Joint attention and communication

ABSTRACT

Joint attention occurs when two (or more) individuals attend together to some object. It has been identified by psychologists as an early form of our joint engagement, and is thought to provide us with an understanding of other minds that is basic in that sophisticated conceptual resources are not involved. Accordingly, it has also attracted the interest of philosophers. Moreover, a very recent trend in the psychological and philosophical literature on joint attention consists of developing the suggestion that it holds partially in virtue of communication: it is because we share our thoughts or feelings about an object that our individual attention becomes joint. This paper unpacks the communicative suggestion in a way that accounts for joint attention's basicness.

KEYWORDS:

• Joint attention
• joint action
• Griceanism
• common knowledge
• theory of mind

Acknowledgments

I am grateful to Nate Charlow, Brendan de Kenessey, Imogen Dickie, Evan Westra, and an anonymous referee for comments on earlier versions of this paper. I am also grateful to John Campbell, Johannes Roessler, and Hong Yu Wong for comments at an APA Pacific Symposium in 2021, as well as participants at the MK40 workshop.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The notion of common knowledge goes back to Lewis (1969) and Schiffer (1972). For prominent uses of it beyond philosophy, see Fagin et al. (1995) and Clark (1996).

2 For recent work challenging the common knowledge orthodoxy, some of which I return to as the paper progresses, see Jankovic (2014), Tenenbaum (2015), Lederman (2017, 2018), and Harris (2019).

3 The view that Peacocke (2005) develops is ultimately non-perceptualist, but in a subtle way, since he appeals to a category of mental states that he labels those of ‘awareness’. This category includes experiences, but also mental states that are more sophisticated than experiences, yet not as sophisticated as beliefs. I do not discuss the subtleties of this view, given space constraints. Campbell's perceptualist view is the focus of this section.

4 For a recent overview of empirical support for both Primacy and Confronting Perspectives, see Tomasello (2018, 2019). As Tomasello outlines, other studies with results complementary to Moll et al.'s (2013) are (Flavell et al., 1981; Doherty and Perner 1998; Rakoczy et al., 2015).

5 The assumption that experiences are non-conceptual is controversial. For early versions of the view, see Dretske (1981) and Evans (1982). More thorough motivation for and clarifications of that view are given by Peacocke (1992) and Heck (2007). A prominent defense of the view that experiences are conceptual is by McDowell (1994), but see Heck (2000) for a response.

6 One immediate worry is that the direct coordination constraint rules out a perceptualist account of joint attention, since the constraint entails that in a basic episode participants must have the belief that theirs and the other's experiences are of the same object. But that entailment does not hold, since the relevant justification is propositional, and not doxastic. The justificandum of direct coordination is the content of a belief, but having justification for a belief-content does not require that one in fact form the corresponding belief.

7 In the terms introduced by Chomsky (1995), this way of drawing the contrast between representationalism and direct realism presupposes an ‘ersatz’ view of some mental representation, since it appeals to metaphorical glosses on the role of subpersonal processing. Compare Egan (2014) and Egan (forthcoming). Though, following Campbell, I am careful to restrict ersatzism to subpersonal mental states.

8 As cited in the introduction, Carpenter and colleagues were the first to provide contrasts of this kind, though they do not specifically target Campbell's perceptualism.

9 Note that Seemann (2010) elaborates Campbell's perceptualist proposal in order to meet the initial objection of underexplanatoriness, and the elaboration might also seem to be able to meet Eilan's more advanced objection. Seemann, drawing on the psychological work of Hobson (2002, 2005), holds that the sharing of feelings about the object is part of joint attention (some other discussions that draw on this material are in Roessler 2005 and Eilan 2005). Crucially, however, the sharing is automatic and perceptual: one perceives another's bodily expression of a feeling towards the object, which is then automatically replicated in oneself. So the sharing of feeling to which Seemann appeals is not a form of communication, since communication is an action. As elaborated much later in this paper, an action is something we do – not something that happens to us – and hence involves personal-level motivational states; whereas, Seemann (and Hobson) seem to be discussing a subpersonal and automatic process. Perhaps appealing to the mechanisms of the automatic echoing of affect makes it plausible that a different kind of subpersonal perceptual processing is being deployed between cases like board meeting 1 and 2. However, it seems plausible that the automatic sharing of affect could occur in cases like board meeting 2, where there is merely symmetrical covert attention: you may, for instance, feel the automatic pull of the other's strange excitement about the presentation, despite the fact that you have only noticed it in a covert manner.

10 A related point has been emphasized by Gilbert (1990, 2013), though she focuses on the claim that common knowledge of interlocked individual intentions is not sufficient for the socio-normative commitments intuitively present in joint action. Following Tenenbaum, I do not take the upshot of the case given here to be primarily about joint commitments.

11 For the latter type of case, see, for instance, Jankovic (2014, 505).

12 This type of case, which appeals to a neuroscientist that can bring about mental states with shocks to the brain, is from Schiffer (1972).

13 This standard view is from Schiffer (1972). An important alternative version of it is given by Sperber and Wilson (1986), who appeal instead to mutual manifestness, which is a weaker notion than common knowledge. But I will not elaborate upon their alternative, since the difference between the two views does not matter for present purposes: the criticism below of the common knowledge orthodoxy I am about to present is that it does not provide sufficient conditions for communication, so appeal to a notion weaker than common knowledge will not address that criticism. This same point applies for the even weaker version of Griceanism in Neale (1992).

14 The problem of sneaky intentions was first raised by Strawson (1964). It was then crucially elaborated by Schiffer (1972), who provided cases of increasing complexity that suggest the addition of common knowledge to fill out the speaker's primary communicative intention. Grice ([1969] 1989), however, suggested that there is in fact something else going wrong in all of the cases put forward by Schiffer. Grice observes that in such cases the speaker intends that the addressee rationally transition from recognition of the informative intention to its fulfillment in a certain way, but at the same intends that the addressee falsely believe that the speaker intends him, the addressee, to not make the transition in that precise way. So, in such a case, the speaker acts with an intention that P, but at the same intends that the addressee think falsely that she, the speaker, has an intention that not-P. So, Grice suggested that, instead of appealing to common knowledge, his account of communication may simply add the requirement that the speaker does not have intentions of that form. But an immediate problem with that suggestion is that it seems ad hoc. So let me suggest that my joint-action-based Griceanism promises to provide a principled way of ruling out the kind of scheme that Grice identified is present in cases illustrating the problem of sneaky intentions. In general, it is plausible that one cannot act with the intention of working together with another, with the aim of making each other's actions mutually intelligible, while at the same intending that the other form a false belief about the intentions behind one's action. Finally, see Jankovic (2014) for a closely allied version of Griceanism, which stresses how communication is a form of joint action, and in which it is argued in detail that such a version of Griceanism can deal with the problem of sneaky intentions. Though it is a subtle question whether I can adopt her solution, since she stresses Bratman's account of joint action, yet I go along with Tenenbaum's alternative. Overall, I do not wish to get into the details of how the present account fully solves the problem of sneaky intentions; instead, what I say in response to the problem here is intended as an indication that the present view is on the right track as a replacement of the common knowledge one.

15 Thanks to Imogen Dickie for discussion on this point.

16 For more developed arguments against Gricean communication along these lines, see Roessler (2005) and Campbell (2018).

17 In particular, there is the still-face effect, which was first established by Tronick et al. (1978) in studies involving infants between 2 and 20 weeks old and their mothers. The behaviour of the infants was compared in two kinds of interaction with their mothers: normal and still-faced. In the normal interaction, the mother played with the infant. Normal playing involves, for instance, miming of the infant's facial expressions, as well as intervening to coax it away from unhappy expressions, or to contain over-excited ones. In the still-face interaction, the mother sits in front of her infant and stares at it with a neutral face. In these cases, the infant attempts to in many ways to engage the mother, but after this is unsuccessful it withdraws and expresses wariness and helplessness. The still-face effect is the significant difference in infants' behaviours between these two different kinds of interactions. According to the meta-analysis of Mesman, van Ijzendoorn, and Bakermans-Kranenburg (2009), similar results have been found in over 80 sudies since Tronick et al. (1978). These studies vary with respect to the age and gender of the children, as well as whether the adult engaging with the child is a parent or stranger. Several studies have been done with non-Western children, and found the effect as well. In addition to the still-face effect, an important source of empirical support for the connection between the social motivation and early episodes of joint attention is that there are negative correlations between Autism Spectrum Disorder (ASD) and engagement in joint attention, among other aspects of social cognition, which supports the present proposal given the recent view that ASD is characterized by a deficit in social motivation (Chevallier et al., 2012; Nyström et al., 2019). For in-depth overview of developmental as well as evolutionary considerations in support of the social motivation's relevance for joint attention, and infant sociality more generally, see Tomasello (2014, 2019).

18 Thanks to John Campbell for discussion on this point.

19 The issue, for instance, of whether the content of a motivational state is an accuracy condition – how things will be – or a fulfillment condition – how things should be. See Dickie (2015, Section 3) and references therein for further discussion of this distinction. For empirical overviews of relevant empirical work on motivation, which support the empirical claims made earlier in this paragraph, see Toates (1986) and Berridge (2004).

20 See especially pages 15–16 of Eilan (2015). I explained above that experiences have non-conceptual content, but that claim is consistent with them having some conceptual content, for it may be that the way experiences contrast with, say, beliefs is that the latter do not have any non-conceptual content. Regardless, I am rejecting Eilan's view, and my view is consistent with the denial of the claim that experiences have any conceptual content.

21 Here I follow the structure of Dickie's (2015) discussion of coordination in lone thought: she is concerned with intrapersonal coordination, in contrast to the present focus on interpersonal coordination. Dickie claims that the immediate justification one often has for taking one's beliefs to be of the same object is practical, given that the activity of forming such beliefs is driven by a need to think about ordinary objects. But note that the correctness of her account of the aboutness of individual thought is orthogonal to the present topic. But, in following the structure of her account, the two ways in which the discussion to come is similar to hers are as follows. First, I also explain epistemic phenomena by likening them to more paradigmatic cases of rational action. Second, in doing so I consider two competing refinements of the general notion of practical justification, one broadly naturalist and externalist and the other broadly non-naturalist and internalist, and ultimately endorse the latter.

22 This account might need to be generalized further, so that the internalist component may also be satisfied by beliefs, since beliefs are also sometimes rationally related to action. For instance, you believe that collecting possessions is the key to happiness, so every day you go to the mall. But I set aside beliefs since they are not relevant to the simple activity relevant to the present topic.

23 Thanks to Johannes Roessler for discussion on this point.

24 This proposal is suggested by Milward and Carpenter (2018) and Roessler (2020).