User:Alindsay9/sandbox

Heading text
This is the course page for Language Acquisition for Winter Semester 2012:


 * Psyc 471

This course is offered at the University of Alberta Augustanacampus, and focuses on the development of language as it begins in childhood.

Practicing References
Seligman and attributional reformulation

The question of meaning, or 'the mapping problem'
In word learning, the mapping problem refers to the question of how infants attach the forms of language to the things that they experience in the world. There are infinite objects, concepts, and actions in the world that words could be mapped onto. So how does the language learner successfully map words onto the correct objects, concepts, and actions?

Several answers to this question have been explored. Some theorists believe that children have built-in constraints that help them to map words onto concepts in the world .Domain-general theories, in contrast, suggest that word learning can be explained through general cognitive mechanisms, like learning and memory, which are not specific to language. Social pragmatic theories propose that caregivers act as experts in order to assist children in the word mapping process. Lastly, an emegentist coalition model has been proposed to consider multiple cues that children might use throughout the word learning process.

Theories of constraints
Theories of constraints argue for biases or default assumptions that guide the infant through the word learning process. These biases, or constraints, are outside of the infant's control and are believed to help the infant limit their hypotheses about the meaning of words that they encounter daily. Constraints can be considered domain-specific (unique to language).

Numerous types of constraints have been proposed over the years. Below, the most prominent constraints in the literature are detailed: reference, mutual exclusivity, shape, whole object assumptions, and taxonomic assumptions.

Reference
Constraints theories argue that word learning depends on a set of principles. The most central principle in word learning is believed to be reference, or the notion that a word symbolizes or stands in for an object, action, or event. According to the principle of reference, words are different from other sounds because they have a special relationship to their referents. Words consistently stand for their referents, even if referents are not physically present in context.

Mutual Exclusivity
The mutual exclusivity constraint rests on the assumption that only one label can be applied to each object. Literature in this field reveals that mutual exclusivity might influence an infant's word learning in four different ways. First, mutual exclusivity might influence the infant's decision about the reference of a new word. For example, if a child is faced with a familiar object for which they already have a label, and an unfamiliar object, which they have not yet labelled, they are likely to map a novel word onto the unfamiliar object. This is called the disambiguation effect. Next, the bias might cause an infant to change the extension of a familiar word. In this case, a child might remove 'wolf' from their extension of 'dog' upon hearing an animal labelled a 'wolf.' This can be called the correction effect. As well, the bias might compel an infant to reject a new word, for instance, rejecting 'wolf' in favour of 'dog.' This is the rejection effect. Lastly, mutual exclusivity might influence word generalization. In this case, if a child knows an object to have a particular name, the child should not generalize other names to it. This is known as the restriction effect.

There are conflicting positions in the literature regarding the mutual exclusivity bias. Some theorists believe that children possess this bias from the start of word learning, others argue that they never have the bias, and yet others believe that the bias is acquired throughout early childhood.

Shape
Research suggests that infants assume that objects that have the same shape also share a name. Shape has been considered the property perhaps most critical for identifying members of an object category. Children's attention to shape can be demonstrated in their early over-extensions. Children will often extend their limited vocabularies by applying the words that they know to objects that are similar in shape to the original object in question. . Therefore, shape plays an important role in both appropriate and inappropriate extensions, and children attend to it from an early age.

Whole Object Assumptions
When infants hear a novel word, and see an object to which the word may refer, how do they know whether the word refers to the object's color, shape, size, texture, part, or to the entire object as a whole? If children entertained all these possibilities, word learning might be a very challenging process. Instead, past research shows that infants assume that labels refer to whole objects instead of parts or properties of these objects. According to constraints theories, this assumption arises because children have a built-in constraint which allows them to interpret names as referring to whole objects. This 'whole object bias' means that children will typically label whole objects first, and parts of properties of objects later in development.

Taxonomic Assumptions
Throughout word learning, infants assume that speakers use words to refer to categories that are internally consistent. In other words, children believe labels to pick out coherent categories of objects, rather than those objects and the things that are related to them. For example, children would assume that the word 'dog' simply refers to the category of 'dogs,' not to 'dogs with bones,' or 'dogs chasing cats'. The taxonomic constraint would limit childrens' hypotheses about language by preventing them from having to consider infinite numbers of combinations of objects within a single category.

Criticisms
Theories of constraints are well documented, however, critics argue that these theories focus on how children learn nouns, ignoring other aspects of their word learning. For example, theories of constraints may not be able to explain how children learn meanings for actions, properties, and relations early on. Constraints are useful in explaining how children limit possible meanings when learning novel words, but these same constraints would have to eventually be overriden because they are not utilized in adult language. For instance, adult speakers often use several terms, each term meaning something slightly different, when referring to one entity, such as a family pet. This practice would violate the mutual exclusivity constraint.

Domain-General Views
Domain-general views of vocabulary development, also known as connectionist or associative models, argue that children do not need principles or constraints in order to successfully develop word-world mappings. Instead, children notice the objects, actions, or events that are most salient in context, and then associate them with the words that are most frequently used in their presence. Word learning can be accounted for through mechanisms such as salience, association, and frequency. For instance, research on word learning suggests that fast mapping, the rapid learning that children display when exposed to a novel word, is not specific to word learning. Children can also successfully fast map when exposed to a novel fact, remembering both words and facts after a time delay.

'Biased' Through General Learning Mechanisms
Domain-general perspectives do not dismiss the notion of biases, but rather, suggest that these biases might develop through general learning strategies instead of existing as built-in constraints. For example, children may begin learning object names through trial-and-error, or associative learning. Children may be exposed to cues associated with categorization by shape, which would draw their attention to shape when they encounter a novel label. Ordinary learning could, then, lead to a shape bias. In addition, the whole object bias could be explained as a general strategy that humans use to reason about the world; perhaps we are prone to thinking about the world in terms of whole objects, and this strategy is not specific to the language domain.

Criticisms
Domain-general views have been criticized for not fully explaining how children manage to avoid mapping errors when there are numerous possible referents that objects, actions, or events might point to. If biases are not present from birth, infants might have to endure lengthy trial and error processes before establishing correct word-world mappings.

Social pragmatic theories
Social pragmatic theories, also in contrast to the constraints view, focus on the social context in which the infant is embedded. According to this approach, it does not matter that an array of meanings could be assigned to a novel word because environmental input removes the ambiguity of the word learning situation. Cues such as the caregiver's gaze, body language, gesture, and smile might help infants to understand the meaning of a word. In particular, social pragmatic theories typically stress the role of the adult or caregiver in talking about objects, actions, or events that the infant is already focused-in upon. Joint attention has been discussed as a crucial context for word mapping. As well, children may adopt some of the same pragmatic assumptions that adults have in order to facilitate the world-learning process.

Joint Attention
Joint attention is an important mechanism through which children learn to map words-to-world, and vice versa. As mentioned previously, adults often make an attempt to establish joint attention with a child before they convey something to the child. Joint attention is often accompanied by physical co-presence, since children are often focused on what is in their immediate environment. As well, conversational co-presence often occurs; the caregiver and adult typically talk together about whatever is taking place at their locus of joint attention. Joint attention is natural and common in the every day routines of caregivers and children, therefore making it an ideal context for children to learn words.

Pragmatic Assumptions
From early on, children assume that language is designed for communication. Infants treat communication as a cooperative process. Specifically, infants observe the principles of conventionality and contrast. According to conventionality, infants believe that for a particular meaning that they wish to convey, there is a term that everyone in the community would expect to be used. According to contrast, infants act according to the notion that differences in form mark differences in meaning. Children's attention to conventionality and contrast is demonstrated in their language use; they direct their early words towards adult targets, repair mispronounciations quickly if possible, ask for words to relate to the world around them, and maintain contrast in their own word use. Infants typically do all of these things before the age of 2 years. The pragmatic assumptions that infants may have seem similar to the principle of mutual exclusivity (ME), as together, conventionality and contrast predict that children will initially only assign one label per referent.

Criticisms
Social-pragmatic perspectives often present children as covariation detectors, who simply associate the words that they hear with whatever they are attending to in the world at the same time.

The co-variation detection model of joint attention, which focuses on the caregiver as actively following in on the child's attentional focus, seems problematic when we consider that about 30 to 50 percent of caregivers' utterances do not refer to things that occupy the immediate attentional focus of infants. Additionally, especially in certain cultures, infants are likely to hear words in the absence of their referent. If caregivers must do most of the work in helping children through the word-learning process, how do children learn words in societies where adults make no direct effort to teach them, or even speak to them ? Assuming the co-variation detector model is correct, these children should make numerous errors in word-mapping. However, evidence to support frequent errors in word-mapping does not exist. Instead, the majority of children in all societies have no problem developing a rich vocabulary.

While emphasis is usually placed on the role of adults in guiding the meaning-making process, some theorists have suggested that the child also plays an important role in this interaction. For example, research shows that children appear to recognize when an attentional focus is shared, as they do not attach a label to a novel object if adults do not appear to be referring to the object and that infants appreciate the intentions of others. When infants are in situations where their own attentional focus differs from that of a speaker, they actively seek out information about the speaker's focus, and then use that information to establish correct word-referent mapping. Joint attention is not solely created by caregivers in order to create the best conditions for infants to engage in co-variation detection. Instead,joint attention can also be created through infant agency, in an attempt to gather information about a speaker's intent. According to this perspective, infants are active participants in their own word learning.

Emergentist Coalition Model
The emergentist coalition model suggests that children make use of multiple cues in order to successfully attach a novel label to a novel object. The word-learning situation may offer an infant combinations of social, perceptual, cognitive, and linguistic cues. While a range of cues are available from the start of word learning, it may be the case that not all cues are utilized by the infant when they begin the world-learning process. While younger children may be able to detect only some cues, older, more experienced word learners may be able to make use of a range of cues. For example, very young children might focus primarily on perceptual salience, but older children might attend to the gaze of caregivers, using caregiver's focus to direct their word-mapping. This model, therefore, argues that principles for word learning may have some foundation at the beginning of word learning, but emerge or develop fully over time.

Towards a Holistic Account of Word Learning
Supporters of the emergentist coalition model argue that, as a hybrid model, this models can holistically explain word learning in ways that other theories can not. For instance, constraints theories typically argue that constraints/principles are available to children from the onset of word learning, but do not explain how children develop into expert speakers who are not limited by constraints. Domain-general perspectives do not address the question of how children sort through numerous potential referents in order to correctly sort out meaning. Social-pragmatic theories claim that social encounters guide word learning. However, these views take for granted that children are capable of utilizing social cues from the onset of word learning, which may not be the case. The emergentist coalion model incorporates constraints/principles, but argues for the development and change in these principles over time, while simultaneously taking into consideration social aspects of word-learning, and other cues, such as salience.

Summarizing 'The Mapping Problem'
While domain-specific accounts of word learning argue for innate constraints that limit infants' hypotheses about word meanings, domain-general perspectives instead argue that word learning can be accounted for by general cognitive processes, such as learning and memory. Yet other theorists have proposed social-pragmatic accounts which stress the role of caregivers in guiding infants through the word-learning process. According to research, however, children are active participants in their own word learning, although caregivers may still play an important role. These theories have all been documented, supported, and also subjected to controversy in the field. Recently, an emergentist coalition model has been proposed to suggest that word learning can not be fully attributed to a singular factor. Instead, a variety of cues, including saliency and social cues, may be utilized by infants at different points in their vocabulary development.