Evidence of a Preference for Teleological Explanations in Patients With Alzheimer’s Disease

Unlike educated adults, young children demonstrate a ‘‘promiscuous’’ tendency to explain objects and phenomena by reference to functions, endorsing what are called teleological explanations. This tendency becomes more selective as children acquire increasingly coherent beliefs about causal mechanisms, but it is unknown whether a widespread preference for teleology is ever truly outgrown. The study reported here investigated this question by examining explanatory judgments in patients with Alzheimer’s disease (AD), whose dementia affects the rich causal beliefs adults typically consult in evaluating explanations. The results indicate that unlike healthy adults, AD patients systematically and promiscuously prefer teleological explanations, suggesting that an underlying tendency to construe the world in terms of functions persists throughout life. This finding has broad relevance not only to understanding conceptual impairments in AD, but also to theories of development, learning, and conceptual change. Moreover, this finding sheds light on the intuitive appeal of creationism. In 1802, William Paley presented a now classic thought experiment. Paley invited readers to imagine coming across either a stone or a watch. One might legitimately ask why either object exists or how it came to be there, but different explanations seem appropriate for the two objects. For a stone, one might be content to conclude that ‘‘it had lain there for ever’’ (Paley, 1802/ 1998, p. 1). But for the watch, argued Paley, this explanation will not do. This is because the complex coordination of the watch’s components, each essential to the watch’s proper functioning, suggests the existence of an underlying design and an accompanying designer. Applying the same reasoning to humans and other aspects of nature, Paley argued for the existence of an ultimate designer: God. Although contemporary scholars debate the merits of Paley’s argument, his reasoning is intuitively compelling. People typically explain the existence and properties of objects such as stones by appealing to proximate causal mechanisms (e.g., geological processes), and the existence and properties of artifacts such as watches by appealing to their functions (e.g., telling time). When one is confronted with objects that appear to have functions, such as hearts, it seems only natural to adopt a functional, or teleological, mode of explanation: Hearts exist and have the properties they do because they are for pumping blood. Inferring the appropriateness of a teleological explanation from an apparent function, which we call the inference to design, is often quite reasonable. The intricate correspondence between watches and telling time provides evidence that watches were designed for telling time, just as the correspondence between hearts and pumping blood provides evidence that hearts resulted from divine creation (for Paley) or natural selection (for contemporary scientists). But the inference to design is not always valid. Mountains support the function of climbing, yet most adults reject the explanation that mountains exist because they are for climbing (Kelemen, 1999c). This is because adults generally restrict teleological explanations to cases in which the function invoked in the explanation played a causal role in bringing about what is being explained (Lombrozo & Carey, 2006): The fact that watches tell time led to the existence of watches, but the fact that mountains support climbing did not lead to the existence of mountains. The fit between the structure being explained and a plausible function provides evidence that this causal condition holds, but background beliefs (e.g., about the origins of mountains) can override the inference to design and lead to the rejection of a given teleological explanation. In this article, we examine the hypothesis that teleological explanations are compelling and pervasive because they reflect an explanatory default: Unless people have evidence to the Address correspondence to Tania Lombrozo, Department of Psychology, UC Berkeley, 3210 Tolman Hall, Berkeley, CA 94720, e-mail: [email protected]. PSYCHOLOGICAL SCIENCE Volume 18—Number 11 999 Copyright r 2007 Association for Psychological Science contrary, they assume that a good fit between an object’s structure and a plausible function licenses a teleological explanation. This explanatory default can be overridden by background beliefs inconsistent with a teleological explanation, as typically occurs when adults explain the existence and properties of nonliving natural objects, such as stones or mountains. Our hypothesis predicts that people with sparse or compromised background beliefs should err on the side of accepting too many, rather than too few, teleological explanations. Evidence from children, who lack many of the background beliefs that prevent adults from accepting teleological explanations, confirms this prediction. In fact, young children have been characterized as ‘‘promiscuously teleological’’: They overwhelmingly accept and prefer teleological explanations for objects like watches, but also for objects like stones and mountains (DiYanni & Kelemen, 2005; Kelemen, 1999b, 1999c, 2003; Kelemen & DiYanni, 2005; but see Greif, Kemler Nelson, Keil, & Gutierrez, 2006, and Keil, 1992). For example, a majority of 7and 8-year-olds endorsed the explanation that ‘‘mountains exist to give animals a place to climb’’ in preference to the alternative that ‘‘mountains exist because volcanoes cooled into lumps’’ (Kelemen, 1999c). The phenomenon of promiscuous teleology in childhood provides support for the hypothesis that the inference to design is an explanatory default. However, developmental evidence is inconclusive about whether this default persists into adulthood. Moreover, the hypothesis is difficult to test in adults, who have deeply held beliefs about the origins of familiar objects and more general beliefs that constrain explanations of novel objects (Keil, 2003; Lombrozo, 2006; Sloman, Lombrozo, & Malt, 2007). Such beliefs restrict the acceptance and preference for teleology (Kelemen & Rosset, 2007; Lombrozo & Carey, 2006) and could thus mask an underlying preference for teleological explanations. A stringent test of the hypothesis that the inference to design is a lifelong default can be conducted in an adult population with impaired or inferentially weakened causal beliefs. One such population is patients with Alzheimer’s disease (AD), a form of dementia that causes serious semantic and conceptual deficits (Silveri, Daniele, Giustolisi, & Gainotti, 1991; Zannino, Perri, Carlesimo, Pasqualetti, & Caltagirone, 2002). Recently, Zaitchik and Solomon (in press) reported that AD patients mirror young children’s ‘‘Piagetian animism’’ in attributing life to inanimate entities such as airplanes and the sun. This finding suggests that relevant biological beliefs are compromised in AD patients. If their causal beliefs about the origins of objects are also compromised, they should—like children—exhibit a promiscuous tendency to accept and prefer teleological explanations. In the experiment that follows, we tested the prediction that AD patients will exhibit promiscuous teleology. In addition, we examined beliefs about the origins of the objects being explained. Kelemen and DiYanni (2005) found that children often invoked a designer, usually God, to account for the existence of entities like mountains, and that the extent to which individual children accepted teleological explanations correlated with how often they invoked a designer. We were interested in whether Alzheimer’s patients might exhibit a similar tendency to promiscuously invoke a designer.