The density of unknown words in a passage may also affect cue utilization. In passages
containing a relatively high number of unfamiliar words, context provides less information
about the meaning of any single word, because the context for unknown word x may
contain words у and z unknown as well. Moreover, a high density of unknown words
tends to increase passage difficulty, which in turn may function to discourage the child
from devoting much time and energy to the passage.
Furthermore, the child’s perception of the importance of the unknown word to
understanding the sentence or passage should influence whether the child allocates
resources to figuring out the word’s meaning. The child may perceive as unimportant an
unfamiliar word that occurs in the description of a setting (e.g. scudding clouds), and
unless the child is curious about the word, s/he is unlikely to expend effort in deriving it’s
meaning. On the other hand, when a word is judged as central to the context, then context
cues may be searched out.
When present, the proximity of relevant contextual cues to the unknown word may be
important. Proximity should raise their salience and increase the likelihood that the child
will recognize them as information that is relevant to the unknown word. The meaning of
some words is more difficult to grasp than the meaning of others. For example, words that
have concrete referents are in general easier to comprehend than ones with abstract
referents. Last, individual differences in prior knowledge about the cued information may
determine how and whether different people take advantage of the available cues.
Experimental evidence: Learning from External Context
To test the explanatory power of this theory, Stemberg and Powell (1983) gave high
school students brief passages, written in various literacy styles (e.g. literary, newspaper,
science, historical), containing one or more unfamiliar words. The students’ task was to
define, as best they could, each of the low-frequency words within each passage. The
independent variables were ratings of the number and quality of context cues along with
the mediating factors delineated in the model. Together these factors accounted for
between 72% and 92% of the variance in deriving definitions from context, depending
upon the particular literacy style of the passage.
59