Wittgenstein, deflationism and moral entities

Abstract

This paper discusses the meta-ethical implications of Wittgenstein’s later moral philosophy. According to Lovibond and Brandhorst, Wittgenstein provided a novel conception of moral facts, properties and objects by adopting deflationism. Lovibond argues that Wittgenstein’s seamless conception of language together with his non-foundational epistemology and non-transcendent understanding of rationality involves a change of perspective towards a plausible and non-mystificatory moral realism. Meanwhile, Brandhorst argues that Wittgenstein’s provides a deflationist conception of moral truths from which we obtain a deflationist conception of moral facts. This paper argues, on the contrary, that the attribution of deflationism does not do justice to Wittgenstein’s later work. It is concluded, therefore, that the appeal to deflationism does not afford or substantiate the exegetical claims made by Lovibond and Brandhorst.

1 Introduction

During the past decades, philosophers have shown a growing interest in understanding Wittgenstein’s later moral philosophy and its implications for contemporary meta-ethical debates. Brandhorst and Lovibond have sought to develop a link between Wittgenstein’s later philosophy and deflationism to shed some light on the meta-ethical implications of his later moral philosophy. Specifically, they suggest that his embrace of fact, property and existence deflationism warrants the postulation of moral entities while avoiding the troubles that have traditionally plagued moral realism.

This paper critically examines their proposed interpretations by discussing the adequacy of attributing fact, property and existence deflationism to Wittgenstein. Section 2 offers an overview of deflationism and some proposals relevant to this paper. Section 3 outlines Lovibond’s and Brandhorst’s proposed interpretations of Wittgenstein’s later moral philosophy. The remainder of the paper discusses the adequacy of invoking Wittgenstein’s so-called fact, property and existence deflationism to postulate the existence of moral entities.

2 Deflationism

Deflationism is a philosophical program that comes in different shades and stripes. Despite admitting a great deal of variation, deflationists can be understood as sharing the following core negative claim: philosophers mistakenly assume that certain properties that are central to philosophy, e.g., truth, reference or existence, are genuine substantive properties that have a nature of the kind one might find out about, investigate and develop theories of by offering informative analysis and reductive generalizations of the form, e.g., x is true iff x corresponds to the facts, x refers to y iff x stands in relation R to y, or Ks exists iff Ks are mind-independent. By contrast, deflationists claim that these properties do not have a deep nature and hold that the relevant concept of truth, reference or existence can be grasped only by certain trivial platitudes that govern the concept.

This paper will not discuss deflationism understood as generic philosophical position acceptable of any notion or property, but rather five particular kinds of deflationism: truth deflationism, syntacticism, property deflationism, fact deflationsim and existence deflationism. Truth deflationism argues that the property of truth has no deep nature that can be informative analyzed in terms of, e.g., correspondence to reality. All that can be said about truth is a relatively trivial and uninformative platitude which captures the meaning of ‘truth’:

(TD): ‘p is true’ = p.

Syntacticism is a deflationary/minimalist conception of truth-aptness, according to which there are two minimal requirements for truth-conditionality: the sentence must be (i) meaningful and (ii) declarative/indicative in form (see Boghossian, 1990, pp. 163–164).

Property deflationism claims that properties do not have a substantive or deep nature that can be theorized about in terms of being mind-independent, physical, observable, and so on. Instead, the following equivalence schema exhausts the meaning of ‘property’:

(PD): The property of being P exists iff ‘P’ expresses (where ‘p’ is a meaningful predicate) (cf. Burgess, 2010, p. 444).

Thus, property deflationists ground facts about property existence in the facts about predicate reference. “Properties are just the shadows cast by meaningful predicates” (Burgess, 2010, p. 444).

Analogously, fact deflationism claims that facts do not have a substantive or deep nature that can be theorized about in terms of being mind-independent, physical, observable, and so on.¹ Instead, the following equivalence schema exhausts the meaning of ‘fact’:

(FD): The fact F exists iff ‘F’ expresses (where ‘F’ is a true meaningful proposition).

Thus, fact deflationists ground facts about fact existence in the facts about proposition reference. Facts are just the shadows cast only by true meaningful propositions.

Finally, existence deflationism is metaphysically deflationary about all entities, not just facts or properties (see Thomasson, 2014 for a detailed explanation).² “On this deflationary view, we should thus give up the search for some ‘criterion of existence’”, in terms of causal powers, mind-independence, physicality or observability, “telling us what it is for something to exist, just as semantic deflationists give up the search for the nature of reference, meaning or truth” (Thomasson, 2014, pp. 196–197).

Thus, arguments about whether or not controversial entities exist are brought into question, since they rest on the mistaken assumption that there is a substantive property of existence to be discovered and informatively analyzed in terms of having causal powers, being mind-independent, being physical, being observable and so on. It is no longer metaphysically controversial, then, to affirm the existence of certain entities (e.g., numbers, moral values, fictional characters) that have troubled ontologists during the past decades, as they can simply be inferred from undisputed claims. For instance, from the true propositions ‘the house is red’ or ‘five is odd’ it is possible to infer that houses (and redness) and the number five (and an odd thing) exist (cf. Thomasson, 2014, pp. 185–197).

3 Lovibond and Brandhorst on Wittgenstein, deflationism and meta-ethics

The attribution of deflationism to Wittgenstein is not a novel claim among Wittgenstein scholars. Many (see e.g., Baker & Hacker, 1980; Kripke, 1982; Blackburn, 2010; Horwich, 2012, 2016b) have identified connections between Wittgenstein and truth deflationism due to his use of the equivalence schema in PI §136.³ Furthermore, at first glance, the attribution of deflationism does not seem far-fetched, given Wittgenstein’s rejection of philosophical dogmatism and philosophical theories (see e.g., PI, Wittgenstein, 1953, §11, §38, §119, §§124–133).⁴ Brandhorst, Lovibond and, to an extent, Horwich have sought to exploit this link between Wittgenstein’s later philosophy and deflationism to shed some light on the meta-ethical implications of his later moral philosophy.

3.1 Lovibond

Lovibond (1983) suggests that Wittgenstein’s embrace of deflationism offers a new way of understanding moral realism.⁵ On her interpretation, Wittgenstein's later philosophy propounds a seamless or homogeneous conception of language, “free from invidious comparisons between different regions of discourse, or (relatedly) between different aspects of mental activity” (Lovibond, 1983, p. 25). On this view all sentences which are (i) meaningful and (ii) qualify by grammatical standards as propositions, i.e., they are declarative/indicative in form, are propositions that must be treated as descriptive, fact-stating and truth-apt (Lovibond, 1983, pp. 25–27). Wittgenstein, consequently, embodies syntacticism.

All regions of assertoric discourse, irrespective of content and subject-matter, embody the same relation between language and reality: the metaphysically neutral idea of ‘talking about something’ (Lovibond, 1983, p. 26). The homogeneity of assertoric discourse, then, is also affirmed at a metaphysical level: the level where empiricists draw a line between fact and value. Wittgenstein’s seamless conception of language denies any metaphysical role to the idea of ‘reality’ to prevent the moral anti-realist from using the notion ‘reality’ as a peg on which to hang their special treatment of moral statements.

Value is thus reabsorbed into the real world from which the anti-realists and non-cognitivists expelled it. “'Fact' and 'value' (in the metaphysical sense of those words) coalesce -and assertoric discourse is now seen to accommodate both impartially” (Lovibond, 1983, p. 27). Reference to an objective reality is now a target hit by all sentences that display the appropriate grammatical syntax. The only way in which declarative/indicative sentences can fail to describe reality is by not being true.

In other words: Wittgenstein embodies a deflationary conception of existence and, by extension, facts and properties (Lovibond seemingly suggests that this is a natural consequence of Wittgenstein’s syntacticism). The existence of moral facts and properties is simply inferred from true propositions. For instance, from the true moral proposition ‘John is good’ it is possible to infer that John, the property of goodness and the fact that John exhibits the property of goodness exist.

According to Lovibond, Wittgenstein’s remarks on language and philosophical method also lead to “a non-foundational epistemology which displays the notions of objectivity (sound judgement) and rationality (valid reasoning) as grounded in consensus” (Lovibond, 1983, p. 40). This “makes it possible again to take a rationalist view of morals and politics, without having to revert to a pre-scientific attitude of mind in order to do so” (Lovibond, 1983, pp. 45–46).

Wittgenstein’s seamless conception of language in conjunction with this non-foundational epistemology and non-transcendent understanding of rationality changes our perspective towards a plausible and non-mystificatory moral realism. “Far from discrediting the critical concepts, allows them to penetrate again into the various regions of discourse from which empiricism has banished them” (Lovibond, 1983, pp. 45–46). Wittgenstein provides a metaphysically unexciting way of talking about facts, properties, truth and objectivity in ethics.

3.2 Brandhorst

Brandhorst (2015, 2017) has argued that Wittgenstein’s embrace of deflationism offers a new way of understanding moral facts, which differs from, and avoids the objectivist commitments of, moral realism. Brandhorst's proposed interpretation stems from an attempt to make sense of the following quote from Wittgenstein’s discussions on ethics with Rush Rhees:

The way in which some reality corresponds—or conflicts—with a physical theory has no counterpart here [in ethics]” (Rhees, 1965, p. 24, my brackets).

On Brandhorst’s interpretation Wittgenstein is not discarding the idea of ethical facts. Instead, he is deflating this notion in order to provide a view apart from those ‘realist’ theories on ethics that mistakenly attempt to establish substantial analogies between ethics and natural sciences.

The basis of Brandhorst’s proposed interpretation is Wittgenstein’s embodiment of a deflationist conception of truth.⁶ Wittgenstein, in his conversations with Rush Rhees (1965, p. 24), deflates the notion of ethical truth “by tying it to a subjective, personal point of view” (Brandhorst, 2015, p. 247). His endorsement of truth deflationism is epitomized by his use of the equivalence schema: “remember that ‘p is true’ means simply ‘p’” (Rhees, 1965, p. 24). “This is consistent with other discussions of truth in the later work, and it embodies what is now known as a deflationary conception of truth” (Brandhorst, 2015, p. 231; see PI, §§136–137; Wittgenstein et al., 2015, p. 29; RFM, Wittgenstein, 1983, §117; for other examples of Wittgenstein’s use of this equivalence schema).

Brandhorst explains that with this deflationary conception of truth Wittgenstein obtains a deflationary conception of facts. “Instead of saying that p is true, we can also say that it is a fact that p, the concepts ‘truth’ and ‘fact’ were made for each other” (Brandhorst, 2015, p. 232) and, thereby, obtain the following equivalence schema:

(BFD): ‘p is fact’ means simply ‘p’.

Unfortunately, Brandhorst’s characterization of fact deflationism is lack-luster. The equivalence schema BFD specifies the meaning of ‘fact’ when used as a logico-linguistic tool in natural language. In other words: this schema delivers a semantic conclusion about ‘fact’, not an ontological conclusion about facts. Thus, it does not deliver the ontological conclusion Brandhorst desires, i.e., that Wittgenstein posits the existence of (deflated) moral facts. To amend this shortcoming I will outline the remainder of Brandhorst’s proposed interpretation by resorting to the deflationist conception of facts presented in Sect. 2.

The point Brandhorst is trying to make by invoking fact deflationism is thus: when we state something is the case we do not “stop anywhere short of the fact” (PI, §95). For instance, the true assertion of an ethical proposition does not stop anywhere short of a moral fact. Wittgenstein, then, has abandoned his earlier views about the non-existence of moral facts by recognizing that talk of facts has no metaphysical depth and that their existence can be simply inferred from true moral propositions (Brandhorst, 2015, pp. 232–233). “If there are truths, there are facts” (Brandhorst, 2015, p. 233).

Brandhorst adduces further evidence for his interpretation in Wittgenstein’s abandonment of his earlier idea that there are no mathematical facts. In mathematics expressions like “2 + 2 = 4” can be used to express propositions that are either true or false.⁷ When true, these propositions do not stop anywhere short of the (mathematical) fact. “In this respect, ethical, logical and mathematical sentences are on a par with ‘snow is white’ or ‘it is raining’” (Brandhorst, 2015, p. 233).

Brandhorst, in consequence, interprets Wittgenstein as propounding a deflationary conception of facts, according to which:

(FD): The fact F exists iff ‘F’ expresses (where ‘F’ is a true meaningful proposition).

Thus, from true moral and mathematical propositions it is possible to infer the existence of moral and mathematical facts, albeit these concessions have no metaphysical depth since they are understood as facts in a thin, attenuated sense.

Brandhorst’s work would have also substantially benefited from including a discussion about property and existence deflationism. Ontological debates in meta-ethics often focus on whether moral values are to be understood as properties/qualities or not, as both Wittgenstein (see MWL, Wittgenstein, 2016, pp. 318–319, 332–333; AWL, Wittgenstein, 1979, §§31–32) and Brandhorst (2015, p. 246) recognize.

In addition, Brandhorst does not limit his ontological claims about mathematics and ethics to the existence of facts. He also discusses the existence (in a deflationary sense) of objects (Brandhorst, 2015, p. 235, 239–240), properties (Brandhorst, 2015, p. 235, 245), numbers (Brandhorst, 2015, pp. 239–240), moral values (Brandhorst, 2015, p. 245) and so on. He goes as far as stating that “a reality corresponds not only to propositions if they are true, but also to words such as ‘two’, ‘red’, ‘rain’ or ‘perhaps’” (Brandhorst’s, 2015, p. 243). Fact deflationism alone cannot deliver these ontological claims, thus Brandhorst must interpret Wittgenstein as an existence and property deflationist to uphold them.

Having detailed the ontological implications of Wittgenstein’s deflationism, Brandhorst returns to the quote presented at the start of this section, which is the focal point of his paper. On Brandhorst’s interpretation, Wittgenstein is suggesting that both moral judgments and physical judgments “can be described as judgments of fact, and the interesting further question is how reference to some ‘reality’ enters the picture. This marks the place […] where his deflationist view of ‘truth’ and ‘fact’ comes to the fore” (Brandhorst, 2015, p. 234).⁸

Physical judgments (or judgments of fact) refer to facts, in some thick sense of ‘fact’, and are objectively true or false in virtue of the objective and mind-independent facts to which they correspond (Brandhorst, 2015, p. 236). For instance, suppose a physical theory claims that gold has better conductivity than silver. If this judgment is true it is because it represents the way things really are in the world with respect to conductivity. Reality can be measured and modeled in the appropriate ways to settle its truth or falsity.

Meanwhile, moral judgments refer to facts, in some thin and deflated sense of ‘fact’. Unlike physical judgments, there is no mind-independent and objective reality that substantiates the objective truth of a moral judgment by being just as the judgment says it is. Instead, moral judgments are true or false in light of our personal commitments and attitudes. Thus, while truths and facts do have a rightful place in ethics, “they come into the picture in an entirely different way than many traditional ethical theories encourage us to think they do” (Brandhorst, 2015, p. 245). Moral realists get it wrong from the outset by modeling moral truths and facts on physical ones. Wittgenstein’s target in his conversations with Rhees is not the idea of ethical facts or truths, but rather the claim to objectivity that characterizes moral realism.⁹

3.3 Horwich

Although Horwich does not discuss Wittgenstein’s later moral philosophy in detail, he does provide a brief investigation of Wittgenstein’s later views on goodness which is sympathetic towards some of the exegetical claims propounded by Lovibond and Brandhorst. Horwich (2012, vii, pp. 60–62) argues that Wittgenstein’s meta-philosophical pronouncements suggest that he defends a deflationary meta-philosophical point of view as the foundation of his treatment of specific issues, such as goodness. This deflationary point of view comes to the fore when we examine the function of declarative moral sentences and moral predicates.

On Horwich’s (2012, p. 60) interpretation Wittgenstein is committed to syntacticism, the deflationist view that all meaningful declarative sentences are fact-stating and truth-apt in virtue of their surface form.¹⁰ Accordingly, declarative moral sentences function logically like other declarative sentences: they are used to relay truthful information about how the world is. Thus, we can properly speak about moral facts and statements and beliefs about the existence of these facts. Likewise, Horwich (2012, p. 60) explains that for Wittgenstein ‘good’ (and other normative terms) “function logically like other predicates (in obeying the predicate calculus)—so that we can properly speak of the property of being good, and of statements and beliefs about which things possess that property”.

Horwich, then, invokes Wittgenstein’s syntacticism to substantiate the claim that moral sentences and predicates are in the business of talking about moral facts and properties respectively. Unfortunately, due to the scarcity of Horwich’s remarks it is unclear whether he interprets Wittgenstein as a deflationist about facts, properties or existence. Although Horwich’s discusses fact and property deflationism elsewhere (see Horwich, 2010, pp. 281–294; 2012, pp. 162–163; 2016a, p. 142), there is insufficient textual evidence to make an informed decision regarding whether he attributes these philosophical programs to Wittgenstein or not. In consequence, I will refrain from discussing his work in detail here, since the purpose of this paper is to examine the legitimacy of resorting to fact, property and existence deflationism to make sense of Wittgenstein’s later views on moral entities. Nevertheless, I believe that my investigation of Wittgenstein’s later work on ethics in Sect. 7 provides sufficient evidence to undermine Horwich’s proposed interpretation.

4 Moral entities

Throughout the remainder of this paper, I set out argue that the attribution of fact, property and existence delfationism does not do justice to Wittgenstein’s later work and it is concluded, therefore, that the appeal to deflationism does not afford or substantiate the exegetical claims made by Lovibond and Brandhorst. First, I discuss some inconsistencies in Brandhorst’s characterization of fact deflationism as a local position. Second, I argue that, while these inconsistencies are solved by Lovibond’s interpretation, the proposed solution fails to capture the intricacies of Wittgenstein’s later work. Thirdly, I examine Wittgenstein’s remarks on the ontology of moral values and the grammar of moral discourse, largely ignored by Brandhorst and Lovibond, in order to demonstrate that there is insufficient textual evidence to uphold their exegetical claims.

5 Local fact deflationism?

Despite the centrality of fact deflationism to Brandhorst’s proposed interpretation, he is unable to spell out its implications and, therefore, misconstrues this philosophical program. Brandhorst distinguishes between empirical facts and, on the other hand, ethical and mathematical facts. The former describe facts in a robust, thick sense of ‘fact’: they are objective, real, observable and mind-independent facts. Meanwhile, the latter describe facts in a thin, attenuated and deflated sense of ‘fact’: they have no metaphysical depth the nature of which can be analyzed in terms of mind-independence, observability or objectivity.

This distinction rests on the on the assumption that fact deflationism is a local philosophical program that is only true of some facts (e.g., moral, and mathematical facts), not all of them (e.g., empirical facts). Thus, on Brandhorst’s interpretation there are both harmless empty facts as the deflationist conceives them (see Brandhorst, 2015, p. 245) and substantial objective mind-independent facts as the objectivist conceives them (see Brandhorst, 2015, p. 236, 245). In drawing this distinction, however, Brandhorst fails to recognize that fact deflationism is not well suited for the exegetical claims he is advancing.

Brandhorst’s localization of fact deflationism requires (i) accepting that deflationism only applies to some facts (not all facts) and, by extension, (ii) acknowledging that there are other explanations of facts which exceed the trivial platitude propounded by fact deflationism. Both of these requirements are incompatible with the central claims that comprise fact deflationism.

Requirement (i) involves a certain kind of line-drawing that stipulates which facts fall within the scope of deflationism (e.g., moral and mathematical facts) and which facts do not (e.g., empirical facts). The terms ‘genuine’, ‘objective’, ‘robust’, ‘mind-independent’ and ‘substantive’ have been used by philosophers like Brandhorst “to mark out the special 'heavy-duty', metaphysical notion of fact (call it "FACT") that is needed to draw the supposed line between" FACTS and mere deflated facts (Horwich, 2010, p. 281). Thus, Brandhorst’s distinction between genuine (or thick) FACTS and deflated facts involves an appeal to a metaphysical notion of FACT.¹¹ In other words: this line-drawing entails accepting that fact is a substantial property whose nature is deep enough as to have metaphysical debates about the substantial differences that exist between FACTS and mere deflated facts. This is inconsistent with the fact deflationist’s claim that there is no substantive property of fact whose nature can be informatively analyzed. On the deflationist view all facts are no more than shadows casted by true meaningful propositions.

Meanwhile, against requirement (ii) fact deflationists hold that all that can be said about facts is exhausted by the trivial platitude FD (or some principle like it).¹² Namely, on the deflationist view all explanations of facts must be instances of FD, thus excluding any alternative conception of facts. Furthermore, even if fact deflationism were compatible with other conceptions of facts, it is necessarily incompatible with all substantive and inflationary conceptions of facts, such as the one propounded by Brandhorst with regards to empirical facts. An inflationary conception of facts claims that facts have a deep nature of the kind one might find out about, investigate and develop theories of by offering informative analysis and substantial explanations in terms of mind-independence or objectivity. This contradicts the core claim of fact deflationism.

It follows from the above that fact deflationism is a global philosophical program insofar as it holds that all that can be said about all facts is exhausted by the unexciting trivial schema FD.¹³ Thus, on the one hand, fact deflationism is inconsistent with Brandhorst’s characterization of empirical facts. On the deflationist view all facts are no more than shadows cast by true meaningful propositions. They cannot have any substantive or deep nature that is theorized about or informatively analyzed in terms of being mind-independent, physical, observable, and so on. By contrast, Brandhorst concedes that empirical facts can be informative analyzed in terms of mind-independence and objectivity, thus accepting something more than it is allowed by fact deflationism.

On the other hand, fact deflationism does not afford the distinction Brandhorst identifies in Wittgenstein’s later philosophy between harmless empty moral facts and substantive objective mind-independent empirical facts. On the deflationist view all that can be said about facts is exhausted by the relatively trivial platitude FD, which does not have the explanatory power to yield Brandhorst’s distinction. Furthermore, this distinction requires these two kinds of facts to have natures that are substantially disanalogous, which, as explained above, entails accepting something more than it is allowed by fact deflationism.

Note I am not arguing against the legitimacy of interpreting Wittgenstein as endorsing a distinction between different kinds of facts or realities whose nature is substantially disanalogous. My point is that this exegetical claim is inconsistent with the attribution of fact deflationism. By resorting to this philosophical program Brandhorst misconstrues Wittgenstein as both a substantialist and deflationist (i.e., insubstantialist) about facts, thereby ignoring that these philosophical programs are contradictory and inconsistent. Thus, a deflationary conception of facts is not well suited for Brandhorst’s exegetical purposes.

By recognizing that fact deflationism is a global philosophical program that reduces all philosophical talk about facts to the trivial platitude FD, Brandhorst is faced with an exegetical choice. He can either uphold the claim that Wittgenstein is a fact deflationist, thereby abandoning the distinction between moral and empirical facts or, alternatively, he can uphold the claim that Wittgenstein distinguishes between moral and empirical facts, thereby abandoning the attribution of fact deflationism. I believe that the latter option is more in line with Brandhorst’s proposed interpretation.¹⁴ Nevertheless, the purpose of this paper is to examine the legitimacy of resorting to deflationism to make sense of Wittgenstein’s later views in meta-ethics. In consequence, only the former option will be discussed.

6 Fact, property and existence deflationism

To amend the localist shortcomings in Brandhorst’s deflationist reading and thus better evaluate the adequacy of interpreting Wittgenstein as a fact deflationist let us recall the work of Lovibond.¹⁵ On Lovibond’s interpretation, Wittgenstein is a deflationist about existence and, by extension, facts and properties. He provides us with a metaphysically unexciting way of speaking about the world, where reference to reality is a target that is hit by all declarative/indicative sentences in the same deflationary way (Lovibond, 1983, p. 26). This is because all facts, properties and other entities are no more than shadows casted by true propositions. The only way in which declarative/indicative moral sentences can fail to describe reality is by not being true. Thus, Lovibond’s reading acknowledges the global implications of fact deflationism, property deflationism and existence deflationism.

6.1 Lectures on the foundations on mathematics

Once the global implications of these deflationist programs are acknowledged, it becomes apparent that the textual evidenced adduced by Brandhorst is self-defeating. Specifically, it highlights how of Wittgenstein’s later remarks on mathematics are inconsistent with existence, fact and property deflationism.¹⁶

In Lectures on the Foundations of Mathematics (Wittgenstein, 1976) Wittgenstein discusses the kind of reality to which mathematics corresponds. One of the recurrent themes in these lectures is the temptation to Platonist views, which model mathematical reality analogously to physical reality. As Brandhorst (2015, pp. 239–242) rightly points out, Wittgenstein rejects these Platonist views on the basis that the analogy they employ is unintelligible and “extremely misleading” (LFM, 1976, p. 240). By contrast, Wittgenstein developed the idea of an enormous difference between reality corresponding to experiential propositions and reality corresponding to mathematical propositions (see Diamond, 1996; Connelly, 2013 and Brandhorst, 2015 for further discussion). Take, for instance, the propositions ‘2 is even’ and ‘The sofa is green’.

Initially Wittgenstein explains that we have certain words, like ‘sofa’, “such that if we were asked, ‘What is the reality which corresponds?’, we should all point to the same thing” (LFM, p. 248). By contrast, “If we were asked to explain what the reality is which corresponds to "two", we should not know what to say” (LFM, p. 248; cf. LFM, p. 251). Being clear about the difference between ‘two’ and ‘sofa’ amounts to understanding that the mathematical proposition is about numbers in the sense that it helps prepare the number-sign for its applications, not that the reality corresponding to the mathematical proposition amounts to some sort of realm of numbers (Diamond, 1996, p. 336; LFM, p. 251).

Thus, against existence deflationists, Wittgenstein does not infer the existence of numbers from true mathematical propositions. Furthermore, if he does infer their existence, it is evident by his critique of the Platonist views that the kind of existence that pertains to mathematical objects is disanalogous to that of empirical objects. For instance, the former are the result of human invention while the latter are the result of empirical discovery (see RFM, I-168, II-2, II-38, V-9, VII-5). The description of this difference entails accepting something more than it is allowed by existence deflationism: the use of a criterion of existence telling us what it is for something to exist, e.g., human invention, observability and so on (cf. Thomasson, 2014, pp. 196–197).

Wittgenstein offers similar remarks to differentiate the realities corresponding to true mathematical propositions and true empirical propositions. The latter denote facts in which objects exemplify properties. Meanwhile, “you will look for the reality corresponding to” a mathematical proposition “in an entirely different place; not in mathematics but in its application” (LFM, p. 251). To say a reality corresponds to ‘two is even’ “is like saying ‘A reality corresponds to 'two'" (LFM, p. 249): they prepare number-signs for their application, not denote mathematical entities. Thus, against fact deflationists, Wittgenstein does not infer the existence of mathematical facts from true mathematical propositions. Furthermore, if he does infer their existence, it is evident by his critique of the Platonist views that mathematical facts are substantially disanalogous to empirical ones. The description of this difference entails accepting that facts are not mere shadows cast by true propositions, which is something more than it is allowed by fact deflationism. As Brandhorst (2015, p. 235, my brackets) himself recognizes: “the formal and empty sense [of facts] fails to yield the distinction that Wittgenstein draws”.

Although Wittgenstein does not discuss the difference between mathematical predicates (e.g., ‘even’) and empirical predicates (e.g., ‘green’), it seems reasonable to assume that he would offer a similar conclusion. Empirical predicates denote properties that are predicated about objects, while mathematical predicates are about numbers in the sense that they help prepare number-signs for their applications. Thus, against property deflationists, Wittgenstein would not infer the existence of mathematical properties from meaningful mathematical predicates. Furthermore, if he did infer their existence, it is evident by his critique of the Platonist views that mathematical properties would be substantially disanalogous to empirical ones. The description of this difference entails accepting that properties are not merely shadows cast by meaningful predicates, which is something more than it is allowed by property deflationism.

It follows from the above that Wittgenstein’s remarks on mathematics are inconsistent with existence, fact and property deflationism. Despite recognizing the existence of true mathematical propositions, Wittgenstein explains through his critique of Platonist views that mathematical propositions do not warrant the postulation of mathematical entities that are akin to those empirical entities that are denoted by true empirical propositions. What ontological status Wittgenstein adjudicates to mathematical entities remains a controversial topic among Wittgenstein scholars. However, this exegetical puzzle and the indeterminacy of its solution pose no threat to the present paper.

If “there is no putative realm of mathematical objects the reality of which either to affirm (realism) or to deny (anti-realism)” (Connelly, 2013, p. 576), then mathematical facts, objects and properties cannot be inferred from true mathematical propositions. Meanwhile, if mathematical entities are to be inferred from true mathematical propositions, Wittgenstein’s critique of Platonist views makes it plain that these entities are substantially disanalogous to empirical ones.¹⁷ The description of this difference is inconsistent with and entails accepting something more than it is allowed by existence, fact and property deflationism and the trivial platitudes they propound.

6.2 Mental phenomena

A similar problem could be raised against Lovibond’s and Brandhorst’s proposed interpretations by drawing attention to Wittgenstein’s remarks on mental phenomena. On Lovibond’s and Brandhorst’s deflationist interpretation, Wittgenstein should be willing to accept that from true declarative psychological sentences one can infer the existence of metal entities that are analogous to empirical ones. However, upon closer inspection, this exegetical claim seems contentious, at best.

Wittgenstein accused the inner/outer conception of mistakenly assimilating the mental entities to physical ones. Specifically, “it construes the relationship between mental phenomena and mental terms 'on the model of material 'object and designation', and thereby turns the mind into a realm of mental entities, states, processes and events, which are just like their physical counterparts, only hidden and more ethereal (PI, §293, §308, §339; BB, 47, 64, 70)” (Glock, 1996, p. 53; see MWL, pp. 318–320; AWL, §18; PI, ix, p. 190; for further examples of Wittgenstein’s critique of the inner/outer conception).¹⁸

Note Wittgenstein does not oppose the philosopher’s picture of material object and designation, but rather challenges the attempt to model mental phenomena on this picture. What ontological status Wittgenstein adjudicates to mental phenomena remains a controversial topic among Wittgenstein scholars. Nevertheless, this exegetical puzzle and the indeterminacy of its solution pose no threat to the present paper, since Wittgenstein’s claims about the implausibility of construing psychological phenomena on the model of physical phenomena suffice for our purposes here.

On the one hand, if Wittgenstein denies the ontological existence of mental phenomena then it is patent that there are no entities to be inferred from true psychological propositions. On the other hand, if Wittgenstein concedes that mental entities exist, his critique of the inner/outer conception makes it plain that mental phenomena are substantially disanalogous to, and should not be modeled on, the picture of, physical/observable phenomena. The description of this difference entails something more than it is allowed by existence, fact and property deflationism. Specifically, it involves recognizing (i) distinct criteria for the existence of mental phenomena and observable physical/empirical phenomena (see e.g., MWL, pp. 318–320; AWL, §18) and (ii) that facts and properties are not mere shadows cast by true meaningful propositions and meaningful predicates respectively.

6.3 A new dilemma

The inconsistencies that surface from the attribution of existence, fact and property deflationism to Wittgenstein make it plain that both Brandhorst and Lovibond are faced with an exegetical choice. They must either concede that Wittgenstein is no deflationist about existence, facts and properties or, alternatively, uphold their proposed interpretation by abandoning their attribution of syntacticism to Wittgenstein and arguing that there are no such things as psychological or mathematical entities.

The rationale underpinning this latter solution is thus: by rejecting the claim that Wittgenstein is a syntacticist, Brandhorst and Lovibond can argue that declarative psychological and mathematical sentences, despite their surface grammar, are not truth-apt propositions.¹⁹ In consequence, since there are no true psychological and mathematical propositions from which we can infer the existence of facts, properties, objects and other entities, the non-existence of psychological and mathematical entities ceases to be problematic.

Notwithstanding the tensions that may arise due to Wittgenstein’s acknowledgment of true mathematical (RFM, App. III, 5; LFM, p. 41) and psychological propositions (MWL, p. 319; AWL, §16), the viability of this proposed solution rests on Wittgenstein’s description of moral discourse. Specifically, it is necessary to demonstrate that Wittgenstein treats moral sentences as truth-apt propositions, thus sanctioning the inference from undisputed true moral claims to the existence of moral entities. Although a detailed study of this topic requires a separate investigation (see Glock, 2015; Fairhurst, 2019 for a detailed discussion) and the burden of proof falls on Brandhorst and Lovibond, in Sect. 7 I examine Wittgenstein’s later remarks on ethics to highlight some of the textual evidence that can dissuade us from accepting this proposed solution.

7 Wittgenstein on ethics

Throughout this paper I have primarily focused on discussing the inadequacy of attributing existence, fact and property deflationism to Wittgenstein as a consequence of his remarks on mathematics and psychology. Little has been said, however, about Wittgenstein’s remarks on ethics and the ontology of moral values. This is partly because neither Brandhorst nor Lovibond pay much attention to them, despite their interest in Wittgenstein’s later views in meta-ethics. During the remainder of this paper I want to briefly explore these remarks, as I believe they provide further evidence against Brandhorst’s, Lovibond’s and, to an extent, Horwich’s proposed interpretations.

Wittgenstein, in Lecture on Aesthetics (Wittgenstein, 1976) and Lectures in Cambridge between 1930–1933 (Wittgenstein, 2016) and 1932–1935 (Wittgenstein, 1979), discusses the grammar of declarative/indicative moral sentences, e.g., ‘John is good’ or ‘Hurting animals is bad’. He explains that, due to their resemblance to empirical propositions, it is commonly assumed that these moral sentences are used to relay truthful information about the world (see LA, 1976, Part I §§1–3; MWL, pp. 318–319, 332–333; AWL, §§31–32). For instance, ‘Hurting animals is bad’, appears to inform us about the fact that the action of hurting animals exemplifies the property of being bad. Wittgenstein contends, however, that it is a mistake to construe moral expressions as denoting moral entities purely due to their surface form. The inadequacy of this picture of moral discourse is attested, first, by the unavailability of certain ontological commitments upon which this picture relies on and, second, by its inadequate portrayal of the grammar of declarative moral sentences. I discuss the former issue in Sect. 7.1 and return to the latter in Sect. 7.2.

7.1 Moral values

In Lectures in Cambridge Wittgenstein (AWL, §§31–32) discusses the widely held assumption that moral values, such as goodness and badness, are genuine properties akin to empirical ones, such as elasticity or greenness, which are displayed or exemplified by actions/substances. This standard account of properties is reliant on two assumptions. First, both the action/substance and the property have independently fixed identities that can be independently grasped. Second, the application of a property to an action/substance is understood as a function applied to an argument. In arithmetic’s, functions remain the same applied to different numbers that serve as arguments. Accordingly, properties must remain the same applied to different actions/substances.

To learn whether an action/substance has a given property it is necessary to examine its features and determine whether they are symptoms of said property or not. For instance, “if I want to know whether a rod is elastic I can find out by looking through a microscope to see the arrangement of its particles, the nature of their arrangement being a symptom of its elasticity, or inelasticity” (AWL, §32). The success of this investigation is dependent on whether there is some independent grasp of what the property is, “otherwise the word ‘symptom’ is meaningless” (AWL, §31).²⁰ As Wittgenstein’s puts it: “a cannot be a symptom of b unless there is a possible independent investigation of b” (AWL, §32) which allows us to examine the nature of b, i.e., the commonality that is shared by all instances of b and serves as its fixed identity. For instance, one must have an independent grasp of the property ‘elastic’ which enables us to recognize the elasticity of an object as an indication of this property (see Kuusela, Forthcoming, p. 3; AWL, §§31–32).

The plausibility of extending these considerations to ethics and postulating the existence of moral qualities is dependent, therefore, on whether there is some independent investigation of goodness that uncovers the commonality that is shared by all instances of goodness and serves as its fixed identity. Wittgenstein, unfortunately, rejects this kind of investigation in ethics (and aesthetics) on the grounds that “it is far too simple” (MWL, p. 332).²¹ He explains that moral and aesthetical terms are used for a thousand different things in a thousand different ways (AWL, §32; MWL, p. 335; LA, Part II §4). For instance, the words ‘beauty’ and ‘ugly’ when applied to a face are not the same as when applied to flowers or trees (see AWL, §32). Although we have in the latter a similar game, it is a mistake to construe moral and aesthetical values as having a fixed meaning covering all its applications (AWL, §29; cf. MWL, p. 325). “Nothing would be more astonishing than if “good” [and ‘beautiful’] had the same meaning always, considering the way we learn it” (MWL, p. 325, my brackets).

Wittgenstein, in consequence, makes it plain that moral values do not abide by the standard account of properties. On the one hand, moral values are not susceptible to a fixed identity. If goodness and beauty can be something different in every particular case, then there is no independently fixed identity worthy of the name.²² Moral and aesthetical values are not definable by reference to an independently graspable quality, which serves as the fixed identity shared by all of its instantiations. Instead, it is only possible to ascertain their meaning by seeing how they are used in each language-game and the actions/subjects they are bound up with and modify (see AWL, §§31–32; MWL, pp. 318–319, 332–335 for a detailed explanation).

On the other hand, since moral values fail to remain invariant when applied to different substances/acts, ascriptions of moral value are not to be understood as functions applied to arguments. Moral judgments “do not simply connect two things, goodness and the action independently understood, as in the case of predication” (Kuusela, Forthcoming, p. 11). Instead, moral values and actions/substances are intimately bound up: moral values modify they actions/substances they are bound up with and viceversa. Consequently, on Wittgenstein’s account, moral adjectives are better understood as having an attributive function, rather than a predicative one (see Kuusela, Forthcoming, pp. 11–13 for a detailed explanation). On this view, moral adjectives do not predicate certain properties about an action/substance, but rather function as predicate modifiers that are not predicates in their own right. “Different kinds of good actions […] may therefore not be good in the same sense” (Kuusela, Forthcoming, p. 11).

It is concluded from the above that “there’s a great confusion in calling beauty [and goodness] a quality” (MWL, p. 333, my brackets) and postulating the existence of moral properties. It is mistaken to construe moral (and aesthetical) adjectives as denoting qualities or properties purely in virtue of their surface form, as it is misleading and not indicative of their actual use (see Sect. 7.2). Thus, goodness and beauty are not properties that are instantiated and exemplified by different actions/objects.²³

Brandhorst does quote two passages of Wittgenstein’s discussion on the misleading surface grammar of moral and aesthetical adjectives:

The use of such a word as 'beautiful' is even more apt to be misunderstood if you look at the linguistic form of sentences in which it occurs than most other words. 'Beautiful' [and 'good'-R] is an adjective, so you are inclined to say: "This has a certain quality, that of being beautiful". (LA, Part I, §1)

If I had to say what is the main mistake made by philosophers of the present generation, including Moore, I would say that it is that when language is looked at, what is looked at is a form of words and not the use made of the form of words. (LA, Part I, §5)

Brandhorst argues that Wittgenstein is not negating the existence of moral and aesthetical properties, but rather deflating it. “That is harmless because it is empty […]. It does not introduce a ‘quality’ as the objectivist conceives of it” (Brandhorst, 2015, p. 245). Thus, on Brandhorst’s interpretation, Wittgenstein only takes issue with the realist’s postulation of objective moral properties (e.g., Moore’s realist view of ethical qualities) insofar as they are thought as mysterious entities that have a “‘shadowy’ existence in a ‘shadowy’ part of the world, and that is a mistake” (Brandhorst, 2015, p. 245).

It is unclear, however, how the quoted passages warrant Brandhorst’s exegetical conclusion. Wittgenstein is highlighting the inadequacy of inferring the existence of moral properties/qualities from meaningful moral adjectives purely due to their surface form. Brandhorst’s appeal to deflationism incurs in this same error. Specifically, it construes meaningful moral adjectives as denoting deflated moral qualities/properties, whose existence is inferred directly from said adjectives, purely due to their surface form. It is another instance of the widespread mistake made by philosophers who look at the form of words and not the use made of the form of words (see Sect. 7.2 for the use of these words).

Moreover, Brandhorst’s deflationist interpretation still fails to overcome the problems raised by Wittgenstein against the idea of moral properties. Wittgenstein does not take issue with the objectivism that moral realists’ attach to moral properties. Instead, he brings out the inadequacy of postulating moral properties insofar as this idea is far too simple to capture the intricacies of our ascriptions of moral value. Thus, postulating deflated moral properties still entails accepting that a deflated property serves as the fixed identity common to all instances of such-and-such moral value, which is something more than it is allowed by Wittgenstein.²⁴ Brandhorst’s appeal to deflationism does not make these properties any less ‘mysterious’ or ‘shadowy’.

The lack of a fixed identity for moral values, i.e., a commonality shared by all instances of such-and-such moral value, also jeopardizes the possibility of conceiving them as moral objects. In other words: Wittgenstein “rejects the assumption that such things as good, evil, and pains must be things, objects more or less like houses, albeit perhaps of a queer sort” (Richter, 2018, p. 167).

In turn, the non-existence of moral properties and objects jeopardizes the possibility of postulating the existence of moral facts. Facts are generally understood as entities comprised by objects and/or properties that stand in a certain relation. For instance, the fact ‘Wittgenstein is mortal’ is composed of Wittgenstein, which is the object, and the property of being mortal. If there are no moral properties or objects it is unclear in what sense a fact can be moral, since there are no moral objects and/or properties that comprise them.

Ensuing the rejection of moral and aesthetical qualities, Wittgenstein briefly considers the idea of moral reductionism, which purports to reveal the nature of moral properties and enlighten what it is for something to be good or bad. Moral reductionists hold that moral facts, properties and objects are identical to the non-moral facts, properties and objects they are reduced to and which can be expressed using non-moral vocabulary.

Wittgenstein (AWL, §§34–36) discusses reductionism and its implications by examining the widespread temptation of reducing ethics to psychology. On this view, moral values are reduced to mental states, e.g., pleasure or pain. This is accomplished by specifying the causal connections that warrant the reductive identity claim between certain values and these non-moral entities. These causal connections are expressed by psychological propositions, such as ‘x is good or beautiful if x gives us pleasure’, which are hypotheses that can be confirmed or disconfirmed through empirical investigation. The question about the goodness of an action, then, is to be answered by psychological propositions which cite causes.

Wittgenstein (AWL, §34; Wittgenstein et al., 2015, p. 27) counters that this causal reductive explanation of moral values does not remove the ethical puzzle one feels when asked what makes an action good—at times Wittgenstein’s remarks are reminiscent of Moore’s Open-Question Argument. The question about the goodness of an action does not demand specifying the physical and mental effects caused by the action on the human body and how they are causally connected with the moral evaluation. Ethical puzzles do not demand causes, but rather reasons. “The difference between a reason and a cause is brought out as follows: the investigation of a reason entails as an essential part one’s agreement with it, whereas the investigation of a cause is carried out experimentally” (AWL, §36).

Ethics, insofar as it demands reasons, is no science nor should it be reduced to science. Accordingly, reducing moral values to non-moral properties does not yield an adequate explanation of ascriptions of moral value that enlightens what it is for an act/substance to be good—as Brandhorst (2015, p. 248) himself recognizes. Furthermore, reductionism is still predicated around the problematic and oversimplistic idea that moral values have a fixed identity.

Wittgenstein’s remarks, then, indicate a dismissive attitude towards the postulation of moral entities. The (dubious) appeal to deflationism does not afford or deliver ontological conclusions desired by Lovibond and Brandhorst, as there is insufficient textual evidence to uphold the claim that Wittgenstein provides a novel way of speaking about moral facts, objects and properties.²⁵ Furthermore, and against Horwich, it seems that Wittgenstein forgoes the postulation of any moral entity, regardless of whether it is deflated or inflated. It is for this reason that Wittgenstein scholars, such as Kuusela (Forthcoming), have recently argued that Wittgenstein’s moral realism forgoes the postulation of moral properties or naturalistic reduction.

7.2 Moral words

The idea that declarative moral sentences are used to relay truthful information about the world is partly rooted in their surface form. According to Wittgenstein, the surface form of moral discourse encourages, rather than forbids, understanding its use by simile to empirical discourse. Thus, it is commonly assumed that moral substantives denote substances, moral adjectives denote properties and declarative moral sentences denote facts (see LA, Part I §§1–3; MWL, pp. 318–319, 332–333; AWL, §§31–32). This assumption is at the heart of both Lovibond’s and Horwich’s attribution of syntacticism to Wittgenstein.

Wittgenstein contends, however, that the surface form of moral words plays us tricks. To get clear about the grammar of moral words, which often lacks openness to view (cf. MWL, pp. 318–319, 332–333; LA Part I, §3), he suggests one must examine the use, rather than the form, of these linguistic expressions. To accomplish this task Wittgenstein recommends studying how these expressions are taught, as it destroys many of misconceptions about their grammar and provides primitive language-games in which they are used (LA Part I, §5).

Wittgenstein initially focuses on moral words that are used as adjectives. He explains that there is a great confusion in saying that moral adjectives, e.g., ‘good’ or ‘bad’, denote or stand for properties in virtue of their surface form (LA, Part I §§1–3; AWL, §§31–32; MWL, p. 333). “We do not as children discover the quality of” goodness or badness in such-and-such action “and find that these are qualities” such-and-such action “has in common with it” (AWL, §32).²⁶ ‘Good’, ‘bad’ and other moral adjectives are roughly learnt (and used) as interjections that substitute certain gestures which express feelings and other natural reactions of approval/disapproval about something (LA, Part I §§5–7, §10; Wittgenstein et al., 2015, p. 30; see Glock, 2015; Fairhurst, 2019 for a detailed study of the meaning of moral adjectives). “In fact, if we do want to be exact, we do use a gesture or a facial expression” (LA, Part I §10). Although this may not be the exact language you speak when you are older, “you get a rough approximation to what kind of language game is going to be played” when using certain moral adjectives (LA, Part I §5).

In like manner, Wittgenstein indicates that one great trouble our language gets us into is that we mistakenly take a moral substantive “to stand for a thing or substance. Ordinary grammar doesn’t forbid us to use substantives in this way: the origin of all use of substantives and verbs is in fact a simile for physical bodies moving” (MWL, pp. 318–319). However, upon closer inspection it is apparent that there are no moral substances to be described.

Accordingly, declarative moral sentences are not used to denote moral facts where substances exemplify moral properties and thus relay truthful information about the world. This is due to the fact that the moral terms that compose these sentences are not used to denote moral entities. Wittgenstein offers an example of this non-descriptivist conclusion when discussing the meaning of ‘This is a noble action’. He explains that this moral sentence simply says ‘ah’ together with a facial expression or gesture of approval (e.g., a smile) about such-and-such action (Wittgenstein et al., 2015, p. 30; cf. LA, Part 1 §7).²⁷ It does not predicate a mysterious moral entity about said action.

Thus, summarizing, it is a mistake to understand moral names as referring to moral objects, moral adjectives to moral properties, and moral sentences to moral facts (see LA, Part I §§1–3; MWL, pp. 318–319, 332–333; AWL, §§31–32). Wittgenstein foregoes the need to postulate moral entities for moral words to be meaningful and useful. “Problems about what these words are about, what their real subject is, [which is called 'beautiful' or 'good'.-R.] don't come up at all” (LA, Part I §8). Accordingly, against Lovibond, Horwich and Brandhorst, it is ill-judged to construe moral words on the model of material object and designation due to their superficial resemblance to empirical discourse. Contrary to what their surface form indicates, moral sentences do not serve the purpose of relaying truthful information about the world by denoting moral facts. An ethical sentence “is a personal act. Not a statement of fact. (Like an exclamation of admiration)” (PPO 2013, p. 85, my emphasis; cf. MWL, pp. 318–333; AWL, §§31–32).

It may be objected that the expressive function briefly discussed above does not exhaust all uses of moral expressions (see De Mesel, 2019 for more on this issue). For instance, Wittgenstein acknowledges that some moral words appear to be prescriptive: they serve the purpose of prescribing certain conducts that are regarded as morally good. Furthermore, in uttering a declarative moral sentence and judging such-and-such as morally good it seems that we also prescribe a certain conduct (see Glock, 2015 for more on this topic). While all of this may very well be the case, first, there are still no moral entities to be described in order to relay truthful information about the word. Again, an ethical judgment is “not a statement of fact” (PPO, p. 85). Thus, these alternative functions in principle pose no threat to my arguments against Brandhorst, Lovibond and Horwich’s interpretations.²⁸ Second, there is insufficient textual evidence to discuss many of the alternative uses of moral discourse since Wittgenstein’s scarce remarks focus primarily on the expressive function of moral words and moral judgments. Thus, rather than trying to infer these other uses from his investigations elsewhere, I have settled for studying the expressive function of some moral words as specific examples of Wittgenstein’s moral non-descriptivism. These expressive uses, however, need not exhaust all Wittgenstein has to say about the non-descriptive grammar of moral discourse. Finally, the purpose of my investigation is to study declarative moral sentences, as it is these sentences that are supposedly used to denote moral entities and thus relay truthful information about the world in Brandhorst, Lovibond and Horwich’s estimation. A complete study of the grammar of moral discourse as a whole exceeds the scope of this paper.

7.3 The need for reality

Brandhorst’s proposed interpretation is predicated around the premise that, for Wittgenstein, correspondence to reality is a requirement moral language must fulfill to avoid being “a game played merely for entertainment” or “some empty formalism without use” (Brandhorst, 2015, p. 244).

We ask or demand or wish for certain things of one another; we praise people for what they do or achieve; we promise to do certain things and accept obligations to others; we criticise and we reproach; […]. All this—and much more could be added—is real. It marks the way we live. It is important to us, shaping our relations to ourselves as well as to others. In this way, it provides the framework for our use of ethical language. So as before, it would be misleading to say that no reality corresponds to that language. (Brandhorst, 2015, p. 244)

Thus, moral language has a use insofar as it corresponds to, and is interwoven with, the tapestry of our life. It is this reality, the reality of human affairs, to which ethics corresponds (Brandhorst, 2015, p. 246).

Despite correctly identifying that moral judgments are not a borderline case of pure non-descriptive sentences, there are certain inconsistencies in Brandhorst’s explanation of the grammar and descriptive content of moral discourse.

First, Brandhorst suggests that moral language has a meaningful use in virtue of its correspondence to the reality of human affairs. However, when “we praise people for what they do or achieve; we promise to do certain things and accept obligations to others; we criticise and we reproach” (Brandhorst, 2015, p. 244) our moral terms do not have the function of describing or denoting the reality of human affairs.²⁹ Instead, they are used to praise, promise, criticize, wish and so on. In other words: the meaning of these moral expressions is not given by the (non-existent) moral entities they are said to describe (see Sect. 7.2). Suppose you utter the sentence ‘Your behavior is wrong’ to criticize a certain behavior. The moral term ‘wrong’ does not denote any moral entity (e.g., the human practice of criticizing). By contrast, it is used non-descriptively to criticize and disapprove such-and-such behavior. For instance, it may substitute a certain gesture that expresses disapproval about said behavior, prescribe the opposite conduct, etcetera.

Relatedly, it is questionable whether descriptions of these human affairs, customs, traditions, rules, behaviors and prescriptions can be regarded as genuine moral sentences. In his conversations with Rhees, Wittgenstein explains: “Supposing you simply describe the Sitten und Gebräuche of various tribes, this would not be ethics” (Wittgenstein et al., 2015, p. 27). For example, describing the human practices of promising, praising criticizing and so on would neither be ethics nor result in descriptive moral sentences. It is a mistake to construe these practices as the moral entities denoted by moral expressions (see note 31). Thus, Brandhorst must clarify in what sense moral language corresponds to reality, given that the description of customs, behaviors and traditions and other human affairs is not ethics.

Second, Brandhorst rightly points out that the reality of human affairs provides the framework for our use of moral language. This idea is already found in Wittgenstein’s notion of ‘language-game’, which brings into prominence the fact that the speaking of language “is part of an activity, or of a form of life” (PI, §23). In other words: all language-games are social linguistic activities that are embedded within the affairs and practices of a form of life and must be understood against this background. Accordingly, clarifying the meaning of moral expressions entails paying attention to the non-linguistic activities with which they are interwoven (see e.g., LA, Part I §35; MWL, pp. 321–322).

The problem arises when Brandhorst concludes from the above that this non-linguistic context is the reality to which moral discourse corresponds. Brandhorst makes this inference in the transition between the two final sentences of the quote supplied above and it is without foundation. Saying that the reality of human affairs is the framework for our use of moral language tells us nothing about the use and grammar of moral language. It only provides the context within which moral language-games take place. Thus, stating that moral discourse corresponds to this reality requires separate argumentation and explanation.

The inadequacy of Brandhorst’s inference can be further attested by Wittgenstein’s remarks on avowals. Avowals are also to be understood in connection to the non-linguistic affairs and practices of a form of life within which they are embedded. We cry; we smile; we learn to express feelings; we learn to recognize how others feel. All of this is real. It marks the way we live, shaping our relations to ourselves as well as to others. In this way, it provides the framework for our use of avowals. However, it is a mistake to infer that avowals correspond to this reality. Avowals describe first-person psychological sentences that are used to express mental states. “Negatively, this indicates that they are not descriptions or reports of private mental entities encountered in an inner realm” (Glock, 1996, p. 50). That is, avowals are not in the business of referring or corresponding to reality.³⁰

Finally, Brandhorst’s work rests on the ill-judged premise that correspondence to reality is a requirement moral language must fulfill to avoid being “a game played merely for entertainment” or “some empty formalism without use” (Brandhorst, 2015, p. 244). Reconsider the case of avowals. Despite their lack of correspondence to reality, they play a crucial role in human life. Verbal expressions of pain, for instance, are crucial for warning others about one’s need to visit the doctors immediately due to illness. Thus, foregoing correspondence to reality does not entrain that avowals are games played for entertainment or empty formalisms without use. Their expressive function alone suffices for them to fulfill their crucial role in human life. It is important to “make a radical break with the idea that language always functions in one way, always serves the same purpose: to convey thoughts—which may be about houses, pains, good and evil, or anything else you please” (PI, §304). Correspondence to reality is a not a prerequisite for language to be useful or meaningful.

Accordingly, moral words need not correspond to reality for them to be interwoven with, and play a crucial role in, human life. Foregoing the postulation of moral entities to which moral words correspond does not entrain reducing these expressions to “a game played merely for entertainment” or “some empty formalism without use” (Brandhorst, 2015, p. 244). (Moral non-descriptivists and anti-realists would be unwilling to accept that their meta-ethical positions entail understanding moral discourse as nothing more than an empty formalism played for entertainment).

Problems notwithstanding, the non-descriptive function of moral words does not rule out the possibility that the moral judgments containing these words are connected to reality. Moral judgments can be about the world in Brandhorst’s sense without denoting mysterious moral entities. Christensen (2011, pp. 807–808) has suggested that for Wittgenstein moral judgments are not a borderline case of pure normative and non-descriptive sentences because they involve relevant descriptions about the features of the actions that make us judge them such-and-so. For instance, our understanding of ‘X is cruel’ involves relevant descriptions about the action X which make us judge it as cruel (Christensen, 2011, p. 807). Thus, on her interpretation, the descriptive and normative content of moral judgments are inseparable.

Building on Christensen’s interpretation, I defend that for Wittgenstein the descriptive and non-descriptive (or, in Christensen’s words, normative) content of moral judgments are inseparable, but, ultimately, distinguishable. Their descriptive content denotes the actions being evaluated and the features of these actions that make us judge them such-and-so.³¹ Meanwhile, their non-descriptive content provides the corresponding moral evaluation. Unfortunately, the descriptive content of moral judgments does not afford Brandhorst’s, Lovibond’s and Horwich’s attribution of moral descriptivism to Wittgenstein. First, the meaning of moral judgments is not purely descriptive, since the moral words contained by these judgments are non-descriptive (e.g., expressive, emotive, evaluative, prescriptive and so on).³² Second, moral judgments are not used to denote moral entities. By contrast, they express a moral evaluation towards the actions they denote (cf. AWL, §32). In other words: they correspond to some part of reality without denoting non-existent mysterious moral entities.

8 Conclusion

Concluding, in this paper I have argued that Wittgenstein does not provide a novel conception of moral facts, properties and objects by adopting fact, property or existence deflationism. First, I outlined some inconsistencies in Brandhorst’s characterization of fact deflationism as a local position. Second, I argued that, while these inconsistencies are solved by Lovibond’s interpretation, the proposed solution does not do justice to Wittgenstein’s later views on mathematics and psychology. In consequence, the appeal to fact, property or existence deflationism does not afford or substantiate the exegetical claims made by Lovibond and Brandhorst. Thirdly, I examined Wittgenstein’s remarks on moral values and the grammar of moral discourse, largely ignored by Brandhorst and Lovibond, in order to demonstrate that there is insufficient textual evidence to uphold their exegetical claims.