As common presumption goes, there is no legal decision without its factual grounds.Footnote 1 What does it mean? When deciding the case, judges (or other officials) have to establish propositions of fact, inferring what counts as evidence in law from objects of sensory evidence and other propositions. This reflects on four basic senses in which the concept of evidence functions: as an object of sensory evidence, as a proposition of fact, as an inferential premise, and as that which counts as evidence in law (Ho, 2021). Our main goal in this article is to refine the understanding of the concept of evidence (i.e. evidence in the fourth sense) by explaining what participants in legal proceedings do while appealing to different propositions of fact (i.e. evidence in the second sense) and inferring from them (i.e. evidence in the third sense). The assumption is that the participants “collect” evidence (propositions of facts) which appears to be useful from the perspective of the participants’ other reasons.

Already in 1985, Robert Allen postulated revolutionising evidence law by understanding the rules of evidence as not only structuring the admission and exclusion of evidence, but also as determining respective roles of the judge and jury and hence granting appropriate attention to the relationship between different actors of the trial. We want to pursue that intuition by providing a description of the relationships between different participants in the legal proceedings, observed through the prism of the reasoning that is performed by the evidence-finding entity. Simultaneously, we will analyse how this reasoning is influenced by actions taken by the participating parties. The novelty of this approach lies in the dynamic picture of legal proceedings, where the participant’s actions take the form of the exchange of reasons. Because of those constant interactions, the current set of relevant reasons—the reasons potentially affecting the outcome of the proceedings—can always change throughout the proceedings’ course. In this picture, introducing new epistemic reasons by the participants amounts to the change of the relevant set of reasons in the proceedings—both as to its content and as to its structure, eventually influencing the final conclusion.

Therefore, we believe that to fully understand what evidence in law is, we need to reflect on the reasons that different agents have for participating in the legal game, so as to the stakes of this game.Footnote 2 This is because the legal concept of evidence is (at least in part) constituted by the role that evidence plays in affecting which reasons for action the participants in legal proceedings choose to act on. For, as we will claim, evidence is not there merely for creating epistemic reasons for the decision-making entity, but also for changing the subject of a game, and hence for altering the outcome of interplays between different types of reasons for action that participants in these proceedings may have.Footnote 3 The second goal of this paper is metatheoretical: we want to show that when analysing what legal evidence is, one should begin from the perspective of a singular instance of legal proceedings, rather than from the perspective of law in general.Footnote 4

Thus, we claim that the process of applying the law is ultimately rooted in the inferential discursive practices of exchanging reasons on the part of the participants in legal proceedings.Footnote 5 We distinguish between three types of these reasons: legal reasons (‘L-reasons’ which derive from legal rules),Footnote 6epistemic reasons (‘E-reasons’, understood as propositions of fact) and stake reasons (‘S-reasons’, rooted in the idea that there are different possibilities to resolve a particular case and different consequences thereof). Special attention is given to the role of the E-reasons, which may support the legal solution (founded on L-reasons) in fitting the S-reasons of the decision-maker. The metaphor of law as a game is employed here as an explanatory tool to describe what different agents do when they engage with these reasons.Footnote 7 From the linguistic point of view, the notion of stake intuitively relates to playing something. Therefore, the perspective of treating legal proceedings as an instance of playing a game draws attention to the two overlooked factors that—as we will argue—are constitutive for the legal proceedings, namely: (1) the interplay between different types of reasons that legal actors may set forth throughout legal proceedings, of which a lawsuit is an illustrative example and (2) to the instance of stake to the game that influences this interplay. The metaphor of the game also serves our metatheoretical purposes—it enables us to show how an instance of playing a game (legal proceedings) relates to a GAME in general (law as such).

The paper is divided into five parts. First, we start by introducing the fact-based story of Mary, which is to illustrate the notion of the parties’ stakes within the particular proceedings. In the second part, we provide reasons for distinguishing between GAME in general and an instance of a game. The third section is where we introduce our model of a particular legal game. The game is to be understood, as already suggested, as an interplay between three types of reasons which we broadly describe. We also try to capture the dynamics of a particular game using our proposed conceptual apparatus. The fourth part is devoted to the metatheoretical project that emerges from our depiction of a game and its relation to the GAME. Since we believe that the adequate direction of analysing legal evidence leads ‘from a game to GAME’, in this section we show the downfalls of adopting a reverse standpoint. To explain this, we juxtapose our approach with that of Andrei Marmor, who also applies the metaphor of law as a game. However—according to our interpretation of his view—he adopts an explanatory perspective different from ours in being directed ‘from the GAME to the game’. In the fifth section, we return to the initial story of Mary to show how the adopted perspective of reasons-relations broadens our understanding of her case.

1 The story of Mary—introducing the Stakes

Let us begin with telling a story that is closely related to the case which took place in Poland in the early 1990s.Footnote 8 Imagine a woman, let us call her Mary, living with her family in the rural areas of the country, close to the woods, rather remote from other premises. One night when Mary is home alone, the lights suddenly go out. Mary takes her husband’s gun and leaves the house, thinking that the blackout was a preparatory move of the thieves aiming to rob her house. At once, she hears some shouts nearby. As a response, Mary fires several shots into the darkness. As it turned out, on the following day, the man who had been trying to steal a copper part of the track, sitting on the electricity pole, was reported dead.

During the criminal proceedings that followed, Mary claimed that she was defending her house from the robbery, reacting to the noises made by the perpetrators. The prosecutor, in turn, argued that the boundaries of the necessary defence had been exceeded; Mary, the prosecutor said, had deliberately aimed at the people who posed no immediate danger to her. It is worth noting that under criminal law at the time, the argument from self-defence could not lead to unconditional acquittal. However, the court ruled in favour of Mary’s exculpation.

Let us add some detail to this simple story. Not long before the events of the case took place, several other burglaries had been reported in Mary’s neighbourhood, but no perpetrators were captured. One such situation had even taken place at Mary’s house. Once, after returning home from a late walk, Mary caught thieves red-handed on her threshold; however, they managed to escape, having heard her from the distance. Additionally, Mary’s husband, just the day before, had replaced the rubber bullets he usually used, with some live ammunition. Moreover, during the legal proceedings in question, an expert witness was called upon, to assess if weather conditions during the night of the event might have prevented Mary from assessing the aim of her shooting. And just to add to this picture, the judge was reluctant to sentence Mary, knowing it would discourage other citizens from making attempts at preventing crimes, which could increase even further the crime rate in Poland in those days.

We believe that all these factors—constituting the dynamics of the legal proceedings in Mary’s case—may be treated as reasons adopted and set forward by different actors, and then assessed by the legal authority while deciding the case. There is also something more to say about evidence in this case scenario—it becomes a function of reasons that the agents adopt. More specifically, there is a story to be told about why epistemic reasons do play a role in deciding the legal case and whether this role exceeds subordination to legal provisions only. And last but not least, there is talk about stakes involved in Mary’s case. We can distinguish at least three of these stakes: Mary’s stake—to be unconditionally acquitted, the prosecutor’s stake—to show that the boundaries of necessary defence were exceeded, and the judge’s stake—to make a justified decision, but also, not to discourage other citizens from acting in self-defence. It is worth noting that the discussion concerning stakes, in a more or less implicit way, takes place in the recent theoretical discourse on evidence, and advocating the game metaphor only makes it more vivid.Footnote 9 Addressing the question of stakes seems even more evident in the broader debate on epistemic justifications—Georgi Gardiner, for example, points out to Laudan’sFootnote 10 discussion on high costs of false acquittal for violent crime, to indicate whether probabilistic likelihood given to the evidence “should be sensitive to practical stakes involved” (Gardiner, 2019). It seems plausible to think of what they have in mind as different stakes of legal evidence-finding.

2 Between GAME and game

Let us take a step back to reflect on the metaphor of the game itself. Legal philosophers (prominent legal positivists included)Footnote 11 often depict law as one of the social games that people engage in. The metaphor of a game, e.g., of chess, typically serves as a heuristic tool for explaining some general features of law as such: its origin, its objective, underlying convention, etc. The point of deploying the game metaphor in philosophising about law rests mainly on the fact that games are played according to certain rules which are constitutive of them. This can also be seen in Black’s analysis of various meanings of ‘should’-statements (see more below on Black’s metaphor of chess playing). Games (board games, sports or games played by children) exemplify a social enterprise virtually identical to their rules. However, there is yet another sense of deploying game analogy in jurisprudence, which the present text explores. This sense has to do with a particular instance of playing a game, so as with a game understood as a particular competition, match, or occasion when people do play a game.

We believe, quite uncontroversially, that there are two separate senses of a game: the first, in which the game is identical with its rules or general practices of playing by such rules, and the second, in which the game is a particular occurrence of people playing.Footnote 12 For the sake of clarity, we will refer to the game in the first sense as ‘GAME’, while the second sense will be referred to as ‘game’. This distinction is important because, as we claim, perhaps more controversially, it is the second sense that provides us with the most interesting resources and insights into both the law itself and the role of evidence in law. The game we will ultimately focus on is analogous to a separate instance of legal proceedings, or a legal case, typically consisting of legal fact-finding and concluded by a final ruling of the decision-maker. In our view, adopting the perspective of a game helps to reveal the assumptions about the roles of different legal actors (participants to the game), the types of reasons they adopt and the relations which hold between them.

Therefore, what we additionally aim at is to suggest an account of law as the GAME in terms of reasons, including practical stakes (that we will call the stake-reasons, respectively), rather than in terms of constitutive rules only. Then, we want to account for a particular game played within the GAME, in order to better account for the GAME itself. Ultimately, these aims are motivated by our determination to salvage the game analogy in law or to make it less misleading. Appealing to reasons in this regard is supposed to give an insight into the role of epistemic reasons in law. This explanatory move seems justified when we take Littlejohn’s (2018) remark on legal consequences seriously: “plausibly legal consequences—even those regulated by the relatively low preponderance standard—rely on establishing claims. [These claims are being treated] as reasons for action”. As Gardiner (2019) points out, for this to be legitimised, some need guarantees, some demand true belief or knowledge. However, our reply concerning such a legitimisation differs significantly from the ones given by Gardiner and Littlejohn: we believe that the legitimisation derives from the dynamics of the practice itself, from the constellation of reciprocal actions that legal actors undertake, in the form of mutual commitments to and acknowledgements of specific claims.

3 Our model: the practice of law as games of reasons

Distinguishing between GAME and game concepts is not only aimed at drawing attention to the fact that the GAME necessarily involves a possibility of its instance in a form of a game (a match). Importantly, these two notions create two possible explanatory directions regarding the act of playing. One of them is based on a direction ‘from GAME to game’, while the opposite leads ‘from game to GAME’. In the legal context, this would imply two different pictures of legal practice. Explaining legal practice—also a practice of evidence-finding—from the perspective of ‘GAME’ has to draw attention to the application of legal rules. This is because, in this perspective, one begins from the set of rules which is then applied during a match. The explanatory direction of ‘from game to GAME’ is more concerned with how a match contributes to the GAME, involving future matches that are yet to be played. We think that there are several benefits that the legal theory based on ‘from game to GAME’ explanatory direction can gain over its alternative account. To illustrate why the ‘game to GAME’ is a more inviting direction to adopt while explaining the legal practice, we turn to an example of a theory that in our view—quite directly and by appealing to the same metaphor—applies in its explanatory scheme the ‘from GAME to game’ direction, which we think Andrei Marmor’s conventionalism is an interesting example of.

By presenting a descriptive model of a singular (particular) game of law, we want to (to use our own nomenclature) change the subject of this explanatory game, for the GAME of law itself can—in our opinion—be characterised as a totality of such games.Footnote 13 In the presented view, the course of the single game of law influences other, future games that are yet to (or might) be played. This is because the reasons adopted or presented throughout legal proceedings can be referred to during these other, future games. In this sense, a single game can influence the content of the GAME of law itself. An adequate presentation of the former can therefore lead to a dynamic account of legal practice. The game we have in mind takes the form of a legal proceeding oriented towards issuing the individual practical legal conclusions, with at least two participants—a decision-making entity and a recipient of this decision. Typically, it is based on the process of legal fact-finding and applying legal rules (including e.g. adjudication and even legislation, that is applying the rules conferring powers to legislate).Footnote 14 Such cases, more or less complex, make the daily life of legal systems. Our model aims at explaining what, next to legal rules under consideration (that we treat as legal reasons or L-reasons), would be constitutive of a coherent, reasons-based account of legal practice.

The conceptual framework that enables a coherent account of practices involving the use of reasons can be found in contemporary pragmatism. According to some pragmatists (e.g. Brandom, 1994; Rouse, 2007), normative discursive practices of exchanging reasons are aimed at philosophical explanation of the meaning of linguistic units, in terms of their socially norm-governed use. This picture of “the social route from reasoning to representing” comprises normative attitudes and corresponding statuses of commitment and entitlement to claims (paradigmatically: assertions) that are implicit in our practices of giving and asking for reasons. Therefore, basic discursive acts—assertions—serve as expressions of reasons that the practitioners adopt, hence committing to their content. Commitments result in the entitlements of other participants to verify them and demand justification (see Brandom, 1994, p. 172). In standard cases, commitments are inherited from premises, due to the character of material inference. Being committed to certain contents may exclude commitments to other contents. Public acts of speech or actions are treated as assertions by those who can assess their inferential significance in the light of their own discursive commitments. Assertions, however, can be supplemented by other acts of speech: deferral, disavowal, query, or challenge. All the above-mentioned discursive moves are normative, precisely because they produce the normative discursive statuses of those who participate in discursive practice. The assumption is that for legal, epistemic or stake reasons to be meaningful, and thereby to enter into the discursive interplay, they must have assertible, propositionally articulated content with regard to which normative statuses can be adopted. Indeed, the content of beliefs and actions is best analysed as propositional. Moreover, as Brandom rightly claims, assertions are not solitary (isolated): they relate and are related to other assertions. Their mutual relation is inferential and as such, it can be viewed as a practice of exchanging reasons. The practice of giving and asking for reasons involves recognising some reasons as correct and other reasons as incorrect, due to their relationships with their premises or conclusions. Such a practice is perceived as normative. Assessing the correctness of reasons adopted in this conceptual scheme is conducted from the perspective of a deontic scorekeeper, who tracks the normative moves undertaken by participants.Footnote 15 In the legal games, this role is played by a decision-making entity.

We will proceed by showing that a particular game of law allows us to discern three different types of reasons while explaining how analytic pragmatism can help to make sense of their interplay (3.1). It will be followed by a characterisation of game and its participants as guided by reasons (3.2). Finally, an abstract game’s dynamics will be described by pointing out some interplays between different types of reasons (3.3).

3.1 A game as an interplay between three types of reasons

Next to the reasons associated with legal rules (those we call legal reasons or L-reasons) that are central to the vast majority of general accounts of law, we want our picture of a game to include two other types of reasons, that is, epistemic (E-reasons) and stake reasons (S-reasons). All these reasons have the capacity of guiding the moves of the participants to the game; they all also have some propositional content that can function in agents’ reasoning both on intra- and interpersonal levels.

3.1.1 L-reasons

Legal reasons (L-reasons) for acting are expressed in the form of legal rules (legal norms)Footnote 16 and are related to the institutional status of an agent which, if recognised, may serve as a premise in practical legal inferences (Dybowski, 2017b, p. 61; Brandom, 2000, p. 91). L-reasons can be expressed in the form of legal rules. In Brandomian inferentialist terms, this means that the legal rules can be understood as ‘codifying practices’ of legal reasons-finding by licensing certain inferences as distinctly legal. L-reasons always come in ‘packages’ because a norm N1 implies further norms as N2, N3…, which becomes apparent in the process of interpretation or reasoning. Practical (legal) reasonings are composed of more than one L-reason, in the vast majority of cases, because some part of the rule usually determines the consequences of its application.Footnote 17

Among L-reasons, one can distinguish between reasons corresponding to the rules that grant legal competences, and the rules of substantive law. The important difference for the reason-oriented account is that while L-reasons of the second kind (the substantive ones) licence practical conclusions for their addressees—whether they draw them themselves or some entities ascribe them to those addressees—the first type of L-reasons (the competence ones) entails specific practical conclusions consisting of recognising and reacting to conventional acts performed by those who exercise their competences. Whether those practical conclusions are treated as intentional actions or as propositions is irrelevant for the present purposes, as long as we assume that intentional actions have either some propositional content or that a capacity to reason practically and a capacity to act on habits and bodily skills are joined in such actions (Levine, 2012).

It is important to underline that L-reasons are the reasons to which other types of reasons are referred and in this sense L-reasons structure the course of the legal proceedings. This may suggest that the three types of reason are intertwined, but they can be at least conceptually discerned. We assume that L-reasons must always figure in the decisions of the decision-making entity.

3.1.2 E-reasons

Epistemic reasons are discerned in relation to their role in evidence-finding. We think that evidence plays a more interesting role in the proceedings than the role it is ascribed by the rules of admission of evidence, which is why we turn to the broader category of epistemic reasons.Footnote 18 E-reasons are those reasons on which our beliefs are based, but the interest of legal theory is restricted to the role that E-reasons play in law or particular games of law, i.e. where decision-makers and other participants deal only with given circumscribed subsets of knowledge. The account of E-reasons, coherent with that of L-reasons, follows Sellars’ claim that the episodes of knowing must have some conceptual content, can therefore be modelled on linguistic performances and on their normative character (see Sellars, 1956, §36). The normativity of beliefs can be explained in Brandomian terms as the doxastic commitment which situates beliefs in an interpersonal inferential structure. This applies also to non-inferential reports which nonetheless must be articulated inferentially (Brandom, 1994, p. xv; 2000, pp. 47–49). The status of evidence is therefore ascribed to the subsets of knowledge (statements of fact) in constituting them as relevant E-reasons through the course of a game.

Brandom’s epistemological views that explain knowledge in inferentialist and normative terms also imply an account of epistemic intersubjectivity that centres around the deontic scorekeeper. Since, unlike in universal discursive practices, decision-makers hold privileged institutional positions in legal discursive practices, these institutional agents’ moves with regard to epistemic beliefs in the particular games in law can also be accounted for in terms of commitments and entitlements.

Dealing with evidence in the theory of law is typically split between the discourse of legal rules and the discourse of facts, and subsequently, it is followed by attempts at bridging the gap between normativity and facticity. This can be exemplified by a tension underlying the concept of ‘application of the law’ in the 20th-century Polish theory of law. On the one hand, the application of the law was understood as a process of fixing legal consequences of some actual states of affairs (Opałek & Wróblewski, 1969). On the other hand, the legal theorists analysed this concept focusing on conventional acts of public authorities (esp. Ziembiński, 1980). The second approach seems to us to be more accurate as it covers not only the declarative judgments (those confirming existing rights and obligations) but the constitutive ones (those creating new rights and obligations) as well. A model process of application of the law distinguishes between two stages: the legal fact-finding and the act of making a justified legal decision based on those findings. Ziembiński notes that such a division is artificial because “in the real life, (…) they are coupled in a feedback loop” (Ziembiński, 1980, p. 459). Importantly, we claim that decision-making entities know the potential legal consequences of some state of affairs and the fact-finding directed at that state of affairs has to justify certain legal decisions. It follows that between epistemic reasons (i.e. reasons used in reasoning about facts) and other types of reasons there is a relevant interplay that has to be taken into account when establishing the model of the game of law.Footnote 19

3.1.3 S-reasons

Stake reasons (S-reasons) have their origin in the existence of the stakes of the game, that is in the different possibilities to resolve the game in one way or another (Rouse, 2007, p. 5). Just like other kinds of reasons, the stakes’ propositional content can be made explicit. Presumably, it can be reasonably asked whether S-reasons are not just a subset of E-reasons. We claim that a standard E-reason is “stuck in the past” from which it is extracted in connection with L-reasons. What our beliefs are based on, in the case of E-reasons, concerns the past, accounted for ex-post through the content of applicable L-reasons. To the extent to which they concern the past, they can be abstractly verified as true or false. The distinguishing feature of S-reasons is that in the exchange of reasons within the game, they have to be preceded by considerations of L-reasons upon which participants build their highly justified beliefs that one or another decision will create a desirable or undesirable practical conclusion. Therefore S-reasons are necessarily “future-oriented”. It is worth noting that the S-reasons do not belong exclusively to the group of legal officials or the decision-making entity, as other participants of the legal proceedings (e.g. other addressees to legal norms) can have their own stakes in the game being played.Footnote 20

From the above description, it can be observed that S-reasons are various reasons distinguished by their relationship to the outcome of the proceedings. They are the reasons why one possible outcome of the proceedings may be favoured over another. S-reasons of different game participants compete in reaching an outcome desired by a particular participant.

As we have already hinted at, the S-reasons, or second-order reasons, are among the reasons for its participants to play the game, and they can lead to an interplay between different reasons for undertaking certain moves in the game. We will try to demonstrate that other models do not account for reasons interchanged by agents in a particular game, as long as those agents have reasons to play the GAME at all. Therefore, we are trying to broaden this perspective, by showing that the second-order reasons do not exhaust their role in motivating participation in the GAME only. In fact, various participants seriously interact with the reasons adopted by other participants during the course of the game. As we claim, this interplay is to be observed in particular games of law, i.e. in singular cases of the process of law’s application.Footnote 21

All three types of reasons enter into various relationships. For us, it is most interesting to observe how both E-reasons and L-reasons affect the S-reasons that determine the outcome of the legal proceedings, which we think was especially visible in Mary’s case, where the external stake to the proceedings led the court to deem applicability of other legal rules than it was initially assumed, and on the basis of epistemic reasons that were not directly provided to the court by the participants.

3.2 A game and its participants

A game itself concerns a particular legal process, within definable time limits, resulting in a legal decision. The process concludes by applying an individual and concrete legal norm. Such an individual norm may be thought of as establishing a licence for practical legal reasoning for that individual (e.g. a litigant in civil proceedings or a defendant). What is essential for the description of a game is the fact that it comprises the exchange of reasons among its participants. The participants can be specified in each particular game, depending on the character and subject matter of the legal actions under scrutiny. However, we claim that it is always possible to discern at least two participants of the game: a decision-making entity and an addressee of its decision.Footnote 22 The legal conclusion issued by the decision-maker becomes at the same time the subject matter of each game.

Despite the fact that the central type of legal proceedings is litigation, where a dispute between two or more parties is authoritatively adjudicated, we chose to set the minimum number of addressees to one. This is because in our view, for an exchange of reasons to take place, it is sufficient to have two participants involved. The proposed game metaphor is also meant to cover such legal proceedings where there could be no more than one such addressee (e.g. administrative or criminal proceedings). Further, even in civil cases, there is typically one party, the plaintiff, who initiates the litigation, and it is ultimately this party’s claims that are addressed by the final decision, whether or not it affects the other party or parties as well. Also, in criminal law, the model of a police-dominated investigation followed by the court’s adjudication must be modified to include those situations in which the investigation is run by prosecutors who can also be treated as decision-makers at some stages of the proceedings. All in all, no matter how many participants of particular proceedings there are and whatever configurations of total exchanges of reasons make up the game, there will always be at least one exchange between a legal decision-maker and an addressee of the decision.

We readily admit that the participants' moves within the game may exhibit different dimensions of reasoned acting. There is a broad spectrum of reasonableness attributable to the participants, but this very fact only shows that some reasons always come into play. At least three dimensions of the use of reasons can be identified, following Brandom’s treatment of action (see Dybowski, 2018, pp. 50–52). First, the agents might act intentionally if their acknowledgement of practical commitment can be inferred by deontic scorekeepers from the context of his or her action or from the speech acts. It is a default presumption of legal practice in genere, which allows characterising an agent's actions in terms of liability or responsibility. Second, they might act with reasons by being entitled to their practical commitments. Such commitments can be made intelligible both to the agent and to others by producing a suitable part of practical reasoning to explain what reasons one had for doing so and so (even though in particular cases one acted intentionally but without reasons). The assumption that in the course of legal proceedings, agents act with reasons corresponds with the agent’s ability to provide explanations, contest factual findings or produce statements about what happened. Finally, they might act for reasons to the extent that their acknowledgement of practical reasoning commitment is elicited by proper reasoning, and particular reasons function as causes for acting. In some situations, participants of a particular game in law assume strategies that may be contradictory. However, even if an agent has no sophisticated strategy, other participants can attribute it to him.

The model assumes universal application among criminal, civil and administrative procedures. It focuses on the interplay between the distinct reasons exchanged in its course.Footnote 23 It has been developed within the legal culture of civil law, but we believe that it could also be adapted to the games played in the common law. Of course, different kinds of legal procedures cannot be treated as the same game, even though they share some ‘family resemblance’ in some features and outcomes. By that, we do not claim that our game model assumes that there is one and only game in the law, nor that there are no significant differences between different legal proceedings or that they can be easily unified. Instead, we wanted to emphasise that all these proceedings share certain features: the participation of a decision-making legal entity; a recipient of the decision, and the orientation of the proceedings towards issuing a legal decision. A game between a plaintiff and an expert judge is certainly different from a game between a defendant and a 12-members jury (or each of the 12 jurors), but the three types of reasons may occur in each of those games. Therefore, there are some common features that all legal proceedings share, despite deep differences between them (also deriving from the specificity of a particular legal system). This intuition—that legal games from both common law and civil law legal systems belong to one family of games—derives from the fact that they all can be referred to as “legal”.

3.3 The game’s dynamics

Our idea is that the dynamic aspect of the reason-oriented description of legal proceedings manifests itself by the different roles that different types of reasons play in the proceedings and by the interplays between them. The key dynamic relation in this picture is the one between the current set of reasons referred to in the proceeding and the stake of the game (taking into account that the stake itself can change, too). The structure of such a set of reasons and their interrelations may change throughout the proceedings, as the new reasons are presented and/or the new stakes become available or vanish. The model we propose is meant to capture those changing dynamics of the reasons at different times of the proceeding.

3.3.1 Starting the game

The game is initiated by an action of an agent. The action can be motivated by any type of reason,Footnote 24 but typically the motivation has to do with S-reasonsFootnote 25 and serves a twofold purpose. First, as an input, the action contributes some E-reasons, by presenting some bits of information (e.g. that this and that happened). Second, the output of this action is such that—from the perspective of the decision-making entity—it produces an L-reason for launching the proceedings. If the input of the game is of such character that it enables further inclusion of additional reasons, the exchange of reasons proceeds.

3.3.2 A course of the game

The decision-making entity’s actions are first aimed at determining the facts that were the grounds for initiating the proceedings. Establishing pieces of knowledge as evidence involves using L-reasons (competence) by the decision-maker in order to obtain E-reasons from the participants of legal proceedings. However, the participants may use their own competence to supply those E-reasons that they deem necessary.

An initial set of collected E-reasons is generated for the moment. L-reasons (substantive) emerge on the part of the decision-maker, and they are typically implicit in the selection of pieces of evidence that are articulated in the initial set of E-reasons. Those L-reasons (substantive) are potentially constituting the subject of the game (i.e. an individual norm to be established). This prompts the processes of examination (and re-examination) of S-reasons by the participants of the game as they are envisaging what difference it would make to decide the case this way or another and thus, they are ascribing certain values to the final outcome. We have to stress that the S-reasons adopted by the participants of the game may remain implicit. It means that—from the external point of view—we may only ascribe these implicit reasons to the participants. Due to the asymmetry of the positions of the participants, a new ascription of the S-reasons is taking place: the participants are developing the views of their own S-reasons but also of the S-reasons of the decision-maker to which they are sensitive in formulating their S-reasons. This leads to the next moves, guided by L-reasons (competence), to provide the decision-maker with additional E-reasons. It might be observed that as a result of the following exchange of E-reasons, the decision-maker may still revise its assertion as to the relevant L-reasons (substantive). The dynamics of legal proceedings require the decision-maker to eventually put an end to this process.

3.3.3 Changing the subject

There is yet another feature of the game that may occur due to its dynamics, that is, changing the subject of a game by one of its participants. As we mentioned, we borrowed the idea of the “change of subject” from Max Black’s (1964) analysis of various meanings of ‘should’-statements. Black considers a hypothetical game of chess where player A asks for advice on the next move. He notes that the purpose of chess is to checkmate the opponent. However, player A may be motivated by some other out-of-the-game premise which is the precarious health of the opponent connected with e.g. the risk of a heart attack in case the opponent would suffer defeat. Therefore, intentionally losing the game, e.g. by not moving the queen in a position in which the opponent’s king is threatened with capture, “involves a change of subject” (Black, 1964, p. 175). Due to the ‘rigid’ character of chess, changing the subject is against the GAME’s nature. When it comes to the law, however, we think that the very possibility to change the subject is not only in harmony with the character of law, but also it is a distinctive feature that distinguishes law from other social practices.

In the legal context, changing the subject is connected to the relevant L-reasons in the legal proceedings, which are the L-reasons that the participants expect to be underlying the game’s outcome. Change of subject occurs when some other reasons (of any of the three types) affect the so far relevant L-reasons in such a way that different L-reasons are called upon in their place (as was the case in the example of Mary, when new evidence and stakes were introduced).

Of course, the change of subject does not occur in every individual game in law. However, in the course of the game (towards the end of the game), when the set of E-reasons has become well-established, the change of relevant and also well-defined, substantive L-reasons is a change of subject of this game.Footnote 26 If the change of subject involves the adoption of a new L-reason (substantive), such reasoning is justified by the suitable selection of E-reasons generated by the participants of the game. The relevance of this change rests upon S-reasons that were deemed by the decision-maker to weigh for one outcome of the game against another. Therefore, the final stake needs to be high enough (have significant weight) to be able to alter the result of the game. It shows how changing the subject becomes a result of an interplay (or influence) of stake reasons with the remaining reasons involved in the game.Footnote 27

3.3.4 Game exit

The game ends with issuing an individual and concrete norm as a licence for reasoning for its addressee. It is established by the decision-maker on the combination of E-reasons and L-reasons that the decision-maker deemed suitable in relation to the final S-reason.Footnote 28

4 “From GAME to game” perspective of law

As already implied, we believe that our model of playing the game of law provides reasons to advocate for the “from game to GAME” perspective of analysing law (and evidence). Since in this article our aim is to present the benefits of a ‘game to GAME’ explanatory perspective, we would like to show what may be the drawbacks of the ‘GAME to game’ perspective and why the focus on the game side of the metaphor can provide a legal theoretician with a wider range of theoretical tools to describe legal practice and why it carries broader implication as to the picture of legal practice. In our view, Andrei Marmor’s deployment of the metaphor of law as a GAME exemplifies the sense that we are interested in (that is of an explanatory tool).

Our focus on the two perspectives that the GAME/game metaphor implies is driven by the observation that drawing a distinction between GAME and game metaphors leads to two different outlooks of legal practice. They differ as to the explanatory perspective (from GAME to game or from game to GAME). Therefore, our criticism is mainly targeted at an explanatory direction: we want to explain why we can gain more in our understanding of law (and evidence) when we focus on the single instances of legal proceedings. For these reasons, the introduction of Marmor’s proposition has mainly metatheoretical purposes. By contrasting Marmor’s metaphor with our reading of GAME/game distinction, we want to underline the consequences of adopting Marmor’s explanatory perspective of the law in general, and what it may miss in comparison with the vision of law being explained from the (game) perspective of a single instance of legal proceedings.

Before others, we think that overly focusing on the GAME side of the metaphor leaves a legal theorist with less adequate tools to describe the course of legal proceedings. From the GAME perspective, the proceedings can only be described chronologically, by a narrative of past instances of the rule applications. We think that the game perspective of law, as a GAME of many games, is better suited for describing the dynamics of legal practice, how it may be influenced by external factors and what—apart from the legal rules—can change the course of legal proceedings or affect its outcome.

Of course, Marmor’s metaphor of law as a GAME of chess (see Marmor, 2006) aims at providing a background narrative for setting forth his proposition about conventional foundations of law. We are therefore aware that his application of the metaphor is different from the one presented in our text. The following discussion of Marmor’s conception is not aimed at criticising Marmor’s usage of the metaphor for his purposes. Instead, we claim that his reading of the metaphor produces an inadequate picture of the legal practice. The ‘from GAME to game’ perspective in Marmor’s account reveals itself when Marmor points out that the game of chess is governed by rules which “prescribe permissible and impermissible moves in the game” (Marmor, 2006, p. 350). He claims that “the rules of chess have a dual function: they constitute what the game is, and they prescribe norms that players ought to follow” (Marmor, 2006, p. 350 and Searle, 1969, pp. 33–34).

Our main reservation is that Marmor’s view does not inform how the course of the game might be influenced by the reasons that different subjects have for participating in it, and how these reasons may affect the game’s outcome (given the assumption that the moves of the participants are exhausted by following applicable rules). To use the metaphor of law as a GAME, this objection refers to the claim that a pragmatist account of normativity (legal normativity included) should involve both the issue (what constitutes the practice) and the stake, understood as the competing outcomes of the game (the difference it would make to resolve the issue one way rather than another).Footnote 29 This distinction is used in practice theory and favours the “game of games” approach, according to which law—as a game with its constitutive rules—comprises a bundle of particular games which produce particular individual results for those who take part in it, but are also relevant for future games yet to be played. ‘The stake’ in question could be thus understood also as a ‘favourable result of the game’ which could be assessed from the perspective of different participants. Marmor’s account lacks this perspective.

4.1 Marmor’s approach: truth-relative

There are several other problems that Marmor’s view entails. The first problem is epistemic in nature and concerns the position of truth-relative considerations in law. What this view presupposes is therefore a specific understanding of the assertive content of an utterance, described in Marmor (2011) as the theory of legal exhortative. The context of a legal GAME, Marmor claims, creates a prefix for statements and utterances, which designates “a constitutive relation to the truth-values of the statements prefixed by them” (Marmor, 2011). He explains further that saying so makes it true, so that if in fiction it is stated that “the moon is green”, it is true that the moon is green in that fiction (Marmor, 2011). Therefore, the conversational context of the game (or of legal practice) creates the relevant truth-value standard for this practice. Marmor explicitly makes it the case, by making an analogy with prefixes of game and law. When a statement is prefixed with a context-indicator, such as: “according to [the rules of] chess it is the case that p”, it means that within the game of chess p is true. If p is true, it is because the truth of p is constituted by the rules of chess (Marmor, 2011). Transferring these assumptions to the grounds of legal practice, Marmor (2011) concludes with the following model of ascribing truth-value in law, which we take to be representative for his understanding of the GAME in law:

According to the law in S (at time t, etc.) {if X [fact] then Y [legal result]}.

(b1) E [something that happened in the world],

(b2) According to the law in S, E counts as X,

therefore, X.

(c) According to the law in S, {Y}.

The obvious concern here is that on this account (move b2) it becomes easy to incorporate to law, somewhat unconditionally, any kind of statement, if the legal consequences are to be based on purely stipulative grounds. Marmor notes this kind of objection, observing that even in fiction there are some factual constraints on the fictional world, such as common knowledge or text-based indication, but he concludes by stating that such restrictions do not apply to law and there is a need to assume that “the law incorporates by implication all the actual facts in the world” (Marmor, 2011). For Marmor, it is enough to assume that the truth-value statements are contestable but contesting in this sense regards rather the plausibility of subsumption made on the grounds of the legal syllogism, not the quality of a statement as factual.Footnote 30

There are certain problems associated with this picture. One of them is hinted at by Marmor himself when he states that “authoritative finding of fact is ambiguous between the finding that something actually happened in the world, and the finding that it conforms to the relevant legal categorization of it” (Marmor, 2011). He claims to be focusing on the latter only, however it seems that for the complete account of evidence-finding in law one needs to account for the relationship between both. Moreover, his view does not distinguish between contexts of the GAME (c) and game (b2) and therefore provides no further grounding for concluding from b2 to c.

The worry with such handling of the relationship between legal rules and facts, although corresponding to some extent with the classical formula of legal syllogism, is that treating legal rules as prefixes to the factual statements makes the relationship of prefixing the only subject of a legal dispute—at least when it comes to the factual findings. In other words, the participants of the proceedings, when disputing over the state of affairs in a legal case, dispute over whether E counts as X according to law in S (to use the wording of Marmor’s truth-value ascription model). We think, however, that the dispute regards both the question whether a fact such as E happened in the world, whether it counts or can be counted as X for the law in S and whether actually the legal outcome of Y is valid. Moreover, we think that sometimes it can be questionable whether E counts as X in a relevant legal system (as was the case in Mary’s scenario). The picture of prefixing is not as good suited as reasons-oriented picture to account for these subtleties.

Marmor’s account also seems to treat the rules as static. If we look at the (b2) move, it seems as if the relationship of ‘counting as X according to the law in S’ is set. Our goal was to underline that in some cases it “is being counted as law in S”, which means that this relationship can be only constituted throughout the legal proceedings and is not a result of an application of a previously set inference (which is what we were trying to show in the narrative about Mary).

There is also another reason why Marmor’s picture may be considered as overly simplified—as Allen (1984, p. 1086) rightfully noted, one of the problems of evidence law results from the fact that “the admission of a piece of evidence can rest upon some fact that is itself relevant to the litigation”. We think that the notion of reasons provides a broader apparatus to account for this ambiguity (e.g. as a relationship between different E-reasons) than the relationship of prefixing does.

Another line of criticism against the view under discussion is that it reproduces the deficiencies of the standard legal-theoretical approaches to the application of the law which describe legal fact-finding retrospectively, by focusing mainly on the following of evidentiary legal rules. The methodology of fact-finding in the legal practice is therefore sometimes accounted for as imitating the process of scientific cognition, albeit with limitations imposed by appropriate legal rules (see e.g. Zieliński, 1979).

Our main concern is, however, that this approach presupposes a certain outlook on what evidence in legal practice is, which is whatever the legal rules claim it be. Therefore, it marginalises the actual role of epistemic reasons in the legal fact-finding. It seems insufficient to account for the fact that a piece of knowledge is being constituted as evidence and ascribed significance when it could play a distinctive role in the reasoning of the decision-maker. When resolving a particular legal case, the decision-maker is relying on reasons for further beliefs (or actions), hence on premises from which legal conclusions are inferred. However, the worry here is not that Marmor accounts for truth in law in a deflationary way but rather that he lacks any account of how the law is actually concerned with statements about facts—or epistemic reasons—that is, the statements with relevant epistemic grounds and how such grounds depend on the actions of different participants of the game.

4.2 Marmor’s approach: rule-oriented

The central feature of the Marmor’s picture is the pivotal role ascribed to legal rules. Despite being aware of there not being an exact identity between a GAME itself and the rules by which that GAME is constituted, Marmor seems to believe that it is ultimately the rules that are constitutive of a GAME: “The rules that determine how law is created, modified, and recognized as law also partly constitute what the law in the relevant community is. They define the rules of the game, thus constituting what the game is” (Marmor, 2006, p. 358).Footnote 31

It is true as long as the GAME in mind is identified with the law as such, perhaps with the concept of law, but not so true about a particular legal system, not to mention a particular game within a given legal system. Therefore, Marmor’s account raises three problems with regard to legal rules, but only two of them are solved. The first problem is what constitutes the GAME of law, and here the pride of place is assigned to the rules. The second problem is what the GAME of law consists of, or what we prefer to call an issue, and there again the rules play a crucial role. The third problem is only raised by Marmor, but remains far from being solved: “Chess is a very complex social practice; it is an elaborate social interaction that embodies certain conceptions of winning and losing, values related to what counts as a good game and a bad one or an elegant move and a sloppy one, and so forth. Following the rules is only part of this complex interaction” (Marmor, 2006, p. 359). Even though Marmor suggests that “the relation between the rules and their emergent social practice is not one of identity” and that “(…) there is more to the practice than just following its rules.” (Marmor, 2006, p. 359), he does not offer an explanation of what it is that exceeds this rule-oriented account of the practice.

Marmor’s overly focus on legal rules creates similar worries to those implied by Ho (2018) when he advocates a turn to the internal point of view of legal fact-finding. Ho’s distinction of the internal and external point of view in legal epistemology is aimed at underlining the role of a fact-finder and its normative consequences, rather than a detached account of the accurateness and effectiveness of rule-application. However, while Ho’s account is more concerned with a moral agency of the fact-finders, our main focus is with the reasons that are decisive for the decision-making entity (also—the extra-legal ones) and that overly stress on legal rules application makes the process of legal fact-finding seem merely mechanical and independent from external factors.

We believe that the third problem, which we defined as that of a stake or, to use Marmor’s phrase, that of ‘winning and losing’, is equally important to our understanding of the law. The picture of law as a game constituted by its rules and the practice of following them is still missing important details which are provided in a distinction that Marmor is aware of, but not exploring, namely the distinction between a game, seen apart from any participants and context, and “this particular game” in which someone is involved at a certain time. The focus on the game side of the metaphor draws attention to its stakes, which indicates that playing is not only concerned with winning and losing, but also with winning and losing to some extent.

4.3 Marmor’s approach: backwards-looking

Because of the rule-oriented depiction of the GAME of law, a particular game modelled along the lines of Marmor’s picture would have to adopt a backwards-looking explanatory perspective on the moves (actions) which the participants undertake.Footnote 32

First, in Marmor’s picture, rules (conventions) are antecedent with regard to moves in the game, and nothing else than rules seems to contribute to the moves. We call this perspective ‘backwards-looking’ in reference to Veronica Rodriguez-Blanco’s (2019) distinction between two opposing models of action: description of an action and the normative characterisation of an action. The latter concentrates on the role of normative standards in guiding the agents’ actions, whilst the former lacks any such reference. Rodriguez-Blanco claims that intentional action is only made intelligible when “we (…) resort to the values or principles or good-making characteristics that the agent aims to bring about in the world” since “there is no ‘brute fact’ or ‘pure facts’ about actions and therefore actions cannot primarily be grasped by descriptors of the world either mental, physical or of a similar sort” (Rodriguez-Blanco, 2019, p. 32). As a consequence, this account of action underlines both the existence of a parallel between practical reason (or deliberative reasoning and intentional action) and the agent-oriented perspective: “intentional action is primarily from the first person or deliberative point of view and therefore it is forward-looking” (Rodriguez-Blanco, 2019, p. 31). This characteristic does not apply to the game players in Marmor’s view, because it is conventionalist in refraining from setting a link between individual and collective action, or from differentiating between two kinds of practical rationality. Although he tries to provide a normative dimension to descriptions of social practices by referring to the deep conventions that underlie them, his account is weakened by the choice to separate individual from collective rationality.

Second, the course of the game cannot be altered by the participants unless they are mistaken in applying the rules. Consequently, such account fails to capture the relation between reasons and the participants of the game (e.g. the normativity of such relation). As a result, it reduces the impact of the game dynamics, expressed in the actions chosen by the participants, in the course of the game itself, underestimating the decisive moments of the game where the choice of a reason for action—from the poll of reasons available at this moment—has been directly affected by reasons for action adopted by other participants, hence altering the course of the game. Also, the rule-oriented model is not sensitive enough to the abilities that the participants need to possess for making significant moves in the game.

Third, this account does not assume any awareness of playing by the rules of the GAME on the part of the players. While the conception of rationality presupposed by Marmor requires that people who follow social rules (conventions) in normal circumstances believe that there is a reason to follow the rules, and hence are rational, this requirement does not assume that they are actually aware of such reasons. As Rodriguez-Blanco comments: “This is puzzling since we follow the rules because of reasons, and it is therefore mysterious how this reason can remain opaque to us” (Rodriguez-Blanco, 2019, p. 36). It is even more puzzling if we bear in mind that Marmor takes reasons as facts that count in favour of an action (Marmor, 2008, p. 102), yet he refrains from linking intentional actions on reasons to the emergence of conventions. We want to underline that Marmor’s insight includes only present actions of legal actors (and their descriptions), without them bearing any consequence for future shape for the social practice, making the causal link between intentional actions and social practice only declaratory.

Fourth, only the prospective dimension of games of law and of law itself explains why those games are also about winning and losing.Footnote 33 This dimension is missing from Marmor’s picture. Moreover, a forward-looking perspective is best fit to capture specific features of legal practices, in analogy to those of a competitive GAME, but having its own aim which is nothing less than managing social coordination at a complex level. Kaluziński (2019), commenting on Marmor’s conventionalism, proposes to distinguish four elements of the deep convention of playing a competitive GAME: “(1) games are established by rules—rules govern the practice and they specify what it takes to win or lose, what moves are forbidden, permitted and necessary, who is viewed as a participant etc.; (2) games are something different from ordinary life or other types of practices such as art or religious rituals (…); (3) the whole rivalry makes sense only when the rivals are in some specific respect similar (this is why there are weight categories in combat sports); (4) that games have certain aim—obtaining favorable result” (Kaluziński, 2019, p. 10). As we argued, an alternative model should not only specify that the aim of the law is to create obligations and duties but also that the law, understood as possessing some features of a competitive game, has in view some specific favourable results (“stakes of a game”). The latter may even be more important because, as we also stated, favourable results may become second-order reasons leading to the choices as to which first-order reasons to act on. Ultimately, specific favourable results may alter the course of a game.

Drawing attention to the winning-losing aspect of a legal game remains adequate also with regard to the legal proceedings where seemingly nobody wins or loses.Footnote 34 Therefore, we do not want to suggest that “winning” is equivalent to “obtaining a decision which is defined by law as preferred by a party in scope”, and “losing” is equivalent to “obtaining a decision which is defined by law as not preferred by a party in scope”. Rather, the “winning” and “losing” aspects of the game are relative to the S-reason adopted by the party concerned.Footnote 35 Of course, usually there are the ‘winners’ and ‘losers’ of the game, but both ‘winning’ and ‘losing’ come in degrees. Sometimes winning means ‘not losing too much’ (as it might have been for Mary, who was interested in receiving as low a penalty as possible). In thinking about this distinction, it may be helpful to have in mind a card game, rather than a game of chess.

4.4 Marmor’s approach: reasons-relation neutral

The backwards-looking perspective, implicit in Marmor’s account, leads to certain conclusions as to its implications on reason-relations: (1) between agents and reasons they adopt, and (2) between reasons themselves—especially as to the reasons for entering the game and the reasons to make certain moves within the game. As we have noted, the commented approach is not attentive to the influence that the conventions of certain GAME exert on the factual behaviour of an agent. It does not follow from the requirement of rationality how it is exactly that certain patterns of behaviour do form conventions, whence the agents remain unaware of the reason (social rule) they act on. Consequently, the model of action inherent in this view, actions in law included, does not allow any interference between reasons the agents have to enter the game (second-order reasons) and the reasons that guide the behaviour within the game (first-order reasons).

A similar argument (albeit for a different purpose) was set by Stefano Bertea, who claims that the difference between the first- and second-order reasons reveals itself in the quality of duties the law imposes on its subjects, leading to a deflationary analysis of legal obligation. Bertea notes that “[t]here are two related questions which Marmor leaves unexplained, that arise out of two basic characteristics which his account ascribes to constitutive conventions and their normativity. The first of these characteristics is that constitutive conventions do not only constitute or define certain practices, but also guide the behavior of those who take part in them. (…) The second characteristic which Marmor ascribes to the normativity of constitutive conventions applies specifically to the master rule itself: this rule sets up an obligation only for persons who already have, from the outset, a reason to participate in the practice of law (…).” (Bertea, 2011, p. 87). In other words, the external second-order reasons are considerations that originate outside the practice and are independent of its conventional dimension. Therefore, only the internal, first-order reasons remain intra-institutional demands, bearing significance as to the content of the social rules. For Marmor, the agents who adopt second-order reasons have first-order reasons (which guide their behaviour) under the presupposed constitutive convention. Hence the role of second-order reasons is exhausted at the very beginning of the practice. The consequence of such move was recognised by Bertea in stating that “the legal ‘ought’ arising under the master rule does not differ in kind from the duties which we incur by taking part in other more circumscribed practices, such as games, which we typically take part in upon a voluntary basis, which makes the duties so incurred conditional to our accepting to play the game as willing participants” (Bertea, 2011, p. 87).

We want to suggest here that it might be possible for second-order reasons (from the outside of the practice) to influence the choice of reasons that the agents decide to act on within the game (first-order reasons), including the choice of the set of E-reasons. The possibility to change the subject within the course of a particular game also ensures that participants recognize the reasons which guide their behaviour. Consequently, this compliments the assumption that within the legal practice the agents act intentionally and their actions can affect the course of a game understood as an individual case of legal proceedings.

This is not to say that the existence of a stake in the game necessarily presupposes that the extra-legal considerations participate in constituting what the game is. Rather, it means that the stake of the game plays a role in favouring certain moves within the game. To use Kaluzinski’s distinction of the types of rules within the GAME, the stakes affect how the peripheral rules of the game are followed (Kaluziński, 2019, p. 1170). Kaluziński divides constitutive rules of the GAME into two categories: “central constitutive rules” and “peripheral constitutive rules”. The rules of the first type form the GAME’s hard core and their violation “is something ‘outside the game’; when we violate the central constitutive rule of a certain game, we risk the termination of the game in question” (Kaluziński, 2019, p. 1170). The violation of the peripheral rules, on the other hand, “is considered as a fault, an ‘illegal move’ within the game” (Kaluziński, 2019, p. 1171). Therefore, if the participant of the legal game changes the subject, it could be considered as a faulty move, but one that remains within the game itself.

These considerations were aimed to show that the model where law and legal evidence are seen from the perspective of the GAME may form an incomplete picture of what legal practice (or law as such) consist of. Therefore, it is a perspective of a game that provides a more dynamic analysis of epistemic propositions in law, being concentrated on reasons for action available for participants of the game, while remaining sensitive to the ‘future-oriented’ considerations as to the possible outcomes of this game.

5 The return of Mary

Let us now return to the initial case of Mary. We will now translate the story of the legal proceedings that were initiated against her into a game of law, and that game into the language of the relevant interchange of reasons in that case. The proceedings in Mary’s case were initiated by a prosecutor, in relation to the relevant L-reasons (due to the prosecutor’s competence), creating an obligation for a court to examine the case. In the beginning, the court is in the process of using L-reasons (exercising the judge’s competence) to hear Mary—the defendant—and the witnesses, trying to establish the set of E-reasons that would serve as material for its further decision. The E-reason delivered by Mary to the court was her claim that she only fired at the burglars because she was defending her house from the robbery, reacting to the noises made by the perpetrators. When other robbers were interrogated, they claimed that their intention was not to rob the house, but the part of the electricity track—establishing thus another E-reason for the court. The substantive L-reason that the court is adopting at this moment is the norm that forbids murder, even murder committed involuntarily, thus creating the subject of the game and its possible result that is a prison sentence for the defendant.

We could observe that in reaction to adopting the L-reason by the court, the set of S-reasons is being shaped: (1) the argumentation of the prosecutor is that the crime was committed and that the boundaries of the necessary defence were exceeded; Mary, the prosecutor said, was fully aware of the fact that she was aiming at the people sitting on the electricity track; S-reason, in this case, being the protection of the victim’s family and the virtue of general prevention; (2) According to Mary, her action was aimed at defending her property, so Mary’s S-reason was future life in the prison or in liberty. The court adopts yet another S-reason: sentencing the defendant would discourage other citizens from making attempts at preventing crimes, increasing even further the crime rate. It has already been mentioned that the argument from self-defence, under the criminal statute from that time, could not lead to an unconditional acquittal.

As a reaction, the court uses its competence (competence L-reason) to appoint an expert witness. The expert creates another E-reason for the court that states that the conditions during the night of the event could prevent Mary from seeing the object of her shoots. Due to the S-reason it adopted, the court decides to change the subject of the game (this instance of legal proceedings): it adopts as its new L-reason a norm that obligates the court to drop the charges, when prohibited action is committed due to the mistake as to its relevant features. It states that the defendant was not aware as to whom she was shooting so she was acting under a relevant mistake.

The final ruling of the court was the exculpation of Mary. The result of the court’s assessment of the evidence was therefore to ascribe the defendant a particular reasoning against Mary’s actual claims. The initial L-reason (substantive) which was the abstract and general norm sentencing to several years of prison for involuntary murder—has been changed for the norm licensing lack of liability. The cause for the change, in the reasoning of the court, was a certain S-reason of promoting active reactions for the prohibited acts.

6 Conclusions

As we have argued, if one takes seriously the metaphor of law as a GAME, one needs an adequate account of what it is to play a particular game of law: a particular case of legal proceedings, finalised with issuing an individual and concrete legal norm. We think that the contrary explanatory perspective—that is a perspective of the „from GAME to game” models—is in fact rule-oriented, backwards-looking, and reason-neutral, and hence it struggles with providing an account of the dynamics of a game. Also, such views marginalise the role of evidence within legal fact-finding, which is no longer at stake, so to speak—the evidence loses its factual dimension (connection to the world of facts), becoming determined by the legal rules only. This description ignores the role of epistemic reasons in shaping the practice, influencing the choice of reasons for action throughout the course of a game which in turn makes the model lose the perspective of a stake of a game itself—the prospective element that has an influence on the reasons adopted in the game.

As a result, these accounts become unable to explain how it is that legal rules have a significant place in the course of the reasoning of an acting agent. For example, Marmor’s model provides only a retrospective account of the moves undertaken by the participants of the legal proceedings. Furthermore, it means that this approach does not allow any interference between reasons the agents have to enter the game (second-order reasons) and the reasons that guide the behaviour within the game (first-order reasons). We have tried to establish a model that would be forward-looking, reasons-oriented and pragmatist in nature, that would address the above reservations. However, this project was not aimed at refusing the legal rules a chief role in the legal game, but rather at a further explanation of the role of rules within the game of law and how they interact with other reasons available to agents.

The presented model reflects reasoning conducted by an official entity –- in Brandom’s terms, a deontic scorekeeper—which is provided with reasons to decide the case by other players of the legal practice (also the ones who are not players in this particular game). To use the terminology proposed in our model, the GAME of law cannot be described as an exchange (game) of L- and E-reasons only. Therefore, there is more to the law than its rules because every game of law involves reactions to its possible outcomes, so as to the S-reasons emerging in a particular game. This conclusion might be inferred from observing the process of evidence-finding that is taking place in the individual game of applying an individual and concrete legal norm as a licence for legal reasoning for its addressee. The eventual picture of legal game that emerges from such approach is dynamic—by providing themselves with different reasons for action, the participants constantly change the structure of the current set of reasons for action involved in these proceedings, as the new reasons are presented and/or the new stakes become available or vanish. Therefore, the model we propose is meant to capture the changing dynamics and interactions between the reasons at different times of the proceedings.

Simultaneously, the above conclusion decides on the singular character of reasons in the legal practice, in which the significance of each reason (undertaken by participants) is being established in the process of reasoning of the decision-making entity. This happens within the course of the exchange of reasons between participants of the legal proceedings. This singularity may be observed in the course of the game, in the situation of the change of subject that is taking place in relation to the relevant reasons. The very possibility of altering the result of evidence-finding upon consideration of the stakes of the game (as presented in the narrative) shows that the L-reasons (and so the legal rules) are insufficient as to determining the outcome of the process of issuing an individual and concrete legal norm. Therefore, the views that adopt the approach ‘from GAME to game’—that equate legal rules with the legal practice itself—miss the perspective of the impact, that the game dynamics may have on its legal outcome.

Our stress on the fact that legal rules do not exhaust the practice of law, since there are other, extra-legal conditions that influence that practice, can be also understood as a variation of Schwyzer’s idea that what distinguishes the game of chess from the rite of chess is the “grammar” of a given practice, and that the chess is a game is not a rule of this game, but a grammatical proposition (Schwyzer, 1969, p. 463; after Lorini, 2012). Therefore, what makes chess-playing an instance of the game of chess is all sorts of things, which in a logical sense are relevant or appropriate to say with regard to chess (Schwyzer, 1969, p. 454; after Lorini, 2012). We wanted to describe that outlook advocating the reasons and conditions of autonomous discursive practice, that the law can be understood as a part of.Footnote 36 The model assumes that all participants act on reasons, ascribing and acquiring normative attitudes as to the reasons they choose to act on, and hence that their intentional action requires conscience as to the reasons adopted for acting. Moreover, the existence of S-reasons in a game (that may or may not be second-order reasons) that motivate agents to choose one first-order reason for action over another, ensures a possibility of an interaction between first- and second-order reasons in a game of law.

Our final and most important point concerning the evidence is shedding a new light on the role of evidence within the legal proceedings thought of as a singular game of law. To relate to the title of this article—our aim of setting evidence “back at the stake of the game” was achieved by both ascribing to evidence a prime place in creating relevant epistemic reasons for the decision-making entity, and showing the possible role of epistemic reasons in changing the subject of a game, hence affecting the outcome of interplays that the other reasons may engage in.