Why we do things together: The social motivation for joint action

Abstract

Joint action is a growing field of research, spanning across the cognitive, behavioral, and brain sciences as well as receiving considerable attention amongst philosophers. I argue that there has been a significant oversight within this field concerning the possibility that many joint actions are driven, at least in part, by agents' social motivations rather than merely by their shared intentions. Social motivations are not directly related to the (joint) target goal of the action. Instead, when agents are mutually socially motivated in joint action this is because they find acting with others rewarding in its own right. Moreover the involvement of social motivation in joint action typically enables individuals to achieve the long-term benefits associated with being part of a social bond. I argue that taking social motivations into account better prepares us for explaining a broader range of joint actions, including those that are of an antagonistic, competitive, or explorative character. Finally, I show that recognizing the importance of social motivations entails that joint actions (in general) should be understood as having the two primary functions of (1) achieving the intended target outcome of an action, and (2) attaining the benefits related to being part of a social bond.

Keywords:

• Joint Action
• Joint Intentions
• Social Bonds
• Social Emotions
• Social Motivation

1. Introduction

Here is a question: why is joint action so prevalent amongst humans? Surely the prevalence is not just because there is just no other way for individual agents to achieve their aims. In fact, to put things crudely, we often do things with others because we like to do things with them. Despite this fact, there has been little attention to the critical social character of joint action. In their recent review article on joint action, Sebanz and colleagues only briefly pause to note that a sort of social or collective experience could be a desired characteristic of at least some joint actions: “musicians in a band might strive to experience a sense of agency that transcends individual boundaries, which is based on what it feels like to produce action effects as a group” (Sebanz, Bekkering, & Knoblich, 2006, p. 75). I will argue that roughly this sort of experience—despite being largely neglected in the research and literature on the subject—is key to characterizing joint action, and that many joint actions are driven by a shared motivation that is essentially social in character.

The structure of the argument is as follows: in section 2, some reasons are offered for why one ought to choose a broad working definition of ‘joint action’. In section 3, four scenarios are discussed, which fall under the guiding definition. I argue that social motivations have an important role in driving and facilitating the behavioral coordination in three of these interactions, and I also show why social motivations cannot be reduced to either individual or shared intentions to achieve the target goal of the joint action; rather, we are socially motivated to engage in joint action simply because we find the social experience rewarding in its own right. In section 4, I argue for the importance of social motivations in joint action from a different direction. I try to show that the social motivations involved in joint action also allow individuals to attain benefits associated with being a part of a social bond. Some examples of the psychological mechanisms responsible for socially motivating us—such as empathy and social emotions—are discussed, and I indicate how their involvement in joint action promotes the formation and preservation of social bonds. In section 5, I respond to a potential criticism and argue that we should adopt a two-function approach to joint action. In sections 6 and 7, I consider whether social motivations can be accommodated within existing accounts of the psychological prerequisites for joint action, and how recognizing their role in joint action might inform future research.

2. Defining Joint Action

The paradigm case of joint action, which research has tended to focus on, is a group of individuals solving or performing some precise task together, such as lifting and moving a heavy sofa (Velleman, 1997) or tidying up the toys together (Behne, Carpenter, & Tomasello, 2005). But of course many of our social interactions with others occur when there is no apparent problem to be solved, and when there is no interest in cooperation per se.

We could of course limit the domain of research and investigation by defining joint action as the class of interactions between agents that are, at least primarily, of a problem-solving and cooperative character. However, this seems unduly restrictive. We have good reason to suppose that the same type of shared attention and coordination in cognition and behavior that occurs in the cooperative settings may also be present when individuals are far less cooperative. Moreover, most researchers concerned with joint action seem reluctant to stipulate that they are only concerned with interactions that are cooperative and/or tend toward a joint goal. On the contrary, researchers seem to favor a broad definition of joint action. For these reasons I propose we adopt the definition that Sebanz et al. propose in their review article:

Joint action can be regarded as any form of social interaction whereby two or more individuals coordinate their actions in space and time to bring about a change in the environment. (2006, p. 75; see also Knoblich & Sebanz, 2008)

This definition has the advantage of not presupposing any particular psychological attitude or mechanism on behalf of the agents.1 It also quite correctly focuses on the fact that joint action requires a high degree of coordination between agents—which arguably is one of the main reasons why researchers got interested in joint action as a unitary phenomenon to begin with.

I will hence adopt this broad understanding of joint action in what follows, and I will try to demonstrate that if we consider the range of social interactions that fall under this guiding definition, we also need to bring in another type of explanation for what is driving these joint actions than is currently on offer. As is often alleged, this explanation does typically appeal to something that can be shared between individual agents; but instead of appealing to a shared goal, intention, or representation, I suggest we should appeal to a shared social motivation driving the joint action.

The nature of social motivations will be further fleshed out in subsequent sections, but it might be helpful to say something about what I take the nature of social motivations to be from the outset. Social motivations, at least in this paper, are a broad group of emotional and affective factors (examples are given in section 4). What distinguishes these emotional factors from other affects and emotions is precisely their pro-social character. That is, the relevant emotions are elicited specifically in social settings involving other people, and they render the social experiences rewarding in their own right. In section 4, I will also argue that there is a second sense in which these motivations are social which is derived from their role in the formation and strengthening of social bonds.

Notice that so far, I have not defined social motivations through their role in joint action, and so there is no danger of circularity (incidentally, social motivations have a role in many social settings other than joint action; see note 8). Nevertheless, their motivational character and their ability to bring about long-term benefits, such as facilitating social bonds, do give us an initial clue to why we should expect social motivations to play an important role in many joint actions that fall under the above definition.

3. Four Cases of Joint Action

To begin to get an idea of how shared social motivations might drive joint actions and how they can be contrasted against joint or shared intentions, I want to consider some scenarios. (A) Let's begin with a typical case of the more “detached” joint action that is typical for the literature on joint action. Two experienced piano movers are brought together for a one-off job to move a piano from the fourth floor to the street by carefully lowering it using a rope and pulley. The task is familiar to both of them and the goal is clear. Moreover both agents are to some degree individually motivated to perform the task. Let's also suppose they have no history of working together and, due to work arrangements, they are mutually aware that there are very slim prospects that they will work with—let alone encounter—each other again. It therefore seems reasonable to suppose that what drives the joint action is the joint target goal and/or a shared intention to get the job done. There hence seems to be no prima facie reason to suppose that social motivations have any role in explaining this case of joint action.

The reason I have included this scenario is that I want to grant from the outset that a social motivation need neither be present nor drive all cases of joint action, as appears to be the case in (A). There might also be joint actions when there is no shared social motivation, such as when only one or a minority of the agents is socially motivated to interact. Neither possibility will affect my argument. What matters, and what I would like to argue for, is rather that there are many joint actions where all the involved agents either are or become socially motivated; in other words, where there is a shared social motivation.

So, let's now move on to three scenarios where there does seem to be a shared social motivation present. Consider a different version of the first scenario. (B) An experienced mover has just taken on a trainee and they embark on their first joint task, which mirrors the first case: to move a piano from the fourth floor to the street by using a rope and pulley. The target goal is therefore the same as in (A). However, in the process of moving the piano, the mover and the trainee find they have the same absurd sense of humor and really get on like a house on fire. The mover and the trainee are then not only motivated to perform the task, they also become socially motivated to engage in the activity with each other; let's say, ultimately leading the mover to want to employ the trainee and vice versa. So, although the joint action is initially only driven by a shared intention, it is increasingly also directed by a shared social motivation.

Now consider a case where, in contrast to scenarios (A) and (B), there seems to be no apparent joint goal or shared intention at all, but which nonetheless falls under the guiding definition of joint action. (C) Two siblings are sitting across from each other at a kitchen table, discussing how to split their mother's inheritance.2 Each sibling wants and feels they deserve a greater portion of the inheritance than the other sibling. Because there is considerable tension between their target goals for the negotiation—one might even say the goals are in direct conflict—we would be hard pressed to posit a shared intention.3 However, let's also suppose the two siblings, without making any conscious note of it, both care for each other and strongly desire that the inheritance issue should not have any severe impact on their long-standing relationship, this mutual emotion representing a shared social motivation. In fact, if the shared social motivation had not been present they would have called their lawyers rather than agreeing to negotiate with one another. Considering the complex set-up, with the siblings having both individual goals and shared social motivations, the joint action in scenario (C) cannot be characterized as purely cooperative or competitive—if anything, the negotiation is explorative. Both the siblings’ individual targets and their shared social motivation impact on the behavioral coordination of the negotiation, in terms of attention, verbal behavior, facial expressions, gestures, body movements, etc., and both have a role in determining the outcome of the negotiation.

Here is another case which is supposed to be a paradigm case of joint action, but where attempts to posit a shared intention seem slightly ad hoc: (D) two infants are playing a game with one another using a kind of trampoline to bounce a ball in the air (Warneken, Chen, & Tomasello, 2006). In recent psychological experiments it has been shown that human infants and young children are much more prone to engage in this form of social play than chimpanzees, who in contrast quickly lose interest in activities where there is no clear objective (Warneken et al., 2006; Warneken & Tomasello, 2007). One of the reasons for this, Tomasello and Carpenter suggest, is simply that “humans collaborate just for the sake of collaborating” (2007, p. 123). Part of the point of this paper is indeed to agree with their observation (at least in the context of collaborative joint action); but in my view they then make a mistake in falling back on a shared goal or intention to collaborate in explaining the successful synchronization in the social game. In contrast, I contend that the synchronization in case (D) can be made sense of purely in terms of the social motivations that encourages the child to engage and coordinate in the game (perhaps, if you like, coupled with individual goals), and that there is no need to posit any shared intention as driving the joint action. What's more, as in the previous case (C), this has the upshot that we need not think of the activity as purely cooperative, i.e., it might be partially competitive or explorative.

The precise role of the shared social motivation in joint action has only been very briefly sketched in the above scenarios, and will be discussed further in the next section; but these scenarios should at least begin to make clear how a shared social motivation can be distinguished from target goals and shared intentions. In particular, I propose that there are two jointly sufficient conditions that distinguish social motivations from shared intentions. The first condition (1) is purely negative and says that a shared social motivation is not directed toward the benefits of attaining a target goal. Moreover, since “shared intention” is typically thought of as precisely being oriented toward the target goal or outcome of the action (see, e.g., Bratman, 1993; Pettit & Schweikard, 2006; Tomasello & Carpenter, 2007), it follows that the shared social motivation cannot be reduced to a shared intention either.4 Of course agents that are socially motivated to joint action can also be motivated to attain individual target goals (as in (C) and (D)) or shared target goals (as in (B)), and I will have more to say about how these different types of motivation interact in section 6.

The second condition (2) gives us a positive reason for distinguishing shared social motivations from both target goals and shared intentions, and is therefore more interesting. This condition is that the social motivations perform a different role than shared intentions, since they make the experience of acting together with others rewarding in its own right, and so do not relate to any potential rewards of the task-related components, such as the goal of the joint action. These social motivations causally influence us to engage in joint actions, either by initiating the joint action in the first place or by being elicited during the course of joint action (or both). In other words, although they may trigger an interaction loop of shared action and shared motivations, they lie outside the interaction and reside on the level of individuals in the joint action.

This is however not to say that the social motivations (the emotional and affective states) driving joint actions are only important within the context of a particular joint action; in fact, I will argue that the social motivations in joint action typically also allow the individuals involved to achieve long-term benefits that lie beyond the joint action and its intended outcome. The next section discusses what these benefits are and how they might be achieved.

4. Social Bonds: Another Upshot of Joint Action

In the opening paragraph of this paper I suggested that we often do things with others because we simply like to do things with them. While I stand by this statement as a rough-and-ready characterization of many joint actions, it does run the risk of underplaying the importance of joint action. Not only does joint action, as is often pointed out, allow us to achieve an increased range of outcomes that we would not be able to achieve individually (e.g., moving a piano or marrying), I will also argue that through the involvement of social motivations, joint action enables the beneficial formation of social bonds. There is also a related possibility that some cases of joint action, such as debating and bullying, play an important part in the establishment social hierarchies.

There are probably many species that can rightly be described as “social,” but it is often thought that only humans qualify as the “hyper-social species” (Boyd & Richerson, 1988). In particular, the human ability to form social bonds both with kin and non-kin is recognized as important from the viewpoint of biological and cultural evolution alike (Boyd & Richerson, 1988; Hamilton, 1964).5 It is of course a near-truism of social psychology and health care that social bonds, whether they be with colleagues, friends, family, or sexual partners, have direct and indirect beneficial effects on human health, reproduction, and general well-being across the life span—and, conversely, that there are significant detrimental effects associated with social isolation (Baumeister & Leary, 1995; House, Landis, & Umberson, 1988; for two recent studies see Cornwell & Waite, 2009; Uchino, 2009). In short, it seems clear social bonds have social and cultural benefits as well as long-term individual benefits (or adaptive value).

So how are such beneficial social bonds formed? It is of course possible that for some key social bonds, like the one between the parent and the newborn, we can just sit back and allow our genes to do most of the work. The formation and maintenance of most other social bonds, however, will require more “action” on the part of individuals—with joint action being a particularly suitable candidate for enabling and facilitating the formation of social bonds. In other words, it may be quite correct to use a commonsense description of many types of joint action as forms of “bonding.”

But now the question becomes: how it is that joint action, as a kind, is able to promote the formation of social bonds? It is at this juncture, I contend, that we should turn to the social motivations that in the last section were argued to drive many cases of joint actions. Yet before we see how social motivations might be able to do this work, I wish to supplement the general characterization in terms of emotional and affective factors with some particular examples of social motivations. I will briefly mention three examples of mechanisms that researchers have recognized as “socially motivating,” but there are probably several other relevant psychological and neurobiological mechanisms that might be worthwhile considering in this respect.6

The first example of a social motivation typically affects us un- or subconsciously: oxytocin. Although it turns out that this hormone can be found even in non-social species like mollusks, in mammals the hormone is presumed to be co-opted for the specific task of managing maternal nurturing behavior (Uvnäs-Moberg, 2003). In humans it is also well established that oxytocin can be released in plasma and in cerebrospinal fluid when individuals partake in social interactions that are not linked to maternal nurturing. Oxytocin has, for instance, been demonstrated to be elicited by simple stimuli such as touch, and to be involved in various kinds of pair bonding (Gimpl & Fahrenholz, 2001). Consequently, it seems reasonable to suppose that for humans, oxytocin might be involved in (and reward) many joint actions where non-aggressive physical contact is involved.

All the same, there also seems to be some reason to be cautious of over-emphasizing the importance of oxytocin and other hormones in relation to social bonds, since, at least for humans, higher cognitive-functioning has enabled us to be socially motivated in many other situations than those which we are hormonally primed for, and hormonal influence is arguably also not sufficient for forming long-term social bonds (Curley & Keverne, 2005). This leads us to a second kind of often conscious social motivations: social emotions. Many psychologists have argued that many emotions are crucial for forming social relationships (Gonzaga, Keltner, Londahl, & Smith, 2001; Griffiths, 1997; Izard, 1978; Keltner & Haidt, 1999). The most prominent account of the social function of emotions is probably given by Frank (1988), who argues that emotional dispositions, such as love and trust, are superior to any prudential reasoning, since social emotions motivate us in a unique way that enables us to accrue the benefits of long-term relationships like friendships and monogamous sexual relationships. Frank also observes that some “negative” social emotions like jealousy are also important in strengthening social bonds.

A final related case of social motivation is empathy, where we feel or resonate with others’ emotions. When people feel, and moreover display, another person's emotions, they communicate that they share a common perspective on the situation, and allow for a continued shared perspective that is arguably required for social bonds to be formed (Anderson & Keltner, 2002). Moreover empathy is often signaled through facial expressions, and so is also quite difficult to feign. Thus, a facial expression of empathy can be construed as a true test of someone's genuine solidarity, and is therefore a means of enhancing both trust and a mutual understanding (Hobson, 1993; Tronick, 1989). Finally, it has been demonstrated that empathy is more likely to be elicited on occasions when an existing bond between individuals already exists, and so would reinforce the bond in question (Hoffman, 2000).

So by which means do social emotions, empathy, etc., in joint action help promote the formation of social bonds?7 First of all, I have previously claimed that the social motivations (the emotional and affective factors) render many joint actions rewarding in their own right. The idea here is in essence Hume's famous observation that “the mind has a great propensity to spread itself on external objects, and conjoin with them any internal impressions which they occasion” (1739/2007, 1. iii. 14, 25). In the context of joint action, this means that certain social emotions, empathy, etc., motivate us to continue with and return to the joint actions that are associated with the emotion (or “internal impressions”). In fact, it is not only the joint action in question that gets positively appraised, but also the other people involved in the joint action. The basic idea is hence that the various types of social motivations described above help shape future behavior by prompting a continued engagement with both the activities and individuals in joint action—in due course enabling or strengthening a social bond.

Moreover, the chances of a social bond being formed or strengthened typically rely on all individuals in the joint action being socially motivated, i.e., the social motivation being shared, generating a sense of mutuality and an ongoing feedback-loop. It is, however, important to remember that even in competitive or antagonistic joint actions, like having a debate or a fight with someone, there are likely to be social motivations involved (for instance, social emotions like resentment and shame). Because of their negative valence these emotions tend to promote the formation of social hierarchies rather than social bonds—though not exclusively so (Keltner & Haidt, 1999).

5. The Dual Function of Joint Action

A very natural objection here would be the following: “regardless of how precisely ‘joint action’ is defined, most researchers in the field are concerned with interactions that are either more straightforwardly cooperative and/or involve the presence of a jointly intended goal.” As I argued initially, there is good reason not to make such assumptions about joint action being restricted to cases where a shared intention is clearly present—but let's leave this aside for a moment. What do I have to say about the more classic cases of joint action where there indeed seems to be cooperation to achieve a joint target goal? In such cases, is it not sufficient that there is a shared intention for the joint action to get off the ground; why would we also need to make reference to social motivations in these cases?

As a preliminary, I have previously pointed out that I am not committed to the claim that a social motivation must be present in all cases of joint action. Still, there are many cases of cooperative joint action where we should expect that there is a shared social motivation that drives and causally influences the joint action. One of the reasons why this is likely has already been implied by what has been said in the previous sections; namely, that the shared social motivations allow joint action to achieve a different goal than the putative target goal, by promoting the formation of social bonds or hierarchies. It then also seems quite natural to suppose that it will have a differential impact on the nature of the coordination, and possibly the outcome of the joint action. At this point it would be nice to appeal to some experimental evidence that demonstrates this differential effect of social motivations when there is a target goal (and likely a shared intention); but as social motivations in general have been neglected in the research on joint action, it is not surprising that there is also a paucity of research addressing this particular issue.

All the same, to get an idea of the differential impact of social motivations we can turn to a related area of research where there is increasing interest in the role of social motivations: imitation (Dijksterhuis, 2005). For instance Nielsen and colleagues (Nielsen, 2006) have tested the differential impact that social engagement has on the child's imitative behavior. They found that 24-month-olds persisted in imitating a demonstrator even when it resulted in frequent failure to use a target object, and there were obvious superior alternatives available. To explain the result, it was hypothesized that the children had persisted in imitating the demonstrator's inefficient object use because they wanted to satisfy social demands. To test this interpretation, Nielsen compared the way 24-month-olds imitated a live versus a videotaped demonstrator opening a box in order to retrieve a toy (Nielsen, Simcock, & Jenkins, 2008). Although the videotaped demonstrator was socially engaging, the medium implied that there was no real opportunity for “live” or spontaneous social interaction with the child. The results showed that children were significantly less inclined to imitate a videotaped demonstrator compared to the live version, and it was concluded that children's imitative behavior is affected by the opportunity for genuine socially interaction.

In line with these experimental results, Uzgiris (1981) notably claimed that genuine social engagement and motivation is important in explaining children's imitative behavior, and suggested that imitation in fact has two functions: (1) acquiring certain behavior and skills through social learning; and (2) experiencing a sense of mutuality which promotes further social interaction with the partner. Uzgiris argued that not only are these aspects theoretically compatible as the dual function of imitation, but when we focus on the interplay between them we have a much better chance of explaining experimental data.

What I want to suggest is that a similar moral holds true for joint action—there are similarly two central functions of joint action. The first function I take to be implicitly or explicitly accepted by most researchers interested in joint action, which is to reap whatever the particular benefits are of achieving the jointly or individually intended target goal. Here it doesn’t really matter if the goal is achieved on one's own or with others; but as we have seen it is a brute fact that to achieve some individual ends, one must jointly act with others. The second function of joint action is the one I have been particularly concerned with in this paper; that is, the function of forming, maintaining, and strengthening social bonds, or establishing social hierarchies.8

I therefore maintain that even in many traditional cases of cooperative joint action, there is also a second function of the interaction; namely to promote or facilitate social bonds. This may not just affect the individuals’ propensities to engage in joint action, but also, when it interacts with a shared intention, make the behavioral coordination more seamless and increase the likelihood that the goal is achieved. More experimental work is of course needed to demonstrate this effect, but for now, take the example of the joint action in scenario (B): in addition to the function of getting the piano to the street, as we have seen the trainee and the employer are also socially motivated to perform the task. The sense of camaraderie and trust that develops between them should help them to achieve a common perspective on the job, their respective roles, and its target goal.

Some qualifications are in order. It should be clear that the second function of joint action (i.e., forming social bonds) does not presuppose that this is achieved through cooperative joint action; this result can equally be attained through explorative and competitive joint actions. Just think about how verbal sparring between two individuals becomes a form of flirtation, possibly resulting in a romantic relationship. I’ve also allowed that the social function of joint action might be less “good-natured” with the result being social hierarchies rather than social bonds; e.g., when one person becomes dominant and starts belittling another. Finally, although the two functions of joint actions are often present and interact, they do not both have to be involved in all cases of joint action (e.g., the moving in (A) does not appear to fulfill a social function, and the negotiation in (C) does not function to bring about a joint target goal).

6. Social Motivations in Existing and Future Accounts of Joint Action?

I hope to have shown that promoting the formation of social bonds is one of two central functions of joint action, and that agents’ shared social motivation is largely responsible for joint action achieving this end. So why then has the literature on joint action almost entirely neglected both the shared social motivations and social function of joint action and focused solely on explaining joint action in terms of agents achieving target outcomes?

One of the reasons might be that it is not only social motivations, but any emotions and affects that tend to be overlooked in accounts of joint action. Researchers have by-in-large focused on various candidate cognitive and sometimes perceptual preconditions of joint action such as shared intentions (Bratman, 1993; Carpenter, 2009; Tomasello, Carpenter, Call, Behne, & Moll, 2005), shared task representations (Sebanz & Knoblich, 2009; Vesper, Butterfill, Knoblich, & Sebanz, 2010), joint attention (Knoblich & Sebanz, 2008; Tollefsen, 2005; Sebanz et al., 2006), action monitoring and prediction processes (Vesper et al., 2010), joint commitments (Gilbert, 1992; Tomasello & Carpenter 2007), abilities to communicate (Gergely & Csibra, 2005a), and abilities for mind/intention-reading (Carpenter, 2009; Tomasello et al., 2005). The omission of social motivations is also evident in the previously mentioned review paper on joint action, where the authors list the following psychological mechanisms as the important ones involved in joint action:

First, joint attention provides a mechanism for sharing the same perceptual input and directing attention to the same events. Second, a close link between perception and action allows individuals to form representations of others’ action goals and to predict action outcomes. Third, by forming shared task representations, it is possible to predict actions based on certain events in the environment, independent of action observation. Fourth, action coordination is achieved by integrating the “what” and “when” of others’ actions in one's own action planning. This affects the perception of object affordances, and permits joint anticipatory action control. Finally, the ability to distinguish between effects of one's own and others’ actions might be reduced in joint actions where the combined outcome of one's own and others’ actions is more important than the results of individual actions. (Sebanz et al., 2006, p. 75)

It seems uncharitable to suppose that researchers want to deny that (social) motivations are present in many cases of joint action; so the neglect must be due to thinking that they are of little explanatory significance. Perhaps the assumption is that motivation just falls out trivially from existing accounts? But considering the cases of joint action that are typically being discussed, this seems tantamount to supposing that agents arrive at joint action with a shared motivation to (jointly) solve a specified coordination problem! As we saw with case (D), it also seems slightly ad hoc to suppose that most joint action can be explained by a shared motivation to collaborate or cooperate. This is further supported by reflection on the possibility of various explorative, competitive, and antagonistic joint actions. In connection with this, it is also worth noting how my account of social motivation differs from the suggestion given by Tomasello and colleagues (2005), who claim that the primary motivation behind collaborative activities is to share intentional mental states with others. They claim that this specific motivation is unique to humans and is a necessary precondition for the collaboration that is ubiquitous in human cultures. I claim that they are over-stating their case, since we are often socially motivated to engage with others despite not wishing to share their mental states per se; but of course that is not to exclude that this particular motivation could be one of the important shared social motivations for joint action amongst humans. 9

Now, it may be that the individual motivation to achieve a target goal, etc., does fit within an account of intention, or shared intention. All the same, it is far less plausible that a social motivation can be accommodated within the existing models of shared intention (for such models see, e.g., Bratman, 1993; Pettit & Schweikard, 2006; Tomasello et al., 2005; Tuomela, 2005; Velleman, 1997). Though I do not want to rule it out in principle, it seems like such an accommodation would be improbable for the reason already rehearsed: shared intentions are currently taken to be directed toward the target outcome of a joint action, and shared social motivations, in contrast, are directed toward the social benefits which lie beyond the token joint action. But there is also a second reason to be skeptical regarding the possibility that the current notion of a shared intention is open to this kind of revision. At least all philosophical accounts of shared intentions demand that there is some common propositional content (e.g., the intention “to form a social bond with X”), which is either instantiated in each individual participant or in the jointly acting group as a whole (Pettit & Schweikard, 2006). In contrast, I have not demanded that a shared social motivation be explicated by the same propositional content in both agents. In fact, I do not demand that we explicate the social motivations by means of representational or propositional content at all. Rather, in section 4, I claimed social motivation should be explicated in terms of emotional or affective mechanisms that, when driving or being elicited by joint action, have the effects of rendering joint action rewarding in its own right.

It would then seem like social motivations are independent from, and not reducible, to any of the psychological or cognitive items that are currently considered to be required for joint action. But of course, in order for social motivations to influence or drive joint action, they must also interact with many of the recognized perceptual/cognitive mechanisms and motor skills. The suggestion here is simply that there is another neglected mechanism (social motivations) involved in joint action that is typically required for explaining why we engage in joint action, as well as what makes joint action successful and also, in the long-term, a greatly beneficial experience for individuals.

In this respect, it is important to stress that investigations into social motivations do not imply a deviation from the aims of developing mechanistic accounts of joint action. On the contrary, I have given examples of how we can use insights from research on oxytocin, empathy, and social emotions to posit relevant mechanisms that are responsible for socially motivating us. This kind of work, which integrates mechanisms for social motivation with cognitive and perceptual mechanisms, already seems to be spearheaded by the growing field of social cognitive neuroscience. For instance, three reciprocally connected neural structures—the ventral striatum, the orbitofrontal cortex and the amygdala—have been demonstrated to mediate both social judgments and basic reward processing. In the case of social coordination, these structures are thought to interact very closely with those regions that are involved in action planning and control, such as primary motor cortex (Adolphs, 2003, p. 170).

In section 5, I also suggested another hypothesis for future research on joint action; namely, that it is likely that social motivations might assist the effectiveness of the coordination in the joint action. What's more, when there is a shared goal, social motivations might increase the likelihood of the joint goal being achieved. We saw that there is some corroborative evidence for such hypotheses from the effects of social engagement on infants’ imitative behavior. Other experiments on non-human primates also suggest that socially motivating cues tend to determine the rate of “gaze-following” (to direct one's attention to locations looked at by others; Teufel, Gutmann, Pirow, & Fischer, 2010). The studies show that subjects are much more likely to follow the gaze of another animal if the gaze-cue is accompanied by a clear facial expression of emotion/affect—which is considered to be a highly socially motivating feature even in primates. At the same time features like sex, rank, and relatedness of the animal did not affect gaze-following rates.

7. Conclusions

There are several positive upshots of acknowledging the importance of social motivations for joint action. First, the appeal to a shared social motivation helps us explain a broader range of phenomena subsumed under a popular definition of joint action. Since social motivations can be regarded as an important driver of joint action in their own right, more attention can legitimately be directed toward explorative, competitive, and even antagonistic joint actions, which often involve equally high degrees of behavioral coordination as do cooperative joint actions.

Second, this paper hypothesizes that the benefits of joint action do not just stem from individuals’ abilities to achieve novel target outcomes that they would not be able to achieve otherwise, but from the social rewards afforded by the experience of acting with others, and the way in which these experiences in turn promote social bonds and social hierarchies. Understanding joint action according to this dual function framework will probably also lead us to look at cases of joint action in a slightly new light. Consider, for instance, a political demonstration that appears to be driven by both the (shared) intention of furthering a cause and by the social motivations that help bring about or enhance a sense of solidarity within the group. As seems clear in this case, the social motivations may furthermore affect the level of successful coordination in such “goal-oriented” joint actions.

Let me finally emphasize that my objective here has not been to challenge the recognized preconditions of joint action. Of course it is necessary to focus on an account of the cognitive and perceptual abilities that underpin joint action, especially as it typically involves quite complex behavioral coordination between agents. Rather, my contention has been that shared social motivations are equally causally significant for most cases of joint action to occur. By leaving them out of the explanatory framework of joint action, one not only will be unable to account for what drives those joint actions where there is no apparent target goal: one will also have an incomplete account of joint action in general.

Acknowledgements

The idea for this paper emerged from a series of socially and intellectually rewarding meetings with other members of the interdisciplinary research group on Social Interaction: The Interplay of Pre-Reflective and Reflective Processes (Marijn van Wingerden, Vivian Bohl, Dr. Mog Stapleton, Dr. Christoph Teufel, Dr. Wouter van den Bos). I would also like to thank Dr. Elselijn Kingma, Dr. Kalbir Sohi, Dr. Anneli Jefferson, Tom Bawden, Paolo Mantovani, and two anonymous reviewers for providing very valuable suggestions and criticism on drafts of this paper. This research was supported by the British Arts and Humanities Research Council [grant number 133751], Volkswagen stiftung, and the Signhild Engkvist Foundation.

Notes

Notes

[1] Of course, this depends on what is meant by “to bring about a change in the environment.” One might take the expression to mean that agents must have a shared intention in order to bring about a change in the environment. Yet nowhere do Sebanz et al. state that this should be part of the definition, and I shall not assume that it is.

[2] Language, in terms of coordination both in gestures and speech, is generally understood as joint action, albeit considered to be a highly evolved kind (Clark, 1996; Sebanz et al., 2006; Tomasello & Carpenter, 2007).

[3] You might think the scenario can be explained by a shared intention to strike equilibrium between their individual aims and their shared interest in not severing their relationship. But the scenario need not collapse into such an account. For instance, let's suppose that one of the siblings has a stronger social motivation and the other will pursue her individual goal further.

[4] There is of course an issue about how to define particular shared intention/joint goal; i.e., in scenarios (A) and (B), does the joint goal refer to getting the piano onto the street, to its new home, or finishing work, etc.? Note however that this issue will not arise in the account of social motivations, since we needn’t conceive of social motivations as propositional attitudes (see section 6).

[5] There is, however, also a growing amount of evidence for the adaptive importance of social bonds in non-human mammals. Social bonds have, for instance, been shown to reduce levels of aggression between females and the risk for infanticide in the house mouse (Dobson, Jacquot, & Baudoin, 2000), and increase the shared care for dependent young in wood mice and prairie voles (Gerlach & Bartmann, 2002; Hayes & Solomon, 2004). In longitudinal studies of the effects of social bonds on wild female baboons’ inclusive fitness, it was concluded that socially integrated female baboons are much more reproductively successful than females who are more solitary (Silk, Alberts, & Altmann, 2003).

[6] If we reject the Humean dictum that “reason alone can never produce any action” (1739/2007, 3. i. 1, 10) we may allow that certain prior knowledge and judgments of others can also socially motivate us, bypassing the emotional/affective level altogether (Koenig & Harris, 2005).

[7] The social emotions and hormones that I have mention may of course be elicited by and reward other social activities than joint action, like planning to see a friend or looking nostalgically at a family photo album and of course such activities may equally help promote social bonds.

[8] Incidentally, joint action can have these two functions for some non-human mammals as well as in humans, although it is not my ambition to discuss this possibility presently.

[9] It also seems plausible that the (human) evolution of collaboration can also be accounted for without the specific motivation to share others’ psychological states (Gergely & Csibra, 2005b).