For some sufficiently long-standing institutions, such as the English Crown, there is no single thread, whether specified in terms of constitutive rules or assigned functions, that would connect the stages of that institution. Elizabeth II and Egbert are not connected by an unbroken chain of primogeniture and they have importantly different powers and functions. Derek Parfit famously sought to illuminate his account of personal identity by comparing a person to a club. If Parfit could use our intuitions about clubs to help motivate his neo-Lockean account of personal identity over time, which resists the idea that personal identity requires a common psychological thread, then I argue that an adapted version of his account of identity might, in turn, be reapplied to clubs and other institutions, such as the Crown.
The English Crown is a circuitous institution that arguably finds its origins in Egbert’s regime (802–839), the first king to rule over the whole of Anglo-Saxon England.  What makes the reigns of Egbert and Elizabeth II stages of the same institution? Even a cursory understanding of English royal history might move one to simply reject this question: their reigns are not stages of the same institution because there does not seem to be a common enough thread, specified either in terms of constitutive rules or functions and their associated deontological powers, that would strongly connect them. And yet we understand the reigns of Egbert and Elizabeth II to bookend the same institution, the Crown. I argue that Derek Parfit’s account of personal identity can be adapted to provide us with an account of institutional identity capable of accommodating the assumed continuity of the Crown.
What relation would ensure that Egbert and Elizabeth II are stages of the Crown? We should more precisely specify the relations and relatia at play in this question. Marriage and monarchy are types of institutions.  The English monarchy is an instantiation of the institution of monarchy. If we follow Francesco Guala’s distinction between institutional types and tokens, so that marriage is a type and Florentine marriage is a token (2016, p. xx, 196–197), then strictly speaking the English monarchy is a token institution. A particular English monarch, such as Elizabeth II or Egbert, is a temporal part or a stage of the token institution, English monarchy, just as a particular Florentine marriage is a part of Florentine marriage.
Our question concerns the criterial identity relation: in virtue of what are Egbert and Elizabeth II stages of the same token institution, the English monarchy (rather than being stages of different token institutions of monarchy)? To answer this question, we have to get clearer about the grounding relation(s) that would metaphysically explain how, e.g. Elizabeth II can be a stage of the Crown. Following Brian Epstein, a fundamental fact grounds a less fundamental fact if the former “metaphysically makes” the latter (2015a, p. 69–72), as clay might be among the grounds of a statue. For institutional facts, the grounds are specified by a “frame principle” (2015a, p. 77); if the grounds are the actual ingredients, a frame principle is the recipe.  In the case of Elizabeth II, the grounds are specified by the principle of primogeniture. But what of Egbert or Henry IV, neither of whom satisfy that rule? Different stages of the Crown appear grounded in importantly different ways, and we need to formulate an identity relation that can accommodate these differences. 
2 Different Grounds at Different Stages
According to John Searle, institutional statuses have three core features: constitutive rules, deontological powers, and functions. Different theorists have appealed to each of these features in answer to the question about what, in general, grounds instances of institutional statuses.
Constitutive rules, according to Searle, have the form “X counts as Y in context C” (Searle 1995, p. 28). Epstein thinks that the grounding relation is what Searle is “trying to capture” with the notion of a constitutive rule (2015a, p. 75). If this were the case, then it would be vacuous to cite constitutive rules in answer to the question of what the grounding relation is for institutional facts. However, as I am using the term, but in accordance with Searle’s principle examples, the notion of a constitutive rule gets at, not the grounding relation in general, but a kind of grounding relation: constitutive rules are frame principles anchored in enactment. Thus, while all social facts have grounds specified by frame principles, only the grounds of highly-structured, legal or quasi-legal institutional facts, such as U.S. dollar or U.S. president are specified by constitutive rules. For Searle, institutional kinds are defined in terms of these constitutive rules (1969, p. 34): a sufficient condition for a person (X) becoming a U.S. President (Y), is that that person is a natural born citizen and wins the electoral college vote (C). In February of 1952, Elizabeth, the heir apparent (X), became Elizabeth II or Queen of England (Y), upon the death of her father, King George VI (C), in accordance with a version of the constitutive rule of primogeniture.
The fact that every institutional status is associated with a function (or a cluster of functions) suggests an additional candidate for what might ground instances of a status.  For example, the function of the present English monarch is largely symbolic and advisory. Guala, in Understanding Institutions (2016), describes “the broadly functionalist conception of institutions that runs through the book” (2016, p. 78). According to this view, institutional kinds are defined in terms of function – in terms of instances being able to solve a class of coordination and cooperation problems. This is in contrast with Searle’s claim that institutional kinds are individuated by constitutive rules.  In Section 4.1, I build on Muhammad Ali Khalidi’s distinction between conventional social kinds and real social kinds to claim that Searle and Guala are not articulating competing accounts of what individuates social kinds, but different ways of being a social kind.
Searle and Guala are principally concerned with answering the question of what it is for something to be an instance of one institutional kind rather than another. However, because an answer to the question of what it is for an instance of that kind to continue to exist over time depends on and is constrained by an answer to the question regarding that thing’s nature,  reasonable inferences can be drawn about what their answers to the identity question would be. For Searle, because the extension of a status term is determined by whatever constitutive rules are associated with the status, it is reasonable to assume that two monarchs at t1 and t2 are both stages of the Crown if they satisfy the constitutive rule of primogeniture and they are part of a sequence of monarchs with those characteristics going back to the monarch which initiated the Crown.  However, the constitutive rule of primogeniture cannot explain why Egbert and Elizabeth II are stages of the Crown. Elizabeth II is not related to Egbert by way of primogeniture. And some English sovereigns, such as Cnut the Great, William the Conqueror, Henry IV, and Egbert himself came to occupy the position, not by satisfying the rule of primogeniture, but by conquest. Moving onto Guala’s proposal, it is also clear that Egbert and Elizabeth II have different functions or solve very different coordination and cooperation problems.  Or, if there is a modest overlap in functions, there are other offices, such as the Prime Minister or the Defence Secretary, that are more functionally equivalent to the office held by Egbert.  Thus, neither functions nor constitutive rules alone would seem to make sense of how Egbert and Elizabeth II are stages of the same institutional token, the Crown.
Perhaps Guala would concede the point and contend that this demonstrates that we should reject the question, what makes the reigns of Egbert and Elizabeth II stages of the same institution? They are not temporal parts of the same institution because their offices solve importantly different problems; there is nothing to explain. This response is consistent with Guala’s claim that “characteristic properties of institutional kinds ought to be discovered, just as in the natural realm” (Guala 2016, p. 175).
Of course, certain theoretically-motivated uses of the concept, institution, can be illuminating. There is merit in classifying social phenomena from the external point of view, as Guala and the structural-functionalists might. However, this is not the stance that Searle would have us adopt when he claims that “[w]e are interested primarily in the internal point of view” (Searle 1995, p. 98). And we think that Egbert and Elizabeth II are stages of the Crown, despite the fact that they are not strongly connected by constitutive rules or functions.
In this paper I develop an analogy that Parfit draws between the survival of persons and the survival of clubs. In both cases, Parfit suggests, mere continuity, rather than strong connectedness, is sufficient for survival. A person can survive radical changes in psychological makeup and a club can survive changes in membership, governing rules, and functions if, in either case, there is continuity. Such an explanation would also explain why we think today’s U.S. Democratic Party is the same party as the 1794 anti-Federalist Democratic-Republican Party (to which it traces its origins), why we think of today’s Freemasons as the same group as the Freemasons of the 14th century, despite the fact that the former is a quasi-religious organization rather than a local trade organization, and why Egbert and Elizabeth II are stages of the Crown. In each of these cases it is unlikely that the relevant stages share much by way of constitutive rules or functions.
My plan is as follows. In Section 3, I review Parfit’s account of personal identity. What matters for survival is not the relation of personal identity as such, but one of its conditions, psychological continuity or “relation R.” I also review Parfit’s claim that, in some cases, the identity of two persons might be indeterminate. In Section 4, I apply a version of relation R to the question of institutional identity over time. In particular, I articulate parallel notions of institutional continuity, strong institutional connectedness, and institutional indeterminacy. In Section 5, I return to the question of what makes the reigns of Egbert and Elizabeth II stages of the Crown.
3 Parfit on Personal Identity and Relation R
Parfit makes several important points about the nature of personal identity in Reasons and Persons. First, he claims that psychological continuity, rather than mere connectedness, is the central feature of personal identity. Second, he argues that this notion of psychological continuity, rather than the containing notion of personal identity, is what matters to us. Finally, he claims that, under certain conditions, the question as to whether two people are the same person is empty. I address each of these points in turn.
3.1 Psychological Continuity
Parfit claims that the identity of two people, X at t1 and Y at t2, consists in psychological continuity – overlapping chains of strong psychological connectedness (with the right kind of cause).
Since the key relation of continuity is defined in terms of strong connectedness, I begin with a discussion of the latter relation. Psychological connectedness starts from a thought Parfit attributes to Locke,  that shared memories are what make a person X and a person Y the same person.  However, Parfit expands the notion of psychological connectedness to include other psychological connections such as when Y realizes the intentions had by X or when X and Y have the same beliefs, desires, and projects (Parfit 1987, p. 205). While memories are important, the other psychological connections can be sufficient for strong psychological connectedness: “We shall then claim, what Locke denied, that a person continues to exist even if he suffers from complete amnesia” (Parfit 1987, p. 208).
Psychological continuity, rather than mere connectedness, is Parfit’s criterion of personal identity. After all, while I am strongly psychologically connected to the person who was writing this essay yesterday, I probably do not share enough memories, beliefs, desires, intentions, etc. to be strongly psychologically connected to my 6-year-old self. And yet I am identical with that person.
3.2 Relation R
Having formulated a criterion of personal identity in terms of the relation of psychological continuity, he goes on to claim that it is continuity – or what he calls relation R – and not identity that matters to us. According to Parfit, we are erroneously committed to construing our survival as a matter of identity because the terms in which we ordinarily think about survival are non-reductionistic. And while a revised, reductionist theory of personal identity, cast in terms of psychological continuity, is possible, Parfit still thinks it would be a mistake to construe our survival in terms of the identity relation. This is because the formal properties of numerical identity, such as symmetry and transitivity, misguide us when assessing fission-type examples. I might be strongly psychologically connected to the two people I branch into by way of fission.  However, because the two resulting people cannot be identical, I might begin to worry about which one is really me.  Parfit argues that if such worries are arbitrary (Parfit 1995, p. 215) then this demonstrates that the identity relation is not what matters in survival. What matters is the identity relation’s key ingredient – psychological continuity.
3.3 Empty Identity Questions
There is another way the stress we place on the identity relation misguides us. Where personal identity is an all-or-nothing relation, the relations of psychological connectedness and continuity are “matters of degree” (Parfit 1995, p. 29, 20). And because identity is an all-or-nothing relation, there will always be a cluster of cases around the demarcation point where the identity question strikes us as difficult to answer. Since relation R is what matters on Parfit’s view, in cases like these, the question of whether X and Y are the same person has no answer (we can, however, stipulate an answer).
Here is one such case: psychological continuity assumes some overlap between chains of strong connectedness. Imagine, then, two persons, X and Y, that would be strongly psychologically connected but for a temporal gap – perhaps X came to suffer and then overcame an acute form of dementia. Is X the same person as Y? Because survival comes in degrees, Parfit urges us to resist the temptation to answer this question. Or, at least, he urges us to recognize that an answer to an empty question is more like a decision than a discovery. Along these lines, Parfit famously compares the identity of a person to that of a club.
Consider, for example, clubs. Suppose that a certain club exists for several years, holding regular meetings. The meetings then cease. Some years later, some of the members of this club form a club with the same name, and the same rules. We ask: “Have these people reconvened the very same club? Or have they merely started up another club, which is exactly similar?” … There would then be no answer to our question. The claim “This is the same club” would be neither true nor false. Though there is no answer to our question, there may be nothing that we do not know.
This is why we would not be puzzled when we cannot answer the question, “Is this the very same club?” We would not be puzzled because, even without answering this question, we can know everything about what happened. If this is true of some question, I call this question empty.
… When an empty question has no answer, we can decide to give it an answer. (Parfit 1987, p. 213–214)
Just because relation R tracks what we really care about does not mean that we should avoid talk of personal identity (or, for that matter, the identity of clubs). In normal cases – cases that do not evoke fission or do not rest near the borderline of the identity relation’s point of demarcation – the judgments that result from appeals to identity neatly coincide with appeals to relation R (because relation R is an ingredient of Parfit’s identity theory).  We can continue to appeal to the identity relation so long as we remain vigilant in those contexts where the formal requirements of the identity relation begin to mislead.
In the next section, I claim that just as Parfit uses the club to illuminate his account of personal identity, his account of personal identity can be reformulated into an account of institutional identity. That account can help us better understand the identity criteria, not just of the club, but of the Crown.
4 Relation C
If persons X and Y satisfy relation R, then they are the same person even if X and Y do not share much by way of memories, intentions, etc. Since relation R is specified in terms of psychological continuity, we need to formulate a relatively analogous relation tuned to address the question of institutional identity. I will call this “relation C” (“club”), or institutional continuity. Relation C should help us understand how X and Y can be stages or temporal parts of the same institutional token, despite differences in the constitutive rules and functions associated with each of these stages.
In what follows I focus on groups, such as a club or (perhaps) the Crown, as a representative subset of institutional or quasi-institutional phenomena.  While the continuity relation is the key innovation in a Parfitian notion of identity, since continuity is defined in terms of overlapping chains of strong connectedness, I begin with a discussion of how the relation of strong psychological connectedness might be transposed into the institutional sphere (4.1) before returning to the core notion of institutional continuity (4.2). I conclude this section with a discussion of institutional indeterminacy and of the significance of identity declarations (4.3).
4.1 Strong Institutional Connectedness
Two persons are strongly psychologically connected if they share enough desires, beliefs, memories, intentions, etc. Clubs, and institutions in general, typically have nested functions or purposes, which are akin to the distinctive intentions or projects persons have. There are social clubs and service clubs, for example.
A maximally connected club is one in which its purposes, deontological powers, and constitutive rules are tightly aligned over a period of time. Such a club has a mission and formal and informal norms that govern the behavior of members and authorized agents in ways that conduce to that mission. Even with an ideal distribution of deontic powers, those purposes may not be realized if roles or offices are not filled with appropriately qualified or skilled people. Thus, in the maximally connected club, the constitutive rules are formulated so as to select the best candidates for the various roles that make up the organization.  A club that requires its president to have been a long-standing member is more likely to realize its purposes than a club that requires its president to, e.g. swim 100 meters underwater (unless, perhaps, the club is a swimming club).
However, seamless alignment between constitutive rules, deontological powers, and functions is too stringent a condition for strong institutional connectedness. People can be strongly psychologically connected even when they have conflicting beliefs and intentions; Philosopher Kings are not the only people who satisfy relation R. Likewise, a club can be strongly connected even if it has unclear missions, distributes deontological powers in such a way that partially undermines the realization of those missions, or has constitutive rules which do not select ideal candidates for key roles. Moreover, some groups, such as a group of street musicians can be constituted in a way that makes no reference to constitutive rules.
Searle and Guala articulate two modes of strong institutional connectedness, each of which corresponds to the categories of real social kinds and conventional social kinds, as broadly described by Khalidi (2015) and elaborated on by Joshua Rust (2017).  The social kind, war, could either be a conventional social kind or a real social kind depending on how the term “war” is used: the Korean conflict is an instance of the real social kind, war, but is (or at least was) not an instance of the conventional social kind, war. A state of affairs is an instance of the conventional kind, war, if certain antecedently-specified, constitutive rules are satisfied (e.g. it is declared as such by the requisite body). Two countries can be at war, in this sense, even if no fighting has occurred. A state of affairs is an instance of the real social kind, war, in so far as participants aim to realize certain ends through deontologically-bound, open-ended, sponsored conflicts.  War, in this sense, is possible without having been declared.
The distinction between conventional and real social kinds implies two modes of strong connectedness, whereby X at t1 and Y at t2 qualify as stages of the same institutional token. Starting with conventional social kinds, just as the death of the sovereign precipitates a new stage of the Crown according to the constitutive rule of primogeniture, there are constitutive rules for creating and maintaining faculty committees. Following Epstein, two collections of faculty members, X and Y (at t1 and t2, respectively), are stages of the same faculty committee if each person in collection X and collection Y has gone through the legislated rotations and processes for its members, in accordance with antecedently specified constitutive rules, and X and Y are parts of an unbroken sequence of stages with those characteristics going back to this committee’s origin (Epstein 2017, p. 17–18).
The possibility of real social kinds, which are defined in terms of the capacity to realize a function, implies a second mode of strong institutional connectedness. An informally organized group of street musicians is an instance of a real social kind, as its creation and continuance only requires a deontologically-bound commitment to perform a function – to play music together; there are no antecedently-specified, enacted constitutive rules that need to be satisfied in order that a group of people counts as a group of street musicians. Two collections of people, X and Y (at t1 and t2, respectively) intentionally performing music with one another are strongly institutionally connected if each person in X and Y is part of an unbroken sequence of stages with those characteristics going back to the origin of the group (Epstein 2017, p. 17). Such functionally-defined social kinds are relatively stable because they help individuals do things they could not do alone – they solve cooperation and coordination problems (Guala 2016, p. xxii).
Each of these two modes  of strong institutional connectedness can fall short of maximal institutional connectedness. On the one hand, instances of real social kinds, such as a group of street musicians, cannot, unlike instances of conventional social kinds, survive certain changes in function or extended periods of inactivity (Epstein 2017, p. 11, 20). On the other hand, conventional social kinds are such that, following Robert Merton, the “unchallenged insistence upon punctilious adherence to formalized procedures” can lead to self-perpetuating, zombie institutions that are divorced from the cooperation and coordination problems they were created to solve (Merton 1968, p. 253). 
4.2 Institutional Continuity
As discussed, the notion of continuity rather than connectedness is the key feature of Parfit’s account of personal identity. Psychological continuity consists in overlapping chains of strong psychological connectedness. The image evoked is that of relatively stable periods of strong connectedness (the chains) punctuated by measured change (the overlap between chains), such that distal and extremely different chains might still be indirectly linked (they are continuous). Having described two modes of strong institutional connectedness, I will now formulate a corresponding notion of institutional continuity.
Elizabeth II and her father, King George VI, are strongly connected stages of the Crown in virtue of the fact that Elizabeth II satisfies the constitutive rule of primogeniture. Sometimes institutional survival is also compatible with changes in the relevant constitutive rules. The Act of Settlement 1701 limited succession to Protestants, resulting in the ascension of George I, despite his having been a member of the most junior line of the House of Stuart. Arguably, George I is strongly connected to previous stages of the Crown, not in virtue of satisfying the same constitutive rule of succession, but in virtue of those different rules satisfying the same secondary rules of change (Hart 2012). Valid modifications of relevant constitutive rules are one source of institutional change. However, as this paper is concerned with institutional continuity, and not just connectedness, it is necessary to see if institutional survival is compatible with even more profound breaks between stages.
Imagine that a collection of people (s0) meet socially over lunch for several years, beginning at t0 and are jointly committed to socializing. This group is an instance of a real social kind. Then, at t1, this collection of people (s1) accepts conditions, specified by constitutive rules, by which membership and leadership roles are determined. At t2 the membership conditions are expanded so as to allow for the induction of additional members. The collection of people (s3) at t3 trace the origin of the club to t0, despite the fact that those lunch-goers had not formulated or committed to membership conditions.  As illustrated in Figure 1, this group is a single, continuous group, bookended by different, overlapping periods of strong institutional connectedness; the group transitions from a real social kind to a conventional social kind. Despite not being strongly institutionally connected, s0 and s3 are stages of the same club in virtue of their being institutionally continuous.
Since each of the stages of the social club before and after s1 perform the same function it might be claimed that the latter stages are also instances of real social kinds. However, I claim that these latter stages are instances of conventional social kinds because the constitutive rules that characterize those stages would allow the club to survive changes in function or extended periods of inactivity. Clubs and corporations might exist on paper only, borders can be ignored, sovereigns can be in exile, and faculty committees can be inert. Following Epstein, instances of conventional social kinds (the faculty committee) “exist in perpetuity or until they are expressly disbanded,” whereas instances of real social kinds either “continue to exist only while an activity continues without break” (the street musicians) or can only survive brief periods of inactivity (the informal social club that meets at lunch) (Epstein 2017, p. 20).
Just as a group can survive the transition from a real social kind to a conventional social kind, a group can also survive the transition from a conventional social kind to a real social kind. Imagine that at t0 a faculty committee (X) is explicitly created with the task of advancing diversity and inclusiveness in pedagogy. The committee is formally dissolved by the provost at t1. At t2, immediately following the dissolution, those who served on the committee continue to meet informally (Y) with the intention of advancing the original committee’s charge. Under certain circumstances, it would seem that groups X and Y are institutionally continuous, and so the same group, despite the fact that Y is not a stage of an instance of a conventional social kind. Especially if those in Y trace the group’s origins to X by way of a declaration.
In the next section I discuss the force and significance of such identity declarations, which are among the factors that can bind and so render continuous two overlapping chains of strong institutional connectedness.
4.3 Institutional Indeterminacy and Identity Declarations
Relation R, unlike the relation of personal identity, does not imply that our continued existence must be all-or-nothing (Parfit 1987, p. 236; 1995, p. 29). Accordingly, there will be cases near the identity relation’s point of demarcation that strike us as puzzling. For these cases, the identity question is empty. “It is not true,” Parfit writes, “that our identity is always determinate” (Parfit 1987, p. 217). Similar remarks apply to instances of social kinds. The two clubs that have identical missions, rules, and members, but are separated by an extended period of inactivity, are among the borderline cases where the identity question is empty (Parfit 1987, p. 243).
Like Parfit’s clubs, no independent Polish state existed between 1795 and 1918, having been partitioned by Austria-Hungary, Germany, and Russia. A question might arise as to whether the state that existed after 1918 was really the same state that existed prior to 1795. It is true that in 1918 relevant officials declared that the state was continuous with the 1795 state, just as the membership of club Y might declare that their club is the same as X. However, if the question is empty, we need to say more about how and under what conditions such a declaration can answer an empty identity question. Here is what Parfit says:
When an empty question has no answer, we can decide to give it an answer. We could decide to call the later club the same as the original club. Or we could decide to call it another club, that is exactly similar. This is not a decision between different views about what really happened. Before making our decision, we already knew what happened. We are merely choosing one of two different descriptions of the very same course of events. (Parfit 1995, p. 214)
Two aspects of this provocative quotation require more careful treatment. First, Parfit misleads us when he says we can “answer” an empty question. Strictly speaking, empty questions can only be corrected or otherwise responded to, not answered (van Fraassen 1980, p. 138, 140); if the identity question is answerable, it is not empty. This terminological point allows me to articulate a second point. An empty identity question can be rendered answerable in two importantly different ways – either by refining the identity criteria and so reformulating the question or by adding to the facts about “what really happened.” In the first case, when all the events relevant to the identity question are known, we can still decide that club Y is (or is not) the same as club X. In so deciding we are not answering the empty identity question so much as simultaneously rejecting and revising (correcting) the old identity question and answering the new question. This happens when, for example, a sociologist or philosopher declares or decides that Y=X. In the second case, I claim that there are scenarios when the response to what was an empty identity question turns out to be among the events relevant to its answerability, so transforming it into an answerable question and answering it. This can happen when, e.g. members of club Y declare that Y=X. While Parfit does not explicitly distinguish this second way of responding to an empty identity question, he indirectly acknowledges its possibility.  In what follows I explore both of these possible responses to an empty identity question in detail.
Following his discussion of the clubs separated by a period of inactivity, Parfit returns to borderline cases of personal identity, wherein “physical and psychological connections hold only to reduced degrees” between persons X and Y (Parfit 1987, p. 213). Here, he makes the noteworthy decision to make X the one posing the identity question: “‘Am I about to die? Will the resulting person be me?’ On the Reductionist View, in some cases there would be no answer. My question would be empty” (Parfit 1987, p. 214).
Parfit’s model of indeterminacy, in this example, appears to be this: following Richard Boyd, in some cases an individual a’s status as an instance of species S can be indeterminate. Perhaps a can only breed with members of S under peculiar conditions or the offspring are infertile. Thus, for Boyd, the question of whether a is an instance of S is empty (Boyd 1991, p. 141–142). Of course, a biologist could decide that a is an S by refining classificatory criteria.  But, in this case, the original empty question is not answered, but rejected, revised, and then answered. Nothing changes about this example if the classifier happened to be a itself, rather than an external observer. Likewise, for the person, who asks, of herself, “Am I about to die?” In this case the facts relevant to the answerability of the question are not changed if the person posing the question asks it of herself. If she does answer the question, she can only do so by refining the identity criteria presupposed by the question. And this is to reject and refine the old question and answer the new one – the first way to respond to an empty identity question, as described above.
The problem with applying Boyd’s model of indeterminacy to social kinds is that it does not account for the difference between indifferent kinds and reflexive kinds (Hacking 1999; Guala 2016, p. 219–223, 132–137).  Instances of indifferent kinds are not affected by our classifications of them, where instances of reflexive kinds, including interactive kinds, are. In general, species are indifferent kinds (if they are kinds at all).  However, institutions are instances of reflexive kinds because their existence is partially dependent on how participants think and talk about them.
The possibility of institutional reflexive kinds introduces an alternative way to respond to an empty identity question. In the case of the clubs X and Y separated by a period of inactivity, “Is X and Y the same club?” is empty. However, in some cases, the response to the question by members of club Y, as when members declare that Y=X, should be added to the events relevant to the determination of the question’s answerability. Without changing the identity criteria (and so changing the question), a declaration that Y=X by Y’s members can, under certain conditions, be enough to transform an empty identity question into an answerable (and answered) question. This is just to say that it can matter that Y’s members think that Y is the same club as X.
In this way, the 1918 declaration aimed at restoring Polish independence by the Polish Regency Council, made with the support of most of Poland’s political parties and the Allied Powers, is relevant to the question of whether that Polish state is the same as the 1795 Polish state. However, the mere fact that members of Y declare that X=Y does not guarantee that the identity question will become answerable. A similar declaration of identity in 1917 by a German puppet regime was not sufficient to reestablish the Polish state.
Recall the case of the faculty committee (X), an instance of a conventional social kind, that became an informal advocacy group (Y), an instance of a real social kind. We are now in a better position to see if and how X and Y could be the same group. While, unlike the club case, there is no temporal gap between X and Y, there are enough differences in terms of X’s and Y’s respective deontological powers and constitutive rules to raise the identity question. In this case, the fact that members of Y declare themselves to be the same group as X is among the facts relevant to both the answerability of and the answer to that question. Such declarations are among the mechanisms that can bind overlapping chains of conventional social kind instances (X) and real social kind instances (Y).
5 Conclusion: The War of the Roses
The proposed account of relation C is compatible with Epstein’s claim that group-stages, X and Y, of a certain kind are the same group when they are “part of an unbroken sequence of stages stretching back to the origin” of the group (Epstein 2017, p. 17). What relation C does is spell out the notion of an “unbroken sequence” that so binds these stages as parts of the same group: X and Y stand in an unbroken sequence if they are institutionally continuous. According to this view, an institutional token’s identity is path-dependent (Hull 1992) in the sense that it can be individuated through time, not in virtue of its intrinsic properties (e.g. functions), but because of a continuous historical relationship to previous states of itself. Since continuity consists in overlapping chains of strong connectedness, and since there are (at least) two distinct modes of strong connectedness, marked by the distinction between real social kinds and conventional social kinds, then the notion of continuity helps us understand how, for example, a social club can, at one stage, be an instance of a real social kind and, at a later stage, be an instance of a conventional social kind. Similarly, the faculty committee that becomes an informal advocacy group begins as an instance of a conventional social kind and becomes an instance of a real social kind. In this section I extend this account to the Crown. In particular, I claim that Egbert and Elizabeth II bookend many overlapping and alternating chains or stages of real and conventional social kind instances.
Sometimes institutional survival is possible through a valid modification of relevant constitutive rules in accordance with effectively accepted secondary rules of change. One example, discussed above, is the Act of Settlement 1701, which restricted succession to Protestant lines. The notion of institutional continuity, however, can tolerate a more profound exception to the primogeniture rule. Despite the fact that Edmund de Mortimer was heir presumptive, Henry Bolingbroke famously deposed Richard II by force in 1399. On Shakespeare’s telling, while Richard II had the right to rule in virtue of being Edward III’s grandson, Bolingbroke had a ruler’s temperament. Perhaps, Bolingbroke became Henry IV the moment he detained Richard II at Flint Castle.
Shakespeare is not incorrect to observe that, in general, sovereignty imperfectly correlates with ability, particularly when the eldest son is young or infirm. And these are among the conditions whereby dynasties collapse and sovereignty reverts to an instance of a real social kind – where the capacity to perform the functions of a sovereign, rather than the satisfaction of enacted constitutive rules, grounds the status. We have seen how a social club (X), an instance of a real social kind, might become a more structured group (Y), an instance of a conventional social kind, and still be the same club. With respect to the Crown we find not just two but multiple, alternating, and partially overlapping chains of conventional and real social kind instances, so that Richard II (X), Henry IV (Y), and Henry V (Z) are stages of the Crown, despite the fact that Y is an instance of a real social kind and X and Z are instances of conventional social kinds.
As illustrated in Figure 2, Richard II’s status as a stage of the Crown is grounded in the rule of primogeniture and Henry IV’s status is grounded, at least initially, not in the satisfaction of constitutive rules, but in an actual ability to perform the functions associated with sovereignty. But Henry V succeeded Henry IV in the conventional way. Thus, Richard II and Henry V – and, by extension, Egbert and Elizabeth II – are institutionally continuous but not strongly connected, by way of alternating and overlapping chains of real and conventional social kind instances.
The description of the Crown as consisting in “alternating” stages or instances of real and conventional social kinds should not be taken to imply that, for example, those, like Henry V or Elizabeth II, who became sovereign on conventional grounds did not also perform their function, or those, like Henry IV, who became sovereign in virtue of performing the functions associated with the Crown did not also seek conventional legitimation. Indeed, it is the co-presence of both these aspects of sovereignty that constitute the “overlap” between chains, when the chains are instances of different social kinds.
If both Henry IV and Henry V fulfilled the functions associated with sovereignty, why is only Henry V characterized as an instance of a conventional social kind? Because, counterfactually, had Henry V not been able to perform the functions associated with the office, he would not thereby cease to be King of England. Had, for instance, Henry V been driven from the country due to a foreign invasion, he would have been a king in exile.
Correspondingly, instances of a real social kind, such as Henry IV, typically seek to further legitimate their status by satisfying whatever constitutive rules they can. While primogeniture is the predominant constitutive rule relevant to coronation, it is not the only such rule. In particular, Henry IV enjoyed a declaration of sovereignty by the assembly of lords and commons (whose independence was questionable). As we have seen in Section 4.3, a perspicuous declaration of identity by interested parties is relevant to the answerability of and answer to an identity question. However, identity declarations are just one of many mechanisms by which overlapping chains of strong connectedness are bound together and rendered continuous. For example, in a ceremony that conspicuously mirrored a long-standing French coronation Ordo, Henry IV had himself anointed with oil supposedly gifted to Thomas Becket by the Virgin Mary.  It also mattered that Henry Bolingbroke could be found within the line of succession and that Edmund de Mortimer, who was first in line, was (falsely) cast as a deserter. However, it is also clear that the satisfaction of these rules represents, at best, a first and insufficient step in the Crown’s transition from a real social kind instance back to a conventional social kind instance. In this way, there is a degree of overlap between chains of real and conventional social kind instances. Had Henry IV not maintained a grip on key brute powers and had his son Henry V not been so popular, he could have gone the way of Louis the Lion, whose year-long rule over England in 1216 is not counted as a stage in the Crown’s history, despite Louis’s having been publicly proclaimed as such at St. Paul’s Cathedral.
I have argued that the Crown is a relatively unbroken and continuous sequence of stages stretching back from Elizabeth II to Egbert. This sequence of stages involves alternating periods of conquest and convention, so that there are different and alternating, but partially overlapping, grounds by which different persons at different times qualify as stages of the Crown. Indeed, this warp and woof of real and conventional social kinds seem characteristic of many long-standing institutions. For example, as the function of mutual support and companionship has become more central to the institution of marriage, the constitutive rule which barred same-sex couples from the status has struck many as increasingly arbitrary. A declaration is required for a skirmish to count as a war until it is not. And counterfeit bills could, under certain circumstances, qualify as money depending on whether they are in fact serving as a medium of exchange. The reason for this productive interplay between real, functionally-described social kinds and conventional social kinds is that real social kinds, unlike conventional social kinds, are ephemeral, going out of existence as soon as they lose the capacity to solve a cooperation or coordination problem. A group of the kind, street musicians, only exists when they are in a position to play music and even the most powerful rulers are not immortal. However, instances of conventional kinds incline toward fecklessness; they can become unmoored from the cooperation and coordination problems that motivated their introduction. These deleterious tendencies are mitigated if an institutional token alternates between these two modalities. That the same institutional token is often describable in terms of both real and conventional social kinds may contribute to its overall relevance, efficacy, and stability. 
Boyd, Richard (1991): “Realism, Anti-Foundationalism and the Enthusiasm for Natural Kinds.” In: Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition 61. No. 1/2, p. 127–48.10.1007/BF00385837 Search in Google Scholar
Epstein, Brian (2015a): The Ant Trap: Rebuilding the Foundations of the Social Sciences. 1 edition. Oxford, UK; New York, NY: Oxford University Press. Search in Google Scholar
Epstein, Brian (2015b): “A Framework for Social Ontology.” In: Philosophy of the Social Sciences 46, No. 2, p. 147–67. Search in Google Scholar
Fein, Helen (1993): Genocide: A Sociological Perspective. New York: SAGE Publications. Search in Google Scholar
Guala, Francesco (2016): Understanding Institutions: The Science and Philosophy of Living Together. 1 edition. Princeton, NJ: Princeton University Press. Search in Google Scholar
Hacking, Ian (1999): The Social Construction of What? Cambridge, Mass: Harvard University Press. Search in Google Scholar
Hart, H. L. A. (2012): The Concept of Law. 3rd edition. Oxford: Oxford University Press. Search in Google Scholar
Hauswald, Rico (2018): “Institution Types and Institution Tokens: An Unproblematic Distinction?” In: The Philosophy of the Social Sciences, November. https://doi.org/10.1177/0048393118799054. Search in Google Scholar
Hull, David (1992): “Individual.” In: E. Fox Keller and E. A. Lloyd (Eds.): Keywords in Evolutionarily Biology, 181–187. Cambridge, MA: Harvard University Press. Search in Google Scholar
Levene, Mark (2008): Genocide in the Age of the Nation State: Volume 1: The Meaning of Genocide. London; New York: I.B. Tauris. Search in Google Scholar
Ludwig, Kirk (2017): From Plural to Institutional Agency: Collective Action II. 1 edition. New York, NY: Oxford University Press. Search in Google Scholar
Merton, Robert King (1968): Social Theory and Social Structure. New York: Free Press. Search in Google Scholar
Parfit, Derek (1987): Reasons and Persons. Oxford: Oxford University Press. Search in Google Scholar
Parfit, Derek (1995): “The Unimportance of Identity.” In: H. Harris (Ed.): Identity, 13–45. Oxford: Oxford University Press. Search in Google Scholar
Rust, Joshua (2017): “On the Relation between Institutional Statuses and Technical Artifacts: A Proposed Taxonomy of Social Kinds.” In: International Journal of Philosophical Studies, 25. No. 5, p. 704–22.10.1080/09672559.2017.1381139 Search in Google Scholar
Searle, John R. (1969): Speech Acts. Cambridge: Cambridge University Press. Search in Google Scholar
Searle, John R. (1995): The Construction of Social Reality. New York: The Free Press. Search in Google Scholar
Searle, John R. (2010): Making the Social World: The Structure of Human Civilization. Oxford UK; New York: Oxford University Press. Search in Google Scholar
“To Have and to Hold” (2015): The Economist, April. https://www.economist.com/news/special-report/21648171-far-declining-family-firms-will-remain-important-feature-global-capitalism. Search in Google Scholar
van Fraassen, Bas C. (1980): The Scientific Image. Oxford, New York: Clarendon Press; Oxford University Press. Search in Google Scholar
Weber, Max (1978): Economy and Society; An Outline of Interpretive Sociology. Berkeley: University of California Press. Search in Google Scholar
Weinberg, Shelley (2016): Consciousness in Locke. Oxford: Oxford University Press. Search in Google Scholar
©2019, Joshua Rust, published by De Gruyter, Berlin/Boston
This work is licensed under the Creative Commons Attribution 4.0 Public License.