Word Learning Lecture 2 PDF

Lecture 2: Word Learning module Developmental psychology https://canvas.liverpool.ac.uk/courses/75783/pages/week-2- materials word-learning?module_item_id=2030953 date @30/09/2024 type...

Lecture 2: Word Learning module Developmental psychology https://canvas.liverpool.ac.uk/courses/75783/pages/week-2- materials word-learning?module_item_id=2030953 date @30/09/2024 type Core lecture Types of communication: Linguistic communication – natural language we use, eg oral, written, sign. Key feature is generativity, we can use the finate units to express an infinate number of ideas. Nonlinguistic communication – gestures, facial expressions Comprehension proceeds production, children typically understand a word before using it. Infant/child-directed speech: (Also known as baby talk) When talking to a child its done with a louder voice, slower speech, we tend to accentuate boundaries between words and phrases, use simpler words, repeat and expand on child's utterances, and recasting. This speech tends to attract children's attention more. To segment words from a speech stream we can use cues/information such as: Pauses between words: Pauses between words not very reliable because theres often no pauses in between words. Stress patterns 7.5 month old infants can segment words using predominant stress pattern of their native language. Infants familiarized with words such as Lecture 2: Word Learning 1  kingdom and hamlet listened longer to passages containing these words than to control passages. 9 month old infants can learn to use a different/new stress pattern in word segmentation. Two categories of words used: trochaic and iambic, trochaic words stress the first syllable, iambic words stress the second syllable. Transitional probability Transitional probability is the probability of one sound following another sound. In the study 8 month olds were familiarised with lots of novel words without pauses for 2 minutes. After this session they get tested using part words. General vocabulary development in first few years: One word utterances: Tend to emerge around 1 year, very basic language Holophrase: a single word standing in for a larger sentence eg) hot, up, stick. Multiword utterances: Tends to emerge around 18 months Tend to follow word orders Learned by rote, they remember the phrase, inflexible Telegraphic speech/texting speech, keeps in the most important words. Mapping words to concepts: Early words tend to be concrete nouns and basic level category names Fast mapping: quickly mapping a novel word to its meaning. Linking words to concepts: Willard quine – referential ambiguity problem/riddle of reference. Word learning heuristics/constraints/bias: Help us to constrain the meaning that we use when we map words. Perceptual heuristics: (the things that you can percieve) Lecture 2: Word Learning 2  Shape bias: children can correctly identify words by shape. By 18 months old most children have a shape bias. In one study children were shown shapes and asked to identify it by name. It was found that when children have more than 50 nouns in their vocab they show a very strong shape bias. Objects with the same shape tend to have the same name. If they have 25-50 words then they tend to randomly choose by shape, colour or material. Conceptual heuristics: Whole-object bias: eg referring to a full object such as a dog rather than parts of its body. Studies show at least by 18 months of age children show whole-object bias. Taxonomic assumptions: objects of the same kind are matched, we tend to assume that objects of the same kind share the same name. This is developed at least by 18 months. Pragmatic and social cues: Many studies have shown that by 12months children use pointing gestures to map objects. By 15 months of age they can use a persons gaze direction in their word learning. Children would map the scene that the person was looking at to the word in the houston-price et al study. Lexical heuristics: mutual exclusitivity: If you show children two objects, one of which they arent familiar with when a word is said they will assume that that is the name of the unknown object/novel thing. Cross situational learning is a new solution to word learning where: Children learn words over time Learning words in ambiguous situation Children learn nouns, verbs and adjectives this is a mechanism for learning the meaning of words across multiple exposures, despite exposure-by-exposure uncertainty as to the word's true meaning. Lecture 2: Word Learning 3  Children can do this by 18 months old Children can use this to learn nouns such as dog and cat (smith and yu) Another study by scott and fisher showed children can also use this learning to learn novel words. Two and a half year olds where shown 2 different videos, one of a women moving side to side and another one of a woman squatting. They then here she's pinning and she's nading but they're not told which is which. In the following trial they see a woman lifting her leg and another one moving side to side, they then hear shes tazzing and she's pinning. And in another situation they hear the terms she's nading and she's rivving. Because the children have already heard these words before they should be able to correctly identify what they refer to through association. External factors influencing vocabulary development Socioeconomic status Hart and risley – the 30 million word gap. Study involved 42 families, 13 professional, 23 working class, 6 welfare. Children and parent interactions were recorded monthly between 7 months and 3 years. For children from a professional background the average different words per hour were 293, in a working class family it was 216, in a welfare family it was 149. 32 million word difference between the two by the age the children were 4. When the kids are 3 year old, children from the proffessional family have a vocab of at least a thousand words (1116) but for a working class child its 749, and 525 for a welfare child. Children from the more proffesional background tend to have higher test performance for receptive vocab, language development and reading comprehension than those from a lower socioeconomic background. Fernald et al looked at the socio-economic status differences in language processing skill and vocabulary present at 18 months. Parents reported on the vocabulary size of the child They also tested childrens processing speed, child was shown two pictures and how fast the child looked to one of the pictures were measured. Children from the higher economic status were faster and more accurate. These differences come from: (fernald et al) Lecture 2: Word Learning 4  Genetics Physical conditions of everyday life Access to crucial resources Social and psychological support Stress level Stability in the family. Child directed speech In the rowe study they reported parent-children interaction when the children are 30,42 and 50 months old. they found that the amount and quality of child-directed speech predicts childrens vocab development. Primary caregiver education is positively related to both quantity and quality measures. Parent education also relates to childrens vocab size. Child-directed speech is not universal, casillas et al recorded data of children from 2months to 3 years from a mexican community and found that there was much less child directed speech. Parents in urban communites tend to use more child-directed speech when talking to children than those in rural communitites. Therefore not essential to developing language Interaction style responsiveness Tamis-lemonda, bornstein, baumwell observed 10 minute parent-child play when infants were 9 and 13 months. They coded the childs behaviour into 4 categories: bidding or looking to mother, exploring an object/toy, vocalizing, playing with a toy or object. Types of responsiveness were coded into one of 6 categories: affirmation of a childs actions, imitations of a childs vocalisation, descriptions of an object, event or activity, questions about an object, event or activity, play prompts or demonstrations and exploratory prompts. They found that overall maternal responsiveness predicts the timing of childrens achieving language milestones: first limitations, first words, 50 words in expressive language, combinatorial speech, and the use of language to talk about life. Children with mothers that are more responsive tend to have their first 50 words at an earlier age than children with mothers that are less responsiveness. Lecture 2: Word Learning 5 Lecture 2: Word Learning 6