2. Streumer’s super-quasi-realism

Before turning to our arguments, let us situate the debate and Streumer’s take on it. Consider a person who endorses the claim, ‘Genocide is wrong’. According to cognitivists, the person then has a truth-apt belief which ascribes a normative property to genocide. Cognitivists who are also realists normally think that beliefs of this kind are sometimes true; that there is a fact of the matter whether genocide is wrong, and (if they are morally right-minded) that genocide has the property of wrongness.

Expressivists, on the other hand, take endorsement of the claim that genocide is wrong to express a non-cognitive attitude (such as approval or disapproval).Footnote ² A quasi-realist is an expressivist who adopts a minimalist view about a variety of philosophically contested notions, such as ‘truth’, ‘property’, ‘representation’ and ‘fact’.Footnote ³ On this kind of analysis, all there is to these contested notions can be expressed by biconditionals including (but not exhausted by) the following: ‘“Genocide is wrong” is true if and only if genocide is wrong.’ On this kind of analysis, any proposition which is truth-apt will express a belief; any true proposition will be suitably factive, and so on for every notion in the relevantly contested domain, suitably interpreted. In this way, the quasi-realist claims to ‘earn the right’ to say all, or most, of what the realist says (see e.g. Blackburn Reference Blackburn1993).

So-called robust realists, however, object that something crucial is missing from the quasi-realist picture, namely how reality – the facts, events and states-of-affairs in the world – determines, fixes, or grounds which judgements are true or false (Streumer Reference Streumer2024: 930–931). In response, Streumer puts forward the strategy of what he calls ‘super-spreading the word’ (cf. Blackburn Reference Blackburn1985), according to which the interpretation of ground-level target claims as expressing non-cognitive attitudes is extended to the interpretation of all meta-level claims as well, resulting in a position he calls super-quasi-realism. Hence, just as ‘genocide is wrong’ is interpreted as (say) disapproval of genocide, claims about truth, fact, reference and properties are given the same non-cognitive analysis. Consequently, a (robust realist) meta-claim such as

is interpreted as expressing a non-cognitive attitude such as:

When, Streumer imagines, the robust realist continues her objection against quasi-realism with a claim such as:

the super-quasi-realist will ‘simply keep going’ (Streumer Reference Streumer2024: 935) by applying the same analysis to this claim as well, thus interpreting (2) as:

The super-quasi-realist makes this interpretative move in response to anything the robust realist claims to be missing from quasi-realism, including:

which the super-quasi realist takes to express:

And so on, apparently without end. If the super-quasi realist thus goes through the process of re-interpreting and endorsing all robust realist claims along quasi-realist lines, Streumer argues, the quasi-realist has turned into a robust realist.Footnote ⁵ Streumer concludes that ‘those who try to endorse super-quasi-realism thereby become robust realists’ which shows ‘that we cannot believe super-quasi-realism, since trying to endorse this view makes us believe a different view instead.’ (Streumer Reference Streumer2024: 939).

3. The triviality argument

Streumer’s argument shares with previous discussions of quasi-realism the claim that in its unrestricted form (what Streumer calls ‘super-spreading the word’), quasi-realism is indistinguishable from realism. Streumer’s original addition to these discussions is the claim that it is impossible for us to believe quasi-realism so construed and that this could be evidence in its favour. This move is too quick. This is because Streumer’s argument operates with a ‘personalized’ understanding of quasi-realism as the project of a thinker who starts off as a committed expressivist and is then described as going through a process of endorsing an increasing range of realist-sounding claims until they end up in a position where what they say is indistinguishable from what realists say. Yet going through such a process is not the only way in which someone might come to believe quasi-realism. On the contrary, a thinker (such as the arbitrary reader of these words) could come to believe quasi-realism by coming to believe the following claim about the philosophically contested judgements in question: ‘The quasi-realist programme can successfully be executed with respect to all concepts in domain X.’ Moreover, in order to believe, or otherwise endorse, this isolated claim, the quasi-realist does not need to have gone through any super-quasi-realist process as characterised by Streumer. Nor does the belief in question have to be true, or even coherent. Hence it is possible to believe quasi-realism even in its super-spreadable form. The canonical evidence for the truth of quasi-realism, and which the quasi-realist can believe obtains, is that the quasi-realist programme can be comprehensively executed. Yet this is not a particularly exciting conclusion. Fortunately, embedded in Streumer’s argument is a more interesting claim about the consequences of actually embarking on the process of super-spreading the word.

It might be immediately objected that genuinely believing that the quasi-realist programme can successfully be executed is not the same thing as believing (a version of) quasi-realism; because believing quasi-realism also involves having beliefs about how the programme can be executed and, crucially, endorsing the results of this execution. In response, we suspect this is a ‘distinction without a difference’ (cf. Köhler Reference Köhler2021). However, let’s assume it is not. In this paper, we consider two ways of coming to believe quasi-realism. The first way, considered in this section, consists in coming to believe that the quasi-realist project can be successfully executed. The second way, considered in the following sections, consists in executing a process of ‘super-spreading the word’ the end result of which is that one comes to endorse the results of this execution. If someone wants to insist that the first way is not a way to come to believe quasi-realism ‘proper’, that’s fine. They can give this another name. For we will shortly go on to consider the second way, which leads to the same conclusion. So either way, our overall argument will stand.

4. The basic argument

According to Streumer, our inability to believe quasi-realism is a direct consequence of the successful execution of the quasi-realist research programme, according to which the quasi-realist can perfectly mimic the realist in virtue of giving a minimalist interpretation of all putatively meta-level terms, such as ‘truth’, ‘fact’, ‘reality', and so on. According to Streumer, the comprehensive execution of the quasi-realist programme in personalized form has the consequence that the quasi-realist ends up making claims indistinguishable from realism; thereby ending up believing realism; consequently excluding themselves from believing quasi-realism which, by hypothesis, the quasi-realist programme was designed to vindicate. Paradoxically, therefore, a super-quasi-realist cannot believe quasi-realism.Footnote ⁶ Yet, as Streumer has previously argued in the case of error theory, ‘just as a view can be true if we do not believe it, a view can also be true if we cannot believe it’ (Streumer Reference Streumer2024: 941; Reference Streumer2017). From which he draws the conclusion that since super-quasi-realism is the best version of expressivism and expressivism could be true, quasi-realism could be true, even if we are unable to believe quasi-realism.

This argument is unsound, even as interpreted in its personalized form. Now, either quasi-realism is true, or it is not. If quasi-realism is false, then the quasi-realist programme cannot be comprehensively executed as embarked on by the super-quasi-realist. Hence, the only possibility we need to consider is that quasi-realism is true.

Suppose, therefore, that quasi-realism is true, and that the super-quasi-realist programme can be comprehensively executed. Yet in that case, pace Streumer, the content of the philosophically contested judgements endorsed by the quasi-realist who comprehensively executes the quasi-realist programme is not what realists and other cognitivists interpret them to mean, but rather something else, which is more accurately explicated as the expression of attitudes, as opposed to cognitive states, at least as this idea of a ‘cognitive state’ is interpreted by realists and other cognitivist theorists. Of course, the nature and status of these judgements as ‘cognitive’ or otherwise need not be phenomenologically transparent to the person making them, quasi-realist or otherwise. But then, the nature and status of our judgements as cognitive or otherwise hardly are.

This direct implication of the success of the quasi-realist programme refutes Streumer’s argument in its original form. This is because the super-quasi-realist who ends up endorsing the relevant philosophically contested claims after the comprehensive execution of the quasi-realist programme has not, in fact, ended up believing realism as initially understood. Instead, they have ended up doing something subtly different; namely, to express a set of attitudes the precise content or nature of which they might, or might not, have a self-consciously reflective understanding. Even sophisticated meta-ethicists, like super-quasi-realists, can be mistaken about the correct interpretation of what they are doing. It follows that if the quasi-realist thinks they have ended up believing realism after successfully executing the quasi-realist programme, they are just mistaken about what they believe. This is the basic argument. Of course, the basic argument does not show that the quasi-realist programme can, in fact, be executed; whether in its super-quasi-realist or in some more restricted form. It is possible to remain neutral about whether the quasi-realist programme can be comprehensively executed while also rejecting Streumer’s conclusion, by arguing against that conclusion on the assumption that it can.

It might now be objected that the basic argument fails to take due account of the fact that the quasi-realist, in addition to endorsing expressivism and minimalism, is also a ‘noncognitivist’.Footnote ⁷ Clarity on this issue is obviously crucial for understanding what is at issue between Streumer’s interpretation of the quasi-realist project and our own. Yet as Streumer agrees, if the quasi-realist project is successful, then the quasi-realist will be able to say everything the realist can say, from which it follows that they will also be able to say that moral and other normative judgements express beliefs. Furthermore, and crucially for us, our contention is that the success of the quasi-realist’s deflationary ‘super-spreading’ project implies that there is no ‘underlying’ metaphysical sense in which they can be said to either be, or not to be, beliefs ‘really’. From which it follows that, pace Streumer, realism and quasi-realism are not to be understood as two ‘meta-level’ views with different ‘truth conditions,’ as opposed to two conflicting first-order normative claims, commitments, positions, views, or whatnot. We elaborate further on this crucial point in the final section of the paper.

5. The no response argument

In this section, we will explore whether a cogent response can be given on Streumer’s behalf to both the basic argument and the trivial argument, by focusing directly on the content of the super-quasi-realist’s commitments. As we rehearsed in Section 2 above, Streumer’s ‘super-quasi-realist’ is a quasi-realist who interprets all claims within and about the normative domain along the ‘non-cognitivist’ and ‘minimalist’ schema. This includes not only first-order moral claims, but also such apparently meta-level claims such as:

And so on, for every first-order or meta-level claim. As Streumer puts it, a super-quasi-realist is thus a figure who will ‘simply keep going’ (Streumer Reference Streumer2024: 935). But is there a way to ‘simply keep going’ that provides a way out for Streumer? We believe not. In what follows, we consider two options, each of which either fails to evade the basic argument or the trivial argument, or both.

On the first interpretation of this locution, which we call the process response, to ‘simply keep going’ is a reflective process that (a) adopts a non-cognitivist and minimalist interpretation of every claim one comes across within the relevant contested domain, and then (b) endorses every (now re-interpreted) claim that a robust realist would accept, and so on indefinitely. Hence, assuming that a robust realist would accept claims (4)–(6) above, the super-quasi-realist would endorse the corresponding non-cognitivist and minimalist variations on those claims, on the interpretative schema described in Section 2. So far so good, since this is in fact the type of claims that the traditional quasi-realist endorses, holding that the possibility of doing so (on a non-cognitivist and minimalist interpretation) speaks in favour of quasi-realism.

Yet the super-quasi-realist then goes on to additionally endorse robust realist claims such as (2) and (3) in Section 2 above (repeated here for convenience).

And once this happens, i.e. once the quasi-realist not only interprets these claims in accordance with the non-cognitivist and minimalist interpretation but also endorses them, ‘the quasi-realist will have to stop interpreting [(2)] as an expression of a non-cognitive attitude’ as well as ‘stop taking minimalism to be true of truths about normativity’ (Streumer Reference Streumer2024: 937). Having started out as a quasi-realist, they have ended up as a robust realist.

To the extent that we are able to give a coherent interpretation to Streumer’s proposal here, we think this process response must be flawed. For on this interpretation, what Streumer has described as the successful execution of a theoretical program – a sophisticated version of quasi-realism – is in fact nothing more than a rough sketch of a change of mind with respect to a range of initial commitments of a certain kind. Moreover, when correctly interpreted, this change of mind on a range of initial commitments of that kind ends up defying the traditional quasi-realist’s theoretical understanding of the nature of the state of mind with which they started.Footnote ⁸ What we are supposed to imagine here is a theorist who starts off thinking and talking about the target domain in terms of expressing attitudes and a minimalist approach to the truth predicate, and then – at some later time – comes to think and talk about the target domain in terms of beliefs and a terminologically more liberal approach to the truth predicate. (And similarly for any other contested concept for which the super-quasi-realist purports to give a minimalist analysis.)

Streumer is obviously right that such a change of mind could take place and arguably occasionally does take place (for example in the context of philosophical discussion). Yet the mere fact that such a change of mind either can or does take place does not show that the change of mind in question is correctly characterized in terms of having as its terminus a theoretical position (i.e. an inflated version of quasi-realism) being true even though we cannot believe it because we have come to believe another theoretical position (i.e. robust realism as traditionally understood) instead. On the contrary, insofar as we are able to give a coherent interpretation to Streumer’s proposal here, what the process response on behalf of super-quasi-realism does is rule out as ill-formed the central premise needed in Streumer’s argument; namely that super-quasi-realism and robust realism are different views.

One crucial claim that Streumer’s conclusion trades on is the idea that in trying to believe super-quasi-realism we are caused to believe robust realism instead (Streumer Reference Streumer2024: 940). To make any coherent sense of this idea, we must assume that this transition can take place because super-quasi-realism and robust realism are distinct meta-level views with distinct truth-conditions (ibid). Yet insofar as we are able to give a coherent interpretation to the super-quasi-realist project, this is simply not true. For on this interpretation, super-quasi-realism is not a view according to which non-cognitivism and minimalism are true, and which can be consistently distinguished from robust realism, according to which non-cognitivism and minimalism are false. Instead, to the extent that the super-quasi-realist project succeeds, the denial of non-cognitivism and minimalism is itself part of that view. Of course, there is still one sense in which robust realism and super-quasi-realism can be said to differ. This is the sense in which the formulation of robust realism per se does not itself include a story about how a theorist is supposed to be able to reach that view, whereas offering such a story is an essential part of the formulation of super-quasi-realism on Streumer’s ‘super-spreading’ story. But all that goes to show is that super-quasi-realism is intrinsically indistinguishable from robust moral realism, except that it has a peculiar ‘how I changed my mind’ story attached.Footnote ⁹

The natural alternative to avoid this problematic outcome is to make only the analysis of the robust realist claim part of thesuper-quasi-realist account. In other words, on this second interpretation, to ‘simply keep going’ is to pursue a theoretical project of continuously applying a non-cognitive-plus-minimalist analysis of all claims within the relevantly contested domain, without necessarily endorsing every given claim thus analysed. Let’s call this the theory-driven response. On this analysis, the project of interpreting ground-level target claims (e.g. genocide is wrong) as expressing non-cognitive attitudes (e.g. disapproval of genocide) is simply extended to the interpretation of meta-level claims about the nature and status of ground-level claims; (e.g. claims about truth, fact, reference and properties involving the wrongness of genocide), and so on, indefinitely. Here, the only constraint on which of the claims so interpreted to endorse is that they must be coherent with the already endorsed theoretical commitments, i.e. with the interpretative scheme and minimalism about the contested concepts.

From a quasi-realist perspective, the theory-driven response might well be perfectly reasonable. Yet it provides no support for Streumer’s argument. Instead, it makes the dynamic process in which the quasi-realist is supposed to seamlessly end up as a robust realist look utterly mysterious. For what exactly, as implied by the content of her theoretical commitments, would rationally force a quasi-realist to go from analysing a given claim in a philosophically contested category as the expression of a non-cognitive attitude to actually endorse that claim? We definitely do not expect the quasi-realist to do so for any arbitrary first-order claim (e.g. ‘Genocide is morally right’). Why so for any particular meta-claim?

Suppose, for example, that the quasi-realist gives an analysis of the aforementioned claim,

as the expression of a non-cognitive attitude. By the same token, she will be committed to analyse the following claim as also being the expression of a non-cognitive attitude:

Although we do expect that most quasi-realists would in most cases endorse the former rather than the latter claim, it clearly does not follow from quasi-realism that she is committed to endorse either one of these two claims (let alone both).

That leaves us with the basic issue with which we started, and what is supposed to be in dispute between the super-quasi-realist and the robust realist, namely the commitment to quasi-realism itself. This is clearly a claim that any quasi-realist must endorse on pains of inconsistency. Yet that is a commitment we can make sense of by forming a theoretical belief that the quasi-realist programme can be successfully executed. That’s just the trivial argument once again.

One might perhaps object to our description of the theoretical alternatives that the metaethical view a person holds should be taken to depend on the sentences they endorse, independently of what those sentences in fact mean. On this view, it is arguably not true that a person undertaking the quasi-realist project ends up endorsing a set of claims whose significance they do not grasp, but rather that they are endorsing a form of robust realism – the very position they began by rejecting.Footnote ¹⁰ In response, we think this objection is committed to a highly implausible view of the content of our beliefs (involving metaethical claims or otherwise). What view a person ‘holds’ (about metaethical matters or otherwise) is generally a matter of what they believe. What someone believes when endorsing a sentence generally depends on what that sentence means. What a sentence means is not generally (and certainly not in the cases at issue here) self-intimating, but a matter of how the sentence in question is correctly interpreted in the relevant context. As all expressivists (and not only expressivists) should agree, how the use of moral and other normative sentences is correctly interpreted is not a simple function of how the people who use those sentences think they are correctly interpreted (if not, how could the truth of a metaethical theory come as a surprise?) To think otherwise would not only undermine the argument in this paper, but the entire debate between quasi-realists and their critics to which it is addressed.Footnote ¹¹

6. The diagnosis

There is a different way in which it would, indeed, be impossible to believe quasi-realism if the quasi-realist programme can be comprehensively executed. This would be because the comprehensive execution of the quasi-realist programme (or ‘superspreading of the word’) implies the impossibility of drawing a coherent distinction between realism and quasi-realism by means of which a belief in either view can coherently be attributed either to the quasi-realist or to anyone else (cf. Akhlaghi Reference Akhlaghi2023). If this were so, then it would be ‘categorically’ mistaken (to adapt a term from Gilbert Ryle) to conclude that quasi-realism is true even though we are unable to believe it. What we should conclude instead is that there is no coherent content to be assigned to the idea of ‘realism’ for ‘quasi-realism’ to be quasi about. Instead, what the comprehensive execution of the quasi-realist programme would show is that we are unable to believe quasi-realism because neither quasi-realism nor realism is true as those views have traditionally been understood. If so, what the failure of Streumer’s argument teaches us is that we should reject the traditional distinction between quasi-realism and realism as based on a mistake.

Following through on the implications of accepting this assumption, what would we actually end up believing at the end of the comprehensive execution of the quasi-realist programme? One obvious answer that should be immediately recognizable to quasi-realists now presents itself. If we take the general strategy embodied in the quasi-realist programme seriously, what we are most likely to take ourselves as having ended up with at the end of the comprehensive execution of the quasi-realist programme is our endorsement of some unusually abstract, and substantive first-order claim. What this claim would express is a distinctive and self-consciously reflective attitude towards one’s own attitudes. Thus, one might employ the vocabulary of ‘truth’ or ‘reality’ to give expression to the seriousness of one’s moral commitments, or one’s disapproval of a moral sensibility that is disordered, unreflective or too easily hostage to the contingency of circumstance. (This is partly what Blackburn, Reference Blackburn2011, has in mind when he speaks of Ramsey’s ladder being ‘flat’.) Alternatively, one might employ the vocabulary of ‘truth’ and ‘reality’ to express a commitment to the value of pursuing agreement on a single answer or policy in the face of conversational friction or resistance. (This is partly what Blackburn, Reference Blackburn1985, has in mind by saying that we are constrained to argue as though the truth is ‘single’).Footnote ¹² Of course, in some contexts the expression of such claims would be surprising, misleading, verbose, or otherwise unwelcome. Yet in other contexts, for example a situation in which conflicting interlocutors are observed to be prematurely giving up on moral argument and discussion in favour of the ‘easy’ option thinking ‘there is no right answer,’ the employment of some unusually abstract and substantive first-order claim could play an informative and corrective conversational role.

We suggest that the underlying theoretical view that emerges from this discussion is one that is most charitably interpreted as a version of pragmatism, along the lines of the position recently defended under the label ‘global expressivism’ (see e.g. Price Reference Price2011). The correct conclusion to draw from this is not, however, that quasi-realism is true although we are unable to believe it. On the contrary, if we are unable to believe in quasi-realism, we are unable to believe it because it is not true. And the reason quasi-realism is not true is that quasi-realists have misunderstood the implications of the explanatory programme on which they are embarked.

Let’s assume for the sake of argument that this is right. What should we then say about what’s been going on in the philosophical dispute between quasi-realists (‘super’ or otherwise) and their critics? Shall we now be forced to conclude that this dispute, as well as other philosophical disputes like it, is either empty, redundant, or nonsensical? We think not.

True, it might be tempting to interpret the minimalist aspect of the quasi-realist programme as implying some reductive redundancy thesis about the meta-level vocabulary that is purportedly deflated on this programme (e.g. talk about facts, objects, properties, reality, etc.). After all, another name for a version of this thesis in the case of the philosophy of truth is ‘the redundancy theory’. Yet as various proponents of minimalist views of truth and comparatively contested concepts have pointed out, this is a mistake (see e.g. Price Reference Price2011). Along similar lines, the comprehensive execution of the quasi-realist programme does not imply a redundancy view about the meta-level vocabulary that is subject to deflation. Although the comprehensive execution of the quasi-realist programme implies that the meta-level vocabulary deflated is extensionally equivalent to the substantial first-order vocabulary in question, it does not follow that the two sets of vocabularies are intensionally, or informationally, equivalent. In other words, although the comprehensive implementation of the quasi-realist programme implies that when employing, alternatively, the substantial ground-level vocabulary and the meta-level vocabulary we are, in some sense, talking about the same things, we are obviously not talking about the same things in the same way. The choice between one set of vocabularies rather than another is therefore a choice between which terms to use when talking or thinking about that same thing in a given dialectical, explanatory, or otherwise informational context. This choice will depend on what is at stake in that context, including what our interests are. It is therefore a choice that depends on what matters in that context, and so on what values are at stake. The choice between employing some piece of substantial, or ground-level, vocabulary, some piece of theoretical, or meta-level vocabulary, or some combination of both, is therefore in part a substantial, or ground-level, question of which terms to employ in addressing whichever issues we are confronted with.

One implication of this response is that we cannot infer from the extensional equivalence between the substantial ground-level and the relevant meta-level vocabulary that the two sets of vocabularies are interchangeable or otherwise equivalent for any given practical or theoretical purpose. On the contrary, it is plausible to assume that for many day-to-day purposes, whether practical or theoretical, the substantial ground-level vocabulary will have default priority.

To take just one example: unless we have some grip on the question of what things are substantially good and bad (perhaps happiness and suffering) we are unlikely to have any grip on the question of what it means for some things to be good or bad in themselves, intrinsically, or objectively. This claim is consistent with the further claim that for certain purposes the employment of some meta-level vocabulary is preferable (or even indispensable) when either abstracting from, comparing, or systematizing judgements made by the employment of the relevant substantial ground-level vocabulary. For example, unless we employ at least some meta-level terms we are unlikely to get any grip on the idea that there are some things that are good or bad (perhaps happiness or suffering) wherever we find them; regardless of context; no matter what; independently of what anyone thinks or feels about it, or objectively, and so on.

On the assumption that the quasi-realist programme can be comprehensively executed, the crucial misunderstanding to avoid is the thought that such an execution makes thinking about our target subject matter in meta-level terms redundant. Nor does it imply that whether we think about our target subject matter in substantially ground-level or in meta-level terms is a matter of indifference. The comprehensive execution of the quasi-realist programme is friendly to the pursuit of a wide range of questions about the nature and status of substantial ground-level claims that have traditionally been labelled ‘philosophical’, even if not as conventionally interpreted or implicitly assumed in Streumer’s argument and much of the literature to which he responds. The endorsement of a pragmatist approach to superspreading the word does not imply a crude form of philosophical quietism.