The Effects of Word Order on Sentence- Pair Comprehension

Abstract

<p>This thesis describes four experiments which make an initial attempt to examine differences in difficulty of comprehension of pairs of sentences which all have the same<br />first sentence. The basic paradigm was that of the threeterm series, as used previously in numerous studies of deductive reasoning (e. g., Huttenlocher, 1968; Clark, 1969), but with no reasoning being required: only one sentence of a pair was tested and correct answers did not require the two sentences to be combined in any way.</p> <p>Three dependent measures were used: (1) likelihood of error in answering the question; (2) time to answer the question; and (3) performance on a secondary, interpolated memory task, designed as a measure of processing capacity left over after sentence processing. The last two measures did not produce consistent results but differences between sentence-pairs in terms of likelihood of error were significant and reliable throughout all four experiments. The order of difficulty of sentence- pair s differed only slightly from that preferred by subjects in a creative discourse situation (Morley, 1971), but was substantially different from that found in deductive reasoning studies using the same material.</p> <p>The presence of different kinds of interpolated material did not interfere with this main effect but the presence or absence of the interpolated memory task produced differences in the function relating the likelihood of error to time.</p> <p>The results were interpreted as consistent with the supposition that the process of understanding as a prerequisite for deductive reasoning is not the same process that was investigated here and that deductive reasoning is probably not a two- stage process. In addition, some common features seem apparent between the writer and the reader in terms of surface structure word orders that are on the one hand most preferably produced, and on the other, most easily understood.</p>