An Abductive Defence of Truthmaker Realism

Abstract

This paper presents an abductive argument for realism and truthmaker realism as follows. A metaphysical theory is better if it ontologically accounts for truths better than its rivals (the Abductive Principle). Truthmaker realism gives us a better ontological account for truths than its antirealist truthmaker rivals (Abductive Step). So, truthmaker realism is better than antirealist rivals. It presents the truthmaker project as an abductive project which asks us what accounts best ontologically for our truths. Antirealisms, especially idealisms, fail against their realist rivals on various abductive criteria.

Truthmaker realism is plagued by three main objections. Presenting an abductive argument for realism does two important things. First, it dissolves the standard objections. Second, it shows how truthmaker realism is overall better motivated than antirealist, pluralist, and neutralist rivals. Simple truthmaker principles added to a plausible abductive package of principles give us a straightforward argument for realism and against antirealism or any neutralist middle ground.

1. Introduction

Truthmaker theory is a family of accounts about a relation between bearers of truth and veridicality (e.g., sentences, experiences, judgements) and the reality their veridicality asymmetrically depends on. Given the asymmetry, the truthmaking reality on which truths depend doesn't generally depend on these truths for their reality. The result: truthmaker theorists have assumed what I call ‘truthmaker-realism’, that they should be realists (rather than antirealists, especially idealists) about truthmakers (i.e., what make truthbearers true), holding, minimally, that there are mind-and-language-independent truthmakers, or stronger: that truthmakers should generally be mind-and-language-independent.¹ For instance, ‘Willows exist’ is made true by mind-and-language-independent willows, which don't depend for their existence on this truth. Early 21st century challenges (e.g., Beebee and Dodd, 2005) to this truthmaking-realism coupling were assumed to challenge truthmaker theory itself. These challenges argued that because truthmaking is neither (a) necessary nor (b) sufficient for realism, and that (c) insisting on truthmaker-realism is question-begging against antirealist alternatives, truthmaker theory itself is unmotivated. This paper dissolves these challenges by presenting a novel abductive defence of ‘truthmaker-realism’ against truthmaker-antirealist rivals. It shifts the debate from requiring, as truthmaking-critics have, that truthmaker-realists should establish truthmaking's necessity and sufficiency for realism to exhorting that truthmaker-realists should defend themselves abductively against rivals. Thereby both truthmaker-realism and truthmaker theory itself are defended.

First, §2 sketches the abductive argument for truthmaker-realism and properly introduces the three challenges (a)–(c). §3 and §4 address (a) and (b) by arguing that even if truthmaking is neither necessary nor sufficient for realism, truthmaking and realism's close tie is maintained on abductive terms. §5 argues against (c), briefly arguing against a rival stance which underlies it: truthmaker-neutralism, that is, either agnosticism or principled neutrality about truthmakers’ metaphysical status (e.g., mind-independence). §6 presents two abductive arguments against truthmaker-realism's main antirealist rival: truthmaker-idealism. Together, §§2–6 introduce a new strategy for defending truthmaking and realism's close connection, truthmaker-realism, against its main rivals.

2. The Abductive Strategy for Truthmaker-Realism

The following are abductively important maxims which any theory should follow:

explanatory power: minimise brute truths, that is, leave as little truth unexplained as possible.

anti-ad-hoc-ery: do not make ad hoc additions to one's theory.

common knowledge consistency: maximise consistency with one's knowledge and our common knowledge.

All three are important maxims of theory choice. According to explanatory power, if theory T₁ accounts for more truths than rival theory T₂, then T₁ is a better theory overall than T₂; T₁ leaves less truth unaccounted for or unexplained, that is, leaves us with fewer brute truths.² anti-ad-hoc-ery is important for it gives us a clear decision procedure: if T₂ makes certain additions merely to deal with some problem or to satisfy a theoretical demand while T₁ does not have to make additions to deal with the problems and theoretical demands, then T₁ is preferable. It is also related to another key virtue:

independent motivations: the more independent reasons for an addition to a theory, the better motivated the addition.

By contrast, if an addition has no independent motivations, it is not properly motivated. anti-ad-hoc-ery and independent motivations go hand in hand.

common knowledge consistency is clearly important: clashing with common knowledge (or one's own knowledge) entails that something in the theory in question is false and there is strong evidence against it. So, a theory should aim to minimise clashes and maximise consistency. Following these maxims is key to the abductive defence of any theory in the face of its various explanatory tasks and in the face of theoretical rivals.

There has been much debate about the truthmaker-project's nature. I propose that it is an abductive project which adds an extra maxim to the abductively important maxims:

truthmaker-maxim: minimise ontologically brute truths, that is, provide a sufficient ontological basis for all relevant truths.

Since not all theories are theories of everything, that a theory leaves some truths without an ontological basis is not a problem for a theory, unless those truths are relevant to the theory. For instance, if it is part of a theory T that s is true, then s should ultimately be given a sufficient ontological basis according to T.³ The truthmaker-maxim is related to explanatory power above. While the latter requires that a theory minimise brute truths by giving an explanation for as much truth as possible, the former is more demanding, asking for an ontological basis for all relevant truths.

Truthmakers are those entities which ontologically explain or account for truths. Asay (2020, §2.1), for instance, divides truthmaking into two rival approaches: explanatory truthmaking and truthmaking-as-ontological-accounting. He defends the latter but argues that even ontological truthmaker theorists can say that truthmaking explains truth (ibid., p. 44). Metaphysical theories are tested on how well they can account for truths: if theory T₁ accounts for truths better than theory T₂, then T₁ is a better theory overall than T₂.

Famously, earlier truthmaker theorists such as Bigelow (1988, p. 123), Armstrong (2004, p. 5), and Heil (2003, p. 61) explicitly tied truthmaking closely to realism, a general metaphysical view according to which there are mind-independent entities. What has worried critics especially is that they never actually argued for this connection. If we understand the project abductively, however, we can construct a novel, abductive way of understanding the famous statements and of arguing for truthmaker-realism and hence realism:

Premise 1: A metaphysical theory is better if it ontologically accounts for truths better than its rivals (the Abductive Principle).

Premise 2: Truthmaker-realism gives us a better ontological account for truths than its antirealist truthmaker rivals (the Abductive Step).

Therefore,

Conclusion: Truthmaker-realism is a metaphysically better theory than antirealist rivals.

Truthmaker-realism is the view that there are mind-independent truthmakers and that truthmakers are on the whole, exceptions aside, mind-independent in the relevant way. Some clarifications: (1) tables, for instance, are ‘mind-independent in the relevant way’ for realism, even though their existence depends on people designing them.⁴ (2) The relevant realism here is general, not for instance relative to a ‘discourse’ (Cameron, 2008, p. 122) or ‘set of claims’ (Asay, 2020, p. 159; 2012, p. 389).⁵ General forms of realism are contrasted with general antirealisms, which aren't discourse-relative but global. For instance, idealism claims that all things are mind-dependent in the relevant way, and globally deflationary neutralisms eschew metaphysics altogether. Defending general realism and articulating truthmaker-realism this way are important because of the recent resurgence of antirealisms, especially idealism.⁶ (3) Truthmaker-realists think that some truths are made true by mind-independent truthmakers (e.g., truthmakers for ‘Some willows are shrubs’), or stronger: truthmakers are generally mind-independent; clear exceptions include truth about mind-dependent reality (e.g., ‘Fear exists’) and truthmaker-free truths (potentially, e.g., negative truths, see Schipper, 2018). Truthmaker-antirealists accept truthmakers but claim they're all mind-or-language-dependent: e.g., truthmaker-idealism takes truthmakers to be ideas. Truthmaker-antirealisms about specific domains (e.g., truths about morals, social constructions, mathematics) take their truthmakers to be mind-or-language-dependent; they are compatible with general truthmaker-realism: some truths may still have mind-independent truthmakers.

Critics of truthmaking, such as Dodd (2002), Beebee (with Dodd) (2005) and MacBride (2005, 2020), argued that truthmaker-realism lacks motivation with three main objections: truthmaking is neither (a) necessary nor (b) sufficient for realism, and (c) insisting on truthmaker-realism is question-begging against antirealist-truthmaking. (a)–(c) are supported with simple, straightforward considerations.

Objection-a

realism without truthmaking: accepting any kind of truthmaker principle is, as Dodd (2002, p. 83) has famously argued, a ‘hugely uncompulsory additional step’ for avoiding idealism or other non-realist views truthmaker theorists have attempted to undermine (e.g., phenomenalism, see Martin, 1984). All we need to accept, Dodd (ibid.) explains, is that ‘the items to which we are ontologically committed by the things we say (viz. objects, events and–arguably–properties) are mind-independent [to] have a version of realism’; Tahko argues that ‘Realism can stand on its own [without truthmaking]’ (2015, p. 218);

Objection-b

antirealist truthmakers: the availability of alternative antirealist and neutral versions of truthmaking; as Cameron (2008, pp. 116ff.) writes, ‘it's definitely not sufficient to be a realist that one be a truthmaker theorist, since one can simply be anti-realist about the truthmakers’, or Tahko: ‘truthmaking is a way to account for truth regardless of one's ontology’ as it can ‘offer an ontologically neutral way to account for truth’ (2015, p. 217, added emphasis); and

Objection-c

question-begging: using a truthmaker-realist-principle, or a truthmaker-principle that is congenial to realism, in arguing against versions of antirealism, is to beg the question against such views.

Presenting the truthmaker argument for realism as abductive, however, does two important things: it dissolves these standard objections, in the sense that (1) even if truthmaking is neither necessary nor sufficient for realism, it neither follows that truthmaker theory as a whole nor truthmaker-realism is unmotivated; and (2) it isn't question-begging: it provides independent reasons to prefer realist over antirealist truthmaking. I shall argue for (1) in §3 (realism without truthmaking) and §4 (antirealist truthmakers); and (2) in §5 (question-begging) and §6 where I illustrate the strategy with two anti-idealist abductive arguments. Doing (2), further, responds to the critics’ anti-truthmaking arguments that use the possibility of truthmaker-idealism against truthmaker-realism and truthmaking generally: truthmaker-idealism's mere possibility doesn't undermine truthmaker-realism, if truthmaker-realism is still abductively stronger. We thus respond to both truthmaker-idealists and anti-truthmaker-theorists.

A final point: truthmaker theorists and those sympathetic to the project have responded to these objections (a)–(c) not by addressing the objections head-on but by retreating to minimal, ‘ontologically neutral’ (Tahko, 2015, §3, pp. 207–11) or pluralist, ecumenical versions of truthmaking (Asay, 2022, passim; 2020, pp. 22ff., §4; Schipper, 2016, §1.6–1.7), explicitly stating that truthmaking is compatible both with realism and antirealism. Since truthmaker theory itself is committed to neither realism nor antirealism, it is argued, truthmaker theorists should in fact be neutral about the nature of any given truthmakers. This is, plausibly, also behind the truthmaker-semantic strategies of Yablo (2014), Fine (2017a), and others. Truthmaker-neutralism is another potential truthmaker-antirealist challenge to which I shall respond, in §5, in defence of truthmaker-realism on abductive grounds.

3. Objection-a: Realism without Truthmaking?

Against objection-a: even if critics such as Dodd are right that realism can be expressed independently of truthmaking, one cannot conclude then that truthmaker theory is not well motivated or that truthmaker-realism is false or not attractive. For an important question remains for any realist metaphysical theory: ‘What ontologically accounts for truth?’ If the theory does not identify the objects in its realist ontology as truthmakers, the ontological grounds, of truths, then what, according to the theory, does ontologically account for truths? Realists who do not also accept truthmaker-realism would leave much unaccounted for. Without such an account, they would score poorly on an important abductive measure for any good metaphysical theory: any ontologist needs ultimately to account ontologically for truths. Further, apple-realists who are not truthmaker-realists accept that there are apples but not that apples make claims such as ‘there are apples’ true. They would (i) need to account for the relationship between ‘there are apples’ and reality in another way, and (ii) explain why their account is better than those who say that apples make such claims true, especially given that their own account already accepts their existence.

They would then also fall short on abductive measures such as simplicity, elegance, and, importantly, relevance. The apple-realist truthmaker view, by contrast, scores high especially on relevance, but also simplicity and elegance (see Forster and Sober, 1994, for a defence of simplicity and its relation to anti-ad-hoc-ery). For apples are clearly the most relevant worldly entities to account for the truth of ‘apples exist’, and appealing to apples as the truthmakers is elegantly simple compared to a complicated story, say, about sense-data or ideas in the minds of deities, especially when apples are available (see §6).

One might respond that this is a shallow response against critics such as Dodd, who accept that ‘apples exist’ is ontologically committed to apples but no truthmaker-ideology. However, this section argues that even if realism doesn't require truthmaking, truthmaker-realism is the abductively better realist position. A realist can reject truthmaker theory (e.g., Dodd), but unless for instance one's realist theory of ontological commitment is tied with a realist truthmaker theory, one still needs to say what accounts for the truths. Rejecting this question altogether seems abductively non-ideal, especially if one already has an ontology which one can use to answer the question ‘What accounts for truths such as “apples exist” or “this apple is spherical”?’. The alternative is to answer this question, saying what accounts for truths, for instance saying why only including spherical apples in one's ontology (as Dodd does) allows us to account for truths such as ‘this apple is spherical’. In the latter case, one is giving a truthmaker theory, despite likely rejecting necessitation and the standard truthmaker-ontology of states-of-affairs (Armstrong, 1997) or particularised properties (Martin and Heil, 1999, p. 45). Say one is realist about spherical apples: they are mind-independent. Then a realist view of spherical apples as truthmakers seems preferable both to rejecting the task of ontological accounting and to an antirealist view about what accounts for such truths. Note, I assume that it's controversial nowadays to insist that truthmaker-ideology essentially involves thinking that truthmakers must be non-mereological unities of objects and their properties – not even Armstrong held this for all truths, and Parsons (1999) and Schipper (2020, §1.1) argue that this view of truthmakers’ ontological category is non-compulsory (though see Dodd, 2002, pp. 74–75 against Parsons). This doesn't affect my defence of truthmaker-realism.

4. Objection-b: Non-realist Truthmakers?

Critics have made clear that precise formulations of the truthmaking relation itself, whether in terms of necessitation (viz., Armstrong) or supervenience (e.g., Bigelow and Lewis), are compatible with some forms of non-realism. According to

necessitation (nec): x makes p true if, and only if, in all the worlds in which p and x exist, p is true (see Merricks, 2007, p. 7).

Most now, except Asay (2020, ch. 3), would take necessitation only to be necessary (e.g., Schipper, 2021, §2; 2022a, p. 997). Supervenience accounts claim that

supervenience-t (sup): truths, the Ts, supervene on reality: so any change in the truths in terms of truth-value entails some change in reality (Bigelow, 1988, pp. 133ff., 158ff.; Lewis, 1992, p. 218).

Antirealists can accept both nec and sup and just claim that all the x's which necessitate truths, and the reality on which truths supervene, are mind-, evidence-, or language-dependent entities.

Some antirealists, in fact, explicitly use the language of truthmaking to articulate their views; for instance, intuitionists following Brouwer and Heyting, such as Göran Sundholm (1994a, pp. 121–22; 2004, pp. 449–51), think that the truthmakers for mathematical claims are proof-constructions.⁷ One might generalise: all claims, also those about chairs and quarks, are made true by constructions, proof-like systems of evidence. Put idealistically: all truths supervene on or are necessitated only by minds and other mind-dependent truthmakers. I shall often focus on the relationship between truthmaking and idealism as the main antirealist rival in this paper, about which Bergmann famously writes: ‘the truth of S must be grounded ontologically. On this first move idealists and realists agree’ (1961, p. 229; see also Dodd, 2002, pp. 83–84; and Daly, 2005, pp.95–97). Note, however, that when this is brought up in the literature (e.g., MacBride, 2020, §3.1), it is never mentioned that Bergmann will argue strongly that the further moves favour realism (see 1964, passim; for discussion, see Allaire, 1974, pp. 60–61; Heald, 1992, pp. 40–43). Fumerton (2002, pp. 5–6) points out that truthmaker theory and alethic realism are compatible with ‘radical metaphysical idealism’. Fumerton (2013, p. 200) also points out that even Berkeley (1713) was ‘a robust realist’ about truth while holding the view that ‘the only truth makers for Berkeley were facts about minds and ideas’ (ibid.). Thus, according to him, Berkeley is a truthmaker-idealist.

It is obvious that just saying that truth depends asymmetrically on, is necessitated by, or supervenes on, reality isn't sufficient for realism. A basic asymmetry principle clearly doesn't rule out that reality is mind-or-interest dependent. Figures such as Berkeley, other idealists, and antirealists of various sorts, have not been unsympathetic to correspondence intuitions. Dummett (1976, pp. 51–55), famously a supporter of antirealism, accepts Principle C (if a statement is true, there must be something in virtue of which it is true, 1976, p. 53), which the intuitionist Sundholm describes as ‘clearly nothing but a formulation of a truth-maker condition on truth’ (1994a, pp. 123–24). See also Daly (2005, p. 96) who argues that the correspondence intuition (that ‘<p> is true if and only if things are as <p> says they are') is compatible with any theory of truth and is ontologically neutral. He (ibid., p. 97) even argues that combining this with the truthmaker principle and Armstrong's states-of-affairs ontology gives us principles which are compatible with pragmatist and coherentist theories of truth.

How should truthmaker-realists (e.g., Bigelow, 1988, p. 123; Armstrong, 2004, p. 5; and Heil, 2003, p. 61), who insist on a strong connection between realism and truthmaking, respond? The paper's novel claim is that accepting truthmaker-principles as part of the best abductive package of principles provides a strong argument for realism as the favoured view. An important part of the way to do this is to be explicit and emphatic about one's underlying metaphysical commitments,⁸ and not just to retreat to a sort of neutralism about one's truthmaking commitments, as some recommend in the face of Beebee and Dodd's (2005, p. 16) criticisms.⁹ Truthmaker-neutralists, who are moved by the critics, will answer the question, ‘What are the natures of truthmakers?’ with ‘We must stay neutral about this, not to beg the question against alternative theories’. Another way to look at such ecumenicality is that it is ultimately question-avoidance.¹⁰ Realists, by contrast, can abductively argue that realism about truthmakers is part of the best version of truthmaker theory. This leaves less brute and unanswered and is, thus, ceteris paribus, more virtuous.

As I argue in the next section, this is not to beg the question against antirealism, to preclude antirealist truthmakers, or to deny that one can have a theory of truthmakers as all mind-dependent. Instead, it is to argue that truthmaker-realism is the best of a variety of alternative theories of truthmaking.¹¹ One is arguing that such antirealists and truthmaker theorists alike do not have the best motivated theory on abductive grounds such as relevance and explanatory power.

In sum, critics of truthmaker theory have argued that truthmaking is neither sufficient nor necessary for realism and thereby rejected truthmaker-realism and truthmaker theory. They aren't arguing that truthmaker-antirealism is better than truthmaker-realism. They assume that the mere possibility of truthmaker-antirealism undermines truthmaker-realism and the realism-truthmaking connection. It was never clear, though, that their targets (e.g., Heil, Bigelow, Armstrong) thought truthmaking was necessary or sufficient for realism. Overall, this paper argues that the critics haven't achieved a rejection of truthmaker-realism. One can either accept truthmaker-realism, truthmaker-antirealism (including truthmaker-idealism or truthmaker-neutralism) or reject accounting for truth altogether. Generally plausible abductive reasons support truthmaker-realism over any of these other alternatives: realism and truthmaking go hand in hand as an abductively better package, or so I continue to argue.

To be clear: I haven't objected to the view that truthmaker theory is neither necessary nor sufficient for realism, but to the view that if truthmaker theory is neither necessary nor sufficient for realism, we shouldn't be truthmaker theorists or truthmaker-realists. I haven't argued that the arguments for truthmaker theory's being neither necessary nor sufficient for realism aren't sound. Rather, as arguments against truthmaker theory and truthmaker-realism the objections are invalid, given the abductive defence of truthmaker-realism. Next, I defend truthmaker-realism further – it doesn't beg the question against, but is instead abductively better than, its truthmaker-antirealist rivals: truthmaker-neutralism (§5.1) and truthmaker-idealism (§6). The abductive defence of truthmaker-realism thus responds both to truthmaking's critics and truthmaker-realism's rivals.

5. Objection-c: Begging the Question against Antirealist Truthmaking

Is an explicit commitment to realism really, as Beebee and Dodd (2005, p. 16) have pointed out and Tahko (2015, p. 211) has repeated, just begging the question against antirealisms and non-realisms?

In response: first, if truthmaker-principles and antirealism are compatible as critics emphasise in objection-b, then truthmaker-realism isn't question-begging. It takes additional argumentation and steps in reasoning to see that realism comes out as the most virtuous – a bias in favour of realism is not merely built into the principle, as one could legitimately argue if the argument were deductive and if antirealist truthmaking were made impossible. Within an abductive strategy, furthermore, truthmaker-realists need not directly rule out non-realist truthmaker theories. Instead, they can propose and defend their alternative as better. This section and the following sections argue for this in detail. Second, they need not appeal to independently plausible accounts of truth and truthmaking, as critics demand, to avoid the charge of begging the question (see Beebee and Dodd, 2005, p. 16; Tahko, 2015, p. 211; and Cameron, 2008, pp. 116ff.). Rather, if one is a realist and comprehensive, comprehensiveness being another theoretical virtue, one will propose an appropriately realist theory of truth and truthmaking (see e.g. Martin, 1984).

Cameron (2008, pp. 121ff.) responds to this problem which he also raised for truthmaker theorists by saying that they need to accept a further theory of fundamentality which they can then combine with their theory of truthmaking to rule out antirealist truthmakers. According to him, it is not the role of truthmaker theories but the role of theories of fundamentality to provide the world's truthmakers. This, however, is not the only attractive option for truthmaker-realists. They have at least two other options:

1. (a) provide a realist ontology, a domain of mind-independent entities, and reasons for identifying truthmakers in that domain (e.g., Bigelow, 1988, p. 123; Armstrong, 2004, p. 5; and Asay, 2020, §8 seem to take this option, though see below); or

2. (b) build into their views of truth and truthmaking a general commitment to realist, mind-independent truthmakers (Russell, 1912, pp. 119–30 clearly takes this route).

By doing either (a) or (b) one provides an account of what truthmakers are as a crucial part of one's truthmaker theory itself, not merely as a supplement – for surely an account of which entities are truthmakers for which truths is part of a truthmaker theory. (Even though he is a critic, MacBride (2020, §0) identifies this as one of truthmaker theory's central tasks, one interconnected with its other tasks.)

Providing a theory of fundamentality is just one route, but it doesn't work in the way that Cameron envisages (see also Schaffer, 2008; Rettler, 2016; and Heil, 2003; for criticisms, see Schipper, 2021; Audi, 2020). The problem with Cameron's view is that even if one provided a separate theory of fundamentality, one would not yet have a theory of truthmakers. This requires in addition a crucial step which is part of truthmaker theory: to say which fundamental entities are truthmakers for which truths (§5.1 defends this in detail).

Importantly, none of these routes is question-begging, including route-(b), if part of an abductive package. Non-realists can provide their own alternatives. Realists provide theirs. The field is then open to argue which theory is abductively stronger.

It is reasonable for realists to give realist accounts of truth and truthmaking. By adopting an overall realist position, one is committing oneself to accepting a systematic account of everything from a realist perspective. This plausibly includes, when providing a theory of truth, pointing to truth-independent objects as the reasons why what we say about them is true. This isn't question-begging. It's a natural consequence of being realist about one's commitments. One who takes everything to be language-dependent (e.g., Goodman, 1978), say, will be committed to answering philosophical questions and problems by analysing language or looking only to linguistic entities. Realists, by contrast, will solve philosophical problems and explain philosophically relevant concepts such as truth by pointing to mind-and-language-independent objects. One can articulate the truthmaking relation as above in terms of nec or sup, but it would be unreasonable to demand that truthmaker-realists should only be committed to something that antirealists can also accept, such as nec or sup, in their account of the truthmaking relation. For instance, one might add:

externality: the truthmakers, those entities on one side of the relation, are, except in exceptional cases,¹² outside of those truths that they make true.¹³

This is a claim about truthmaking being an external relation to outside entities. It is consistent with truthmaking being an internal relation in Heil's (2012, pp. 158–61) sense, on which all it takes for the relation to hold is for the truthmaker, e, and the truthbearer, s, to exist, for even though e is external to s, when they both exist, e makes s true. Externality in our sense is directly relevant to the realism debate – and has to do with relations, e.g., spatial, mereological, or instantiative. If x is external to y, then x does not occupy the same space as y, is not a proper part of y, and is not an intrinsic property of y. Idealists, pragmatists, linguistic constructivists, and other antirealists might not accept this additional principle, for they might accept principles on which only either s, a part of s, or some property of s (such as its expediency) makes s true. This principle is not just consistent with realism but reasonable for realists to accept.

The charge against realists of begging-the-question made by supposedly neutral, neutralist critics seems motivated by something like the following principle:

fair play: when giving an account of some subject-matter m, one must put one's account of m and the principles regulating m in such terms with which all parties in the debate can agree, including one's opponents and those who are yet undecided, or else one is begging the question against those opponents, that is, or else one isn't playing fair. We must all play fair!¹⁴

Critics seem to apply this supposed principle of philosophical disputes to the debate about truthmaking and charge realists with playing unfairly by begging the question against truthmaker-antirealists. Realists demand, minimally, that some truthmakers are mind-independent. Pluralists, such as Asay (2020) and Schipper (2016), can accept this. But not all parties in the debate can accept even this. Beebee and Dodd (2005, p. 16) put it starkly when they say that antirealists can successfully respond to realist arguments thus:

If truthmaker theory enshrines a commitment to realism, then presumably the appropriate anti-realist reaction to such an argument is simply to deny whatever truthmaker principle is being used as a premise in that argument.

However, is it reasonable to demand that realists follow fair play and to demand that antirealists can just respond by denying realists’ truthmaker-principles? What account of truthmakers should they give that allows them to play fair?

It would be equally unfair and unreasonable to demand that all truthmakers must be idealist and antirealist. The only alternative is some neutral, middle-ground conception of truthmakers potentially acceptable to everyone. But is such a conception a feasible alternative? I now briefly argue that neutralism is not an attractive alternative, especially if one needs to play fair against realism or, more generally, if one is moved by considerations of begging the question.

5.1 The Feasibility of Neutralism

To flesh out a middle ground, neutralist conception of truthmakers, one might draw on Cameron's (2008, pp. 115–16) other famous suggestion that we should separate (i) questions about whether truths have truthmakers from (ii) questions about what the right ontology is. Any commitments about the nature of entities would come at the level of ontology, leaving one's truthmaker theory neutral about the nature of truthmakers (e.g., Fine, 2017b, §2). Realists and antirealists can, then, both accept that some truth, T, is made true by reality. They might even agree on what entity, e, in reality makes T true. Where they disagree is on the nature of e. The existence and truthmaking role of e, as long as we are neutral about the nature of e (as the suggestion goes), is something about which both sides can agree. For instance, apples make ‘apples exist’ true. But according to neutralists, we can stay neutral about the nature of apples. To avoid begging the question against even truthmaker-antirealists and non-necessitarians, we should be maximally neutral about the natures of truthmakers, including whether they are mind-independent: truthmaking apples’ natures need merely allow them to be truthmakers – most minimally, they just need to exist. Both realists and antirealists can accept this maximally neutral truthmaker-principle:

tmp: truths T are made true by truthmakers e, except in exceptional circumstances.

Compare tmp with Tahko's (2015, p. 216) proposed ‘ontologically neutral’ truthmaker-principle:

tm*: Necessarily, if a proposition <p> is true and has a truthmaker, then there is some α in virtue of which it is true.

tmp has, by contrast, no commitment to propositions or to necessitation, which arguably, leads one to a theory of universals, acceptable to Armstrong but not to nominalists (see Tahko, 2015, p. 217). tmp allows for rejecting truthmaker-maximalism, on which all truths require truthmakers. tmp is perhaps neutrality-maximalised. Most importantly for our purposes, tmp has no commitment to realism.

No one need endorse tmp as enough for truthmaking. Its relevance here is that it takes the reasoning regarding the charge of begging the question to its natural conclusion: leaving us with a maximally neutral truthmaker-principle, which begs no questions against any potential truthmaker theory (including antirealist, non-maximalist, and non-necessitarian truthmaking). Similarly, truth-minimalists (e.g., Horwich, 1998) strip the truth-property of all metaphysical substance: ‘<p> is true iff p’ captures everything about truth. tmp is a minimalist truthmaker-principle, and truthmaker-neutralism is minimalist-truthmaking, stripped of all metaphysics.

Cameron's suggestion and this excision of metaphysical commitments from one's truthmaker-principle miss the point of providing a theory of truthmakers in the first place for at least two interrelated reasons. These reasons support the realist approach of option-(b) above, e.g., of Russell, rather than option-(a), chosen, e.g., by Armstrong and Bigelow.

Reason-1: contrary to what Cameron and Asay (2020; see below) argue, defending some ontology O as the right ontology is not the same as giving an account of truthmakers. Realists defend the view that at least some of reality is mind-and-language-independent (see what Brock and Mares (2007, ch. 1) call the existence and independence theses). But one can consistently be a realist while also arguing that because of the nature of truth, that is, because of something to do with what truths mean and how truthbearers can come to have the property being true, all truths can only be made true by mind-and-culture dependent practices. In this case, one will still be a realist since one defends the view that there are mind-independent entities; they just are not among the truthmakers. Also, one can be a thorough-going idealist and think that everything is mind-dependent but think that truth requires mind-independent entities. Depending on the full story of truthmaking one provides, one might then be an error theorist about most of what we say and believe (see Unger, 2006, ch. 7).¹⁵ Again, there is no guarantee that defending any ontology, or account of what is fundamental, entails any account of truthmakers or the truthmaking relation. Pace Cameron, and others such as Asay (2020, §1.3), one's ontology doesn't automatically give one one's account of truthmakers. More work is needed to connect one's theory of what there is (or one's theory of what's fundamental) with one's theory of truthmakers.

According to Asay's ‘conception of truthmaking as fundamentally an exercise in ontology’ (2020, cover blurb), ‘to be is to be a truthmaker’ (2020, §1.3). Asay is here clearly echoing Quine's (1948) famous criterion of ontological commitment and giving his own truthmaker-based criterion; since the equivalence goes both ways, he is also giving us a criterion of truthmaker-commitment, that is, to identify the truthmakers, identify what exists. Asay (2020, pp. 22–23) rejects truthmaker-fundamentalism (unlike Cameron),¹⁶ but accepts a similar relationship between truthmakers and ontology; he argues: ‘one doesn't “complete” the task of truthmaker theory – i.e., specifying what all the truthmakers are – until one has a completed ontological inventory. Since the truthmakers aren't a proper subset of what exists, they cannot be thought of as some special ontological category’ (ibid., p. 23). When one has a complete ontological inventory, one has a complete list of truthmakers, and vice versa.

But this conception of the relation between truthmakers and ontology is not necessarily correct, for instance if there are ineffable entities, about which one cannot articulate any truths and which thus cannot be truthmakers,¹⁷ or if truthmakers need not be entities (e.g., Melia, 2005, pp. 67ff.; Schipper, 2020, §1.1). Either way, having a complete ontological inventory or an inventory of all the fundamental entities, pace Cameron and Asay, doesn't allow us to conclude yet that these are all the truthmakers, the only truthmakers, or even any of the truthmakers. This requires further argumentation specifically linking what we can say and think truly and one's metaphysics. And one definitely needs further argumentation to establish that the truthmakers are the only entities in the inventory. These arguments depend on the resolution of difficult questions, for instance, about the limits of thought. Ultimately, I suspect that Asay's overall metaphysical view is realist (see also his useful account of realism in terms of truthmakers, 2020, §8.3) – my point is that theirs is not a view of the relation between truthmaking and ontology that truthmaker theorists must hold.¹⁸

The upshot is that to defend truthmaker-realism one needs a more direct argument within truthmaker theory. It is not enough merely to defend a realist ontology. Demanding that all truthmaker theorists must embrace an independent neutral principle such as tmp is just not feasible. This would, further, problematically rule out even other natural specifications of the truthmaker-principle, e.g., necessitation's necessity for truthmaking. Although this doesn't without further argumentation entail truthmaker-realism, it does have consequences for truthmaking's categorial ontology, e.g., a states-of-affairs or trope ontology. Just because accepting principles like necessitation would (potentially) beg the question against rival views about ontological categories (e.g., individuals-only ontologies, see Parsons, 1999), this doesn't mean that truthmaking itself is not a substantial relation, like necessitation. Taking the begging-the-question objection to its natural conclusion yields non-committal, tmp-like principles, but is not something truthmaker theorists should accept. Worse, it stacks the cards against truthmaker-realists and -antirealists alike, whose truthmaker accounts rely not just on proposing the right ontology but on an account of the relation between this ontology and truth, namely truthmaker theory. Doing ontology is part of doing truthmaker theory, but they are not the same – truthmaker-realists should do the latter: present a realist truthmaker theory – and this involves presenting a realist truthmaker-principle.

Also, Reason-2, although tmp is acceptable to the various sides of the discussion, it constitutes no account of truthmakers. It just says that some truths have truthmakers without saying anything about what they are. However, if one is a realist, why would one stop at tmp in one's account of truthmaking? A realist shouldn't shy away from giving an all-encompassing, comprehensive account from a realist perspective, as long as it remains theoretically virtuous; this includes giving realist accounts of truthmaking and of the nature of truthmakers. It is a natural extension of the aims of realists, when giving an account of truth and the truthmaking relation itself, to add that truthmakers are, exceptional cases aside, external to and independent of truthbearers and to minds.

This seems to follow too when one adds an abductively attractive relevance or aboutness-constraint on the truthmaking relation itself, as Smith (1999), Merricks (2007, ch. 2), Griffith (2015, pp. 326–27), and Schipper (2018, §4; 2022a, p. 998; 2022b, pp. 4–7; 2022c, §2) do, thereby moving beyond tmp in minimal abductively well-motivated ways. It is not just via one's account of one's ontology that one demands that the truthmakers must be mind-independent. When we talk about mind-independent entities, it is most plausible that only mind-independent entities make what we say true. But, in the exceptional circumstances that one is talking about mind-dependent cases, say if one is a psychologist who often talks about psychological states, then much of what one says is made true by mind-dependent reality. Similar points apply to linguists and language-dependent reality. One's commitment to realism can be part of one's account of truthmakers not only via one's realist ontology, but it can be introduced in one's account of the truthmaking relation itself, for instance, when one demands that it is constrained by what the truthbearers are about. In the cases where we are trying to say or believe truths about mind-independent reality, no truthmaking relation applies to those truthbearers unless realism is true.

Also, realists would be giving a more unified overall theory of everything if they were to use realist truthmakers to account for what makes a truthbearer true. The other possible options, recall, are:

1. (a) to give an antirealist account of truthmakers,

2. (b) to give some neutral account of truthmakers, or

3. (c) not to account for truthmakers at all.

We can rule (c) out as theoretically not virtuous for truthmaker theorists.¹⁹ If (a), then truthmakers are not mind-independent. If (b), then the (neutral) truthmakers are neither mind-independent nor mind-dependent. In this case, they are not mind-independent. In both cases, (a) and (b), realists would be accepting an account on which truthmakers are not mind-independent. Both are possible accounts even for a realist to accept, given the reasoning in Reason-1 above. However, neither is congenial to realists, who, once they go in for truthmakers, should aim for a unified account of reality and truthmakers.²⁰ Realists should not agree to either. Therefore, neither option is playing fair.

Another neutralist (b)-option would be to demand that all parties be non-commital or agnostic about the nature of truthmakers. But this merely kicks the can down the road. Ultimately truthmaker theorists must commit, otherwise they fail on comprehensiveness. Comprehensive neutralists aim to remain neutral about all ontological matters.²¹ But realist accounts would score poorly abductively if they had a realist ontology but ontologically neutral truthmakers, e.g., by claiming that all truthmakers are abstract, semantic posits (Fine 2017a, p. 561). For then one may ask whether the semantic posits are mind-and-language-dependent or -independent. If the former, neutralism collapses into truthmaker-antirealism, and if the latter, truthmaker-realism. Even if some form of neutralism doesn't so collapse, metaphysical realists would need good reasons why one's available realist ontology does not provide the truthmakers.²² One's realist theory would be simpler and more comprehensive if one provided a realist account of truthmakers. It is not question-begging for one to defend a unified view which is more congenial to one's overall philosophical outlook and makes stronger, more informative claims, actually telling us something positive about truthmakers. Therefore, it is not question-begging for realists about ontology also to be realists about truthmakers, even though weaker claims, in option-(b), are available to them. Realisms about truthmakers are not compatible with either neutralism about the nature of all truthmakers or the existence of all truthmakers: they need at least to say that some truthmakers exist and are mind-independent. In fact, beyond this, they could remain neutral about both the existence and natures of truthmakers, though on abductive grounds, a more definitive and precise answer would on the whole be preferable.

In fact, one should be sceptical about principles such as tmp without further explication of what all the terms are about and, crucially, what truthmakers are. tmp captures a core commitment of truthmaker theory but it is also vague and open-ended; so why think that it is at all satisfactory as even part of one's account? For instance, the (commonsense) realist answer to the question ‘What are truthmakers?’ is: truthmakers are just the ordinary mind-independent entities which we all thought we were talking about,²³ except potentially where matters are more complicated (e.g., talk about mathematical objects and science's theoretical entities) or specifically psychological (e.g., talk about emotions). This, then, gives us an answer to a question that must be forthcoming. Not to answer the question, as neutralism suggests (b and c), is to prevaricate and to leave one's truthmaker theory incomplete (and vague). Of course, one may take a more piecemeal approach and build one's theory by giving answers truth-by-truth or by leaving the ontological work to others.²⁴ But, ultimately one should opt for as much completeness as one can get. The realist answers are not so much question-begging as giving some of the more plausible, informative, general answers to pressing questions.

The neutralist alternative is not abductively feasible. Now, we have a defence of truthmaker-realism specifically in the face of antirealist and neutralist alternatives which addresses the classic objections head-on.

6. Two Abductive Arguments for Truthmaker-Realism against Idealist Rivals

As we saw, antirealists, including idealists, can accept tmp and build their own truthmaker theory. Truthmaker-idealists might accept tmp and add that all truthmakers are mental states such as beliefs, ideas, and experiences, and their properties.

However, in this section, I shall argue that:

1. (1) if one is an idealist, developing such an account of truth in terms of truthmakers is not attractive, and, specifically,

2. (2) on closer inspection, such antirealist truthmaker views, in particular the best forms of idealism, are not the abductively most attractive views to hold.²⁵

Although realism also doesn't entail tmp (or vice versa), I argue here that even a neutral principle such as tmp, ultimately, abductively favours truthmaker-realism. So, whether or not we are forced to accept a principle such as tmp, truthmaker-realism comes out better motivated. Even though idealism is prima facie compatible with tmp, closer examination reveals the unattractiveness of truthmaker-idealism. To show this, I present two abductive arguments against truthmaker-idealism: (a) the Perceptual Argument (against Berkeleyan truthmaker-idealism), and (b) the Argument from Unobserved Entities (against truthmaker-phenomenalism). These anti-idealist abductive arguments illustrate how the abductive defence gives us arguments for truthmaker-realism and against specific antirealisms, giving models for future arguments against other specific antirealisms.

6.1 The Perceptual Argument

First, Berkeleyan truthmaker-idealism is problematic because it fails to distinguish between the objects of perception and the perceiving of these objects, that is, the perceptual experiences themselves.²⁶ This is especially pressing for truthmaker theorists, for it is natural to try to use truthmaking as an account of the veridicality of perception – and the wider a theory's explanatory strength, the better.

But the structure of perception according to Berkeleyan idealists is not congenial to truthmaker theory, which needs minimally to distinguish between truthbearers, or more precisely veridicality-bearers,²⁷ here the perceptual experiences and what the truthbearers are about, the objects, that is, the truthmakers. If there is no distinction between perceptual experiences and the objects of perceptions, then one must give an account of the veridicality (i.e., truth) of perceptions differently. One option is to account for veridicality not in terms of the existence of a perception's objects. But this severs the link between perceptions and what they are about, opening their truthmakers to the charge of irrelevance and their theory to ad-hoc-ery. Truthmaker-realism, by contrast, can easily be applied to the standard accounts of perception, which are either directly or indirectly realist, by saying that the objects of perception make the perceptual experiences (or their contents) true – see Snowdon, 1981, p. 186; Johnston, 2006, pp. 278–79; and Brewer, 2011, p. 143, who employ making true and truthmakers in their favoured account of (veridical) perceptual experience.

When we look closely at the structure of the most plausible form of Berkeleyan idealism, a clear, more radical option emerges: we see that it should instead accept one of truthmaker theory's rivals, namely the identity theory or approach (e.g., Hornsby, 1997; Dodd, 2000; Johnston, 2013), on which experiences are veridical (true) when they are identical to their objects.²⁸ This fits the structure of perception according to Berkeleyan idealists perfectly. Truthmaking's asymmetric dependence relation is in fact not compatible with such idealisms. Therefore, Berkeleyans will have to account for the veridicality of perception in some other way within the resources of their view but without truthmaking, making at least Berkeleyan truthmaker theory inconsistent or unattractive.

There may be possible responses to this problem and the next, but these explanatory challenges clearly lead to complications that make the accounts potentially, unnecessarily complex and, thus, abductively weaker than their realist rivals.²⁹

6.2 The Argument from Unobserved Entities

Now, the (Abductive) Argument from Unobserved Entities.³⁰ Phenomenalism is faced with the following problem about unobserved entities: (a) according to phenomenalism, all objects, including ordinary objects such as watermelons and tins-of-beans are constituted by mental states – normally perceptions or experiences, of such objects; (b) phenomenalism needs to account for objects that are not perceived; but (c) it would be counterintuitive and against the common knowledge virtue (§2) just to reject that there is anything to be accounted for.

The standard non-theistic, non-Berkeleyan phenomenalist (most famously, Ayer's 1954) response to the problem is to say that an unobserved tin-of-beans's reality consists in something like this counterfactual being true: if a perceiver were in the right conditions, for instance, properly situated in relation to the kitchen cupboard, then the perceiver would experience tins-of-beans-like sensations. However, what accounts for the truth of these phenomenalist-enabling conditionals? They are contingent truths, so not true by logic alone. Therefore, there must be something in the world in virtue of which they are true – something that accounts for their truth.³¹ Normally, the tin-of-beans itself, if available, is what most plausibly accounts for our experience of the tin-of-beans, specifically the tin-of-beans being in the right conditions for being perceived. According to phenomenalists, however, the tin-of-beans doesn't just exist only when one has tins-of-beans-like experiences, but the tin-of-beans is nothing more than such experiences. Without there being an experience-independent tin-of-beans, however, one is hard-pressed to find what makes the conditional true. By stipulation, there just aren't any tin-of-beans-experiences now. Therefore, currently, there is no tin-of-beans yet, according to phenomenalism. Phenomenalism, then, given the resources of its theory of the world, is hard-pressed and seemingly unable to give further explanations and an ontological account when they need to be given. Phenomenalists cannot explain the truths that they posit to account for the intuition and datum about unobserved tins-of-beans.

Realism (about tins-of-beans), on the other hand, can easily say what in the world accounts for the truth of the counterfactuals: the tin-of-beans, its dispositions, and the dispositions of perceivers. Phenomenalism, thus, posits at least one series of truths as brute that realists need not posit as brute. Therefore, by abductive principles such as explanatory power, the truthmaker-maxim, and common knowledge, realism is abductively stronger than phenomenalism, at least on this point.

Rejecting the truthmaker-principle or the need to explain this truth won't help phenomenalists. Doing so (i) still leaves the truth unaccounted for, which, in the wider debate held on abductive grounds, leaves realists on firmer explanatory ground. And, (ii), denying that it is a problem that phenomenalism accepts more brute truths than other theories is not plausible, for it is a central abductive principle that theories are more virtuous if they posit fewer brute truths. Even if Bueno and Shalkowski (2020, p. 458) are right that this doesn't guarantee that the theories are true, at least they are more likely to be true based on all the relevant information that we have.

Phenomenalists might, instead, reject the datum about unobserved objects. They would, then, still have further explaining to do, which the realist does not: that is, phenomenalists would need to give a compelling explanation of why we think that objects don't just disappear when no one is perceiving them and that there are objects that exist unperceived, even though, according to their rejection of the datum, there aren't any such objects.³² Both options leave phenomenalists on abductively weaker grounds than realists.

This abductive argument for realism can also be extended to similar non-realist views which accept the truth of counterfactuals without having the resources internal to their theories to say what in the world accounts for their truth. Such theories are less attractive than those that have such resources. Although I won't spell this out here, this abductive strategy should also apply to truthmaker arguments against non-realisms about the past (see Parsons, 2005), against behaviourism (Liggins, 2005), against Rylean (Ryle, 1949, p.120), counterfactual accounts of dispositions (Martin, 1994/2008, p.19; Armstrong, 1969, p.23; 1989) and against various other similar views which posit brute counterfactuals to account for the data of ordinary experience (Armstrong, 1997, 2004).

7. Conclusion

If Williamson (2016, passim) is right that the argumentative approach of all philosophy is ultimately abductive, even if philosophers think otherwise about what they are doing, then philosophical disputes are not resolved by piecemeal argumentation. Rather, whole pictures need to be compared and weighed up for their virtues, not just piecemeal definitions (see also Quine, 1951). Abductive strategies have prominent advocates in metaphysics (see, e.g., Lewis, 1986, p. 4; Sider, Hawthorne, and Zimmerman, 2008, p. 6; Paul, 2011). Critics such as Bueno and Shalkowski (2020, passim) have argued that theoretical virtues are not truth-conducive, thus providing no reason to believe a theory. However, not even they deny that virtues provide reasons to accept the virtuous over their less virtuous rivals (ibid., p. 465).³³ Such strategies in defence of some theory T about m would minimally involve

1. (a) proposing a theory T of m, and

2. (b) defending T by arguing that it has explanatory and theoretical virtues over its rivals, such as parsimony, modesty, simplicity, elegance, relevance, and completeness.

Realist truthmaker theorists are, then, not begging the question against antirealists, nor is rejecting truthmaker-principles, as Beebee and Dodd (2005, p. 16) suggest, then, enough to nullify the challenge to non-realisms. Understood as an abductive challenge, truthmaker-realists, as Beebee and Dodd also demand, will ‘have reasons, independently of [their] commitment to realism, for believing that the principle is true’ (ibid.). Rejecting any of these generally attractive abductive principles and virtues would need to be well motivated and comply with further principles such as anti-ad-hoc-ery.³⁴ The main reason, I have argued, for accepting an explicitly realist principle that can be wielded in arguments against antirealism is this: realist truthmaker-principles have more theoretical virtues than antirealist ones. The relevant virtues are not merely that truthmakers explain truths, as many – for instance, Daly (2005) – have assumed is the only theoretical role that truthmakers play according to truthmaker theory.³⁵ Truthmaker-realism has a wider theoretical role and has more theoretical virtues that are relevant to its success. For instance, as I've argued, it gives relevant ontological accounts and leaves less unexplained than phenomenalism and other similar non-realist views do. So, truthmaker-realism is more plausible than these alternatives.

Indeed, if one utilises an abductive strategy to argue for truthmaker-realism and against -antirealist rivals, then one can see that there is no particular truthmaker-principle that antirealists and other, perhaps more neutral, critics may point to as question-begging. Understood in this light, one now has a response to the critics and a new argumentative strategy in defence of truthmaker-realism.