Research Information System
METU Workshop on Conditional and Causal Reasoning (METU-WCCR), Muğla, Turkey, 22 - 23 March 2023
Conditional presuppositions arise when a presupposition embedded in the consequent of a conditional sentence projects contingent on the truth of the antecedent of the conditional. In (1), for instance, whether John has a surfboard depends on if he is a surfer. Traditional accounts have explained how conditional presuppositions may arise in mainly two different procedures, either by producing the conditional presupposition first (Gazdar, 1979; Geurts, 1999) or producing the non-conditional presupposition first (Beaver, 2001; Heim, 1988). Then, pragmatic considerations determine whether a conditional or a non-conditional presupposition projects (van Rooij, 2007). However, a probabilistic procedure is also possible in determining the final presupposition in which case the prior beliefs probabilistically determine the projection of presuppositions embedded in conditional structures (Lassiter, 2012). Previous experimental research has not considered such an option. This study aims to explore how prior beliefs influence the projection of conditional presuppositions with possessive pronouns as their triggers.
We conducted two rating studies with three types of situations that would make the conditional operator correspond to Karttunen’s (1973) filters (entailing situations), as in (1), and holes (related and unrelated situations), as in (2) and (3) respectively, allowing us to vary the relationship between the antecedent and the embedded presupposition. The truth of an entailing situation makes the possession of a certain object highly likely (e.g., a surfer owning a surfboard), the truth of a related situation makes it moderately likely (e.g., an adventurous person owning a surfboard), and the truth of an unrelated situation does not affect whether someone possesses that object (e.g., someone who likes coffee owning a surfboard).
In Study 1, to obtain individuals’ prior beliefs, we asked participants to rate the probability of someone in a certain situation owning an object, as in (4). The obtained data were analyzed with a linear-mixed model, which showed that prior beliefs about owning an object varied depending on the relationship between the situation and the object (Table 1). Specifically, the falsity of entailing situations significantly lowered the probability of owning the object compared to the falsity of unrelated situations. The ratings for the related situations, on the other hand, did not differ from the ratings for the entailing situations or the unrelated situations, suggesting a gradual influence of prior beliefs.
In Study 2, we adapted the scenarios from Study 1 into conditional sentences in dialogues, as in (5), to test the projection of conditional presuppositions. The results showed a similar gradual pattern of probabilities for the projection of the presupposition when the antecedent was denied (Figure 1). This finding supports that prior beliefs play a crucial role in the projection of conditional presuppositions. Additionally, when the non-conditional presupposition was denied, only entailing antecedents received a very low score, so participants accommodated a conditional presupposition only in conditionals with high probability antecedents (Figure 2).
Overall, our findings show that prior beliefs influence how presuppositions embedded in conditional structures are processed. Only entailing situations lead to conditional presuppositions. This is in line with probabilistic approaches (Lassiter, 2012).