Language Acquisition

Language Acquisition

1. Adults struggle to segment speech stream of unfamiliar language into words

   • Adults can learn to segment the sounds into words through more input, repetition and practice
     • "Mondegreens"
       • Misinterpreting lyrics
       • Mishearing/misinterpretation of a phrase into a new meaning
     • If adults struggle so much, how do infants learn to segment speech sounds into words when they acquire their L1?
       • What abilities or knowledge do infants have to be able to extract words out of speech stream?
         • Could it be an innate ability?
         • Saffran et al. 1996 study about whether infants segment speech stream into words and how
           • Method used: Head-turn preference paradigm
             • Linking hypothesis: infants are interested in novel words
           • Results: infants were more interested in "part-words" than in the words that they had heard previously in the speech stream, showing that infants can segment speech stream into words
             • How is this possible?
               • Infants are statistically analyzing the transitional probabilities
                 • What is the probability that certain syllables follow each other?
                   • Even monkeys and rats have shown to use statistical probabilities in experiments
                   • Could this also apply to other areas of life? Musical tones, shapes
                 • TRANSITIONAL PROBABILITIES
                   • Ability to statistically analyze the probability of which constituents group together (e.g. syllables into words)
       • Mattys and Jusczyk (2001) study of whether infants segment speech stream into words, or into reccuring closely related patterns
         • First: familiarize the infants with a speech stream
           • Secondly: Run a sequence of words that were in the speech stream, and run a sequence of words that were NOT in the speech stream, but are very similar
             • Method used: Head-turn preference paradigm
               • Linking hypothesis: infant will look longer in the direction of the speaker if they are interested, and they are interested in familiar words
                 • Do they prefer words that they have heard before or novel words?
             • Shows if Infants segment the speech stream into words or not
             • Results: infants were more interested in familiar words that they had previously heard in the familiarization fase, and could thereby distinguish between real words and non-words
               • Inftants segment speech stream into words. They do this based on transitional probabilities
     • Holorimes
       • Two very similar sequences of sounds can compose phrases that have two very different meanings
         • Works especially well in French

2. How well can infants pick up phonemic variations?

   • Phonemic sounds exist in a continuum, and not as a phonemic category (either /b/ or /p/)
     • Eimas et. al. study testing how well infants can distinguish between fine-grained differences in stop consonants, using VOT
       • Test in terms of habituation: is this something new or something familiar?
         • Result: Infants develop categorical perception very early because they tune in to the patterns of their native language
           • Method used: High Amplitude Sucking
             • Linking hypothesis: more sucking = more interested (in novel sounds)
     • Werker 1989 article about early phonemic contrast
       • Infants are able to distinguish between many more phonemic contrasts than adults (and even 1 year olds)
         • This is because infants quickly tune in to the language that is relevant to them, and because of this, their ability to perceive more fine-grained distinctions disappear
           • Other studies have tested place of articulation, manner, aspiration, and lip-rounding and concluded the same
           • This loss in categorical distinction works against us when learning an L2
             • But useful in terms of our L1 to use as a basis for learning about other structures of our language
   • Adults have a categorical perception of phonemes and distinguish them as either/or and categorize phonemes differently depending on what language they speak.
     • CATEGORICAL PERCEPTION
       • Perceiving non-categorical stimuli - here "phonemes" as "either/or" when exposed to something in between /t/ and /k/ or /s/ and /z/ etc.
       • Voice Onset Time (VOC) is the time it takes for the vocal chords to vibrate. Changes in VOT changes our perception of what phoneme is produced