Under titles such as “collective intention”, “shared intention”, or “joint intention”, attitudes – or states of mind – of the form “we intend” have received increasing attention in recent philosophical research. The accounts given in the literature differ widely from each other, but there are also some points of agreement. It seems to be generally accepted that attitudes of this form need not necessarily be fully worked out thoughts. “We intend” is not necessarily a linguistic entity – need not come in words, as it were –, and may be pre-conceptual. For there to be an attitude of this form, all that is needed might be some sort of sense of what has to be the case for us to have done what we intended to do (conditions of satisfaction).

Attitudes of the form “we intend” include some action, φ, that is intended in “we intend”, and a further point of agreement in large parts of the literature seems to be that there are cases in which the φ-ing in question is collective rather than distributive, and that these cases are not (or not easily) reducible to distributions. In the collective case, the φ-ing that is intended in “we intend” is not a type of action of which each of us intends to do a token. Rather, it is one token action all of us intend to do together. The difference between a distribution and a collection seems to be under-analyzed in the literature, but the core idea is clear enough: where the action intended in “we intend to φ” is collective, it is one rather than many: one token action, many participants.

Though it is rarely spelled out in exactly these terms, this focus on the collective case seems to be at the heart of the current debate on collective intentionality. The nature and structure of such intentionality, however, is controversially discussed. When Schweikard and Schmid (2013) wrote the Stanford Encyclopedia-article on the topic, we decided to use a distinction from earlier analysis of intentionality to identify the different “camps” in the current debate and to capture what the controversy might be about. The distinction is between intentional content, intentional subject, and intentional mode. In the case of mind of the form “we intend to φ”, the “φ” is the content, the “we” is the subject, and the “intend” is the mode.

Put in these terms, the main controversy in the current debate is this: is it just the content that is collective about collective intentionality (content-accounts, e.g. Michael E. Bratman’s), or is there something collective about subject and mode, too (mode- and subject-accounts, e.g. John Searle’s for the former, and Margaret Gilbert’s for the latter)? While everybody seem to agree that the intentional content is collective, content-accounts argue that it is just the content that is collective. In this view, the the φ in “we intend to φ” is collective (one token joint action rather than just a distribution of individual actions), while the “we” is distributive. The view is thus that in “we intend to φ”, each of us intends our collective φ-ing. A strong argument against the content view is that intention is action self-referential – subjects intend their own actions only –, and that therefore the subject of the intention has to be the subject of the intended action. In the same sense in which I cannot directly intend your action, but only intend to bring it about that you act (where it is my “bringing it about” that I intend), I cannot intend our action, and neither can you; if each of us can intend only what he or she herself can do, neither of us can intend our φ-ing, as neither you nor I can do it. Our collective φ-ing – as one token joint action – is therefore not something we can intend distributively. The easiest way out of this predicament seems to be this: If you and I cannot intend our φ-ing distributively, we, collectively, can. As intention is action self-referential, it takes a collective subject for a collective action to be the content of intention. Subject accounts thus argue that both content and subject of attitudes of the form “we intend to φ” are collective.

The big problem with this easy solution is the following. From the early times of Wilfrid Sellars’ thinking on the matter (for references cf. Schweikard and Schmid 2013) up to current research, the idea of a collective subject has been viewed with great suspicion. It has often been portrayed as a spooky entity, and this suspicion has considerable initial plausibility. After all, it seems that if you and I intend to go for a walk together, or if you and I are walking together intentionally, there are just you and I, there is no such additional thing as “the we” having the intention or intending the walking. There are just two intentional subjects intending and intentionally doing the walking together, not three. There is no third subject involved in the planning and carrying out of our joint intentional action, coming along with the two of us somehow hovering above our heads.

Based on the view that content-accounts are insufficient and that subject-accounts are metaphysically spooky, many authors have endorsed a mode-account of collective intentionality. The view is that the correct analysis of “we intend to φ “is really as follows: “we we-intend to φ “(where the first “we”, the subject, is distributive, and “we-intend” is the mode). According to the mode-account, each of us has to we-intend to φ. Of intentions of this kind it is assumed that their mode includes the collective subject concerning which intention is action self-referential, so that there is, in this view, no problem in assuming that collective actions can be individually intended.

In another paper, I argued that mode-accounts are really subject-accounts – after all, what the mode really modifies is the subject (Schmid forthcoming). The following aims at developing a positive view of the subject of “we intend”, and to show how that subject-account survives the allegation of spookiness. This requires us, first, to consider what exactly a subject of intention is. As attitudes of the form “I intend” seem to be less controversial, especially concerning the intentional subject, I will fist focus on attitudes of the singular form and examine the ways in which they involve a subject. The way I suggest to put this question is: how is “I intend” my intention, and how do I figure in attitudes of this kind? (1.) Moving from the singular to the plural, the question is: are attitudes of the form “we intend” ours in the same way as attitudes of the form “I intend” are mine (according to the understanding developed in the previous section)? (2.) The concluding section finally addresses the allegation of spookiness. It is argued that this rests on a fundamentally mistaken concept of the subject, or “the self”. In the same way that it is often claimed that there is no such spooky thing as “the we” in shared intention and joint action, there is no such thing as “the I” in individual intention and action (3.).

1 The singular subject

Since the question of the subject of intention is at issue, and since this question seems to be a bit less controversial in the singular case than in the plural, let us first examine the subject of “I intend”. How is it that “I intend” is my intention? The aim in this section is not to say anything original, but to develop an ecumenical view on which many philosophers could agree. One way to approach this which figures prominently enough in the relevant literature to be promising for this venture is from the way in which I, under some circumstances at least, know “I intend” it is my intention. For this approach to be helpful for our understanding of the subject of “I intend”, we need not discard the possibility of ignorance (e.g., in the case in which I find myself opening the fridge without knowing what it is I wanted there) or misinterpretation of intention (e.g., in the case of implicit bias), just as long as a point can be made that such cases are deficient rather than paradigmatic with regard to the features in virtue of which “I intend” is, or can be, my intention. The view that I submit is acceptable to many philosophers of intentionality is that it is crucially in virtue of my capacity to know that I intend to φ that ny such attitude is really mine.

The interesting point in this way of analyzing an attitude’s being mine is that the knowledge at stake here is of a special form. This special “form of knowledge” has been analyzed under a wide variety of labels, including “self-knowledge”, “self-consciousness”, “self-awareness”, or “de se-knowledge”. Though few authors use those terms interchangeably, and though many draw clear distinctions, there is a wide overlap between the general views advocated under these labels. Therefore, I use “self-knowledge” as an umbrella term in the following. The widely shared consensus is that self-knowledge (in the umbrella term sense) differs profoundly from knowledge of anything else. It is often from this difference (and thus negatively) that the topic is approached in the literature. I shall chose a positive approach instead: let us see what the special knowledge that is “I intend” does. Following are four functions (I do not wish to argue for a particular view of how they relate to each other, though I sense that the list starts with the most basic and primordial function, and proceeds to more complex functions that somehow build on the simpler functions). The four functions of self-knowledge of “I intend” are (a.) self-identification, (b.) self-validation, (c.) self-commitment, and (d.) self-authorization.

  1. a.

    The first function is that qua self-knowledge, knowledge of “I intend” self-identifies. Its function is that it establishes identity in an incontestable or perhaps infallible way. In virtue of its being self-knowledge, “I intend to φ” secure its own standpoint, as it were. If there is an attitude of the form “I intend to φ”, there is simply no question of whether or not I exist. Attitudes of the form “I intend to φ” are, in other words, immune against challenges concerning the identity of the intender. It is easy to overlook how essential self-identification is for mind. It is certainly important for our understanding of the way in which mind is subjective, that is, the way in which mind involves being somebody’s mind. The way in which mind is subjective is in virtue of its being self-identifying, and this self-identifying function may indeed be basic for subjectivity. It is, of course, tempting to read too much into this, e.g., along Cartesian, Kantian, or Hegelian lines. Self-identification, by itself, does not secure self-predication, and may thus not be the base of epistemology, ethics, or ontology some philosophers have taken it to be. So far, I may still be mistaken in what exactly it is I intend, or perhaps even in assuming that my attitude is really one of intending (rather than, say, one of mere wishing). There is no way to go from self-identification straight to substantive insights. But in virtue of self-identification, mind of the form “I intend” constitutes its own standpoint, which is remarkable enough in its own right, independently of further philosophical ambitions.

  2. b.

    The second function is that qua self-knowledge, knowledge of “I intend” self-validates. The issue at stake here is not the identity of the knower itself, but the how of the knowing in question. Knowing what’s on my mind may well be a cognitive achievement, but compared to the tough business of knowing anything else it is, under normal circumstances, quite easy. Where mind is self-known, it does not usually involve registering evidence and inferring. It is true that looking at the mirror or reading my diary sometimes gives substantive knowledge of what it is I truly intend. But this is not the usual way of knowing “I intend”. The usual way is self-knowledge, and this is non-observational and non-inferential. Other knowledge usually needs some external validation. Such knowledge comes with knowledge of how you know, and validation is always an issue. Knowledge of the form “I intend to φ” is not exposed in this way. It is valid not because of observation or inference, but just in virtue of the peculiar kind of knowledge it is. Knowledge of the form “I intend to φ” self-validates. It is immune against challenges concerning its source. There is no question concerning how you know. You know “just like that”. It is simply in virtue of the form of knowledge (or whatever we may wish to call it) that you know.

  3. c.

    The third function is that qua self-knowledge, knowledge of “I intend” self-commits. In virtue of this knowledge’s being self-knowledge, no motivational gap between the known attitude and motivation is left open. I may not care about what I know it is you intend, but I simply cannot not care about what I know it is I intend. The way “I intend” is known to me explains why. “I intend” commits me. “I don’t care” denies commitment, and is thus incompatible with “I intend”. The way to know “I intend” is thus self-committing. It leaves no room for questions of the sort of: “O.k., this seems to be what I intend, but why should I bother?” “I intend to φ, but I don’t care at all” is not a viable position, but a contradiction of commitments. In virtue of its being self-knowledge, the known attitudes are the subject’s commitments, and no external motivational factor or commissive element of sorts are needed. This may not give us the foundations of a substantive ethical theory, but at least it explains what is wrong about the exalted view of evil attitudes sometimes found in the literature, that is, attitudes of the form “φ is bad under any description, but I intend it”. Again, this is a contradiction of commitments (the practical equivalent of Moore’s paradox). Focusing on our attitudes’ being self-known thus gives a perspective on a fundamental way in which the “guise of the good” operates in intention (similar to the way the “guise of truth” operates in belief; cf. Green and Williams 2007).

  4. d.

    The fourth function is that qua self-knowledge, knowledge of “I intend” self-authorizes. This is the feature in virtue of which such knowledge comes with the maker’s privilege. “I intend to φ” constitutes the knower’s standpoint as the position of the one who usually knows best. It is certainly true that I do not always know what I’m really up to, and as an outside observer who knows me well, you may know more than I myself do. But this cannot be the normal case, and it’s usually the self-knower who knows best. This privileged position – or first-personal authority – is not due to the fact that you’re usually closest to yourself, and thus in the best position to observe. It is not that you know yourself better than others do. Rather, first-personal authority comes from the fact that the knowledge at stake, qua self-knowledge, is of an altogether different kind. From this perspective, “self-knowledge” is deeply misleading, as a label, because it comes with a cognitivist view of how mind is subjective. The way the attitude in question self-authorizes makes clear that the feature in question is as practical as it is cognitive – or rather, that it is beneath the distinction between cognitive or theoretical and volitive or practical attitudes. It is self-determination, and the privilege that it entails is the authority of the one who has made up his or her own mind.

It is this structure of self-identification, self-validation, self-commitment, and self-authorization that I take to be at the heart of the idea of an intentional subject. Not all features may be fully realized in any case of “I intend to φ”, but any instance of mind of the form “I intend to φ”, I take it, realizes some of these features, and it is in virtue of this that mind is subjective.

If this – or some not too remote version thereof – seems plausible as an account of how mind of the form “I intend to φ” involves an intentional subject, it is now time to turn to mind of the form “we intend to φ”, and examine how it involves an intentional subject. How are the four functions that we have identified realized in this case? Is there a sense in which mind of the form “we intend to φ” is knowledge that self-identifies our (collective) identity, self-validates our attitude, self-commits us to act, and self-authorizes us (together) as the makers of our mind? To put it in a word, is mind of the form “we intend to φ” plural self-knowledge, or is it just a combination of individual self-knowledge, together with regular knowledge of other people?

2 The plural subject

Let’s play the role of the devil’s advocate and make the case against the plural subject as strong as possible. After all, there seems to be ample room to make the differences between the kind of knowledge which mind of the form “I intend to φ” and “we intend to φ” is, respectively, so strong as to make it appear that the only self-knowledge (or self-determination, in the above sense) there is in the latter case is singular self-knowledge. Put bluntly, the devil’s advocate’s point is this: whatever self-knowledge there is in, or of, “we intend”, it cannot differ from the way “I intend” is self-known, because whatever else “we intend” involves other than you yourself obviously concerns other people, and you can’t self-know other people: they’re not you, and you is all you can self-know. Cast in terms of determination rather than knowledge, the contradiction becomes even more blatant: about self-determination, you have to keep it to yourself – you can’t self-determine other people. Thus it seems obvious in this line of reasoning that “we intend” does not involves a plural subject in the way “I intend” involves a singular subject.

Let’s look closer and see how this spells out in each of the four functions found above:

  1. a.

    In “I intend”, the identity of the agent comes with the form of the thought, as it were; there is no room for the question “who am I thinking about?”. This is obviously different in the plural case. “We intend” is very much in need of an answer to the question “who am I thinking about?”. The identity of one’s partners does not seem to come with the thought itself. It has to be established independently. Another (and more Wittgensteinian) way of putting this is that “I intend” does not “pick out” or “represent” a person, and thus cannot misidentify. Yet “we intend” obviously seems to pick out or represent agents – one’s prospective partners in joint action. And it can misidentify rather easily indeed – I might assume that you’re in, when you’re really not. As I represent you in the attitude “we intend to φ”, I may misrepresent you. I might even mistakenly assume that there are cooperators which don’t even exist. Whatever there is to the identity of “the we” does obviously not come with the form of the thought “we intend”; “we-intend”, it seems, does not self-identify.

  2. b.

    Similarly with regards to the question of validation: knowledge of what it is I intend is not in need of observational evidence, and it cannot be derived from knowledge of the form “Hans Bernhard Schmid intends”. Conversely, knowledge of what it is we intend obviously seems to need a great deal of observational support. I better recall correctly what we’ve agreed to do together, and I better watch what you’re doing, as knowledge of the form “we intend” involves – and is partially derived from – knowledge of the form “XY intends”, where XY is one’s prospective partner. Attitudes of the form “we intend” do not self-validate. Participants in joint intentionality need to listen to and observe each other to know what it is they’re up to together; they do not know “just like that”.

  3. c.

    “I intend to φ, but why should I bother” is a contradiction in commitments that undermines the status of the intention. “We intend to φ, but why should I bother” is nothing of the sort, and it does not undermine the status of the intention, if there is a sufficient number of participants which are suitably motivated so that my potential non-participation does not undermine our activity. Indeed, asking ourselves whether or not we, individually, should do our part in what we, jointly, are up to is often a very reasonable position instead of a contradiction of commitments. This seems to make it likely to assume that wherever there is mind of the sort “We intend to φ”, this does not self-commit to anybody’s action, but is in need of external considerations – such as my thought that what we’re up to is good, and that I should promote it. No such additional consideration is needed in the singular case, as individual intention is per se under the guise of the good in the same way as belief is under the guise of truth. “Φ-ing is utterly bad, but I intend it” is paradoxical in the same way as Moore’s “p is false, but I believe it”; “φ-ing is utterly bad, but we intend it”, however, is often a reasonable position rather than a paradox, as any participant in democratic procedures knows well.

  4. d.

    “I intend to φ” is by no means infallible predication-wise, but it privileges over observers; I might be mistaken, but I usually know best, as mind of that form is my authority of the maker of the attitude. Mind of the form “we intend to φ”, by contrast, does not seem to self-authorize. Thinking of what we, the Philosophy Department of the University of Vienna, intend, I might suspect that you, an outside observer who happens to be on a visit, might know better what has recently decided behind my back. Who’s in a better position to know – me, the member, or you, the non-member – seems to be entirely a circumstantial matter, and not a matter of the kind of knowledge, as in the individual case. It is true that some members may be authorized by the group to make “our” decision, and if I’m that member, it may seem that my thinking “we intend to φ” comes with the authority of the maker of the attitude. But this is conditional on my being authorized, and on the other member’s continued willingness to go along with that arrangement. Again, I might think that based on my authority within my group, my thought of the form “we intend to φ” settles the matter with first-personal authority, while you, as a non-member, might have better information, and know that the way in which I make up my mind does not carry over to the collective. Whatever authority there might be in thought of the form “we intend to φ” about what it is we intend is conditional, and does not come with the form of thought itself, as it does in the individual case.

Knowledge of intentional joint action, it thus seems, is each participant’s self-knowledge of his or her own doing plus mutual ordinary knowledge (e.g., the common knowledge appealed to in received accounts of collective intentionality) of what the respective partners are doing. If this line of argument is sound, it leaves us with what we might call the singularist view – the view that the only subjects that exist are singular subjects, and that whatever plural attitudes there are have singular subjects. As knowledge of the form “we intend” does not self-identify, self-validate, self-commit, and self-authorize, and as these four features of self-knowledge (self-consciousness, self-determination) constitute subjectivity, it seems that there simply is is no plural subject in attitudes of the form “we intend”.

We might be happy with this result, and leave matters at that. But we will also have to consider the costs. Recall the previous steps we have taken. We are interested in attitudes of the form “we intend to φ” where the φ-ing in question is collective rather than distributive, that is, one token action with many agents. Assume that the arguments presented so far are by and large correct, that is, that for there to be one token plural action it has to be collectively intended, and that the distinction between content-, mode-, and subject-accounts of collective intention is exhaustive in terms of possible analyses of collective intention (after all, what else could it be in virtue of which intention is collective, if not some feature of intention?). Assume further that our above argument against content accounts of collective intentionality is right, and that the claim that mode-accounts are really subject accounts is true. Accepting all of that (as I submit we should) makes it a really costly move to deny the plural subject; for if if we do so, we’ll have to accept that there simply is no such thing as a token joint intentional action. Let us spell the consequences out in practice: Whenever you think it is actually one token tango dance which you and your partner intentionally perform together, you’re under an illusion – all there really is your intentional part and your partner’s, perhaps with some structure of mutual knowledge so that it (hopefully) adds up to something that looks like that one token action. But there is never one collective dance with many participants, but just several suitably combined individual dances. All that you’re doing with your partner is to “emulate” a plural subject of a joint performance (to use Margaret Gilbert’s [2013] term). If you happen to be more into disco dance anyway, this may not be particularly hard for you to accept, but the same is true for such activities as, e.g., conversations: there is never one conversation we’re having together, but just your saying what you have to say and my saying what I have to say, interconnected by some structure of belief and expectation. There is never anything we’re really doing together – all there is are combinations of individual activities.

This, I take it, amounts to a reductio of singularism, which calls for a pluralist alternative. Following is a moderate defense of pluralism. In this view, intentional joint action is action which is collectively intended in terms of an intention of which there is plural self-knowledge. Plural self-knowledge is knowledge that shares the features of singular self-knowledge; it comes with self-identification, self-validation, self-commitment, and self-authorization, though since the identity, validation, commitment and authority involved is plural rather than singular, it should not come as a surprise that these features differ from the singular case. In other words, the claim is that while there is some good sense in what has been said in points a.-d. in this section, each conclusion was mistaken: there is plural self-identification, plural self-validation, plural self-commitment, and plural self-authorization after all, though it is, as plural, not of the singular kind – the plural is no collective singular (for a more extended version of some parts of the argument cf. Schmid 2016).

  1. a.

    Concerning the question of identification, it is a mistake to assume that “we intend” presupposes prior representational “picking out” of the other members of “the we”. We don’t have to look around and establish who’s there before thought of the form “let’s φ” can occur, as cases of spontaneous joint action show. Rather, there is often a pre-intentional “sense of ‘us’” in a given situation (Searle 1990), and this “sense of ‘us’” operates in situations of joint action in the same general way as the individual “sense of ‘self’” in the case of individual action (Schmid 2014). Insofar as this is true, there is a sort of self-identification in thought of the form “we intend” after all. An important way in which self-identification in the plural differs from the singular, however, seems to be that it is not infallible. It is true that the identification in thought of the form “we intend” can – and often does – go wrong and thus does not seem to secure a plural standpoint in the way “I intend” secures a singular standpoint. E.g., thought of the form “we intend to φ”, as occurring in my had, may dispose me to cooperate with what turns out to be a display dummy; I had a “sense of ‘us’” that went wrong in a way that seems impossible in the singular case. In this case, thought of the form “we intend” does not establish a plural identity, as there is no “we”.

There are at least two ways to react to this challenge. The first is to say that the possibility of mistake does not rule out the view that “we intend” is self-identifying where there is no mistake. As self-identification is in virtue of the mere form of thought, however, this requires to assume that mind of the sort of “we intend” has a different form if it occurs in my head only, and that “we intends” are plurally self-identifying in virtue of a form that is relational – thought of that form is not realized in individual heads independently of each other (Schmid 2012).

Another strategy to counter the challenge involves a tu quoque-argument against the assumption of singular infallibility, i.e. the view that self-identification is in virtue of the mere form of thought in the singular case. Imagine the following scenario. In an intense discussion between you and me on a conference, an exciting idea comes up, and we form the long-term plan of writing a paper together. Soon after my return home, and as the egoist and insensitive person who I happen to be, I somehow forget that the plan is ours, and I come to see our idea as my idea, and our project as my project. There is now thought of the form “I intend to write this paper”. Imagine you call me on the phone after a couple of weeks, tell me about the work on the joint project you have been doing, and ask about my part. I now realize that something has gone wrong at my end, and I’ll have to apologize. One way of putting my apology would be for intending to write the whole paper rather than just the chapter assigned to me, and thus overextending the content of my intention. But you might feel that any such apology does not really touch at the root of the problem. The problem is not just that I started to write your parts. The problem is that the writing of the paper is ours rather than mine. How am I not, in my egoist frame of mind, simply mistaken about the subject of the intention to write that paper in the same way thought of the form “we intend to φ” misidentifies the subject of intention if no-one else is in on the φ-ing?

  1. b.

    Concerning self-validation, the briefest way to state the pluralist view is that it is not true that I need to guess or watch what you’re up to in order to know what it is we intend to do together anymore than I need to speculate about my own mind in order to know what I’m doing. In this regard, I find the analogy between one’s relation to one’s own future intention in the case of individual intention, on the one hand, and one’s relation to one’s partner’s intentions, on the other hand, instructive. The way in which “we intend” involves a sense of what you’re up to is the same as the way in which “I intend” involves a sense of what I’m going to be up to in a moment. As it is in virtue of that sense that “I intend” self-validates, there is self-validation involved in “we intend”, too. Our access to the content of our shared intentions is not observational; once we intend to φ together, we “just know”; we don’t infer from what we observe about our partners. Rather, we correct them in the light of what it is we intend if what we observe gives reason to assume that they may be up to something else (cf., e.g., Gilbert 1989). The “direction of fit” is world-to-mind in the case of mind of the form “we intend” just as it is in the case of mind of the form “I intend”.

  2. c.

    About self-commitment, it is certainly true that attitudes of the form “φ is utterly bad, but we intend it” are no paradox of commitments, but often a rather reasonable position to take. Yet it would be wrong to think that there is no conflict of commitments involved here. Such attitudes involve a tension of commitments that calls for further joint deliberation. If there is a disagreement between what you think, and what your group thinks, there is something you and your partners should do: talk it over, and though this often can’t be done under real life conditions – democratic groups have to live with dissenting members – it would be wrong to ignore the normative pressure towards consensus. “We intend” is not constitutively sub specie boni in the way “I intend” is, but it is normatively sub specie boni: “We intend” commits to joint deliberation that makes it the case that “we intend” is sub specie boni. Thus even though thought of the form “we intend” does not commit in the exact same way as thought of the form “I intend” does, there is a sense in which “we intend” does self-commit after all. There is some commitment in the form of the thought itself.

  3. d.

    Even the briefest way to counter the singularist challenge concerning self-authorization has to start with an admission. There is, it seems, no integrated collective standpoint from the perspective of which the common mind is made up and that has thus first person singular authority. This should not come as a surprise, as “the we” is no integrated Rousseauvian “collective I” – the “nous” is no “moi commun”, and it is a rather fatal misunderstanding that singular self-authorization should translate to the plural in the way Rousseau, at some places of his work, imagined the state to work. First-personal authority works differently in the singular and the plural, yet there certainly is a plural version. “We intend” does not fail to come with some form of first-person authority: the authority is not the authority of the autocratic “maker of the mind”, but that of the democratic participants in the making up our mind about what we intend. This is first-person plural authority: it is democratic rather than autocratic in nature (“democratic” not in the sense of votes, elections etc. under some majority rule, of course, but in the much more basic sense of the members’ participation in the joint acceptance in virtue of which whatever democratic or un-democratic institutional authority structures exist, an attitude of which the members are the co-authors). That such authority exists, and that it does come with some privilege (though not with the privilege of the singular), is supported by the following consideration. If we want to know what a group’s attitude is, our first instinct would usually be simply to ask a member. The reason why we privilege group members over outside observers in this matter is not only that we assume them to be in a better position to observe what’s going on within their group; rather we privilege the participant perspective simply because it is a participant’s in the making up of the group’s attitude. Thus it seems that “we intend” does self-authorize after all, even if it does so not in the autocratic sense in which “I intend” self-authorizes, but rather in a participatory – or, if you wish, democratic – way.

Thus the emerging view is that mind of the form “we intend” does have a plural subject after all. Joint activity is intentional in virtue of our plural self-knowledge (self-awareness, self-consciousness, apperceptive co-awareness, self-determination, etc.) in the same way individual activity is intentional in virtue of singular self-knowledge (etc.). What’s collective about collective intentionality is that it is plurally self-known. This form of the intentionality in question is the subject, and it is plural. Subjectivity does not only come in the singular, but in the plural, too. The differences between the ways in which subjectivity is realized in the singular and the plural and that are obvious in the different ways in which the functions of self-identification, self-validation, self-commitment and self-authorization are realized can only surprise on the base of the mistaken assumption that subjectivity is always singular.

3 Metaphysical qualms

A reason why many authors resist a plural subject account of collective intentionality is that it seems metaphysically suspicious, and perhaps ideologically charged with the collectivist ballast of potentially totalitarian politics. This seems to be something like the sense in which John Searle has claimed that any conception of mind that is a group’s rather than single individuals’ is “abominable” and a “perfectly dreadful metaphysical excrescence” (Searle 1998, 150). Other authors have made similar claims, calling the idea a “spooky” notion. This seems to be the strong intuition that supports the view that as far as subjectivity is concerned in joint intention, there is no such thing as “the we” – it’s just you and I, in the dual case. In other words, singularism seems to be driven by metaphysical qualms. Here are some ways to put the metaphysical side of the worry:

If you and I intend to φ together, it is not the case that “the we” (rather than you and I) intends to φ. There is no additional subject; if you and I go for a walk together, it’s just the two of us out there, not three: it’s not that there’s a third subject (“the we”) somehow coming along with us. There is nothing “over and above” our heads that keeps us together and somehow pulls off our φ-ing.

If these are indeed the worries that drive singularism, it seems easy enough to dispel them. These worries arise from a mistaken view of the way in which mind is subjective that is just as wrong in the singular case as in the plural. Our defense of the plural subject view can therefore be presented as a squeeze, as it were. If you’re worried about the plural subject and find it spooky, you should be worried about the individual subject and find it spooky, too. The way in which the plural subject is metaphysically suspicious in this view is the exact same way in which the singular subject is metaphysically suspicious. If there is no “we” – no plural subject, no “collective self” – in shared intention, there is no “I” – no individual subject, no individual self – in individual intention either:

If you and I intend to φ together, it is not the case that “the we” (rather than you and I) intends to φ – in the exact same way, if I intend to φ alone, it is not the case that there is something of the sort of “the I” that has my intention.

If you and I go for a walk together, it’s just the two of us out there, not three: it’s not that there’s a third subject (“the we”) somehow coming along with us – in the exact same way, if I go for a walk alone, it’s just me out there, all by myself, there’s no such thing as “the I” coming along with me.

In collective intention and joint activity, there is nothing “over and above” our heads that keeps us together and somehow pulls off our φ-ing – in the exact same way, it is not the case in individual intention and individual activity, there’s no homunculus with or within me that keeps me on course and pulls off my φ-ing.

The spookiness comes from a misconception of what it is to be a subject, in general, not from its particular plural form. That misconception consists of a “reification” or “substantivization” that is perhaps suggested by the very talk of “the subject” or “the self”. We should always keep in mind that whatever there is to “the subject” is not a substance that has intentions as properties. “Subject”, “self” are adverbial in nature. The subject is the way in which φ-ing is intended subjectively, it is a feature of intention, and not an object, or a thing of any sort. For the singular case, the category mistake involved in “reifications” of the subject has often been identified and discussed in a wide variety of ways and in several philosophical traditions in the history of philosophy. Given the fact that talk of “the I”, “the self” sometimes seem to a bit loose in recent philosophical research, it might be good to recall some of those insights. David Hume has pointed out that he can’t find anything of the sort of “a self” in his experience (cf., e.g., Pitson 2002); Johann Gottlieb Fichte has made an elaborate point about how to understand “the I” as a feature of a relation (cf., e.g., Henrich 1966); after resisting the idea of “the ego” for a long time, Edmund Husserl came to adopt a view according to which “the ego” is the form of unity of the stream of consciousness, not an object of experience (cf., e.g., Marbach 1974); Jean-Paul Sartre has argued for the transcendence of “the ego” and argued that whatever there is to “the self” is a matter of pre-reflexive self-awareness (cf., e.g., Priest 2000); Ludwig Wittgenstein (and countless followers) have argued that the subject use of “I” does not refer or “pick out” some entity (cf., e.g., Chauviré 2009).

Applying this to the line taken in this paper, the litmus test for seeing whether or not one agrees with the view that I take Hume, Fichte, Husserl, Sartre, and Wittgenstein to support is this. It was claimed above that “I intend” as well as “we intend” are subjective in virtue of their being self-known (in terms of self-conscious, self-awareness etc.). It was claimed that self-knowledge is self-identification, self-validation, self-commitment and self-authorization, and that these features come in the singular as well as in the plural.

As to the difference between a substantive and an adverbial reading of subjectivity, there is an easy litmus test, and it is in whether or not the following objection against the line of argument taken in this paper seems plausible. We have explored several functions of self-knowledge, and compared ways in which it works in the singular and the plural. But the question concerning the subject does not seem to have been addressed at all, and that question is: who is it that has that self-knowledge, singular or plural?

Learning the lesson from the likes of Hume, Fichte, Husserl, Sartre and Wittgenstein is to take this question to rest on a fundamental misunderstanding of subjectivity. The subject is an adverbial feature of mind, not an entity that has the property of being self-conscious. The subject does not have that knowledge – subjectivity is intending, believing, etc. self-knowingly, and this adverbial feature of intentionality that is the subject comes in the plural as well as in the singular.