Review of Realisms Interlinked by Arindam Chakrabarti

Review of Realisms Interlinked by Arindam Chakrabarti
London: Bloomsbury, 2020

(Forthcoming in Mind)

1. Introduction

Realisms Interlinked is a sublime work. It reanimates theoretical philosophy with a distinctive synthesis of ideas and methods drawn from the commonsense metaphysics, epistemology, philosophy of mind, and philosophy of language of classical India (mainly via the Nyāya school) and $20^{\text {th }}$ century analytic philosophy (mainly via Strawson). In pursuing this agenda, the book lights up a new possible future for the study of world and mind, rooted in a borderless history of philosophy that should hopefully be better known among coming generations of analytic philosophers. Partly for this reason, it provides a refreshing contrast to other recent metaphysical and meta-metaphysical atlases like Sider (2011), Chalmers (2012), Thomasson (2014), Hofweber (2016), and Bennett (2017). It deserves to be read alongside these works; while Chakrabarti doesn’t discuss them and they don’t discuss him, a conversation between the traditions would, l’ll suggest, prove illuminating.

Before turning to a discussion of some key ideas in this wide-ranging book, let’s consider its overall shape. Although Chakrabarti is a realist, the primary goal of the book is to defend certain realism-links, which he first explored in his 1992 ‘I Touch What I Saw’ (which reappears as the first chapter). These links are neutral on the status of realism, serving instead to constrain one’s overall set of commitments, leading to consistent realism or anti-realism. Hence the book could have been given the clunkier title ‘(Anti-)Realisms Interlinked’. But while that alternative title would have better conveyed the primary goal, it would have failed to signpost the aims of many chapters, which are often friendlier to realism, and consistently opposed to subjective idealism. For besides defending metaphysical linking theses in chapters 1-2, 7, 16, and 24, the book also supports key tenets of Naiyāyika and Strawsonian epistemology, and critically discusses non-realist classical Indian schools (especially Buddhism and Advaita Vedānta) and anti-realist British philosophers (especially Dummett). This current in the book includes chapters defending the outer-sense perceivability of properties (ch.3) and the inner-sense perceivability of self and mental states (chs.12-14), a chapter opposing dream-based arguments for idealism (ch.8), chapters defending testimony as a basic source of knowledge (ch.19) and noninferential knowledge of other minds (ch.17), a chapter containing an approving discussion of ‘Idealist Refutations of Idealism’ (ch.6), and a chapter opposing Dummettian semantic claims from the case of our understanding of future-tense sentences (ch.23). The book also features chapters friendly to Nyāya’s more surprising realist commitments, including realism about absences (ch.25).

Bearing its primary and secondary goals in mind, the book can be read as favouring an integrated realist approach, in two stages. Once the metaphysical chapters give us the realism-links, the epistemological chapters support using these links to move from one

realism to another, by clearing space for non-inferential knowledge of various parts of the common-sense world (including properties, self, and other minds). Once aware of some part of reality (as commonsense epistemology permits), we can gain support for others via the realism-links and consolidate on a sweeping realist picture through the link-contributed coherence. In this way, realist light dawns gradually over the more neutral metaphysical core of the book. We might reason from realism about self to realism about world (with objects, properties, and a primitive tie of inherence), to realism about other selves, and perhaps even to realism about God (ch.24) and absences (ch.25).

Or so it may seem. But the undertones of the book are not unequivocally realist. Chakrabarti hints at the end that once common sense discloses the known world, what we mainly find is that the ‘world of ordinary objects and conventional truth is left un-denied as an object of non-attached double denial’, so that one ‘denies the denial of the world of ordinary experience’ (292). Upon reaching this stage, one achieves a ‘wonderful lightness’, and can shake off a desire for either ‘world-affirmation or world-denial’ (292), and hence decline to fly the flag for either realism or anti-realism.

This conclusion-a contented anti-anti-realism-is also available to a sophisticated non-realist or objective idealist, as Chakrabarti allows in discussing transcendental idealism and Advaita non-realism (p.75), despite his more Moorean tone elsewhere. This ambivalent side of the book could have been signposted more. It raises questions. If the most important links we get are between different applications of anti-anti-realism, one might expect cracks to appear once we consider possible links between different strengths of realism. Hence the book may be compatible with a more fragmented metaphysics than it seemed intended to support.

With that theme in mind, here is the plan for the rest of the review. In §2, I consider what Chakrabarti takes realism to require at different points, mention some distinctions between strengths of realism that seem underacknowledged, and argue that the realismlinks are only strong if the linked realisms are weak; here I focus on aspects of chapters 1-2, $5-7,21-23$, and 25 . I then note in $\S 3$ that there are different linking relations that should be more explicitly distinguished. I argue that the book gives insufficient reason to forge the strongest link between the realisms it considers. Here I describe some more fragmented outlooks (mixing realism of one kind with anti-realism of another) that seem consistent with Chakrabarti’s arguments, but conflict with Naiyāika-Strawsonian (meta-)metaphysics.

2. Grades of Realism and Some Subsurface Fragmentation

Part I defends the book’s core realism-links: (a) realism about the self is linked to realism about objects, (b) realism about subjects and objects is linked to realism about universals, and © these realisms are also linked to realism about inherence, the primitive tie of Naiyāika metaphysics. How are these realisms linked, and what does ‘realism’ involve in each case? I doubt that we should accept a uniform answer. If so, some significant disunity may outlast Chakrabarti’s arguments.

To uncover this disunity, let’s consider what realism and the links involve in each case, starting with objects and the subject-object realism-link. In the setup on pp.10-11, realism about objects is understood minimally, as a view that involves denying that objects are nothing over and above collections of experiences or qualities. So, following Chakrabarti’s presentation on p.11, realism about apples requires holding that they are (A) distinct from experiences of apples, (B) distinct from their qualities, © re-identifiable over time and across sense-modalities, and (D) intersubjectively identifiable. Realism about the self is understood on p. 11 as a related set of claims.

But Chakrabarti elsewhere strengthens this characterization, without enough support. For example, revealing his allegiance to the Nyāya school, he says that we must accept ‘dualism, that is, a full-fledged realism about substantial selves and material bodies’ if we

reject bundle theories (8). This is a starker choice than the choice between (A-D)-realism and its negation. A one-world transcendental idealist or Advaitin non-realist could hold (A-D)-realism while rejecting dualism. The chapter’s focus on the debate between Buddhist phenomenalism and Nyāya realism makes the discussion feel biased toward heavyweight realism. Consider how, when the foregrounded opponent is the Buddhist, Chakrabarti draws support from Frege, whose objects included third-realm entities about which objective idealism may be defensible (see Resnik (1979) and Sluga (1980)), and also from Kant’s refutation of subjective idealism. Yet Chakrabarti here takes realism about self and world to exclude transcendental idealism and non-dualism. For objects and subjects to be real according to Nyāya, it isn’t enough for there to be objective truths about them-not if ‘objective’ could merely mean intersubjectively valid and truth is minimalist truth (or if truth here can be modelled by superassertibility (Wright (1992)).

To understand the significance of this point for the dialectic, we can compare ideas in recent meta-ethics. Normative non-naturalism can be framed to require substantial truthmakers (e.g., Platonic reason-relations), but it can also dispense with substantial truthmakers, detaching objectivity from correspondence truth. These are different outlooks, as Parfit (2017) and Skorupski (2011) emphasize. Mimicking them, we might say that it is objectively true that there are selves and particulars, not just bundles of experiences or qualities, but remain unsure about whether the objectivity of this truth is explicable in the way ‘kneejerk’ realists like Sider (2011: 18) prefer. Hofweber (2016)'s and Thomasson (2014)'s metametaphysics may support this kind of line.

More concerningly for the realism-links, we might uphold heavyweight realism for objects but minimalist realism for subjects. This view is not excluded by ch.1. And it receives support, I think, toward the end of the discussion of self-reference in ch. 11 and in the discussion of other minds in chs.17-18. Consider how Chakrabarti approvingly discusses Abhinavagupta’s non-objectifying view of knowers (selves and others), writing that ‘[The ‘I’] exists, objectively, that is for all of us, but not as an object.’ The link between self and other minds in chs.17-18, together with the non-objectifying treatment of other minds in ch.18, further supports adopting different kinds of realism about objects and other minds, at least if ‘objects’ can be fully known in an objectifying way. Given a consistent view of self and other minds, that suggests treating the self differently from objects. One might take this view about God too (who enters in ch.24).

Let’s now consider some related observations about the discussion of universals and inherence. Chakrabarti links realism about subjects and objects with realism about universals in ch.2. I agree that, minimally understood, realism about one rationally requires realism about the other. But must we accept the same strength of realism about both? It seems coherent to be a heavyweight realist about objects but a minimalist realist or transcendental idealist about universals, considered apart from that in which they inhere. Indeed, the argument Chakrabarti uses to support this link only seems designed to exclude reductions of universals to sets of particulars. One can reject this reduction while taking a more minimalist view about universals (e.g., a ‘pleonastic’ view (cf. Schiffer (2005)) or transcendental idealist view). Another fragmented view would treat universals as less fundamental than thick particulars, or as ‘abstractions’ in Armstrong (1997)'s sense from other non-repeatable entities (e.g., events or states of affairs). Here one might consider a hylomorphic view that treats substances as basic, with universals and bare particulars viewed as abstractions from substances (as form and matter respectively), thus getting what Armstrong called the ‘victory of particularity’.

Of course, like Armstrong, one might hold these views while resisting any ‘anti-realist’ billing. But it seems coherent, even promising, to combine objective idealism about the world’s form with heavyweight realism about its matter. This view is untouched by ch.2’s arguments.

Related points arise with the next link, to inherence (ch.7). If we regard states of affairs as fundamental, we might view inherence as an abstraction, like universals and thin particulars. We might go farther, treating universals and inherence as having a different status from particulars while remaining within the Nyāya framework. As Phillips (1997: 50) notes, Udayana recognizes three grades of reality, with substances, qualities, and motions coming first, universals and inherence occupying a middle sphere, having being (bhāvatva) but not existence (sattā), and absences lying in an outer circle, amounting only to something which is real in the sense of being a padārtha (a nameable). If we took states of affairs to obtain wherever inherence relations obtain, the depiction of Nyāya metaphysical structure in Ganeri (2001: Ch.3) would provide a window onto a more Armstrongian view, which treated the nodes in the network of inherence and the network itself as an abstraction from the structured whole of the world. Important differences in ontological status may remain, then, if we accept links between the minimal forms of realism.

Let’s take stock before considering some final thoughts. The foregoing discussion suggests that different strengths of realism might be appropriate to different subject matters, and that the only realism that holds in all cases will be minimal. A corollary is that one may legitimately incline more toward anti-realism in some cases (self, universals, and inherence) than in others (objects or states of affairs). If our bugbear is the subjective idealist, we might correctly say we are ‘realists’ across the board. But different bugbears may loom for different subject matters. Objective idealism may fit with commonsense when applied to alleged third-realm entities like propositions (where Chakrabarti resists realism (see p. 18 and p.222)). It is, however, contrary to commonsense when applied to concrete objects. Commonsense hence seems not to be uniformly realist. Chakrabarti partly agrees: he is not a realist about third-realm entities. But there can coherently be more gradations than he believes.

Now, Chakrabarti does at different points give weaker and stronger characterizations of realism. At least six characterizations of realism appear:
(1) a minimal characterization as the negation of subjective idealism plus the view that there are ungrasped truths about the world (ch.1),
(2) a Dummett-style semantic characterization (p.26), which Chakrabarti later rejects (ch.23),
(3) a characterization requiring irreducibility (p.266),
(4) a characterization that entails the negation of objective idealism (ch.6),
(5) a characterization like the Lewisian understanding in terms of naturalness, suggested by a discussion of ‘universal blockers’ in Nyāya as disqualifiers of fullblooded realism about some abundant properties (see p.30), and
(6) an implicit characterization on p. 274 of realism as opposing all the anti-realist views discussed there, which include mere skepticism (following Kant, Chakrabarti counts this as ‘problematic idealism’).

It would have helped to get these distinctions up front, with some stock-taking of whether a given realism-link is supposed to hold for every strength of realism, or just some strength (p.284).

Although six characterizations might sound like plenty, others need attention. Ideas from the meta-metaphysical turn matter here. One might combine the first-order realism that Chakrabarti favors with something less than full-blooded meta-metaphysical realism. We

could agree that objects, subjects, universals, inherence, and even absences and God exist, while opting for what Hofweber calls an internalist reading of ‘exist’ in some cases. Or one might take all these things to exist, regarding existence questions as easy, but hold that some are more fundamental than others. Ganeri (2001)'s graph-theoretic portrayal could help in sorting out degrees of fundamentality: those entities that inhere in nothing (substances) could be treated as absolutely fundamental, qualities and parts less fundamental, and universals even less fundamental.

The basic observation is that when more distinctions are drawn, there is a risk to the significance of the linking project. Suppose we should only be lightweight realists about universals and selves. There must then be a thinning of the links: robust realism about objects will not entail a correspondingly robust realism about universals. To get a whole network of links, we may hence need to invoke very weak forms of realism. In the metametaphysical age, this weakening reduces the interest of the links: the hard questions in metaphysics are not about the truth of ‘realism’ understood so minimally.

3. Types of Linking Relation: More Subsurface Fragmentation

There is a further problem to discuss, which is that it is not obvious that the same linking relation is supported at all stages in the book. The strongest link is mutual entailment. But some of Chakrabarti’s arguments don’t establish this link. Defending this link between even weak forms of realism sometimes depends on questionable epistemological and methodological assumptions. Since a version of this point already appears in the literature (see Ganeri (2000) and Chadha and Nichols (2020) responding to Chakrabarti (1992)), I will be briefer with it.

Two kinds of links receive support in the book, though Chakrabarti assumes that they converge, given his epistemology and philosophical methodology. To see the first, weaker link, consider Chakrabarti’s first kind of argument. In chs.1-2, we get arguments from the unity of experience for the necessity of positing subjects, objects, and universals. Chakrabarti summarizes the idea as follows:

Both realisms derive support from the plain fact of perceptual re-identification of objects across times and senses. In the face of that fact, realism about a persistent substantial self would turn out to be inconsistent with anti-realism about external physical objects. (9)

The argument resembles the Kantian thought that the unity of experience requires an ‘I’ to do the unifying, though Chakrabarti goes beyond Kant in regarding subjects as substantial. Since the unity of experience partly consists in our stable awareness of objects, the facts of experience support belief in objects only if they also support belief in subjects. We get a different but related argument for the converse claim: to experience an objective time-order that could sustain one’s psychological continuity, there must appear to be a stable world from one’s perspective.

The link that is most directly supported by these arguments is the following:
Rationality Link: Experience rationalizes belief in objects iff it rationalizes belief in subjects.

As Chakrabarti also nicely puts it:
[A] realist conception of material objects over and above their properties and independent of experience of them compels us to believe in the existence of substantial selves. (14; my italics).

Hence these arguments are not directly arguments for:
Reality Link: Objects are real iff subjects are real.
Yet Chakrabarti makes stronger claims:
Selves over and above experiences are real if and only if objects over and above experiences are real. (10).

If one is a realist about objects, taking external things that many of us perceive directly to be independent of our perceptions, then, this book argues, one is logically committed to being a realist about subjects, selves as enduring substances owning mental states. (3)

Consider also, for example, the relation between realism about inherence and realism about the world:

Realism about the world logically entails and crucially requires realism about inherence. (77)

These claims matter for the overall linking project, which does not merely defend relations of mutual rational requirement between beliefs in different forms of realism.

How does Chakrabarti get the stronger claim for subjects and objects? Ganeri (2000: 646) suggests he gets it through a fallacy. I think, however, that the argument is more complex than it appears. First, Chakrabarti has independent arguments for the reality of selves (ch.16), and so takes himself to have sufficient independent reason to believe in selves, which gives reason to believe in objects via the Rationality Link. Secondly, Chakrabarti assumes that assertibility and rational belief go together: ‘To assert the existence of entities of a type is to pragmatically imply that the asserter knows that some things of that type exist and also knows what type he is referring to’ and hence it is ‘incoherent to add to the minimal existence-claim the strongly agnostic claim that all those recognitiontranscendently existing things of type $M$ might turn out to be…not of the type we know and refer to as M-type at all’ (24). Finally, Chakrabarti has a commonsense epistemology that entitles us to take experience at face value, barring defeaters. Hence, if experience supports one realism iff it supports the other (per the Rationality Link), and we take experience at face value, one might think realism about subjects and objects will stand together.

There is, however, a problem with this case for the stronger claim. Experience only provides defeasible reason for taking its content at face value. There could be defeaters for rational belief in some objects that are not defeaters for belief in subjects. For example, one might worry that common-sense realism about ordinary objects could be undermined by debunking arguments from the dependence of intuitions about object persistence and composition on folk teleology (compare Kovacs (2021)). Yet even if it were undermined, experience would retain the requisite unity, and unity-based arguments would continue to support the self. To accept such arguments, we needn’t be realists about objects in any stronger way than believing that veridical experience has intentional objects that are not private. So, one can take an ultima facie rational realist attitude about subjects but not ordinary objects.

Related concerns bedog the link with universals. Consider this argument for the objects-universals realism-link:

[I]f ordinary perceived particulars have to exist as the physical objects we often know them to be whether or not we can apprehend them correctly in detail on every occasion, they must have (or lack) some features objectively; that is whether the ones we identify as wooden tables are actually so or not, woodenness and tableness must be actually exemplifiable. For a non-particular individual like tableness, exemplifiability is enough for existence. Hence tableness has to exist independently of our awareness of it. Since bare or propertyless particulars would not be the macroscopic objects of ordinary perception, a world devoid of such actual features like cowness and treeness would not be a world where cows and trees can exist unperceived. (25)

As Chakrabarti acknowledges in mentioning Nyāya’s commitment to ‘universal-blockers’ (pp.30; 57-58) and attitude toward ‘titular properties’ (p.22), there is not sufficient reason to be uniformly realistic about all the ‘features’ objects appear to have. These blockers provide theoretical defeaters of the experience-given reason for robust realism about features. Yet all that is necessary for objects to appear in experience, according to the foregoing argument, is that they are experienced as having some features, not that these features be genuine universals.

If this is true, experience may provide undefeated support for objects without providing undefeated support for the universalhood of the features that objects appear to have. Hence we can at most get links of the following kind for objects, subjects, and universals:

Prima Facie Rationality Link: It is prima facie rational to believe in X’s reality iff it is prima facie rational to believe in Y’s reality.

If different ‘blockers’ to $X$ and $Y$ can be drawn from metaphysics, then we do not get the relevant instance of:

Ultima Facie Rationality Link: It is ultima facie rational to believe in X’s reality iff it is ultima facie rational to believe in Y’s reality.

If we can’t assert this link, we also cannot assert the relevant instances of:
Reality Link Schema: $X$ has reality iff $Y$ has reality.
To be sure, if we have default reason to take experience at face value, the burden is on the selective anti-realist to give selective defeaters. But there is insufficient attention to defeaters in the book. Most notably, the tension between the manifest image of descriptive metaphysics and the scientific image of many contemporary metaphysicians is neglected. The one discussion (pp.180-181) appears during a defence of folk psychology against fictionalism about the mental. While Chakrabarti reasonably worries that this fictionalist view is self-undermining, there is no general parallel response to defeaters of folk metaphysics from the scientific image, like recent debunking arguments against folk mereology and commitment to substances. More radical scientific challenges to descriptive metaphysics also needed attention; Chakrabarti’s claim that ‘[f]aithfulness to common linguistic usage is undeniably the starting point for all realists and descriptive metaphysicians’ overlooks, for example, Maudlin (2007)'s realist challenge to descriptive metaphysics. Unchecked by science or theoretical virtues besides intuitive adequacy, descriptive metaphysics risks spinning in the void.

4. Conclusion

My discussion has been selective-inevitably so: the book condenses in twenty-five chapters decades of work in every area of theoretical philosophy. My recommendation for anyone who has read this review is to get a copy of the book and read it right now. It is a remarkable achievement, and deserves very wide attention.

References

Armstrong, D. M. 1997, A World of States of Affairs (Cambridge: Cambridge University Press)
Bennett, Karen 2017, Making Things Up (Oxford: Oxford University Press)
Chakrabarti, Arindam 1992, ‘I Touch What I Saw.’ Philosophy and Phenomenological Research 52: 103-117
Chadha, Monima and Nichols, Shaun 2020, ‘Experiential Unity without a Self: The Case of Synchronic Synthesis.’ Australasian Journal of Philosophy DOI: 10.1080/00048402.2020.1836007

Chalmers, David 2012, Constructing the World (Oxford: Oxford University Press)
Ganeri, Jonardon 2000, ‘Cross-Modality and the Self.’ Philosophy and Phenomenological Research 61: 639-657.
—_2001, Philosophy in Classical India (London: Routledge)
Hofweber, Thomas 2016, Ontology and the Ambitions of Metaphysics (Oxford: Oxford University Press)
Kovacs, David Mark 2021, ‘Intuitions about Objects: From Teleology to Elimination.’ Mind 130: 199-213.
Maudlin, Tim 2007, The Metaphysics within Physics (Oxford: Oxford University Press)
Parfit, Derek 2017, On What Matters, Volume 3 (Oxford: Oxford University Press)
Phillips, Stephen 1997, Classical Indian Metaphysics (Delhi: Motilal Banarsidass)
Resnik, Michael 1979, ‘Frege as Idealist and then Realist.’ Inquiry 22: 350-357.
Schiffer, Stephen 2005, ‘Pleonastic Propositions’ in Beall, J. C. and Armour-Garb, B. (eds.) Deflationary Truth (Chicago: Open Court)
Sider, Theodore 2009, ‘Ontological Realism’ in Chalmers, D., Manley, D. and Wasserman, R. (eds.) Metametaphysics (Oxford: Oxford University Press)
—_2011. Writing the Book of the World (Oxford: Oxford University Press)
Skorupski, John 2011, The Domain of Reasons (Oxford: Oxford University Press)
Sluga, Hans 1980, Frege (London: Routledge)
Thomasson, Amie 2014, Ontology Made Easy (Oxford: Oxford University Press)
Wright, Crispin 1992, Truth and Objectivity (Cambridge MA: Harvard University Press)

KURT SYLVAN

University of Southampton, UK
k.l.sylvan@soton.ac.uk