Skip to main content
Log in

The origins of concepts

  • Published:
Philosophical Studies Aims and scope Submit manuscript

Abstract

Certain of our concepts are innate, but many others are learned. Despite the plausibility of this claim, some have argued that the very idea of concept learning is incoherent. I present a conception of learning that sidesteps the arguments against the possibility of concept learning, and sketch several mechanisms that result in the generation of new primitive concepts. Given the rational considerations that motivate their deployment, I argue that these deserve to be called learning mechanisms. I conclude by replying to the objections that these mechanisms cannot produce genuinely new content and cannot be part of genuinely cognitive explanations.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1

Similar content being viewed by others

Jürgen Jost

Notes

  1. The most prominent recent discussions of concept learning and the arguments against it are those of Cowie (1999), Margolis (Laurence and Margolis 2002; Margolis 1998), and Rupert (2001).

  2. By ‘concepts’ here I mean to restrict the focus only to lexical concepts (e.g., cat, glass, table, etc.). In accordance with most who adopt the Representational Theory of Mind, I treat concepts as a kind of mental representation (in opposition to Fregeans, who treat concepts as extramental objects such as senses).

  3. The most recent attempts at solving the problem of the primitives have been empiricist in character (Barsalou 1999; Prinz 2002). These views attempt to resuscitate the classical notion that all concepts are ultimately composed of copies of perceptual representations. While I won’t take up the empiricist case in detail here, it is not at all clear that all concepts can be adequately reduced to percepts (Weiskopf 2007b, forthcoming). Alternative, non-empiricist views tend just to supplement the perceptual basis with a set of relatively abstract concepts such as cause, event, and agent (Jackendoff 1992a; Miller and Johnson-Laird 1976). Again, it is not at all clear that this simple form of supplementation will do the required trick of capturing the content of all possible lexical concepts.

  4. The ‘environment’ should be broadly interpreted to include other creatures—teachers—who guide the learner by providing specially chosen examples, delivering comprehensible error signals, and drawing attention to contrasts in classes that are not antecedently obvious to the individual.

  5. It’s important to note that this sense of ‘adaptation’ has nothing to do with the term’s use in evolutionary biology. Not every thing learned in the adaptationist sense needs to improve a creature’s fitness or help it live longer or make it happier. It’s perfectly coherent that what is learned might have a negative effect on a creature’s fitness, lifespan, or happiness. You could, for instance, coherently hone your ability to prevent pregnancy (fitness-reducing), or your ability to commit suicide (lifespan-reducing).

  6. A remarkable real-life example along these lines is song acquisition in the juvenile canaries studied by Gardner et al. (2005). If these young birds are isolated from their conspecifics and played samples of artificially generated irregular songs, they can learn to imitate these songs reasonably well. But if they are given a testosterone injection to stimulate the onset of adulthood, their song shifts rapidly over to the normal adult song pattern, although they have not been exposed to this previously. The presence of the hormone, then, appears to induce rearrangement of the song patterns the birds can sing. But they don’t thereby learn their song. (I heard of this study from Ariew (2007, in press).)

  7. The specificity of the occasioning event, however, differs in each case. Imprinting requires a stimulus having certain perceivable characteristics, but may allow for significant variation within broad constraints. Parameter-setting, on the other hand, is highly constrained. It is only by hearing a Head-Initial language, for instance, that one comes to speak such a language. It is interesting, given the relatively constrained nature of the inputs in each case, that both have been held up as good examples of ‘triggering’ processes. Neither truly seems to possess the required degree of arbitrariness that is allowed in genuine cases of triggering. This has sometimes been interpreted as casting doubt on triggering explanations of these phenomena (Cowie 1999; Fodor 1998; Sterelny 1989).

  8. While parameter-setting approaches are commonly adopted by syntacticians working in the Chomskyan tradition, linguists working in different theoretical frameworks may reject the assumption that there exists such a thing as Universal Grammar. I want to leave aside the question of whether this is in fact the right sort of approach to take to syntax acquisition. The point of evaluating the case is just to see whether parameter-setting is a form of learning on the present criterion, not to settle the issue of what linguistic information is innate, or what form that information takes.

  9. This point is made by Jackendoff (1997), who notes that within the Universal Grammar itself there may be both (i) mechanisms for constraining the search space of possible grammars, and (ii) learning mechanisms that settle on a particular grammar given the linguistic evidence. However, he notes that most principles-and-parameters advocates have aimed to “[constrain] the search space of possible grammars, so that the complexity of the learning procedure can be minimized” (p. 6). In the limiting case, nothing resembling learning happens at all. The switches totally determine the grammar. But, as he goes on to note, this is not the only possibility. So there may be space for learning procedures even within UG itself.

  10. While these expressions indirectly represent entities, they are at the same time direct representations of (respectively) the property of being the unique F, or the property of being the local or currently demonstrated F. Direct and indirect representation can come apart, since being G is typically not the same thing as being the F; see Sect. 6.1 for discussion.

  11. This isn’t to imply that primitive concepts represent their content under no mode of presentation. Rather, it is to draw a contrast between complex representations—which necessarily represent their contents in a certain way, as fixed by their compositional structure—and primitives, which, insofar as they lack any such structure, don’t. This distinction is largely independent of one’s view about what actually determines a primitive concept’s MOP. To see this point, consider a conceptual role-based theory, on which a primitive’s MOP is determined by certain of its inferential liaisons. It might be that when a primitive is first introduced, the only such liaison that it possesses links it to its introducing descriptive information. However, we shouldn’t assume that this link is inviolable. Later it might acquire more links to other bodies of information, and might even lose its link to the initial description. This is particularly likely if, as I have been emphasizing, the properties that occasion a concept’s being coined are often fleeting ones. So the only view that is strictly inconsistent with the claim that primitive concepts can present properties in a way that differs from how they are presented by these descriptions is a view on which primitive concepts are permanently tied to the MOP given by their introducing description. But without further argument, I can see little reason to adopt this view. These descriptions may even be long forgotten for most of our concepts. Conceptual role theories of MOPs can, then, accept that primitives may later acquire different MOPs than they begin life with. Complex concepts, however, can’t change their constituents—which are at least partial determinants of their MOP—without changing their identity. For more on content and modes of presentation, see Weiskopf (2007a, forthcoming).

  12. A similar proposal is explored by Rey (1992), although he doesn’t distinguish among these possibilities. He suggests that we imagine the human language of thought to contain an operator like Kaplan’s ‘dthat’ for rigidifying definite descriptions, and that new concepts might be introduced by combining “previously uninterpreted predicates” (p. 323) with such rigidified descriptions. Note, though, that on my account the input description need not contain any such symbol as dthat[the f] or the actual f. Concepts such as dthat and actual may be more sophisticated intellectual achievements. The inputs to coining, then, are descriptions and demonstratives that need not contain references to the actual world. The semantic function of coining captures the rigidifying effect of these concepts without requiring that the concept learner possess them.

  13. See Jeshion (2004) and Reimer (2004) for recent discussion of the semantics of descriptively introduced names.

  14. As Devitt and Sterelny (1999) argue, both causal chains and descriptions may be required for introducing a term with a certain reference.

  15. This account of our epistemic warrant for introducing new concepts is closely paralleled by Kroon’s (1985) account of epistemic warrant for introducing new terms. Kroon emphasizes that term introduction is warranted when we believe that there is some entity that deserves being singled out by a term of its own, and when we take ourselves to be capable of determining further truths about the object to be named. This latter condition, which Kroon calls the Fact-finding condition, dovetails nicely, on the present account, with the emphasis on the role of newly coined concepts in generating hypotheses.

  16. I don’t mean to imply that this reasoning is conscious or that the child could articulate it in precisely the way laid out here, only that it gives a more or less accurate characterization of some of the child’s psychological state as she tries to figure out how to respond in the experimental situation.

  17. Viger (2005) notes that the notion of expressive power is central to the dialectic here. He argues that natural language can expand the expressive power of mentalese by introducing logical operators into it. He seems, however, to agree with Fodor that the expressive power of a creature’s repertoire of predicate concepts cannot be increased. This latter point is the one that I take issue with here.

  18. At least, not the direct product. Intentional acts like getting drunk or hitting myself in the head with a hammer can certainly change the structure of my functional architecture—not always for the better.

References

  • Antony, L. M. (2001). Empty heads. Mind and Language, 16, 193–214.

    Article  Google Scholar 

  • Ariew, A. (2007). Innateness. In M. Matthen & C. Stephens (Eds.), Handbook of the philosophy of biology (pp. 567–584). Amsterdam: North Holland.

  • Barsalou, L. W. (1999). Perceptual symbol systems. Behavioral and Brain Sciences, 22, 577–609.

    Google Scholar 

  • Bateson, P. (1990). Is imprinting such a special case? Philosophical Transactions of the Royal Society of London B, 329, 125–131.

    Article  Google Scholar 

  • Bateson, P. (2000). What must be known in order to understand imprinting? In C. Heyes, & L. Huber (Eds.), The evolution of cognition (pp. 85–102). Cambridge: MIT Press.

    Google Scholar 

  • Bloom, P. (2000). How children learn the meanings of words. Cambridge: MIT Press.

    Google Scholar 

  • Bolhuis, J. J. (1991). Mechanisms of avian imprinting: A review. Biological Reviews, 66, 303–345.

    Article  Google Scholar 

  • Carruthers, P. (2002). The cognitive functions of language. Behavioral and Brain Sciences, 25, 657–674.

    Google Scholar 

  • Cowie, F. (1999). What’s within? Nativism reconsidered. Oxford, UK: Oxford University Press.

    Google Scholar 

  • Devitt, M., & Sterelny, K. (1999). Language and reality (2nd ed.). Cambridge: MIT Press.

    Google Scholar 

  • Fodor, J. (1975). The language of thought. Cambridge: Harvard University Press.

    Google Scholar 

  • Fodor, J. (1981). The present status of the innateness controversy. In Representations (pp. 257–316). Cambridge: MIT Press.

  • Fodor, J. (1994). Concepts: A potboiler. Cognition, 50, 95–113.

    Article  Google Scholar 

  • Fodor, J. (1998). Concepts. Oxford, UK: Oxford University Press.

    Google Scholar 

  • Fodor, J. (2001). Doing without What’s Within: Fiona Cowie’s critique of nativism. Mind, 110, 99–148.

    Article  Google Scholar 

  • Gardner, T. J., Naef, F., & Nottebohm, F. (2005). Freedom and rules: The acquisition and reprogramming of a bird’s learned song. Science, 308, 1046–1049.

    Article  Google Scholar 

  • Gibson, J. J., & Gibson, E. J. (1955). Perceptual learning: Differentiation or enrichment? Psychological Review, 62, 32–41.

    Article  Google Scholar 

  • Goldstone, R. L. (1998). Perceptual learning. Annual Review of Psychology, 49, 585–612.

    Article  Google Scholar 

  • Hampton, J. A. (1995). Similarity-based categorization: The development of prototype theory. Psychologica Belgica, 35, 103–125.

    Google Scholar 

  • Jackendoff, R. (1983). Semantics and cognition. Cambridge: MIT Press.

    Google Scholar 

  • Jackendoff, R. (1992a). What is a concept, that a person may grasp it? In Languages of the Mind (pp. 21–52). Cambridge: MIT Press.

  • Jackendoff, R. (1992b). Word meanings and what it takes to learn them. In Languages of the Mind (pp. 53–67). Cambridge: MIT Press.

  • Jackendoff, R. (1997). The architecture of the language faculty. Cambridge: MIT Press.

    Google Scholar 

  • Jeshion, R. (2004). Descriptive descriptive names. In M. Reimer & A. Bezuidenhout (Eds.), Descriptions and beyond (pp. 591–612). Oxford, UK: Oxford University Press.

    Google Scholar 

  • Juszyk, P. (1997). The discovery of spoken language. Cambridge: MIT Press.

    Google Scholar 

  • Katz, J. J. (1992). The new intensionalism. Mind, 101, 689–719.

    Article  Google Scholar 

  • Kroon, F. W. (1985). Theoretical terms and the causal view of reference. Australasian Journal of Philosophy, 63, 143–166.

    Article  Google Scholar 

  • LaBerge, D. (1973). Attention and the measurement of perceptual learning. Memory and Cognition, 1, 268–276.

    Google Scholar 

  • Landy, D., & Goldstone, R. L. (2005). How we learn about things we don’t already understand. Journal of Experimental and Theoretical Artificial Intelligence, 17, 343–369.

    Article  Google Scholar 

  • Laurence, S., & Margolis, E. (2002). Radical concept nativism. Cognition, 86, 25–55.

    Article  Google Scholar 

  • Levinson, S. C. (2003). Language and mind: Let’s get the issues straight! In D. Gentner & S. Goldin-Meadow (Eds.), Language in mind: Advances in the study of language and thought (pp. 25–46). Cambridge: MIT Press.

    Google Scholar 

  • Margolis, E. (1998). How to acquire a concept. Mind and Language, 13, 347–369.

    Article  Google Scholar 

  • Miller, G. A., & Johnson-Laird, P. N. (1976). Language and perception. Cambridge: Harvard University Press.

    Google Scholar 

  • Millikan, R. G. (2000). On clear and confused ideas. Cambridge: Cambridge University Press.

    Google Scholar 

  • Nazzi, T., & Gopnik, A. (2001). Linguistic and cognitive abilities in infancy: When does language become a tool for categorization? Cognition, 80, B11–B20.

    Article  Google Scholar 

  • Nosofsky, R. M. (1986). Attention, similarity, and the identification-categorization relationship. Journal of Experimental Psychology: General, 115, 39–57.

    Article  Google Scholar 

  • Piattelli-Palmarini, M. (1989). Evolution, selection, and cognition: From “learning” to parameter setting in biology and the study of language. Cognition, 31, 1–44.

    Article  Google Scholar 

  • Prinz, J. (2002). Furnishing the mind. Cambridge: MIT Press.

    Google Scholar 

  • Pustejovsky, J. (1995). The generative lexicon. Cambridge: MIT Press.

    Google Scholar 

  • Pylyshyn, Z. W. (1984). Computation and cognition. Cambridge: MIT Press.

    Google Scholar 

  • Reimer, M. (2004). Descriptively introduced names. In M. Reimer & A. Bezuidenhout (Eds.), Descriptions and beyond (pp. 613–629). Oxford, UK: Oxford University Press.

    Google Scholar 

  • Rey, G. (1992). Semantic externalism and conceptual competence. Proceedings of the Aristotelian Society, 82, 315–334.

    Google Scholar 

  • Roelofs, A. (1997). A case for nondecomposition in conceptually driven word retrieval. Journal of Psycholinguistic Research, 26, 33–67.

    Article  Google Scholar 

  • Rosch, E., Mervis, C. B., Gray, W. D., Johnson, D. M., & Boyes-Braem, P. (1976). Basic objects in natural categories. Cognitive Psychology, 8, 382–439.

    Article  Google Scholar 

  • Rupert, R. (2001). Coining terms in the language of thought: Innateness, emergence, and the lot of Cummins’ argument against the causal theory of mental content. Journal of Philosophy, 98, 499–530.

    Article  Google Scholar 

  • Samet, J. (1986). Troubles with Fodor’s nativism. In P. French, T. E. Uehling Jr., & H. Wettstein (Eds.), Midwest studies in philosophy: Studies in the philosophy of mind (Vol. 10, pp. 575–594). Minneapolis, MN: University of Minnesota Press.

    Google Scholar 

  • Schank, R. C., Collins, G. C., & Hunter, L. E. (1986). Transcending inductive category formation in learning. Behavioral and Brain Sciences, 9, 639–686.

    Article  Google Scholar 

  • Schyns, P. G., Goldstone, R. L., & Thibaut, J.-P. (1998). The development of features in object concepts. Behavioral and Brain Sciences, 21, 1–54.

    Google Scholar 

  • Staddon, J. E. R. (2003). Adaptive behavior and learning. Retrieved 5/1/2007, from http://psychweb.psych.duke.edu/department/jers/abl/TableC.htm

  • Sterelny, K. (1989). Fodor’s nativism. Philosophical Studies, 55, 119–141.

    Article  Google Scholar 

  • ten Cate, C. (1994). Perceptual mechanisms in imprinting and song learning. In: J. A. Hogan & J. J. Bolhuis (Eds.), Causal mechanisms of behavioural development (pp. 116–146). Cambridge: Cambridge University Press.

    Google Scholar 

  • van Kempen, H. S. (1996). A framework for the study of filial imprinting and the development of attachment. Psychonomic Bulletin & Review, 3, 3–20.

    Google Scholar 

  • Viger, C. (2005). Learning to think: A response to the Language of Thought argument for innateness. Mind and Language, 20, 313–325.

    Article  Google Scholar 

  • Waxman, S. R. (1990). Linguistic biases and the establishment of conceptual hierarchies. Cognitive Development, 5, 123–150.

    Article  Google Scholar 

  • Waxman, S. R. (1999). The dubbing ceremony revisited: Object naming and categorization in infancy, early childhood. In D. L. Medin & S. Atran (Eds.), Folkbiology (pp. 233–284). Cambridge: MIT Press.

    Google Scholar 

  • Waxman, S. R., & Gelman, R. (1986). Preschoolers’ use of superordinate relations in classification and language. Cognitive Development, 1, 139–156.

    Article  Google Scholar 

  • Weiskopf, D. A. (2007a). Atomism, pluralism, and conceptual content. Philosophy and Phenomenological Research (forthcoming).

  • Weiskopf, D. A. (2007b). Concept empiricism and the vehicles of thought. Journal of Consciousness Studies (forthcoming).

  • Xu, F. (2002). The role of language in acquiring object kind concepts in infancy. Cognition, 85, 223–250.

    Article  Google Scholar 

Download references

Acknowledgments

I would like to thank Edouard Machery, Rob Rupert, Robert Thompson, Chase Wrenn, and Tad Zawidzki for comments on earlier versions of this material. Thanks also to audience members at George Washington University, Rice University, UC Davis, University of South Florida, and York University, where versions of this paper were read. Finally, an anonymous referee for this journal provided extremely thoughtful and detailed comments that spurred many improvements to the paper, for which I’m grateful.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Daniel A. Weiskopf.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Weiskopf, D.A. The origins of concepts. Philos Stud 140, 359–384 (2008). https://doi.org/10.1007/s11098-007-9150-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11098-007-9150-8

Keywords

Navigation