Toward a theory of the empirical tracking of individuals: Cognitive flexibility and the functions of attention in integrated tracking

Abstract

Abstract How do humans manage to keep track of a gradually changing object or person as the same persisting individual despite the fact that the extraction of information about this individual must often rely on heterogeneous information sources and heterogeneous tracking methods? The article introduces the Empirical Tracking of Individuals (ETI) theory to address this problem. This theory proposes an analysis of the concept of integrated tracking, which refers to the capacity to acquire, store, and update information about the identity and location of individuals in our environment. It hypothesizes that certain functions of attention are a key to explaining how the cognitive flexibility of the human mind overcomes the heterogeneity of sources and methods in integrated tracking. At least two premises lend support to this hypothesis. First, heterogeneity of tracking sources is overcome by the combination of information from multiple perceptual modalities and a phenomenon of multisensory 'transparency'. Second, heterogeneity of tracking sources and methods may also be overcome by inferences that combine information across domains to acquire reasons to believe propositions about the target's location and identity. Keywords: AttentionCognitive FlexibilityCross-domain InferencesFile SemanticsIndividualIntegrated TrackingIdentificationLocationMultisensoryReference Acknowledgments I would like to thank Anina Rich, Mohan Matthen and the referees of Philosophical Psychology for helpful comments about this work. John Sutton and Max Coltheart also provided me with useful discussions on the project of a theory of integrated tracking. The preparation of this article has benefitted from a Fellowship awarded by the Social Sciences and Humanities Research Council of Canada and a Tübitak Fellowship awarded by the Council for Higher Education of Turkey. Notes Notes 1. In this formulation of the problem, the term heterogeneous refers to differences between genera of (1) specialized cognitive systems (or information sources) such as perceptual modules (Carruthers, Citation2006; Coltheart, Citation1999; Fodor, Citation1983), sensory-motor systems or memory systems (Schacter & Tulving, Citation1994); (2) domain-specific knowledge (Hirschfeld & Gelman, Citation1994) and (3) information provided by the use of distinct methods of tracking. The concept of tracking method refers to orderly procedures of identification and localization grounded in learned expertise, the application of scientific knowledge or of technological instruments. For instance, legal identification through Bertillon's system or Galton's system of fingerprints (see below, section 1), biometrical procedures of identification (see below, section 1), localization through radar and satellite systems (see below, section 1) or auscultation (see section 4) are specific tracking methods. 2. In a Kantian tradition, Strawson (Citation1959) has notoriously described this unified scheme as follows: "I have suggested that the fact that material bodies are the basic particulars in our scheme can be deduced from the fact that our scheme is of a certain kind, viz. the scheme of a unified spatio-temporal system of one temporal and three spatial dimensions" (p. 62). The characteristics of the unified spatio-temporal system are also discussed, namely, by Campbell (Citation1993), Evans (Citation1982), and Quinton (Citation1973, pp. 57–80, Citation1979). 3. In the text, the term allocentric refers to a kind of spatial representation that locates places in a framework external to, and independent of, the position of the observer. On the distinction between ego-centered and allocentric spatial representation, see Bloom, Peterson, Nadel, and Garrett (Citation1996), Newcombe and Huttenlocher (Citation2000) and McNamara (Citation2003), O'Keefe and Nadel (Citation1978), and Paillard (Citation1991). 4. For a philosophical discussion of this point, see Evans (Citation1982) and the debate between McDowell (Citation1990) and Peacocke (Citation1983, Citation1991); for empirical evidence on this kind of flexibility in human spatial cognition, see, e.g., Hermer-Vazquez, Spelke, and Katsnelson (Citation1999). 5. On spatial representation, see Eilan, McCarth, and Brewer (Citation1993), Newcombe and Huttenlocher (Citation2000), and O'Keefe and Nadel (Citation1978); on spatial representation and language, see Bloom et al. (Citation1996), Hermer-Vazquez, Moffet, and Munkholm (Citation2001), Hermer-Vazquez et al. (Citation1999), and Landau and Jackendoff (Citation1993). 6. See, e.g., Gattis (Citation2001). 7. For epistemological analyses of the conditions in which perception may provide reasons to believe, see, namely, Brewer (Citation1999), Dretske (Citation1969, Citation1995, Citation2000), Markie (Citation2005, Citation2006). 8. On the notion of truth makers, see Mulligan, Simons, and Smith (Citation1984) and Armstrong (Citation1997). 9. On philosophical ontology or metaphysics, see, e.g., Gaifman (Citation1975), Quine (Citation1960, Citation1969), or Strawson (Citation1959). 10. Pylyshyn (Citation2007) holds that the capacity to individuate and track several independently moving things is accomplished by a mechanism in the early vision module that he calls FINSTs (for "FINgers of INSTantiation"), which functions to track 'visual objects' or 'FINGS'. Pylyshyn holds that FINSTs 'attach to' 'FINGS' through a non-conceptual and causal link but admit that cognitive science has not already completed the task of specifying the ontology of FINGS (see Pylyshyn, Citation2007, pp. 94–98). If his analysis is correct, in the terms of Principle 2, it means that visual FINSTs have a non-conceptual 'ontological commitment', which is to select and refer to FINGS. 11. See Bullot and Droulez (Citation2008), Ganea, Shutts, Spelke, and DeLoache (Citation2007), and Strawson (Citation1959). 12. Perceptual-motor tracking is studied namely in experimental psychology (e.g., Adams, Citation1961; Cavanagh & Alvarez, Citation2005; Craik, Citation1947; Poulton, Citation1952; Pylyshyn & Storm, Citation1988) and the philosophy of demonstrative identification (e.g., Evans, Citation1982; McDowell, Citation1990; Peacocke, Citation1991). 13. I use the concept of episodic memory in Tulving's sense (Schacter & Tulving, Citation1994; Tulving, Citation2002). 14. See, e.g., Gibson (Citation1979) or Warren, Kay, Zosh, Duchon, and Sahuc (Citation2001). 15. This ability to perceive and locate one individual as the same (and distinct from all other individuals) is a form of demonstrative identification in Gareth Evans' sense (1982, pp. 143–203). In his analysis, demonstrative identification depends on the perceptual and conceptual grasp of the properties that distinguish the target object from all other objects of the same kind. This is what Evans (Citation1982, p. 107) terms the fundamental ground of difference of that particular individual, which is reminiscent of Leibniz's Principle of the Identity of Indiscernibles (Hacking, Citation1975; Leibniz, Citation1764/1916). A fundamental identification in Evans' sense is that which correctly attributes a fundamental ground of difference to an individual. A fundamental identification of a material individual (e.g., a perceptual-demonstrative identification) is a form of integrated tracking of an individual because it combines perceptual with epistemic tracking (e.g., deploying the recognitional concept that this individual is your mother and connecting her with a memorized history and some expected future). 16. This argument for this view has its origin in the psychology of vision (Kahneman, Treisman, & Gibbs, Citation1992; Pylyshyn, Citation2007; Treisman, Citation1988, Citation1996); it is adopted in distinct versions by several philosophers (Campbell, Citation2002; Clark, Citation2000, Citation2004; Matthen, Citation2005). 17. The view that the human brain is predominantly visual may be found, under different guises, in Kosslyn (Citation1994), Milner and Goodale (Citation1995), Posner (Citation1994), Pylyshyn (Citation2007) along related philosophical works by Campbell (Citation2002), Dretske (Citation1969), Gendler and Hawthorne (Citation2006), Jacob and Jeannerod (Citation2003), and Matthen (Citation2005). 18. This view is expressed by Campbell (Citation2002, pp. 63–64, 115–116) and Matthen (Citation2005, pp. 282–289, in press); other authors, such as Quinton (Citation1979), hold similar views. 19. In the present text, I use the adjective 'pure' as a qualification of a particular sensory modality (i.e., a presumed uni-modal perceptual system) to indicate that the sensory modality is conceived of in isolation of other sensory modalities, for instance because it is conceived of as a domain-specific module (Coltheart, Citation1999; Fodor, Citation1983). The concept of a pure sensory system is problematic and the notions of 'sense' or 'sensory modality' are especially difficult to characterize. This difficulty has been a long-lasting concern for philosophers (e.g., Keeley, Citation2002; Nelkin, Citation1990; Roxbee-Cox, Citation1970), psychologists (Driver & Spence, Citation1998b; Gibson, Citation1966; Shimojo & Shams, Citation2001; Stoffregen & Bardy, Citation2001) and neuroscientists (Calvert, Spence, & Stein, Citation2004; Churchland, Ramachandran, & Sejnowski, Citation1994). 20. For an experimental study of a task of epistemic tracking of invisible individuals in which subjects must ascribe and update current status predicates, see Bullot and Droulez (Citation2008). 21. As a cognitive function of the faculty of language may be to endow a tracker with cross-domain thinking and cognitive flexibility (Carruthers, Citation2002; Hauser, Chomsky, & Fitch, Citation2002; Hermer-Vazquez et al., Citation1999; Mithen, Citation1996), the possession of language may determine the capacity to perform the integrative tracking of individuals. This hypothesis could justify the intuition that sophisticated forms of integrated tracking are uniquely human. I cannot assess this issue in the limited space of the present article, which focus on the role of attention in integrated tracking. 22. Here recalibrate refers to adjustments performed by the brain to compensate for delays or discrepancies among inputs of distinct sensory modalities and, subsequently, to track the spatio-temporal congruency of physical events or individuals. For instance, detecting the simultaneity of two events across separate sensory channels is a challenge for the brain in audiovisual integration because the temporal congruency at the source of an audiovisual event is polluted by delays in the physical and neural transmission of signals. See, e.g., Fujisaki, Shimojo, Kashino, and Nishida (Citation2004); Vroomen, Keetels, and De Gelder (Citation2004). 23. Introductions to the philosophical debates about transparency are found, namely, in Campbell (Citation2002), Gendler and Hawthorne (Citation2006), Harman (Citation1990), and Martin (Citation2002). 24. Campbell (Citation1993, Citation1995) has proposed to analyze objects or persons as individuals possessing causal connectedness; see also Cassam (Citation1997) for a discussion of Campbell's proposal. 25. It has long been described by phenomenological analyses that attention can undergo voluntary and involuntary shifts (see Hatfield, Citation1998, p. 10, for an historical overview). This has led to the distinction between (i) automatic or reflex and (ii) voluntary attention within various lexical idioms. See James (Citation1890, pp. 416–417), Titchener (Citation1899), or Wundt (Citation1896/1897, pp. 217–218). Experimental cognitive sciences also bring this distinction into play, but use the phrases exogenous attention and endogenous attention respectively (e.g., Driver & Spence, Citation1998a, Citation2004, p. 189; Jones, Citation2001; Spence, Citation2001). 26. See Bregman (Citation1990), Handel (Citation1995), McAdams (Citation2000), McAdams and Bigand (Citation1993). 27. Alain, Arnott, Hevenor, Graham, and Grady (Citation2001), Arnott, Binns, Grady, and Alain (Citation2004), Clarke et al. (Citation2002), Kubovy and Van Valkenburg (Citation2001). 28. Kunkler-Peck and Turvey (Citation2000), Lakatos, McAdams, and Caussé (Citation1997). 29. Audition can inform on events such as vibrations between solids (e.g., scraping, rolling), motions of gases (e.g., exploding balloons, wind), and impacts involving liquids (e.g., splashing or pouring). 30. The argument is thus about what Block (Citation1995, Citation2001) and others term access consciousness in contrast to phenomenal consciousness; the argument presented in the text is neutral with regard to the debates about phenomenal consciousness, and focuses on access consciousness (supra-modal and global accessibility) and reflective consciousness (a special kind of access in which a conscious mental state—e.g., a state of episodic memory—is the object of another state). 31. Posner (Citation1994), Schneider and Shiffrin (Citation1977), Shallice (Citation1988). 32. The workspace model of consciousness has been introduced by Baars (Citation1988)—see also Baars (Citation1997) and Baars, Banks, and Newman (Citation2003)—and is currently developed by Dehaene and his colleagues (Dehaene, Changeux, Naccache, Sackur, & Sergent, Citation2006; Dehaene & Naccache, Citation2001, p. 14). Numerous theories are compatible with the hypothesis of consciousness as an integrative structure; see, namely, Damasio (Citation2000), Dennett (Citation1991), LaBerge (Citation1997), Lycan (Citation1995), Singer (Citation2001), or Tulving (Citation1985). 33. Another reference includes Abeles and Morton (Citation2000). 34. For similar approaches to metacognition, see also Fernandez-Duque, Baird, and Posner (Citation2000), Metcalfe and Shimamura (Citation1994), and Proust (Citation2007).