Pragmatic Errors

國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

15

other footwear which are objects that children act upon. Moreover, objects like furniture seldom appear in children‟s early vocabularies because those items are static and children just see them but do not have interactions with them. Thus, Nelson does not agree with Clark‟s (1973) view. She thought that the dominant feature children notice should be function rather than semantic features.

2.1.2 Pragmatic Errors

The second kind of overextension errors happens when children use certain known words to replace the names of some unknown objects. Thus, some researchers pointed out that this may be a strategy for children to compensate for their insufficient vocabulary, and is called „pragmatic errors‟ in the previous literature (Bloom, 1973;

Hoek, Ingram & Gibson, 1986). Hoek, Ingram and Gibson (1986) supported the view that children‟s insufficient vocabulary indeed will affect the phenomena of overextensions. In their study, children overextended known words to unknown words in 33% of comprehension, and 45% of production. Notably, this was just one third of the whole data. Thus, Hoek, Ingram & Gibson (1986) proposed that still other factors may result in children‟s overextension. And there are at least six factors involved as follows:

(1) Some unknown words would be substituted for other known words by children. This is the pragmatic error in question.

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

16

(2) Children would substitute some words acquired earlier for some words acquired later. This is called „retrieval errors‟, which will be introduced in the next error category.

(3) The fact that children have not acquired the whole knowledge of the words‟

criterial features yet would lead to their wrong use of two or more words with similar meanings. And this is so-called the „category error‟ which has been mentioned before.

(4) Children may in some cases use their preferred word to overextend. This phenomenon which was called „Preferred Word Hypothesis‟ in Hoek, Ingram and Gibson‟s study (1986) means that children may have preference toward certain word, thus using it to overextend to various contexts.

(5) Some words are phonologically difficult for children to pronounce. Thus, children may use some phonologically easier words instead of harder ones.

(6) Children may substitute a word for a more natural class than its meaning in adults‟ lexicon. This is the „Semantic Naturalness Hypothesis‟ that Hoek, Ingram and Gibson (1986) proposed in their study. This hypothesis stated that some words should be part of a more natural semantic class than some other words for some reasons which researchers have not been able to find out. And those words with more natural classes would be less vulnerable to be

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

17

overextended.

Accordingly, Hoek, Ingram and Gibson (1986) contended that the factors to cause children‟s overextensions may not be just one, but should be a combination of those factors.

The phenomenon that children may choose a known word to replace another unknown and novel word is because they have finite resources available to say the objects in the world. Therefore, they use an old and acquired word from a different category to refer to another new object despite the fact that they may have the same cognition of this object as adults. However, many researchers thought that some overextensions actually should be classified as a metaphor (Carlson & Anisfeld, 1969;

Bloom, 1973; Thomson & Chapman, 1977; Nelson et al., 1978; Winner, 1979; Billow, 1981; Hudson & Nelson, 1984). When children call apple as ball, they are expressing that apples are like balls (Naigles & Gelman, 1995). This is a different but related phenomenon with children‟s overextensions. Winner (1988) pointed out that it‟s hard to distinguish between an overextended error and a metaphor since they are both

based on similarity. For instance, in Winner‟s study (1988), a child may say the

“skywriting” as “scar” since the skywriting reminded him/her of the scar on his/her

mother. He/she may want to express that the “white line in the sky with its adjacent dots” is very like “the white line on her mother‟s body with its adjacent stitch marks”.

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

18

Thus, this is not a kind of overextension error. Instead, it should be a metaphor based on the perceptual similarity between two objects.

Hence, some criteria are needed to distinguish overextensions from metaphors.

Winner (1979) explained that the differences between an overextension error and metaphor. An overextension usage is from children‟s belief that two objects should have the same name and belong to the same category. For example, children may think that string and tail are the same thing due to their similar appearances. This kind of situation should be an overextension error. On the other hand, if children extend one word to another „on purpose‟ and they know the correct name of both objects respectively, this should be deemed as a metaphor. Besides, Clark (1983) proposed that as children have acquired the correct name of an object as adults, its wrong and overextended name will be abandoned. Thus, an overextended error fills children‟s lexical gap while a metaphor does not. Winner (1988) exemplified that as long as a child knows the name of “apple”, the overextended name “ball” will be no longer used. On the contrary, metaphorical names are still used even if children have acquired a correct name of that object.

Children‟s metaphors can be classified into two categories (Winner, 1979;

Winner, 1988). The first one is “symbolic play metaphors”. This kind of metaphors often happens when children are doing symbolic plays. When playing, they pretend

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

19

some objects just like another one through actions. For example, if a child put his leg into a wastebasket and says „boot,‟ wastebasket becomes a boot because of the child‟s pretending action. However, in other situations, wastebasket will not be a boot.

Furthermore, for the symbolic play metaphor, the main properties on which children based are the object‟s function, or mainly on its function but accompanied with physical similarity at the same time (Winner, 1979; Hudson & Nelson, 1984; Winner, 1988). Thus, the renaming of an object does not necessarily take the features of it into consideration. The features here are neutral for naming.

The second one is “sensory metaphors,” which means that children do metaphors based on perceptual resemblance rather than pretending actions. For this kind of metaphor, children turn their focus from actions to the features of the object itself.

Children can do metaphors just based on the properties of the thing solely rather than with the support of pretending actions. Thus, the features of objects here are essential for naming. And the features on which children based are most frequently shape (Winner, 1988). Taken together, children‟s early metaphors result from either pretend action or physical similarity. The function and sensory resemblance are very important for children to classify the world (Clark, 1973; Nelson, 1974).

Gershkoff-Stowe et al. (2006) pointed out that the overextension and metaphors together show not only that language has open and flexible properties, but also that

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

20

words‟ boundaries are not unchanged. Instead, they undergo a fast and imaginative alteration.