According to Kant, human rationality has to be understood in terms of self-legislation. Rational animals, i.e., human beings, establish norms for themselves and subordinate their behavior to these norms. Kant thinks that this fundamental structure explains what freedom is. But one might suspect that in doing so, he falls into what has been discussed, in recent years, as the “Kantian paradox”:  reason binds itself through norms that it gives to itself; thus, these norms are founded in a lawless act. Out of an exercise of what Hegel calls “absolute freedom,”  reason submits itself to norms which then are supposed to bind reasonable behavior. But how is it possible for norms to bind rational animals if these norms are founded in an act of unboundedness and arbitrariness?
The Kantian paradox confronts us with the general question: how is it possible to treat the structure of being bound by self-given norms as a conception of freedom? Or, in short: within an explanation of reason, how are boundedness and freedom compatible with one another? It is the aim of this article to show that the compatibility in question can be explained if one does not conceive of reason’s self-legislation in terms of an initial submission to norms, but as a submission to them that is renewed again and again in the course of rational actions. Accordingly, self-legislation has to be understood within the ever-ongoing renovation and change of norms to which one subordinates one’s own behavior. This will allow us to implicitly show that the Kantian paradox results out of an untenable conception of rational norms. According to this untenable conception, rational norms are unchangeably valid once they are established. But, as we will argue, the norms that guide rational practices are subject to constant revision. These norms – and the examples are countless: eating rules, greeting rules, traffic regulations, dress codes, linguistic customs, conversational conventions, artistic styles, etc. – are constantly developing, transforming, and changing.  A famous dictum of Wittgenstein nicely captures what might be called the flexibility of our rational behavior: “we make up the rules as we go along.” 
The theoretical foundation of this account of human reason lies in the concept of improvisation. Making up the rules as one goes along is, by definition, an improvisatory practice. Our main claim is that human beings establish normative practices by improvising within the situations they are living through, a phenomenon well exemplified by artistic improvisation.
However, the aim of the following reflections is not to offer a detailed account of rational practices as improvisational practices. The article’s aim is more modest and preliminary: it consists in providing an investigation into the conceptual conditions necessary for understanding rationality in terms of improvisation. The most serious theoretical problems here are posed by misconceptions that hinder us from seeing how it could be possible to explain human practices as improvisational. In our view, of particular relevance are four misconceptions related to the concepts of (a) improvisation, (b) normativity, (c) habit, and (d) freedom.
(a) According to the misconception of improvisation, improvisations involve events that are not prepared and not defined. Such an understanding relies on a rigid distinction between a plan, such as a musical composition, and improvisation. In contrast to performances of compositions and to actions executing detailed plans, it is assumed that improvised performances are not rule governed, are spontaneous, and come, as it were, out of the blue.
(b) A common, but in our view misguided, perspective characterizes normativity like this: human practices are guided, regulated, and stabilized by norms. Within these practices, improvisation only comes into play if no guiding norms are available. Kant’s distinction between determining and reflecting judgments offers a good philosophical example of this kind of thinking. According to Kant, determining judgments apply preestablished norms, whereas reflecting judgments establish norms in cases where no guiding norms are available. 
(c) The third notion we want to criticize is that habits are automatisms that have the function of easing our everyday life. This view is often meant to avoid an overintellectualized conception of everyday practices.  For its proponents, habits are embodied, unconscious dispositions to repeat the ever-same series of actions, which leads them to the conclusion that habits are free of all reflections. A (for us unwelcome) consequence of this idea is that it sets up a sharp contrast between habits and improvisations: as one might guess, in contrast to habits, this view implies that improvisations do not make our existence easier, that they consist in making something new and in constantly redefining practices, and thus that they involve constant, strenuous reflection on our relation to a changing world.
(d) Finally, according to the misconception of freedom, we act freely if we act in a self-determined manner. In this view, freedom has its source in the acting subject. Acting freely, so the story goes, means spontaneously determining the course of one’s action. As a consequence, this view draws a rigid distinction between acting in line with externally determining forces and acting spontaneously on one’s will. Accordingly, improvisation is seen as an expression of this second kind of action and is thus seen as not being bound up with external constraints (which is evidently false).
In our article, we want to show how these misconceptions can be overcome in order to clear the path toward a plausible view of human practices as rational, yet improvisational practices. First, we will offer a more informed, realistic understanding of improvisation as a highly prepared undertaking. The next three steps will then consist in sketching out conceptions of normativity, habit, and freedom that are able to adequately account for their intrinsically improvisational aspects, properly understood. In turn, we argue that important dimensions of human practices (normativity, habit, and freedom) can only be adequately understood by taking into account their constitutive connection to improvisation. In the end, these corrected conceptions of normativity, habit, and freedom lay the basis for an explanation of rational practices as improvisational practices. As we shall see, this will require making explicit the dialectical structure underpinning both improvisation and the three other dimensions of human practices.
We first need to clear up some misunderstandings about improvisation, since there can be no doubt that widespread assumptions about the nature of improvisation pose the most significant obstacle to understanding the essentially improvisational nature of human practice. Two points are particularly important. First, we should reflect on the ways in which improvisation is intrinsically bound up with unexpectedness, as the dialectics of preparedness and unpreparedness are key to improvisation. Second, we want to illuminate the inherently normative dimension of improvisation. This means properly grasping the evaluative aspect of improvisational interactions.
(1) Improvisation is generally understood as the coincidence of invention and performance. A performance (a musical performance, for instance) is improvised if that which is produced by the performers’ actions (sounds, gestures, movements, etc.) does not result from executing a predetermined plan of action (a musical score, for instance) but is rather invented “on the spot.” However, this aspect of improvisation is misunderstood if one takes it as implying that improvisation is a creatio ex nihilo, an expression of unbounded creativity.  Escaping the hold of this misconception demands that we shed light on the preconditions required for successful improvisation (whatever that might mean). Improvisers must have expertise within cultural referential frames.  In the context of improvised performances, the creativity of invention can only emerge if preparation and a series of cultural, social, technical, and material constraints open a proper space for it. In this sense, “improvisation is not improvised.”  This is what saxophonist Lee Konitz meant when he said that “you have to be prepared to be unprepared.”  For instance, in the world of music, improvising musicians must know how to react to musical events that are, to a certain extent, unpredictable. Hence, preparation – which includes but is not limited to the acquisition of specific skills by training, repetition, and imitation, and familiarity with a practice and its specific cultural context – is (a) at once a precondition and an inherent aspect of improvisation. However, (b) the very skills that are preconditions of improvisation are not external to the improvisational practice as some kind of self-standing foundation. Rather, they are formed and, at the same time, put into question and transformed, within and through the practice of which they are the precondition. As we shall see, this allows us to understand (c) the ways in which improvisation generates unexpectedness (and does not presuppose it).
Preparation is a precondition of improvisation, since without preparation improvisers could not properly cope with the unexpected situations that emerge during a performance.  Thus, improvisers have to be “well-equipped” with skills; this applies to both everyday and artistic improvisation. For example, improvising musicians have to be able to play a musical instrument. Moreover, they should know the styles, conventions and tricks of a particular artistic practice or genre. But this training alone does not help them to cope with the unexpected events that might occur during an improvisation. Thus, skills should be conceived of as consisting not only in technical and stylistic training but also in the ability to adequately react to the unexpected. These abilities are obtained through active and repetitive exposure to a practice, i.e., through practical training, rather than through theoretical learning. They are competencies acquired through performance: “learning through doing.”  Hence, we conceive of a skill as embodied “procedural knowledge” (in the sense of a “knowing how” ). The acquisition of skill takes place within the practice and as practice. 
(a) However, if improvisers only repeat preestablished, stereotypical formulas, we consider them unsuccessful. Improvisers’ skills thus have to be plastic,  precisely because they have to adapt to concrete, unforeseen circumstances. Naturally, a skill is learned by means of performing it. However, this insight makes clear that performing a skill also involves constantly (un)learning that skill because concretely dealing with specific and unforeseen performing situations has a recursive effect on the development of the “know-how” of improvisation. In short, it is part of its development as artistic practice. Having a specific know-how is not only a precondition of the practice but is (re-)shaped by improvisational practices that, in this way, put into play their own preconditions. The way performers cope (more or less successfully) with the (more or less) unexpected situations of each performance retroactively contributes to the (trans)formation of their skills. Briefly stated, improvisers do not learn a technique once and for all, for the simple reason that each concrete situation in which a certain skill is applied is specific and cannot be anticipated. As each situation calls for a specific response, it requires a specific adaptation and thus development of the skill. A trumpet player, for example, finds herself playing with a new saxophonist. For this reason, she must, however slightly, adjust her way of playing to this new situation (for example, by altering her instrument’s intonation), perhaps abandoning conventions and techniques on which she used to rely. Not only that: in order to be creative, improvisers also intentionally try to stop using performing skills that have crystallized into behavioral habits. As in the case of Marcel Duchamp’s attempt to “forget with [his] hand” in order to create his mature works, this act of unlearning oneself can sometimes serve as a conscious and explicit artistic resource (even in arts that are not specifically improvisational). 
(b) Another aspect of the dialectical relation between preparedness and unpreparedness that is of particular importance is what one might call the epistemic paradox of improvisers:  improvisers are at once knowledgeable and destitute of knowledge. Before the performance, they know how to do, but they do not know what to do (despite the fact that, as we have just seen, even the skills constitutive for improvising are reworked through improvisations). They know how to play an instrument, recite, or dance, but they have no predetermined knowledge of what they will concretely do while playing, reciting, or dancing. Otherwise, they would not be improvising but executing a plan. Within an improvisation itself, too, knowing and not knowing stand side by side. Improvisers have, as we have seen, a procedural knowledge of the performing practice they are involved in, but they do not know what exactly the right action in a concrete moment of the improvisation is.  Knowing how to do something is not the same as knowing what one concretely has to do. She who truly improvises does not have a fully determined conception of what she is supposed to do, even if, possessing the appropriate skill, she knows how to do it. Hence, concrete aims and intentions do not precede the improvised action but rather arise and develop through, and in virtue of, the action itself. In improvisation, the plan of action is shaped through the action itself, in accordance with the specificity of the situation to which performers must adapt themselves. 
(c) This explanation allows us to account for the most important aspect of the dialectics of preparedness and unpreparedness in improvisation. Improvisation allows for the generation of unexpectedness, of the new.  The misconception of improvisation discussed above makes it seem as if the unexpected were simply one of the preconditions of improvising. Accordingly, improvising would require not only knowing how to perform within given constraints but would also imply facing situations we did not expect and do not know how to deal with, which would in turn trigger improvisational actions. But this is mistaken.
The basis for improvisation is not knowing what to do, but rather learning how not to know what to do, developing skills that enable one to recognize and appreciate performative occurrences not as external disturbances but as affordances.  The skills necessary for improvising thus have to be understood as entailing a sensibility for unanticipated events. If performers lack this sensibility, they can only repeat stereotypical, practiced patterns. This sensibility is needed for recognizing something unexpected and for making it productive for the development of the improvisation. Since something can only be unexpected in relation to some kind of expectation that is in play, we have to conceive of the unexpected as being constitutively related to an ability to recognize and react to it, i.e., to a skill that does not simply express itself in stereotypical repetition. Thus, like the cognitive and practical resources of improvisation, the unexpected (or the real “beginning,” as Jankélévitch would have had it) , too, is not an external precondition of what it means to improvise (which, by the way, is confirmed e contrario by the fact that many improvised performances are, in certain respects, quite predictable). Rather, the unexpected has to be understood as an outcome of improvisation that feeds back into the practice of improvisation itself.  In this sense, improvising should be understood as an exercise of phronesis: an ability to practically deliberate well about how to act according to the specific demands of the concrete situation, an ability that has the potential to transform the course of the improvisation. 
The dialectical relation at issue can be summed up in an oxymoron: the unexpected is what we expect to be confronted with in an improvisation. Less oxymoronically, this means that it is the concrete development of the improvisational process that allows or does not allow performers to treat “something that happened”  as an affordance for taking the performance in a new direction, integrating it into the performance, and thereby producing the unexpected. In Hegelian terms, the presupposition(s) of the process work(s) as such only if they are posited by/within the process.  The concrete development of the performance loops back into its preconditions, transforming their significance within the context of the specific performance. Thus conceived, improvisation is an autopoietic process that produces its own elements.  This is why the concrete sense of an improvisation, its “normative content” (as we will call it in Section 5), emerges out of interactions of different kinds. Hence, in order to get a better understanding of what is required such that something unexpected can serve as the foundation of a particular action, we have to shed some light on the interactive dimension of improvisation.
(2) Improvisations are made up of interactions.  Even soloists have to interact with the contextual circumstances of the performance, such as the space in which the performance takes place and its acoustics, the audience, and the cultural frame of the event, which generates expectations as to the kind of music that will be performed. And they have to interact with the resources of their practice, such as their own skills, style, artistic vocation, and professional career. For example, soloists interact with the musical instrument they are playing, variously reacting to the affordances it offers and searching for different ways to (ab)use it.  But they also interact with the cultural traditions of their artistic practice, variously referring to it, continuing it, appropriating it, (ab)using it, spreading it, or betraying it.
However, the most evident and obvious interactions within improvised performances are those that occur between performers in a group improvisation. Performers are at once producers and receivers: while they act, they also have to perceive what their fellow players are doing. Their actions are re-actions to what others do, and they in turn provide affordances for the others’ actions. Moreover, their actions, precisely as responses to and affordances for their fellow players, are evaluative.  For example, a certain melodic line the pianist plays in response to the rhythmic pattern offered by the bassist can be seen as a performative evaluation of the bassist’s action, inviting her to continue along this line, or to change direction, or to stop. The bassist, also evaluating what the pianist is doing, can choose to accept or decline his invitation, collaborating or contrasting with his musical offer. In the course of the performance, continuous evaluations of its different moments are expressed by improvisers in the form of performative contributions. These evaluations are not contemplative but have performative power:  they produce effects on the performance, guiding the decisions taken (often unconsciously) and driving the performance forward.
The intrinsic connection between performance and evaluation can be conceived of as an expression of the normative dimension of improvisation. Improvisers share a (more or less) common (cultural, artistic, and stylistic) background that provides them with conceptual schemas by virtue of which they are able to represent and evaluate the ongoing situation and solve problems of coordination.  However, the referential framework that guides performers’ (and audiences’) expectations and sets norms for the meaning of what is happening is one of the preconditions of the improvised performance. And, as we have just seen, improvisation performatively feeds back into this framework. This means that the latter is not stable and static. Rather, it is dynamic and changes (or may change) during the performance. In fact, it emerges during the performance in a way that goes beyond performers’ subjective intentions. But why does the normative framework of the performance change? The reason for this dynamic character of the normativity of improvisational practices is precisely the evaluative dimension of the performance itself. What fellow players do places constraints on those who respond to it during the improvisational interaction, thereby functioning as a (kind of) norm. Each action and gesture (e.g., in a musical performance, a particular accompanying figure played by the drummer) contributes to shape the normative context of the improvisation, in that it becomes a constraint for the activities of other players. In turn, its sense and normative force is configured by the ways the other performers react to it. 
So we are faced with the same conceptual structure already emphasized in relation to the links between preparedness and unpreparedness, plan and action, and improvisation and unexpectedness. Norms guiding improvised performances do not simply precede them but are established within their concrete development, which means that the referential context of the performance is continuously re-created in the course of the performance itself.  Every act and gesture constrain and anticipatively shape the meaning of future acts and gestures and, reciprocally, the meaning of every past moment or sign may change depending on what happens later.  Hence, the normativity of the performance is continuously in flux. In order to better understand this, we have to analyze the concept of normativity itself.
3 Normative practices
In Section 1, we claimed that a common misconception of normativity asks us to distinguish between practices that are guided by stable norms and practices in which no guiding norms are available, so that it is necessary to first establish them in order for the practice to work.  Our position against this misconception goes as follows: the guiding norms of human practices are based on dynamic relationships between norm-oriented subjects and the world and, in particular, on reciprocal relationships between varying constellations of individuals and groups (or collective agents). Normative practices function in as much as they are open for developments and transformations prompted by the world and especially by other individuals and collective agents. If normative practices lack this openness, they “freeze,” thereby losing their normative force because their guiding norms no longer cope with reality and thus become inapplicable. In order to give support to these claims, we will (1) defend the notion that the world plays a role in the dynamic constitution of norms and (2) retrace the importance of interactions between individuals for the dynamics of the constitution of norms. Against this background, we will then (3) be able to illuminate the improvisational dimension of normativity.
(1) Every conception of normativity has to deal with the problem of application. Wittgenstein famously explained this problem by pointing to the threat of infinite regress. If we conceive of a norm’s application as being separate from the norm applied, we need a second norm to bridge the gap between a norm and its application, which in turn demands a third norm that bridges the gap between this second norm and its application and so on ad infinitum. In order to avoid such a regress, norms should be conceived of as being enforced in and through their very application. According to Wittgenstein’s lesson, norms are only effective and actually real through their application. Consequently, the applications of a norm are essential for establishing and reactualizing the norm. Norms are what they are through their application, and unapplied norms, i.e., norms valid only “on paper,” are not norms properly speaking because they norm nothing. Thus, an explanation of what norms are has to focus on what it means to apply a norm. 
Now, every application of a norm has to deal with the actual specific situation the norm is applied to. However, norms ensue from specific situations and cannot be applied to new situations in the same way. They are not programs that are automatically – and successfully – activated in the same circumstances forever and ever, if only for the simple reason that circumstances are always changing and every new situation to which a norm is applied changes the norm itself. The norm has to be adapted to the changing situation: hence, we have to think of chains of applications as constantly developing and transforming the norm at issue. This means that norms are not indifferent to the world. On the contrary, norms are determined by the world and its changes, thereby undergoing changes themselves. Let us call this dimension of the norm the worldliness of norms.
(2) A second constitutive source of the development of norms lies in interactions between the individuals who apply the norms in question. This is especially clear in the case of judges who apply norms to specific cases. If a judge claims that a norm is relevant for a specific case (in short, “first judgment”), other judges will evaluate what has been judged. If these other judges find the first judgment to be convincing, they will affirm it and make it into a precedent for their own judgments. They continue applying the norm as it was shaped by the first judgment. But if the other judges disagree with the first judge, they will not let the first judgment inform their own judgments – or, as a different reaction, they will only refer to it negatively. In such cases, the first judgment will be excluded or remain without resonance. 
What holds for judges applies as well to other norm-applying subjects in all normative fields, including the understanding and use of concepts.  Someone who speaks of “poverty” with regard to specific individuals or situations is confronted with others’ reactions to what she says. Either they will agree with the application in question and will take it as relevant for their own ways of using the concept “poverty,” or they will reject the use of the concept in the specific situation and will not rely on it when they themselves use the concept in question. For instance, one person may have a certain standard of richness and poverty in terms of goods and money individuals possess that is informed by their own specific existential situation, and someone else may have a different standard that is informed by their specific life situation and lifestyle. So, using the same expression, they may mean two different things. Moreover, one can also use the notion of “poverty” in reference to nonmaterial goods, such as intelligence or variety of cultural experiences, or in association with notions such as “innocence,” as in the Christian saying “poor in spirit.” These different applications of the norm (the concept) in interaction with others change the context required for the enforcement of the norm: in this case, the meaning of the word “poverty.” Like in the case of judges, norm-using individuals interact with one another, and their interactions are an essential aspect of the developments taken by norms. Let us call this dimension the intersubjectivity of norms.
(3) Both the worldliness and the intersubjectivity of norms are constitutive for a view of normativity focused on norms’ applications. In addition, it is important to stress that no norm stands on its own. Norms develop in relation to each other. Once again, Wittgenstein can help us understand why this is so. In his famous argument against the possibility of a private language, Wittgenstein stresses that norms function as criteria for one another.  The application of a norm is always related to applications of other norms, whether implicitly or explicitly. Norms follow from each other, they contrast with each other, they rule each other out, and they explain one another. In many different contexts, and even cross contextually, norms are bound up with other norms to the extent that the development of a norm is constitutively influenced by other norms (and by their own developments). The example of the concept “poverty” we just touched on is telling in this regard. Once the concept is connected with “innocence,” it changes. It changes according to how we understand needs, too. In short, norms form webs of (interacting and changing) norms – not only in language but in bodily gestures, in the arts, and in all the symbolic articulations of practical and cultural life.
On the basis of the foregoing explanations, it is easy to see that the dynamics of norms have an improvisational dimension. This is clearly shown by their intersubjectivity: norms’ applications and developments are affirmed, rejected, or modified in the course of interactions of which norms are elements. Even though in some human practices (like legal practices) the way norms are applied is more explicit (even publicly declared) and controlled, normative practices always have an interactive dimension of an improvisational kind. The application of the norm relies on the development the norm has taken thus far in its complex position within a web of norms; however, it is also exposed to improvisational interactions, since there is no way to prevent the emergence of applications of the norm that might change it. This means that no norm-using individual has the last word as to how the norm has to be applied. Such a last word would be contradictory in itself. It would be a way of stepping out of the practice that, at the same time, would control the practice, therefore being, paradoxically, still involved in it, which is logically impossible. To put it bluntly, nobody can invoke something like a super-criterion to determine how to understand and apply a specific norm correctly. The meaning of norms and what constitutes their correct application are established by practical negotiations that take place in the course of the norm’s application itself. So, it certainly may happen that, for pragmatic reasons, interactions are cut off (as in the case of courts of appeals). But even in such situations, interactions continue – perhaps in other contexts or at different levels – thereby further developing the norms at issue in more or less unpredictable ways. This is why every engagement with a norm and its past applications is improvisational in nature. 
Let us recall what we explained in Section 1 of our article. Whoever engages with a norm is prepared in many ways. She masters the norm in question and its connections to other norms and has acquired the capacity to – whether explicitly or implicitly – performatively evaluate norms and their applications. But all these preparations do not guarantee that she has an absolutely certain way of dealing with the norm in question or that she is able to apply the norm suitably in every single concrete situation. Perhaps she does not see the concrete specificity of the situation or she resorts to the wrong norm. Every way of commenting on and continuing a normative practice has to be able to bear uncertainties and irritations. Consider the example of the judge again. Even if a judge is the best in a specific field and even if she has done everything to attain an ideal preparation for the cases she is dealing with, she still always has to face the risk of erring. Like in an improvisation, it is always possible not to succeed. Someone who perfects herself in her capacities still opens herself to making mistakes. For comparative purposes, consider the person who just acts blindly and thus outside of a normative realm: she cannot make mistakes. Thus, a symptom of the improvisational character of normative practices is that we always have to face the possibility of making mistakes within these practices. It is not possible to diminish this possibility, no matter how well one is trained. Even though one might learn to take, for instance, lots of exceptions into account, the possibility of error – i.e., of the unpredicted exception that is not taken into account – holds in general. To put it with Derrida, we might say that it is not possible to calculate the risk of failure. 
However, there is another point to consider here. “Mistakes” may turn out to be affordances for positively changing the norm in an unexpected situation, thereby inventing a new norm in which the “wrong” application of the (old) norm is salvaged as a good application of the (new) norm. Mistakes are not always simple disturbances of normative practices. They are disturbances of normative practices that, by contributing to the transformation of the norms guiding the practice, may turn out to be successful outcomes of the practice. Artistic cases are well-known: artistic performances considered mistakes in reference to established artistic norms may turn out to be affordances for creatively changing the norms and moving the practice forward.  But while crucial for understanding the normativity of human practices, artistic cases ordinarily do not face the risks encountered in everyday life, where unforeseen events and errors can generate significant problems for people’s lives. In such cases, errors, as “deviations” from established norms of action, can be fatal; nonetheless, sometimes they may turn out to be right after all.
In this regard, let us think about the real history of “Sully” Sullenberger, the pilot who in January 2009 was able to come up with a way to save the passengers of US Airways Flight 1549, which encountered an unpredictable accident. As shown in Clint Eastwood’s movie Sully (USA 2016), if the pilot (played by Tom Hanks) had followed the routine emergency procedures by trying to land at a nearby airport, the plane would have crashed, as experts were able to later demonstrate. Instead, going against the established norms – that is, by intentionally committing an error – he succeeded in saving all the passengers and the crew. Having understood the specificity of the situation, Sully invented a plan of action while carrying out the action itself, succeeding, for the first time in history, in ditching an Airbus A320 in a river.
Sully’s performance was, in a way, a paradigmatic case of phronesis (see Section 2). We might better explain it by saying that he was in the same situation as improvisers who, as we stressed, are in the double position of knowing and not knowing: they know how to perform a practice (in his case, the practice of flying an airplane, even in difficult circumstances), but they do not know what exactly they have to do in a specific situation (since routine emergency instructions were unsuitable for that situation and until then nobody had ever landed a passenger jet on the Hudson River). This was an extreme case.  But norm users are often involved in analogous situations, even if they are, fortunately enough, usually less dramatic. Norm users are not always in a position to know what they have to do with the specific norm they are dealing with.  In this sense, norm users have no plan for what they are doing. Or, more precisely, they develop a plan while applying a specific norm.  At the end of the day, this is what the ability to judge concretely consists in. It is exactly here that the aforementioned Kantian distinction between determining and reflecting judgments collapses: every determining judgment is bound up with reflective aspects and thus includes a (trans)formation of the norm applied.
At this point, a potential objection might be raised about the way we are linking normativity and improvisation together. In Section 2, we described improvisation as normative and habitual. Here, we describe normativity as improvisational. However, as we discussed earlier, the reason normativity has an improvisational character is not that improvisation is habitual, but rather that improvisation is unpredictable, since it is always bound to the specificity of the performing situation. So, the objection might continue, we are connecting the two concepts equivocally. The reason why normativity is improvisational seems to be exactly contrary to the reason why improvisation is normative in character. The best way to counter this objection is to make the dialectical character of the conceptual connection between improvisation and normativity explicit. Novelty and unpredictability (in terms of emergence, change, irritation, transformation, etc.) are not opposed to norms as guiding criteria and, as we shall see in Section 3, as habits. In other words, novelty, so conceived, is not contrary to the way norms organize human practices. Rather, it is an essential element of how human practices work normatively, which is to say, by coping with the specific reality of concrete situations. Norms only work (are applied and enforced) if they can be (creatively) changed, which is to say, if their sense and/or their normative power can be (trans)formed through interactions between individuals and through confrontations with a changing reality. And norms are again and again opened to the new because only in this way can they guide practices (in their ever-changing situations) and be effective as norms. Correspondingly, as we saw earlier, improvisation does not rule out guiding norms of different kinds but rather requires them in order to be successful. But, as we are now better able to understand, norms guiding a practice, including a practice like improvisation (musical improvisation, for instance), are themselves developed within the practice, since there is no extrapractical super-criterion for regulating how the norm should work as a norm.
This allows us to take a last step in our explanation of normativity. It consists in explaining how normative practices can freeze. If the application of norms becomes automatic, normative practices become impoverished in as much as they tend to lose grip on the world and cannot deal with the unexpected. If norms are applied in a readymade way, they do not develop further, which means that they cannot cope with the world. In the movie Sully, after the incident, the pilot was accused of not having followed the usual procedures, i.e., of having broken the rules, thereby endangering the passengers’ lives.  But, as was later demonstrated, the exact opposite was the case. If the unexpected event had not invited him to change the norm in its application, making up the rule as he was going along (or, less elegantly but perhaps more precisely, integrating the unexpected into the norm, thereby producing an unexpected norm), the standard norm would have been dramatically ineffective.
Even though what norm users are confronted with in the world depends upon cultural and social structures and general patterns of expectation, sometimes what happens here and now calls for a response that could never have been planned. Thus, human rationality is not reducible to rigid (algorithmic) rules and inferential thinking.  For rationality is responsiveness to potentially unforeseeable worldly and social facts and events. Rigid, unchanging rules tend to lose touch with what is. By failing to respond to the specific nature of the situations to which they are applied, they fail to achieve their real purpose, instead calcifying into algorithmic codifications of procedures considered as ends in themselves. Indeed, this is a common deficiency of normative practices, as norms tend to become stereotypical, automatized, and estranged from the world. If we think of norms as providing stability, we tend to understand normativity on the basis of a defective version of it. This brings us to the third misconception we have to deal with: the misconception of habit.
The dynamic conception of normativity elaborated thus far can be misunderstood as implying that normative beings are constantly reflecting on how to apply norms. So, since we have rejected the view of normativity as an automatism, it would seem that we are endorsing an overintellectualized conception of human practices according to which human practices, as normative practices, are laden with intellectual reflection. However, as phenomenology  and pragmatism  have taught us, this notion of human practices would be highly questionable. Human beings are not continuously and intentionally reflecting on what to do while applying norms in their actions. Thus, our task is now to demonstrate that, as improvisational artistic practices paradigmatically show (see Section 2), the dynamics of human practices are not bound up with constant intellectual reflection: we perform normative practices without continuously questioning them explicitly. And this is so, not in spite of, but in virtue of their improvisational aspects. The point is that normative practices, as rational responses to worldly facts, are habitual practices; they are grounded in habits. The skills discussed that serve as preconditions of improvisational activities are basically realized in habits, i.e., in dispositions or, as Peirce felicitously called them, embodied “rules of action.” 
It is generally accepted that habits give a rhythm to everyday life and that they allow us to perform in a regular, quick, and efficacious way operations that otherwise would require too many resources and too much effort. So, as important philosophical traditions have pointed out, habits are key elements for the formation of personal identity, interpersonal relationships, and the development of cultural processes.  But how are habits to be understood? As outlined in Section 1, a widespread view sees habits as automatic and repetitive patterns of behavior that have the function of easing the everyday through the formation (or inculcation ) of behavioral routines that have to be applied automatically to all situations of the same kind. As discussed by Sawyer,  attempts have been made to conceive of social situations – everyday situations like eating in a restaurant – as “scripts,” understood as internalizations of “social texts.”  According to this dramaturgical model of social life, when we enter the situation “eating in a restaurant,” we simply apply the corresponding ready-made algorithm of behavioral habitual patterns prescribed by the internalized scripts. This view, we claimed in the introduction, implies a sharp contrast between habits, misunderstood as unreflective and uncreative, and improvisation, also naively misconceived of as spontaneous creatio ex nihilo. In order to avoid these misconceptions, we have to revise our concept of habits by endorsing a view of habits that has been defended by philosophers like Hegel, Ravaisson, Bourdieu, De Certeau, and Malabou.  We will do so by explaining (1) the inherent dynamics and (2) the inherent reflective and plastic dimension of habits. This will allow us to understand the skills and activities that make up normative–improvisational practices as practices based in habits.
(1) Our main point is that habits have to be distinguished from algorithms. Algorithms regulate specific functions in a fixed way; thus, they either work or they do not. If the situation to which they are applied is inappropriate, they simply do not work. Adapting an algorithm to a different situation means replacing the first algorithm with a different one. There is no transformation, only substitution. Conversely, habits do not work in a fixed way like this. Habits can perform their function of regulating human and social behavior and simplifying (better: making possible) everyday life only if they are adaptable to “disturbances” that ensue from the different specific situations in which they take place. Thus, when something is not as it used to be and expected to be, habitual behavior, differently from an algorithmic program, is able to plastically adapt itself to the changed situation, coping with it. In this sense, habits “make possible a future.”  For, again in contrast to algorithms, habits are not simply executed in automatic and stereotypical ways but can be modified according to a shifting reality. In addition, they can also be intentionally shaped by subjects in order to change reality in a proactive way. 
This sensibility for disturbances and changes has to do with the evaluative and normative dimension of human practices we discussed in the previous sections. After all, habits are not only routine rules for accomplishing things in an “easy” way; they are also responsible for the generation and realization of values that give orientation to human practices in specific situations. Hence, although human beings certainly typify, individually and collectively, the situations of their daily as well as of their cultural life, and although habits not only constitute a useful and rational conduct of life but – as already mentioned – the very condition of possibility of human personal life and culture, typified situations, just like the norms and values that regulate human actions, are not fixed and unchangeable. Rather, they are dynamic, in that each singular concrete situation is specific and therefore cannot be entirely anticipated. So, while a behavioral routine of a typified situation is applied to a specific situation, this specific situation retroactively influences the routine, impacting, if ever so slightly, its “structure” (or its “script”). In other words, in order to be efficient and to adequately cope with the concrete and unprecedented aspects of a specific situation, habitual routines of typified situations have to be plastically adapted to it.  Moreover, habits arise, thanks to their being continuously, and inventively, adapted to the specificities of different situations. Human practices, as De Certeau wrote,  work by appropriating and adapting institutionalized norms and habits to particular situations in unprecedented ways, inventing, through the exercise of habits, spaces of freedom and novelty that (trans)form the habits, thereby creatively producing novelties. Thus, habits are always in flux and continuously (if slowly) changing.
In short, habits are plastic configurations – generated through repetition and training – of the essential dispositions and skills of individuals (and cultures). They are not a priori and readymade properties of human beings but come into existence a posteriori through the continuous evaluative, adaptive, and creative interaction between human beings and their changing natural and cultural environments. They are the essence of individuals but are nonetheless modifiable and generated by continuous changes. In Hegelian terms, we can say that habits “enable the Bildung of the self”:  they are dispositions through which individuals shape themselves and their interrelations, becoming and developing the particular individuals they are in continuous dynamic interaction with reality.
This understanding of habits makes explicit the dialectical relation between habits and improvisation. Not only, as previously seen (see Section 2), are improvisational performances made possible by habitual skills, so too are habits, as “the intentionless invention of regulated improvisation,”  constitutively bound up with improvisational practices. The plasticity of habits can only be realized if the individuals whose character and identity are formed through habits are constantly improvising in their interactions with others and with the world. Habit is both a result of improvisation and a precondition of it. It is a “moment of improvisation,” since “it is responsible for rather than resistant to the transformation of human action and the creativity associated with that.” 
(2) The plasticity of habits has to be explained in terms of inherent processes of reflection. But in which sense are habits reflective? Habits are reflective, we contend, because, as previously suggested, anyone who carries out a habitual routine is not simply executing the rule of an algorithm but also exposes her routine to the situation in which she is acting. Hence, in habitual practices, habits themselves are constantly evolving and reflecting the ways in which they have to be plastically adapted to an ever-changing reality. However, it would be mistaken to take this reflective quality of habit as being merely cognitive in character. Human beings do not simply theoretically represent changing situations and their habits and then try out different ways of matching the former with the latter. Rather, evaluative and practical sensibility for emerging and unforeseen disturbances of reality is constitutive of habits. Someone who acquires habits equips herself with this evaluative sensibility by exercising it in her practices.
The inherent reflexivity of habits as practical norms of action allows us to deepen our understanding of their improvisational nature. We have stressed that improvisation is bound up with continuous evaluation. Improvisers take evaluative stances toward what their fellow players bring into play. However, this is usually not a conscious process. Rather, evaluations are generally realized by bodily reactions. Thus, we follow Bourdieu in thinking that habit “is a spontaneity without consciousness or will.”  Habits are a spontaneity without consciousness or will in the sense that they entail an implicit evaluative awareness of aspects of the world but do not involve any conscious evaluation or decisions. In short, habit, as disposition or embodied rule of action, is embodied reflection.
Finally, habits should not be misconceived as being exclusively private dispositions. Although they shape the character and the lifestyle of individual subjects, their formation as well as their continuous development take place in interaction with others.  Habits, as practical and plastic dispositions of actions through which human beings organize their behavior while adapting themselves to an ever-changing reality, have a social dimension: they are acquired by integrating other agents’ reactions and habits into one’s own habitual practices. In particular, social practical norms are incorporated into individuals’ actions as social behavioral habits.  Briefly stated, the interactive dimension of human practical life discussed above is constitutive for the shaping, the development, and the application of habits. Hence, although it may be maintained that habits shape the character of individuals, which, albeit in flux, is more or less “contoured” and publicly recognizable, habits bind together, rather than isolating, individuals in that they emerge out of dynamic interactions with others.
The reconception of habit developed thus far gives us a good foundation for, finally, explaining how we have to think of freedom in order to open space for a conception of rational practices as improvisational practices. It seems obvious that we have to treat improvisation as a realization of freedom. Improvisers are free to react to what they are confronted with in the course of a performance. However, as we pointed out at the beginning of this article, this can too easily lead us to misconceive freedom as having its source in the spontaneously acting subject. But this is a one-sided conception that should be overcome because, in actuality, we have to negotiate our freedom in and through interactions with others and with the world. The foregoing reflections on normativity and habit allow for a better understanding of how freedom is realized.
According to a commonplace, we have to distinguish between two conceptions of freedom, namely, negative and positive freedom. Many see exclusively negative conceptions of freedom as problematic,  claiming that freedom is not really actualized if we are only free of external constraints. Freedom, they think, needs to be understood as being positively manifested in a realization of something specific. However, the conception of positive freedom is itself a source of further theoretical problems that are pertinent to our discussion. By stressing that freedom has to be conceived of as positive freedom, one tends to understand the dialectics of freedom and constraints in too narrow a way, namely, by stating that freedom is realized through constraints one gives to oneself. I am free if I constrain myself through norms that I take as binding me. According to such an explanation, self-binding is the source of freedom.  However, this understanding of positive freedom is mistaken. In order to explain why, we need to return to improvisational practices, which we will do in two steps.
(1) Let us go back to the example of interacting musical improvisers from Section 2. A pianist, for instance, reacts to the saxophonist with a new rhythmic pattern. What she does is an expression of a claim to freedom. But is it already a realization of freedom? We have to understand the rhythmic pattern established by her reaction as a proposal that she is offering to her fellow players. It is up to them to make something out of the proposal in question. It may be that the fellow players do not react to the rhythmic pattern, which they might find not promising for further development. But it might also happen that they react intensely and make the pattern at issue the “new norm” of what they are playing further. These different possibilities should make clear that it is only through the – at once evaluative and performative – reactions of others that the pattern at issue becomes what it is. The pattern is determined through the development it takes in the practices that respond to it.
We can explain how practices respond to specific proposals in terms of normative content. The normative content of the pattern is determined through interactions within the performance. It is not determined by the mere action of being played once. One may object that, in “Sully’s story,” the pattern of action was “played” only once. However, it obtained its normative force, as well as what we might term its socially performative power, retroactively, through the evaluative reactions made during the investigations conducted by aviation agencies. In this way, Sully’s decision not only was successful but also became the right decision. Its normative force was established in the course of a practice of evaluative interactions.
In other words, freedom is not realized by binding oneself to some norm. Its realization requires others’ reactions to what one realizes. This allows us to better understand how to conceive of the dialectics of freedom and constraints: to realize freedom one needs the constraints of others. If I want to realize a free action, others have to react to my action by explicating in some way that it makes a valuable contribution to whatever practice is in question. This does not imply that they have to agree with what I do. They may criticize me or answer with a different proposal or with a variation of what I have done. Still, it is through reactions like this that what I have done gains the status of something I have done in a self-determined way and thus freely. The most significant factor in my actions not being granted such a status is that others simply remain indifferent to them.  If they do not respond to me in any way, if they do not give me any attention by how they (re)act, I am not reflected in doing something out of self-determination. If nobody in the band reacts to what the pianist plays, her acts would not contribute to shaping the musical meaning of the performance and she won’t be realized in her freedom. The realization of freedom requires the recognition of our freedom by other subjects with whom we interact, sometimes conflictingly,  as can also happen in an improvisational performance. 
One might object that it is easily possible to do a lot of things alone, without receiving others’ evaluative reactions. Isn’t it absurd to state that the status of being someone who does something in a self-determined manner is obtained through reactions of others? Doubtlessly, it is an expression of freedom as self-determination to prepare a coffee in the morning in solitude, or to think through an objection in solitary reflection. But what one does in cases like this depends indirectly on others. What it means to drink a coffee, to appropriately clean a cup, to buy coffee, is determined by interactions with others. The status of being someone who does something out of self-determination is dependent on others’ participation in these interactions. Even though it is not necessary that others react to what one does when one prepares a coffee in the morning, one has to stand in relation to others by which they react to what one does if one cleans a cup, buys coffee, etc., as something that one does out of self-determination. Thus, the status of doing something in a self-determined way is not realized through one single action. Rather, like in a solo performance, its realization depends on webs of action that bind the performer to others. Some actions within these webs are not directly bound up with others in the sense that they do not as such prompt reactions from them. But the status of doing something out of self-determination is dependent on the reactions of others to one’s actions.
(2) However, another relevant aspect of the realization of freedom has to be addressed. Up to this point, we have described the example of the jazz musician as if her reaction comes out of the void. But as we have stressed many times, it would be a mistake to understand it this way. Jazz musicians need intense training to be able to react to others in apt ways. They must develop skills and learn how to interact with others successfully. As we have shown, performative skills can be understood as habits acquired by means of practice. Such habits do not execute themselves automatically but are internally bound up with forms of critical reflection and require attention in order to be adapted to concrete situations. At least in this sense, they are creative. Hence, habits are, on the one hand, an essential precondition for the freedom improvisers realize through their (re-)actions, and, on the other hand, are themselves generated in a performing practice.
Seen in such a way, freedom is constrained not only by the reactions of others with whom one interacts but also by habits one has acquired. Habits are an indispensable foundation for the realization of freedom,  for without habits, we could not successfully interact with others and with our natural and cultural environment. Hence, habits ground free practices. However, habits are not an aspect of human existence that, while making free practices possible, might stand outside the actual realization of these practices. As already seen, they are not automatic algorithms that “free” our everyday existence from the burden of freedom. Rather, they are involved in the realization of free practices. We can get a deeper understanding of this idea by considering the connection between freedom and unexpectedness (see Section 2). If we are only confronted with aspects of the world that we already know through and through, we cannot act spontaneously. We act automatically, since we are prompted to simply repeat what we have already done in the past. Hence, in order to be able to act freely, we need to confront aspects of the world that we do not know in advance. Put differently, we need to be confronted with something unexpected. It is this structural aspect of what free action implies that helps us understand the relevance of habits. As we have shown, habits have to be understood as dispositions to act in varying circumstances, and they shape themselves by means of plastically adapting themselves to an unexpectedly changing reality. Because of this, habits enable us to deal with something unexpected and are shaped through this confrontation with the unexpected. Habits open us to experience unexpected aspects of the world. In this sense, habitual behavior is an inherent part of our practice of freedom.
To sum up, the dialectics of freedom and constraints entail at least two aspects. (1) The realization of freedom needs the constraints provided by interactions with others. Through these interactions, binding intersubjective relationships are established because the interactions concern what we have called normative content. In webs of actions and reactions, the normative content of practices is determined through evaluative performances, binding those who, through their actions, place themselves in relation to the developing normative content. Moreover (2), the realization of freedom is constrained by habits that are preconditions for dealing with unexpected aspects one is confronted with in the world and, at the very same time, are performatively shaped through the realization of free practices. Briefly, freedom is realized by dealing with relations of dependency in which subjects’ stand that, at the same time, are not static constraints but dynamically develop as we go along. We need a conception of positive freedom in this vein in order to properly account for the improvisational nature of rational practices.
As we have highlighted in the different steps of our argument, the outcome of this investigation of the concepts of improvisation, normativity, habits, and freedom has a dialectical structure. On the one hand, improvisational practices are misunderstood if they are taken as relying on a continuous invention ex nihilo and, as such, as being wholly different from the everyday habitual norms and dispositions that organize our activities and interactions. On the other hand, human normative practices are misunderstood if they are taken as rigid repetitions of “frozen” behavioral patterns insensitive to the changing concreteness of reality and, as such, as being opposed to improvisation as a form of spontaneous, creative and adaptive agency that is capable of generating and dealing with unexpectedness while interacting with the affordances of the world. Improvisational practices are based on extensive preparatory work resulting in skills and habits; the normativity of human rational practices, precisely by relying on habits, should not be understood as fixed but as plastically and creatively changing in order to cope with reality. It is in this sense that they make human freedom real.
Thus, our considerations have an anthropological dimension that might allow us to overcome the Kantian paradox mentioned at the beginning of the article: the paradox of a freedom that binds itself to rigid norms. The rigid norms explanation of rationality that defines it as self-legislation independent of external influence falls short because it fails to explain the ways in which human beings interact with one another and in which “the space of reasons”  is responsive to a changing world. In the face of these shortcomings, philosophers might want to seek out an alternative explanation of how rationality realizes itself in the world. We have tried to articulate a possible way out of this impasse that can be articulated as follows: an essential feature of human rational practice is its openness for practical–reflective reactions to unexpectedness and freedom, which, in turn, are not simply “externally” presupposed givens, but are “internally” generated by our practices.  Such an understanding is directed against the idea that human beings, as the “flawed” animals, are fundamentally open, an idea that entails that reason consists in developing its determination out of itself.  Thus, we maintain, understanding human practices as improvisational means interpreting openness and freedom as something that is realized, and not as something that is given in advance. Our ability to perform the unexpected is an achievement, a result of practices developed as responses to the states and circumstances of human life in its natural and cultural environment. Thus, that “we make up the rules as we go along” not only means that we shape specific norms of our practice in and through our practice. It also means that we shape and determine our freedom through our practices, thereby realizing key (pre)conditions of our practice. In other words, human rational life is improvisatory because the conditions of the practice are won within and by the practice itself – freedom maybe being the most important of them.
However, the following objection might be raised: doesn’t our explanation of the improvisational dimension of normativity, habit, and freedom imply a conflation of artistic and nonartistic practices? While understanding rational practices as improvisational practices, are we not taking company with Joseph Beuys who famously claimed that every human being is an artist?  This would be a profound misunderstanding of our thinking. Our explanations do not imply that rational practices have an artistic dimension per se, although art does contribute constitutively to human rationality.  Rather, the improvisational dimension of normativity, habit and freedom is, in our view, primarily to be understood as nonartistic. It is nonartistic in that ordinary rational practices have an improvisational dimension, without them necessarily being (connected to) artistic practices: they can be connected to art, but they don’t have to be. 
Nevertheless, this improvisational dimension certainly is the basis of artistic forms of improvisation. Thus, we take our argument to imply that rational practices are constitutively open to artistic impulses developed by specific forms of improvisation in art. Hence, we consider improvisation as a common feature of rational and ordinary practices, and we consider specific forms of artistic improvisation as being rooted in this common feature. Surely, further analysis is necessary to explain the differences between everyday forms of improvisation and their artistic siblings. Such an investigation would have to meet the challenge of explaining both the distinction and the connection between everyday rational practices and artistic improvisational practices. However, for our present purposes, it suffices to underscore that nothing of what we have argued for commits us to the idea that improvisation is necessarily artistic in nature. In the way we have developed it here, improvisation has to be understood as a basic trait of ordinary rational and everyday practices, since, as we have shown, the distinguishing feature of human rationality is that “we make up the rules as we go along,” i.e., that we follow and (trans)form our norms of action while coping with the world and with others.