importance of constraints for lexical acquisition, how best to characterize their nature and
how do they originate (Behrend 1990; Nelson 1988). Moreover, Keil (1989) proposes that
there are still strong disagreements about whether they are innate or acquired, domain
specific or domain general.
Some scholars such as Nelson (1988, 1990) have argued that the constraint approach is
fundamentally misleading. First of all, constraints have been criticised for failing to
explain how children acquire names for parts, substances, or abstract entities, as well as
the meanings of the verbs, prepositions, determiners and so on.
A second objection focuses on the claim that these constraints are present prior to word
learning, perhaps as a part of a special acquisition device. Many investigators have
suggested that children go through a stage (lasting 6-12 months) where they use words
in ways that violate proposed constraints. Barrett (1986) and Lucarriello & Nelson (1986)
have observed one-year-olds to apply words only in highly restricted contexts; for
instance, only using the word “car” when watching cars move on the street from a certain
location. Children might also use words in “completive” ways; for example, a child might
use the word “clock” to refer to clocks, to dials and timers, to bracelets, to objects that
make buzzing noises, and so on, suggesting that the word refers not to a kind of object,
but to “an associative complex of features” (Rescorla, 1980). Only when these usages
largely disappear does the naming explosion begin.
Nelson (1988) suggests that this is the point when the child seems to have accomplished
the understanding that words name categories of objects and events, which implies that
the constraints are the result of early lexical development. If this were true, then they
clearly cannot serve as an explanation for how children acquire their first words.
Finally, Nelson claims that the results found by Markman and other researchers suggest
more of a bias than a constraint. For example, children’s performance on the type of
forced-choice task used by Markman and Hutchinson (1984) has traditionally been
described as preference for thematic relations over categorical ones. Although Markman
and Hutchinson also speak of children’s conceptual preferences for thematic relations,
they formulate children’s taxonomic choices in terms of constraints. Rather than
4 ■" L ,
i у yΓ'√V,
`` "ɪ