Human thinking beyond the brain

Human Thinking beyond the Brain Frédéric Vallée-Tourangeau Department of Psychology Kingston University Kingston-upon-Thames Surrey, KT1 2EE T: +44 (0)20 8417 2005 F: +44 (0)20 8417 2388 and Stephen Cowley School of Psychology University of Hertfordshire College Lane Hatfield, Hertfordshire, AL10 9AB T: +44 (0)1707 284000 F: +44 (0)1707 284115 1  Abstract It was long assumed that thinking goes on ‘in the head’: indeed, as recently as twenty years ago, many would have regarded it as absurd to examine thinking with reference to events beyond the brain. The chapters in Cognition beyond the brain adopt a different perspective: In thinking, people use dispositions from both sides of the skull. Readily observed phenomena—including neural activity—constitute the object of thinking, which relates conceptually to the construct ‘thinking’. Like all folk concepts, ‘thinking’ is a second-order construct used to ‘explain’ observations or, specifically, how action is — and should be—integrated with perception. As attested in each of the chapters, bodies co-orient to cultural resources while using bidirectional coupling. The focus thus falls on what can be learned about thinking by studying world-side activity. The chapters report empirical, observational and theoretical studies of how people use circumstances (and objects) to act alone, in dyads and in groups. People manage and track attention as they integrate speech and action with gestures, gaze and other bodily activity. In making interactivity part of thinking, a broad range and assortments of parts, procedures and modes of operation are invoked. 2 Thinking in Action It was long assumed that thinking goes on ‘in the head’: indeed, as recently as twenty years ago, many would have regarded it as absurd to examine thinking with reference to events beyond the brain. Not only did behaviourists reject this idea but when the cognitive (counter) revolution arrived, most were enthralled by models that described task performance in terms of computation. Using what philosophers termed ‘functionalism’, this legitimised science based on formal models that were implemented on von Neumann machines. Thus, problem solving, linguistic-analysis and making up 3D visual sketches all came to be pictured as occurring ‘within’ an algorithmic processing system. By the 1990s, however, the climate had changed. It was increasingly recognised that action, perception and attention affect language and thinking. Pursuing this, Cognition beyond the Brain presents studies of how cognitive skills are deployed in a range of complex tasks and activities. While neurally enabled, cultural and bodily dispositions contribute to human action, people exploit sense-saturated coordination or interactivity, a modus operandi based on coordinating with people/objects while orienting to the cultural environment. From this perspective, the heuristic power of symbolic, connectionist, robotic or dynamical models can be separated from normative assumptions. This is done by setting out, not to model hidden states, but to understand how bodies (plural) coordinate. Human interactivity enriches action and collective life by connecting norm-based procedures with the statistical power of information compression. Though people use experience and feeling, they do so in a world where contexts tend to be highly predictable. As a result, a wide range of cultural institutions (e.g., the family, farming, science) exert effects on human brain-body systems. Further, because these evolved to draw on movement and affect, there is no need to reduce thinking to Shannon information that correlates with semantic content. Rather, because thinking uses evaluative functions that have developed over long time scales, formalizations apply to second-order phenomena. Indeed, it is because these are abstracta that that no one claims that the Internet thinks. Rather, like brains, it serves a human community as a cognitive resource. In thinking, people use dispositions from both sides of the skull. Readily observed phenomena — including neural activity — constitute the object of thinking, which relates conceptually to the construct ‘thinking’. Like all folk concepts, ‘thinking’ is a second- order construct used to ‘explain’ observations or, specifically, how action is — and should be — integrated with perception. Our strategy is to seek its basis in adaptive primate behaviour. However, it is also emphasised that, especially when young, human individuals show remarkable flexibility. This appears, above all, in the use they make of language, artefacts and culture. As attested in each of the chapters, bodies co-orient to cultural resources while using bidirectional coupling. The focus thus falls on what can be learned about thinking by studying world-side activity. Though this almost certainly reconfigures plastic brains (e.g., Maguire, Gadian, Johnsrude, Good, Ashburner et al., 2000; Anderson, 2010), that is not the topic of the volume. Rather, the chapters report empirical, observational and theoretical studies of how people use circumstances (and objects) to act alone, in dyads and in groups. People manage and track attention as they integrate speech and action with gestures, gaze and other bodily activity. In making 3 interactivity part of thinking, a broad range and assortments of parts, procedures and modes of operation are invoked. Cognitive Science: The Last 20 Years Work on robotics and the brain shows that models of how sensory input can be processed by von Neumann machines fail to capture the dynamical complexity of self- organizing biological systems. Yet, aspects of thinking can be modelled by algorithms whose strings ‘correspond’ to semantic content. For example, vision can be described by generating a 3D sketch or, alternatively, designing machines that use pattern recognition to act as if they ‘see’. Using Shannon-information produces results that carry functional information for human observers (but not, of course, artificial systems). To suppose that there is a parallel between these models and what brains do is to adopt what Cowley and Spurrett (2004) call an “epistemic conception of mind”. Ultimately, Humean and Cartesian traditions led to epistemic focus on disembodied, disembedded representations that can be modelled on alphabetic or imagistic patterns (i.e., associated with invariant ‘content’). These can function as ‘input’ that is decoded and mapped onto physical symbols that are processed by algorithms that generate various kinds of output. As formal models, such approaches have enormous power. Problems arise from making thinking representational. First, when modelled as information-processing, systems depend on design. Because they are separated from the designer’s world, they face the notorious symbol-grounding problem (see Harnad, 1991; Belpaeme et al., 2009). This is the converse of what we find in nature. There, complex ecosystems sustain living brains, bodies and people. Indeed, representational models even treat sensory systems as akin to computer peripherals. Further, by the 1990s, it had became abundantly clear that, instead of relying on algorithms, robots could use the world as its own representation. They relied on materials—(nonliving) bodies and environments—and modes of engagement that used non-linear mathematics. Much of nature’s complexity depends on flexible brains that adapt as they grow together with moving, perceiving bodies. Unlike programs that use invariant input-patterns, selective processes function to reshape sensorimotor experience. The concept of ‘input’ served to rethink behaviourist theory. While Turing treated computation as extending human powers, psychologists tended to conceptualise computation around actual von Neumann machines. By linking these to the epistemic conception of mind, they were able to ignore the objection that biological systems lack equivalents to keyboards. This is because philosophy often assumes that, while animals are automata, human flexibility arises in dealing with propositional attitudes (or similar). In fact, even brainless systems act flexibly: bacteria can choose between sugar solutions (Monod, 1942; Shapiro, 2011). This is possible because dynamics generate functional information for the bacteria. They depend on mechanisms or effectivities that cannot be reduced to information processing. These systems arose as ecosystems co-evolved with organisms that realize values or, simply, act adaptively. Of course, once brains evolved, new possibilities came into being. Organisms began to link perception with action and, with learning, to use statistical indicators of environmental valences. In starting with biology, mind ceases to resemble the ‘filling’ of the input-output sandwich (Hurley, 2001). As Turing thought, computation extends, but does not explain, human powers 4 (Wells, 2006). An epistemic conception of mind is thus, as Lyon (2006) suggests, anthropogenic. The alternative is to take biology seriously. Although increasingly based on the study of living systems, the shift to biology began with Varela, Thompson and Rosch’s Embodied Mind (1991). Second-generation cognitive science focused attention on lived experience. From a first-person perspective, life is embodied and, in some sense, embedded. Further, if one seeks to avoid dealing with reports of experience, one can turn to bodily ‘modalities’. Given the ambiguities of appeal to ‘embodiment’, many focus on how sense-specific histories contribute to later action and perception. On such an approach, cognition is seen as grounded (Barsalou, 2010): humans become multimodal situated agents. Rather than debate the value of connectionist, robotic or dynamical models, weight falls on how tasks are accomplished. Neuroscience is thus increasingly complemented by work on how organisms (including people) act in the world (Thompson, 2007; Robbins & Aydede, 2009; Chemero, 2009; Stewart et al., 2010). Moving beyond the negative claim that cognition is not brain-bound, new debates have flourished. On the one side, many propose embedded and/or extended functionalism; on the other, another grouping build on The Embodied Mind to propose that cognitive science adopt the enactivist paradigm. Philosophy is plagued by the so-called mind-body problem. In first-generation cognitive science, this was avoided by positing that intelligent behaviour depends on mental states. On a functionalist view, these states (or systems) play a causal role. They can, for example, link a physical injury (input) to the belief that one is in pain (or, rather, a representation of that belief). While traditionally based on input, the model can include the body and, in extended mind, external vehicles. Today functionalists debate whether brains alone carry ‘the mark of the cognitive’ (Adams & Aizawa, 2010). Offering an alternative, Clark and Chalmers (1998) influentially argued that artefacts extend the mind. Using the parity principle they suggest that if external vehicles have the same functional role as mental states, they too serve as part of the cognitive process. However, both sides place the organism at the core of cognitive systems. Indeed, Clark (2008) views this as uncontroversial. Rather than scrutinise how organisms use the environment to anticipate what may happen, the brain is seen as a predictive engine. Further, mechanism is closely identified with function –no attempt is made to decompose these into parts, operations and modes of organization. Functionalists thus retain the classic view that language is a verbal system that serves, among other things, to specify propositions. Like a phonetic alphabet, language consists in symbols whose structure appears in analysis. Written language – and translation– show that language can be seen as a system unto itself –a means of mapping forms onto meanings (and vice versa). On the extended mind view, brains are able to learn the properties of material symbols; for active internalists, as Dennett (1991) phrases it, they ‘install’ a virtual language machine. To the extent that thinking draws on words, therefore, it depends on plugging beliefs into how language is ‘used’ in accordance with grammar and logic. Others emphasise phenomenological experience. While Searle’s (1980) Chinese Room thought experiment challenged functionalism to explain how consciousness arose in biology, more recent work looks elsewhere. Instead of treating mental states (or representations) as biological, the enactivist seeks a common basis for first and third person views. Living systems are taken to possess the functionality that allows cells to engage with the world (though structural coupling), maintain self-organization (by means 5 of internal reorganization) and, where necessary, deal with change (though adaptivity). The enactivist paradigm for cognitive science (see Stewart, Gapenne, & Di Paolo, 2010), has used artificial life in building cognitive models. Their successes mean that discussion of representation and vehicles is giving way to interest in how sensorimotor activity links perception-action with neural function. Recently, this has been linked to participatory sense-making (De Jaegher & Di Paolo; 2007), a form of social behaviour whose complexity resembles that of bacterial communication/cognition (emergent patterns of interaction between agents affect decisions –and trigger unforeseen consequences). In experiments on what is known as perceptual crossing (Auvray et al., 2009) participants engage with three identical stimuli of which one is controlled by a person. Crucially, using interactivity, they identify this more often than would be the case by chance. Cognition connects material entities with how agents orient to each other in what is called a consensual domain (Maturana, 1978). However, it is also possible to allow activity to reiterate linguistic patterns by drawing on motor movements or phonetic gesture. On this view, as people learn to speak, they discover the effects that arise from what Bottineau (2012) calls linguistic techniques. Much work in cognitive science fits neither of these categories. Using cognitive ethnography, Hutchins (1995) proposed functional models of how individuals and groups use artefacts during complex tasks (e.g., in landing a plane). In this, he treated representations as entities that are propagated in public. While often seen as a variant of extended mind, the approach is more radical. This is because Hutchins’s public representations link experience with physical patterns. Far from reducing to material symbols, they are embedded in cultural process. The radicalism comes to the fore when one recognises that non-linguistic experience is bound up with wordings. Thus, while part of action, language is also part of history. This insight shapes a view where, in Love’s (2004) terms, second-order cultural constructs or verbal patterns (‘words’) are perceived as part of first-order activity (or action-perception). During talk, people draw on interactivity to create and construe wordings. First-order language is thus measurable whole-bodied activity that, oddly, evokes second-order patterns (including ‘meanings’). Full-bodied metabolic activity therefore enacts sociocultural patterns. The resulting distributed view of language thus blends with ecological psychology and Chemero’s (2009) ‘radical embodied cognitive science.’ For Cowley and Vallée-Tourangeau (2010), this ‘more subtle’ challenge to the epistemic view of mind builds, in Hollan, Hutchins and Kirsh’s (2000) terms, on how the products of earlier (cultural) events transform later events. In linking neural function and with the slow dynamics of linguistic and cultural change, interactivity makes human cognition central to how people live temporal experience. Systemic Cognition: Making a Start Cognition Beyond the Brain had its beginnings at a symposium on ‘Distributed thinking’ in Hertfordshire in 2009. By connecting thinking with action, participants addressed how cultural organization influences people’s dealings with both objects and each other. In presenting some of the resulting papers in a special issue of AI and Society, Cowley and Vallée-Tourangeau (2010), emphasised that people often manage cognitive tasks by drawing on co-acting assemblages. We hoped that identification of such assemblages could be used to place bounds on thinking in action. The chapters by Baber, 6 Perry, Ben-Naim et al., Spiekermann and Jones all began at the Distributed Thinking Symposium and can be read as describing how cognition spreads beyond the brain. However, a striking problem arose. Above all, the work on computer-mediated trust (see, Ben-Naim et al., 2013) shows that an assemblage based view of distributed cognition reaches far beyond the domain of ‘thinking’. Baber and Jones thus concur that the approach loses sight of what is truly cognitive: In Jones’s (2013) terms, the only active part of a cognitive system is a living human being. Accordingly, in a later symposium at Kingston University (‘Rethinking problem solving’), new importance was given to questions of human agency. Building on this meeting, Ball and Litchfield present evidence supportive of the view, first argued by Kirsh, that interactivity is central to cognition. They show that, in reading X-rays, novices draw on how the image guides real time saccading. They connect what can be seen with the coupling of organism- environment relations. In development, Neumann and Cowley (2013) argue, similar processes result in allowing people (and their brains) to make use of cultural resources. As individual agents become increasingly rational, they learn to participate in a cognitive-cultural network. Brains and genes function in an extended ecology where interactivity contributes to learning to talk and, indeed, prompts discovery of how to seek, identify and solve problems. In placing emphasis on how humans contribute to co-acting assemblages, emphasis fell on interbodily activity (i.e., ways of engaging with objects and people). Having established its importance beyond the skin, we elicited a paper that made comparison with molecular processes (Markos et al., 2013) and one on how a brain-damaged person acts to construct the now (Hemmingsen, 2013). Finally, new papers explore the role of interactivity in language (Steffensen, 2013) and in problem solving (Vallée-Tourangeau & Villejoubert, 2013). Kirsh’s (2013) ‘Thinking with external representations’ shows how interactivity opens up the previously unthinkable. This depends on the remarkably large number of ways in which even simple physical entities can be used as affordances. While they can be seen as a stable form of output or fixed patterns, they also serve to set off incremental processes that create (shareable) thoughts. They promote human projecting and materialising that drive both efficiency/effectiveness and other cognitive processes. Vallée-Tourangeau and Villejoubert cash out Kirsh’s (2013) dictum “there is magic in actually doing something in the world” by illustrating in a range of laboratory-based problem solving tasks that interactivity with a modifiable problem presentation produces unanticipated paths to solution. This augmented creativity defuses mental set and facilitates representation restructuring in overcoming impasse in insight problem solving. Exploring related ideas in a professional setting, Steffensen’s (2013) cognitive event analysis offers thick description of problem solving ‘in the wild’. He shows that what people do and say depends heavily on very rapid events. What he terms ‘insight’ arises as full-bodied activity is concerted in the pico-scale of tens of milliseconds. Like problem-solving, language arises as we manage action and perception, Interactivity thus becomes an ‘ontological substrate’ for language and cognition (Steffensen, 2013: pp). Much depends on how people coordinate and, as a result, manage events (and each other) in different time scales. Baber’s (2013) paper on ‘Distributed cognition at the crime scene’ emphasises change over time. While investigation begins by using various resources to create narratives– as gaze, for example, shapes interactivity, there are gradual changes in the 7 kinds of thinking that are required. In developing wearable technology, there is a need to give ‘structure’ to what occurs in court. Captured affordances, evidence, must sustain objective judgements based on individual ‘mental activity’: implicitly, early on decisions emerge and, later, decisions must be made. Cowley and Vallée-Tourangeau (2013) argue that, like language, thinking has a strange duality in that it too is grounded in activity while managed in line with cultural and organizational constraints. Paradoxically, on the systemic view, certain ‘forms of cognition’ (Loughlin, 2012) depend on how individuals draw on available linguistic and institutional resources. Ball and Litchfield (2013) show much the same for expertise. They demonstrate how interactivity is deeply implicated in gaze. Far from relying on physical invariances, people exploit cues during goal directed behaviour. Like heuristics, these extend adaptive processes. Further while cues can function consciously — as when a pulsating image invites conclusions — they also function in more subtle ways. Hinting at a transfer of ‘intentionality’, they show that the performance of novices who interpret a lung x-ray improve when they are covertly cued by expert gaze. While calling this grounded cognition, simulation of sensorimotor action and input-output processes, Ball and Litchfield (2013) suggest that people draw on external resources by using sensorimotor control to decide how to make the most of what can be seen. Computation, Interactivity and Human Artifice On a systemic view, the focus shifts from modelling hidden mechanisms to the investigation of how results (Järvilehto, 2009), or the products of the organism- environment system, are obtained. Results “embod(y) both preceding organization of the system and its potential for future behaviour and future results” (Järvilehto, 2009, p. 116). The volume presents much empirical work showing that action and perception direct attention to structures whose origins lie in both biological and cultural time-scales. While brains contribute to cognitive tasks, over time, people learn to use cultural products to affect later action and perception. As Järvilehto stresses, joining organs prompt people to anticipate what is likely to occur not only immediately but also in longer scales. Typically, interbodily activity fine-tunes how people engage with each other, artefacts and how language links embodiment with verbal patterns. Although the results can often be explained (or explained away) by the concept of “thinking”, they appear in observable activity. Indeed, we may come to believe in mind and languages through linking a history of acting and perceiving to folk explanations of our doings. In learning to conform/ strategise, or act in rational ways, we discover a partly shared world. Where one understands other peoples’ perspectives, much can be gained by orienting to likely judgements. While this bears on debate between functionalists and enactivists (Cowley & Vallée-Tourangeau, 2013), in this context we focus on simpler matters. Sixty years of computational and robotic modelling show that information- processing models do not clarify how individuals think and perceive. Appeal to Shannon- information and statistics fails to clarify how, in practice, we implement thinking and language. This is because all such models overlook functional information. Living systems depend on, not input (or pure statistics), but bidirectional coupling. For an astute observer, even bacteria choose what to do. On the other hand, computation comes into its own in seeking to understand more abstract patterns (e.g., numbers, syntax). As Turing (1936) thought, it may function by extending cognition (not ‘explaining’ it). Populations, 8 not individuals, use functional (Shannon) information. Not only does this appear in computer generated trust and judgement aggregation but also in second-order aspects of language. Though bound up with action/perception, human populations use verbal patterns with striking predictability. In the extended ecology, eusocial primates draw on a selection history that compresses functional information and, by so doing, makes it increasingly Shannon-like. In the domain of language, at least, these resources allow them to bring their doings under a degree of collective control. Human encounters with the world are embodied and embedded or, more precisely, based in sense-saturated coordination. Interactivity matters to human action and perception. Even if bidirectional coupling is (almost) as old as life itself, humans use felt activity to manage attention and perception. They link bodily encounters with experience of what is likely to happen between people: much thinking has an interbodily basis. As people perceive, they act and, as they act, they perceive: in Gibson’s (1979) sense, learning is the education of attention. Gradually, people discover the duality of language –they shape a flow of first-order activity while drawing on a background of second-order patterns. Using reiterated activity, the phenomena of decision-making are automatized. To do this, however, people need new forms of cognitive control. Individuals develop a sense of how circumstances can be used as they develop a singular identity. People depend on traditions, practices and wordings –modes of life where circumstances evoke cultural products that have the power to alter later activity. The capacity to link circumstances with a flow of acting, feeling and thinking is often called creative. To our knowledge, the only established alternative is Peirce’s (1891) objective idealism. However, Cognition beyond the Brain hints at another view. Novel thoughts and actions are traced to changes in the flow of action. Often, people use the products of past events to set up what Steffensen (2013) calls problem seeking and solution probing. While not goal-directed, this behaviour enables people to coordinate in ways that realise values. They enact reiterated activity but not, of course, perfect repetition: as Heraclitus saw, there is only flux. Accordingly, they satisfice (or improvise). Perhaps this is because brains work on the edge of chaos or, in Wallot and Van Orden’s (2011) terms, in states of continuous frustration. Sometimes felt movement produces valued results; sometimes these arise from inhibition. This may seem miraculous. Of course, the basis lies in experience of using interactivity while orienting to second-order constructs that are evoked by circumstances. If we are correct, our human thinking is inextricable from the history of human artifice. 9  References Adams, F., & Aizawa, K. (2010). Defending the bounds of cognition. In R. Menary (Ed.) The extended mind (pp. 67-80). Cambridge MA: MIT Press. Anderson, M. L. (2010). Neural reuse: A fundamental organizational principle of the brain. Behavioral and Brain Sciences, 33, 245-313. Auvray, M., Lenay, C., & Stewart, J. (2009). Perceptual interactions in a minimalist virtual environment. New Ideas in Psychology, 27, 32-47. Barsalou, L.W. (2010). Grounded cognition: Past, present and future. Topics in Cognitive Science 2, 716-724 Belpaeme, T., Cowley, S. J., & MacDorman, K. F. (2009). Symbol grounding. Amsterdam: John Benjamins Publishing Company. Bottineau, D. (2012). Remembering voice past: languaging as an embodied interactive cognitive technique. In: Proceedings of the Humanities International Forum, Russian State University for the Humanities, Moscow. Chemero, A. (2009). Radical embodied cognitive science. Cambridge MA; MIT Press. Clark, A. (2008). Supersizing the mind: Embodiment, action and cognitive extension. Oxford: Oxford University Press. Clark, A., & Chalmers, D. (1998). The extended mind. Analysis, 58, 7-19. Cowley, S., & Spurrett, D. (2003). Putting apes (body and language) together again. Language Sciences, 25, 289-318. De Jaegher, H., & Di Paolo, E. (2007). Participatory Sense-Making: An enactive approach to social cognition. Phenomenology and the Cognitive Sciences, 6, 485- 507. Dennett, D. (1991). Consciousness explained. Boston: Little Brown. Gibson J. J. (1979). The ecological approach to visual perception. Boston, MA: Houghton Mifflin. Harnad, S. (1990). The symbol grounding problem. Physica D: Nonlinear Phenomena, 42, 335-346. Hollan, J., Hutchins, E., & Kirsh, D. (2000). Distributed cognition: Toward a new foundation for human-computer interaction research. ACM Transactions on Computer-Human Interaction, 7, 174-196. Hutchins, E. (1995). How a cockpit remembers its speeds. Cognitive Science, 19, 265- 288. Hurley, S. L. (2001). Perception and action: Alternate views. Synthese, 129, 3-40. Järvilehto, T (2009). The theory of the organism-environment system as a basis of experimental work in psychology. Ecological Psychology, 21, 112-120. Loughlin, V. (2012). Mark Rowlands, The new science of the mind: from extended mind to embodied phenomenology. Phenomenology and the Cognitive Sciences, 1-7. Lyon, P. (2006). The biogenic approach to cognition. Cognitive Processing, 7, 11-29. Maguire, E. A., Gadian, D. G., Johnsrude, I. S., Good, C. D., Ashburner, J., Frackowiak, R. S. J., & Frith, C. D. (2000). Navigation-related structural changes in the hippocampi of taxi drivers. Proceedings of the National Academy of Sciences, 97, 4398-4403. 10 Maturana, H. (1978). Cognition. In P. Hejl, K. Wolfram, K. Köck, and G. Roth (Eds.), Wahrnehmung und Kommunikation (pp. 29-49). Frankfurt: Peter Lang. Monod, J. (1942). Recherches sur la croissance des cultures bactériennes. Paris: Hermann. Peirce, C. S. (1891). The architecture of theories, The Monist, 1, 161-176 Robbins, P., & Aydade, M. (Eds.). (2009). The Cambridge handbook of situated cognition. Cambridge: Cambridge University Press. Shapiro, J. A. (2011). Evolution: A view from the 21st century. Upper Saddle River NJ: FT Press. Searle, J. R. (1980). Minds, brains, and programs. Behavioral and Brain Sciences, 3, 417- 457. Stewart, J., Gapenne, O., & Di Paolo, E. A. (Eds.). (2010). Enaction: Toward a new paradigm for cognitive science. Cambridge, MA: MIT Press. Thompson, E. (2007). Mind in life: Biology, phenomenology, and the sciences of mind. Cambridge, MA: Belknap Press. Turing, A.M. (1936). On Computable Numbers, with an Application to the Entscheidungsproblem. Proceedings of the London Mathematical Society, ser. 2, 42, 230-265. Varela, F. J., Thompson, E. T., & Rosch, E. (1991). The embodied mind: Cognitive science and human experience. Cambridge MA: MIT press. Wallot, S., & Van Orden, G. (2011). Grounding language performance in the anticipatory dynamics of the body. Ecological Psychology, 23, 157-184. Wells, A. (2006). Rethinking cognitive computation: Turing and the science of the mind. Basingstoke UK: Palgrave Macmillan. 11