Spelling Patterns, Letter Cancellation and the Processing of Text

Two experiments are reported in which an attempt is made to identify those parts of words and sentences that readers pay particular attention to during fluent silent reading. Using a technique first reported by Corcoran (1966), we asked university students to cross out all the letter e’s appearing in a text they read. In the first experiment the nature of the text was systematically varied (there were various degrees of difficult, easy, and nonsensical texts); in the second experiment the instructions to subjects were varied (to pay attention to or ignore meaning). The primary results are 1) contrary to Corcoran’s (1966) findings, there is not an acoustic factor in this task (i.e., silent e’s are as readily detected as pronounced e’s); 2) there is a strong tendency, even with the most nonsensical texts, to miss a greater proportion of e’s at the end of words; 3) e’s in unstressed syllables are more likely to be missed than e’s in stressed syllables; 4) there are several grammatical and lexical effects (i.e., the linguistic function of the e to a large extent determines the probability that it will be missed); and 5) the position of a word within a sentence and its position on each line of text also contribute substantially to the probability that any e’s in the word will be missed. The implications of the above results for a theory of reading are discussed. This paper has two aims: to attack the idea that acoustic factors play a significant role in fluent silent reading, and to demonstrate that one of the methods that has been used to investigate acoustic factors is in fact a sensitive tool for studying the micro-structure of the reading process.