Psychologists will probably come to view The Origin of Concepts (Carey Reference Carey2009) as a landmark in the history of psychology, as important as Piaget's (Reference Piaget1954) The Construction of Reality in the Child. Among other virtues, it illustrates how extraordinarily successful the nativist research program in developmental psychology has been since the 1970s.
That said, The Origin of Concepts is not without shortcomings. Here, I focus on its significance for a general theory of concepts. In my view (Machery Reference Machery2009; Reference Machery2010a), philosophers and psychologists have usually focused on two distinct issues (a point Carey acknowledges; pp. 489–91):
-
1.
The philosophical issue: How are we able to have propositional attitudes (beliefs, desires, etc.) about the objects of our attitudes? For example, in virtue of what can we have beliefs about dogs?
-
2.
The psychological issue: Why do people categorize, draw inductions, make analogies, combine concepts, and so forth, the way they do? For example, why are inductive judgments sensitive to similarity?
Psychologists attempt to solve the psychological issue by determining the properties of the bodies of information about categories, substances, events, and so forth, that people rely on when they categorize, make inductions, draw analogies, and understand words.
In the introduction (p. 5) and in the last chapter of The Origin of Concepts (particularly, pp. 487–89, 503–508), Carey claims that the study of conceptual development casts light on the philosophical issue. However, the reader is bound to be disappointed, for Carey does not let the philosophical theory of reference she officially endorses – informational semantics – determine what babies' and toddlers' concepts refer to; instead, she relies on her intuitions to determine what their concepts refer to, and she uses philosophical theories of reference to justify her intuitions.
Two aspects of Carey's discussion support this claim. First, quite conveniently, the philosophical views about reference Carey discusses never lead her to conclude that babies' and toddlers' concepts refer to something different from what she intuitively takes them to refer to.
Second, although Carey endorses an informational semantics, she in fact appeals to several distinct theories of reference, and she switches from one theory to the other when convenient. When Carey discusses the reference of the concept of object, she appeals to Fodor's informational semantics (pp. 98–99), according to which a concept (e.g., the concept of dog) refers to the property that nomologically causes its tokening (e.g., the property of being a dog). Elsewhere (pp. 17, 99), she seems to endorse some (quite unspecified) version of a teleological theory of reference (of the kind developed by Millikan and Neander): a concept refers to a particular property because the evolutionary function of this concept (roughly, what this concept evolved to do) is to be tokened when this property is instantiated. According to this view, the concept of dog refers to the property of being a dog because its function is to be tokened when dogs are perceived (whether perceiving dogs actually causes its tokening or not). Finally, when she discusses the reference of the analog magnitude representations (e.g., pp. 293–95), she implicitly endorses an isomorphism-based conception of reference (of the type developed by Cummins), according to which the reference of our thoughts depends on an isomorphism between the laws that govern them and the laws that apply to some domain of objects. On this view, thoughts about natural numbers are about natural numbers if and only if they obey laws that are isomorphic to the arithmetic operations defined over natural numbers.
These theories of reference are fundamentally distinct. They disagree about what determines the reference of concepts (their current nomological relations with properties outside the mind, their past evolutionary history, or the isomorphism between the laws that govern their use and other laws), and they also occasionally disagree about what concepts refer to. Furthermore, it would do no good to propose to combine these theories into an encompassing theory of reference because it would be unclear why this encompassing theory should be preferred to each of these theories considered on its own. Nor would it do to simply state that different theories of reference apply to different types of representations because one would then need to explain why a particular theory applies to a particular type of concept.
Carey also holds that the study of conceptual development casts some light on the psychological issue (p. 487), but what is curious is that, she overlooks much of the psychological research on concepts (for review, see Murphy Reference Murphy2002; Machery Reference Machery2009), and she promptly dismisses the principal theories of concepts developed by psychologists working on categorization, induction, and concept combination (pp. 496–99). Her main argument is that one needs to distinguish people's concepts from their beliefs or conceptions, that the psychological theories of concepts (prototype, exemplar, and theory theories) were developed to explain categorization, and that research on categorization casts light only on the nature of people's conceptions because categorization is holistic (pp. 490–491, 498). She also holds that psychologists' theories of concepts are descriptivist and that descriptivism is false.
I will briefly deal with Carey's second argument. Psychologists' theories of concepts are not committed to descriptivism because they can be combined with any theory of reference (Machery Reference Machery2010a, p. 235). In addition, Carey's appeal to Kripke's and Putnam's anti-descriptivist views is problematic in light of the cross-cultural variation in intuitions about reference (Machery et al. Reference Machery, Mallon, Nichols and Stich2004; Mallon et al. Reference Mallon, Machery, Nichols and Stich2009; Machery et al. Reference Machery, Olivola and de Blanc2009).
I now turn to the first argument. A theory of concepts that attempts to explain categorization (as prototype, exemplar, and theory theories do) is able to distinguish concepts and conceptions, for one can, and should, distinguish the information that is used by default, in a context-insensitive manner in categorization (people's concepts) from the information used in a context-sensitive manner (their conceptions; Machery Reference Machery2009, Reference Machery2010b). Furthermore, the bodies of information that are used by default in categorization are also used by default in induction, in concept combination, and so forth. For example, typicality effects found in categorization, induction, and concept combination show that prototypes are used in the processes underlying all these cognitive competences. Therefore, the main psychological theories of concepts can not only distinguish concepts from conceptions, they are also essential to solve the psychological issue: explain how we reason, make inductions, draw analogies, in brief explain how we think.
Psychologists will probably come to view The Origin of Concepts (Carey Reference Carey2009) as a landmark in the history of psychology, as important as Piaget's (Reference Piaget1954) The Construction of Reality in the Child. Among other virtues, it illustrates how extraordinarily successful the nativist research program in developmental psychology has been since the 1970s.
That said, The Origin of Concepts is not without shortcomings. Here, I focus on its significance for a general theory of concepts. In my view (Machery Reference Machery2009; Reference Machery2010a), philosophers and psychologists have usually focused on two distinct issues (a point Carey acknowledges; pp. 489–91):
1. The philosophical issue: How are we able to have propositional attitudes (beliefs, desires, etc.) about the objects of our attitudes? For example, in virtue of what can we have beliefs about dogs?
2. The psychological issue: Why do people categorize, draw inductions, make analogies, combine concepts, and so forth, the way they do? For example, why are inductive judgments sensitive to similarity?
Psychologists attempt to solve the psychological issue by determining the properties of the bodies of information about categories, substances, events, and so forth, that people rely on when they categorize, make inductions, draw analogies, and understand words.
In the introduction (p. 5) and in the last chapter of The Origin of Concepts (particularly, pp. 487–89, 503–508), Carey claims that the study of conceptual development casts light on the philosophical issue. However, the reader is bound to be disappointed, for Carey does not let the philosophical theory of reference she officially endorses – informational semantics – determine what babies' and toddlers' concepts refer to; instead, she relies on her intuitions to determine what their concepts refer to, and she uses philosophical theories of reference to justify her intuitions.
Two aspects of Carey's discussion support this claim. First, quite conveniently, the philosophical views about reference Carey discusses never lead her to conclude that babies' and toddlers' concepts refer to something different from what she intuitively takes them to refer to.
Second, although Carey endorses an informational semantics, she in fact appeals to several distinct theories of reference, and she switches from one theory to the other when convenient. When Carey discusses the reference of the concept of object, she appeals to Fodor's informational semantics (pp. 98–99), according to which a concept (e.g., the concept of dog) refers to the property that nomologically causes its tokening (e.g., the property of being a dog). Elsewhere (pp. 17, 99), she seems to endorse some (quite unspecified) version of a teleological theory of reference (of the kind developed by Millikan and Neander): a concept refers to a particular property because the evolutionary function of this concept (roughly, what this concept evolved to do) is to be tokened when this property is instantiated. According to this view, the concept of dog refers to the property of being a dog because its function is to be tokened when dogs are perceived (whether perceiving dogs actually causes its tokening or not). Finally, when she discusses the reference of the analog magnitude representations (e.g., pp. 293–95), she implicitly endorses an isomorphism-based conception of reference (of the type developed by Cummins), according to which the reference of our thoughts depends on an isomorphism between the laws that govern them and the laws that apply to some domain of objects. On this view, thoughts about natural numbers are about natural numbers if and only if they obey laws that are isomorphic to the arithmetic operations defined over natural numbers.
These theories of reference are fundamentally distinct. They disagree about what determines the reference of concepts (their current nomological relations with properties outside the mind, their past evolutionary history, or the isomorphism between the laws that govern their use and other laws), and they also occasionally disagree about what concepts refer to. Furthermore, it would do no good to propose to combine these theories into an encompassing theory of reference because it would be unclear why this encompassing theory should be preferred to each of these theories considered on its own. Nor would it do to simply state that different theories of reference apply to different types of representations because one would then need to explain why a particular theory applies to a particular type of concept.
Carey also holds that the study of conceptual development casts some light on the psychological issue (p. 487), but what is curious is that, she overlooks much of the psychological research on concepts (for review, see Murphy Reference Murphy2002; Machery Reference Machery2009), and she promptly dismisses the principal theories of concepts developed by psychologists working on categorization, induction, and concept combination (pp. 496–99). Her main argument is that one needs to distinguish people's concepts from their beliefs or conceptions, that the psychological theories of concepts (prototype, exemplar, and theory theories) were developed to explain categorization, and that research on categorization casts light only on the nature of people's conceptions because categorization is holistic (pp. 490–491, 498). She also holds that psychologists' theories of concepts are descriptivist and that descriptivism is false.
I will briefly deal with Carey's second argument. Psychologists' theories of concepts are not committed to descriptivism because they can be combined with any theory of reference (Machery Reference Machery2010a, p. 235). In addition, Carey's appeal to Kripke's and Putnam's anti-descriptivist views is problematic in light of the cross-cultural variation in intuitions about reference (Machery et al. Reference Machery, Mallon, Nichols and Stich2004; Mallon et al. Reference Mallon, Machery, Nichols and Stich2009; Machery et al. Reference Machery, Olivola and de Blanc2009).
I now turn to the first argument. A theory of concepts that attempts to explain categorization (as prototype, exemplar, and theory theories do) is able to distinguish concepts and conceptions, for one can, and should, distinguish the information that is used by default, in a context-insensitive manner in categorization (people's concepts) from the information used in a context-sensitive manner (their conceptions; Machery Reference Machery2009, Reference Machery2010b). Furthermore, the bodies of information that are used by default in categorization are also used by default in induction, in concept combination, and so forth. For example, typicality effects found in categorization, induction, and concept combination show that prototypes are used in the processes underlying all these cognitive competences. Therefore, the main psychological theories of concepts can not only distinguish concepts from conceptions, they are also essential to solve the psychological issue: explain how we reason, make inductions, draw analogies, in brief explain how we think.