A. Idealizing away constraints on reflection

The examples of controversial idealization just given all concern features of the world that relate to the task of instantiating the conclusions of some piece of moral or political philosophy. Yet the philosophy itself is also part of the world — it is a construction of thought, constructed out of a particular intellectual inheritance, constructed by a thinker with finite powers of cognition and reflection. No contributors to current debates would deny this, yet the point receives surprisingly little attention. ^{Footnote 7} This is unfortunate, as capacities for reflection, and the form reflection takes, are also aspects of the world, and theorizing may idealize both.

To help vindicate this thought, and to make it less obscure, I would like to briefly illustrate how it applies to a position claiming the label of “nonideal theory.” Liam Murphy’s Moral Demands in Nonideal Theory addresses a moral question that arises when some agents do not comply with moral obligations. Because such a world is less than ideal, Murphy describes his position as an exercise in nonideal theory. The problem he is concerned with is that there is something unfair about beneficence under such conditions: when some do not do their bit to help those in need, the virtuous must do even more. Suppose (though I do not agree) this is a problem. ^{Footnote 8} Murphy’s solution is to insist that each is only obligated to do as much as would be required if everyone fully discharged their obligations of beneficence. ^{Footnote 9} Yet this solution involves a thoroughly idealized way of thinking.

Consider what is intrinsic to ethical reflection as Murphy understands it. It is intrinsic to Murphy’s reflection that moral standards involve a metric for calculating the amount of good one must do. To answer the question “How much does beneficence require of me?” Murphy would have the reflective subject calculate what would be required to bring about the state of affairs in which the global demands of beneficence are satisfied, and then divide this sum fairly among all natural and artificial agents, to determine each’s fair share. Murphy acknowledges that a precise calculation is probably beyond us. But as long as individuals make “good-faith” attempts to perform the calculation using “rough guesses” supplemented by “informed estimates of experts,” the principle is sound as a “background standard against which agents assess their standing motivations and rules of thumb.” ^{Footnote 10} I am not so sure. It strikes me that Murphy must make significant idealizations about reflection itself before his calculation is even intelligible.

One cannot ask “How much does beneficence require me to do?” as if the metric of beneficence could be calculated in isolation from other moral considerations. If I am not acting well, I am not acting beneficently, and so to think about what it is to act beneficently I must think about what it is to act well generally. ^{Footnote 11} Furthermore, the claims of need and assistance that the beneficent agent responds to typically have incredibly complex causes in which an array of injustices are implicated. ^{Footnote 12} Many quotidian areas of moral life will be shaped in subtle ways by the effects of such injustice: the personal is, after all, the political. This adds considerable complexity to what is involved in acting well.

With these points in view, Murphy’s conception of beneficence would have to involve a calculation of how much one needs to do to act well generally. Here, the idealized nature of Murphy’s thinking is revealed. A calculation of how much one needs to do to act well generally could only make sense if the limits to the set of possibilities for acting well could be given. Murphy’s thinking must presuppose a standpoint from which these limits are known; he admits we are not good at making the calculation, but insists it is in principle possible. As human powers of cognition and reflection are finite, however, the possibilities for human interaction — and thus the possibilities for acting well — must be understood to be indefinite. ^{Footnote 13} The limits of the set of possibilities for acting well are impossible to cognize, and it is only when one idealizes away our finite capacities for reflection that Murphy’s moral mathematics becomes something that could intelligibly apply to human interaction. If this is right, we cannot know what Murphy’s proposed principle of beneficence could mean. The reflective standpoint Murphy presupposes is not one a reflective subject could take up, and so is no reflective standpoint at all.

My aim here is only to use Murphy’s position to illustrate the point that even if a certain feature of the world, such as noncompliance, is not idealized away, one’s theorizing about how to account for that feature may be thoroughly idealized. A defender of Murphy’s position would no doubt wish to contest the objection I have briefly raised against it. Such a debate would be about whether the idealization Murphy makes use of is justifiable. If we limit philosophical scrutiny of idealization to features of the world to which the conclusions of theory are to be applied, that debate cannot even register. That would be unfortunate.

B. The form of reflection is of central importance

I now turn to a second reason for expanding the scope of the debates over idealization and ideal theory to include the form of reflection itself. Philosophical thought is, after all, inherently reflexive: to reflect philosophically about something is at the same time to reflect on the nature of philosophical reflection itself. Because assumptions about the form of reflection color answers to the question of which other features of the world can be justifiably idealized away, it is essential to give philosophical scrutiny to the form of reflection itself.

Again, it may be helpful to illustrate the point with an example. Consider Onora O’Neill’s claim that practical reason is necessarily abstract, but cannot support idealization. ^{Footnote 14} Idealization, as noted earlier, falsifies some feature of the world. By contrast, abstraction does not deny the existence of predicates, but brackets them: “[r]easoning that abstracts from a predicate makes claims that do not depend on that predicate holding, or on its not holding.” ^{Footnote 15} O’Neill’s conceptual distinction is clear. A statement that brackets some feature of the world is not the same as a statement that falsifies that feature of the world.

Nevertheless, Alan Hamlin and Zofia Stemplowska have recently found O’Neill’s distinction “murky in practice,” while Holly Lawford-Smith declares it “flimsy at best.” ^{Footnote 16} The problem they find is this: if an idealized model of justice/right action does not deliver false results, there is no straightforward sense in which the idealization problematically falsifies. ^{Footnote 17} As Lawford-Smith claims, “it doesn’t matter if the [theory’s] assumptions are literally false.” As long as such idealizations remove barriers to truth, and allow for the possibility of reintroducing real-world complexity at some later point of deliberation, the theoretical model can deliver conclusions that are truly normative. ^{Footnote 18}

This claim that falsification is unproblematic strikes me as the standard defense of idealization in modern moral and political philosophy. Reflection on justice, or on any other ethical concept, is understood as aspiring to accurately capture what that concept is, by way of a theoretical model of right action (and, in the case of social justice, of the institutional structures that mediate right action). It is then argued that theoretical modeling of justice and other values can make use of simplifying idealizations to better illuminate what those values are, so idealization is not problematic if the theoretical model does not deliver false results. Thus Stemplowska, in an important related paper, insists that unless simplifying idealizations are employed, reflection on justice will not be able to get a grip on its object: “the complexity of the problem will be simply too great.” ^{Footnote 19}

Yet note how both Lawford-Smith and Stemplowska take for granted a focus on results delivered by a theoretical model of what justice requires. I find it striking that both explicitly appeal to the use of idealization in scientific theorizing to vindicate points about idealization in philosophical theorizing about justice. ^{Footnote 20} Neither critic places significance on the fact that O’Neill’s is a claim about practical reason. Instead, Lawford-Smith and Stemplowska assume that the same form of reflection is appropriate to scientific theorizing and to reflection on how we should live. If this assumption is accepted, practical reason can be nothing more than reasoning on the topic of practice. I think it is only in light of this assumption that Stemplowska and other defenders of idealization can offer the defense that idealization aids clarity. Moral and political theories should, just as scientific theories do, employ idealizations to screen out (what my scientist friends typically refer to as) the distracting and irrelevant noise arising from the complexity of the real world.

My impression is that this assumption about the shared form of scientific and normative theorizing is widely held in modern moral and political philosophy. ^{Footnote 21} It may (I am honestly unsure) be the orthodox view. But I am sure it was not always the orthodox view. And so it can be challenged by the ancient distinction, sustained in Aristotelian and Kantian thought, between the forms of practical and theoretical reason. But, again, this challenge cannot even register unless we expand the scope of debates around idealization and ideal theory to include the form of reflection itself.

A. Practical reason is such as to produce its object

A man is putting items into a basket in accordance with a shopping list, while a detective is following the man about recording what he is placing into the basket. Anscombe draws our attention to a difference between the two:

This example is not, contrary to some interpretations, meant to illustrate the distinction between the mental states of desire and belief. It illustrates a distinction between two forms of reason.

If the detective’s record does not match what the shopper purchases, then the detective will have made a mistake because the detective’s reasoning is such as to arrive at an accurate representation of how the world is. If the shopper’s purchases do not match his shopping list, then he too will have made a mistake in virtue of the form of his reasoning. Yet the shopper’s mistake would be of a very different sort, and in grasping this we start to grasp the distinction between two forms of reasoning. The shopper’s mistake would lie “in the man’s performance,” and this is because the shopper is engaged in an exercise of reason that guides the action of the reflective subject. Should he buy margarine instead of butter, the shopper’s action will have failed to instantiate the standards deployed in his reasoning. So here we have a distinction between two forms of reasoning. Practical reasoning is not such as to represent what is independently the case. Practical reasoning, unlike theoretical reasoning, is such as to produce that which it is reasoning about.

The key to understanding the distinction between the forms of practical and theoretical reason lies in grasping the idea that practical reason is such as to produce its object. In the application of a concept to an object in an exercise of theoretical judgment, the object of the judgment is already given to the reflective subject independently of the movement of thought constituting the judgment. A reflective subject either subsumes a given particular under a given universal, or seeks the appropriate universal for a given particular. So, for example, observing someone shopping I might inquire, “Was that butter she purchased?” or (if, say, I were a visiting anthropologist set down in an unfamiliar culture), “What is she doing?” In either sort of case, the exercise of theoretical reason is such as to determine what the object of the reasoning is, but obviously is not such as to be capable of producing the object of the reasoning.

In contrast, the object of a practical judgment is not independently given to the reflective subject. The object of a practical judgment is what should be done, and this is not a feature of the world that is ready to hand. When, for example, a subject judges that the bribe should be declined, the subject understands that it is through this very judgment that she is to decline the bribe. Her judgment is such as to produce what should be done. So here we have reasoning and knowledge that cannot be theoretical in form, for if the reasoning and knowledge is such as to produce its object, its object cannot be something given independently to the subject.

Thus far, my remarks may suggest that the object of practical reasoning is really the doing of what should be done, where what should be done is ready to hand. The reader should banish any such notion from his or her mind! The object of a practical judgment is what should be done. The act instantiating that “what” cannot be antecedently ready to hand, for it is something yet to be performed. So in deploying concepts in practical judgment, the reflective subject is not applying those concepts to a particular act that is given independently to the subject. A shopper, for example, will be deploying a complex constellation of standards. Perhaps she is aiming to get the week’s shopping and also to get value for money and avoid supporting exploitative companies and not disappoint her wife and be home by 6 pm and so on. Each standard by itself is not the object of the practical judgment. Rather, the object of the practical judgment is the unity of all of these standards, and it is only through an exercise of practical judgment that the standards are unified. In this way, the object of practical judgment is what should be done. ^{Footnote 24}

A practical judgment, then, is a specification of action through the unification of a plurality of standards the reflective subject identifies with (insofar as unity is possible — hard cases can, of course, exist). These standards are standards of reason. The subject’s knowledge that this should be done is delivered by a practical judgment combining a plurality of reason-giving standards, and so the subject’s knowledge that this should be done includes knowledge of the force of the reasons for doing this. ^{Footnote 25} In this way, practical judgment is such as to determine its object. I use the locution “ . . . is such as to . . . ” to signal that it is always possible the subject will not do what should be done. A subject who judges that the bribe should be declined may still take the bribe through weakness of will, just as Anscombe’s shopper might absentmindedly buy margarine. This possibility represents a mistake in virtue of the very form of the subject’s reasoning.

Because the object of theoretical reason is given independently to the subject, whereas practical reason is such as to produce its object, these are two clearly distinct forms of reason. Yet it is easy to think that the distinction is in fact blurred. One might, for example, point to reasoning that is characteristic of medical practice. Medical science is clearly theoretical, and yet (one might say) is such as to produce good health, which is its object. ^{Footnote 26} So the distinction in form that seemed clear in the abstract now appears blurred in reality.

The apparent blurriness of the distinction in form between practical and theoretical reason dissipates once one grasps that the two forms of reason do not operate in isolation from one another. To continue with the example of medicine, it would be a mistake to characterize so complex a human undertaking as an instance of either theoretical or practical reasoning. The practice of medicine displays both forms of reasoning. There is an audible difference between the judgments “Does this cure cancer?” and “How should we respond to the presence of cancer in this patient?” In the former case, the object of the judgment is independent of the reflective subject, and so the reasoning cannot be such as to produce its object. In the latter case, the object of the judgment is what should be done, understood as the unity (insofar as possible) in action of a complex constellation of standards. The surgeon should, for example, act both ethically and effectively. And in understanding that this is what should be done, the surgeon understands she is to act accordingly.

A surgeon’s practical judgment “How should we respond to the presence of cancer?” will of course make use of the deliverances of theoretical judgments about, for example, effective treatment, just as Anscombe’s shopper needs to know where the supermarket is. If an agent is to successfully instantiate in her action the standards that guide it, she needs to know a good deal about how the world relevantly is. But these pieces of theoretical knowledge are not the objects of practical judgments, and so the distinction in form is not blurred. Nor do these pieces of theoretical knowledge feature in practical judgments in quite the same way as they do in theoretical judgments. For example, a theoretical judgment that a patient has a bad reaction to anesthetics would simply identify the relevant fact. In the practical judgment “How should we respond to the presence of cancer in this patient?” by contrast, such a fact is relevant because it is an obstacle. Practical reasoning involves an understanding of the relevant features of how the world is precisely because those features are relevant to shaping action that instantiates the standards deployed in that reasoning. ^{Footnote 27}

With the distinction in form between practical and theoretical reason clarified, I would like to signal one implication of it. That the form of practical reason is such as to produce its object entails that practical reasoning is an exercise of the productive faculty, the will. Any exercise of the will must be constrained by the conditions of its own productive possibility; it could not otherwise be productive, and so mere wishful thinking is not an exercise of the will. ^{Footnote 28} So it must also be intrinsic to practical judgment that it acknowledges the conditions of the will’s productive possibility. In Section III, I attach considerable significance to this point.

B. The scope of practical reason

The assumption noted in Section I.B, that reflection on standards of justice and morality is fundamentally theoretical, may seem untroubled by the distinction in form that I have just drawn attention to. A familiar line of thought holds that reflection on the nature and justification of the standards one deploys in practical reason is outside the scope of practical reason: reflection on the nature and justification of those standards is an exercise of theoretical reason. ^{Footnote 29}

Further reflection on the standards of practical reason can, however, undermine this familiar view of the scope of practical reason. Standards of practical reason admit of different types. Following Sebastian Rödl, ^{Footnote 30} some ends are finite, insofar as an action instantiating such an end brings about the completion of that end — for example, the typing of this sentence. Other ends are by contrast infinite, insofar as an action instantiating such an end manifests one of an indefinite number of possibilities for exemplifying the particular way of acting prescribed by that end. Think, for example, of acting justly or dressing stylishly or keeping fit. Having typed this sentence, I can in the future only type other sentences; having acted justly, I can at any point in my future act justly again. Infinite ends do not come to a limit in any individuated temporally finite actions. Rödl’s distinction helpfully marks out how some standards of practical reason possess much greater generality than do others. But as the label “infinite” has the potential to distract from this point, I shall refer to such standards as general standards.

Within the class of general standards, a further distinction can be drawn. Some general standards are extremely unremarkable: consider typing in English. Other general standards are more interesting, in that they are instantiated in the shape of one’s life. This point unfortunately requires yet more jargon. Reflection on what I should do, in the sense of a temporally finite action, can at the same time be reflection on what I should do in the broader sense of my ongoing activity. By “activity” I mean a way of going on that is continuous — temporally finite only insofar as it might be stopped (by, for example, incapacity or rejection), not in that it comes to completion — and that is determined by the active subject’s reflection. So, for example, the life cycle of a mayfly, although continuous, would not be an instance of activity in my jargon. A human living a meaningful life, by contrast, would be the broadest possible example of an activity. ^{Footnote 31}

The general standards instantiated in the shape of one’s life provide action guidance in a complex way. General standards provide guidance with respect to individuated, temporally finite actions, where practical judgment shapes the action to be done in light of the various standards to be instantiated in that action. General standards also guide activity, embodying the central ordering concepts that both constitute an agent’s understanding of what is worthwhile, and serve to connect and cohere the various standards that are particular to the various domains, roles, or facets of an agent’s life.

A practical reasoner must possess some general orienting standards. It is certainly possible to imagine someone who could reason about quotidian tasks (for example, cooking), but who had lost the central ordering concepts that make their overarching activity intelligible. Such a fate has befallen many colonized peoples for periods of their history. ^{Footnote 32} Yet such a person would be capable of practical reasoning only to a severely attenuated degree. Having lost the central ordering concepts that make their activity intelligible, such a person could only understand quotidian tasks as prolonging life, but not what life amounted to and so why it should be prolonged. If that is correct, it is intrinsic to (nondefective) practical reasoning that the reflective subject possesses general orienting standards.

Finally, the standards that orient activity are subjective, in the peculiarly Kantian sense of belonging to the reflective subject. They are standards I myself, the subject, have made my own by adopting them as maxims, and in doing so committing myself to applying the standards, when appropriate, in practical judgment. ^{Footnote 33} Or — synonymously — it is to adopt the reason-giving concept (justice, kindness, honesty, and so on) embedded in the standard as a central ordering concept that orients my activity: a concept I am to live by. ^{Footnote 34} I cannot reasonably do this unless I take the standards I am adopting as maxims to pass a test of reflective scrutiny. And so deliberation about which general standards should orient my activity just is the task of determining which standards are philosophically justified.

To accept the justification for some standard is to adopt the relevant principle as a maxim that orients my activity. In making a standard my own, the deployment of that concept with respect to my activity is such as to produce the activity in question (in judging that I should live justly, I understand that it is through this judgment that I am to be just). And so it appears — doesn’t it? — that reflection on the nature and justification of the standards of practical reason is itself an exercise of practical reasoning: an exercise of a thoroughly reflexive aspect of practical reason, but which, for all that, is still reasoning about what one should do and is doing.

A. Against the priority of theoretical reason

One ground for holding that reflection on the nature and justification of standards of justice and morality is an exercise of theoretical reason, is to suppose that a (theoretical) account of what those standards are must be in place before any (practical) account of how we should act is given. In modern political philosophy, perhaps the most famous assertion of this claim is G. A. Cohen’s. Cohen, if I understand him, insists that we must first inquire into what the correct principles of justice are before we can consider the question of how justice should guide action. Cohen writes:

To borrow Cohen’s own misappropriation of Justinian, the idea is that justice is each getting their due, which can be determined independently of any considerations of how — or even whether it is possible — that each might get what they are due. ^{Footnote 36} This is to think of the concept of justice as, ultimately, denoting a state of affairs. This strikes me as a very popular thought: Stemplowska, for instance, asserts that normative theory must enable us “to judge what we have already achieved against a final landmark of where we ought to be.” ^{Footnote 37}

Cohen elsewhere asserts the priority of theoretical reflection by saying that a reflective thinker might care about what justice is without caring at all about how standards of justice should guide practice. ^{Footnote 38} This has invited the skeptical response from David Miller, that it is hard to understand why anyone would care about what justice is if they did not care about how we should act or the structure of the social institutions through which we interact. ^{Footnote 39} I think a stronger variant of that skeptical response should be offered: it seems impossible to understand what justice is in isolation from understanding how standards of justice guide action.

Consider again Cohen’s gloss of justice as what each is due. Perhaps this is a failure of imagination on my part, but I cannot understand what “due” could mean if it does not signal that something must be done; a pattern of action produced. To say that something must be done is to identify a constraint, in the form of an obligation, on the exercise of the will of one or more agents. To think about what someone is due is to think about how others are required to act. It is to locate the possessor of the entitlement within a particular sort of practical nexus with one or more others: a nexus constituted by the obligations to which the entitlement is correlative. I say practical nexus because what is joined in this connection is the activity of the different agents in question. Accordingly, unless one can account for both the identities of obligation-bearers and the content of their obligations, thought about entitlement is empty: such thought would be unable to locate the entitlement holder within any practical nexus of the correct sort. ^{Footnote 40}

So there can be no account of what justice is isolated from an account of how agents ought to act. It is worth asking what would be needed for an account of the latter. It might be thought that all one needs is a sufficiently substantive specification of what ought to be done — which a theoretical model could provide. But if that is all that is required, the account will have left agency out of the picture altogether. Such an account would treat the doers of what ought to be done as functional components of a process, and in this respect they could be replaced by complexly programmed robots with no difference made to the theory. To think of others in this way is not to think of them as acting at all. Acting involves the reflective exercise of the will. An account of how we should act must recognize that the enacting of what ought to be enacted involves exercises of the productive faculties of those who enact it. So any such account must provide a vindicatory explanation of how the standards of justice can be willed by those who are to enact them.

Going further, and connecting the point to Section II.B. above, to will a principle is to make it one’s own — it is to adopt the principle as a guide to action on the basis of that principle passing a test of reflective scrutiny. For reflective agents, who are free to adopt any standards they see fit, action guidance cannot just be a matter of the standard’s having sufficient content. No standard could be action guiding for a reflective agent if unsupported by reasoning that vindicates the adoption of that standard. ^{Footnote 41}

When considering a plurality of agents, the reasoning that justifies those standards must be able to coordinate the wills of all members of that manifold. I use “coordinate” in a specific and perhaps uncommon sense here. I am not thinking of the way artifacts of game theory and the like may be used to model a pattern of behavior in solving “coordination problems.” I am thinking of what it would take for a standard to survive the reflective scrutiny of a plurality of agents, such that each can adopt the standard as their own. To adopt a standard as one’s own is, I have argued, an exercise of practical reasoning. So the test of reflective scrutiny must have the form of practical reason: it is reasoning about which central ordering concepts shape one’s activity. (I will say a little below about the criteria reasoning must meet if it is to be capable of coordinating a plurality of wills).

Let me bring these points together. It is impossible to say what each is due without saying how others should act. It is impossible to say how others should act without saying what others should will. And it is impossible to say what others should will without saying how it is possible for reasoning to coordinate the wills of a plurality of agents. So reasoning about what justice is must be such as to produce its object: it is practical reasoning. The nature of justice is captured by a constellation of principles embodying standards that bind both the activity of agents, and the institutions through which that activity plays out, into a practical nexus of obligation. Criteria of justice are not derived from an ideal state of affairs.

Indeed, a proper grasp of what, in Section II.B, I called an infinite or general orienting standard cannot require any depiction of an ideal state of affairs. When we think of the indefinite scope of possibilities for human interaction, it appears that the reason-giving concepts of justice, compassion, honesty, benevolence, and so on (and all the negative ones: cruelty, neglect, deceit . . . ) must have the form of general standards. To grasp such a general standard is to understand the orienting role the standard plays in the open-ended activity of life. Precisely because such activity is indefinite, a proper grasp of any such standard cannot involve the specification of a final landmark, an ideal state of affairs. To properly grasp such a standard is to grasp how a reflective subject could adopt it as something to be unfailingly observed in their open-ended activity of life. This is to grasp how justice operates as a standard of practical reason, not to grasp any final landmark depicted by an ideal state of affairs. ^{Footnote 42}

B. Against the use of idealization

One could agree that justice and other moral standards are general standards, and that it is impossible to grasp the nature of these standards in isolation from considerations of practice, and still hold that idealization is essential to reflection on the nature and justification of such standards. For accepting that justice is a general standard, we must understand the complex nature of this standard, and its relation to other standards, and the broad structure of the ethical outlook in which this standard is located, if reflective scrutiny is to recommend justice as a central ordering concept. This thought returns us to the defense of idealization noted in Section I.B: on all these points, idealization is either essential or extremely useful.

Stemplowska provides the most powerful version of this argument. We must consider how standards of morality and justice interconnect, and, Stemplowska claims, it is far easier to “tease out the relationship between two principles” in the context of “a radically simplified universe.” Further, “by assuming the fanciful we can see more clearly how crucial certain constraints are to shaping what we consider desirable or just when they are present.” ^{Footnote 43} Thirdly, we must also give thought to what our standards demand of us, and idealization is essential to prevent “existing constraints on what we can achieve” from distorting “our understanding of what a given value requires.” ^{Footnote 44} If these claims are accepted, idealization to eliminate distracting noise is equally appropriate in reflection on how we should live as it is in scientific reflection.

Yet once it is accepted that reflection on how we should live has the form of practical reason, the claim that idealization aids clarity must be abandoned. To see why, it is necessary to return to several conclusions that emerged in the course of explaining the form of practical reason. These are: that the object of practical reason is what should be done; that the distinct forms of reasoning are not isolated from each other; and that if practical reasoning is such as to produce its object it must acknowledge the conditions of the will’s productive possibility.

The object of practical judgment is what should be done. This object is necessarily abstract, because there is always a gap between action and norm: any practical judgment can only shape the action to be done in such a way that leaves indeterminate the precise character of that action. ^{Footnote 45} A multitude of patterns of action may instantiate the standards involved in, for example, getting the week’s shopping or respecting others’ rights. Practical reason is also such as to produce its object, and the action, once done, is necessarily an entirely concrete historical event. So at the sharp end of practical reasoning, there must be a movement of thought from abstract standards to historically concrete action. ^{Footnote 46} Here, sanity requires a considerable number of assumptions, on the subject’s part, about the state of the world. Flukes aside, it is a condition of the will’s productive possibility that things are as one takes them to be. It therefore belongs to the form of practical reason, considered under this aspect, that beyond a certain point the more assumptions one makes, the more tenuous that reasoning becomes as a guide to action. Idealizations are impractical here: they constitute defective reasoning.

And this point — that idealizations are impractical — applies equally to the highly reflexive aspect of practical reasoning concerning the nature and justification of standards of justice and morality. Considered under this aspect, practical reason coordinates a plurality of wills. I must postpone until another occasion a full discussion of what it would take for a principle to pass the reflective scrutiny of a plurality. But at least the following can be said. Reasoning that is such as to pass the reflective scrutiny of a plurality must be followable in thought by all to whom it is addressed. ^{Footnote 47} This claim involves a modality that is notoriously difficult to pin down. Yet it must also be true that whether a subject can reflectively identify with a particular reason-giving concept depends on which concepts that subject already reflectively identifies with. ^{Footnote 48} So this too can be said: reasoning that is followable in thought must establish connections that enable a movement of thought from the concepts one identifies with to the concepts that the reasoning promulgates identification with. (This movement of thought by no means must preserve the starting concepts unchanged). This must be a condition of the will’s productive possibility.

The conditions for the productive possibility of a plurality of coordinated wills are accordingly complex. Human capacities for reflection and cognition are only actualized through habituation into a particular historically and culturally contingent intellectual inheritance. The concepts any particular individual identifies with are a matter of history and culture, with all the dizzying diversity that implies. They are also a matter of politics: think, for example, of the ways in which complex histories of injustice in a world disfigured by centuries of colonization charge certain concepts with political significance. ^{Footnote 49} Reasoning that is capable of coordinating a plurality of wills must account for the cultural and historical deposit embedded in the standards it promulgates. Such reasoning must also account for how that deposit stands in relation to the concepts held by the domain of subjects it is addressed to. If reasoning does not do this, it cannot show how the concepts it promulgates can connect to the concepts its audience identifies with.

These historical and cultural contingencies are exactly what simplifying models of justice and right action typically idealize away. For example, theoretical models of the behavior and deliberation of an ideal community of agents, or of contracting parties, operate with a homogenous picture of the intuitions and motivations of agents in an ahistorical context. The heterogeneity of cultural outlooks and thick ethical concepts, as well as the complex relevance of histories of injustice, is idealized away. Likewise, elaborately constructed thought experiments, typically involving fanciful scenarios, aspire to isolate and test moral intuitions in complete detachment from the constellation of historically and culturally contingent thick ethical concepts that would give the intuition content and reflective significance to any particular subject. None of what is being idealized away here is noise. These are essential materials for determining the conditions for the productive possibility of a coordinated plurality of wills. Far from aiding clarity, simplifying idealizations of the sorts just described make reflection on the nature and justification of standards of justice and morality — reflection that is practical in form — impossible.

This objection also shows what is fundamentally wrong with the claim that idealization is unproblematic so long as real-world complexity is reintroduced at some later stage of reflection. To think this is to understand reflection on standards of morality and justice as a process of several steps. The most fundamental step is the identification of first principles or considered intuitions, and it is here that idealizations to screen out complexity are deemed, for reasons of the sort given by Stemplowska, either essential or extremely helpful. Having reached a philosophically clear and rigorous account of these standards, subsequent reflection can reintroduce idealized-away features of the world in order to determine what these standards require of us in the here-and-now. ^{Footnote 50}

Practical reasoning typically exhibits a process that at first glance may seem identical to the multistage process just described. This is most obviously the case in reasoning about what one is doing, which — as noted — must involve a movement of thought from abstract standards to the performance of a particular historical event. But it is also the case in more reflexive practical reasoning about what should be done. O’Neill captures the point well:

Note, however, that with respect to both dimensions of practical thought, this is a process of the increasing concretization of initially highly abstract standards. It is not a process of reintroducing real-world complexity after making simplifying idealizations to arrive at clear first principles or bedrock intuitions.

Indeed, reflection on the nature of justice and other ethical standards cannot be a process of first employing idealizations to clarify first principles and then subsequently reintroducing what has been idealized away. Standards of practical reason must be capable of coordinating a plurality of wills, and this makes the idealization of real-world complexity impractical at the first, fundamental step of reflection. To repeat the three examples given above, the depth of cultural diversity, the complex relevance of historic injustice, and the historical, cultural, and political deposits contained within the reason-giving concepts one is reflecting upon, all fundamentally bear on the question of whether some proposed practical standard could pass the reflective scrutiny of a plurality of agents. For this reason, none of these considerations can be idealized away, only to be reintroduced at some later stage. The identification of first principles must be the identification of principles/reason-giving concepts that can coordinate a plurality of wills. Here, again, idealization is impractical. Stemplowska’s claim that without simplifying idealizations the complexity is too great should be turned on its head. Without an accurate understanding of how complexities of culture and history feature in human reason, reflection on standards of morality and justice will be unable to get a grip on its object. ^{Footnote 52}