I. Philosophy and the social sciences

Many psychological questions belong to what I would like to call micro-sociology—with planned action and joint cooperation as topics, including general attitudes to others and oneself. These topics do not belong to the biological sciences of mere behaviour and animal cognition. In fact, much talk of dispositions in the context of human actions is a methodological blunder and belongs not to the sciences, but to scientism, which is a kind of modern superstition. Psychology thus splits up into a hermeneutic micro-sociology and a natural science which is systematically close to physiology and medical psychiatry.

Philosophical anthropology tries to make general structures of individual and joint action explicit. Since this may include the forms of their representation in thought, a virtually indefinite regress of reflective explications can follow. The topic of such a phenomenological hermeneutics is the subjective and personal form of human life—which Martin Heidegger notoriously called Dasein. A phenomenological analysis of Dasein investigates, for example, forms of subjective reasons for individual attitudes and individual and joint actions, with or without speculative narratives about alleged phylo- and onto-genetic causes, most commonly known from their Freudian background. Science-affined psychology investigates, instead, the brain and neuro-physiological dispositions in animal and human sensation and reactive behaviour empirically, chiefly with the help of statistics. Freud, for one, had wanted to merge the two perspectives, or rather, had claimed that his hermeneutic narratives have a causal core.

In the following, I do not go into the general debate about the still underdeveloped scientific self-consciousness of psychology. Rather I want to show how we can, and why we must, make implicit forms and norms of practical actions explicit. Such forms play crucial roles not only on the meta-level of reflection but also already on the object-level of one's own actions. This has important consequences for the methods and tasks of the social sciences altogether, in their relations to society and economy, state and community.

A phenomenological approach of the kind proposed here is not merely descriptive. Descriptions refer to empirical, i.e., historical, facts, depicted from subjective viewpoints. However, any subjective experience already presupposes some competence to act and speak, understand and comprehend, far transcending animal sensation and enactive, i.e., behaviour-oriented perception (see Noë Reference Noë2004). In phenomenological hermeneutics, also, we do not develop theoretical images of ourselves from the outside, in a fictitious alienation from our own practices, as if we could ‘explain’ our own culture from a Martian's-eye or divine point of view. We rather explicate practically known, but frequently only implicit, forms of action by talking about forms as generic objects of reflection.

Since the time of Plato, the difference between forms-in-performance and forms-as-objects of reflection stands in the centre of methodological reflection on knowledge in general and knowledge about ourselves in particular. The difference is difficult to grasp because understanding forms of being and action, and the relation between objects of reflection and their instantiations in performances, is itself not theoretical knowledge. It is practical competence. A paradigm case is rule-following. Wittgenstein, for one, helped us to differentiate between 1. understanding a rule as an instruction represented by words, 2. performances of actions according to a rule, and, finally, 3. assessment of correctness. All three moments of rule-following are governed by an implicit form, the practical norm of how to do it properly. Understanding, for example, can show up in proper commentaries, but also in proper instantiations of following a rule practically. A performance can accord with a rule either by chance or through wilful action. According to the dialectics of assessments of norms, we can, and sometimes must, control assessments as well.

It is practical knowledge that distinguishes reflective knowledge in a phenomenological and hermeneutic setting from theoretically generalized statistical observations. David Hume had already seen that we cannot or should not jump immediately from observations of relative frequencies into general theories, even if these already work with probability values.Footnote ¹ At least if joint actions are involved, we have to account for their presupposed logical structures. A self-conscious social and behavioural science (deserving its name) therefore needs structural maps of human practices and institutions. They are transcendental or a priori forms that orient our actions and guide our individual decisions. Explicating the sociality of general practices and institutions, as they are presupposed in psychological reactions and individual rationalities, is therefore methodologically different from empirical, i.e., statistical, induction. That kind of induction can work well only on the grounds of some generic knowledge about normalities.

An investigation of the practical forms of joint action and the general structure of sociality, which we can or should label ‘personhood’, turns, in the end, into a logic of the words ‘us’ and ‘we’, together with ‘I’ and ‘you’.

II. Individual and plural subjects of actions

The word ‘I’ does not name a body. The word ‘we’ does not name a set of persons. Rather, such words are used to express anaphoric relations to, say, me as a real speaker or us as a we-group of persons engaged in a certain form of cooperation. Such a we-group is not usually defined as a prefixed class of persons. More often than not it depends on a system of we-properties and we-actions that locally define the extension of the ‘we’. If we recite a text as we sing a song in a choir, each member speaks for the group and utters, perhaps, the word ‘we’. In normal cases, we do not say ‘we’, but, rather, a singular speaker speaks for the group, for example I or you.

Some of the relevant properties and actions that determine the group can be made explicit by the corresponding sentences; some are left presupposed as practically known. Let us look at an example. When I say that we understand the concept of moral autonomy, I mean, in a sense, that any person can or could understand it. Even if some individuals do not understand the concept, the sentence is not wrong. The ‘we’ in it expresses a certain generic universality of a possible understanding of this principle; we do not quantify blindly over all people.Footnote ² The ‘we’ refers, nevertheless, somehow to all of us, namely as persons. That is, insofar as an individual is a full person, she understands moral duties—and acts according to them. If she does not, we may judge it as a privation of her personhood.

Even if it were true that only thinkers like Kant ‘really’ understood the principles of morality (although it is not true), our judgment above still did not allow, say, for a Eurocentric reading. There is always the possibility of enlarging the actual set of people taking part in a practice like that of moral autonomy, i.e., being full persons—whatever this means in detail. Virtually, we can enlarge it to all personal subjects, just like literacy or the faculty to count and calculate. Or rather, to say that any Muslim or Chinese person, to name just these examples, could (learn to) take part in the respective practice if she wished. These examples show how the word ‘we’ can be used without presupposing a contrasting differentiation between ‘us’ and ‘them’. Moreover, in canonized reflections, we should differentiate the notion of a person from the notion of a human individual and of a subject, despite the polite address of people in elevators as persons, not as individuals. Humans and animals both are subjects in the sense that they perform certain types of behaviour. However, no animal is a person, either in the sense of having personal competence or that of having personal dignity, whatever some friends of animal rights believe or say. No personal right is a right of the animal merely for itself. It is, at most, a duty for us, imposed on us by ourselves. This should count as a truism; it also shows the importance of Hegel's phrase ‘for itself’ [für sich] in the sense of Latin pro se, expressing a relation only to itself, not to others.

In what I call a Logic of ‘Us’, we deal with the proper understanding of the different usage of the words ‘we’, ‘us’, and ‘our’. The most important feature of this usage is its dependence on predicates. The ‘we’ can refer to an open we-group as in our example about us moral persons; or it can refer to a closed or finite quantification as in a sentence like ‘you and I, we both, went for a walk’ or ‘we all have to die’. The last sentence has the unproblematic universally quantified meaning that we are all mortal. Some of the predicates are distributive, as I would like to call it, and some are not; some have two readings. In the distributive case, sentences of the form ‘we do x’ or ‘we are y’ mean ‘each of us does x’ and ‘each of us has the property y’.

The non-distributive case is logically more complicated than the distributive. In such cases, we usually refer to a joint and shared generic form of action or practice with variable persons playing some roles. Well-known examples are ‘we sing a duet together’, ‘we are a boys’ choir’, ‘we build a house together’. The last sentence clearly does not mean that each of us is building a house for himself. The distinction becomes clear if we look at the inference rule of distribution. In non-distributive cases, its (schematic) use would yield nonsensical conclusions like ‘I sing a duet (together)’ or ‘he is a boys’ choir’.Footnote ³ The differences in these schematic inferences are, of course, only a first sign that there are some interesting features of the logic of ‘us’ in non-distributive cases.

The last three examples of non-distributive we-sentences are closed in the sense that the set of persons addressed is a closed set of all the members of a we-group—if we abstract from the fact that a choir can be enlarged and usually changes its members. The following we-sentence is non-distributive and open: ‘We have a practice of moral judgment’. It does not make sense to say that I, you, he, or we, the Europeans, have a practice of moral judgement. Nevertheless, it might be true that many people actually do not take part in the practice of moral judgement. Perhaps they are too young or sick, mentally disabled or badly educated.

The basic idea of how to read words like ‘we’, ‘us’, and ‘our’ in all four classes of contexts, distributive and non-distributive, open and closed, is making them depend on the predicates used. That is., the corresponding verb-phrases express the relevant joint actions, properties or practices. However, it is also important for determining the we-group, corresponding to a sentence S of the form ‘we do x’ or ‘we are x’, to answer the following questions:

1. 1. Who utters S? The speaker should belong to the we-group, just as I do, whenever I say ‘we’ in an object-level mode.

2. 2. Who would accept what is said as correct and proper, in such a way that he or she might either say ‘I (want to) do x (too)’ or ‘I (also) have the property y’ (in the distributive case) or ‘I take part in x’ in the non-distributive case?

3. 3. Who would count himself as belonging to us?

4. 4. Whom would we count?

The fact that a we-group often is not defined by distributive properties leads us to say, with Margaret Gilbert (Reference Gilbert1989), that a we-group is a plural subject of a shared intention and of a joint action or practice. There is no mystery about such plural subjects. We should get rid of the myth that we have to avoid such plural subjects defined by the corresponding expressions. We simply have to comprehend the logico-linguistic technique of talking about such a We as a plural subject.

A pair singing a duet, a choir performing a mass in Church, or an entire nation are thus no mystical super-subjects above, and beyond, the individual subjects that comprise them. There is also no mystical super-intention above or beyond the intentions of the participants of the we-group, even though a systematically coherent will is altogether different from a merely empirical aggregate of individual intentions, actions, or sheer behaviour. Nevertheless, there is nothing wrong when we talk about ‘us Germans’—so long as we are careful enough in dealing with such an expression.

The peculiarity of a joint intention consists in the respective content of we-intentions: Each of us may have the intention to take part in the joint action H. The joint action H is defined by what each of us is expected to do in case we really take part.

Raimo Tuomela (Reference Tuomela1984) speaks of group action and intentional group action. In joint actions, individual actors act in a certain way and fulfil certain conditions or norms defining the joint action. In this sense, we have to reduce collective (or plural) subjects of joint actions to individual actors. This much is true in so-called methodological individualism. The problem is that methodological individualism does not stick to this rather moderate claim. In an exaggerated fear of falling into some kind of Platonism by accepting mystical super-subjects, there is a tendency to reduce co-operative action to collective behaviour. The result is that we overlook the particular ‘moral’ form of free co-operation.

Gilbert doubts that joint intention can be reduced to individual intentions and John Searle speaks about we-intentions in the head of the individual actors (1995: 26). What Searle means seems to be this. Each participant of a we-group has the corresponding we-intention. Unfortunately, Searle does not tell us enough about what is in the head or what is represented in the mind of the individual people involved. He does not even seem to realize that he, like Gilbert, uses mystifying pictures and metaphorical expressions for what I want to make explicit in the following exploration—that is, the difference between collective action and free co-operation.

III. Collective action and mere coordination

A collective action results from what each individual in a set of people does. That can happen by chance. But it might also follow from conventions that coordinate individual behaviour and actions, or from forms of free cooperation. Let us start with the case of conventions. David Lewis presented an analysis of this type of joint action in his important book Convention (Reference Lewis1969). My main concern in the following is to show the limits of this promising approach.

Conventions in this particular sense of the word solve problems of coordinating individual actions. We want, for example, to avoid running into one another, or just do not want to disturb the other's actions, at least if this is possible without extra costs. A basic example is the convention of driving on the left in England and on the right in America. After a convention like this has been established by chance, by some implicit practice or by some joint decisions, there is no high risk that follows. The reason is that it is in the interest of the individual actors to accept and stick to the convention. This is a defining feature of mere problems of coordination. We thus define problems of mere coordination by the fact that they have a conventional solution. Such solutions, if established, normally do not run the risk that anybody defects intentionally from the norm or rule: virtually each of us is better off when he or she sticks to the given convention after it is given, at least in cases where there are no extra gains that result from breaking the rule or trespassing the norm.

In an evaluation of the coordinating scheme, each of us gets maximum gains. Therefore, there are no free riders in such a case, or almost none. Of course, we can make coordinating schemes even safer by some threat of sanctions against possible non-compliance. Sometimes, we also systematically control compliance, as in the case of the police controlling the traffic. If we include the threat of sanctions in our calculations, this can change some of our prior evaluation of benefits. Sanctioning state laws can thus produce what Adam Smith, following an idea of Bernard Mandeville, called the invisible hand. My own interest can point, then, in roughly the same direction as what is most fitting to common interests. Threats of sanctions can thus diminish the danger that some expected individual advantage may provide motives not to follow a common scheme of coordinating our actions.

Conventional solutions of coordination problems of the Lewis-Type, as I would like to call them, cohere with our individual advantage. This is so, as I have said, by definition: A mere coordination problem is an ‘easy’ problem because it has a ‘convention’ as a possible solution that allows each individual to maximize her gains and minimize her losses. In doing so, she considers what other persons will most surely do and think and what they probably think about her probable judgments and actions.Footnote ⁴ Certain knowledge about collectively shared convictions can help too. By considerations of this sort, I might expect the others to use well-known conventions that coordinate our behaviour and actions in such a way that we all achieve most of our goals.

However, a reduction of coordinating action to individual intentions and beliefs is possible only when the persons are already able to have well-defined intentions and contentions. This is no harmless precondition. It is not harmless because animals, for example, do not have the means to reflect in a Lewis-type way. They cannot represent the content of what other creatures might think. In fact, they cannot think in the relevant way at all, because thinking presupposes a practice of speaking a full language. In such a language, we can represent non-present possibilities, whereas the signals of animals only coordinate their present behaviour. Unfortunately, Lewis wants to explain the emergence of linguistic meanings on the ground of conventions. Such an approach is circular. It already presupposes complex linguistic representations of intentions and contentions. Grice's intentionalism (Reference Grice1957) is circular in the same way. We can have a determinate contention or intention only if we already know the corresponding fulfilment- or truth-conditions of the utterances as types that express the contention or intention. Intentional semantics and ‘utterer's meanings’ do not help us to get off the ground from animal behaviour at all—and they spread the myth that animals have determinate intentions instead of merely present desires.

Having well-determined intentions and contentions presupposes cultural competence in a practice of expressing intentions and contentions. This practice already contains a practice of reflection in which we control the correctness or propriety of individual speech acts (and other acts) with respect to situation and context, including what the person later does and says. Hence, intending and believing are already social competences.

We have all arrived at a stage in which we are already able to talk to ourselves, just as people have learned to read books silently. Moreover, each of us can already play different roles in a certain form of intra-mental language game which tradition has called ‘representations’. To be able to think in this way presupposes not only linguistic competence but also competence in representing contents, possible speakers and hearers, actors and persons or things acted upon.Footnote ⁵ This is why there can be a kind of fight between different forces in my soul, for example between my immediate animal appetite or desire on one side and my longer lasting will and duty as a full person on the other. The fight is, metaphorically speaking, sometimes a struggle of life and death, namely if desire or duty has to give in (see also Stekeler Reference Stekeler2014: I, 663ff and II, 86ff).

IV. Cooperative actions and the division of labour

There is a general problem of presuppositional analysis here. It concerns implicit forms of actions in general, and of joint actions in particular. The problem is how to put forms that are implicit in a practice (or in ways of acting cooperatively) into language. If we talk, for example, of a division of labour, we take a special case as a paradigm and already speak in metonymic mode. If we use rules for making implicit forms explicit, we have already changed the implicit form of practical norms and have turned them into explicitly described forms: We already change a norm by naming it explicitly by a title, or sketching its form by a sentence, rule or principle. The peculiarity of rule-following thus consists in the fact that the rule is a verbal form (like, for example, ‘add the numbers n and m’) that still needs an implicit norm that ‘tells us practically’ what to do to follow the rule correctly.

If we want to understand rule-like representations of implicit forms in a proper way, we therefore have to bracket the surplus of the explicitly verbalized form, so to speak. More precisely, we have to read grammatical or logical commentaries on the form of a linguistic practice as hints, as Frege had said about his way of making a semantic structure explicit. The hints refer to what we already know how to do implicitly, as we say.Footnote ⁶ We can explicate implicit norms of good action post hoc in some way or other not only by rules or instructions, but also by title-words and head phrases.Footnote ⁷

In his Republic, Plato had already used an analogy in order to make explicit basic forms of cooperation, joint actions and institutions with their complex structure of division of tasks and merits. At the same time, Plato had shown how this structure relates to personal competence or virtue. In the following, I implicitly refer to this Platonic analysis and introduce some terminological distinctions.

We usually think of a system of division of labour as a paradigm case of a joint practice. This fits well in cases in which we produce some object such as a house together. The structure is, however, much more general; it even applies to the roles we play in giving promises, for example. You cannot promise me anything if I do not agree, at least silently, to the promise. It is, therefore, a mistake to assume that it is just up to the speaker whether he has given a promise by saying ‘I promise to φ’. ‘Saying so makes so’ does not work without additional acceptance on the side of the addressees. The same holds for many other joint actions and practices. In the following, I propose to articulate some of the important moments or partial forms of cooperative actions and practices.

1. Cooperative actions are actualizations (or token-realizations) of cooperative action-types, schemes or, what is the same, of generic forms of cooperative actions H. Such a cooperative action-type is an abstract division of individual generic actions or typical tasks. I tacitly include a division of merits and gains Z, if there is one. We can use a formal expression H=<h_j: j∈J> to articulate the fact that H splits up into these tasks h_j (and merits Z_j). Michael Bratman (1999) has also seen that we have to split up a joint action into what he calls individual procedures. However, we must distinguish between such a system of procedures in the abstract, as a merely possible way of arriving at a cooperative action H, and as a concrete division of labour, in which the persons P_j who do or have to do h_j are already named. We may think of doing some work together or playing games with different players, of singing in choirs or performing symphonies with different or fixed persons.

2. An institution consists of a complex system of roles and partial actions. In the institution of jurisprudence, for example, there are the roles of judges and lawyers, of state attorneys and defendants. For each of them there is a whole system of sequences of possible cooperative actions, more or less pre-fixed. In a mathematical short-hand, we could write P=<P_J|J∈K> and P_J=<H_j: j∈J> in order to express that the practice consists of a family of joint actions. How to read such notation is known from mathematical contexts. We use it here in order to sketch the form of the practice. It is a structured system of typical personal roles or tasks,Footnote ⁸ such that we still have to distinguish between the form of actions and their actualization by this or that group of persons.

A formal institution as well as an informal praxis consists of whole families of cooperative actions: Sometimes we use the word ‘justice’ for the formal institution of jurisprudence, and sometimes the word ‘justice’ is a title for the informal practice of moral judgment, made explicit by a system of moral principles. We appeal to this informal notion of moral justice when we criticize a positive and explicit legal system. Legalism is the claim that there cannot be justice outside a positive system of legal rules. It rests on a deep misunderstanding of the relation between informal practices and formal institutions. When we identify the rationality of understanding [Verstand] with the faculty of following rules correctly, reason [Vernunft] goes far beyond rationality since it includes free reflective evaluations of given norms or sometimes even new constructions of rules for general or particular cases.

We have a similar situation in the sciences as well. There is always some informal general knowledge by which we can criticize formal science. Scientism is the mistaken hope of making any informal knowledge more exact by rational rules. Science is rather a collective work aiming at canons of norms and rules of conditioned conceptual inferences. Mathematics can help, even though mathematically controlled calculations do not and cannot replace good judgement and informal norms of proper thinking altogether. Legalism and scientism are two versions of the ideology of ‘regulism’, i.e., the assumption that there is no reason above, and beyond, the mere rationality of following explicit rules properly (see Brandom Reference Brandom1994).

3. In the descriptions of forms in cooperative actions (or institutions or practices) H, proper names stand for specific actors, variables for possible actors. A role is, in a sense, an abstract person represented by a variable in a practice H. The variable stands for possible actors. Its defining property is given by a description of the specific task H_j or its modification in a particular h_j, which is still an action type and not yet a ‘token’ or empirical performance here and now or then and there. Therefore, it may still also be left open who will have to undertake the role or task of performing h_j.

Abstract descriptions of joint actions, practices or institutions as (possible) divisions of labour (and merits) H make it possible to judge about a form of cooperation without already knowing who amongst us will play which roles in the end. In the following, I use the helpful proposal made by Hegel and call such an abstract form the practice as such or in itself [an sich]. It refers to a type or form of practice. It is, as such, generic and therefore ‘abstract’—in contrast to its concrete empirical performances.

A generic form of action is determined in its very identity by a system of contrasts. As forms-in-performance, we can reproduce them. An action in-and-for-itself is an actual performance of such a type in its empirical identity as a token. Being for itself thus refers to empirical identities, but usually presupposes the relevant types or forms.

In the case of a proposal for a new form of action or practice, we usually at first talk about the possible action or practice as such, as a type, which we frequently just describe by words in typical narrative stories, in which even proper names can stand for abstract persons. In such a case, the seeming proper names just function as variables and do not name ‘real persons’. This holds, for example, for the names in fictional stories or novels. The same holds, interestingly, for our talk about things as such, as Hegel has seen. There we talk about generic objects or types of entities. We do this in reflections on forms and in thinking about whole structures of things. A thing, matter or case in-and-for-itself is, then, an empirical instantiation or indexical manifestation of a form. It is a concrete case here and now or then and there. Any concrete thing in the empirical world is thus of a generic type or eidetic form, which belongs to its being-as-such or, what is the same, its general or particular being-in-itself; and it has an empirical being for itself.

In order to show the importance of the distinction between abstract talk about things and persons as such and concrete talk about actual things and real personal subjects in-and-for-themselves, we can, perhaps, make use of John Rawls's A Theory of Justice. Here we can see at least some significance in talking about a practice as such, i.e., about its generic, hence abstract, form. Rawls's talk about a ‘veil of ignorance’ in effect means that we should make judgments about a practice as such, not just about its actual reality and empirical appearances. When we talk about concrete cooperative actions and practices under the veil of ignorance (or ‘as such’), we talk about it as if there were a possibility of a new social contract. This way of talking is important in order to articulate the difference between acceptance of the scheme in its abstract form and in its concrete realization.

4. It is clear why a fair division of tasks and merits is in general better than an unfair or non-proportional distribution. When we discuss a plan for joint actions before putting it into practice, a fair division of tasks and merits is obviously a precondition for general and unconditioned acceptance of the proposed scheme of joint action. Without it, the cooperation or social contract most probably will never leave the stage of an abstract plan, the level of an action-type merely as such.

Let us assume now that we have agreed, in principle, that a certain joint action as such should be carried out, or that some joint practice or institution should be established. How do we figure out in such cases who has to do what? The question is not trivial at all. There are cases in which we all agree that it would be nice and good if we had an institution of a certain form H—but nobody wants to do the necessary work in practice.

In the abstract world of mere possibilities, a scheme of cooperation may exist, or rather, a corresponding plan H can be described and may even get our verbal approval ‘in principle’ even if we still do not have any idea yet how to realize the plan. We say, then, that we support H ‘as such’. Existence as such is, obviously, no more than existence in principle.

Nevertheless, support for a possible practice as such, like, for example, a practice that would reduce CO₂-emissions, can already create some real commitment. If we recognize that we have to save the planet and do something against global warming, our energy policy cannot stay just the same as before. Nevertheless, we still have to consider who of us will have to fulfil which task, carry which burden, and who will get which benefit or share Z_i of the end Z we wish to attain.Footnote ⁹

The problem is usually that we do not find an appropriate way to divide tasks and merits effectively, as we say. Therefore, to say that we would support H in principle often does not mean yet that we support any of the known versions of a concrete division of labour and merits H* for H. In H*, the tasks (and merits) are already distributed between real persons. When we want to turn an abstract scheme of cooperation H into a concrete plan H*, perhaps nobody wants to undertake the most difficult tasks and carry the heaviest burden. Many of us do not want to take up a task before we feel sure that others will also do their share.

Usually, we are not satisfied with an unfair distribution of merits and gains. Fairness is thus a precondition of acknowledgement. Recognizing a meshed plan of joint action thus comes in two stages: recognizing it as such and recognizing it in-and-for-itself, in its concrete version, in which it is already clear who has to do what, who will carry which costs, and who will gain which benefit.

This acknowledgement or recognition is not a relation between personal subjects as living individuals in an empirical group of people. It is a relation between me as member of a we-group and the whole practice in question. In other words, personal relations in a we-group are mediated by the practice H.

It is now important to see that Hegel's reflections on ‘recognition’ in the Phenomenology of Spirit have put the mediation of personal practices into the centre of the analysis of personhood. We do not just recognize other human beings as individual subjects, as we do when we deal with dogs or cats. Recognizing other human beings as persons and viewing oneself as a person is already recognizing personhood as it goes beyond mere subjectivity. The step from consciousness to self-consciousness is thus at the same time a step from subjectivity, which we share with animals, to personhood. If a person is recognized or if I refer to myself as a person, the recognition is mediated by personal roles. In other words, fighting for recognition is—at least most often—an inner fight between a mere subject with its desires guiding the immediate tendencies of (perhaps my own) animal behaviour and the person (that I am also). Plato's word for the person in this sense had been psyche.

If the mere subject, with its enactive animal dispositions, wins the contest, the person loses; or rather, the person collapses back into a pre-personal subject, in the worst case: without stable intentions and, therefore, partially without free will. It is, therefore, wrong to assume that Hegel talks of fights between different individuals, as if I had to fight for my life against you—who perhaps do not recognize me without such a fight of life and death. The case in which a master esteems a slave that fights against him more than a servant who does not is not a good paradigm for understanding Hegel's simile, even though it lies behind all ‘social’ readings of the struggle for recognition, from Kojève via Habermas to Honneth. Instead, I read the fight for recognition in the context of Hegel's version of the categorical imperative: ‘be a person and respect the others as persons’ (PR: §36).Footnote ¹⁰ It says that we have to develop our personal skills, undertake a whole system of personal roles and recognize that all other human individuals are also persons. They stand, as such, on the same footing that I (and we) do. To recognize oneself or another subject as a person thus means to take oneself, her or him, as a real or possible member in a we-group of joint action and cooperation.

In reality, the situation is more complicated. The reason is that in the real social world no one has an unlimited choice as to which role he wants to play in a pre-given division of tasks and merits. Our choices are limited, namely by a path-dependent development of human practices and institutions in cultural history. The complex scheme of a division of tasks and merits in a society is not at all a result of explicit design and explicit acknowledgement. It is a result of implicit tradition. Actors that are least favoured by the division of tasks and merits usually accept it not because it is the best way to cooperate, but because the costs of defection are (or seem to be) higher than grudging acknowledgment of social reality or social necessity.

5. Only in the happy case in which all participants accept (and not only verbally) a concrete distribution H* of tasks and merits in an actual we-group is there no problem. Here, we all want to achieve the common end Z; and each of us wants to do his share h_i—and hopefully does it. These non-problematic forms of we-actions have the following form:

1. 5i) Each of us is acquainted with the general form of the relevant joint action H as such, accepts it as a good means for a jointly acknowledged end Z, and agrees that H is in principle a good form of cooperation.

2. 5ii) Each of us acknowledges a concrete distribution H* of tasks *h_i and merits *Z_i between us persons P_i in our we-group.

3. 5iii) Each of us does what he or she has to do according to our own acknowledgement and verbal undertaking. That is, each of us fulfils his or her task *h_i. By acting in such a way, we realize the joint generic action or the cooperative scheme H. Only now do we perform or actualize the action or practice as such and for itself. The phrase ‘for itself’ refers again to the empirical instantiation—according to the meaning of the Latin word ‘pro’ that expresses a relation, which includes here the relation between a type and a token. Being for itself means that there is such a relation in the empirical world of appearances defined; it ‘tells’ us which different appearances are instantiations of the same action or thing.

We can now describe the happy situation of a sufficiently well-actualized cooperation or practice thus: We all want to attain a certain (common) end Z by our cooperative action H, if each of us (at least prima facie) wants this. Each of us has to accept the (common) end Z and the cooperative scheme H=<h_i|i∈I> as such. Moreover, we must also recognize the concrete scheme H*. That is, we evaluate not only H but also H* as a means to Z; and we accept the concrete distribution Z* of the benefits Z. In these cases, we can understand the ‘we’ in a distributive sense and do not need a mystical ‘we-intention’.

The question is whether or how we know of each other that we agree in our acknowledgment of H in respect of H*. Sometimes we know it, sometimes not. To know it is not even always important. It may not be important in cases in which the implicit acknowledgment of H and H* just shows itself in our performances. It will be important, however, in cases in which we try to set up a new scheme of cooperation. Then, individual persons will take part only if they expect that the cooperative action will succeed in some sufficient way. For this, they reflect about what the others might think about H, H* and about what they should and what they will do.

V. Commitment in free cooperation

The crucial problem of free cooperation is therefore this. Even if we already have verbally or in principle accepted a concrete division of tasks in some we-action H*, and even if this acknowledgment is subjectively sincere, we still cannot be sure if all of us really will do what we now should do. We may have pledged to do it; but still we do not know if you will do it.Footnote ¹¹ Therefore, we are uncertain whether our free cooperation H or its concrete version H* will bring about the end Z or not; nor is each of us (or even any of us) certain to get the share *Z_i she hopes for.

In such cases, we talk about moral commitment or moral duty. Its autonomy, as Kant was the first to see clearly, consists in the conceptual (‘logical’) fact that each person is bound by his (or her) own (free) acknowledgements in a cooperative practice or institution. However, such acknowledgements are usually implicit. They normally show up by what the persons do, rather than by what they actually say or even would say if asked. Making one's acknowledgements explicit can even mean distancing oneself theoretically or verbally from silently presupposed recognitions of practices and their forms. However, even in the implicit case of practical recognition, the commitment to perform certain forms, hence the recognition of the norms in the practice, always-already rests on the very idea of personal autonomy.

Obviously, we use the word ‘morality’ for a whole family of cases. All of them, though, have something to do with free commitments and, therefore, with our duty as persons in real or possible relations to other persons. This holds even for any possible duty I could have to myself. Any such duty is mediated by personhood; and personhood as being a person in-and-for-itself (not just as such or in principle) is, in a way, the subjective moment of an objective, inter-subjective, practical system of social, interpersonal, relations in an actual practice of cooperation and living together, with traditions in their path-dependent genealogy. This was, in my contemporary language, the deepest insight of Hegel that goes beyond Kant.

The following principle resembles Kant's categorical imperative: If I can will that a possible maxim, i.e., a possible form of joint action H (in itself) be a general law, and if a practice H* (in-and-for-itself) defines a ‘meshed plan’ (Bratman Reference Bratman1999) of an already actually recognized we-intention, then I have to perform my corresponding moral duty h*. Our analysis shows that Kant does not account for the difference between H and H* nor for the crucial difference between my recognition of H, H* and our joint recognition of them. The same holds for Bratman, in the end. My special duty h* is not just a result of my generalization of a possible maxim, which I read as a generic action as a form to act. We can see the problem when looking, for example, at the regime of private property. Kant argues much too rashly when he thinks that his categorical imperative forbids me to ‘steal’, i.e., to appropriate the property of others. As Hegel sees and says against Kant, we can, in principle, will that all property so-called is, as in an ideal communist society, common property, such that the very notion of ‘stealing’ radically changes. It only reduces, then, to a maxim like this: We should not interfere with what others are doing with things at hand. The example also shows why earnestly wanting the corresponding maxim to become a general law is not enough for moral entitlement.

In other words, Kant's talk of ‘general law’ is much too unclear to fit all relevant cases. However, in some interesting types of cases, Kant's principle is a sufficient criterion for distinguishing between morally coherent and incoherent schemes of action. Any action that falls under the label ‘free rider’ is incoherent, hence morally wrong; but moral coherence in one's theoretical will and one's doings is not good enough, as the case of property clearly shows. As in many other cases, Kant appeals silently to merely historically given forms of coordination and cooperation. The grammatical problem of understanding the generic I and We as standing logically for an open ‘we-group’ of persons obviously acquires practical significance here: Kant's transcendental I must be reconstructed as a We, defined by the jointly accepted maxim H in respect of H*, not by my acceptance of the maxim as a general law. Put into a ‘we-form’, a group-related joint action H in respect of H* as such (i.e., in its genericity) must replace my merely subjective maxim to act according to a particular action-type h. The result, then, is this: If we want a certain form of (free) cooperation H (for whatever reasons or goals) as such, and if we (can) accept a concrete division of tasks H* for us, and if my role or task is to perform the individual action *h_i then I have to do it, i.e. I have to perform *h_i.

Of course, any such principle, which obviously lies outside a set of mathematical axioms, is somewhat vague. This is a conceptual statement. Principles are vague because they are short, gnomic, expressions of rather complex generic forms. They must be short because only in their short versions do they fulfil the mnemonic task of sketching relevant features of a complex form. A moral principle is, therefore, no schematic rule or procedure. Rather, it is similar to a title. It usually refers to a whole realm of procedures and rules, forms of judgment and action that are in accordance with the principle, fall under it, do not contradict it. In other words, any proper understanding of such a principle, with or without further explanation, already presupposes quite some experience, competence in understanding high-level reflective language, power of judgement and autonomous thinking. The situation is similar to the case of a proper use of prototypes, paradigms, metonymies, analogies, and other not fully schematized figures of speech or tropes of discourse.

A first principle concerning the question what it means that we want H and (can) accept H* follows almost immediately from our analysis of we-intentions and we-actions. It means that each of us, I, you, and all the others, want to perform the joint action or practice H in itself and (can) accept the corresponding division of task H*, the concept or form of the action for us. A second (much more important and difficult) question is why Kant formulates his moral principle from the perspective of the first (and second) person singular and why I (and hopefully we) do not. As a first answer I would like to say that Kant's formulation is appropriate for all cases when I am in a position to know in a sense immediately that I, as an arbitrary representative of us, would acknowledge the action H and H* as a good form of cooperative practice.

In order to explain what these situations are, the example of the famous prisoner's dilemma is most helpful (if we do not shift too much attention to formal or mathematical technicalities). The paradigmatic situation is this. Two prisoners, both accused of murder, have the option to bear witness against the other or to keep silent. If one ‘confesses’ by accusing the other we speak of defection. If he keeps silent, we say that he cooperates. If both confess, they get, say, a five-year sentence. If they cooperate, both get a one-year sentence for illegal possession of weapons, as the story goes. If I accuse you, and if you keep silent, I am freed as a witness—and you get ten years. The same holds vice versa. Game theory predicts that in such cases both of us will defect on the grounds of the following rational consideration: if you cooperate, it is better for me not to cooperate; if you do not cooperate, it is still better for me to defect. The dilemma is this: our defendants both know—if they think about it—that in principle it would be best to cooperate, i.e., not to defect; but in a ‘rational calculation’ of expected costs and benefits, it seems best to defect. This sounds strange. The word ‘seems’ contradicts, in fact, the definition of rationality in game theory and makes the dilemma manifest. The dilemma disappears if we use Kant's formulation of the moral principle in an appropriate way. The principle says, in short, that if I can accept H (or H*) as a general law or scheme of joint action, then I must act according to it. This ‘must’ is the ‘must’ of morality. As such, it is the duty of being a full moral person. The prisoner's dilemma is, in a sense, a dilemma for criminals especially, because a criminal acts against ‘moral’ principles if he expects higher gains without taking the risk of ‘trust’.

A moral defector is still a human subject. We have to respect him or her as a person in the passive sense of the word, referring to personal dignity, and we may attribute to him or her the active capacity for instrumental reasoning and rational calculation; but, by defecting, the subject loses, at least in principle, the right to be trusted as a free cooperative person. It is this conceptual fact of practical moral attitudes that lies at the ground of a current distrust in so-called economic man, the homo economicus as the character role of how pure capitalism runs its business.

The moral principle also helps us to understand the contrast between moral reason [Vernunft] and mere rationality [Verstand]. We do not, and cannot, jointly agree to any other scheme of joint action than cooperation. We do not intentionally want not to cooperate, because this would mean to want worse results than those we could get by cooperation. Nevertheless, we seem to think cautiously and rationally only when we do not cooperate, when we do not trust in cases in which trust comes with some high risks: In a case of the type of the prisoner's dilemma, I cannot really want that only I cooperate and you do not cooperate. I can wish in such a case that you cooperate—even though I defect. However, this is by no means a general law or rule in the Kantian sense that we should accept together.

Each of us knows this quite well; children learn it already. It is interesting that adults forget it. We frequently do not even have to talk to each other to know what we would have to do to cooperate in the corresponding we-group. This shows why the implicit case without any explicit promise or contract is often no different at all from the case in which we make explicit promises to cooperate. Our duty is grounded in the fact that we both know that we would accept only the cooperative scheme of action H=<h₁, h₂> as a good scheme and not any other. This is the case Kant has in mind. Therefore, we both know that, morally, i.e., in view of our general duty of safeguarding free cooperation, we must do h_i no matter how sure we are about the success of the cooperation.

We do not need a high-school diploma to know this practically. On the contrary, the most down-to-earth people, even people living in slums, know this. Kant's moral philosophy makes this form of morally coherent judgments explicit. However, like the much weaker version in the gospel, Kant can account only for some free cooperative schemes of actions between individual people, and not, as Hegel and John Rawls do, for a complex division of labour and the generic benefits of institutions. Kant neglects, in particular, the non-trivial significance of traditional solutions to coordination problems—and the knowledge about how they already reduce the risks of merely morally controlled free cooperation.

Some will certainly object that morality is not to be confused with group-solidarity; or rather that Kant did not intend his principle of morality as a norm for a we-group of delinquents to escape justice. The point, however, is that we have to take the general form of the principle in view. There is no problem, at any rate, with saying that moral duty in the universal sense is a duty resulting from implicitly accepted cooperative action schemes for all of us in an open and all-inclusive sense of all humankind. Examples of such cooperative action-schemes are not difficult to find. In fact, they are everywhere. Telling the truth and keeping promises are the most prominent examples. Plato was already aware that a person who defects is acting immorally and by this is hurting her own soul or personality. We could add that she does not deserve respect as a person as long as she does not show remorse and ask for forgiveness.

Threats of sanctions by state power are needed because we know that persons are seduced by the immorality of free-riding. Nevertheless, only on the ground of trusting the morality in free cooperation can we be free from fear of getting robbed or killed at almost any minute. In order to make the problem clear, Plato used the story of Gyges, who, by chance, did not have to fear the threat of sanctions. Indeed, the cases in which we must trust that other persons use moral judgement and act according to schemes of free cooperation are much more common than we may think. Plato was well aware of this. As an ardent supporter of rational decision theory and game theoretic economics, Lewis on the other hand is wrong to believe, together with other defendants of the alleged rationality of homo economicus, that virtually any relevant case of cooperative action can be turned by a threat of sanctions into a mere coordination problem.

My point is that we have to replace Kant's ‘I-thou-version’ of the moral principle, which stresses the subjective point of view of judgment too much, with a more complex ‘we-version’. The reason is this. For Kant, the objects of moral judgments are subjective maxims. Such maxims are rules or rather schemes of individual action h. I think, instead, that the object and topic of moral judgments are (at least sometimes if not always) schemes of joint actions H as such, and H* for us, as Hegel proposed.

My claim is now that the general law in Kant's formulation refers in almost any relevant case to we-actions or cooperative schemes H and H*, not directly to individual action schemes h. What I should do, then, is to act according to H or H*. Therefore, it is wrong to assume that according to moral duty all have to do the same. Rather, each person has to do her duty, as Plato had already seen. Of course, my duty and your duty are not given by ‘nature’. They are given by our shared intention or joint will, by an already implicitly or even explicitly accepted form of cooperation H. In other words, I have used the we-version in order to criticize the following surface readings of the categorical imperative: I must perform an action h if I can accept that everybody has to do h; I am morally allowed to do h if I can will that everybody is allowed to do h; I am morally forbidden to do h if I can assent to the moral norm that it is forbidden to do h.

VI. The moral duty of trust and its limits

We have already touched on the crucial problem of the moral law of free cooperation. It is the problem of trust. It is concerned with the hope that the other participants of our we-group will do their duty, too. This trust and hope are necessary since we are actually interested in the success of our free cooperation. In distinction from Kant, Hegel acknowledged this clearly. Trust and hope are constitutive for the success of (free) cooperation. If I do not trust that others will cooperate, I will most probably also not do what I am supposed to do according to the implicitly known ‘best form’ of cooperation.Footnote ¹²

The problem of trust shows that moral duty is not unconditional. But it is not conditional on hoped-for gains either, even though duty still depends on the internal end of the form of cooperation in question. This might be the deeper reason why Kant himself had talked about what we can hope for when we fulfil our duty. In his widely contested philosophy of religion, he accepts that we may have a subjective interest in the overall good consequences of fulfilling our duty. Nevertheless, we should do our duty without thinking about the individual gains we hope for, even though we may think of generic gains Z_i. The reason for this (only seemingly strange) prohibition on thinking rationally about the expected outcomes of fulfilling my duty for my future life by doing h_i is this. If each of us were to maximize the expected gains and minimize the expected risks, free cooperation would collapse; it would not take place at all. We would return to what Hegel ironically calls the rational animal kingdom (as I would propose to translate geistiges Tierreich) of homo economicus. As a result, the only ‘rational’ way out seems, at first sight, to be Lewis-coordination, which can be seen as making the form of thought of Thomas Hobbes explicit: we need the state to coordinate our rational behaviour. Only if sanctions change the matrix of benefits is the rationally best choice to ‘cooperate’, or rather, to coordinate our actions accordingly. Only threats of sanctions on earth (or in heaven, as still in the case of Plato's myth, copied by Christianity, and even Hobbes) reduce the risks of trust in view of free ridership. However, it is precisely coping with this risk even without reliance on such (assumed) sanctions which defines the morality of a really adult and, as such, really morally autonomous person in his or her free participation in free cooperation. This is the very reason for the whole practice of morality as a practice of free evaluation of a person's free cooperativeness.

The most important we-group, however, is indeed a society in a state. The state turns the society into a closed we-group. This has advantages and disadvantages. It is not —at least not easily—possible to leave a state for good without the consent of another state to take me in. The reason is that we have divided the whole earth geographically into states and their distinct societies as we-groups with corresponding institutions, tasks and merits; even though there are some exceptions, namely ‘stateless’ individuals who do not belong to any state or country in ‘our’ political system of states, countries and nations.

It is, therefore, not idiosyncratic love of the state, but opposition to romantic and sentimental theories of freedom and natural law, that explains why Hegel is so concerned about the state. He agrees with Hobbes that there is no legal law outside the state system; and he agrees with Rousseau, against Locke's natural property law, that property-rights presuppose some threats of force against possible trespassers. In contrast to Rousseau, however, Hegel agrees with the tradition of liberalism that property rights are most important for individual freedom and autonomy, knowing, however, that there can be highly problematic consequences of property laws for large parts of the populace. The fate of African Americans is just one example; the fate of the proletarian work force in the nineteenth and early twentieth century another. If we know that many people can by no means agree freely to the imposed form of cooperation in the state, then we must evaluate the specific form of the law as unjust and try to improve it. On the other hand, we have to agree that there are limits to reasonable change to established schemes of social cooperation. There are always dangers in fast and violent change.

That danger is connected in some interesting ways with what I would like to call the dialectics of self-consciousness. Explications of practical forms and norms turn historically pre-given systems of cooperation and coordination for itself by thought into a mere possibility in itself. In fact, Hegel sees that the very method of thinking as making forms and norms explicit consists in turning actualities into possibilities. Such thinking can become a danger for a merely historically caused earthly power (and religious belief) in a society: since there is no immediate justification of the norms as fair (and religion is just a way to articulate the general form of compliance), we always have to expect unwelcome results, such as those that showed up historically as results of so-called enlightenment. We might withdraw the acknowledgment of a given tradition or practice before having a better one established. This can already happen when we look at the structure of a society as such, under the veil of ignorance. We then talk about it as if there were a kind of free choice between different versions of a social contract.

Free speech is the medium in which we make pre-given schemes of social cooperation explicit and think about possible alternatives. We propose such alternatives, evaluate them in comparison to others and to existing forms—and become self-conscious about ourselves. Hegel is therefore right to identify freedom with self-consciousness and to attack deficient forms of merely individualistic, parochial and sentimental self-consciousness. Introspective, psychological self-analysis and empirical investigation of real situations are just not enough.

Hegel's thinking leads to a critique of fundamentalist liberalism.Footnote ¹³ The important difference from Marx and Marxism lies in Hegel's conservative principle of dialectics. The political prototype is the principle of the constructive vote of distrust: only the establishment of a new government (for example by election) should force a government to resign. In general, the principle says that we should not give up or destroy established institutional practices H* solving coordination problems if we do not already agree upon some ‘better’ replacement H**. Of course, the generic openness of the ‘We’ is one problem here; another is that without civil unrest or even war many necessary changes to ‘the better’ seem to be impossible. Hegel was the only thinker who was realistic in his attitude to these problems and who dared to say so against ‘the good people’ of an abstract, romantic or idealistic, common sense.

For Hegel, the liberal state as a safeguard of property rights must be defended. We need the central power of the state, if ‘we’ want to defend personal freedom and autonomy. However, this also entails for Hegel the duty to provide social security for those who benefit least from a ‘liberal’ property regime, enforced by the state. Manchester liberalism is inconsistent insofar as it has no moral argument against revolutionary civil war and just relies on brute force.

In contradistinction to Hegel's realism, which stands in some parts in the tradition of Hobbes and Kant, Marx, like many anarchists of his time, dreams of a society that rests totally on free cooperation, neglecting the organizational problems of large-scale economic production. Even though Marx is right in his analysis of surplus value in capitalism, he thinks that we can, at least in the event of a rapid development of automatization, do without any hierarchy in our design and control of our division of labour and merits by property rights and the state. If we abstract from the problems of its realization, the abstract idea of communism as such, in principle, is certainly appealing. Unfortunately, this is only so in theory, not in practice.

Communism does not only presuppose the autonomous insight and acknowledgment of all participants with respect to their duties in a complex division of labour and merits. It also presupposes the moral power of really fulfilling tasks without any sanctions. A communist society would be a free we-group with free cooperation that rests only on pure morality and absolute freedom. Just as its opposing bourgeois version, namely fascism, Marxism takes the family, as in Ferdinand Tönnies's analysis of ‘community’ [Gemeinschaft] ([1887] Reference Tönnies2010) in contrast to the ‘society’ of property-rights and contracts, as a paradigm of a communitarian state. It is, therefore, no accident that Stalin, Mussolini, Hitler, even Pétain, Franco or Salazar, and all similar populist leaders in later times sound so similar in their family rhetoric. Marx had been an ardent defender of radical personal freedom against the state and against its legal support for private power, mediated by property rights. However, he totally underestimated the role of tradition in a path-dependent normativity of social facts (Habermas, Brandom), the needs of division of labour (Hayek), of security and, hence, of legal justice (Kelsen and many other Neo-Kantian ‘positivists’). In the end, we should not mistake so-called communitarianism for a Hegelian position, which by no means opts for blind faith, either in liberal property rights or in the state.