Mandler also writes that a few of the earliest image-schemas infants acquire are the ones of containment, support and contact. Children learn the prepositions in and on very early, and hardly make errors when they use them. Some languages, like Spanish, make no difference between in and on, and other languages, like German, Dutch or Korean, make more complex distinctions (Mandler 1996). For example, Korean has a morpheme for fit together tightly, and Korean children seem to acquire this morpheme effortlessly. Image-schemas seem to be universal: only the way they are ’repackaged linguistically’ (Mandler 1996: 378) varies.
Image-schemas help children by further constraining what is possible to learn in language. By building image-schemas, children are taking another step towards lexical acquisition. Once they are old enough to speak, they will be matching new words with their appropriate image-schemas.
- Constraints and assumptions
Quine’s (1960) ’fundamental problem of induction’ states that ’for any set of data there will be an indefinite number of logically possible hypotheses that are consistent with it’ (Markman 1993: 62). It has been suggested that children are biased to some of these hypotheses. This enables them to map the words they learn onto their pre-existing cognitive structures (or image-schemas, as Mandler (1996) would call them), while eliminating the least probable hypotheses.
- The Taxonomic Constraint
Markman (1993) states that young children are naturally more enclined to look at thematic relations between objects than taxonomic relations. However, they become focused on categorical organisation when they are being taught new words. This is known as the taxonomic constraint.
Thematic relations are more prominent than taxonomic relations: they describe an event happening between the two objects, something a child can easily see. A child might for example put a hand and a mitten in the same category because a mitten is something one puts on one’s hand. However, as Markman says, ’single nouns rarely encode thematic relations’ (1993: 63). Therefore, when confronted with the task of lexical acquisition, the child must put aside his preference for thematic relations, because he knows that they are not possible meanings for words.
A series of studies by Markman and Hutchison (1984) proved the existence of the taxonomic constraint (Markman 1993). They used triads of words: one standard basic-level object (for example, blue jay), one taxonomic choice (duck), and one thematic choice (nest). A puppet would point at the ’standard object’ card and say either (1) ’See this? Find another one that is the same as this.’ or (2) ’See this? It is a sud [or another novel word]. Find another sud that is the same as this sud.’ (Markman 1996: 64-65). It was found that when given a novel word (such as ’sud’), the children picked the card that was the taxonomic choice. When they were not given a novel word, they chose the thematic choice. These studies were repeated with superordinate objects and artificial objects, and the results were very similar (Markman 1996).
The taxonomic constraint theory is not without weaknesses. Are children truly more interested in thematic relations? Barrett (1995) reports results of a case study by Anglin (1983), in which a child overextended the use of the word dog to referents such as lambs, cats, wolves and cows. This seems to indicate that children may be more sensitive to taxonomic relations than one would think. Objects of a same taxonomic category often share perceptual features, and overextensions are mostly based upon these (Barrett 1995). This does not discredit the taxonomic constraint theory, but it is a point to keep in mind. A study by Bauer and Mandler (1989) also showed that young children are capable of classifying objects taxonomically, without having been given a novel word (Markman 1993). Markman (1993) explained these results by pointing out that the children had been pre-trained to categorise objects according to their taxonomic relations, and that the thematic relations they could select were not salient enough for eighteen-month-old children.
Despite this, the taxonomic constraint theory remains well-proven, and it has been speculated that its emergence coincides with the lexical spurt children go through at approximately 18 months of age (Markman 1993, Barrett 1995).
- The Whole Object Constraint
As it has been mentioned before, children tend to assign a novel word to an object which is taxonomically related to another object which they already know. They also tend to apply this new label to a whole object, and not to one of its parts. This is called the whole object constraint, and it has been described by Markman (1993). Markman states that they will only give a label to a part of an object when they already know one word to refer to that object. This will be explained in more detail below.
- The Whole Object Constraint
According to Markman (1993), children tend to assume that words can only have one label. Markman writes that this is especially true at the beginning of a child’s lexical acquisition: once they know a term for an object, they will reject all other possible terms for that object, even though they might be correct. A child may call the family dog ’Rupert’ and refuse to call it by any other name, such as ’the dog’. When children receive enough evidence that more than one term can describe one particular object, they will give up the mutual exclusivity assumption. This assumption helps them narrow down the amount of possible words for one object.
Clark (1995) has her own theory which resembles Markman’s Mutual Exclusivity Assumption. It is called the Lexical Contrast Theory (Markman 1993, Clark 1995, Barrett 1995). There are a few differences between the two. For example, Clark argues that children are used to hearing a variety of terms for a given object, individual or event, and that they fully accept all words. For example, they can hear themselves being referred to as the baby, my son, her grandson, the boy, and other terms (Clark 1995: 406). She states that this goes against the view that children will choose only one word by referrent: in other words, she believes that her theory challenges Markman’s mutual exclusivity assumption. Markman’s answer is that terms such as the baby and my son are not mutually exclusive: the concepts are different, and ’evidence in favour of one is not necessarily evidence in favour of the other’ (Markman 1993: 73).
It could be that children are constrained by the mutual exclusivity assumption only when they produce language. They might realise that ’Rupert’ and ’the dog’ are terms that refer to the same pet, but they make a (probably subconscious) decision to stick with one term until they are more familiar with the production of language. This is a flaw with the theory that is mentioned by Markman (1993). She designed a series of experiments to find out the mutual exclusivity assumption was also present in children’s comprehension.
What Markman did was to give children a novel word and point at an object for which the child already had a label. According to the whole object constraint, the child should assume that the new word is describing the whole object, but this would be violating mutual exclusivity. Therefore, the children applied the new word to a salient part or property of the object. This is evidence for the mutual exclusivity assumption, showing that children are still constrained by it even when they do not produce the word themselves.
Tomasello (2001) questions this theory, saying that it does not explain how the children pick which property or part of the object to which they will apply the new label. An object may have several prominent properties, so what guides them towards a particular one?
I would be interested to know if the results of Markman’s study would have been the same if she, when pointing to an object for which the children already had a label, had given them a synonym instead of a novel word. For example: a child already knows the word ’cup’. Markman points to a cup and says ’Find the mug’. It is possible that the child would try to remember if he had heard the word before: maybe the child’s grandfather always uses ’mug’ to describe this object. I think it would be reasonable to speculate that the child would be able to apply the label ’mug’ to the whole object. This does not mean that he would have started to use it as a synonym: the mutual exclusivity assumption might still rule the production aspect of his language acquisition.
- The acquisition of verbs
The constraints outlined in the past section have mostly been concerned with the acquisition of nouns. When and how do children acquire verbs?
- Are verbs acquired later than nouns?
Much evidence shows that verbs tend to be acquired later than nouns. Barrett (1995) writes that children tend to mostly acquire common nouns in the earliest phase of lexical development. Then, they start acquiring verbs and adjectives as well as more common nouns. Gleitman and Gillette (1995) write that the early vocabularies of children contain very few, if any, verbs. Gentner’s (1982) many studies show the same.
- Why is this the case?
Many reasons can explain this phenomenon. For example, ostensive naming is much more common with nouns than with verbs: parents will more often point at a squirrel and say ’That’s a squirrel!’ than point at somebody performing an action and say ’That’s jumping!’. Moreover, as it has been mentioned before, young children are likely to apply a new word to a whole object: it is probable that they would have difficulty assigning ’jumping’ to the action it describes. Sometimes, other events might be taking place at the same time, and the child must zoom in on the appropriate action, making the task even more complicated. Also, when describing an action, a problem called interleaving arises (Gleitman & Gillette, 1995: 415): many times, an adult will use a verb to refer to an action before or after it has taken place, making it harder for the child to pinpoint exactly what the verb means.
- Syntactic Bootstrapping
An experiment by Naigles (1990) reported in Gleitman and Gillette (1995) proved that young children know the difference between transitive and intransitive verbs. In her experiment, Naigles used two videos: the first one showed a rabbit and a duck waving their arms, and the second one showed the rabbit pushing the duck, making it bend over. Two groups of children were tested. Both videos were shown simultaneously. While watching the videos, group (1) heard ’The rabbit is gorping the duck’ and group (2) heard ’The rabbit and the duck are gorping’ (Gleitman and Gillette 1995: 419). Then, a voice asked them to ’find gorping’. The children in group (1) looked the longest at the video showing the rabbit pushing the duck, and the children in group (2) preferred the video of both animals waving their arms. Naigles concluded that the children’s interpretation of the meaning of ’gorping’ was clearly influenced by the sentence’s syntax. This is an example of syntactic bootstrapping.
Studies by Lederer, Gleitman and Gleitman, described in Gleitman and Gillette (1995) demonstrated that frame-range information, i.e. syntactic information, helps the children the most in determining the meaning of a verb. Syntactic information seems to help children with lexical acquisition, by narrowing down the possible meanings of a given verb.
- Conclusion
In this essay, I have described and discussed the main principles and constraints that faciliate lexical acquisition. Children acquire fundamental knowledge as infants which will help them greatly once they start producing language. They are biased to certain hypotheses, something which makes it easier for them to map a new word onto one of their pre-existing image-schemas. They are also sensitive to syntactic and morphological information. This enables them to pinpoint a verb’s correct meaning. I believe there is more that helps children to acquire a vocabulary than cognitive biases. I tend to agree with Tomasello (2001), who says that while these constraints are useful in lexical acquisition, the role of social interactions should not be underestimated.
2477 words
- Reference list
Barrett, M. (1995) Early Lexical Development. In P. Fletcher & B. MacWhinney (eds.) The handbook of child language. Oxford: Blackwell.
Clark, E. (1995) Later Lexical Development and Word Formation. In P. Fletcher & B. MacWhinney (eds.) The handbook of child language. Oxford: Blackwell.
Gentner, D. (1982) Why nouns are learned before verbs: linguistic relativity versus natural partitioning. In S. Kuczaj (ed.) Language development Volume 2: Language thought and culture. Hillsdale, New Jersey: Lawrence Erlbaum.
Gleitman, L. & J. Gillette (1995) The role of syntax in verb learning. In P. Fletcher & B. MacWhinney (eds.) The handbook of child language. Oxford: Blackwell.
Mandler, J. (1996) Preverbal representations and language. In P. Bloom, M. Peterson, L. Nadel and M. Garrett (eds.) Language and space. MIT Press.
Markman, E. (1993) Ways in which children constrain meaning. In E. Dromi (ed.) Language and cognition: a developmental perspective. New Jersey: Ablex.
Spelke, E. (1994) Initial knowledge: six suggestions. Cognition, 62, 431-445.
Tomasello, M. (2001) Perceiving intentions and learning words in the second year of life. In M. Bowerman & S. Levinson (eds.) Language acquisition and conceptual development. CUP.
