Conditional Promises and Threats – Cognition and Emotion

Conditional Promises and Threats – Cognition and Emotion Sieghard Beller (beller@psychologie.uni-freiburg.de) Department of Psychology, University of Freiburg D-79085 Freiburg, Germany Abstract Conditional promises and threats are speech acts that can be used to manipulate the behavior of other persons. Although reasoning studies have been able to reveal some peculiarities of these concepts, the explanation has remained fragmentary. To fill in this gap, a theoretical analysis of conditional inducements is proposed, which integrates cognitive as well as emotional aspects. An experiment – focussing on linguistic, pragmatic, emo- tional, and deontic consequences – corroborates the anal- ysis and shows that persons have a clear understanding of conditional inducements. Introduction Conditional promises and threats are speech acts (Searle, 1971) uttered by a person to manipulate an addressee’s behavior by setting up consequences of his behavior (cf. Conison, 1997; von Wright, 1962). Walk- ing home from school, Henry may make the following proposal to his classmate Bob: (1) “If you lend me your bike, then I will help you with your homework”. Bob can infer from this statement that Henry would like to borrow his bike, and that Henry believes that he (Bob) needs Henry’s help. Research on human reasoning usually focusses on the inferential use of conditional promises/threats within the normative framework of propositional logic: Which inferences do people draw from conditional arguments? Suppose, (2) Bob lends Henry his bike. What follows then from (1) and (2)? Will Henry help Bob? – Although we cannot know for sure which action a person will actually take, we would expect that Henry will help him, or at least we think that he will have to. This answer is logically warranted by Modus Ponens (MP). But what should be concluded from the promise (1), if (3) Bob does not lend out his bike? Usually, persons answer that Henry will not help Bob either. This inference, however, corresponds to a logically invalid pattern known as Denial of the Antecedent (DA). In general, the valid MP and MT (Modus Tollens) as well as the invalid DA and AC (Affirmation of the Con- sequent) are drawn more frequently from conditional promises and threats than from universal conditionals (e.g., Fillenbaum, 1978; Markovits & Lesage, 1990; Newstead, Ellis, Evans, & Dennis, 1997; but see Evans & Twyman-Musgrove, 1998, for a diverging result). Most people accept the complementary conditional ‘If not-P, then not-Q’ as following from inducements of the form ‘If P, then Q’ (Fillenbaum, 1978). Geis and Zwicky (1971) speak of “invited inferences” in normal linguistic usage. Accordingly, the associated truth tables often reflect an equivalence relation instead of an impli- cation (Newstead et al., 1997). Treating conditional inducements in this way enables one to detect effects of different propositional contents on reasoning, but it is not sufficient to explain the under- lying causes. What is the reason for the DA inference in the introductory example? Furthermore, a simple truth- table analysis is much too restricted to capture concep- tual aspects of conditional inducements beyond the if- then relation. Why, for example, is Henry obliged to help Bob under certain circumstances? How will Bob react emotionally if Henry does not help him? A detailed theoretical analysis is presented to overcome these limi- tations. It integrates several aspects on different levels: goals and incentives on the motivational level, formula- tions and inferences on the linguistic level, obligation and permission on the deontic level, action sequences on the pragmatic level, and finally, affective reactions on the emotional level. The multi-level analysis explains the phenomena observed in reasoning studies; new phe- nomena are predicted and experimentally confirmed. Levels of Conditional Inducements (1) Motivational level: The basic level of analysis con- cerns the motivational situation in which a person utters an inducement. It is determined by expectations, goals and incentives. The speaker (S) wants an addressee (A) to show a certain goal-behavior (i.e., to perform a cer- tain action or to refrain from performing an action) with a positive value for himself, the speaker (S+: Behavior A ). In the introductory example, it was Henry who wanted Bob to lend him his bike. Henry must expect that the addressee is not willing to show this behavior voluntar- ily, otherwise an inducement would not be necessary. Thus, the speaker has to induce a behavioral change: Expected behavior (grey boxes) S–: ¬Behavior A Goal of the speaker S S+: Behavior A This change can be motivated in two ways: First, the speaker may promise to reward the desired goal behav- ior S+ with a positive consequence for the addressee