On Knowing What We’re Doing Together
On Knowing What We’re Doing Together
Groundless Group Self-Knowledge and Plural Self-Blindness
Abstract and Keywords
An influential view argues that in order to act intentionally, the agent needs to know what he or she is doing. Such self-knowledge, it is claimed, is epistemically distinctive in that it is ‘groundless’—non-observational and non-inferential. This chapter investigates how this view relates to the theory of intentional joint action. Is our knowledge of what we are doing together with others—collectively, as a team or a group—of the same groundless kind? The chapter is divided in three sections. The first section unpacks the idea of groundless (individual) self-knowledge, as developed by G. E. M. Anscombe, and in more recent philosophical research. The second section plays the part of the individualist’s advocate. The third section argues against the individualist view, and defends (a moderate version of) groundless group self-knowledge. The knowledge in question is plural pre-reflective and non-thematic self-awareness of what it is the participants are doing together.
An influential view argues that in order to act intentionally, the agent needs to know what he or she is doing. Such self-knowledge, it is claimed, is epistemically distinctive in that it is ‘groundless’—non-observational and non-inferential. This chapter investigates how this view relates to the theory of intentional joint action. Is our knowledge of what we are doing together with others—collectively, as a team or a group—of the same groundless kind? The chapter is divided in three sections. The first section unpacks the idea of groundless (individual) self-knowledge, as developed by G. E. M. Anscombe, and in more recent philosophical research. Plausible features of self-knowledge include first-person identity, first-person perspective, first-person commitment, and first-person authority. The second section plays the part of the individualist’s advocate. In order to know what we are doing, it is obviously not enough to know what I am doing; I need to have an idea of what my partners are up to, too, and it seems that observation and inference are the only sources of knowledge of other people’s actions. I might not know exactly what it is what my partners are doing (for example, in a company with a high degree of division of labour), but for me to (p.52) know what it is we are doing, as a company, I need to have some knowledge of the others’ doing, and it seems that any such knowledge cannot be of the non-observational or groundless kind. The third section argues against the individualist view and defends (a moderate version of) groundless group self-knowledge. The knowledge in question is plural pre-reflective and non-thematic self-awareness of what it is the participants are doing together. Groundless group self-knowledge, however, differs from groundless individual self-knowledge in the kind of identity, perspective, commitment, and authority it involves.
1. Groundless Self-Knowledge Unpacked
The claim that intentional action implies knowledge of what it is one is doing may not initially strike one as plausible. After all, large parts of intentional action are performed unthinkingly—we’re not monitoring it actively. Nevertheless, we’re acting intentionally even if we’re not paying any attention to it. In that ambitious sense, we certainly do not constantly need to ‘know’ what we are doing. In a more basic sense, however, it seems that even in such cases of routine action we do, and our agency breaks down if we don’t. That more basic ‘knowledge’ involves a sense in which, even in routine behavior, it has to be apparent to us what it is we’re doing, or else intentional action breaks down.
Such breakdowns sometimes happen. Imagine that during a short break after some hours of intense work on a paper at your desk, still thinking about your paper, you find yourself in the kitchen, opening the fridge, not knowing what it is you’re doing there. Perhaps your cluelessness does not run all the way down to your present bodily movements—you know perfectly well that you’re opening the fridge—but you have no idea as to the question of why you’re doing it. Were you about to get something from there, or put something back? You still feel utterly lost and rather stupid for a moment, and intentional action has broken down, until you realize that you’re holding a cup of black coffee in your hand. It is coffee time. And you don’t drink your coffee black. So that’s what you were about to do: get some milk, and there it is, the milk, right in front of you. You’re now able to resume your intentional action—you’re back on track. The crisis is over.
A philosophical question is, what exactly has gone wrong here? One potential answer might simply be that you have lost for a moment the very element that brought you back on track at the end of the crisis: you were lost in your thoughts about your paper, failed to watch what you were doing, and you were thus unable to jump in when your auto-pilot somehow failed. This implies a certain view of how intentional action works: if the action is not of the altogether spontaneous sort, there usually is a decision to act that is then kept in mind, and that places (p.53) constraints on the future behavior which the agent has to observe with sufficient attention. Another view argues, however, that a crucial element is missing in this account. An agent needs to know what it is he or she is doing ‘just like that’, without resorting to recollection that a decision has been made sometime in the past, and perception of the stage of execution in which he or she is currently engaged. The shortest argument for this latter view is that if somebody else asks you what it is you’re doing, you don’t normally have to observe your behavior in order to come up with an answer. The rare cases in which you do are typically cases in which your intentional action has already broken down, such as in our fridge example.
This is similar to the view argued for by Anscombe in her Intention:
Say I go over to the window and open it. Someone who hears me moving calls out: What are you doing making that noise? I reply ‘Opening the window.’ I have called such a statement knowledge all along; and precisely because in such a case what I say is true—I do open the window; and that means that the window is getting opened by the movements of the body out of whose mouth those words come. But I don’t say the words like this: ‘Let me see, what is this body bringing about? Ah yes! the opening of the window’. Or even like this ‘Let me see, what are my movements bringing about? The opening of the window’. To see this, if it is not already plain, contrast this case with the following one: I open the window and it focuses a spot of light on the wall. Someone who cannot see me but can see the wall, says ‘What are you doing making that light come on the wall?’ and I say ‘Ah yes, it’s opening the window that does it’, or ‘That always happens when one opens that window at midday if the sun is shining.’
(Anscombe 1957, 51)
Thus the kind of knowledge Anscombe places at the heart of her account of intentional action is not based on observation and inference. It involves ‘opinion’ that, as she famously puts it, ‘is held without any foundation at all’ (Anscombe 1957, 50)—if it is knowledge, it is groundless. Anscombe claims that it is knowledge of this kind that characterizes an action as intentional.
What, then, is this groundless knowledge? Anscombe further characterizes it in contradistinction to the observational knowledge one may have about one’s actions:
[I]f there are two ways of knowing here, one of which I call knowledge by observation of what takes place, then must there not be two objects of knowledge? How can one speak of two different knowledges of exactly the same thing? It is not that there are two descriptions of the same thing, both of which are known, as when one knows that something is red and that it is coloured; no, here the description, opening the window, is identical, whether it is known by observation or by its being one’s intentional action.
(Anscombe 1957, 51)
Anscombe is very explicit in her rejection of the view that it is only the intention (and nothing about the actual behavior) that is ‘known without any foundation’, (p.54) but she says surprisingly little about what this groundless knowledge really is. Reconstructive work needs to be done here, a great deal of which can be found in the relevant literature. In what follows, I’ll consider and explain four features that self-knowledge of the relevant kind seems to have:
(a) First-person identity. To use Anscombe’s example, the knowledge in question is whatever is lacking when I hear that somebody is ringing the bell, but I do not realize that it is me who is ringing the bell—I hear that the doorbell is ringing, but I’m not aware that it’s me who’s pushing the knob by leaning against the wall. This characterizes my ringing the bell as non-intentional, even though it is me who does it (Anscombe 1957, 51). The way in which I become aware of the fact that it is me who is doing the pushing is not by learning it is the person who is leaning against the wall who is doing it, because that does not, in itself, constitute the kind of self-knowledge needed here. I still need to know that I am that person, and no observation alone can yield that knowledge. Anscombe’s point is closely related to Wittgenstein’s distinction between the subject use and the object use of ‘I’ (Wittgenstein 1958 [1933/4], 66 ff.). The ‘groundless knowledge’ in question is knowledge of what it is I am doing, and it involves self-knowledge that is radically different from the way in which knowledge of what somebody is doing is knowledge of that agent. Self-knowledge of the ‘groundless’ kind does not refer to oneself, while observational self-knowledge does. Groundless self-knowledge does not ‘pick out a person’, to use Wittgenstein’s expression. It establishes the agent’s identity in a way that is not open to any question concerning the referent of the ‘I’. Perhaps it is even wrong to say that groundless self-knowledge identifies the agent. If the question ‘is it really me of whom I’m thinking when I think that I’m doing such-and such?’ is nonsensical, this is not because the reference of that use of ‘I’ is infallible, but rather because it does not refer at all. This kind of self-knowledge is not ‘of’ the self in that referential way.
(b) First-person perspective. Sidney Shoemaker spells out the unique way in which groundless self-knowledge establishes the agent’s identity. He, too, approaches the topic ex negativo, with an analysis of what an agent without that kind of self-knowledge would be missing. Such an agent would be ‘self-blind’ (Shoemaker 1996, part 1). ‘A self-blind creature would be one which has the conception of the various mental states, and can entertain the thought that it has this or that belief, desire, intention, etc., but which is unable to become aware of the truth of such a thought except in a third-person way’ (Shoemaker 1996, 30 f.). Self-blindness would not simply result in some limitation on the agent’s rational capacities, he claims. Rather, it would disqualify the agent in question as a rational altogether. A rational creature is a creature that is capable of reasoning, (p.55) and reasoning has to start from the reasoner’s own attitudes. For a creature that has only perceptual access to mental states, however, there is no way to know of a mental state that it is its own in the way required for reasoning.
Here is a version of the story that is usually used in the literature to illustrate the point (it is the same point as in the Anscombian ringing of the doorbell, only that it is attitude rather than action that is at stake here). Imagine a highly perceptive agent who knows what a person’s intentions are simply by looking at that person’s face. However, looking at faces is the only way she knows about intention. Luckily, she has a mirror, and she has her own face in view. How can she know that whatever attitude she reads off that face is hers? She knows that mirrors usually show the faces of the people standing in front of them. To make sure that it is her who is standing in front of the mirror, she can resort to another mirror that shows herself standing before the first mirror. Now she sees not only the expression in the mirror; she also sees whose expression is reflected in the mirror. But this pushes the question only one level further back: how can she know that the person whose expression is in the first mirror is herself? Further mirrors can’t help. ‘Self-blindness’ cannot be cured by self-observation of whatever high a level, Shoemaker argues, because no such observation provides first-personal access. Thus the attitudes the person observes herself to have are not available to her as her own in the way that is necessary to engage in reasoning.
(c) First-personal commitment. A self-blind person fails to see his own attitudes as his commitments, that is, as accepted constraints on his thoughts and actions. This may not seem particularly important, though. Louis C. K. once said about his most cherished beliefs and intentions—such as the belief that other people’s well-being matters, and the intention to do good—that he is very proud of the fact that he has them, but that he is certainly not planning on acting on any such attitude. He likes the ‘having’ part, he says, but that does not amount to any sort of commitment. But insofar as this is true, it seems that these attitudes are not really his after all, or not his in the right way. He may like to think he has them, but he really doesn’t. It seems plausible that an attitude that does not come with any sort of corresponding commitment to think and act cannot be an agent’s attitude.
This is brought to the fore rather forcefully in Shoemaker’s discussion of Moore’s paradox (cf. Shoemaker 1996, 34 f., 74 ff.). There is nothing formally wrong with views such as ‘p is true, but I believe it is false’, because both conjuncts may very well be true: it may be raining, while I believe that it is not raining. However, the position taken by somebody uttering such a sentence is obviously not viable—and neither is the position taken by somebody having such an attitude, and the point at stake is that such an attitude fails to take the subject’s (p.56) own attitudes as his or her commitment. The attitude may be formally sound, however, it contains contradicting commitments. By thinking ‘p is true’ the subject is committed to the truth of p, but by thinking ‘I believe p is false’, the subject is committed to the falsity of p.
A similar line is pressed by Shoemaker, who argues that perceptual models of self-knowledge fail to account for the way in which an agent’s attitudes are his or her commitments, and for the way in which Moore’s paradox is a paradox. A self-blind person with the above mind-reading capacities may look in the mirror and see that she believes that it is not raining, and then look out of the window and believe that it is raining, amounting to the belief that it is raining, and that she does not believe it. Another look in the mirror would of course quickly reveal to the self-blind person that she has changed her opinion—the person in the mirror has changed her views on the weather. But any such move reduces Moore’s paradox to a mere case of error—after checking the weather, and before looking into the mirror once more, the person mistakenly believes of herself that she does not believe that it is raining, while she has changed her opinion on the weather at the moment she has looked out of the window. Even if we remove the temporal lag by assuming an inner sense of self-perception that works simultaneously with the ‘outer’ senses, the gap between what a person believes to be the case and what a person takes herself to believe is not closed in the right way—Moore’s paradox is not just some sort of mistake, after all, but a paradox.
This carries over to Louis C. K.’s case to the degree to which a similar point can be made for the case of intention. Assume for a moment that intention is always under the guise of some good—whenever we do things intentionally we do so because we see the doing as worthwhile under some description. The practical version of Moore’s paradox would then be an assertion or attitude of the sort ‘p is utterly bad, but I intend it’. Just as in Moore’s cognitive case, that is an attitude that is unproblematic when it refers to other people’s intention, but it is not an attitude that can be taken towards one’s own intentions.
A self-blind person would not be committed to his or her own views in a way that would make any form of reasoning possible, because he or she would not be bothered by any inconsistency between his or her own attitudes. The third-personal insight that he or she has it wrong cognitively, or that his or her intentions are thoroughly misguided, would amount only to the view that there is somebody who got it wrong and who is thoroughly misguided—which is none of his or her own business. There is simply no bridge between the insight that somebody obviously has a problem to the insight that this problem is his or hers.
(p.57) (d) First-personal authority. A further point to be made for the case of groundless self-knowledge is that it seems to explain the special authority agents seem to have over the aspect under which it is intentional. This authority consists in the way in which the agent’s view of what it is he or she is doing settles the question of under which description it is intentional. This authority may not seem obvious. You think you’re calling your friend, but you type in your parents’ number. Where is your authority? An outside observer may well be in a better position to know what it is you’re doing. Still, you may claim that calling your friend was what you intended to do. But even authority over intentionality may not seem obvious. To modify an example discussed by Velleman: You quarrel all evening with your friend, and it is only later on that you realize that you intended to break up that friendship—a fact which you denied when your friend brought it up; another person was obviously in a better condition to say what it is you were doing intentionally. Yet again, any such case is a challenge to the range of first-person authority, not to its existence. After all, it was intentionally, and known to you to be so at the time, that you brought up all these sensitive issues in the discussion—you did not just sit there and listen in surprise at these words coming out of your mouth. Thus there was something you did know non-observationally, even though you may have been unaware of, or mistaken about, the reasons why you did what you did. Quarrels over what it is you’re doing may arise in the interpretation of the relevant description of your action, but any such dispute does not displace you as an authority over what it is you’re doing. Within his discussion of Anscombe’s groundless self-knowledge, Richard Moran (2001, 124 ff.) argues that the ‘knowledge’ in question should be more adequately conceived of as a ‘making’. It is no wonder that people should seem to have special epistemic authority over what it is they intend, Moran argues: they are the authors of these states after all, that is, the ones who have made up their own minds accordingly. One special feature of Anscombian ‘self-knowledge’, in other words, is just the fact that it is a product of agency (cf. O’Brien 2003). According to the view that emerges from Moran’s (and others’) view, first-personal authority is the feature in virtue of which the agent is ‘in control’ of his or her attitudes (Moran 2001, 147). Thus, first-personal authority marks the agent’s place in action. It does not exclude the possibility of a great deal of estrangement, as is the case in weak-willed action. But such estrangement cannot be total, or agency breaks down.
Moran and others have attempted to demystify ‘groundless belief’ by arguing that it is no wonder that our relation to our own mental states should be special, given we are the ‘authors’ or makers of our own attitudes by making up our mind (Moran 2001)—a view that seems especially plausible where intention is concerned, because it may appear that whereas we’re not free to choose what to (p.58) believe, intention in terms of decisions what to do is ‘up to us’ to a somewhat wider degree (cf. Paul 2009. Other authors attack the view that self-knowledge is somehow basic and ‘groundless’ head-on. Shoemaker has been accused of conflating introspective awareness and self-knowledge, and of implying that a quasi-perceptual ‘inner sense’-view of introspection can be maintained, amounting to the view that the self-knowledge in question is perceptual after all, only that the sense in question is of a particular kind (Kind 2003; Armstrong 1981). Looking at the literature, it is far from obvious that the ‘Anscombian view’ is generally accepted, even if it is as loosely conceived as in the remarks made in this section. Many of the arguments sketched above are controversial. But I think it is plausible to say, at least of most of the competing accounts, that their aim is not to deny first-person identity, first-person perspective, first-person commitment, and first-person authority as such, but rather to account for it in a different way—for instance, by relying on introspection rather than on non-observational self-knowledge. This is enough for the line of argument of this chapter to proceed, because it seems that any account of self-knowledge that features first-person identity, perspective, commitment, and authority is likely to run into at least some of the problems to be discussed in the next session. The problem is that any knowledge of what it is we are doing together involves knowledge of other people’s actions. And it seems that any such knowledge needs to be of the observational or inferential kind. You may know what you are doing ‘just like that’—but not what others are doing. Looking at the matter this way, it does not seem difficult to argue that knowledge of what we are doing together does not come with first-person identity, perspective, commitment, and authority.
2. How Do We Know What We’re Doing Together?
Typical cases of joint actions are intentionally joint. When we go for a walk together we do not coincidentally walk alongside each other. The joint walk is intended by us as a unit. The fact that a joint action is one token action with many participants is sometimes downplayed in the literature, and the reason may be that it is felt that this compromises the status of the individual participants as agents. But there is no reason for such worries, as the analogy to complex individual actions makes clear. Complex individual intentional actions often consist of different parts that are themselves intentional actions. Filling the coffee machine with beans and water, switching it on, taking the milk can from the fridge are intentional actions that combine to constitute one token intentional action, the making of the coffee, without thereby losing their status as intentional actions (cf. Anscombe 1957, §§22–3). Analogously, the status of the participating (p.59) individuals as agents is not undermined if their individual contributions combine to a single token joint action. But it is clear that not just any aggregate of individual intentional actions will do. Just that I happen to fill the coffee machine, you happen to switch it on, and he happens to get the can of milk, together with similarly coincidental other actions, does not mean that there is one token action that is jointly intended. What is it, then, that provides the form of unity that is necessary for an aggregate of individual acts to be components of a joint intentional activity?
At least one attempt has been made in the received literature to give a straightforward Anscombian answer to that question, applying her approach to intentional action to collective action (Laurence 2011). Ben Laurence uses Anscombe’s view on how single acts may be parts of one and the same intentional action of making tea as the starting-point. The focus is on the non-accidental nature of the actions, which is explained with the special form of mutual explanatory relations between the overall intentional action of making tea, and the component actions (warming the water, preparing the cups etc.), as well as between those component actions. ‘The several actions thus share an explanatory unity’ (Laurence 2011, 277), and it is not difficult to show that the same kind of ‘non-accidental fit’ is also in place between the participant actions of the members of a joint action—if we are intentionally walking together, I walk and you walk because we’re walking together, and vice versa. This is the basic line of this ‘Anscombian Approach to Collective Action’, and it is summarized as follows: ‘What emerges from the comparison of individual to collective action is that when a group is acting together, the actions of the individual group members can be straightforwardly instrumentally rationalized by an action being performed by the group’ (Laurence 2011, 281); if this ‘straightforward instrumental rationalization’ is possible, the participants can be said to share one token intention (Laurence 2011, 283). This approach nicely explores the analogy between individual action and joint action, and it does place the group in the role of the agency in a way that does not displace the participating individuals from their role as agents. But this does not so far place the core Anscombian idea in the analysis: what about the particular ‘groundless’ way of knowing what it is we’re doing together? Laurence argues that it exists, even though this may not be true of all members. ‘Some suitably placed person(s) must know without observation what purposes the group is pursuing. A group cannot be said to φ intentionally, if none of its members know that it is φ-ing, or even if some do but only through observation.’ The argument concludes by claiming that group groundless self-knowledge of that sort does not imply any group mind. Collective action is action by a unity of embodied individual minds: ‘In light of this, why not say that a (p.60) collective agent knows what it is doing, and why, without observation.’ Laurence is, however, not entirely comfortable with that claim in the end. He emphasizes that ‘knowledge without observation’ is a purely philosophical term invented by Anscombe for a specific purpose, and he admits that it might be ‘laying it on a bit thick to speak in this way’ of groups (Laurence 2011, 293). Groundless group self-knowledge, in the view argued for in Laurence’s paper, simply amounts to the claim that practical reasoning works in the same formal way in collective action as in individual action.
Let us grant that we reason from a we-perspective if we act together. But as the above discussion has suggested, we need to dig somewhat deeper in order to be justified to call any such we-reasoning ‘groundless self-knowledge’ of joint intentional action. After all, we have encountered a whole series of features of what groundless self-knowledge seems to imply. Before jumping to the conclusion that the same self-knowledge applies in the case of joint action, we better examine how this works with each of these features. What about first-person identity, first-person access, first-person commitment, and first-person authority in the plural case? It seems that if an account of collective action aspires to be straightforwardly ‘Anscombian’, these issues have to be addressed.
This is where the problems with the assumption of groundless self-knowledge of joint action start. None of the features encountered in the analysis of groundless self-knowledge carries over smoothly to the case of joint action, as shall be argued below. It seems that whatever first-person identity, first-person perspective, first-person commitment, and first-person authority there is in the plural, it works nothing like the singular case. To the degree to which this can be made plausible, it seems that groups cannot really act intentionally in a straightforward sense, and that what groups can do is really more like an emulation of one token intentional action that individuals can pull off if they combine in a suitable manner. The joint action in question will then not really be one token intentional action, but a complex of actions patched together so as to look as token-like as possible. I will make the case against groundless group self-knowledge in this chapter, arguing against the possibility of something like the singular kind of first-personal identity, perspective, commitment, and authority in plural case, before turning the tables in the concluding section.
(a) First-Person Plural Identity. The first difference seems rather obvious. Self-knowledge of one’s own individual actions may establish the identity of the agent in an immediate way that involves no referential ‘picking out’ of any kind. This does not seem to work in the plural. As a participant, it appears that you don’t know who’s with you in the same way in which you know you yourself are (p.61) involved. The only non-referential, immediate, and perhaps doubtless feature of any ‘we’ is the ‘I’. Any attitude of the form ‘We are doing x’, ‘we intend to x’, ‘we believe that x’, is clearly quite heavily involved in the picking-out business, and may well go wrong accordingly. There is no question to whom the ‘I’ in the ‘I’m doing…’ refers—because it does not refer, in Anscombe’s view, and even if the line is taken that it does refer after all, the reference in question can plausibly be claimed to be of the kind that is not open to doubt. Any plural version of such attitudes, however, does leave ample room for mis-identification, and is thus subject to reasonable doubt. No plural identity is established ‘immediately’ by somebody having an intention of the plural form. I might think that ‘we intend…’, but I’d better know who’s in and who’s out, and that knowledge does not come ‘just like that’. I may mis-identify the members of the relevant group (I mistakenly take you to be one of those of whom I think in terms of ‘we intend…’), or there may not be any such identity at all, such as when I’m dreaming that we’re doing such-and-such with nobody around. While the self of groundless individual self-knowledge is somehow special with regard to its identity, no plural self that figures in some individual’s attitude can aspire to any such privilege.
This carries over to the way in which the knowledge of what we are doing combines to constitute knowledge of a single token action. It seems that joint intentional activity resembles much more the initial fridge case, where I know ‘without ground’ the ‘current bit’—my opening the fridge—but have to resort to recollection and perception as to the question of what it is I’m doing. If we are walking together, I may have whatever epistemically groundless (or perhaps introspective) knowledge of what it is I am doing. As far as your part is concerned, however, I’d better recall correctly what exactly it was we’ve agreed to do together. But that’s not enough. Just that I recall our agreement, and thus normatively expect you to do your part, or that I am engaged in some such instrumental or explanatory reasoning, doesn’t make it the case that this is in fact what we’re doing. I may still be mistaken. Eventually, I’ll simply have to watch and see; what I take ‘our intention’ to be does not settle the question of what it is we’re doing together in the same way it does in the case of my own intentions. The question ‘how do we know what we are doing?’ does not seem to be epistemically on a par whether it is asked in the distributive or in the collective sense. Distributively, each of us may know ‘without ground’ what he or she is doing; collectively, however, as a team, we do not seem to have any such knowledge.
If you’re hiking up a narrow path, and I ask you what it is you’re doing, you’ll be able to answer, ‘I’m climbing that mountain’, just like that. If, however, you’re climbing that path together with a friend who is walking behind you, and I ask (p.62) you what you’re doing, you’d better turn around and see if he’s still with you before answering, ‘we’re climbing that mountain’. This may not seem much of a difference in principle. The fact that the other is still going along with you—a matter of perceptual knowledge—may seem to matter no more to your climbing the mountain together than the existence of the mountain—another matter of observation—matters to your individual climbing venture. Yet an important difference that is at stake here is brought to the fore when intention rather than action is at issue. You may know what it is you intend, individually, and perhaps you know it just like that. As to the intentions you take to be jointly held, with others, you’d better be able to quote some evidence for what it is the relevant other people have in mind. One can be painfully mistaken in assuming of an intention that it is ‘ours’ in a way that does not seem—at first glance at least—to have any correspondence in the individual case. You may mistakenly think that you’re sharing an intention in a way that differs from individual intention. Perhaps you misunderstood the other person whom you have met on the trail, and you assume now that the climbing is something you’re doing together, because that’s what you have agreed to do, while he still thinks of the hiking that is going on just in terms of a coincidental case of one person walking behind another, and not in terms of one token action. In that case, you’re wrong at a level that does not seem to have a parallel in the singular case.
(b) First-person plural perspective. The difference is further highlighted if we focus on the ‘how do you know?’ question. If you tell me what you intend to do, individually, it does not make much sense for me to ask how do you know what it is you intend. You just know—that’s it. But if you tell me, that is what you intend to do together with your partner, no such reply seems to be possible. You don’t ‘just know’. You’ll have to quote some evidence, and you are likely to reply with: ‘That’s what we have agreed to do, and here we are’, or some such. In that sense, too, joint intentional activity seems to be deeply different from individual intentional activity. Whatever knowledge of what it is we are doing together cannot be basic, but implies individual self-knowledge and observation. Neither of us has immediate awareness or introspective access to our intention to go for a walk together, but only to whatever individual contributive intentionality we have, individually. As no opinion an individual might have of whatever she thinks she is doing together with others guarantees joint activity, and as our knowledge of what we are doing, in typical cases at least, involves a great deal of observation and inference, there is a sharp contrast between individual and joint action (remember, though, that this section is playing the individualist’s advocate, and that a different line of argument will be developed in the next section).
(p.63) (c) The same applies to the question of first-person plural commitment. If a team or group of agents jointly settle on a view of what is true and what is good, this places the participants under some normative constraints—if ‘we’ believe that x is true or good, each of us can be normatively expected to treat these attitudes as valid in our reasoning. Whatever normativity there may be in attitudes that are ‘ours’, however, it is clearly not of the kind of commitment that is involved in one’s own intentional states. This comes to the fore in Moore’s case. ‘P is true, but I believe it is false’ is not a viable position; ‘p is true, but we believe it is false’, however, is actually quite frequent in our lives, as anyone who has ever participated in a department meeting knows (for example, your department has settled on a view of how its problems can be solved—you, however, know perfectly well that the planned measures will get your department only deeper into trouble). And it is not conceptually incompatible with the capacity to reason; in fact, it is often by the most reasonable group members that this attitude is taken. This does not only apply to belief, but also to intention, if we accept the analogy I’ve sketched above, based on the sub specie boni-principle. There’s something fishy with the attitude expressed by ‘x is bad, but I intend it’, while it is deplorably frequent that we find ourselves, as members of a team or some group, in a position to think ‘x is bad, but we intend it’—this is especially obvious to citizens of direct democracies, or perhaps to teenagers who have tried in vain to convince their parents of what makes for a truly good holiday. The decision we’ve made is ours, and thus is an intention. A proponent of some minority position is not thereby excluded from the group that has the intention; whatever commitment is involved in the group attitude involves them as well. But any such member may be firmly convinced that the intended course of action is utterly bad. Tough as it may be to be in this situation, there is clearly nothing paradoxical about it. Again, there is no parallel between the case of individual and collective intentional action, as far as the features of the ‘groundless knowledge’ of what it is we’re doing is concerned. Right as Laurence may be in observing that joint action involves reasoning from a shared point of view, it is obvious that joint intentions or shared beliefs do not commit us in the same way as individual intentions and beliefs do, and the difference at stake here seems to be a fundamental one, not one of degree.
(d) This brings us to the last feature of the ‘deep structure’ of groundless self-knowledge, the issue concerning authority. In the above sketch, a self-blind person was characterized as a person who refers to him- or herself only in the third person. Now it is obvious that even though we do have a first-personal way of referring to our groups—the collective use of ‘we’—the third-personal approach to a group to which one belongs seems quite normal, such as when (p.64) we say ‘the department does x’ rather than ‘we do x’, and that way of expressing group attitudes seems to reflect a difference between two ways of thinking of those attitudes. If we call illeists (from latin: ille) those individuals who refer to themselves only in a third-personal way, we should call illiists (from illi, the plural of ille) those who refer to their groups only in a third-personal way (cf. Schmid 2014b). Illeism does occur—American basketball players are famous for it, and parents seem to be prone to it, too (‘Daddy’s going to pick you up’)—but it is rare, and if it were not just a manner of speaking, but a question of the kind of self-knowledge a person has, it would certainly be pathological. Illiism, however, seems to be frequent, and perhaps the standard case. Some people know that they are members, but they simply do not take a first-person-plural stance. They may go along with whatever group attitudes require them to do, but they do not ‘identify’ with the group perspective in the way that is expressed in the subject use of ‘we’. It is not initially obvious at all that such attitudes on the part of the group members should make it impossible for agents to act jointly. Thus it seems that any such attitude is not necessary for joint action. Granted, there are cases of individual action that are similarly alienated—for example, when a person is coerced by another person, or acts on desires which she disowns, such as in the case of strong addiction. In such cases, the agent does not see herself as the ‘author’ of her action in the way expressed in the subject use of ‘I’.1 But this parallel highlights the basic difference between the singular and the plural cases—what is quite normal in joint action, if the above remarks are correct, is certainly an unusual condition in individual action.
Another question is if group identification, where it occurs, mirrors first-person-singular authority. Imagine a person who is closely identified with his group and does a lot of his reasoning in plural terms. He does not see his group membership in terms of an external circumstance, but as a part of his identity. It seems clear, however, that he does not thereby acquire any sort of special epistemic authority concerning what it is the group thinks, intends, or is doing. His position with regard to the group’s intention is not privileged in principle, compared to an outside observer (for an extended version of the following argument, cf. Schmid 2014b). Even a person who conceives of his whole life in terms of his membership may misunderstand his group’s attitudes, as any church member who has made the effort to find out what it exactly is that ‘we’ believe knows all too well.
If group identification does not constitute first-person plural authority, however, perhaps authorization does. List and Pettit (2011, 35 f.) argue that some (p.65) groups, or corporate agents, have spokespeople that are endowed by the group with a special authority to express the corporation’s attitude. Of such people we assume that they are in a privileged position to know what the corporation believes and intends. But there is serious reason to doubt that authorization by a group leads to first-person plural authority. There may be good reason to trust a spokesperson more than any other observer, but that does not imply a difference in the kind of knowledge a spokesperson can rationally be assumed to have, as opposed to an observer. Even the best-informed spokesperson for a business enterprise has to rely on information concerning the board’s decision, and any journalist in the audience of his or her press conference may happen to have better information.
The only promising candidate for being in the kind of ‘know’ that is of truly first-person authoritative kind would be some sort of dictator, or perhaps an all-powerful CEO who also happens to own the company and dominates its board of trustees. In that case, it is really up to him or her what the corporate attitude is. He or she is calling all the shots concerning what it is the group is doing. In that case, the corporate intentional action really does reflect his or her rational stance, and the authority structure highlighted by Moran and others is fully in place. Such cases may be possible, but by the same token, any such example would not be a case in point; it would not be a case of joint intentional activity. Rather, it would be a case in which one agent’s capacity for intentional action is extended; in the way in which it is intended by that authoritative individual, the action in question is not the action of a plurality of participants teaming up in acting together, but rather the individual intentional action of one individual that has increased powers because he or she can count on other people to do whatever she or he wants. Thus, instead of an analogy we seem to end up with a contradiction: wherever there is first-person plural authority, there cannot be joint intentional activity.
3. Plural Pre-Reflective Self-awareness
In the light of the above considerations, speaking of groundless self-knowledge of intentional joint action, or groundless group self-knowledge, does not just appear as ‘laying it on a bit thick’ (Laurence 2011, 293). It seems outright wrong, and whatever remains of the ‘Anscombian approach to collective action’ may seem like it cannot amount to much more than a faint analogy. I think, however, that we should resist this conclusion, and that it is now time to turn the tables. The previous section zoomed in on the dis-analogies between the singular and plural case with regard to the features we have identified as elements of groundless (p.66) self-knowledge. But there is a positive side to explore, too. True as it may be that the knowledge of the participants in joint action cannot be of the exact same sort of knowledge of individual action that Anscombe labels ‘groundless’, it seems equally wrong to equate the knowledge in question with pure perceptual knowledge. And with a bit of good-will, one can even find some leads towards an account of genuine groundless group self-knowledge in the received literature. That it is not exactly the same should not come as a surprise. After all, the plural is not just another singular.
(a) First-person plural identity: the paper in which the label ‘collective intentionality’ was introduced in the debate ends with a section on ‘Presuppositions’ (Searle 1990, 414 f.). Searle here argues that it is not enough for people to engage in a joint intentional activity just to know one’s own intention, and to know what it is other people intend, and to know that there is some structure of mutual awareness between the agents. Collective intentionality, he claims, involves a pre-intentional ‘sense of “us”’, an ‘awareness of oneself and others as potential or actual cooperators’, and he even uses terms such as ‘sense of community’ or ‘communal awareness’. I have followed these leads elsewhere, and I have argued that the ‘sense of “us”’ in question here is plural pre-reflective self-awareness, that is, the participants’ awareness of what it is they’re doing together as their joint action, collectively. The ‘us’ is not of the sort that it figures in the content of the attitude in question. Rather, it is the subjective aspect of the consciousness in question, that is, the way in which whatever ‘what it is like’ there is to consciousness also involves a ‘for somebody’. In the case of plural consciousness, that subject is not ‘me’ but ‘us’ (Schmid 2014a). Pre-reflective self-awareness is explained in contradistinction to self-knowledge, and the decisive difference is in the way each of the attitudes is ‘of’ oneself. Self-knowledge is ‘of’ the self in terms of the self being the object or ‘theme’ of one’s attitude, such as in the case in which one thinks how a new hairstyle would suit one’s features, while self-awareness is of the self in the way of the subjective aspect of consciousness. The distinction at stake here has been analyzed most clearly by Jean-Paul Sartre in his 1947 paper on ‘conscience et connaissance de soi’. Sartre’s distinction fits Anscombe’s distinction between the groundless and the perceptual kind of knowing oneself seamlessly.2 The central idea that I take to be implied in Searle’s remarks on the ‘sense of “us”’ is that Sartrean pre-intentional self-awareness does not only come in the singular. It sometimes comes in the plural, too. Sometimes it (p.67) is our pre-intentional sense of our individual self that comes with whatever it is we’re experiencing—if you’re reading a book, the reading experience is transparent to you as yours. If we’re engaged in joint action, however, the pre-intentional sense in question is likely to be of the plural sort: if we are walking together, the walking experience is transparent to us as ours, collectively. And true as it may be that no such ‘sense of “us”’ is infallible, it is immediate, nevertheless. It is that pre-reflective sense itself that answers the question who it is that the agent takes to be in and out, not some knowledge of other people.
(b) First-person plural perspective. In the previous section it was argued that Shoemaker’s problem of self-knowledge does not carry over to the plural case. But that is only half of the story. There is such a thing as ‘plural self-blindness’, that is, the incapacity to see the attitudes one shares with others first-personally. Let us draw the parallel as closely as possible, and define as plurally self-blind a creature who has the conception of the various mental states a group may have, and can entertain the thought her group accepts this or that view as true, and has this or that intention, but who is unable to become aware of the truth of such a thought except in a third-person way (say, for example, by checking the minutes of the last group meeting). Other than what she knows in that way she cannot ‘see’ as a jointly accepted attitude. She has no other sense of what her group wants and accepts, that is, she has no pre-reflective ‘sense of “us”’, but only a sense of her individual self, together with observations of decisions and actions of her group, which she takes to be ‘hers’ only in the way of an illiist sketched above.
Contrary to the singular case, first-person-plural blindness does certainly not disqualify an agent as a reasoner. However, it seems that any such condition does disqualify him or her as a joint reasoner on two accounts. First, joint reasoning is a cooperative process of discussion over what it is we should accept as true, and what we should do. However, the question of what should be done cannot be approached cooperatively from the point of view of a person who is plurally self-blind, because that person can take only his or her own attitudes as a base of whatever reasoning process in which he or she is engaged. For a plurally self-blind person, other people may be valid sources of information, and the discussion with them may serve as an important means to test and improve his or her own attitudes, as well as to influence those other people’s views. However, using others as sources of evidence, or as targets of one’s influence, is not partnership in a cooperative reasoning process.
From the point of view of a plurally self-blind person, the only possible answer to the question what a group should do is the following: the group should do whatever I think is best (for the group), in the light of the reasons I have. In (p.68) contrast, joint reasoning is predicated on the assumption that what we (together) should do is what we think is best in the light of the reasons we have. From a first-person plural perspective, the participants should articulate what they individually think is true or good, but it is only on the base of agreed-upon attitudes that the group has reasons to act, and this is what a plurally self-blind person cannot see. However benevolent towards others a creature with plural self-blindness may be, and however rational he or she may be as an individual, he or she simply cannot engage in joint reasoning—in that sense, the singular and the plural case seem to be parallel after all.
(c) First-person plural commitment. The claim that a plurally self-blind person cannot be a joint reasoner becomes clearer if we move on to the issue of commitment. Take again the person who is plurally self-blind, and has only third-personal knowledge of the attitudes of the group of which he is a member—for example, he registers the majority decisions that are made in his group, and he knows that the way in which group decisions are made is based on that process. In many cases he will find himself to have views of the Moorean kind: ‘p is wrong, but we believe it’; ‘p is bad, but we intend it’. For a person who is plurally self-blind, this is just another case of different and perhaps mutually incompatible attitudes held by different agents. Thereby, he fails to see the situation for what it really is: a problem that needs to be worked out in joint reasoning. It is true that divergence in opinion between individuals and groups always exist, even in healthy democracies; but it is equally true that a healthy democratic culture cannot survive for long if nobody cares in the slightest about whether or not what the group believes or intends and what he or she believes or intends are in line. Whoever thinks that what we believe or intend is wrong or bad should speak up and initiate a process of joint reasoning. Attitudes like ‘p is bad, but we intend it’ are not paradoxical. But among joint reasoners, such cases certainly mark a tension, or a deficiency of the relevant reasoning process. The situation is not unlike the case in which an individual reasoner makes up his or her mind, but finds some parts of his or her own mind remaining unconvinced of the stance he or she has taken. A plurally self-blind person fails to see a situation in which his or her view differs from the group’s in the Moorean way as being of that kind of a problem. It is true that any real democratic process, especially in larger groups, demands of us that the joint reasoning be terminated at some point, and that the joint decision be accepted even by those who disagree. But joint deliberation would not even get off the ground if nobody perceived a contradiction between what he or she thinks and intends and what his or her group accepts and intends as more than just any divergence of opinion. For a plurally self-blind person, there is no way to see that what ‘we’ think and intend should ultimately be what (p.69) each of us has reason to think and intend—without any further reason such as fear of nonconformity, loss of reputation, sanctions, or some such, but simply in virtue of the attitude in question’s being ours, collectively.
A plurally self-blind person identifies with the group only in a third-personal way: ‘There is the group (of which I know I’m a member). And here is where I stand’; and she fails to identify with the group in a first-personal way: ‘The group is us.’ One of the places at which this issue of first-personal identification with the group has been addressed in the received literature is in the last chapter of Philip Pettit’s and Christian List’s book on group agency. The authors argue here that group members need to think of their groups in we-terms. They need to collectively self-identify as a group agent, that is, to recognize that the group agent is ‘us’ (List and Pettit 2011, 186 ff.). They need to see the group attitudes as their (joint) commitments in a way that does not open what Pettit and List call the ‘identification gap’. The identification gap is between what one believes the group to intend or believe, and what one takes to be one’s own concerns: this is what that group intends to do—but why should I bother? Pettit and List see two ways of closing the identification gap. The first is based on a further thought. It requires of the individual members that they are disposed to do what the group attitudes require, based on the knowledge, ‘well, I’m a group member after all’. According to Pettit and List, this is not enough. Identification cannot be left to any such extra ‘cognitive achievement’. The identification has to be a ‘by-product’ of the group attitude itself—the ‘by-product model’ is the preferred way to close the identification gap. In that model, there has to be a feature in virtue of which it is a direct consequence of the attitudes of the group that the participants are disposed to go along. Put in other words, the group’s attitude has to figure as an element in the individual’s deliberative base in a way in which it is not dependent on the agent’s thought that he is a member of that group.
Here is how Pettit and List imagine such a ‘by-product model’ identification with the group to work:
An analogy may help to explain the idea. Consider how pilots relate to cockpit instruments that give information on the plane’s altitude, orientation, speed, and so on. Beginner pilots may consult the instruments and let the evidence count in their own reasoning process. If the horizon is out of sight, they may let panel information weigh against the information from their own senses; the latter may be misleading, since the senses cannot distinguish, for example, between gravity and acceleration…Rather than sustaining a relationship whereby the instruments are consulted, experienced pilots develop a different, direct connection with them. They let their own bodily cues go off-line, and hitch their intuitions and instincts directly on the instruments. They let the instruments guide them without the intrusion of thoughts about the evidence provided. Or at least they do this when the ‘red lights’ are off, and nothing indicates that things are (p.70) amiss. Just as pilots can connect in this direct way to the instruments on the cockpit panel, so the members of a group may connect themselves directly to the attitudes of the group. They do not treat the group attitudes as mere indicators of what the group is to do, asking themselves explicitly whether they wish to identify with the group, and acting only if they have this wish. Rather, the individual attitudes are under the automatic guidance of the group, so that they can respond as spontaneously as pilots do when they take their cue from the panel before them. Or at least they may do this where there are no ‘red lights’ that suggest they should hesitate and take stock.
(List and Pettit 2011, 192)
The basic idea of the problem of the identification gap, and of the ‘by-product’ model of closing it, is exactly right and important, but the way this is further analyzed in the above quote is certainly problematic, and here is why: to the degree to which an individual is accustomed to base his reasoning exclusively on another agent’s beliefs and desires, he or she is not intentionally autonomous—his or her behavior does not instantiate his or her own action, but rather the action of the agent whose primary reasons rationalize that individual’s behavior (cf. Schmid 2009). The way in which the participant’s disposition to go along cannot be that sort of a by-product of the group’s attitude, or else only the group, and not the participating individuals, would be agents. And even in List’s and Pettit’s account, this is not really what is meant, because List and Pettit emphasize that it is only in the light of the individual’s knowledge that there are ‘no red lights’ on that they simply trust the group attitude. Therefore, what List and Pettit are suggesting here is not really a by-product model after all, but rather a version of the cognitive achievement model. The idea is that whenever people think it is okay for them to go along, they are disposed to act on the group attitudes. But this model uses the very element of which the by-product model was supposed to be free, namely, an individual attitude in light of which it makes sense for the individual to go along. So the dilemma is: either this account undermines intentional autonomy, which makes it implausible, or it isn’t really a by-product model of collective identification.
The question is: is there a way of conceiving of collective identification without further cognitive achievements that is compatible with individual intentional autonomy? I submit that there is. The way collective self-identification is a by-product of an attitude is in virtue of the individuals’ awareness of an attitude as theirs, collectively. The plural self-consciousness involved in this is not a further attitude that is added to the awareness of whatever it is that is in the focus of the consciousness in question. If I’m aware of our walking together in the right way, that is, first-person plurally, the awareness of that activity as being ours is part and parcel of the awareness of the walking that’s going on. At the same time, it clearly does not undermine the status of my contributive action as mine if I do my (p.71) part based on that first-person plural knowledge of what it is we’re doing. I’m not acting on the ultimate base of an attitude that is another agent’s, and not my own. I’m acting on attitudes that I share, that is, attitudes that I have, only that the way in which I have them is jointly with others. In that way, plural self-awareness of an attitude as ours closes the identification gap without further cognitive achievement, without thereby disqualifying the participating individuals as proper agents.
(d) First-person plural authority. The last point of comparison is the problem of first-person authority. This may seem to be the most serious obstacle to the idea of groundless group self-knowledge. As argued above, no single individual in a group can rationally aspire to being in a position to express her group’s attitudes in a way that comes with true first-person authority. All a member can express with authority is that she thinks that the group wants, or believes, or intends to do—because whatever it is she pre-reflectively takes to be the attitudes she has jointly, with others, is not only up to her, but up to others as well. If a single individual were in a position to ‘make up‘ her group’s mind, as it were, and thus speak that mind with first-personal authority, that group would not be the agent of a joint action, but simply the extension of her individual agency. One way around this would be to claim that while no individual can express (and thus know) her group’s attitudes, perhaps the group itself can. Ideas such as Raimo Tuomela’s ‘chorus sense of “we”’ come to mind here, the case in which the group’s attitude is expressed jointly by the group members, instead of just a single individual. Among the doubts I have concerning the view that some such collective speech act expresses a group attitude with first-personal authority is that it is not clear by the mere fact of the collective speech act that the participant individual speakers thereby have the same ‘we’ in mind (for a more careful argument, cf. Schmid 2014b). And that, of course, carries over to the kind of knowledge at work in these cases. No fully authoritative knowledge of what it is we’re doing together is available to us, however aligned our views thereof may be. But the fact that there is no such thing as full first-person plural authority certainly does not prove that any knowledge we have of what it is we’re doing together is third-personal. It does not even show that knowing what we’re doing together comes with no first-person plural authority at all. I submit that a weaker form of first-person plural authority is widely accepted even in our everyday practice. We certainly do not grant the members of another group a privileged position as to the question of what it is their group is doing simply because we assume they are in a better position to observe what is going on among them. Rather, we grant them that privileged position because the attitudes of their group are what they, jointly, make them to be, and are, thus, they are co-authors (p.72) of their attitudes in quite a literal sense. The disanalogy shrinks to the insight that the ‘making’ of a group attitude does not come with the sort of centeredness that is typical of rational individual minds.
One reason why the first-person singular viewpoint is granted special authority is the alleged impossibility of misidentification. My predication of what it is I think or intend may be wrong (such as in the case in which I mistake my temporary infatuation for true love), but the identification that it is my attitude cannot be mistaken. And it seems that there is no analogy for this in the plural case: I may easily be mistaken in assuming that an attitude is really ours. Here, my strategy is that of companionship in guilt: just as a person may be mistaken in taking an action or intention to be joint, when it is really just his individual own, he may be mistaken in taking an action or intention to be his individual own, when it is really joint. Take the case of you and me deciding to write a paper together. The process is a prolonged one, and we rarely talk about the progress we’re making in developing the argument of our respective parts. As I am a bit egocentric, however, I somehow come to see the project as my individual own after a while, and that is how I see what I am doing: ‘I am writing this paper.’ Yet I’m simply mistaken in the identification of the subject of what I am doing; I am misattributing the intentional action, or even just the intention, to me, when it is really ours, as I come to realize on the occasion of our next conversation. In that way, I was mistaken in the subject of the action in question. Mis-identification does happen in the plural, yet thinking about joint action makes us realize a way in which it may happen in the singular, too. If first-person authority is an impossibility of mis-identification, it exists neither in the plural nor in the singular.
It is time now to come to a conclusion. Where does all of this leave us? I have argued that the idea of groundless self-knowledge of what it is one is doing can be spelled out in terms of first-personal identity, perspective, commitment, and authority. I have argued that, in spite of the obvious differences between the singular and the plural cases, the idea that the participants have groundless knowledge of what it is they are doing together makes good sense. There is first-person plural identity, perspective, commitment, and authority at work in joint action, and thus groundless group self-knowledge. Yet there are remarkable differences between the singular and the plural form of groundless self-knowledge. This should not be surprising. After all, the ‘we’ is no collective ‘I’.
I wish to thank Miranda Fricker, Michael Brady, Katharina Bernhard, the participants of the Social Ontology Colloquium at the University of Vienna, as well as an anonymous referee for helpful comments.
(1) I owe this example to Miranda Fricker.
(2) Moran has clearly missed a chance here—he draws heavily on Sartre, but he quotes only from his earlier opus magnum.