A re-evaluation of story grammars

Abstract

Black and Wbnsky (1 979) haw mode serious methodological errors in analyzing story grammars, and in the process they haw committed additional errors in applying formal language thswy. Ow arguments imdw clarifying certain as- Particular criticisms focus on the following shwtcomings of their prosentotiom 1) an error#ouI statement from formal langwga hoary, 2) misapplication of formal language theory to story grammars, 3) unsubstontioted and doubrfvl analogies with English grammar, 4) various llon sequiturs concerning tha Q.cHration of non-stories, 5) a false claim bawd on the artificial distinciion belwoen syntax and semantics, and 6) misinterpretation of the role of story grammors in story understanding.We conclude by suggesting appropriate criteria for the evaluation of story grammars.John Black and Robert Wilensky (1979) (henceforth B & W) argue against story grammars as germane to the study of story understanding. Their arguments strike us as seriously inadequate and in several instances based on a fallacious application of results from the theory of formal languages. It seems to us that lack of a well-defined specification of the story-grammar approach in their article has led to a confused methodological 'outlook which we hope to clarify. Our criticism will be organized into the following parts:1. We review some theory of formal languages and, in particular, show how languages and grammars are related. This is then used to correct a theorem misstated by B & W. 2. We point out several instances where B & W have misapplied the theory of formal languages in attempting to show that certain story grammars are inadequate.

Language, Culture, Computation. Computing of the Humanities, Law, and Narratives

“…1975 Propp envisaged twenty-two tale functions, and nine more in the "preparatory section" of a tale, a part that is sometimes absent. 12 From the controversy concerning story grammars in cognitive science (which was influential in AI), see Black & Wilensky (1979), Rumelhart (1980b), Frisch & Perlis (1981), Wilensky (1983b), and Garnham (1983). Not always Propp's transformation function is intended as a disguise disabling recognition on the part of other characters.…”