The predictability of implicit causes: testing frequency and topicality explanations

ABSTRACT

In discourses involving implicit causality, the implicit cause of the event is referentially predictable, that is, it is likely to be rementioned. However, it is unclear how referential predictability is calculated. We test two possible explanations: (1) The frequency account suggests that people learn that implicit causes are predictable through experience with the most frequent patterns of reference in natural language, and (2) the topicality account asks whether implicit causes tend to play topical roles in the discourse, which itself may lead to the perception of discourse accessibility. With two text analyses we show that implicit causes are frequently rementioned, but only if we consider a narrow set of discourse circumstances, which would require comprehenders to track contingent frequencies. We found no evidence for the topicality account: in two experiments, implicit causality affected predictability but not topicality, and in a corpus of natural speech, implicit causes tended to not occupy topical positions.

Acknowledgments

We thank Irene Tang for her work on the Google analysis and Michaela Neely, Grant Huffman, Elise Rosa, Simon Wolf, Leela Rao, and Ana Medina Fetterman for their work on the Fisher corpus analysis. We are grateful to UNC’s statistical consulting services for assistance with our models.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. It is also well established that comprehension is facilitated by the predictive activation of sounds, words, and syntactic structures (inter alia Coulson et al., 2005; Falkauskas & Kuperman, 2015; Federmeier & Kutas, 2001; Kochari & Flecken, 2019; Kowalski & Huang, 2017; Kutas & Hillyard, 1984; Levy, 2008; Pickering & Garrod, 2007; Ryskin et al., 2019; Smith & Levy, 2013; Viebahn et al., 2015).

2. For a similar question based on the role of connector words, see Mak et al. (2013).

3. This is comparable to the analysis in Arnold (2001), which included 174 tokens.

4. But see for example, Koornneef and Sanders (2013), Experiment 2; Majid et al. (2006), and Van den Hoven and Ferstl (2018) for a similar stimulus feature.

¹We are very grateful to Hannah Rohde and Andrew Kehler for sharing their coding schema, which is based on the inventory of relations in Kehler (2002).

Additional information

Funding

This work was partially supported by NSF grants 1651000 and 1348549 to J. Arnold and by the 2018 Swarthmore Career Services Summer Experiential Fellowship to S. Guan.