Kemper, S. (1988). Inferential complexity and the readability of texts. In A. Davison, & G. M. Green (Eds.), Linguistic complexity and text comprehension: Readability issues reconsidered (pp. 141-165). Hillsdale, England: Lawrence Erlbaum.

Kemper, S.

1988

Kemper, S. (1988). Inferential complexity and the readability of texts. In A. Davison, & G. M. Green (Eds.), Linguistic complexity and text comprehension: Readability issues reconsidered (pp. 141-165). Hillsdale, England: Lawrence Erlbaum.

1

Readability formulas are criticized because they do not consider the contributions of the readers' background or expertise and ignore the meaning and content of texts. A new approach is proposed based on the analysis of texts as causally connected chains of actions, physical states, and mental states. A formula for measuring the inference load of texts is presented which uses multiple regression techniques. The inference load of texts reflects the difficulty readers have in inferring the causal connections necessary to recover the event chains underlying the texts. Using the inference load formula, the difficulty of texts can be adjusted for readers differing in skill or knowledge.

The results confirm both implications of the event-chain approach to analyzing the difficulty of texts: (1) The time required to read a sentence is determined, in part, by whether it refers to an action, physical state, or mental state and, in part, by how easily the reader can causally link the action, physical state, or mental state to preceding ones. (2) Abridging a passage also affects the speed, but not the accuracy, with which factual questions about a text can be answered. Although inference questions are answered more slowly and less accurately than factual questions, text abridgements do not lower readers' accuracy in answering inference questions.

20

2