These two major routes to recognizing words, whole-word vs lettersound sequence, can be demonstrated. In one experiment, Grades II, IV, VI schoolchildren and college students were asked to judge whether a set of English words belonged to animal or non-animal categories (Samuels et al. 1978). Words varied in length from three to six letters (e.g., hog, pony, whale, cattle). The younger the children, the more slowly they judged the words. Further, the young children's judging times increased as the word length increased, suggesting that they recognized words as letter sequences. But college students' judging times were more or less the same, independent of word length, suggesting that they recognized a word as a whole pattern, regardless of its length. Note that the longest test words were reasonably common and short, as compared with words such as arachnophobia and tachistoscope, for which even college students are less likely to use whole-word recognition.
Logographic Chinese characters and phonetic English words appear to be learned in a different manner. In a cross-cultural study, Stevenson and Stigler (1992) compared the ability of first graders in Beijing and in Chicago to read words. The Chicago children were over-represented at the top and bottom groups; that is, some could read words above their grade level, and others below it, suggesting that English-speaking children who catch on how to break down words by sound can read new words not yet taught, whereas those who do not acquire this decoding ability tend to be poor readers. Once they can sound out new words, they can understand the meanings of these words, which are likely to be in their oral vocabulary. By contrast most Beijing children read at their grade level: they can read words already taught but not those not yet taught, because they cannot sound out new words in Chinese characters. Figure 6-1 shows the two different patterns of word reading by Chinese and American first graders.
Figure 6-1. Chinese and American first graders'
ability to read words at andabove different grade levels (Stevenson and
Stigler 1992:47; included in the book with permission)
If the character 
means 'mountain' and is given the
sound shan (in Mandarin) or yama (in one of several Japanese
sounds associated with it), then a reader has to recognize it as a whole
pattern; there is no point in decomposing it into a series of strokes, which
would represent neither its sound sequence nor its meaning. In learning
simple characters, each character as a whole pattern is associated with
its tone syllable and morpheme. But characters that are composites of two
or more simple characters (e.g., 
'person + mountain' ='hermit')
may be analyzed into components, each of which is often a simple character
that is learned as a whole pattern.
So, a logographic Chinese character, or a component that makes up a composite character, is recognized as a whole pattern, whereas an English word may be recognized as a whole pattern or as a sequence of lettersounds, depending on, among other things, its familiarity to a reader.
<-Previous sample | Table of Contents | Next Sample->
