
doi: 10.1007/bf01067207
pmid: 24197821
The present study investigated the effects of deep, lexical, and surface structure relationships between sentences on judgments of these sentences' semantic similarity. Ten sentence conditions, four paraphrases and six nonparaphrases, were derived from a base sentence. The four paraphrase types weretransformational (T), a passive form of the base,lexical (L), containing synonyms for base content words,formalexic (F), a combination ofT andL types, andparasyntactic (P), one of several alternative interpretations of the base. The six nonparaphrases consisted of three sets of two sentences each: the falsepermutation sentences retained the base lexicon, thefalse synonymous sentences contained synonyms, and theunrelated sentences' lexicon was completely unrelated to the base. One sentence in each nonparaphrase set retained the base surface form and the other, a passivization, did not. Using a modified paired comparisons task, the following rank order of conditions, in terms of preference, was obtained:T>L>F>P>false permutation>false synonymous>unrelated. It was concluded that deep structure similarity had potent effects but that a more complete description of the data required the postulation of additional factors such as "propositional structure" and "semantic structure."
| selected citations These citations are derived from selected sources. This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | 6 | |
| popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network. | Average | |
| influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | Top 10% | |
| impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network. | Average |
