Developmental Lecture 1-5 PDF

Developmental -- Lecture 1 to 5. **LECTURE 1:** **Language basics (1)** - Language is generative. - Language is comprised of small units that are combined (phonology) - Language conveys meaning (semantics) Language Basics (2) Languages have rules about how words go together (syntax)...

Developmental -- Lecture 1 to 5. **LECTURE 1:** **Language basics (1)** - Language is generative. - Language is comprised of small units that are combined (phonology) - Language conveys meaning (semantics) Language Basics (2) Languages have rules about how words go together (syntax) - A bites the dog man. - The dog bites a man. - The man bites a dog. - Subject-verb-object in English Language is social. **[Infant designs ]** **Preference studies** - With no training, what do infants want to listen (or look) to. **Habituation/familiarisation studies** - First, we train infants and then measure what they prefer. **Change detection studies** - We train infants to respond to a change (can infants tell the difference between two things). **What sounds are in my language?** Prosody - The pattern of stress and intonation in a language. - Languages have different prosodic patterns Phonemes - The perceptually distinct units of sound in a language that distinguish one word from another (e.g., p, b, d, t) - Pat, bat, bad, pad - Languages differ in the sounds that they use as phonemes. **Early Phonological Development: Prosody (2)** The foetal auditory system is fully functioning during the last trimester. Newborns: - prefer their own mother's voice (De Caspar & Fifer, 1980). - discriminate languages with different prosody (German/Spanish) but not languages of similar prosody (English/Dutch) (Nazzi et al., 1998) - prefer their native language compared to a foreign language (Moon et al., 1993). - cry with an "accent" (Mampe et al., 2009). **Early Phonological Development: Phonemes (1)** - Phonemes are the sounds that distinguish words (e.g., pat/bat, sip/zip). - Other languages have phoneme contrasts that are not in English. - Across the world's, there are about 600 consonants and 200 vowels. But any language uses about 40. - Children's babble -- initially wide range of sounds. In first year move towards producing only sounds of target language (Levitt and Wang, 1991). **Early Phonological Development: Phonemes (3)** - At 1-2months, infants can discriminate between all sounds, even foreign ones. Adults only discriminate those in their language (Eimas et al, 1979; Miyawaki et al., 1975). - Between 7-11 months, systematic decline in ability to distinguish sounds from nontarget language and increase for target language. (Kuhl et al, 2006) **Early Phonological Development: Phonemes (2)** **Finding the words** - Infants can segment words from their language at \~ 7.5 months, but not 6 months (Jusczyk & Aslin, 1995). **Statistics (1)** - Infants are surprisingly sophisticated statisticians. - Track the co-occurrence of syllables. - Syllables that co-occur often are likely part of the same word. - H-A-P is always followed by P-P-Y - P-R-E is always followed by T-T-Y - B\--A is always followed by B-Y - D-O is always followed by G-G-Y **Statistics (2)** Experiment with 8-month-olds using highly controlled made-up language (Saffran et al., 1996). - A few minutes of listening to words in a randomised order. - Infants listen longer to part words, suggesting they found the words in the stream. **Finding the words: Prosody (1)** **Infant-directed speech (IDS) (Christia, 2013).** - Higher pitched - Slower speaking rate. - Important words are generally at the end and are exaggerated more. - The boundaries between phrases are enhanced, making it easier to segment speech. - Infants prefer to listen to IDS and interact with people who use IDS. - More attentive around IDS **Finding the words: Prosody (2)** - Infants segment speech better with IDS than adult-directed speech (ADS) (Theissen, Hill & Saffran, 2005). **Finding the words: Frequency (1)** - Highly frequent salient words (e.g., Mummy, child's name) - Highly frequent linguistic words (e.g., the, he/she) - These words act as an "anchor". - If you can identify a word in the speech stream you can identify one boundary of the adjacent words **Finding the words: Frequency (2)** - Highly familiar words (own name, "Mommy") help 6-month-olds segment words (Bortfeld et al., 2005) - Baby Maggie recognized words next to the name \"Maggie\" and baby Hanna recognized words next to the name \"Hanna\" - Recall that 6-month-olds fail in the Jusczyk & Aslin (1995) study. **Finding the words: Frequency (3)** - Some linguistic categories of words (e.g., articles the, a, his, hers, conjunctions and, or) are highly frequent. - Infants can use "the" to segment nouns at 8 months (Shi & Lepage, 2008) - At test, infants listened longer to an isolated word that was taught with a real function word. **How are words organised?** - The dog chased the squirrel in the park. - The squirrel chased the dog in the park. - Chased in squirrel the park in dog the. - Order matters! **Finding the patterns: Frequency (1)** Many highly frequent words are function words. Grammatical function How would you describe this stream to someone else? - Su is beginning - Su is in the middle - Su is at the end **Finding the patterns: Frequency (2)** In English, function words tend to go to before (articles, pronouns, prepositions): - An apple - The dog - You ran - They swam - On the table - Under the chair But not in all languages In order to learn syntax, infants need to learn the word order. **Finding the patterns: Frequency (3)** - In Japanese, the order is switched - Articles are after the noun, postposition rather than prepositions. - Infants are sensitive to this by 8 months (Gervain et al., 2008) - Italian is a frequent-first language; Japanese is a frequent-final language. **Finding the patterns: Frequency (4)** - Italian 8-month-olds listen longer to frequent-first - Japanese 8-month-olds listen longer to frequent-final - By 8 months, infants have started to learn some of the ordering rules for their language. **Finding the patterns: Rule-learning (1)** Syntax requires learning the abstract rules of a language. But can infants learn abstract rules? - The red car (grammatically correct) - \*The car red (grammatically incorrect) - The boy is jumping (grammatically correct) - \*The boy is jump (grammatically incorrect) - The girl eating cake is happy (grammatically correct) - The girls eating cake are happy (grammatically correct) - \*The girls eating cake is happy (grammatically incorrect) - \*The girl eating cake are happy(grammatically incorrect) Finding the patterns: Rule-learning (2) Syntax requires learning the abstract rules of a language. But can infants learn abstract rules? - 6-month-olds could learn an abstract rule with linguistic stimuli (Marcus et al., 1999) - Those familiarized to ABA pattern listened longer to ABB - Those familiarized to ABB pattern listened longer to ABA **Conclusions and Issues** Infants make rapid strides in language acquisition over the first year. Identifying sounds, statistics, patterns, word boundaries Early preferences and skills for starting to crack language complexities. Infants tune the specifics of their language (sounds, order) well before they begin to speak. LECTURE 2: - Two main stages of early social skills (primary and secondary intersubjectivity). - Two key modes of communication are important for language and acquisition (turn-taking and joint attention). Early socialisation (1): - Pre-linguistic communication is more than just crying. - Primary intersubjectivity (Trevarthen, 1979): First months: attention to faces, eye contact, produce vocalisations, imitate sounds and gesture. - Secondary intersubjectivity : Older infants: more sophisticated, pointing, turntaking, shared attention. **Early socialization: Primary intersubjectivity:** - First months: attention to faces, eye contact, produce vocalisations, imitate sounds and facial gestures, one-at-a-time interactions. - Caregiver and infant share experiences in face-to-face interactions. But these interactions are dyadic (baby and caregiver, baby and object). - No assumption of the perspective of others. - These interactions are not intentional. **Motivation -- dyadic mimicry:** - Infants imitate (Kuhl, & Meltzoff, 1996; Meltzoff & Moore, 1977). - Newborns mimic facial expressions. - 3-4-month-olds imitate sounds. - Limited form of imitation. - No understanding of others' intentions. - But shows that infants motivated to engage with others. **Preference for Faces** - From birth, infants prefer to look at things that are "facelike" Goren et al.. (1975). **Early Socialisation: Attention to faces and eye gaze** - Newborns prefer to look at direct (compared to averted) gaze (Farroni et al., 2002). - Senju and Csibra (2008) = 6mo infants only follow the gaze to the object if preceded by mutual eye gaze. Same results found for IDS (but not ADS). The communicative signal (eye gaze/IDS) encourages infants attend to the same object. **Early Socialisation: Secondary Intersubjectivity** - Older infants (from around 9 months): more sophisticated, pointing, turn-taking, joint attention. - Caregiver and infant share experiences and these interactions start to become triadic (the infant and caregiver interacting together with a toy; social referencing). - The interactions become intentional, and infants start to assume that others have their own perspective. **Secondary Intersubjectivity (2)** - Coordinate emotional response with another person. - Still Face Experiment (Adamson & Frick, 2003). - Parent \"freezes\" and stops responding - The interaction breaks down - Attempts to repair the interaction (social engagement cues) **Secondary Intersubjectivity (3)** Coordinate emotional response with another person. Still Face Experiment (Adamson & Frick, 2003). - Parent \"freezes\" and stops responding - The interaction breaks down - Attempts to repair the interaction (social engagement cues) Social referencing and the Visual Cliff example (Sorce et al., 1985). - Visual cliff (depth perception) - Infants will look to the parent for an emotional cue of how to respond - Shared attention to the situation, transfer of information **The visual cliff:** - Social referencing and the Visual Cliff example (Sorce et al., 1985). - Visual cliff (depth perception) - Infants will look to the parent for an emotional cue of how to respond - Shared attention to the situation, transfer of information. Secondary Intersubjectivity (5) Beginnings of intentional communication by the infant signified by: - Use of eye contact/pointing to direct another's attention. - Consistent use of vocalisation to indicate specific goal. - Evidence of child waiting for response. - Persistence if not understood. Two key modes of communication important for language acquisition = turn-taking and joint attention. **Modes of communication** Turn-taking. Joint Attention. - Sharing a focus of attention. - Following attention. - Directing attention **Turn-taking (1):** - Young infants (from around 3 months) alternate vocalisations with their caregivers (Stern et al. 1975). - By 12 months, very few overlaps between 'speakers' (Schaffer et al, 1977). - Proto-conversations (Bruner, 1975) - similarities between turn-taking in early vocalisations and later conversation. **Do infants really have sophisticated turn-taking skills?** - Interruptions suggest not until 3rd year can children control turn-taking in language (Rutter & Durkin, 1987). - In the early stages the caregiver ensures a smooth interaction between speakers. - Difficult to establish exactly when mutually intentional. **Joint Attention (1)** - Initial interactions incorporate either: The child and adult or the child and an object. **Joint Attention (2)** - Joint attention = triadic interaction involving child, adult, and object/event. - Shared awareness of the shared attention. **Joint Attention (3)** Joint attention = triadic interaction involving child, adult and object/event. - Sharing Attention. - Following Attention. - Directing Attention. **Joint Attention: Sharing Attention (1)** Social referencing (Sorce et al., 1985). - Visual Cliff. By 9 months, children look to adult in unfamiliar or threatening situations to gauge emotional response. **Joint attention: Sharing attention (2)** Topic Comment - At 9 months, child and adult interact over an object. Child switches gaze between adult and object (Carpenter et al, 1997). - Caregivers talks about object of joint attention.(West & Iverson, 2017) **Joint attention: Sharing attention (4)** Topic Comment - Joint attention skills predict later language skills (e.g., Tomasello & Farrar, 1986). - Children better learn the names for objects better when they are attending to the object when it is named (Pereira, Smith & Yu, 2014). **Joint attention: Sharing attention (5)** Routines (Bruner, 1983) - Much of early language is learnt in routines - Caregivers structure routines around child. - Routines create a shared context. The child knows what comes next. - Highly repetitive routines provide a scaffold for language learning. - Routines differ in the types of words used (Tamis-LeMonda et a., 2018). **Joint attention: Sharing attention (6)** - BUT\... - During 1st year, mothers constantly monitor the child's line of regard. When a child's attention shifts from the desired object of attention, mothers attempt to regain the child's attention (Collis & Schaffer, 1975). - The mother is initially solely responsible for establishing a shared topic and providing relevant language. - Mother's sensitivity to a child's focus of attention is related to the child's vocabulary development - children are more likely to learn the referent for an object they attend to than for one their attention is directed to (Tomasello & Farrar, 1986). - Twins often show language delay -- linked to amount of time spent in joint attention episodes with mother (Tomasello et al, 1986), highlighting the mother's role. **Joint attention: The following points** - 9 months: can follows point in front of another person. - 12 months: begin to check back with the pointer. - 14 months: follows point across the line of sight. **Joint attention: Following attention - Gaze Following (1)** - Gaze following allows us to track where someone else is looking and join them, engaging in joint attention. - By 9 months, infants will turn to follow an adult\'s gaze and share an object of attention with another (Scaife & Bruner, 1975). **Joint attention: Following attention Gaze Following (3)** - Infants aren't tracking the GAZE specifically until around 18 months (Corkum & Moore, 1995; Moore & Corkum, 1998). **Joint attention: Following attention Gaze Following (4)** - 12-month-olds will follow a head turn....even if the person is blindfolded! - 14-month-olds will only follow when the eyes are visible (Brooks and Meltzoff, 2002) **Joint attention: Following attention Gaze Following (5)** - But, 12-month olds will gaze follow if the partner has their eyes open, but not if the eyes are closed! (Brooks & Meltzoff, 2002). **Joint attention: Following attention Gaze Following (6)** - Infants follow gaze behind barriers (Moll & Tomasello, 2004). **Joint attention: Following attention Gaze Following (7)** - It can be tricky to determine to motives for infant gaze following. - Conflicting evidence about when children are following because they think the looker sees something interesting. - 18 months, but some evidence in infants as young as 12 months. **Understanding communicative intentions (1)** - Can children understand that adult intends to communicate information to them? - Behne et al. (2005) look at whether 14-, 18-, and 24- month-olds will follow a helper\'s point. Do they understand that there is a shared goal of finding the toy? **Understanding communicative intentions (2)** - Infants follow both point and gaze direction to retrieve object of interest (Behne et al., 2005). **Understanding communicative intentions (3)** - Infants do not follow non-communicative points and gaze direction (Behne et al., 2005). **Joint attention: Directing attention (1)** Pointing - Imperative --to get adult to do something. - Declarative -- to direct adult's attention to something. - 9 months, child points to object then checks mother's line of regard, by 18 months child checks mother's line of regard before pointing to an object. **Joint attention: Directing attention, Criticisms** - **Imperative pointing:** infant learns that if she points, she gets what she wants (Camaioni, 1993). - **Declarative pointing:** infant learns that she gets more attention by pointing at things (e.g. Moore & D'Entremont, 2001). - BUT 12month-olds indicate when an adult finds 'wrong' object (e.g., Liszkowski, et al., 2006) and respond negatively when attention is directed to the infant and not the object (e.g., Boundy et al., 2019) **Conclusions** Two main stages of social development - Primary Intersubjectivity - Secondary Intersubjectivity Key communication skills - Turn Taking - Joint Attention Development of social skills and understanding of communicative intent **LECTURE 3:** **Word Learning is HARD:** - It can't just be "point and name" - Point and name is not common (and not universal) - When pointing and naming, usually only nouns **Word Learning is HARD (2)** A picture of a woman petting a dog. The word &\#34;gavagai&\#34; is shown along with possible meanings of the word (e.g., dog, ear, sitting, brown, paw, stroke, bark, fast) **Word Learning is HARD (3)** - Getting meaning right - Under-extension  **Word Learning is HARD (4)** - Getting meaning right: - Over-extension The word dog; with arrows pointing to a picture of a dog, a picture of a lion, and a picture of a horse. **Early Word Knowledge: Comprehension (1)** - Comprehension precedes production - 2-year-olds comprehend 2-3x as many words as they produce (Goldin-Meadow et al., 1976) - Infants appear to start to comprehend nouns as early as 6 months (Bergelson & Swingley, 2012) - Infants start to comprehend verbs (e.g., eat, hug) later \~ 10 months (Bergelson & Swingley, 2013) **Early Word Knowledge: Comprehension (2)** - Between 18- and 24-months, infants get much faster on the looking-while-listening task (Fernald et al., 1998). - By 18 months, they don't even need the full word (Fernald, Swingley, & Pinto, 2001) **Early Word Knowledge: Production (1)** - Around 12 months, first words, by 24-30mths around 500 words - (however, lots of variability) - First words from a range of categories - Nouns (objects -- dog, cat & proper names - Mummy) - Verbs (action words -- jump, get) - Social routines (bye, hello, please) - Adjectives (descriptions -- cold, dirty) - Lack of things like articles (a, the) **Early Word Knowledge: Production (2)** **[Early Noun Bias]** - Cross-linguistically, predominance of nouns in early vocabularies (e.g., 40% of English-speaking children's first 50 words -- Nelson, 1973) - More nouns even in \"verb-friendly\" languages - Natural partitions hypothesis (Gentner, 1982) - Early nouns denote concrete objects easily individuated from surroundings. - Actions, states etc. tend to apply TO entities labelled by nouns, less clearly defined in space & time. **Early Word Knowledge: Production (3)** - Early Noun Bias - Socially mediated word learning (Tomasello, 2003) - Not all early words are nouns (hello, bye) - Not all early nouns are discrete objects (breakfast) - Learning occurs in situations where easiest to read adult's intentions, irrespective of word class. - Happens often with nouns **Early Word Knowledge: Production (4)** - Words used in a variety of situations (Tomasello, 1992) - Names for people and objects -- e.g. Daddy, spoon - Names for actions -- e.g. open to request that a door be opened, a jar be opened etc.  - Names for properties - gone, more, dirty **Early Word Knowledge: Production Errors (1)** - **Under-extension:** word used only in specific context or specific exemplar - Words used in specific contexts where adults would use in a wide range of contexts. - *bye* only when putting the telephone receiver down (Bates et al, 1979) - *there* only when putting an object in a location (Barrett, 1982) - Refer to the word *flower* only to mean a rose and NOT other flowers (Fernandez & Cairns, 2011) **Early Word Knowledge: Production Errors (3)** - **Overextension:** a word used beyond its true meaning. - Overextension errors are frequent. E.g. calling a ball an apple. - Generally, make errors until the age of 2.5 years. - Category error (the concept of ball is incorrectly in the same category as apple). - Vocabulary limitations (lack the word "ball"). **Innate constraints on early word learning (1)** **Gavagai --** how do children know what a word refers to? 1. **Object constraint =** words refer to objects, explains early noun bias. 2. **Whole object constraint =** words refer to whole objects rather than their parts. Gavagai = whole animal, not tail, ears, legs. 3. **Principle of contrast =** no two words have exactly the same meaning, explains how the child overcomes overextension. 4. **Mutual exclusivity =** no object has more than one name. Helps children override the 'whole-object constraint' and learn the names for parts of objects. **Problems with constraints theories:** **Do constraints explain word learning or just describe it?** - Non-noun words? **Are constraints innate or learned via experience?** - Little research on infants. **Are constraints specific to language?** - My uncle gave me this (show object). - Give me the one my dog likes to play with (from array). - 3-year-olds select the new object, social inferencing on intention unrelated to the meaning of words. **Structural cues to word meaning (1)** Syntactic bootstrapping hypothesis. 3- to 5-year-olds shown a picture of someone kneading a substance in a bowl. - Do you know what it means to sib? - Do you know what a sib is? - Have you seen any sib? Task -- pick sibbing, a sib, or a sib from a selection of pictures depicting several actions, substances, and containers. - Sibbing = picture of kneading. - A sib = picture of a bowl - Sib = picture of substance. **Structural cues to word meaning (2)** Nouns refer to objects/categories and adjectives refer to properties. Nouns and adjectives differ in how they are used in sentences. - "Can you hand me the X?" = noun - "Can you hand me the X one?" = adjective - "Can you hand me the red X?" = noun - "Can you hand me the X pen?" = adjective **Structural cues to word meaning (3)** - Gelman & Markman (1985) - 4-year-olds pick a different object of the same kind when asked to **find the fep one**, but a different object when asked to **find the fep.** **Structural cues to word meaning (4)** Waxman and Booth (2001) looked at how 14-month-olds extend novel nouns and adjectives. With **nouns,** children extend the noun to the category but not the property. - Children see objects (e.g., a purple elephant, purple dog, purple bear, purple lion). - Children are told \"Look! These are blickets! This one is a blicket and this one is a blicket.\" - When shown a purple horse and a purple plate and asked to give a blicket, they give the horse - When shown a purple horse and a blue horse and asked to give a blicket, they give randomly **Structural cues to word meaning (5)**  **Structural cues to word meaning (6)** - Waxman and Booth (2001) looked at how 14-month-olds extend novel nouns and adjectives. - With **adjectives**, children do not extend to the category OR the property. - Children see objects (e.g., purple elephant, purple dog, purple bear, purple lion). - Children are told \"Look! These are blickish! This one is blickish and this one is blickish.\" - When shown a purple horse and a purple plate and asked to give the blickish one, they give randomly - When shown a purple horse and a blue horse and asked to give the blickish one, they give randomly - Children DON'T extend it to the category, but also don't extend it to the property.  They seem to understand that it is not a noun, but don't quite get what it actually does **Structural cues to word meaning (7)** Visual Depiction of the Adjective condition of the Waxman and Booth (2001) paper. **Structural cues to word meaning (8)** - Structural cues to nouns seemed to be learned early, but structural cues to other words appear later. - Adjectives - 18-month-olds show the same pattern (Booth & Waxman, 2009) - 21-month-olds are getting better but not great (Waxman and Markow, 1998). **Structural cues to word meaning: Verbs** - This can be used to narrow down verb meanings.  Two-year-olds use structural cues to narrow down verb meanings (Naigles, 1990) - A man is running with a dog right behind him: - The dog is **meeking** the man. - The man is **meeking** the dog. - The man is **meeking**. - What does **meeking** mean? **Structural Cues to Word Meaning: Issues** - Children sensitive to some aspects of sentence structure, but not clear exactly what and when. - The chicken and the egg...Some knowledge of words and word categories is needed to understand their structure. - Do experimental studies reveal something about long term learning of word meaning, or immediate problem-solving task? - Structural information can't solve all the problems - The man's tamming over the bridge - Tamming = walking, strolling, going **The social-pragmatic approach** - Proposed by Tomasello (2003) - Children learn words and word meaning from pragmatic cues in the environment which remove ambiguities around word meaning. - Word learning is constrained in two main ways: 1. The social world is structured: - Routines,  games, patterned social interactions 2\. Social-cognitive skills the infant has: - joint attention, intention reading. **The social-pragmatic approach: Scaffolding and routines** - Children learn language in familiar social contexts in repeated daily routines. - Young children learn almost all their early language in cultural routines, e.g., feeding, games, book reading (Ratner & Bruner, 1978) - Cross-culturally, children are engaged in a wide range of social routines and learn most of their early words in familiar contexts (Lieven, 1994). **The social-pragmatic approach: Social-cognitive skills** - Social revolution at approximately 9 months. - During joint attention, adults use language and children attempt to interpret the communicative intent. - Word learning occurs when children attempt to interpret the communicative intentions as expressed in the utterance.  - The shared common ground reduces the possible referents (Baldwin, 1993). **The social-pragmatic approach: Social-cognitive skills (2)** - Children use things like eye-gaze and joint attention to identify referents from adults. - 18- to 20-month-olds learn names for objects better when the speaker and infant are jointly attending to the object (Baldwin et al., 1996) - Gaze-following behaviour at 10 months predicts language skills at 18 months (Brooks & Meltzoff, 2005) **The social-pragmatic approach: Intention Reading (1)** - Children use speaker's intentions to infer meaning (Baldwin, 1991, though see Jaswal & Hansen, 2006) - Child already knows the name of the familiar object on the table & she knows the adult knows this too.  - So she can assume (using intention-reading) that the adult intends the novel object when the adults asks her to "show me the modi." - 2-year-olds understand that a novel referent refers to object adult looking for rather than objects they have rejected (Tomasello & Barton, 1994). **The social-pragmatic approach: Intention Reading (2)** - Acquisition of verbs -- children able to interpret adult's anticipation of what will happen and learn verbs which relate to forthcoming action (Tomasello & Kruger, 1992) - Children can differentiate between intended and accidental actions when learning new verbs (Tomasello & Barton, 1994). **The social-pragmatic approach: Issues** - What kinds of inferential skills does the child bring to the task of language acquisition? - Can this process of learning account for the acquisition of complex syntax? **Word Learning Summary** - Language learning is not isolated - Cognitive, social, environmental factors - Children show key patterns in how they use language - Comprehension precedes production - Early noun bias - Children have access to a variety of sources of information when learning word meanings. - Different theories to try to explain how children learn word meanings. **Critical Evaluations** - Unclear when different information is used at different stages of development. - Would we expect the same strategies across languages and cultures? - How do children learn less salient words (e.g., the) or words that are more abstract (e.g., happiness, justice)? **LECTURE 4: EARLY MULTI-WORD SPEECH: CONTRUSTIVIST APPROACHES.** **What is syntax?** - The ways in which a language allows words to be combined: - Enables understanding between speakers, e.g. 'Who did what to whom' - Allows productivity -- with a finite set of words we can produce an infinite number of possible sentences. **What needs explaining?** - Language is: Species-specific: Little evidence other primates can acquire syntax even with intensive training. Species-universal: virtually all children have acquired the majority of the grammar of their language by 5yrs. **What are early word combinations like?** - Mainly content words - Refers to here-and-now, easily understood in context. - Creative More sing, All gone sticky, other one spoon - Observes adult word order truck gone vs. gone truck **Lexical (word-based) rules?** - Rules item-specific - based on individual words or schemas (sets of words). - Limited variety of utterances until children are able to generalise between schemas. **Syntactic (grammatical) rules?** - Rules abstract -- based on grammatical categories. - Rules not restricted, therefore allow all utterances possible in the adult language **Interim summary** - Children's early multiword utterances are not random, nor simply imitations of what they have heard - Children learn language, but other species do not demonstrate the same impressive abilities. - So how are children able to put words together into sentences, what kind of knowledge do they need to be able to do this, and where does it come from? **Part B: Constructivist approaches: theory & initial evidence** **What is the Constructivist approach?** - Grammar is used for communication - Infants are motivated to learn to communicate - Grammar can be learned using general cognitive learning mechanisms - Communicative intention-reading - Drawing analogies - Distributional learning **The role of routines** - Routines allow children to predict what happens next and therefore what the language they are hearing might refer to. - Repetitive chunks of language can then be learned in context where the relation between linguistic form and meaning is more transparent. **What kind of evidence would support the approach?** - Children begin with lexically-based linguistic representations - High frequency items are learned early - Only gradual generalisation across exemplars to create more abstract syntactic categories and rules **Evidence for lexically-based (word-based) learning** **The verb island hypothesis** - Knowledge of grammar tied to individual verbs until 2½- 3yrs. - Child initially unable to generalise between verbs with similar meanings or used in similar sentence types. **Experimental evidence** - With familiar verbs (e.g. chasing), 2-yr-olds able to describe actions correctly to explain who is chasing, and whom is being chased. - But with unfamiliar (novel) verbs (e.g. weefing), before 3yrs children struggle to explain who is doing what to whom. **Evidence: Limited (lexical) constructions** - Argue children's early utterances based around individual lexical items (words) but not exclusively verbs - I + X Can I + X - Where's X gone? X + go - More + X Don't + X - 'X' represents a set of possible words used in the slot - Any high frequency word / group of words can form the basis for organisation of the child's linguistic system. - The constructions children learn reflect the frequency of particular patterns in the input. **Interim Summary** - Good evidence that children's early utterances are more restricted than those of adults. - How do children start to 'join up' the different parts of their developing linguistic knowledge? **Part C: Building an adult-like grammar** **How do children link up their lexically-based constructions to form a more adult-like grammar?** **1. Structure combining** **How do children's utterances build on what they have previously said?** - Dense diary study of a single child for 6 weeks at 2;0 - Recorded for 5 hours/week, and written diary of all new utterances kept by mother. - All utterances on last hour-long recording noted -- 'Target'. - All previous recordings searched for closest match -- 'Source'. **Method** **Identify:** - What changes required to change closest matching utterance - the 'source' - into the 'target' utterance (operations) **Results** 295 multiword utterances 186 repetitions (63%) - 158 repetitions of something child said previously - 28 immediate repetitions of mother 109 novel utterances (37%, of these ¾ single operation change) - 68 substitutions + 12 add on + 1 drop - 22 utterances required 2 operations, e.g. add on + substitutionWhere's Daddy's work? Where's my Daddy's cup of tea - 6 utterances required 3 or more operations (substitute, drop, add) I can't put it back on I don't put it (\_\_\_) on there **Conclusions** - Many of the child's apparently complex utterances are based around repetitions or small changes to what she has said before. - Most changes involve simple substitutions within a lexically-based frame, or the addition or subtractionof a single word. - Suggests child is operating with an extensive inventory of specific utterances, and fairly limited mechanisms for altering these utterances to match the demands of the discourse context. **2. Semantic analogy** - Children need to learn a number of verbs before they can recognise similarities between them and begin to build more general schemas. Commonalities reinforced, differences forgotten. **Evidence: repeating sequences** - 2 & 3-yr-olds asked to repeat 4-word sequences - \[FRAME\] \[SLOT\] - Back in the 'box/case/town' (higher similarity) - It's time for 'lunch/soup/drums' (lower similarity) - Manipulated 3-word frame by similarity of meaning of items in 4th 'slot' - Children made fewer errors when items that normally occur in the slot are more similar -- suggests overlap in meaning helps build flexible constructions **3. Distributional learning** - The ability to learn the co-occurrence characteristics of the input, i.e. which words occur together or in similar contexts. **Experimental evidence for distributional learning** - 2-yr-olds exposed to multiple transitive sentences of form X is Verb-ing Y with familiar verbs - Noun Phrase only condition -- all Xs and Ys are lexical nouns The cat is chasing the mouse The bear is hugging the fox - Mixed condition -- Xs and Ys are combination of lexical nouns and pronouns The cat is chasing the mouse / He is chasing himThe bear is hugging the fox / He is hugging him. **Experimental evidence for distributional learning** - Children taught novel verb (This is called dacking) to describe a new action between two participants. - Asked What's happening here? to elicit description - In which condition are children more likely to show generalisation of the X is V-ing Y lexical frame? i.e. to say 'The dog/he is dacking the lion/him' - 'This is called dacking, What's happening here?' - Pronouns helped children extract a more abstract representation of the SubjectVerb-Object sentence structure for use with novel (unfamiliar) verbs. **Interim Summary** - Studies of children's language production suggest early language not organised around same categories and rules as used by adult speakers. - Evidence for gradual generalisations based on similarities in form and meaning of sentences. **Overall Summary** - Children begin to combine words together at 18-24 months - Constructivist theorists argue that children access meaning and learn to combine words by interpreting the intentions of their interlocutors -- from hearing language used in predictable contexts - Children build up grammar by starting with more limited scope rules (e.g. lexical rules) than those used by adults and using general cognitive mechanisms to generalise **Critical Evaluation** - Production studies are difficult for children -- significant memory load in remembering and recalling novel words, planning entire sentences. - Do production studies underestimate how abstract children's knowledge of sentence structure really is? - Exactly how sentence structures become gradually more abstract over development is not clearly specified. **LECTURE 5:** Early Multi-Word Speech: Nativist Approaches Part A: Background In contrast to Constructivist Approaches, Nativist (or generativist) approaches assume that children approach the task of learning language with innate machinery that is specific to language, sometimes described as a Language Acquisition Device or Universal Grammar (UG). **[Nativists]** - argue that children\'s utterances are creative because they have access to innate grammatical rules. - children observe adult word order because they have an abstract rule - generalisations (e.g. adding inflections to words, wug -\> wugs) provide evidence of abstract (innate) rules **[Constructivists]** - argue that ν children\'s utterances are creative because creativity is based on the use of lexical frames learned from the language children hear, with new items inserted into variable \'X\' slots, e.g. I want X. - children observe adult word order because they pick up highly frequency lexical frames from their input (which, of course, follow the adult word order) - generalisations demonstrate that children learn these patterns gradually from distributional analysis of the language they hear **1. Nativist assumptions** ♣ Assume that grammar is a symbolic computational system which processes the relationships between abstract variables (e.g. Chomsky, 1995; Marcus, 1998). ♣ Assume that grammatical categories and rules are given apriori in the child's brain from birth (UG). ♣ Predict that the acquisition of a particular aspect of grammar should have an all-or-nothing quality. 'As soon as an item is assimilated into a class, that item automatically inherits the privileges of that category.' (Marcus,1998: 250) **General predictions** - Radford (1990: 61): '\...Once a child is able to parse an utterance such as \"close the door!\", he will be able to infer from the fact that the verb \"close\" in English precedes its complement \"the door\", that all verbs in English precede their complements...' - Prediction 1: children should learn these innately specified aspects of grammar very early on - Prediction 2: children should show consistent treatment of members of a particular grammatical category. **2. The nature of UG: Principles and Parameters** - All the possible rules for languages are innate. - Grammar is universal (UG) -- the rules of grammar apply in all languages. - Where the rules of grammar differ across languages, they do so in highly constrained ways which are encoded by parameters. - Children need to work out which parameter settings apply for the language they are learning. **Examples of parameter settings** Word order -- Verb-Object (English) or Object Verb (Japanese) - I eat sashimi - "watashi-wa sashimi-o tabe-tai-desu" = I sashimi eat-want Subject use -- In some languages subjects are obligatory (English), in others subjects are optional (Italian). - It is raining Sta piovendo = is raining **3. Theoretical advantages of UG** - Avoids problem of explaining how children acquire complex grammatical rules - Allows a unified theory of acquisition across languages whilst explaining how languages differ. **Empirical Evidence for Principles & Parameters** - Children's early utterances (usually) observe adult word order -- taken as evidence the relevant parameter is set. - Children are productive from early on (allgone sticky) --taken as evidence they are applying rules of grammar. - Some evidence that children understand the role of word order (Subject-Verb-Object transitive construction) from age 2 years or earlier from preferential looking studies... **Preferential looking & pointing studies** - Children aged 1;9 can identify the correct picture to match Subject-Verb-Object sentences from a choice of 2 causal actions - Taken as evidence for setting the word order parameter - BUT - disagreement from constructivists as to what these results mean -- comprehension vs. production **4. Theoretical problems for UG** - Parameters not specified. - How many parameters are there? - Which aspects of language are coded by parameters, and which are not? - Unclear how children avoid setting parameters incorrectly. E.g., Want a drink?, Got to go now - Bilingualism -- how do children set two (or more) versions of same parameters? **Empirical evidence against P&P** - Children display limited knowledge of SVO word order in production and act-out studies (e.g. Akhtar et al, 1997; Akhtar, 1999; Matthews et al, 2005; Chan et al, 2010). - Naturalistic data studies provide evidence of partial, lexically specific knowledge within a grammatical category -- verbs, auxiliaries, determiners (e.g. Pine et al, 1998; Lieven et al, 1997; Wilson, 2003) - Many studies show a very close relation between what children hear, how often, and what and when they learn(e.g. Ambridge et al., 2015) **Interim summary** - Nativist approaches provide an account of children's early multiword utterances that emphasises their similarity to adult language - Continuity accounts (that posit grammatical rules from the outset) explain development in terms of limitations on performance rather than limited knowledge - Next: maturation accounts to explain why children's language develops, while maintaining innate knowledge **Part C: Maturational Models** - children's language develops (changes over time), so many researchers argue that this provides evidence that they do not start out with a full innate UG (contra continuity accounts). - One solution to this problem is to build in a part of UG that matures over time according to a biologically-determined timescale (e.g. Radford, 1990). **Radford's (1990) maturational model** - At the Lexical Stage of development (around 20 months), children\'s utterances consist of mainly content words such as nouns, verbs, adjectives and prepositions, with other parts of the corresponding adult utterance omitted. - At the Functional Stage (around 24 months) the child\'s innate grammar \'matures\' and the parts governing the use of more complex grammatical components switch on; for instance auxiliary verbs (e.g. to mark modality, certainty, futurity - can, will, might), determiners (to distinguish definite and indefinite referents - a/the), and inflections (to mark tense and agreement - watch/watched, I watch/He watch-es) **Distinct stages of development?** **Lexical utterances** - Kathryn no like celery - Hair wet - Pig say oink - Mummy doing? - Hands dirty **Functional utterances** - I'm pulling this - I don't need that - Will you help me? - She likes ice-cream - I watched the ducks **Theoretical Advantages & Evidence** - Explains why early utterances are not fully grammatical. - Allows for development over time so more likely to fit the empirical data. - Some have claimed a similar trajectory of learning for typically developing children with normal hearing, deaf, blind (e.g. Gleitman, 1981), despite their experiences of the world being different. **Theoretical & Empirical Problems** - Difficult to identify precise points in development when maturing aspects of the grammatical system come 'on-line'. - From earliest stages, children show some use of most grammatical functions, although inconsistent, and vary across languages - At around 24 months, children's use of many 'functional' words related to lexical frames - Can I X?, Don't X etc. (Lieven et al, 1997) **Interim Summary** - UG approaches claim innate abstract grammar but explain changes in children's language over development in terms of biological maturation of parts of the grammatical system. - Do changes in language over development reflect a process of gradual learning from the input combined with the ability to be productive in limited ways? (constructivist account) - How can we distinguish these two approaches? **Part D: The linking problem** **1. What is the linking problem?** How do children link up their innate knowledge of grammatical categories to the words they are hearing? - Caregivers don't label particular words as nouns, verbs etc. - But Universal Grammar -- what children are hypothesised to have innately - is defined in these terms **2. A proposed solution -- semantic bootstrapping** Assumes: - Grammatical (syntactic) categories and rules innate - Children use semantics (meaning) to map words in the input onto these innate syntactic categories by using innate Linking Rules to map semantics onto syntax **Linking rules between meaning and syntax** - The child 'links' individual words to innate grammatical categories (e.g. noun, verb, adjective, preposition,...) **Word meaning** **Child assumes that the word is a...** ----------------------------------- ----------------------------------------- Attribute Adjective Person, thing Noun Action, change of state Verb Spatial relation, path, direction Preposition **Linking rules cont.** - The child can also link semantic roles: - Agent -- the person carrying out the action - Patient -- the person or thing affected by the action to syntactic roles - Agent = Subject of sentence - Patient = Object of sentence **How does linking work?** **IMAGINE YOU HEAR:** "wug tamo pim" - 'Wug' means 'dinosaur' - 'Pim' means 'puppet' - The agent in this picture/event is the puppet ('Pim') - So, the puppet is the Subject - So, the Subject comes after the Object in this language -- Object-Verb-Subject word order **The problem -- it's not always easy to work out grammatical categories from meaning...** - Not all verbs are actions (Believe, want, need) - Not all nouns are concrete objects (Idea, dream, justice) - Not all subjects are agents (She wants a drink) **Solution:** - Use a form of distributional analysis to determine word order for the language from prototypical sentences. Then apply knowledge of word order to work out grammatical categories of more abstract terms. **[Using a prototypical sentence to work out word order and grammatical categories for non-prototypical examples]** **Prototypical transitive sentence** - The cat \[Agent=Noun=Subject\] - chased \[Action=Verb\] - the mouse \[Patient=Noun=Object\] **3. Advantages of semantic bootstrapping** - Explains how children break into innate system. - Explains why early utterances follow adult word order. - Explains how children learn verbs which are not actions, nouns which are not objects etc. **4. Problems for semantic bootstrapping** - Many of children's early lexically-specific utterances are not semantically prototypical, and therefore are unlikely to be based on innate knowledge of semantic linking rules - I want a drink, I don't like it (Lieven et al, 1997) - In passive sentences, the noun phrase (NP) which is usually the object of an active transitive becomes the subject **[ACTIVE:]** The cat (agent:Subj) chased the mouse (patient:Obj) **[PASSIVE:]** The mouse (patient:Subj) was chased by the cat (agent:Obj) The problem of passives - If the child hears passive utterances (e.g. the postman was bitten by the dog) early on, she may use semantic bootstrapping to conclude that her language is object verb-subject problems parsing other utterances - Some nativists propose that the passive 'Parameter' doesn't mature until later (5yrs) so passives are learned late (e.g. Borer & Wexler, 1987) - Children do hear and use passive sentences from fairly early on, (e.g. in Tomasello's verb island study), especially in some other languages (e.g. Demuth, 1989). **Interim summary** - Nativist approaches claim an innate abstract UG. - But how do children map words onto grammatical categories? - Semantic Bootstrapping proposed as a solution based on children's perceptual understanding of the world around them and innate linking rules - But the approach faces problems in terms of fit to the empirical evidence from children's early utterances **Overall Summary** - Nativist account introduced to explain how children learn how to put words together into sentences. - Assumes children operate with innate knowledge specific to grammar. To account for differences between languages, grammar encoded in Principles and Parameters. - Continuity accounts assume children start out with full grammatical knowledge - Maturational accounts assume parts of the grammatical system \'switch on\' at different stages in development based on a predetermined biological timetable. - Key challenge - how do children link up the language they hear with their innate grammar; the Linking Problem. **Critical Evaluation** - Children show improved performance on comprehension tasks compared to production -- does this demonstrate that they have innate abstract knowledge? - How can we differentiate between maturation of innate grammar and development due to gradual learning? - And for those of you interested in neuroscience, how might innate grammar be represented in the brain?