Conversational Implicatures in Counterexample Retrieval: Working Memory and Crystallized Pragmatics Niki Verschueren (Niki.Verschueren@ppw.kuleuven.be) Walter Schaeken (Walter.Schaeken@psy.kuleuven.be) Sara Verbrugge (Sara.Verbrugge@psy.kuleuven.be) University of Leuven, Laboratory of Experimental Psychology, Tiensestraat 102 3000 Leuven – Belgium Abstract Conditional inference patterns are influenced by the availability of counterexamples. We show that aside of the semantic counterexamples that are stored in long-term memory, there is an additional pool of counterexamples that stem from conversational implicatures. Experiment 1 showed that for young adults the retrieval of pragmatic counterexamples requires working memory resources. Experiment 2 showed that for senior adults (aged 58 to 88) working memory resources were significantly lower. Nonetheless, seniors retrieved at least as many pragmatic counterexamples as young adults. As pragmatic counterexamples are key examples of cultural conventions, it is argued that senior adults compensate for limitations in working memory by relying on their acquired cultural expertise. Introduction You can’t live a long life without aging, and aging comes with a cost. Cognitive deficits start occurring at early adulthood and increase with age. Extensive research established robust age-related deficits in working memory functioning (Fisk & Sharp, 2004; Salthouse, 2001) and working memory function is shown to be on its turn a crucial determinant of performance in several cognitively complex tasks (see e.g., Kyllonen & Cristal, 1990). While laboratory tasks involving higher-level cognition show a pronounced age-related decline, seniors perform at level with young adults in everyday life. It is argued that senior adults cope with the cognitive shortcomings by adapting their behavior. We examine the hypothesis that elderly are more versed in cultural conventions and can compensate the atrophy in their cognitive hardware by relying on crystallized conversational pragmatics (Baltes, 1997). One of the capital domains where pragmatics are manifestly present is everyday communication and reasoning. The current research focuses on the use of pragmatic, conversational implicatures in everyday causal conditional inference making. First, we discuss the nature of conversational implicatures and the derivation of pragmatic counterexamples during reasoning. Next, we turn to the working memory involvement for pragmatic counterexample retrieval and compare performance of young adults with senior adult reasoners. Pragmatic counterexamples Grice coined the idea of ‘conversational implicatures’ in 1975 and it is still a hot topic for linguists and psychologists alike (see e.g., Bezuidenhout & Cutting, 2002; Garrod & Pickering, 2004; Levinson, 2000; Noveck, 2001). Conversational implicatures refer to subtle yet strong implicit inference processes that govern all human communication. They arise because the interlocutors customarily obey the conversational maxims governing the efficient co-operative use of language. These maxims state that every speaker should make a contribution that is: 1. adequately – not overly- informative (quantity maxim) 2. build on adequate evidence (quality maxim) 3. relevant (maxim of relevance) 4. clear, brief and orderly (maxim of manner) Every speaker is believed to follow these maxims, which enables listeners to make specific inferences about what is implicated. For example, whenever someone says: ‘There is a man in my office’ this triggers the implicature ‘This man is not the speakers husband, boyfriend or friend’. Following the quantity maxim, the speaker had otherwise provided this more specific information (Levinson, 2000). Despite its compelling nature, the role of these implicatures in reasoning has been widely neglected. We argue that reasoners can rely on conversational implicatures to derive counterexamples in conditional reasoning. Even for causal conditionals —where the felicity conditions relating to the adequate utterance of a speech act are minimal— conversational implicatures play a significant part in the inference process. We briefly discuss the principles of causal conditional reasoning and the role of counterexample retrieval. In conditional reasoning, reasoners are given a conditional sentence as major premise, while the minor premise affirms (denies) the antecedent or consequent clause of the major. We focus on the two affirmation inferences: Modus Ponens (MP) and Affirmation of the Consequent (AC). Some examples: MP: If you water a plant well, then the plant stays green. A plant is well watered. Will it stay green? AC: If a dog has fleas, then it will scratch itself. A dog scratches itself. Does this dog have fleas? Whether reasoners accept the given conclusion depends on their consideration of disabling conditions (MP) or alternative causes (AC). For MP, when reasoners think of ‘deprivation of sunlight’ they do not conclude that the plant stays green. Likewise for AC, when they think of ‘an allergic reaction’ they do not accept that the dog has fleas. Disabling conditions and alternative causes are together referred to as counterexamples. The probability of retrieving a counterexample is related to the number of counterexamples that a reasoner can retrieve from background knowledge. When there are more disablers available, less MP are made, when there are alternatives available, less AC are made. The robust impact of available counterexamples on conditional 2311