When does hearing begin?
- weeks 23-28 in pregnancy = no response in fetal HR to an external sound meaning they can’t hear it
- around 29 weeks in pregnancy there is an obvious response
Speech processing before birth, De Casper & Spence (1986)
- showed that babies are actively processing speech before birth.
- this is called “transnatal learning”
- 12 pregnant women read a passage from The Cat in the Hat 2x per day for last 6 weeks of pregnancy
- Story chosen because it has a very regular rhyme
- 2/3 days after birth babies were tested for recognition using a sucking response
- Sucking played either a recording of The Cat in the Hat or another (unfamiliar) passage
- Babies altered sucking pattern to hear the familiar passage but not the unfamiliar one
- Change in sucking occurred irrespective of whether
Telling languages apart, Christophe and Morton (1998)
- Presented 2-month-old English babies with two different language comparisons
- English vs Japanese (different rhythmical - or prosodic - pattern)
- English vs Dutch (more similar in prosody)
- Babies could tell the difference between English and Japanese but not English and Dutch
- Babies can use prosody to distinguish languages
what is a phoneme?
the smallest sound unit that carry distinctions between one meaning and another e.g. /b/ and /p/
what are phoneme boundaries?
where a physical parameter, such as voice onset time, changes perception from one phoneme /b/ to another /p/
what is it crucial to tell different phonemes apart?
It’s crucial to perceive different variants of the same phoneme as the same (i.e. to perceive all instances of /p/ sounds as /p/)
Infants’ early discrimination of speech sounds, Eimas et al (1971)
- pioneered the High Amplitude Sucking (HAS) paradigm to test infants’ discrimination of speech sounds
- Phase 1: babies aged from 1 to 4 months presented with the single sound /ba/
- Babies increased their rate of sucking then their sucking rate settled back to the baseline.
- Phase 2: Once the babies had habituated to /b/ a new but similar sound was played.
- Half the babies heard a different phoneme /p/
- Half heard a variant of /b/
RESULTS
- Babies who heard /p/ increased sucking rate
- Babies who heard /b/ did not increase sucking rate
who can make better phonetic discriminations, adults or babies?
Babies. This is because a baby can be born into any country speaking any language
Perceptual narrowing of speech in infancy
- Infants are initially universal language perceivers
- Language-specific experience provides infants with continuing exposure to native contrasts and no exposure to non-native ones
- System becomes fine-tuned to relevant contrasts
- Become specialists in the phonology, semantics and grammar of their native tongue.
Developmental changes in language acquisition, Werker & Tees (1984)
- Compared babies from language communities where the phonemes differed
- English and Hindi
- Wanted to look at whether babies can discriminate between Hindi /Da/ vs. /da/?
- Infants are trained to look at a toy when there is a change in sound
- Then the target items are played, and an observed (who cannot hear the sounds) judges whether the infant heard a stimulus change based on the infant’s actions.
Experience of foreign language can reverse decline in non-native speech perception (Kuhl et al., 2003)
- a group of english speaking babies and Chinese speaking babies
- Chinese speaking babies are better at distinguishing between Chinese sounds than English babies who have only been exposed to English
- Found that English babies who have been exposed to Chinese sounds performed just as well as Chinese babies
Diversity and representation in studies of infant perceptual narrowing of speech, Singh et al. (2022)
- Darker shading represents greater relative density of study populations sampled in that region.
- Language background of monolingual participants.
-This depicts the proportion of monolingual infants’ language background in prior studies on perceptual narrowing.
Infants’ ability to extract words from fluent speech, Jusczyk & Aslin (1995)
- Familiarised 7.5 month infants to repetitions of sentences.
- Containing two target words.
- Then, tested on target and novel individual words using the preferential listening (head turn preference) paradigm.
- By 7.5 months, infants have at least some rudimentary ability to detect words when they occur in fluent speech contexts
what is the preferential listening paradigm?
- Infants sit on their caregiver’s lap in a test booth.
- On each trial, one of the side lights flash, and when the infant orients to the light, sounds come from that speaker.
- The experimenter records how long the infants looks at the “source” of the sound (the flashing light) as a measure of the infant’s preference.
How can babies tell where one word ends and the next begins?
- They are helped by Infant directed speech (ID)
- Implicit discovery in the language input: using prosodic cures, attend to transitional probabilities
How can syllable stress affect how babies tell when words begin and end?
- In English 90% words have the stress on the first syllable:
strong-weak stress pattern - babies could use presence of stressed syllable as a guide to the beginning of a word
Infants’ sensitivity to syllable stress, Jusczyk, Cutler & Redanz (1993)
- Compared strong-weak with weak-strong words
- 6- and 9-month English infants
- At 6 months showed no preference but by 9 months spent longer listening to strong-weak lists
Statistical learning: Transitional probabilities (TP)
- TP = probability of one syllable following another
- Certain sequences of syllables will occur more often than others
- More commonly occurring sequences are likely to be words
e.g. pretty baby: Saffran et al. (1996); Johnson and Jusczyk (2001)
- can be spilt into pre + ty,
ty + ba , ba + by
- Transitional probability that ty will be followed by ba is lower than either of the two word-internal transitional probabilities (i.e. pre followed by ty or ba followed by by)
Statistical learning: Transitional probabilities cont. Saffran et al. (1996); Johnson and Jusczyk (2001)
- Investigated the use of transitional probabilities in the detection of word boundaries by 8-month-old infants
- Invented ‘words’ by taking 12 syllables and combining them into four sequences to make
Phase 1 (familiarization phase ) - 8 month old infants heard the ‘words’ repeated over and over in random order with no pauses between them
- In all, they listened to the syllables for 2 minutes
- In this way, transitional probabilities could be used to distinguish ‘words’ from other syllable sequences
Phase 2 (test phase ) - Infants were presented with the 4 ‘words’
- They also heard part-words where syllables from two words were recombined e,g, tudaro, was formed golatu + daropi
- Attention to words and part-words was measured by orientation to loud speaker
- they have a novelty preference
implications
- At 8 months, infants were able to segment a continuous stream of speech based on statistical cues alone
- Suggests infants have a powerful mechanism for the computation of statistical regularities in the language input
- Bias to selectively attend to certain properties in the acoustic signal
What is a word?
In comprehension
- Consistent and specific response to the use of a word
- Must be response to word itself rather than nonverbal cues
In production
- Consistent use of a sound sequence in a consistent and specific context
what are the 3 ways to measure comprehension?
- Parental reports - Communicative Development Inventory (CDI)
- Home observations/video recordings
- In lab, ask infants to choose named object from an array (e.g. Golnikoff et al., 1987) or use preferential looking paradigm
Early word comprehension for socially salient words, Tincoff & Jusczyk, (1999)
- 6-month old infants
- Hear recordings of a voice saying ‘mummy’ or ‘daddy’ while viewing two monitors, one showing video of their mother, one of their father
- Infants looked more at the video matching the word heard
Early word comprehension, Bergelson & Swingley, 2012; 2015)
- By 6-9 months infants know the meaning of many common nouns such as food and body parts
- Demonstrate greater looking toward object in picture that matches the word in the phrase: “Look at the X”
- parents underestimated the words their baby would know
