Timmons initiates his presentation of his positive view in a thorough chapter on contextualist moral semantics, first providing a general introduction to the semantic program. In contrast to the correspondence view of truth (the view of realists), for a sentence to be true according to contextual semantics is for it to have correct assertibility, where correct assertibility is the complex interaction of the norms and practices that govern a certain type of discourse and the world. Norms and practices vary between discourses, such that some discourses may require a tight connection between norms of use and the actual world (scientific discourse), while other discourses do not require such a strict connection (art discourse does not require the actual existence of such things as symphonies for our discussions about Beethoven to have correct assertability). An important point to note here is that we can take different perspectives on the same object through different types of discourse and we can and do make this switch in perspective – as music enthusiasts we may speak of symphonies as if they exist, but as philosophers we may speak of symphonies as not truly existing. Contextual semantics allows that our talk about Beethoven’s symphonies to be true without symphony types existing out in the world.
The moral realist could make use of contextualist semantics by arguing that the connection between semantic norms and the world is tight, thus giving objective credence to the semantic content of statements, but Timmons argues that his Moral Twin Earth thought experiment shows that there is no such strict regulation of word by world. The other descriptivist route is relativism, which when making use of contextualism, holds that the semantic norms align with the world to make statements true or false, relative to that speaker’s outlook. Timmons doesn’t think there is much plausibility to this view, as it goes against the categoricalness of moral judgments, so, having exhausted descriptivist options for contextual semantics we move to non-descriptivist options.
However, Timmons wants to pursue a new non-descriptivist option, given his focus on moral judgments as being categorical, and thus indicative of judgments being assertions, which seems to be a feature only accommodated by descriptivists. Timmons makes a brave proposal by challenging the common assumption that all assertions are descriptive and not evaluative, an assumption that led those moved by the categoricalness of moral judgments to be descriptivists, and those wanting a naturalistic moral theory to deny that the surface features of moral judgments are their most real features – a move taken by non-descriptivists. For Timmons, we should embrace the idea that there are two different types of assertions, neither more basic than the other, descriptive and evaluative assertions, where evaluative assertions “do not describe the world or purport to describe it, but play a fundamentally different sort of role in human discourse – broadly speaking, an action-guiding role” (129). Crucially, if moral statements are assertions then they are truth-apt, but since they are non-descriptive they make minimal demands of what must be in the world for those statements to be true. This is the minimalistic spirit of Timmons’ moral semantics, which he embraces in regards to the minimalistic demands moral discourse makes of the world, and in regards to the limited expectations of analysis of moral statements, as Timmons writes that “one is not going to be able to provide such an analysis, and so the appropriate response to questions of the form ‘what is the content of p?’ is simply the minimalist response: p” (133). The turn to minimalism is motivated, for Timmons, by the general failure of reductive irrealism (specifically in regards to moral error).
Given this minimalist bent, insight into morality will not be gained by the analysis of moral terms into non-moral terms, but by looking at the point and purpose of moral discourse, leading Timmons to look into moral judgments. Moral judgments seem to involve bringing ones’ moral outlook into play at the situation at hand, where a moral outlook includes sensitivity to morally relevant factors, emotional responses to evaluations, familiarity with particular moral examples of general moral types, knowledge of frequently relevant moral features of acts, and basic patterns of moral reasoning. Taking a moral stance on an issue, then, is coming to a moral judgment by thinking about the matter at hand in a morally engaged way, that is, by employing one’s moral outlook.
Timmons considers some important features of moral judgments, which include being reasons-based, categorical, having cognitive content (we judge that murder is wrong), and being commitments to action. This last feature takes Timmons into the internalist/externalist discussion of moral motivation, and Timmons characterizes his preferred version of internalism as the theory that moral judgments are sufficient in action guidance through motivation, but have defeasible conditions due to other psychological causal tendencies. Thus, “a moral judgment, then, is a certain contentful psychological state that is implicated in a web of defeasible psychological tendencies aimed primarily at choice and guidance of action, not representation” (143). This allows us to put into focus exactly what is asserted in a moral statement, per Timmons’ “assertoric non-descriptivism”: the moral stance held by the agent is asserted in a moral statement, that is, a moral statement asserts an evaluative stance.
In contrast to this morally engaged perspective of employing a moral outlook there is the option of taking a morally detached perspective, where “ ‘true’ and ‘false’ are used in such a way that they are only applicable to statements governed by tight semantic norms” (151), and given that there are no moral properties and facts to create tight relations between word and world, from a morally detached perspective moral statements are neither true nor false. The same is true of moral properties and facts themselves, in the sense that though they cannot be said to exist from a morally detached perspective, from a morally engaged perspective we can make reference to them, because loose semantic norms and practices do allow for them.
Timmons goes to great lengths to emphasize that his view is not relativism because his view does not equate correct assertability with truth, which seems sufficient to handle that objection. To me, his view seems closer to an error theory, but instead of the claim being that moral statements assume the existence of, and refer to, sui generis moral properties that don’t actually exist, moral statements assert evaluations of actions, persons and institutions without referring to properties at all.
Timmons is attempting to craft a metaethical theory that makes sense of two theses: that moral statements are assertions, and that moral statements are action-guiding. Traditionally, non-descriptivists rejected the former to embrace the latter, while non-naturalists accepted both theses, holding that moral assertions describe non-natural properties that are intrinsically motivating and thus action guiding. We must understand Timmons as an irrealist who, due to skepticism about non-naturalism, wants a different way of accepting both premises – that moral statements are evaluative assertions that have minimalist truth conditions from a morally engaged perspective, but none from a morally detached perspective.
The problem is that Timmons doesn’t tell us much about how evaluations can be assertions, just that evaluative sentences do have assertoric content because they take the form of an assertion. But the move from assertoric form to content is under-motivated here. What is more, while Timmons thinks that it should be unproblematic that moral statements have minimalist truth conditions because they are assertions, we should be skeptical that a statement being an assertion automatically sets it up for minimal truth conditions, as it seems that is the fact that a sentence is descriptive that sets it up for minimal truth conditions, not its assertoric nature. So even if we grant Timmons the point that assertions can be broken down into two unique classes (descriptive and evaluative) it hardly follows from this bifurcation that evaluations qualify for minimalism.
But granting Timmons the possibility of minimalist truth conditions for evaluative assertions only reveals more troubles for the idea of minimal truth conditions for non-descriptive assertions – it’s difficult to see how there could be minimalistic truth conditions for moral statements made from a morally engaged perspective without there also being the same sorts of minimalist truth conditions for moral statements made from a morally detached perspective, if moral statements only aim to assert the moral stance of the speaker and not describe the world. What this suggests is that the difference in moral perspective (detached or engaged) is central to the different possibilities for truth conditions.
Now, the morally engaged perspective is the perspective from one’s moral outlook, which is the point of view where we are sensitive to what we take to be morally relevant factors. If I understand Timmons correctly, moral statements have minimalist truth conditions because they are made from this moral outlook, where there are morally relevant factors in states of affairs, but moral statements do not have minimalist truth conditions when not made from this moral outlook, because there are no morally relevant factors to be considered – such factors do not exist. Now, it seems possible that we can “turn off” our sensitivity to morally relevant factors, and thus step out of the morally engaged perspective and into the morally detached perspective, but from there it doesn’t seem possible to make moral statements at all – if we’re not thinking within the boundaries of the concepts of morality then it seems impossible to make statements using those moral concepts. Timmons was trying to provide a reason for why moral statements have minimalist truth conditions in the engaged perspective but not in the detached perspective, which requires that we be able to make moral statements in the detached perspective. If the proposed solution entails that we cannot make moral statements from the detached perspective then it is no solution at all. So, the reason why statements from the engaged perspective qualify for minimalist truth conditions while statements from the detached perspective do not must have to do with something else that is different about the perspectives.
I think this general point can be pushed for all reasons a Timmons supporter can offer – the reason for the difference in minimalist truth-aptness cannot be because of differences in the statements themselves, but because of the intentions behind the statements. Given this, it seems the only move is to admit that the difference in mindset from the detached and engaged perspectives is that from the detached perspective moral statements are not meant as assertions, whereas they are meant as assertions from the engaged perspective. This would allow Timmons to maintain that moral statements have minimalistic truth conditions from the engaged perspective given their context as being placed in an assertoric discourse with loose semantic constraints, but as lacking minimalistic truth conditions from the detached perspective because such moral statements are not assertions at all, they are just emotings or prescriptions, in the expressivist/non-cognitivist tradition, lacking truth conditions.
Perhaps it is quite obvious that this is no solution either. Aside from the fact that it seems quite dubitable that a statement is or is not an assertion depending on how we mean it (given that what makes a statement an assertion is its grammatical form), we would be committed not just to the strange thesis that moral statements are evaluative assertions, but that there can also be evaluative non-assertions, and to providing non-circular definitions of each evaluation type that would clearly map these differences. So even if we grant Timmons the existence of evaluative assertions, we still need to maintain the existence of evaluative non-assertions for his view to work. Given the dubiousness of there being both assertoric and non-assertoric evaluations, and the plausibility of non-assertoric evaluations (as manifest by the popularity of expressivism for moral semantics), we would be well advised to reject the notion of assertoric evaluations, and Timmons contextualist semantics with it.