Contractualism as Restricted Constructivism

Abstract

Metaethics is often dominated by both realist views according to which moral claims are made true by either non-natural or natural properties and by non-cognitivist views according to which these claims express desire-like attitudes. It is sometimes suggested that constructivism is a fourth alternative, but it has remained opaque just how it differs from the other views. To solve this problem, this article first describes a clear constructivist theory based on Crispin Wright’s anti-realism. It then outlines an argumentative strategy that can be used to argue against constructivist views about practical reasons. The rest of the article explains how the outlined constructivist metaethical framework, reasons, and contractualism in normative ethics can still be used to create a new viable metaethical constructivist position about right and wrong.

1 Introduction

Metaethical theories attempt to understand the nature of moral properties, language, and judgments. The three main alternatives are (i) non-naturalist realism, (ii) reductive naturalism, and (iii) different varieties of expressivism.¹ Contractualism is often understood as an ethical theory, but sometimes it has also been seen as a metaethical position called ‘constructivism’ (Southwood 2009, p. 926).² I aim to show that contractualism can be used to formulate a new metaethical position only in a restricted sense.³ Contractualist theories must rely on practical reasons and constructivism about them fails. Yet, if we accept a non-constructivist view of reasons, we can use these reasons and contractualism to create a new restricted constructivist metaethical position about right and wrong.

Section 2 outlines realism and reductivism as broad worldviews. Section 3 then creates a third constructivist alternative based on Crispin Wright’s epistemic notion of truth. Section 4 explains the difference between global and restricted versions of constructivism. Sections 5 and 6 outline an argumentative strategy that can be used to argue against constructivism about reasons. Sections 7, 8, 9, and 10 finally sketch how contractualists can use unconstructed reasons to create a new restricted constructivist metaethical view.

2 Realism and Reductivism

According to Extreme Realism, the world consists of independently existing layers of reality. At the bottom-level, there are space–time points, basic physical particles, and their properties. In virtue of this layer, there are atoms and molecules; in virtue of them, biological organisms; in virtue of them, economic, political and social systems, and evaluative and moral facts; and so on.

Here the ‘in virtue of’ locution describes a grounding relation between two independent levels of existence. Imagine that God is creating a world and stops at level n. According to Extreme Realists, at this point the world does not contain the objects and properties of the level n + 1: if God stops at the biological level, the world does not yet contain moral properties. In order to get those properties, God has to create again. Extreme Realism is thus maximally ontologically committed: it posits many independently existing layers of reality.

According to Austere Reductivism in contrast, the most fundamental natural science—physics—captures everything that exists. In explaining everything that takes place, physicists only posit the existence of space–time points, their material occupants instantiating arrangements of localised intrinsic properties, and the external space–time relations between them. Austere Reductivists then claim that all truths are about those objects, properties, and relations. Take any simple true sentence of the form ‘X is F’ where ‘X’ is a name of an object and ‘is F’ a predicate which picks out a property such as feeling pain, being a cat, or the like. On this view, these true sentences can always be translated into true sentences about the space–time points, their material occupants and their properties, and the relations between them. The different ways of arranging the previous objects, properties and relations create a space of possibilities of how the world could be. Any sentence of the form ‘X is F’ is then taken to divide the space of possibilities to conditions under which the sentence is true and to ones in which it isn’t (Lewis 1970, Sect. 3).

In this framework, any truth of the form ‘X is F’ can be captured by sentences which only describes the basic particles, their properties, and the relations between them (Jackson 1994, pp. 31–32), because the latter sentences divide up the space of possibilities in the very same way as the original sentence. According to Austere Reductivism, there are thus no truths about our world that cannot be accounted for in terms of differences and similarities among patterns of localised particles and their properties.

3 Constructivism

Constructivism is located somewhere between metaphysically committed realism and metaphysically parsimonious reductivism (Lenman 2010, p. 178; Timmons 2004, p. 93). How could there be room for such a view? This section describes Wright’s (1992) answer to this question.

According to the previous views, true sentences correspond to wordly facts that make them true. These positions only disagree about the nature of the truth-making facts. According to the Extreme Realists, these facts consist of sui generis objects and properties, whereas, according to the Austere Reductivists, they consist of the space–time points, the basic particles, and their properties.⁴

In order to formulate a third alternative, we must give up the correspondence theory of truth (Dorsey 2012, pp. 100–108) and adopt instead an epistemic conception of truth.⁵ In any area of discourse, there are epistemic norms that describe when a speaker is warranted to assert its sentences. They can be used to define a property of sentences called ‘superassertibility’: a sentence is superassertible if and only if there is an accessible state of information which would justify asserting that sentence and continue to do so no matter how much more humanly possible investigation were done in the future (Wright 1992, pp. 44–48).⁶

At least in some discourses, superassertible sentences have all platitudinous features of true sentences. The sentence ‘X is F’ is true if and only if X is F, truth is a timeless property of sentences, truth doesn’t come in degrees, and a sentence is true only if it corresponds to facts. Wright has shown that, in some discourses, the superassertibility property of sentences satisfies these same platitudes: ‘X is F’ is superassertible if and only if X is F, superassertibility is a timeless property of sentences, it doesn’t come in degrees, and the superassertible sentences correspond to the relevant facts when the latter is understood in a deflationist way (Wright 1992, pp. 44–61).

Given that superassertibility is an epistemic property of sentences, it is the most minimal property of sentences, which satisfies the previous truth-platitudes. We can thus assume that at least in some discourses nothing further is required for truth than superassertibility (Wright 1992, pp. 71–78). The previous platitudes specify a role which truth plays in a discourse and so if superassertibility satisfies these platitudes in a discourse we can assume that it realizes the role of truth in that discourse.⁷ Until we have good reasons to believe that truth in a discourse consists of more robust correspondence, we are entitled to believe that in that discourse truth is realized by superassertibility. Wright has also helpfully outlined the ways in which realists about a given domain can argue that the property of truth is realized by some more robust property in that domain (Wright 1992, Chs. 3–5).

If being true is realized by superassertibility in a discourse, then the superassertible sentences of that discourse are its true sentences. So, if the truth of colour-sentences is realized by their superassertibility, then the superassertibility of the sentence ‘X is red’ entails that it is true that X is red, which also entails that X really is red. Furthermore, if we can talk about fact and properties in a deflationist way, then the superassertibility of the original sentence in addition entails that it is a fact that X is red and that X has the property of being red (Wright 1992, pp. 24–27).

Constructivism about colour thus means that redness is a projection from the relevant superassertible sentences. This view is a third, ‘constructivist’ alternative. It is not a realist view because it does not assume ontologically heavyweight sui generis colour facts and properties. The view isn’t a form of reductivism either. You have warrant for asserting that an object is red on the basis of some information specified by the epistemic norms of the colour discourse. For example, perhaps you have warrant for claiming that an object is red when you know that normal observers are disposed to call it ‘red’ in broad daylight. Yet, even if this information warrants your statement, it is about a quality of the object—its redness—rather than about the relevant dispositions of the normal observers.

4 Global and Restricted Constructivism

Global constructivism claims that superassertibility realizes the truth of all true sentences (Wright 1996, Sect. 2). If you are a constructivist about some property F, then you must specify what information makes the sentences about Fness superassertible. If you are a global constructivist, then, for every property, you have to find what further information makes the sentences about that property superassertible.

This view might be thought to lead to a vicious regress. One would need to find more and more new information that makes the sentences of the previous domains superassertible. Such a regress could, however, be avoided if all discourses formed a large interrelated network. Yet, even if such networks avoid the problem, there are still discourses that are difficult to understand in the constructivist way. They include, for example, discourses about basic particles, ordinary objects, phenomenal experiences, and semantic properties.

This suggests that we should be constructivists only in certain domains. On this view, truth in some domains is realized by robust correspondence whereas in others it is realized by superassertibility. This view then claims that the information about the former truths makes the sentences of the latter discourses superassertible. People who accept this position are realists about the basic truths and constructivists about the truths in the other domains. Where should we then draw the line between the basic and constructed truths? One natural suggestion is that we should draw it between ‘the natural domain’ and ‘the normative domain’ (Dorsey 2012, p. 109).

We should thus be realists about all the domains that can be investigated scientifically. This natural domain can be taken to include the basic particles and the laws of nature, atoms and molecules, biological systems, social, economic, and political systems, and psychological properties. According to this proposal, the natural domain doesn’t contain any normative properties, but information about it makes normative sentences superassertible. Restricted constructivist could then argue that the truth of the latter sentences is realized by their superassertibility, and that the normative properties are thus projections from minimally true sentences. This constructivist metaethical position is not a form of realism, reductivism, or expressivism. Normative facts would not have distinct, independent existence, nor would they be reducible to natural facts either given that the information about the natural facts only makes the normative sentences superassertible.

5 Constructivism About Reasons

Sentences of the type ‘that X is F is a reason for A to φ in C’ are paradigmatic sentences of the practical reasons discourse. They specify what is the reason, what that reason is a reason for, for whom it is a reason, and the circumstances in which the reason-relation holds. Constructivism about practical reasons claims that the truth of these sentences is realized by their superassertibility.

Such proposals need to identify the information that makes sentences about reasons superassertible. Constructivists emphasise the importance of deliberation procedures (James 2012, p. 60; Lenman 2012, p. 215). We can thus formulate these views with the following schema:

Constructivism about Reasons: Information about the outcomes of the deliberation procedure R makes claims about reasons superassertible.

Constructivists then disagree about R. According to a simple Humean view, information about whether a statement about a reason seems true from the agent’s own deliberative perspective is sufficient to make the statement superassertible (Street 2008). Humeans can also rely on information about more idealized deliberation procedures: they could claim that information about what a fully rational version of A would want A to do in C makes the relevant sentences about A’s reasons superassertible.⁸

Another, Sidgwickian alternative would be to argue that one can subject an intellectual seeming about a reason to critical reflection in which one must pursue conceptual and factual clarity, stability in reflection, internal coherence and interpersonal agreement (Sidgwick 1884, pp. 339–342; Scanlon 1998, pp. 65–69). Information about candidate reasons surviving these stages of critical reflection could then be claimed to make the relevant sentences about reasons superassertible.

Similarly, Kantians could argue that information about whether A could consistently universalise her maxim of action makes the relevant sentences about A’s reasons superassertible.⁹ Aristotelians, in contrast, would be give even more substantive descriptions of R. They could argue that sentences about reasons are made superassertible by information about the practical judgments of a wise person or by information about the outcomes of good dispositions of practical thought (Setiya 2007; Tiberius 2012).

6 An Argumentative Strategy

Constructivists and realists about reasons disagree about which property realizes the truth of the sentences about reasons. Constructivists claim that their is realized by superassertibility, whereas realists claim that it is realized by correspondence to facts. Realists therefore claim that the satisfaction of the epistemic norms in the reasons-discourse only at best tracks the reason-relations. These views thereby also disagree about the modal status of the connection between the satisfaction of the relevant epistemic norms and the reasons-facts (Wright 1992, p. 112; James 2012, pp. 61–65). According to constructivism, because the superassertibility of the reasons-sentences constitutes their truth, necessarily, if a sentence about an agent’s reason is superassertible, it is fact that the agent in question has that reason. In contrasts, realists must accept that it is possible that a sentence satisfies the relevant epistemic norms and yet the corresponding fact about reasons fails to obtain. We can therefore decide which of these views is correct by considering which one of these modal relations is more plausible.

There are Moorean facts about reasons (Dworkin 1996; James 2012, p. 63; Nagel 1997). In 2012, a 23-year-old female was raped by six men in South Delhi. The victim sadly died from her injuries two weeks later. It is a Moorean fact about reasons that the six men had sufficient reason not to rape the victim. This is one of the ‘things we know better than we know the premises of any philosophical facts to the contrary’ (Lewis 1996, p. 549).¹⁰

The Moorean facts about reasons have a strong modal status (Schroeder 2007, p. 106). In a wide range of possible worlds closest to ours, it remains the case that the six men had sufficient reason not to rape the victim. In these worlds, rape still causes suffering, it violates sexual autonomy, it has other horrible consequences, and the rapists have all the same normative capacities. The attitudes people have in these worlds towards rape, their beliefs about rape, and the relevant social conventions can, however, differ in these worlds.

The constructivists about reasons face a dilemma when they attempt to accommodate the modally strong Moorean facts about reasons.¹¹ They need to guarantee that there is a necessary connection between the satisfaction of the epistemic norms of the reasons-discourse and the relevant facts about reasons. The first way to secure this connection is to specify the epistemic norms in a way that is conceptually tied to antecedent facts about reasons (Wright 1992, p. 112).

The constructivists can thus specify R so that it is impossible for the sentence ‘the six South Delhi men had no reason not to rape the victim’ to pass this procedure. The Sidgwickian proposal could, for example, state that the previous sentence could not be stable in reflection for those who understand its concepts. Here we assume the Moorean fact about reasons first and then define the epistemic norms so that the satisfaction of those norms guarantees the Moorean fact in the relevant worlds. This proposal, unfortunately, collapses into realism (Wright 1992, p. 112). The epistemic norms of the reasons-discourse now conceptually depend on facts about reasons, and so they cannot play an explanatory role in the construction of the reasons.

This means that constructivists must formulate the epistemic norms of the reasons-discourse by relying only on concepts that don’t make a reference to substantial facts about reasons.¹² The problem here is that there are situations in which a sentence about an agent’s reason satisfies the resulting independently specified epistemic norms but the truth of the sentence conflicts with the relevant Moorean facts about reasons (Wright 1988, pp. 23–24; Miller 2003, pp. 136; Shafer-Landau 2003, p. 42). The constructivists about reasons cannot thus vindicate the strong modal connection in this way either.

Consider the Smith-inspired constructivism about reasons. We must now understand full rationality (being fully informed and having maximally coherent and unified desires) as a purely formal quality. As a result, there could be individuals whose advisors would have substantially different motivational sets even if they would all be equally fully rational. This type of formal maximal coherence and unifiedness would only require structural isomorphism between those sets. However, if the ideal advisors can have substantially different but equally rational sets of desires, then at least some of the fully rational advisors could have had no desire for the actual South Delhi men not to rape the 23-year-old woman (Parfit 2011, vol. 1, pp. 72–82).¹³

If the constructivists formulate the epistemic norms of the reasons-discourse independently of substantial facts about reasons, there won’t thus be a strong modal connection between the satisfaction of the epistemic norms and the corresponding Moorean facts about reasons.¹⁴ As a result, these epistemic norms could at best imperfectly track the independent facts about reasons, which entails realism. Simply put, the Moorean facts about reasons would fail to obtain in all the relevant worlds if the constructivist views about practical reasons were true. Yet, their obtaining doesn’t depend on whether certain statements about these reasons happen to satisfy the independently specified epistemic norms.

This section has described an argumentative strategy against constructivist views about reasons. We can always argue against such views by showing that they either (a) accommodate the strong modal status by relying on antecedent facts about reasons or (b) they fail to accommodate the strong modal status of the Moorean facts. In contrast, non-naturalist realist views about reasons can arguably support the strong modal status of the Moorean facts about reasons (Parfit 2011, Vols. 1, 2, Chs. 2, 31). Recently also reductivist naturalists and expressivists have argued that their views too are compatible with the modally strong Moorean facts about reasons (Schroeder 2007, Ch. 6; Lenman 2009).¹⁵

7 Constructivism About Wrongness and Contractualism

This section motivates the idea that we should be constructivists about right and wrong even if we reject constructivism about reasons.¹⁶ If you are a non-naturalist, naturalist, or an expressivist about reasons, you could first think that claims about wrongness just are about reasons understood in your favoured non-constructivist way. Perhaps the claim that an act is wrong just means that there are decisive reasons not to do it (Parfit 2011, vol. 1, p. 166). The main problem with this simple proposal is that it is not sufficient for an act to be morally wrong that there are decisive reasons not to do it. I can have decisive reasons not to go out without an umbrella even if doing so isn’t morally wrong.

In response, it could be claimed that claims about wrongness are about a whole set of reasons: about (i) strong reasons not to do an act, but also (ii) about strong reasons to accept moral principles that forbid doing it, and (iii) about reasons for having negative reactive attitudes towards its performance (Parfit 2011, vol. 1, pp. 65–66). The problem here is that all these reasons are not always necessary. There are cases in which an agent can act wrongly even if there are no reasons to blame her (Parfit 1984, pp. 33–34). It is hence difficult to specify a set of reasons that would be both necessary and sufficient for an act’s wrongness. Claims about wrongness cannot therefore be about reasons in a direct way.

The other alternative for the non-constructivists about reasons is to think that wrongness is a distinct property, both metaphysically and semantically independent of reasons. One could then give one account of reasons and another of wrongness with one’s metaethical framework. The problem here is that there are inferential connections between wrongness and reasons. If we know that an act is wrong, we can expect that there are at least some reasons related to that act. We can expect that, by default, there are reasons not to do the act and to blame those who commit those acts. If some of these reasons are not present in a particular case, there is a special explanation for this. The problem is that, if reasons and wrongness are understood separately in non-naturalist, naturalist or expressivist way, then this type of inferential connections are difficult to explain.

The main advantage of constructivism about wrongness is that it can avoid the previous difficulty. It can explain the inferential connections between reasons and wrongness without reducing wrongness to reasons. According to this view, information about reasons makes claims about the wrongness superassertible and thus minimally true. The fact that the epistemic norms of the wrongness discourse are tied to information about reasons then explains why we can expect that reasons are related to wrong actions even if claims about wrongness are not about reasons (Sect. 3).

This view can be formulated in two stages. Firstly, we can begin from the idea that we are justified to claim that an act is wrong when we have enough information about the cluster of reasons (i)–(iii) that is related to the act. These reasons need not be overriding reasons and not all of them need to exist in a specific case just as long as a sufficient number of them are present in each case. Information about reasons can thus be argued to make claims about wrongness superassertible.

At the second stage, we can try to discover general information about the previous type of clusters of reasons that are related to different acts. One could then claim that such more general information about reasons too makes wrongness-sentences superassertible and as a result also minimally true. My suggestion is that, as an ethical theory, contractualism enables us to locate the previous information about the relevant cluster of reasons for the purposes of the general constructivist view about wrongness.¹⁷ The relevant information turns out to be about personal reasons for rejecting moral principles.

At the core of contractualism is the idea that some acts are forbidden by the set of moral principles which no one could reasonably reject (Scanlon 1998, Chs. 4–5). Call this set ‘the contractualist principles’. There are many alternative sets of moral principles which we could internalize. For every such set, there is a world in other ways like the actual world in which that code has been internalised by an overwhelming majority.

Whichever moral code has been internalised in a world determines what kind of lives individuals come to live. Call these lives of the individuals their ‘standpoints’ (Scanlon 1998, pp. 202–206). Some qualities of the standpoints make them choiceworthy. These are the qualities of the standpoints which give us reasons to prefer to occupy them. Other qualities of the standpoints provide us with reasons for not wanting to occupy them. Call the latter qualities of the standpoints ‘burdens’ (Scanlon 1998, p. 195). As examples of such burdens, Scanlon lists bodily injury, not being able to rely on the assurances of others, not having control over what happens to oneself and not being able to pursue our own projects or to have family and friends (Scanlon 1998, p. 204).

We can then stipulate that individuals whose standpoints are burdensome can present their burdens as personal reasons for rejecting the principles under which they live (Scanlon 1998, p. 195). How strong such reasons are depends on the seriousness of the burdens. These personal reasons can be used for defining the reasonable rejectability of sets of moral principles. From a world in which certain principles have been internalised, we take the individual whose life is the most burdensome. This individual can reasonably reject the principles under which she lives if there is an alternative set that doesn’t create as burdensome standpoints to anyone in the world in which it has been adopted (Scanlon 1998, pp. 195–196).

As a consequence, there will be one set that no one can reasonably reject. It is such that under all other sets live some individuals whose lives are more burdensome than the lives of the people who live under it. The individuals who have to live such unnecessarily burdensome lives have stronger personal reasons for rejecting the principles under which they live than the individuals living under the non-rejectable set have for rejecting their code. In this way, the property of being forbidden by the contractualist principles is a function the personal reasons which different individuals have for rejecting the sets of principles.

8 Warranted Assertion

My proposal then is that information about which acts are forbidden by the contractualist principles makes sentences about wrongness superassertible. For this proposal to work, (i) the information about whether an act is forbidden by the contractualist principles needs to justify asserting sentences about the wrongness of acts, (ii) such information needs to be accessible, and (iii) it would also need to continue to justify claims about wrongness no matter how much further investigation we did. I will consider the reasons for accepting (i) in this section and the reasons for accepting (ii) in the next one. (iii) is, of course, a matter of how the future investigation pans out.

As explained above, we are justified to assert that an act is wrong whenever we are aware of strong reasons not to do the act, to accept and follow principles that forbid it, and for having negative reactive attitudes towards the performance of the act. Does being forbidden by the contractualist principles entail the existence of such a cluster of reasons?

Firstly, there are strong reasons not to do acts that are forbidden by the contractualist principles. By following these principles we can form relationships with others that enable us to recognise their ability to evaluate and to act on reasons (Scanlon 1998, p. 162). If we didn’t follow such principles, we would express our willingness to ignore the personal objections of others. Furthermore, we tend to be offended when we realise that others do not see us as someone to whom justification is owed. In contrast, if we are able to justify our actions on the basis of the contractualist principles, we can stand by our actions (Pettit 2000, p. 231). This relation to others constitutes a concrete good in our lives. We also have reasons not to do these acts because of the first-order burdens which they tend to cause to others.

Secondly, we have reasons to accept the contractualist principles—to let them guide our moral deliberation. In the same way as merely conforming to these principles, accepting them too helps us to form valuable relationships with others. It expresses that we are concerned about their potential objections. The acceptance of the contractualist principles also has further advantageous expectation effects. If we accept those principles, others are more likely to treat us in ways which they can justify to us on the basis of the contractualist principles (Lenman 2006, p. 12). This will create an atmosphere of trust which will support mutually beneficial co-operation.

We also have reasons for adopting reactive attitudes such as blame and resentment when we and others act in the ways that are forbidden by the contractualist principles. When someone fails to follow the contractualist principles, this expresses lack of respect for our rational capacities. This is because, if an agent violates those principles, she does not care about what kind of objections we might have to her actions. In such cases, the agent does not think that we deserve to be treated in the mutually justifiable ways, which is a way of denying us an equal moral status. By violating the contractualist principles, the agent thus reveals that she has attitudes towards us which impair the valuable moral relationship that existed between us (Scanlon 2008, pp. 140–141).

This is why we have reasons to blame those who fail follow the contractualist principles (Scanlon 2008, pp. 166–179). Blame and resentment are reactions that enable us to distance ourselves from those who have impaired their relationship to us through their actions (Scanlon 2008, p. 141). Such reactions are a reasonable response to the fact that others have excluded us from the valuable moral relationships described above. We also have good reasons to blame others for the first-order burdens that we have to bear when the contractualist principles have been violated.

If the previous arguments work, then if an act is forbidden by the contractualist principles there are strong reasons not to do the act, to accept and follow principles that forbid it, and also for blaming and resenting others for doing that act. This entails that information about an act being forbidden by the contractualist principles justifies us to claim that the act is wrong.

9 Accessibility

In order for the previous information to make the relevant sentences about the wrongness of different acts superassertible, this information would also have to be accessible. Firstly, we would need to be able to discover what kind of standpoints the adoption of different sets of moral principles creates for different individuals. Much of this information can be acquired through thought-experiments. If the principles encouraged people to kill one other, presumably our lives would be shorter. Furthermore, we can acquire information about the consequences of different principles by observing different communities in which they have been adopted and we could also empirically test what consequences different sets would have. The consequences of different sets of moral principles are thus knowable.

We would also need reliable information about how strong reasons different burdensome qualities of individual lives provide objecting to the principles that are responsible for them. Is it possible to compare the choiceworthiness of different individual lives? Firstly, we often make choices that require evaluating the choiceworthiness of different lives. We have to choose which career to pursue and what kind of relationships to form. When we make these decisions, we have little information about the future circumstance which will shape our lives, desires, and tastes. We cannot therefore ground our decisions on how many of our special tastes and desires will be satisfied, but rather we must assess in more general terms whether the new circumstances would enable us to pursue our projects, to form special relationships, to have sufficient wealth, and so on. If we can compare our future lives on these general grounds, it must be possible also to compare the choiceworthiness of the possible lives which could be ours (Broome 2004, pp. 94–95).

It could be objected that there are no good standards for evaluating the correctness of the previous judgments (Wallace 2002, p. 432), but this objection is unduly sceptical. Not all judgments about reasons can be accepted as equally correct. We do recognise certain ‘standards for arriving at conclusions about reasons’ as better than others (Scanlon 1998, p. 63). These standards are the same epistemic standards which we use to critically evaluate the credentials of our beliefs generally.

When we assess whether some intuitively plausible claim is true, we must consider whether we have clearly understood the concepts which compose the proposition which is expressed by that claim. We must also consider the epistemic origins of our initial intuition. Can the source of our intuition be found from undermining sources such as momentary emotions, distorting social factors or the like? In addition, we need to make sure that our intuition is stable under reflection and that it coheres with our other beliefs. Furthermore, we must consider whether other rational inquirers share our intuition (Sect. 5).

If we can in this way assess how strong objections different individuals can present to different sets of moral principles, then the information needed for considering which acts are forbidden by the contractualist principles is accessible. That information thus also satisfies the second requirement for making the relevant wrongness-claims superassertible. And, given that we do not have good reasons now to believe that this information would stop justifying assertions about wrongness in the future, we can conclude that information about being forbidden by the contractualist principles makes sentences about wrongness superassertible.

10 Conclusion

This means that we have all the required elements for a new restricted contractualist form of constructivism.¹⁸ We first need to understand reasons in a non-constructivist way. We can then think that information about reasons to reject different sets of principles makes claims about right and wrong superassertible. This allows us to argue that, by default, the truth of these claims is realized by their superassertibility, and therefore the property of wrongness is a projection from the minimally true statements about wrongness.¹⁹ This means that there is finally a distinct coherent constructivist metaethical proposal on the table.