After reading through enough philosophical debates, it’s natural to wonder whether philosophers are sometimes just talking past one another. Most terms have more than a single meaning, and in philosophy in particular, the different meanings associated with a term are not always as easy to distinguish as, for example, the different meanings of the word ‘table’ (which can be used as a noun—‘I placed my mug on the table’—or as a verb—‘I propose that we table the motion’). Consider the word ‘freedom.’ Philosophers disagree over what conditions must be met for a person’s will to be free, and thus for the actions that spring from it to be free. However, ‘free will’ is not the only sense of the word ‘freedom.’ Other senses include ‘negative freedom’ or the absence of interference; ‘capability freedom’ or the sort of freedom one has when one has the power to perform an action or achieve a functioning^{Footnote 1}; as well as the sort of freedom John Rawls has in mind when he calls citizens ‘free and equal’: a sense of freedom that looks like freedom of the will but which is allegedly independent of metaphysical debate.^{Footnote 2} With so many kinds of freedom, it must be hard for philosophers to avoid talking past each other sometimes (as David Hume arguably does when he argues for the compatibility of ‘freedom’ with determinism by appealing to a kind of ‘freedom’ that sounds rather like negative freedom).^{Footnote 3}

The conventional view is that disagreement, when traceable to different uses of a word, is philosophically uninteresting. Such disagreements are dubbed ‘merely verbal’ or ‘merely linguistic’ and resolving them is thought to merely require disambiguating the different concepts associated with the word, as well as perhaps determining which use of the word is more consistent with common usage.^{Footnote 4} Philosophically interesting disagreements are thought to be specifically about conceptions rather than concepts, i.e., about how to correctly fill out the content of a shared concept that disputants are mutually discussing.

In some verbal disagreements, however, it is wrong to say that the nature of the dispute is merely linguistic. Sometimes divergent uses of a word reflect a theoretical disagreement, one that manifests itself linguistically but is deeper than a dispute over standard usage. Moral theorists’ use of the word ‘morality’ is arguably a good example of this.^{Footnote 5} For most moral theorists, a conception of morality is a conception of right action. Proponents of different conceptions, such as utilitarians and deontologists, disagree over the principles that spell out morality’s content, but they implicitly agree that the concept of ‘morality’ is to be understood as ‘right action.’ The proponents of virtue ethics are ostensibly an exception to this. For virtue ethicists, a conception of morality is a conception of virtuous character. Different virtue ethicists sometimes disagree about how particular virtues should be understood or about the relationship between them, but they implicitly agree that the concept of morality is to be understood in terms of the virtues that constitute a virtuous character. To treat these different uses of the term ‘morality’ as merely a verbal disagreement is to miss that they reflect a significant theoretical disagreement. Virtue ethicists and other moral theorists disagree over how the relationship between character and action should be understood.^{Footnote 6} For the virtue ethicist, right action can only be understood with reference to an account of moral character. Right actions are those that flow from virtue, and thus identifying them presupposes an account of the virtues. Other theorists, by contrast, see moral character as parasitic upon right action. For them, the value of a character trait depends upon the value of the actions it is conducive to, making an account of right action more fundamental than an account of moral character.

In my book, A Conceptual Investigation of Justice, I argue that a verbal disagreement, similar to the one between virtue ethicists and proponents of the other major moral theories, is present in contemporary political philosophy.^{Footnote 7} On one side of the disagreement is a view shared by many contemporary liberals, most notably Rawls, which understands justice to be about the moral rightness of institutions (I call it the ‘contextualist view’). Put another way, they think of it as the output of practical reasoning about how institutions ought to be designed.^{Footnote 8} By contrast, many value pluralists, most notably Isaiah Berlin, understand justice to be one fundamental value among a plurality of fundamental values.^{Footnote 9} When justice is understood as one fundamental value among many, to say that something is a requirement of justice is not to say that it is required, all things considered, as the demands of justice are often not entirely feasible and must sometimes give way to the demands of competing values. Thinking of justice this way means thinking of it as an input in practical reasoning, rather than the output of it.^{Footnote 10}

The details of my book are discussed in the insightful critiques that comprise this special issue, as well as in my replies to them. Put very briefly, though, I argue that two prominent debates in contemporary political philosophy—the debates over luck egalitarianism and over the scope of principles of justice (do they apply to personal choices?)—are traceable to the disagreement between pluralist and contextualist understandings of justice, and that the plausibility of different positions available in those debates depends upon whether the contextualist view or the pluralist view is correct.^{Footnote 11} Furthermore, though the disagreement between contextualists and pluralists may, at first glance, appear to be merely a conceptual dispute over which use of the term ‘justice’ is more consistent with standard usage (is ‘justice’ normally used to refer to an input in practical reasoning or to the output of it?), I argue that there’s a philosophically correct answer to the question ‘how should ‘justice’ be used?,’ one that turns on issues concerning the sense in which justice has primacy in the institutional context,^{Footnote 12} as well as on whether fairness is primarily substantive rather than procedural.^{Footnote 13}

In their insightful papers, the authors in this special issue take issue with a number of the above-mentioned claims, as well as with the arguments I offer in favour of them. One of their targets is my claim that luck egalitarianism (i.e., the view that inequalities traceable to choice are just, but that those traceable to luck are not) is compelling if understood as a theory of a particular value (substantive fairness), but that it’s implausible if understood as a theory of all-things-considered institutional rightness. In support of this claim, I argue that we should distinguish between intuitive judgements that are internal to substantive fairness and intuitive judgements that are external to it. On my view, luck egalitarianism, because it coheres quite well with our intuitive judgements of fairness, is an excellent theory of fairness but also a rather poor theory of institutional rightness, as some of our judgements of fairness are at odds with what’s right, all things considered, e.g., the judgement that it is fair not to compensate those who, because of their imprudence, are unable to meet their basic needs. Colin Macleod, however, worries that any attempt to draw a boundary between judgements that do and judgements that don’t belong to fairness will inevitably be controversial. And Matthew Palynchuk worries that some of the fairness judgements I draw upon, specifically those about disability, presuppose the problematic ‘medical model’ of disability, and that adopting a (in his view, superior) ‘social model’ of disability is not a costless move for me.

In her article, Kristin Voigt targets my treatment of the debate over the scope of distributive justice. There, I argue that the view that principles of justice appropriate for institutions are also appropriate for the context of personal decision-making is implausible if principles of justice are all things considered. Justified all-things-considered principles are necessarily sensitive to the facts that differentiate contexts, e.g., the fact that people, unlike institutions, have personal lives; and thus all-things-considered principles suitable for the institutional context are presumably different from those suitable for the personal context. Furthermore, I argue that G.A. Cohen, in his efforts to defend his claim that the difference principle extends to the personal context, inadvertently treats the difference principle as if it’s a principle that expresses the content of a fundamental value, rather than an action-guiding regulatory principle, and that this undermines his critique of Rawls’s restriction on its scope. Voigt, however, maintains that I’m confusing Cohen’s internal critique of Rawls for Cohen’s own view of the relationship between justice and personal choice. Though Cohen thinks that Rawls, as a matter of consistency, is committed to applying the difference principle to citizens’ personal choices, Cohen himself is not, external to his critique of Rawls, committed to the view that the difference principle supplies the content of a justified ethos of justice. For Cohen, a personal commitment to justice is a commitment to a fundamental value, not a commitment to action-guiding principles, and the content of that fundamental value is, for Cohen, supplied by luck egalitarianism.

In his article, Louis-Philippe Hodgson claims (among other things) that I’m wrong to maintain that the issue at stake between pluralist and contextualist understandings of justice is whether fairness is primarily procedural. Throughout my book, a central claim I make is that all-things-considered institutional principles can only tell us what justice is if fairness is primarily a procedural concept. When their fairness is derived from a fair procedure, justified all-things-considered principles are appropriately thought of as fully fair. If, by contrast, they instead represent a compromise between substantive fairness and the other political values in conflict with it, then they can only be imperfectly fair, as their all-things-considered status then requires some deviation from fairness. Principles that represent a compromise between substantive fairness and other values may only be thought to express what’s institutionally required of us, all things, including justice, considered. In my book, I argue that a fundamental principle of substantive fairness is ineliminable. More specifically, I argue that a fundamental principle of substantive fairness is needed to specify the content of procedural fairness, and that, when substantive fairness is used to evaluate the principles our fair procedure selects, those principles necessarily fall short of it. Hodgson, however, notes that Rawls has his own, substantive understanding of fairness (fairness as reciprocity), and that Rawls too directly applies substantive fairness to the principles his procedure selects. According to Hodgson, this casts doubt on the claim that there’s an underlying theoretical disagreement between pluralists and contextualists, which in turn supports the view that their dispute is merely a verbal one.

My commentators express various other worries related to my treatment of the relationship between substantive and procedural fairness. Macleod, for example, worries that substantive, distributive principles don’t have the right form to be directly applied to procedures. And Hodgson worries that my use of specifically luck egalitarian fairness at the procedural level is problematic, as luck egalitarianism is, among other things, a rather controversial view.

At the end of this special issue, I will address many (though, due to limited space, not all) of the criticisms raised by my commentators. For the moment, however, I would like to turn away from my commentators’ thoughts and briefly address a puzzled response I sometimes receive when explaining my book’s main argument. In particular, I’m sometimes asked what it means to say that a dispute is verbal, but not merely verbal. How, exactly, should such disputes be understood? On the one hand, the presence of underlying theoretical disagreements entails that verbal disputes of this sort are not simply a competition between different concepts or uses of ‘justice.’ On the other hand, the fact that the word ‘justice’ is being used differently by different camps entails that disputes of this sort should not be understood as a competition between different conceptions or theories of ‘justice,’ since two conceptions, to be in competition with each other, must be theorizing the same concept. Upon reflection, I think that the traditional concept-conception distinction is too simple, and that a third, middle category is needed to account for disputes that, though verbal, are more than that. We might perhaps call the theoretical entities that occupy this middle category ‘referential identities,’ and speak of justice as having multiple, competing referential identities in the contemporary literature.^{Footnote 14}

To better understand what disputes that lie between the concept-conception distinction involve, it will be helpful to discuss a canonical case from the philosophy of language. In his work on linguistic meaning, Gottlob Frege is famous for showing that a term’s meaning is not exhaustively derivable from that to which it refers. To demonstrate this, Frege asks us to reflect on the terms ‘morning star’ and ‘evening star.’^{Footnote 15} Though we now know that the morning star and the evening star are actually the same celestial body (Venus), the ancient Greeks thought them to be different bodies. The morning star appeared in the morning, whereas the evening star appeared in the evening, and astronomers understandably thought they were looking at two different celestial bodies (and hence created two different names). Though the terms ‘evening star’ and ‘morning star’ had the same referent (both referred to Venus), it’s clear that they didn’t originally mean the same thing. When using the term ‘evening star,’ people had a very different understanding of its referent than they did when using the term ‘morning star,’ and this different understanding created a difference in meaning between the two terms. Frege concluded that, in addition to reference, meaning must also have a cognitive component, otherwise there would be no explanation for why the terms ‘morning star’ and ‘evening star’ had different meanings. This cognitive component is normally referred to as a word’s ‘sense’ or ‘intension.’

I believe that the disagreement between pluralist and contextualist understandings of ‘justice’ involves the reverse of what we see in the morning and evening stars case. With respect to the morning and evening stars, we have two different words with two different intensions that nonetheless refer to the same thing (Venus). With respect to the pluralist and contextualist understandings of ‘justice,’ we have the same word (justice) being used to refer to two different things (an input in practical reasoning and the output of practical reasoning about institutions), but there’s significant overlap between each word’s intension. Contextualists understand ‘justice’ to be the first virtue of institutions, as well as the part of moral rightness that pertains to fairness in matters of distribution and rectification. Similarly, pluralists understand ‘justice’ to be the first virtue of institutions, as well as the part of moral rightness that pertains to fairness in matters of distribution and rectification. This overlap between intensions is significant, and it’s what makes the disagreement between pluralists and contextualists more than just conceptual.

In light of the above, we should distinguish (at least) three different ways in which disputants using the same term may disagree about how the term should be understood. First, there are disagreements in which a term is being used to refer to two different things, each with its own distinct intension, i.e., traditional conceptual disagreements. Second, there are disagreements in which disputants are using the same term to refer to the same thing, and their basic understanding of that thing is also the same, even though they disagree about the specifics, i.e., a traditional disagreement between proponents of opposing conceptions or theories. Third, there are disagreements in which disputants are using the same term to refer to different things, but they share the same basic understanding of their respective referents, i.e., a disagreement in which each disputant’s use of the word possesses the same intension, even though they’re referring to different things. Let’s call such disagreements ‘referential disputes.’

I don’t have space here for a lengthy discussion of what this third category of dispute implies. I will note, however, that it’s important not to incorrectly categorize the disputes that belong to it. Incorrectly categorizing a referential dispute as a traditional conceptual dispute leads to dismissal, since traditional conceptual disputes are philosophically uninteresting. By contrast, incorrectly categorizing a referential dispute as a dispute between competing conceptions can lead to disputants talking past one another, as disputants will be under the impression that they share more assumptions than they actually do. For example, an argument between theorists who agree that ‘justice’ is all-things-considered institutional rightness, but who disagree about what the requirements of institutional rightness are, should, if conducted well, look different from an argument between theorists who disagree about whether justice is institutional rightness. Theorists who agree that justice is institutional rightness can agree that justice, because it incorporates a great many political considerations, is weightier than other matters that bear on institutional design, and they can thus turn their attention away from that matter and towards other issues they disagree about, e.g., the question of which institutional principles would be selected in the original position, or the question of whether the original position is the fairest procedure for selecting institutional principles. By contrast, theorists who disagree about whether justice is institutional rightness cannot, in the context of their dispute with one another, take it for granted that justice trumps other political considerations, or that principles of justice are even the sort of thing the original position selects, since the referential identity one assigns to justice affects whether these assumptions are plausible. For disputants who disagree about that to which the word ‘justice’ should be used to refer, the issues to address are those that pertain to the shared intension their divergent uses of the word possess. For example, since both can agree that justice has primacy in the institutional context, they might engage in a debate over how primacy should be understood. Though contextualists prefer to understand primacy as ‘weightiness,’ it’s sensible for pluralists to defend a different interpretation, given their view that justice is defeasible.^{Footnote 16}

For his part, Phil Smolenski is doubtful that justice can be thought of as anything other than a complex, ‘weighty’ concept that incorporates a variety of political considerations. He argues that a number of important issues of justice, such as the preservation of minority cultures or the protection of sexual minorities against discrimination, involve more than just considerations of substantive fairness. Values such as autonomy and liberty underlie their importance, too, and are a part of what makes these issues pressing matters of justice.

Smolenski is right to note that autonomy and liberty underlie many issues of justice, though I think he’s wrong to claim that this observation undermines the view that justice is a single, fundamental value. My replies to both him and my other commentators will have to wait until the end of this special issue, though.