When children are more logical than adults:experimental investigations of scalar implicatureIra A. Noveck*Institut des Sciences Cognitives, Centre National de Recherche Scienti®que, Lyon, FranceReceived 10 December 1999; received in revised form 24 June 2000; accepted 1 September 2000AbstractA conversational implicature is an inference that consists of attributing to a speaker animplicit meaning that goes beyond the explicit linguistic meaning of an utterance. This paperexperimentally investigates scalar implicature, a paradigmatic case of implicature in whicha speaker's use of a term like Some indicates that the speaker had reasons not to use a moreinformative term from the same scale, e.g. All; thus, Some implicates Not all. Pragmatictheorists like Grice would predict that a pragmatic interpretation is determined only after itsexplicit, logical meaning is incorporated (e.g. where Some means at least one). The presentwork aims to developmentally examine this prediction by showing how younger, albeitcompetent, reasoners initially treat a relatively weak term logically before becomingaware of its pragmatic potential. Three experiments are presented. Experiment 1 presentsa modal reasoning scenario offering an exhaustive set of conclusions; critical among these isparticipants' evaluation of a statement expressing Might be x when the context indicates thatthe stronger Must be x is true. The conversationally-infelicitous Might be x can be under-stood logically (e.g. as compatible with Must) or pragmatically (as exclusive to Must).Results from 5-, 7-, and 9-year-olds as well as adults revealed that (a) 7-year-olds are theyoungest to demonstrate modal competence overall and that (b) 7- and 9-year-olds treat theinfelicitous Might logically signi®cantly more often than adults do. Experiment 2 showedhow training with the modal task can suspend the implicatures for adults. Experiment 3provides converging evidence of the developmental pragmatic effect with the French exis-tential quanti®er Certains (Some). While linguistically-sophisticated children (8- and 10-year-olds olds) typically treat Certains as compatible with Tous (All), adults are equivocal.These results, which are consistent with unanticipated ®ndings in classic developmentalpapers, reveal a consistent ordering in which representations of weak scalar terms tend tobe treated logically by young competent participants and more pragmatically by older ones.This work is also relevant to the treatment of scalar implicatures in the reasoning literature.q 2001 Elsevier Science B.V. All rights reserved.I.A. Noveck / Cognition 78 (2001) 165±188 165Cognition 78 (2001) 165±188www.elsevier.com/locate/cognit0010-0277/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved.PII: S0010-0277(00)00114-1COGNITION* Institut des Sciences Cognitives, 67 Boulevard Pinel, 69675 Bron, France. Fax: 133-437-911-210.E-mail address: [email protected] (I.A. Noveck).Keywords: Development; Implicature; Pragmatics; Reasoning1. IntroductionA participant in a laboratory reasoning experiment is comparable to a listener in aconversational exchange. Both are interlocutors who process incoming linguisticcues before eventually offering a reply. It is not surprising then that reasoningresearchers often cite conversational, i.e. pragmatic, factors as sources of variabilityin their experiments. The best known of these factors is arguably conversationalimplicature (or Gricean implicature). This refers to an inference that consists ofattributing to a speaker an implicit meaning that goes beyond the explicit linguisticmeaning of an utterance. While the term implicature refers to a wide range ofphenomena the present paper focuses on implicatures linked to critical logicalterms (in this case, Might and Some). By focusing on a few terms in a structuredcontext one can better determine the in¯uence of implicature in general. This paperendeavors to show how implicatures can be investigated experimentally and tohighlight what such inferences can tell us about cognition.Although the notion of implicature may come up often in the reasoning literature,it is typically treated dismissively in one of two ways: (a) as a kind of high-mindedexplanation for unanticipated responses or; (b) as a phenomenon worth recognizingand then minimizing. For an example of (a), consider the following heading andparagraph from a paper on propositional inferences (Braine, O'Brien, Noveck,Samuels, Fisch, Lea, & Yang, 1995).Nature of unpredicted responses¼A second major category of ªotherº responses can be interpreted as conver-sational implicatures or invited inferences. Under this heading we includedinferences of the following forms:If not p then not q and If q then p from If pthen q, inferences of not both p and q and If p then not q from porq, andinferences of porqand If not p then q from not both p and q. All of these werefairly common.While the variety and extent of implicatures were duly cited, not much more aboutthem was said (for another similar treatment of implicature see Johnson-Laird &Bara, 1984, p. 23).When implicatures have been directly investigated, it is to report that they areultimately not relevant, thus falling into category (b). For example, Newstead (1995)reported that implicatures ªseem to be virtually non-existent in syllogistic reasoningtasksº. Newstead's negative claim about implicatures comes from valid syllogismslike the following:All Nobel Prize winners are eminent scientists.Some Nobel Prize winners are economists.Some economists are eminent scientists.I.A. Noveck / Cognition 78 (2001) 165±188166That is, participants' not concluding that Some economists are not eminent scien-tists, which is arguably an implicature of the conclusion Some economists areeminent scientists, is taken as evidence for the non-existence of conversationalin¯uences in syllogistic reasoning. How Some are not can be inferred from Someand how this inference rears it head will be taken up in the paper. The point for nowis that work like Newstead (1995) leaves the impression that implicatures are notubiquitous in laboratory tasks because participants do not mechanically produceconversational implicatures as errors in syllogistic reasoning tasks.Not all cognitive scientists treat implicatures as phenomena of limited or ques-tionable value. Linguists have been investigating implicature since it was introducedby Grice in the 1967 William James lectures and a small number of