it's not the native language - at 4 days French native infants can distinguish syllabic changes (French rhythm unit), but not mora changes (Japanese rhythm system). STRESS AND WORD RECOGNITION (tests on American babies) - stress seems to be independent of tone, and therefore can be present in tonal languages (maybe?) - English uses dynamic and quantitative accent for stress - English words tend to have predominant stress patterns (strong followed by week) - Norwegian uses pitch accent for stress - "Phonological Strategy" - Certain allophones, even phonetic features, occur only at word or morpheme boundaries, infants recognize these and learn to bind their perceptual words with them - Metrical Segmentation Strategy (hypothesis) - infant assumes a 1 to 1 relationship between strongly stressed syllables and words. - "Because the uterine wall acts as a low-pass filter, the information about speech that is most likely to be transmitted prenatally is that having to do with prosodic features." (pg 87) - Strong stress may be in any constant position, not just predominant, but penultimate, or final syllable. - Only predominant stress has been tested (English). Jusczyk admits to difficulties with this theory where the stress is not mainly in the first syllable. From English Speaking Environments - by 4 1/2 months - infants show recognition of their names (which are usually presented in isolation and have iambic stress) - by 6 months - infants pay more attention to English than Norwegian prosodic characteristics. - 6 to 9 months - infants pay more attention to predominant stress patterns, compared to other English words - by 7 1/2 months - infants recognize words with iambic stress in fluent speech acts, not just stressed monosyllables, but not trochaic stressed words - by 9 months - infants judge stress by more than just syllable weight - by 10 1/2 months - infants recognize trochaic
