Role of Contrastive and Noncontrastive Associates in the Interpretation of Focus Particles

Abstract

The present research examines the mechanisms underlying the comprehension of focus alternatives. In particular, we investigate whether listeners determine alternatives based on general semantic priming mechanisms or whether they only consider contrastive alternatives, elements that can replace the expression in focus. In a probe recognition experiment, we compare the retrieval of contrastive alternatives with that of general associates of a focused expression across conditions with focus particles (only or also) and a condition without a particle. The rationale is that focus particles like only necessarily involve the computation of alternatives and should therefore affect the retrieval of relevant alternatives. The results of the study show that focus particles influence the retrieval of unmentioned contrastive alternatives but not that of general associates. Therefore, we suggest that alternatives are determined by a specialized mechanism that takes into account whether or not an element can replace a focused expression and only these possible replacements are relevant for the processing of focus particles.

Acknowledgments

We are grateful to Anton Benz, Mary Byram-Washburn, Stephen Crain, Jesse Harris, Matthew Husband, Manfred Krifka, Christina Kim, Brian Leahy, Edgar Onea, Uli Sauerland, Stephanie Solt, Carla Umbach, and Duane Watson for helpful comments.

Notes

¹ The experiment reported here is part of the first author's dissertation with some modifications (see Gotzner, 2015).

² This memory representation needs to represent, at the very least, superficial traces of the words in the stimulus material. However, several authors have argued that participants' performance in a probe recognition task not only reflects superficial knowledge of a text (e.g., the features of the text itself) but rather the underlying structure of the events described (e.g., Gernsbacher & Jescheniak, 1995; Glenberg, Meyer & Lindem, 1987; McKoon & Ratcliff, 1980).

³ We also carried out a number of acoustic analyses to compare the pitch maximum, pitch minimum, mean pitch, duration, and relative points of pitch maximum and minimum across conditions. A series of ANOVAs showed no significant differences across conditions (all ps>.28).

⁴ The noncontrastive associates could be related to the focused element in different ways: either by a part-whole relationship, spatial association, produce-material association, or instrument association (see Appendix A for examples).

⁵ An anonymous reviewer pointed out the heterogenity across noncontrastive probes, and we therefore ran an additional analysis of different types of noncontrastive probes (part-whole relation, instrument, product-material, a.o.). The analysis showed that in neither of the different asssociative categories, there was a difference across particle conditions. Table A.1 in Appendix A shows an overview of the different categories.

⁶ However, the fact that probe recognition was slower for noncontrastive associates than contrastive associates in the bare focus condition suggests that noncontrastive associates were primed less than contrastive associates.

⁷ In Braun and Tagliapietra (2010), most contrastive targets were of the same semantic type as the prime words but some of the items used by Husband and Ferreira (2016) compared entities of different semantic types (e.g., SCULPTOR and STATUE).

Additional information

Funding

This work was supported by the German Research Foundation (DFG) as part of the Collaborative Research Centre SFB 632 “Information Structure.” NG is currently supported by the Xprag.de Inititative. We are grateful to Janine Klose, Vinzent Müller, and Linda Giesel for assistance with recording the stimuli and running the experiments.