Order of Mention in Causal Sequences: Talking about Cause and Effect in Narratives and Warning Signs

ABSTRACT

Causal sequences can be segmented into cause and effect. However, some argue causal relations in discourse are by default in effect–cause order. Others claim cause–effect order is easier to process and the default way of expressing causality, due to iconicity. We conducted experiments testing participants’ production choices in two different contexts—narratives and safety/warning signs—to see whether genres/discourse types differ in their preferred cause–effect order. We find that while narratives (which involve temporally anchored events) elicit iconic cause–effect order, safety signs (with generic statements rather than specific temporally anchored events) show a bias toward effect–cause. The present work highlights the importance of differences in text type and communicative purpose and suggests that there is no single answer regarding the primacy/salience of cause versus effect.

Acknowledgments

Thanks to Evangeline Alva, Emily Fedele, Josephine Lim, Curran Mahowald, Iris Ouyang, and Caitlin Wilhelm for help with data collection, coding, and experiment setup. Part of this work was supported by a University of Southern California Provost Undergraduate Research Fellowship. Thanks to audiences at the 26th Annual Meeting of the Society for Text & Discourse and the 2016 AMLaP conference, where earlier versions of some of this work were presented. I would also like to thank three anonymous reviewers from Discourse Processes for their helpful feedback and comments.

Notes

1. As noted by Knott et al. (2001), in these kinds of relations we “understand the second part as a conclusion from evidence in the first, and not because there is a causal relation between two states of affairs in the world: it is not because the lights are out that the neighbors are not at home” (p. 202).

2. Some minor wording changes were made, for example, changing MIT to Caltech to be more regionally appropriate to southern California (where the study was conducted). In addition, pronouns were changed to full nouns or names to avoid biasing people toward a certain order.

3. If participants changed the names in the sentences (e.g., item had mentioned Sharon but participant talked about Samantha) or paraphrased the sentences or used close synonyms, these utterances were included in the final dataset.

(i) Shown on screen: Cathy was carried unconscious to a hospital. / Cathy felt very dizzy and fainted at work.

(ii) What participant said: Lisa felt woozy at work and fainted and was then carried to the hospital unconscious.

4. Due to the memory-based nature of the task and the brief wording of the sings, if participants paraphrased the sentences or used synonyms, these utterances were included in the final dataset. Examples are given below:

(a) (i) Shown on sign: do not enter // shock hazard

(ii) What participant said: You should stay away because there’s a shock hazard.

(b) (i) Shown on sign: corrosive chemicals // face shield required

(ii) What participant said: Because of corrosive chemicals, please wear a face mask.