![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
Abstract
The theory of mind (ToM) framework has been criticised by emerging alternative accounts. Each alternative begins with the accusation that ToM's validity as a research paradigm rests on the assumption of the ‘unobservability’ of other minds. We argue that the critics' discussion of the unobservability assumption (UA) targets a straw man. We discuss metaphysical, phenomenological, epistemological, and psychological readings of UA and demonstrate that it is not the case that ToM assumes the metaphysical, phenomenological, or epistemological claims. However, ToM supports the psychological UA as a claim about cognitive processes responsible for mindreading. The latter can be interpreted as a claim that (a) neither the other's ‘mindedness’ in general nor the other's particular mental states are observable (i.e. apprehended perceptually); (b) particular mental states are unobservable, whereas some aspects indicative of ‘mindedness’ are observable; (c) some mental states are unobservable but some are also observable. Whereas the critics tend to attribute (a) to ToM, most ToM accounts actually take positions (b) or (c). We conclude that the allegations against ToM for positing UA are seriously misdirected. We further bring out an important stipulation of any account of observability of mental states: mental states are not observable in the same way as the sensory properties of physical objects.
Acknowledgements
We would like to thank two anonymous reviewers for their detailed reading and insightful suggestions. We are also grateful to Alexander Stewart Davies, Alan Page Fiske, Bruno Mölder, and Søren Overgaard for their helpful comments on earlier versions of the paper. The work of V.B. was supported by the Volkswagen Stiftung as part of a network grant from the European Platform for Life Sciences, Mind Sciences, and the Humanities, and by the Estonian Science Foundation grant ETF9117. The work of N.G. was supported by Marie Curie Actions- Intra-European Fellowship (FP7-PEOPLE-2011-IEF, project no. 298633).
Notes on contributors
Vivian Bohl is a PhD student at the University of Tartu, Estonia. She has done part of her graduate studies at the Center for Intersubjectivity Research in Copenhagen and at the University of California, Los Angeles. Her current research focuses on the intersection of theories of mindreading and phenomenological approaches to social cognition.
Nivedita Gangopadhyay is a Marie Curie Fellow at the Center for Mind, Brain and Cognitive Evolution, Institut für Philosophie II, Ruhr-Universität Bochum, Germany. Her research focuses on (i) philosophy of mind; (ii) social cognition and intersubjectivity; and (iii) the relation between perception, action, and cognition.
Notes
1. According to Michael (2011, 259–260), interactionism unifies different approaches “by their commitment to the claim that social understanding and interaction do not require mindreading because various embodied and/or extended capabilities sustain social understanding and interaction in the absence of mindreading.” See also Bohl and van den Bos (2012) for an outline of the interactionist position.
2. “[…] a core assumption of both theory theory and simulation theory is that we perceive or experience only the physical movements of another person, since the mental states of others are unobservable“ (Zahavi and Gallagher 2008, 237).
3. By “non-perceptual mechanisms” we mean both mechanisms that do not use perception at all and mechanisms for which perception is merely a way of delivering raw input which then needs to be processed by the non-perceptual mechanism for generating meaningful content.
4. In the context of social cognition, “direct perception” is often used to refer to perception as such, not in contrast to indirect perception, but in contrast to something other than perception (Gallagher 2008a).
5. The roots of ToM are to be traced back to Sellars’ “Empiricism and the Philosophy of Mind” (1956), where mental states are conceived as theoretical entities identified with functional states.
6. Under certain conditions, people tend to attribute mental states to cartoon characters, machines, supernatural beings, and even moving geometrical figures (Heider and Simmel 1944). It is doubtful whether machines have minds or whether supernatural beings exist, while cartoon characters and geometrical figures most certainly lack a mind; however, the practice of mental-state attribution in such cases can be explained by a ToM mechanism.
7. For example, according to the hypothesis of mindshaping (Mameli 2001; Zawidzki 2008), mindreading is not an epistemological endeavour at all: we attribute mental states to others to shape their minds and thus to influence their behaviour rather than to know their mental states. Most proponents of ToM may not agree with the mindshaping thesis, but it illustrates that the issue of the cognitive processes for mindreading and the issue of the epistemological status of our attributions of mental states need to be distinguished.
8. We do not think that the current confusion concerning UA is to be blamed solely on the critics of ToM; in the ToM literature, the position concerning UA is often vaguely expressed, which greatly increases the chances of misunderstanding.
9. We follow the distinction between personal and sub-personal levels of description as outlined in Dennett (1969). The personal level concerns persons and their behaviour, which can be described in intentional vocabulary (e.g. “a person is feeling pain”), whereas the sub-personal level deals with cognitive and neural mechanisms that serve various functions in a cognitive system and that can be described in a non-intentional vocabulary (e.g. “C-fibres are activated”). Whereas explicit ToM operates on the personal level, implicit ToM can serve as a middle ground between the personal level and the sub-personal neural level by giving functional explanations on the sub-personal level in an intentional vocabulary (e.g. “mirror neurons are mechanisms for simulating the actions of another individual”). The exact relation between different levels is a much debated issue, but a detailed account of it is beyond the scope of this paper.
10. For example, Gallagher (2007) claims that what theorists of implicit simulation call “simulation” is not simulation in any genuine sense of the word – neither according to the “pretense definition” (doing something as if doing something else) nor according to the “instrumental definition” (using something as a model for something else). He argues that it is absurd to say that (parts of) the brain simulate(s) another person's mental processes, because only persons can pretend or use models: “neurons either fire or do not fire. They do not pretend to fire” (p. 361). He also contends that the mirror system is neutral with respect to the agent and therefore it is not possible that the mirror neuron system registers my mental states as pretending to be your mental states (Gallagher 2007).
11. Lavelle's example may at first sight seem to speak against the idea of the theory-ladenness of perception. As one of the reviewers pointed out, identifying red spots as shingles may look like a case of postulating a specific cause for these spots that is hidden from view as such. However, when we look at the classic works on theory-ladenness of perception in philosophy of science, such as Hansen (1958) or Kuhn (1962), we discover many similar examples. For instance, Hansen (1958, 22) writes that “We do not ask ‘What's that?’ of every passing bicycle. The knowledge is there in the seeing and not an adjunct of it.” People who have the concept of bicycle tend to see bicycles as rideable objects. This does not mean that every time we see a bicycle, we make an inference that we see a rideable object. However, let us imagine two people seeing a bicycle. One of them comes from a culture that lacks the concept of bicycle, so she has no idea of what bicycles are. Both of them have the same sensory information (presuming that their visual systems function properly), but (as Hansen would argue) they see different things: one of them sees a bicycle; the other one sees an unidentified object of a certain shape. There is no principle difference between this example and the example of a lay person seeing unidentified red spots and an experienced doctor seeing the same spots as shingles.
12. Zahavi (2011, 555) writes, Consider, moreover, the standard argument for why we need to stipulate something like inner simulations or theoretical inferences in the first place. The explanation is typically that we need such processes in order to move from the input which is psychologically meaningless – such as the perception of physical qualities and their changes, say, a distortion of facial muscles – to the output – which is an ascription of mental states, say, joy or happiness, to the other. In short, the processes are needed in order to supplement the input with information coming from elsewhere in order to generate the required output. If phenomenological analysis tells us that the perceptually observed expressive phenomenon is already saturated with psychological meaning and that this is the explanandum, we should reconsider postulating mechanisms supposed to bridge a nonexistent gap.
13. We define “attribution of mindedness in general” as apprehending some entities as potential carriers of at least some mental states. This definition leaves open what the crucial cues indicative of mindedness are, e.g. whether attribution of mindedness is primarily attribution of agency, (motor) intentionality, or something else.
14. The issues concerning neonate imitation are controversial and the results of imitation studies are still open to interpretation (see, e.g. Anisfeld 1996; Jones 2009). Therefore, it might be more adequate to speak about “behavioural matching” rather than “imitation”. However, the thesis of neonatal imitation has been widely accepted within the ToM approach.
15. Gallagher (2008a) claims that ToM in general adheres to a view of perception as a “not-so-smart” process whereas direct perception considers perception to be “smart”. He (2008a, 537–538) argues,[t]he smarter the perception is, the more work it does; the dumber it is, the more it requires extra cognitive processes (theory, simulation) to get the job done. The direct perception theorist is claiming that social perception is very smart and that in the usual circumstances of social interaction it does most of the work without the need of extra cognitive (theoretical or simulationist) processes.He points out that his approach does not deny that while perceiving “all kinds of complex processing is going on in my brain” (Gallagher 2008a), but he rejects the characterisation of these processes as “subpersonal inferences”, thereby implying that the structure of the subpersonal cognitive processes of information organisation for “smart” perception is radically different from that discussed by theories in mainstream cognitive science. But what exactly is the nature of such information organisation? Gallagher leaves the reader wondering by simply saying, “[t]he subpersonal sensory processing that underpins it [direct perception] follows a complex and dynamic route” (Gallagher 2008a).
16. For example, Gallagher (2008a) coined the concept of “smart perception”; Zahavi (2011) extensively discusses “empathy” as immediate knowledge of other minds, being ambiguous about characterising it as perceptual all the way through. However, neither provides an account of any cognitive mechanisms for perceiving other minds. At the same time, some proponents of mindreading (Gallese 2003, 2005) avow the use of Husserlian ideas to develop an account of cognitive, and even neural, mechanisms of social cognition.
17. The English terms “appresence”/“appresentation” and “co-presence”/“co-presentation” are synonymous in Husserlian discussion of intersubjectivity (Welton 2003; Smith 2010)
18. Husserl distinguished between bodies as objects (Körper) and living bodies (Leib) or bodies as subjects. Because of the lack of respective notions in English, in some English translations of Husserl's texts, the Körper/Leib distinction is marked by the use of an uppercase letter for Leib (Body) and lowercase letter for Körper (body).
19. This is not to say that such an account cannot in principle be given without referring to ToM mechanisms. If, for example, one relies on the Gibsonian concept of affordance, one might buy into the above-explained stipulation by stating that the affordances of mind perception are of a unique type. In some more recent papers, e.g. in Gallagher and Povinelli (2012), an enactivist account that goes in this direction is discussed. However, while explaining how the concept of enactive perception could be applied to social cognition, Gallagher gives up his earlier claim (cf. Gallagher 2008a) that perception provides us with an access to others’ mental states. He shifts the central question of social cognition from knowing minds to simply interacting with another agent. Thus, social cognition is no longer about knowing other minds at all; it is merely an interaction between agents, and enactive perception is simply the perception of agency with no reference to mental states. However, the criticism of UA against ToM by direct perceptionism is precisely about how mental states are known and not merely about how agents interact.
20. For example, McNeill (2012) points out the conceptual and empirical problems that beset any attempts to separate perceptual processes from processes of theorising and higher-order belief formation. Taking the scepticism of isolating special cognitive processes of mindreading to a greater length, Apperly (2010) argues that it is virtually impossible to empirically determine whether theorising and/or simulation enable any given instance of mindreading. If one follows Apperly's line of thought, it is no longer only perceptual processes that face a fight to qualify in a unique cognitive role of mindreading without theorising and/or simulation; the whole concept of mindreading by theorising and/or simulation is at stake. In a recent paper, Butterfill and Apperly (forthcoming) have proposed an alternative account of social cognition based on an understanding of the relation between agents, objects, and locations that requires only a minimal ToM. In an earlier paper (Butterfill and Apperly 2009), they outline a “two-systems” account that may be interesting to consider in conjunction with the claim that perception picks up more than sensory properties. A two-system view on mindreading may translate into the claim that system one mindreading is perceptual whereas system two mindreading is not. See also Bohl and van den Bos (2012) for a dual process account of social cognition.