The approach described by Cimpian & Salomon (C&S) presents very promising steps towards a better understanding of the cognitive underpinnings of and interindividual differences in intuitive reasoning and explanation seeking by adults. However, it is less convincing as an explanatory approach for the developmental and evolutionary foundations of essentialist reasoning. One of C&S's main claims – that the so-called inherence heuristic (IH) is the developmental foundation for psychological essentialism – has little plausibility for at least two reasons: First, the IH is built on a complex inferential machinery (e.g., the storyteller) that might not be available early in development (and evolution), among other reasons because it appears to rely heavily on linguistic capacities. Second, C&S consider psychological essentialism to be a late-developing phenomenon. In their view, essentialism first appears around age 4, when children master verbal tasks that require them to distinguish between essential and superficial features of animals and other objects. In classical transformation and adoption vignettes (“Will a squirrel painted like a raccoon and growing up among raccoons turn out to be a squirrel or a raccoon?”), children from this age base their judgment of the identity of animals exclusively on essential features and neglect superficial ones (Keil Reference Keil1989).
Unfortunately, however, C&S fail to take notice of earlier cognitive capacities that share some of the essential properties of essentialist reasoning and should thus be considered the ontogenetically and evolutionarily primary forms of psychological essentialism. In particular, even human infants and nonhuman primates engage in a form of object individuation – kind-based or sortal object individuation – that embodies essentialist modes of thinking. Conceptually, sortal object individuation is the capacity to conceive of objects as objects of certain kinds X, Y, Z (dogs, cats, bananas…) by using sortal concepts (“X,” “Y,” “Z” – “dog,” “cat”…) that supply criteria for individuation (“How many Xs are here?”) and identity (“Is this the same X that I saw before?”). Empirically, this capacity has been studied by confronting infants and apes with, say, a box into which an X enters at time 1, followed at time 2 by a Y coming out of the box. Subjects' numerical expectations as to how many objects are in the box are then measured, as indexed in looking and searching behavior: If they think of the objects as objects of distinct kinds that cannot turn into each other, they should expect that there must be (at least) one object, the X, still in the box.
Basic versions of such tasks in which an X (e.g., a banana piece) and a Y (e.g., a carrot piece) are used that differ both in essential and in superficial properties are mastered by human infants from around 12 months (Xu Reference Xu2007) and by nonhuman primates (Mendes et al. Reference Mendes, Rakoczy and Call2008; Reference Mendes, Rakoczy and Call2011; Phillips & Santos Reference Phillips and Santos2007). Because these findings – due to the confound between essential and superficial feature differences – remain somewhat inconclusive regarding the question of whether infants and primates really solve these tasks by using sortal concepts and essentialist reasoning, we recently designed a modified version of individuation tasks. This version was inspired by the classical verbal essentialism tests and enabled us to deconfound essential and superficial property differences: Infants and nonhuman primates saw one object with surface features SF1 enter into a box at time 1 and then at time 2 an object with different surface features SF2 coming out of the box. In fact, however, the object with SF1 was identical to the object with SF2. (For example, in the infant studies, there were stuffed toy animals that could be turned inside out, appearing as a bunny in one form and as a carrot in the other.) The crucial variation was whether the subjects were aware of this. The results showed that those infants and apes that were unaware of the dual identity of the object took the superficial feature differences as diagnostic for their numerical expectations: They thought that there still must be an object in the box – as indicated by their significantly longer searching in the box as compared with events where the superficially identical object with SF1 appeared at time 2. In contrast, those infants and apes that knew about the object's dual identity disregarded the superficial feature differences, did not expect a second object in the box, and did not search differently in the two events. That is, given the requisite background knowledge, infants and apes disregarded the superficial feature differences in very much the same way that older children disregard the superficial feature differences between a normal squirrel at time 1 and a raccoon-looking squirrel at time 2 (after it has been painted etc.) when it comes to the question of the animal's identity (Cacchione et al. Reference Cacchione, Schaub and Rakoczy2013; submitted).
So, what these studies – together with other infant studies (Newman et al. Reference Newman, Herrmann, Wynn and Keil2008) and primate work (Phillips et al. Reference Phillips, Shankar and Santos2010) – suggest is that basic forms of psychological essentialism appear to be present much earlier in development and evolution than assumed by C&S. And this has important implications for the type of account C&S are putting forward – suggesting the following slightly different picture:
-
1. The developmentally and evolutionarily primary forms of psychological essentialism lie in sortal object individuation: Distinguishing between essential properties of an object that determine criteria of identity and countability, and merely superficial features that do not, is already the simplest form of essentialist reasoning. From an evolutionary perspective, such a basic distinction between deep and superficial properties seems to be a fundamental design feature of higher cognition given the need to distinguish between differences in essential (and therefore identity-preserving) properties and merely superficial properties in so many domains (e.g., regarding natural food items that constantly change form and color, or regarding conspecifics that change appearances by growing and decaying).
-
2. Out of this general capacity, then, more specific forms of conceiving of the (essential vs. merely superficial) properties that are relevant for a given kind of object emerge, possibly based on some domain-specific sensitivities to which kinds of properties might be relevant for which type of object.
-
3. Over development, subsequently, the IH – as a broader and more complex cognitive tendency– might emerge out of and hook onto this developmental basis.
The approach described by Cimpian & Salomon (C&S) presents very promising steps towards a better understanding of the cognitive underpinnings of and interindividual differences in intuitive reasoning and explanation seeking by adults. However, it is less convincing as an explanatory approach for the developmental and evolutionary foundations of essentialist reasoning. One of C&S's main claims – that the so-called inherence heuristic (IH) is the developmental foundation for psychological essentialism – has little plausibility for at least two reasons: First, the IH is built on a complex inferential machinery (e.g., the storyteller) that might not be available early in development (and evolution), among other reasons because it appears to rely heavily on linguistic capacities. Second, C&S consider psychological essentialism to be a late-developing phenomenon. In their view, essentialism first appears around age 4, when children master verbal tasks that require them to distinguish between essential and superficial features of animals and other objects. In classical transformation and adoption vignettes (“Will a squirrel painted like a raccoon and growing up among raccoons turn out to be a squirrel or a raccoon?”), children from this age base their judgment of the identity of animals exclusively on essential features and neglect superficial ones (Keil Reference Keil1989).
Unfortunately, however, C&S fail to take notice of earlier cognitive capacities that share some of the essential properties of essentialist reasoning and should thus be considered the ontogenetically and evolutionarily primary forms of psychological essentialism. In particular, even human infants and nonhuman primates engage in a form of object individuation – kind-based or sortal object individuation – that embodies essentialist modes of thinking. Conceptually, sortal object individuation is the capacity to conceive of objects as objects of certain kinds X, Y, Z (dogs, cats, bananas…) by using sortal concepts (“X,” “Y,” “Z” – “dog,” “cat”…) that supply criteria for individuation (“How many Xs are here?”) and identity (“Is this the same X that I saw before?”). Empirically, this capacity has been studied by confronting infants and apes with, say, a box into which an X enters at time 1, followed at time 2 by a Y coming out of the box. Subjects' numerical expectations as to how many objects are in the box are then measured, as indexed in looking and searching behavior: If they think of the objects as objects of distinct kinds that cannot turn into each other, they should expect that there must be (at least) one object, the X, still in the box.
Basic versions of such tasks in which an X (e.g., a banana piece) and a Y (e.g., a carrot piece) are used that differ both in essential and in superficial properties are mastered by human infants from around 12 months (Xu Reference Xu2007) and by nonhuman primates (Mendes et al. Reference Mendes, Rakoczy and Call2008; Reference Mendes, Rakoczy and Call2011; Phillips & Santos Reference Phillips and Santos2007). Because these findings – due to the confound between essential and superficial feature differences – remain somewhat inconclusive regarding the question of whether infants and primates really solve these tasks by using sortal concepts and essentialist reasoning, we recently designed a modified version of individuation tasks. This version was inspired by the classical verbal essentialism tests and enabled us to deconfound essential and superficial property differences: Infants and nonhuman primates saw one object with surface features SF1 enter into a box at time 1 and then at time 2 an object with different surface features SF2 coming out of the box. In fact, however, the object with SF1 was identical to the object with SF2. (For example, in the infant studies, there were stuffed toy animals that could be turned inside out, appearing as a bunny in one form and as a carrot in the other.) The crucial variation was whether the subjects were aware of this. The results showed that those infants and apes that were unaware of the dual identity of the object took the superficial feature differences as diagnostic for their numerical expectations: They thought that there still must be an object in the box – as indicated by their significantly longer searching in the box as compared with events where the superficially identical object with SF1 appeared at time 2. In contrast, those infants and apes that knew about the object's dual identity disregarded the superficial feature differences, did not expect a second object in the box, and did not search differently in the two events. That is, given the requisite background knowledge, infants and apes disregarded the superficial feature differences in very much the same way that older children disregard the superficial feature differences between a normal squirrel at time 1 and a raccoon-looking squirrel at time 2 (after it has been painted etc.) when it comes to the question of the animal's identity (Cacchione et al. Reference Cacchione, Schaub and Rakoczy2013; submitted).
So, what these studies – together with other infant studies (Newman et al. Reference Newman, Herrmann, Wynn and Keil2008) and primate work (Phillips et al. Reference Phillips, Shankar and Santos2010) – suggest is that basic forms of psychological essentialism appear to be present much earlier in development and evolution than assumed by C&S. And this has important implications for the type of account C&S are putting forward – suggesting the following slightly different picture:
1. The developmentally and evolutionarily primary forms of psychological essentialism lie in sortal object individuation: Distinguishing between essential properties of an object that determine criteria of identity and countability, and merely superficial features that do not, is already the simplest form of essentialist reasoning. From an evolutionary perspective, such a basic distinction between deep and superficial properties seems to be a fundamental design feature of higher cognition given the need to distinguish between differences in essential (and therefore identity-preserving) properties and merely superficial properties in so many domains (e.g., regarding natural food items that constantly change form and color, or regarding conspecifics that change appearances by growing and decaying).
2. Out of this general capacity, then, more specific forms of conceiving of the (essential vs. merely superficial) properties that are relevant for a given kind of object emerge, possibly based on some domain-specific sensitivities to which kinds of properties might be relevant for which type of object.
3. Over development, subsequently, the IH – as a broader and more complex cognitive tendency– might emerge out of and hook onto this developmental basis.