L24 - Introduction to language development

Question 1

Why does language get special attention?

Answer 1

• Critical for thinking & problem solving, socializing, cultural transmission
  
  • Important in helping establish cognitive science
• Some have thought language is the critical difference between humans and other animals.
  
  • The structural difference between human and animal communication is proposed to either be or reveal the capacities that have enabled civilization, culture, etc.

Question 2

What are the different aspects of language?

Answer 2

Phonology = Physical signal or sensory signal

Question 3

What is prescriptivism vs descriptivism?

Answer 3

• Most people’s experience with “grammar” is being told they are not using it properly
  
  • “don’t end a sentence with a preposition” or “don’t use the passive”
• (Psycho)Linguistics is not concerned with “proper usage” in this sense
• Our concern is how people actually do represent, process, and use language, not how they should
  
  • Not that there is anything a priori wrong with improving one’s writing to meet professional standard’s or greater aesthetic value, it’s just not what this field of research is about.
  • Use more descriptivism - how people use language

Question 4

What is the problem of language learning?

Answer 4

• Infants must learn what sounds their language uses
  
  • R/L are two distinct sounds in English, not in Japanese
• Parse the continuous speech stream into words
  
  • We think there are gaps between words: there are not - need reliable cues or learn the words to understand when one word ends and another begins
• Learn the meanings of 20,000+ words
• Learn the rules of putting words together
  
  • In English, verbs tend to be in the middle of sentences, in Korean, at the end
• Learn the social conventions of language
  
  • Literal vs. figurative, polite vs. rude
• Learning a hierarchical system
• Language is componential and compositional
• At each level of description there is a finite set of units that get combined productively with a finite set of rules
• Speech sounds into syllables into words into sentences into discourses
• Not any two sounds can combine to make a syllable
  
  • St = ok, df = not ok
• Some combinations of words do not go together
  
  • Words combinations some of go not do
• Not any two sentences can go together to make a coherent discourse
• Because of these constraints = language is infinitely productive (they can be recombined in a multiple number of ways)

Question 5

What are infant vocalisations (how are they different in 2 months, 6 month and 12 month olds)?

Answer 5

• 2 month-olds coo:
  
  • Produce simple speech sounds (gooo, aaahh) and vocal gymnastics (smacks, clicks, bubbles)
  • Improved motor control of vocalizations
  • Imitate sounds of their partners, high pitched for Mom and lower for Dad - social learning aspect
  • Increase in vocal complexity
• 6 month-olds babble:
  
  • Repeated consonant vowels patterns that become more varied
  • By 10 months: native language specific babbling.
• 12 month-olds: first words

Question 6

What is newborn speech perception like?

Answer 6

• Recognize the prosody of their native language!
  
  • Prosody = pitch contours
• Prefer speech to non-speech
  
  • Vouloumanos & Werker, Dev Sci, 2004; 2007
    
    • High Amplitude Sucking Procedure
    • Alternating minutes, speech & non-speech
    • More HA sucks to speech than to non-speech auditory stimuli
      
      • Already heard language through the womb

Question 7

What is infant speech perception like (e.g. in categorical perception)?

Answer 7

• Acoustic properties of speech sounds vary
  
  • Some variance is important, some not - some are meaningful
• Not all languages use the same contrastive sounds
• Child’s task: Figure out what sounds their native language uses contrastively.
• Categorical perception of phonemes: some variance generalize across, others draw boundaries
• Across-category discrimination is easy to distinguish (between sounds)
• Within-category discrimination is harder
• Categorical perception - Eimas, Siqueland, Jusczyk & Vigorito
  
  • Measured by habituation tests
  • Tested category discrimination in infants = Showed organization from the first days of life

Question 8

What changes in infant speech perception?

Answer 8

• Infants show categorical perception across the world’s language sounds in first half of 1st year
• By 10 months of age, infants are no longer discriminating phonemic contrasts irrelevant to their language
  
  • e.g., Japanese infants no longer discriminate between R and L
• Why is this adaptive?
  
  • Every time you hear a sound, it has slightly different qualities (e.g., across speakers)
  • Learners need to learn which contrasts are functionally important, i.e., signify different words

Question 9

What is infant speech segmentation?

Answer 9

• Will discuss further in oncoming lectures but:
  
  • Host of properties
    
    • e.g., stress patterns
    • Any utterance with only one stressed syllable (Man; Dusty; Spaghetti) is a single word, no matter how long it is. - have tone contours in every word - can start to learn how to parse the speech stream

Question 10

How does word learning occur?

Answer 10

• 1st words at 12 months
  
  • Vocab explosion (approx 2-6 years)
    
    • Children’s vocab grows from 100-200 words at age 2 to 5000 - 20,000 words at age 7

Question 11

When is the beginning of syntax?

Answer 11

• Prior to 18 months typically just say one word at a time
• 18 months: 2 word combinations
  
  • The two words chosen are not random (e.g., “Daddy went to work” becomes “Daddy work”) - respecting the order/syntax of their native language
    
    • Some kind of bottle neck limiting the amount of words they can get out
  • Child maintains most important content words and word order
  • Grammar ability correlates with size of vocabulary

Question 12

What is morpho-syntax?

Answer 12

• Morphemes: meaningful units of language
• Some words have just 1: sport, man, bag
• Others are compounds of single morpheme words: tugboat
• Syntactic morphemes have syntactic function and add meaning
• Inflectional morpheme: changes in number and tense
  
  • e.g., plural “s”, past tense “ed”
• Derivational morpheme: change grammatical category,
  
  • e.g., -tion turning verb into a noun “destruction”

Question 13

How do you test they have actually learnt the understanding rather than copying of morphemes?

Answer 13

• 2-3 children show productive mastery of basic inflectional morphology as shown by applying them to novel words (e.g. wug test - see if they can apply the rule to new word)
• Further, overgeneralization:
  
  • tooths, mouses
• Often kids imitate irregulars
• Then learn the rules and overextend
• Then re-learn irregulars
  
  • We’re going to focus on syntax more later, but from 2 – 5 the complexity of syntactic production rapidly increases

Question 14

What are pragmatics and how do they develop?

Answer 14

• Sophisticated pragmatics develops through middle childhood along with general social cognitive (e.g., advanced ToM abilities) and metacognitive development (ability to reflect upon language use)
• Adjusting kinds of language for different contexts fluidly
• Metaphor, irony, sarcasm by age 8
• Vocabularies and pragmatic skills just keep on growing and developing

Question 15

What happens if you don’t have access to language?

Answer 15

• Many deaf children are born to speaking parents: what happens?
  
  • If they do not have interaction with any sign language they invent simple gestural systems to communicate some basic stuff: “home-sign”
  • If not exposed to sign language by puberty, they most likely will never get beyond home-sign
• If parents learn to sign, then children learn it from them better then the parents know it themselves!
  
  • Singleton & Newport (2004), case study
    
    • Hearing parents would use correct inflectional and derivational morphemes ~70%, the rest a mix of errors
    • 7 year-old child of deaf parents used correct morpheme >80%, same as sample of native signers
      
      • If deaf children have native signing parents, there are no language delays compared to speaking parents.
    • Children found the signal in the noise and exploited it so could filter out the errors
• Hudson Kam & Newport (2005) show this pattern with large sample of hearing children learning stochastic artificial grammar
  
  • Artificial grammar - hear nonsense syllables & words and there are rules within the fake language
  • Adults reproduce variability of input grammar, 5-7 year-old children find the most common linguistic form and produce that even more. What if you don’t have access to language?