A Comparison of Implicit and Explicit Discourse Relations in Middle School Science and Social Studies Textbooks

• In Event: 29.082 - Talk That Talk: Making the Most of Text, Speech, and Silence for Emergent Bilinguals in the Classroom

Abstract

Extracting information from textbooks is challenging because readers must understand novel concepts and interpret the language used to convey concepts (Fang & Schleppegrell, 2008). Students must comprehend the coherence relations between discourse segments —those that hold within sentences and across sentence boundaries (Prasad et al., 2008).

Some discourse coherence relations are explicitly signaled via use of discourse markers (e.g. for example) (Fraser, 2006), but some are implicit and not marked at all (Taboada & Mann, 2006). Following Fraser (2006) and an adaptation of the framework provided by The Penn Discourse Treebank [henceforth PDTB] (Prasad et al., 2008), this study explored the proportions of four types of discourse coherence relations (i.e. temporal, inferential, contrastive, and elaborative) between discourse segments in a corpus of textbooks and the proportions in which these relations were explicitly or implicitly marked by discourse markers.

The corpus consisted of twenty life science and nineteen social studies chapters from eighteen California adopted middle school textbooks. The chapters addressed issues of human impact on the environment. All the clauses in the selected chapters were identified and coded. The entire corpus consisted of 127,697 words across 12,936 clauses (7,038 for science and 5,898 for social studies).

A clause-unit-based definition was used to identify adjacent paragraph-internal discourse segments. Two analysts read and coded the same texts until they reached at least 80% agreement. Then, separate regression models were run in order to test whether subject and grade level influence the percentage of each variable analyzed.

Results showed that science textbooks had a higher percentage of clauses with explicit discourse markers (24.5%) than social studies textbooks (23.2%). Although the difference was not significant, this finding suggests that the low frequency of explicit markers is not a feature exclusive to the register of science as has been posited by Halliday and Martin (1993).

Only 20% of all discourse relations were marked via explicit discourse markers, but approximately 80% of the discourse relations were implicit. When an overt connective is absent, readers are expected to identify logical connections between discourse segments.

The findings suggest that strategies only teaching students to look for specific vocabulary features such as discourse markers might not be effective as these occur in only 20% of the clauses analyzed. Thus, we need subject-specific strategies that ask students to comprehend the coherent relations between discourse segments especially when they are unmarked.


References

Fang, Z., & Schleppegrell, M. (2008). Reading in secondary content areas: A language-based pedagogy. Ann Arbor: The University of Michigan Press.

Fraser, B. (2006). Towards a Theory of Discourse Markers. Approaches to Discourse Particles. Ed. Kerstin Fischer. New York: Elsevier. 189-204. Print.

Halliday, M. A. K., & Martin, J. (1993). Writing science: literacy and discursive power. PA: University of Pittsburgh Press.

Prasad, R., Dinesh, N., Lee, A., Miltsakaki, E., Robaldo, L., Joshi, A. & Webber, B. (2008). The Penn Discourse Treebank 2.0. In Proceedings of the 6th International Conference on Language Resources and Evaluation (LREC). Marrakech, Morocco.

Taboada, M., & Mann, W. (2006a). Rhetorical Structure Theory: looking back and moving ahead. Discourse studies, 8(3), 423-459.

Authors

• Diego Roman, Southern Methodist University

• Stephanie Hironaka, Stanford University