Learning Word Meanings

Date First Published: October 6, 2008

Late emergence of spoken words is an early sign of autism spectrum disorders, ^{1 2} but most children with ASD acquire at least some spoken language, with approximately 80% producing more than five words. ³ Expression of words may be stronger than understanding. ^{4 5 6 7}       

Words are the essential building blocks of spoken communication; therefore, it is important to understand how verbal children with ASD best learn words. In this paper, we describe the components and processes of word learning and summarize relevant evidence from studies of children with ASD.

What is involved in learning a word?

To know a word is to know its sound structure, its grammatical role, its meaning, and its appropriate usage. In our laboratory, we are particularly interested in what children know about the meanings of words and the processes involved in acquiring this knowledge.

The initial process of word learning is termed “fast mapping” because, indeed, both children and adults typically are quick to attach meaning to a new word. ⁸ However, this new knowledge is fragile, in that it is easily forgotten, ⁹ and incomplete, in that it does not encompass a full appreciation of the word’s meaning. ¹⁰ The subsequent process of word learning is referred to as “slow mapping.” Given additional experiences with the word in meaningful environments, the learner establishes a stronger memory and a deeper knowledge of the word meaning.

A third process of word learning is extension. ¹¹ With the exception of proper names, words extend to members of a category, not just to a single item. For example, dog refers to all dogs, not just the family pet. Typical children as young as 12 months have a basic understanding that words can refer to categories, and they can extend a new word to multiple items based on that understanding. ¹² Even with this general appreciation, some additional experience with a given word in multiple contexts may be required before knowledge of the conventional category boundary is acquired. The learning of a word extension, then, may involve both fast mapping and slow mapping. The learner may assume from his first meeting with the word that it refers to a category of things, not just to a single thing (e.g., dog refers to all dogs), but he may need additional experience to establish the limits of that category (e.g., Siberian husky but not wolf) and the extent to which that category overlaps with or is encompassed within other categories (e.g., animal).

How do children accomplish fast mapping, slow mapping, and extension? We view these processes as the outcome of dynamic interactions between the child and his linguistic, social, and physical environments. Both the child and his communication partner -- be it a parent, teacher, therapist, sibling, or peer -- play a role in the creation and manipulation of those environments.

Fast mapping

Let us consider one simple, but not uncommon, environment. Imagine a child who is playing with two toys. One, a truck, is familiar; the other, a puzzle, is new. The child hears her father use a word she has never heard before, puzzle. How will the child know that the new word refers to the new toy? From a very young age, a typical child will figure this out by excluding the familiar toy as a potential reference -- she already knows that toy is called a truck so she infers that the other toy must be the puzzle. ¹³ Can children with ASD apply such principles? In 2005, Preissler and Carey ¹⁴ demonstrated that they can. Five- to 9-year-olds with ASD were as able as their younger vocabulary-matched peers to make such inferences. This is an example of how the child with ASD and that child’s environment interact during the process of word learning. It was the father (or the examiner) who provided the word at a time when the correct physical reference was present, but it was the child who applied her previous knowledge to finish the solution to the problem.

Let us imagine a slightly different environment. A child is again playing with two toys, but this time both are unfamiliar. When he hears his mother produce a new word, how will he know which toy the mother is naming? A typical child will look toward his mother and read her eye gaze or other body gestures. If she is gazing toward one of the toys and not the other, the child will infer that the word she spoke names the toy to which she is attending. ¹⁵ Do children with ASD use eye-gaze cues as a basis for inferring word meaning? Two studies, one by Baron-Cohen, Baldwin, and Crowson in 1997 ¹⁶ and the other by Preissler and Carey in 2005, ¹⁷ suggest that this is difficult for them. These two teams observed children, with ASD, developmental delay, or no impairment, in the type of word-learning environment described previously. When the new word was uttered, most of the children in the developmental-delay or no-impairment groups monitored the eye gaze of the speaker, shifted focus to the object, and mapped the word. Most of the children in the ASD group failed to monitor the eye-gaze cue, and so they failed to map the word.

This environment could be restructured to facilitate fast mapping. In 2006, McDuffie, Yoder, and Stone  ¹⁸ presented one possibility. They heavily supported attention to a novel object as it was named by presenting only one object at a time, placing the object close to the child and gesturing to and moving the object. Under these conditions, children with ASD did look at the object, and they did so more often in conjunction with naming than in conjunction with other sorts of comments. Simpler and more redundant environments may help to focus attention and reduce demands on executive memory, processes known to relate to language learning  ¹⁹ and known to be challenging for children with ASD. ^{20 21}

By gesturing toward and moving the object, McDuffie and her colleagues ²² also enhanced the prominence of the attention-directing cue. In 2007, Parish-Morris, Hennon, Hirsh-Pasek, Golinkoff, and Tager-Flusberg ²³ also tried this strategy, this time in the context of two unfamiliar objects. They found that when the examiner pointed to or touched, as well as gazed at, one of the two objects, the children with ASD successfully oriented to the object and fast mapped its name. But this happened only if the object was interesting to the child. If it was boring, then, even with the support of the point and touch gestures, the children with ASD were not successful.

Why do children with ASD fail to use eye gaze in service of fast mapping if it is not supported by extra gestures and simple or prominent object presentations? Eye gaze and other directional gestures such as pointing serve both to attract attention and to indicate intent. When the child sees his mother gaze toward an object while saying a word, he may follow the physical direction of her gaze and attend to that object; he also may infer that she intends to name that object. Therefore, we might ask whether the failure of children with ASD to use these cues is a failure to exploit their attentional or intentional meanings. Increasing evidence supports the latter explanation. Parish-Morris and her colleagues ²⁴ attempted to isolate social-intent cues from social-attention cues. Three- to 7-year-olds with ASD and two groups of typical children matched on the basis of either nonverbal intelligence or vocabulary level participated in a word-learning task that depended upon reading a social intent. An examiner pretended to unsuccessfully hunt for a parlu in a sack. She then asked the child to hunt through the sack and find it. The child had to infer that the unfamiliar object in the sack must be the object she intended, even though he had never seen the object while simultaneously hearing its name. Unlike the two comparison groups, the ASD group did not perform better than chance on this task. This finding demonstrates that children with ASD have difficulty using sophisticated combinations of cues, such as the language and facial expressions that might indicate frustration about not finding an object, to infer intent.

Slow mapping and extension

Whereas we are beginning to understand the environments that promote or impede fast mapping by the child with ASD, we know little about his slow mapping and extension of words. Two studies from the 1980s suggest that children with ASD, like other children, appreciate that words extend to all members of a category. In those studies, children affected by ASD performed just as well as language- and IQ-matched peers when asked to sort items into categories or to name categories and category members. ^{25 26} A more recent study confirmed good performance by adults and adolescents with ASD on a task requiring generation of members of the category “animals.” ²⁷

In our own laboratory we are collecting evidence on the outcomes of the slow-mapping process, if not the process itself. Specifically, we are finding that some children with ASD, aged 9 to 13 years, have a rather shallow understanding of word meanings. It isn’t that they do not know the words; rather, they do not know them completely. That said, we find that other children with ASD have rich, age- and IQ-appropriate knowledge of word meanings. In 2005, Norbury ²⁸ drew similar conclusions. Rather than investigating depth of word knowledge, she focused on flexibility of word knowledge. Specifically, she asked whether these children aged 9 to 17 years with ASD knew not only dominant word meanings (e.g., bank as a place to keep money) but also secondary meanings (e.g., bank as the edge of a river). Like us, she found that some children with ASD were limited in their knowledge of secondary meanings but that others had age-appropriate knowledge.

These are hopeful findings in that they suggest that the social deficits characteristic of autism do not necessarily limit successful slow mapping and extension of word meanings. But what of the child who does have these limitations? Can the environment be structured to support learning? Research on typically developing toddlers suggests that increasing the prominence of cues can enhance slow mapping and extension. Horst and Samuelson ²⁹ conducted four experiments in which typical 2-year-olds were presented with names for unfamiliar objects, tested immediately, and tested again after a 5-minute delay. In all four experiments, participants had no difficulty fast mapping word-object pairings. On the other hand, they were able to retain or extend novel names at better than chance levels only in the experiments that involved manipulation of the novel objects and direct naming. In other words, if the examiner was very direct in showing, pointing to, and labeling the object, the child was able to extend that name to other category members and to retain its name following a 5-minute interval. Booth and her colleagues ³⁰ presented typical 2-year-olds with two unfamiliar objects and named one of them. For some children, the examiner gazed toward the target object as she named it, for others, she gazed and pointed, and for others still she gazed and touched or manipulated the object. As the prominence of the gesture increased, from gaze, to point, to touch and manipulate, the toddlers’ extension and retention of the new words during a period of 3 to 5 days increased, with the biggest difference being the benefit of gaze and point vs. gaze alone. In 2005, Capone and McGregor ³¹ found that toddlers recalled new words more readily and had a deeper knowledge of the word meanings if the words had been taught in the context of a gesture that emphasized the meaning (i.e., shape or function) of the object than when the same words and objects were taught without gestural support. Although it is a leap to assume that these same strategies would support extension and slow mapping for children with ASD, these strategies do exploit the visual modality, a relative strength for many children with ASD. ^{32 33 34} 

Conclusion

In summary, to the extent that word learning depends on the ability to read the social cues of others, children with ASD will face challenges in word learning. We know that these children have difficulties with fast mapping in situations where subtle social-attention cues like eye gaze or more sophisticated social cues to intention hold the key to success. We have little evidence to suggest that the inability to read social cues limits their slow mapping and extension of words; but, then again, we have little evidence about these processes period. We have found that some children with ASD can and do develop age-appropriate knowledge of word meaning, and, fortunately, other investigators have found the same. ^{35 36 37 38 39}  We do not know for certain the best ways to help the children who do lag in the processes of word learning, but the literature demonstrates that children with ASD, like young typically developing children, learn better in environments where the complexity of competing information is reduced and the prominence of important attentional cues is enhanced.

Related Articles

• IAN Research Findings: Speech and Language Therapy
  IAN shares what participating families report about their experiences with speech and language therapy.
• Speech and Language Problems in Autism Spectrum Disorders
  Neurology professor and a father of a teenager with autism, Barry Gordon, M.D., Ph.D., explores the complexities of speech and language problems in children with ASDs.
• Communication and Language Issues
  Part of IAN's Challenging Behaviors series, this article delves into what the research says about the particular speech difficulties that characterize children with ASD.

View References for this Article

4 people have found this article useful.