Truth For The Rest Of Us: Conventional Truth In The Work Of Dharmakīrti

Philosophy ETDs Electronic Theses and Dissertations

Laura Guerrero

Recommended Citation

Guerrero, Laura. "Truth For The Rest Of Us: Conventional Truth In The Work Of Dharmakīrti." (2013). https://digitalrepository.unm.edu/phil_etds/3 This Dissertation is brought to you for free and open access by the Electronic Theses and Dissertations at UNM Digital Repository. It has been accepted for inclusion in Philosophy ETDs by an authorized administrator of UNM Digital Repository. For more information, please contact disc@unm.edu.

Laura P. Guerrero

Candidate Department of Philosophy

Department This dissertation is approved, and it is acceptable in quality and form for publication:

Approved by the Dissertation Committee:

Prof. John Taber, Chairperson

Prof. Kelly Becker

Prof. Jay L. Garfield

Prof. Russell Goodman

Prof. Richard Hayes

Prof. Paul Livingston

TRUTH FOR THE REST OF US: CONVENTIONAL TRUTH IN THE WORK OF DHARMAKĪRTI BY LAURA P. GUERRERO B.A., Philosophy, Willamette University, 2000 M.A., Philosophy, University of Hawai‘i at Mānoa, 2005 DISSERTATION Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy Philosophy The University of New Mexico Albuquerque, New Mexico July, 2013 DEDICATION For all my teachers, first and foremost my parents. TRUTH FOR THE REST OF US: CONVENTIONAL TRUTH IN THE WORK OF DHARMAKĪRTI By Laura P. Guerrero

B.A., Philosophy, Willamette University M.A., Philosophy, University of Hawai‘i at Mānoa Ph.D., Philosophy, University of New Mexico ABSTRACT

It is common in Buddhist philosophical literature to differentiate between two different types of truth: ultimate truth and conventional truth. For the philosophers of the Mahāyāna tradition of Buddhism, it is difficult to give an account of conventional truth that is both consistent with their anti-realist metaphysics (their ultimate position) and also robust enough to support truth as a normative concept. This dissertation addresses this problem by offering a deflationary interpretation of truth in Mahāyāna that is supported by a pragmatic account of intentionality and meaning. This account of meaning is developed from the work of the 7th Century Buddhist epistemologist Dharmakīrti. A careful reading of the Sanskrit source texts reveals that Dharmakīrti was alive to the problems of truth and objectivity in his tradition and sought to address them in his work. Dharmakīrti’s work can be read as offering a Carnapian-type solution to the problem of truth and meaning by way of an account of conventional knowledge that is grounded in what he calls arthakriyā – goal-driven human activity. Such an account is consistent with Mahāyāna anti-realist metaphysics, while at the same time providing an account robust enough to retain a sense of objectivity and to preserve a normative role for truth.

REFERENCES………………………………………………………………………....152

CHAPTER 1 INTRODUCTION This essay is concerned with truth. Specifically it is concerned with what Buddhists call conventional truth (saṃvṛtisatya), which is truth as it is ascribed to ordinary propositions asserted and believed in the course of our everyday lives. In the Buddhist tradition, it is common to distinguish this kind of truth from what is called the ultimate truth (paramārthasatya). Ultimate truth pertains to claims that reflect the actual nature of reality, or in other terms, that describe what is ultimately real. While the understanding of what is ultimately real changes throughout the course of the historical development of Buddhism, there remains in every tradition the idea that the actual nature of reality, the way it ultimately is, is not reflected in our ordinary discourse or in the way we commonly, i.e., conventionally, conceive of it. Despite this, however, we are told that we can still ascribe a qualified sense of truth, “conventional truth,” to such discourse insofar as that discourse plays an important practical and/or soteriological role in our lives.

Drawing a distinction between ultimate and conventional truth results in the problem of explaining the sense in which conventional truths are true. The two main general traditions in Buddhism, the Abhidharma and the Mahāyāna, have different conceptual resources at their disposal with which to address this problem. As a result, the kind of answers they each can give are very different. The Abhidharma tradition of Buddhism can give a reductive account that explains conventional truths in terms of their indirect reference to entities that are more fundamental and thus ultimately real. The Mahāyāna tradition, on the other hand, cannot employ such a strategy because Mahāyānists deny that there is any ontologically basic level to which conventional entities could be reduced. Therefore, the Mahāyānists have to give a different kind of account of truth at the conventional level than the account given by the Ābhidharmikas.

The concern that is the focus of this project is that of identifying an account of conventional truth that is both consistent with Mahāyāna ontological commitments and that is yet robust enough to support truth as a normative concept. In other words, I seek here to provide an account of conventional truth for Mahāyāna Buddhism that can explain the sense in which such truth is true.

The account I present here develops the suggestion made by Priest, Siderits, and Tillemans (2011) that the sense in which conventional truth is true for Madhyamaka is a deflationary sense. Expanding this suggestion to apply to all of Mahāyāna Buddhism, not just Madhyamaka, I argue that the answer that Priest, Siderits, and Tillemans offer to the question of conventional truth is a plausible one when it is appropriately supplemented with an account of meaning that is consistent with Mahāyāna anti-realist commitments. Deflationary accounts of truth, while not uncontroversial, have been rigorously defended by various philosophers as providing a minimal yet robust account of truth that preserves a sense of objectivity and can support the norms associated with truth. Presuming that such an account is defensible, it satisfies one of the criterion for an adequate account of truth for Mahāyāna: a deflationary account is robust enough to support truth as a normative concept, as Priest, Siderits, and Tillemans claim.

A deflationary account of truth also appears, at least on the surface, to satisfy the other criterion of an adequate account of truth for Mahāyāna, namely that of being consistent with Mahāyāna ontological commitments. It is consistent with these commitments because, as Priest, Siderits, and Tillemans point out, it is ontologically neutral; it does not presuppose any particular metaphysics. However, as Paul Horwich, a defender of a version of the deflationary account of truth, points out, deflationary accounts of truth presuppose that an independent account of meaning can be provided. Therefore, a deflationary account can be shown to fully meet this criterion only if it can be adequately supplemented by a theory of meaning that is consistent with Mahāyāna anti-realist commitments.

In order to show that the deflationary theory of truth does indeed satisfy this anti-realism criterion, I argue that the required anti-realist account of meaning can be provided by the work of the 7th Century Buddhist philosopher Dharmakīrti. His work provides an anti-realist account of meaning, grounded in a more fundamental account of intentionality, that can be used to supplement a deflationary account of truth. I explain that Dharmakīrti, focusing on ordinary empirical claims about things and events in our environment, provides an account of how it is that the mind can generate meaningful, intentional mental states about ordinary things and events without presupposing an ultimate ontology. This account, which describes how conceptual cognitions can be generated from non-conceptual ones, explains concept formation in pragmatic terms. According to Dharmakīrti, we form concepts in order to use them to help negotiate and predict our environment. This emphasis on use, I argue, makes Dharmakīrti’s account effectively a use theory of meaning. In fact, I argue that like Carnap, Dharmakīrti takes concept and language use to determine a conventional ontology, instead of presupposing one. Thought of in this way, I take Dharmakīrti to be, in effect, providing an account of what Carnap calls the “framework of things.”

I argue that Dharmakīrti’s anti-realist account of meaning provides the resources necessary to supplement a deflationary account of truth in such a way as to ensure that such an account is consistent with Mahāyāna anti-realism. Provided that a deflationary account of truth is robust enough to support truth as objective and normative, as philosophers such as Paul Horwich have argued it is, then I conclude that the suggestion made by Priest, Siderits, and Tillemans to understand conventional truth as true in a deflationary sense is a good one. When such an account is supplemented by the account of meaning provided by Dharmakīrti, it proves to be both robust enough to support truth as a normative concept and consistent with Mahāyāna ontological commitments.

Outline of Chapters

In chapter two I trace the development of the distinction between the two truths from the earliest sūtra literature up to the Mahāyāna tradition of Buddhism. I explain that the nature of the relationship between the ultimate and conventional truths for the Abhidharma tradition is reductionist and then show why explaining conventional truth in this way is not possible for the Mahāyāna given that tradition’s rejection of the Abhidharma account of the nature of ultimate reality. The contrast of the Mahāyāna with the Abhidharma makes clear the problem the Mahāyānists have of explaining the sense in which conventional truth is true: without any foundational ontological ground, how can the Mahāyāna draw a distinction between truth and falsity?

Focusing on attempts that have recently been made to understand the sense in which conventional truth is true from the perspective of the Madhyamaka tradition of Mahāyāna Buddhism, in chapter three I address in particular the fictionalist account of truth that is most often thought to reflect the Madhyamaka position. I argue that even though such an account does present a way of preserving truth in a context in which ultimate ontological considerations are irrelevant, a Carnapian approach, such as the one advocated by Finnigan and Tanaka (2011), is better. Without providing a non-arbitrary account of what governs the norms within a fiction, in particular what marks the distinction between truth and falsehood, a fictionalist account fails to give an adequately robust account of fictional truth. Furthermore, I argue that, since the Madhyamaka rejects any attempt at giving an ultimate ontology, there is no objective frame of reference to which the conventional could be contrasted in order to be understood as a fiction. In a context where everything is a fiction, it is unclear that the analogy with fiction helps to clarify the sense in which the conventional can still be true.

For these reasons, after addressing the fictionalist reading, I proceed to address the suggestion made by Finnigan and Tanaka (2011) to interpret the distinction between ultimate and conventional truth in terms of Carnap’s distinction between claims that are external and those that are internal to a framework. I argue that this approach has all the advantages of the fictionalist account but without having to adopt a pretense stance at the level of conventional truth; on this approach truth can be taken straightforwardly and literally.

A notion of truth that is general enough to operate across frameworks, and is thus consistent with a Carnapian-type account, is provided by deflationary theories of truth. In the remainder of chapter three I develop Priest, Siderits, and Tillemans’ deflationary account of conventional truth, highlighting its attractive features while drawing attention to Horwich’s observation that such an account can only be successful when supplemented with a truth-independent theory of meaning. This requirement motivates an examination of Dharmakīrti’s work in search of the necessary supplementary account of meaning, which I develop in chapter five along

Carnapian lines.

Before arguing that Dharmakīrti’s work provides a compatible theory of meaning, in chapter four I first defend my reading of Dharmakīrti as a Mahāyānist. Many scholars, while acknowledging that Dharmakīrti appears to ultimately endorse a Yogācāra (a school within the Mahāyāna tradition) position, interpret his account of pramāṇa, and its associated theory of meaning, from a realist, Sautrāntika Abhidharma perspective. These scholars argue that Dharmakīrti adopts both realist and idealist perspectives in different places of his texts and that his accounts of intentionality and meaning are given from the realist perspective. I resist the multiple perspectives interpretation and make a case for understanding Dharmakīrti as presenting a unified account that is given from the anti-realist, Yogācāra perspective. I contend, against the prevailing opinion, that Dharmakīrti’s account can be understood as internally consistent and thus that the multiple perspectives interpretation is unnecessary. In chapter four I also explain the sense in which I take all Mahāyāna Buddhists to be anti-realists.

Having made a case for reading Dharmakīrti as a Mahāyānist, in chapter five I present Dharmakīrti’s theory of meaning interpreted exclusively from an anti-realist perspective. Using the Carnapian strategy advocated by Finnigan and Tanaka (2011), I argue that in giving his account of pramāṇa, Dharmakīrti in effect offers an account of what Carnap calls the “framework of things.” This account, which is specifically focused on explaining how we come to understand the empirical world in terms of discrete and classifiable objects and events, is supported by a use theory of meaning. This account of meaning, which is grounded in a more fundamental account of intentionality, explains that the mind generates conceptual cognitions from nonconceptual ones for the purpose of using those concepts to guide behavior. I argue that because it is a context of pragmatic engagement that fixes the meaning of concepts in functional terms, the result is a pragmatist account of meaning as use. The use of concepts can then generate a framework that can support deflationary accounts of truth and reference.

I then in chapter six conclude the essay by summarizing the argument and drawing out some of its implications. In general, I take this essay to show that the account of meaning in Dharmakīrti’s work, when understood from his Yogācārin perspective, along with a deflationary account of truth, amounts to an account of conventional truth that is both consistent with Mahāyāna anti-realism and is robust enough to support the norms that we ordinarily associate with truth. In meeting these two criteria I conclude that the goal of explaining the sense in which conventional truth is true for Mahāyāna Buddhism can be achieved by appealing to the work of Dharmakīrti. In his distinctively Buddhist way, Dharmakīrti can thus explain “truth for the rest of us.” Giving an account of such truth is valuable for the Mahāyānist insofar as it gives her the resources to defend the Buddhist soteriological project, and more broadly to allow for the possibility of meaningful inquiry and knowledge even within an anti-realist context. Dharmakīrti

Before beginning the main argument, I would like first to introduce the main protagonist of this work: the philosopher Dharmakīrti. Dharmakīrti, who is commonly thought to have lived in India in the 7th Century, is one of the most influential thinkers in the history of Buddhist thought. Developing the work of his predecessor Dignāga, Dharmakīrti established the foundations of a system of logic, rhetoric and epistemology that came to be considered the definitive Buddhist position on these matters and helped to shaped the way in which philosophy was conducted in India during what Taber (2005, xi) describes as its “golden period” from approximately the fifth century to the thirteenth century.

Dharmakīrti’s principal text, the Pramāṇavārttika, is intended as a commentary to Dignāga’s Pramāṇasamuccaya. Dignāga’s work attempts to explain what constitutes a valid means of knowledge (pramāṇa) in order to correct the erroneous views on the matter presented by non-Buddhists (Nagatomi 1957, i-v). Dharmakīrti, picking up where Dignāga left off, uses the Pramāṇavārttika to defend the Buddha and his teaching against rival religious-philosophical positions. The style and method whereby Dharmakīrti defends Buddhism, like Dignāga, involves making clear the standards according to which the validity of claims is to be assessed. These standards are based on a sophisticated account of the means by which individuals come to have knowledge (pramāṇa). These means, which for Dignāga and Dharmakīrti are limited to perception (pratyakṣa) and inference (anumāna), are explained and defended in their works so as to be acceptable to anyone, that is, they do not rely on scriptural authority. In fact, Dignāga and Dharmakīrti use the doctrinally neutral standpoints that they develop to defend the authority of the Buddha and his teaching.

This method of defending Buddhist philosophical positions by grounding them in a clearly articulated account of the means of knowledge had significant impact on how philosophy was conducted in India. The polemical nature of Dharmakīrti’s work, along with its method, spurred on a lively period of interdoctrinal debate between Buddhists and non-Buddhists that is characterized by its emphasis on issues of epistemology and logic (McCrea & Patil 2010, 5). Dreyfus (1997) explains that the work of Dignāga and Dharmakīrti

helped create an epistemological turn that was part of a broader movement affecting the whole of Indian philosophy. These two authors were part of a vast movement of thinkers, Hindu and Jain as well as Buddhist, who turned away from traditional metaphysical questions such as the nature of the self and the path to liberation. Instead they devoted themselves to the study of epistemology. After this turning point, the philosophical quest seemed to bear less on metaphysical truth claims than on the basis for making such claims. The new concern with validating philosophical and religious claims through argumentation motivated thinkers to elaborate theories about the sources and types of knowledge. (16)

Throughout this period characterized by Dreyfus as an “epistemological turn”, it is the views of Dharmakīrti that are taken to represent the authoritative Buddhist position on these matters. The debates that continued long after Dharmakīrti’s death were conducted by Buddhists who carried his torch.

As the representative of Buddhism with respect to epistemological and logical matters, Dharmakīrti’s work had a profound and lasting influence on the development of Mahāyāna traditions of Buddhism. Dunne (2004) notes, for example:

Following upon the work of his predecessor Dignāga, Dharmakīrti addressed at length numerous questions that are of central concern to Buddhist thought and practice. The impact of his views on Buddhist theories of perception, inference, and language is difficult to overestimate. Indeed, it would not be outlandish to claim that his ideas are repeated in every Mahāyāna Buddhist philosophical work written after his time in South Asia. (2) Dharmakīrti’s views, while accepted as authoritative, were developed and changed by subsequent Buddhist commentators as they worked to address the various objections raised in the course of their exchanges with their non-Buddhist interlocutors. Differences among commentators gave rise to different epistemological views, all of which claim to represent Dharmakīrti’s true philosophical intention. Some of the characterizations of Dharmakīrti’s position made by subsequent commentators even go so far as to attribute to him positions that are the complete opposite of the views articulated by him (Dunne 2005, 9- 10). Despite these differences, however, Dharmakīrti’s views serve as the authoritative foundation for the development of all new epistemological positions within Buddhism, which inherit his vocabulary and general framework. And for those schools, such as some Madhyamaka schools that deny epistemology altogether, it is Dharmakīrti’s formulation of it that serves as the object of refutation (Dreyfus 1997, 19-20).

Because of the central position that Dharmakīrti occupies in the Buddhist philosophical tradition, his work is well positioned to provide a solution to the problem I am concerned with in this essay, namely that of providing a robust account of conventional truth that is consistent with Mahāyāna commitments. As a Mahāyānist himself (as I argue he is), he already shares the relevant commitments with his co-religionists and it is thus natural and charitable to read his work in a manner consistent with those commitments. When this is done, as I hope to show, an important result is an account of conventional truth and meaning that can be acceptable to even the most skeptical of Mādhyamikas.

In presenting the solution that I see in Dharmakīrti’s work to the problem of conventional truth, I take myself to be joining the efforts of the many commentators that have come before me, both ancient and modern. Like them, I aim to use the insights provided by Dharmakīrti’s work to solve philosophical problems that arise for Mahāyāna Buddhism. Unlike many of those commentators, however, I do not contend that the views I present will be the ones that Dharmakīrti himself intended; I recognize the futility of attempting to divine his true intention. My goal is only to arrive at a historically informed interpretation of Dharmakīrti’s work that thinks along with him, takes his views seriously, and can be used to develop and clarify his position in light of new objections, concerns, or contexts. In the case of the present project, the ‘new’ concern is really an old one -- that of explaining conventional truth -- made new by being framed within the context of twenty-first century analytic philosophy. The result of bringing Dharmakīrti into dialogue with Western analytic philosophy, as I do in this essay, is undoubtably something new, but it need not be the worse for that. The interpretation I provide may not be what Dharmakīrti could have himself intended, but it is consistent with his texts and, most importantly, in keeping with the spirit of his work. What I present here is what I imagine Dharmakīrti would say, were he able to engage in this inquiry in my stead.

CHAPTER 2 TWO TRUTHS

In this chapter I will briefly trace the historical development of the distinction between the two truths in Buddhism in order to highlight a fundamental difference between how the Abhidharma and Mahāyāna traditions of Buddhism understand the sense in which conventional truth is true. This difference results from the radically different ontologies that these two traditions have. The Abhidharma tradition maintains that reality is ultimately composed of entities called dharmas. In virtue of having this kind of realist ontology the Abhidharma tradition can account for truth at the conventional level by appealing to some kind of reductionist account that explains the referents of conventional discourse in terms of the ultimately real dharmas that constitute them.

The Mahāyāna tradition on the other hand, which is characterized in part by the rejection of the view that dharmas (or anything else) constitute the fundamental building blocks of reality, cannot in that way explain truth at the conventional level by appealing to some ultimate ontology. The anti-realist metaphysics of the Mahāyāna precludes giving such a reductionist account of conventional truth.

This important difference in ontology between the Mahāyāna and the Abhidharma helps to make clear the problem that is the impetus for this essay: given that a Mahāyāna Buddhist cannot appeal to some ultimate reality to explain the sense in which conventional truth is true, how can she explain it? In the chapters that follow I will examine various attempts to answer this question, as well as provide one of my own that is inspired by the work of Dharmakīrti.

Two Truths in Early Buddhism

The Buddhist distinction between ultimate and conventional truth is almost as old as Buddhism itself. Initially this distinction was drawn as a hermeneutical device to resolve apparent contradictions in the Buddha’s teaching by differentiating those parts of the teaching that require interpretation from those that could be taken literally (Newland & Tillemans 2001, 5). This first sense of the two truths doctrine originally stems, as has been argued by Karunadasa (2006), from a distinction drawn in the early canonical sūtra texts between discourses “whose meaning needs to be inferred” (neyyattha) and discourses “whose meaning has already been fully drawn out” (nītattha) (1). For example, in the Neyyattha Sutta of the Aṅguttara Nikāya it is stated that: Monks, these two slander the Tathagata [the Buddha]. Which two? He who explains a discourse whose meaning needs to be inferred as one whose meaning has already been fully drawn out. And he who explains a discourse whose meaning has already been fully drawn out as one whose meaning needs to be inferred. These are two who slander the Tathagata.

In this sūtra the distinction between what Karunadasa calls “direct and indirect meaning” is clear (1). A disciple of the Buddha, is expected to appreciate the difference between those teachings that require interpretation and those that do not. The doctrine of two truths is not explicitly present in this sūtra, or anywhere else in the sūtra literature; however, the distinction between direct and indirect meaning that is mentioned here is interpreted by subsequent commentators as alluding to a distinction between two kinds of truth. Karunadasa explains that in commentaries, such as the Aṅguttaranikāya Aṭṭhakathā, the passages that are to be taken literally, whose meaning has already been drawn out, are understood as those that express the ultimate truth (paramatthasacca). Those passages whose meaning needs to be inferred and thus require interpretation are then taken to be conventionally true (sammutisacca) (2).

To see the problem and the solution this initial distinction provides, take for example two separate passages from the Saṃyutta Nikāya, one of the texts from the Buddhist canon. In one passage, when one of the Buddha’s disciples, the bhikkhunī Vajirā, is asked by the Evil One, Māra: By whom has this being been created? Where is the maker of the being? Where has the being arisen? Where does the being cease? the bhikkhunī Vajirā responds in the following way: Why now do you assume ‘a being’? Māra, is that your speculative view? This is a heap of sheer formations; Here no being is found. Just as, with an assemblage of parts, The word ‘chariot’ is used, So, when the aggregates exist, There is the convention ‘a being.’ It’s only suffering that comes to be, Suffering that stands and falls away. Nothing but suffering comes to be, Nothing but suffering ceases. (Bodhi 2000, 230)

Here in this passage we have the bhikkhunī Vajirā giving an articulation of what is considered to be one of the Buddha’s most fundamental teachings: the doctrine of noself (anattā). This doctrine maintains that there is no underlying enduring substance, entity, self, or soul that makes a person the particular person that he or she is. Instead, the Buddha teaches that each “person” is really nothing more than a bundle of impermanent, dependently originating, and constantly changing factors he calls “aggregates” (khandha in Pali or skandha in Sanskrit) that neither individually nor collectively can constitute a being that endures through time and is the subject of afflictions like suffering. Now compare the previous passage with a different passage appearing earlier in the same text, where the Buddha is giving a lesson to a King:

If one regards oneself as dear One should not yoke oneself to evil, For happiness is not easily gained By one who does a wrongful deed. When one is seized by the End-maker As one discards the human state, What can one call truly one’s own? What does one take when one goes? What follows one along Like a shadow that never departs? Both the merits and the evil That a mortal does right here: What is truly one’s own, This one takes when one goes; This is what follows one along Like a shadow that never departs. Therefore one should do what is good As a collection for the future life. Merits are the support for living beings [When they arise] in the other world. (Bodhi 2000,168)

In this passage it appears that the Buddha is referring to the very entity -- the person or being -- that his doctrine of no-self articulated in the first quoted passage denies. In order to explain away this apparent contradiction, the commentarial tradition argues that passages like the first passage express ultimate truth while in passages like the second passage the Buddha expresses something that is only conventionally true and therefore cannot be taken literally as positing the existence of the self as an entity that persists through time and is the owner of its deeds. While the reflective position of the Buddha is that such a self does not exist, it would be inappropriate and confusing for him to say to the King in the passage above instead that “a set of aggregates at a future time will arise having the nature they do at that time because of the nature of the set of aggregates that exist now, to which they are causally connected and among which is included factors involving either evil or good action.” Such a statement might be closer to what is literally true, but it would be less effective at helping the king understand the importance of present action in determining future happiness or suffering. What the Buddha says is not false, it is just less literal.

In light of these considerations Karunadasa, agreeing with Jayatilleke (1963, 52), is keen to point out that as this distinction was initially understood neither of the two truths took priority over the other (10-11). At first, as its connection with the distinction between neyyattha and nītattha might suggest, the distinction between the truths was intended simply to reveal two different ways that the Buddha presented his teaching. In fact, that he was able to cater his teachings to the abilities and propensities of his students, as well as to the context of discussion, was taken to be indicative of the Buddha’s great skill as a teacher (Karunadasa 2006, 12). Teaching in terms of the ultimate truth and thereby being more literal is not, therefore, taken to be better, as the above example demonstrates. It is simply a more technical and specialized way of discussing what might otherwise be discussed more conventionally.

In order to illustrate this point, Karunadasa provides a “free translation” of some relevant passages in three Pali commentaries: Herein references to living beings, gods, Brahma, etc., are sammuti-kathā [[[Wikipedia:conventionally|conventionally]] true speech], whereas references to impermanence, suffering, egolessness, the aggregates of the empiric individuality, the spheres and elements of sense perception and mind-cognition, bases of mindfulness, right effort, etc., are paramattha-kathā [ultimately true speech]. One who is capable of understanding and penetrating to the truth and hoisting the flag of Arahantship when the teaching is set out in terms of generally accepted conventions, to him the Buddha preaches the doctrine based on sammuti- kathā. One who is capable of understanding and penetrating to the truth and hoisting the flag of Arahantship when the teaching is set out in terms of ultimate categories, to him the Buddha preaches the doctrine based on paramattha-kathā. To one who is capable of awakening to the truth through sammuti-kathā, the teaching is not presented on the basis of paramattha-kathā, and conversely, to one who is capable of awakening to the truth through paramattha-kathā, the teaching is not presented on the basis of sammuti-kathā. There is this simile on this matter: Just as a teacher of the three Vedas who is capable of explaining their meaning in different dialects might teach his pupils, adopting the particular dialect, which each pupil understands, even so the Buddha preaches the doctrine adopting, according to the suitability of the occasion, either the sammuti- or paramattha-kathā. It is by taking into consideration the ability of each individual to understand the Four Noble Truths, that the Buddha presents his teaching, either by way of sammuti, or by way of paramattha, or by way of both. Whatever the method adopted the purpose is the same, to show the way to Immortality through the analysis of mental and physical phenomena. (10)

According to the views laid out in these commentaries, whether the Buddha taught in terms of ultimate truth or conventional truth depended on the capabilities of his students. And since either way of teaching could be effective at helping the student progress along the path towards becoming an enlightened Arhat, neither was taken to be a superior form of truth.

Thus on this view there is nothing objectionable in talking in conventional ways, just so long as we don’t let our ways of talking confuse us into reifying the objects of our discussion. The technical and specialized mode of discussion is intended to help, in more reflective philosophical and perhaps meditative contexts, to remove such possible confusion. But it is not considered, at least initially, to be thereby a superior form of truth. In the context of early Buddhism it would therefore be helpful to think of the terms ‘ultimate truth’ and ‘conventional truth’ as differentiating between truth that is conveyed by means of conventions and truth that is conveyed by means of more discursive reasoning. This distinction indicates two different methods of arriving at the truth rather than as differentiating between two distinct kinds of truth. The Development of Abhidharma

After the Buddha’s death, the Abhidharma tradition of Buddhism developed out of an attempt to systematize the teachings presented in the sūtra literature. Noa Ronkin (2005) explains that: While the Nikāyas [[[sūtra]] literature] present the Buddha’s teachings as addressed to specific audiences at specific times and locations, the Abhidhamma [Skt. Abhidharma] seeks to describe the structure underlying the Buddha’s Dhamma [[[teaching]]] fully, in ultimate terms that apply in all circumstances. In this sense it marks the attempt to establish Buddhist thought as a comprehensive philosophy. (1) The Abhidharma, as Ronkin here explains, is concerned with distilling the Buddha’s core and “ultimate” teaching from out of the contexts in which the various sūtra teachings are embedded. By removing the conventional and context dependent aspects of the particular teachings presented in the sūtras, the Abhidharma thus hoped to preserve the essence of the Buddha’s teaching. Using the distinction between those teachings whose meaning must be drawn out and those whose meaning is explicit, the Abhidharma aims to systematically draw out the meanings of the neyyattha teachings so as to make explicit all of the Buddha’s teachings. In doing so the the Abhidharma presents all of the Buddha’s teachings in ultimate, as opposed to conventional, terms. This goal of presenting the Buddha’s teachings in ultimate terms is reflected in the name given to this tradition of Buddhism, abhidharma, or abhidhamma in Pali, which means “higher teaching” (Gethin, 203).

The way in which the early Abhidharma preserves the core of the Buddha’s teaching is basically by compiling detailed lists of the various things that the Buddha taught. These lists were probably developed from summary lists that were already present in the sūtra literature referred to as mātṛkā (mātikā in Pali) (Ronkin 2005, 28; Frauwallner 1995, 3). These lists, which were probably used as aids for memorization, included items like the four noble truths, the five aggregates, the four states of meditative absorption, the five hindrances, the six sense faculties along with the six corresponding object fields, the eightfold path, the seven factors of awakening, the twelvefold chain of dependent arising, etc. (Ronkin 2005, 27). These various elements all represent things that the Buddha taught and thus represent in shorthand his teaching as a whole. This teaching is taken to consist of both descriptive and prescriptive claims that reflect the Buddha’s understanding of the way things really are, the way things work, and the path an individual ought follow in order to escape from the cycle of rebirth, saṃsāra.

The early canonical Abhidharma developed, categorized, and expanded these mātṛkā so as to make explicit the details that are only implicit in the sūtra literature. Initially this amounted to nothing more than enumerating the various teachings that were scattered throughout the sūtra literature in order to preserve them (Frauwallner 1995, 8), however, as the Abhidharma tradition developed, more and more emphasis was placed on systematizing the teachings -- on explaining how the various elements fit together to form a systematic whole. This resulted in the proliferation of lists that became more analytically detailed both with respect to the basic elements of the teaching and how those elements relate to one another. The divisions of these lists are finer and more exhaustive (Karunadasa 1996, 4). For example, items such as the five aggregates come to be understood as general categories for what are considered to be more fundamental elements. Rūpa, the material form aggregate, in the Theravāda tradition, for example, is taken to be a class that contains 28 different kinds of more fundamental elements (Gethin 1998, 210).

The most fundamental elements of the Abhidharma lists are called dharmas (dhammas in Pali). Initially, these elements were so called simply because they constitute the elements of the Buddha’s teaching in general, which is also referred to as (the) dharma. This use of the the term to refer to things the Buddha taught as well as to the teaching in general is consistent with a use of the term dhamma in the the original sūtra literature. Ronkin (2005) explains that the term dhamma in the sūtra literature is ambiguous and that “[a]mong those meanings are included Buddhist teaching in general, any doctrine which forms part of that teaching, an element of experience, principle, phenomenon, nature, mental object, idea and others” (34). Picking up on these uses of the term dhamma, the early Abhidharma uses the term to refer to the wide variety of things referred to in their lists, which collectively constitute the content of the Buddha’s teaching in general -- his Dhamma. As the Abhidharma canons developed and the tradition attempted to exhaustively list and characterize all the dharmas and to explain their interaction, the term dharma came explicitly to take on ontological significance. Since the Abhidharma understood itself to be presenting the Buddha’s ultimate teaching, and the Buddha’s teaching is understood to present the the way things really are, it is not surprising that the Abhidharma comes to view the elements in terms of which it analyzes that teaching to be the ultimate constituents of that reality. The account that develops thus comes to understand the dharmas to be the basic building blocks of the experienced world. Paul Williams (2000) explains that,

Abhidharma texts set out to offer a list of all the types of factors into which experiences can be analyzed when we aim to find what is ‘really there’. They also explain how these link up causally and relate to each other in order to provide us with the actual world of lived experience. (90) The dharmas are understood to be those elements of experience that cannot be further analyzed. They are the elements that are explanatorily basic and that in conjunction with one another explain and account for the totality of experience. Different schools of Abhidharma give different lists of dharmas and understand their nature in different ways. The Sarvāstivādins, for example, maintain that there are 75 types of dharma while the Theravādins maintain that there are 82 (Gethin1998, 210; Williams 2000, 87). Despite their differences, however, all schools, at least in the post-canonical period, “attempt to give a systematic and exhaustive account of the world by breaking it down into its constituent physical and mental events (dharma/dhamma)” (Gethin 204). According to the Abhidharma, it is through attaining insight into the nature of reality -- both physical and mental -- understood in terms of its most fundamental elements that an individual can come to eradicate the ignorance that according to the Buddha is the fundamental cause of suffering. Of particular concern is the experience of the Buddhist practitioner at various stages of meditative attainment, the general patterns of which are laid out in the Buddha’s teaching (Ronkin 2005, 37). The dharmas, the vast majority of which are mental, account for the various experiences that may possibly occur along the path to complete spiritual attainment (Ronkin 50).

Two Truths in Abhidharma

The two truths in early canonical Abhidharma were probably understood in much the same way as they were in the sūtras, as two different ways to express and present the same reality. The Abhidharma is, on this understanding, exclusively concerned with the “ultimate,” in the sense of more technical and literal, manner of expression (Karunadasa 1996, 26-28). However, as the Abhidharma developed, both in the Theravāda and (especially) in the Sarvāstivāda Abhidharma traditions, the understanding of the two truths changed as the dharmas came to be understood in explicitly ontological terms. The two truths came to be understood in terms of two distinct kinds of reality: ultimate reality (the dharmas) and conventional reality (the things composed of dharmas). On this view, the ultimate comes to be understood as that into which the conventional gets analyzed -- an idea that is absent in the sūtra literature (Karunadasa 2006, 3-4).

Take for example the following exposition given of chapter 1, section 2 of the Abhidhammattha Sangaha, an Abhidharma commentary from the Theravāda tradition. In that section of the text, it is explained that in the Abhidharma, ultimate reality is constituted by four main classes of dharmas: consciousness, mental factors, matter, and nibbāna (Skt. nirvāṇa). Commenting on this section, Ven. Rewata

Dhamma and Bhikkhu Bodhi (2000) explain that:

According to the Abhidhamma philosophy, there are two kinds of realities -- the conventional (sammuti) and the ultimate (paramattha). Conventional realities are the referents of ordinary conceptual thought (paññatti) and conventional modes of expression (vohāra). They include such entities as living beings, persons, men, women, animals, and the apparently stable persisting objects that constitute our unanalyzed picture of the world. The Abhidhamma philosophy maintains that these notions do not possess ultimate validity, for the objects which they signify do not exist in their own right as irreducible realities. Their mode of being is conceptual, not actual. They are products of mental construction (parikappanā) not realities existing by reason of their own nature.

Ultimate realities, in contrast, are things that exist by reason of their own intrinsic nature (sabhāva). These are the dhammas: the final, irreducible components of existence, the ultimate entities which result from a correctly performed analysis of experience. Such existents admit of no further reduction, but are themselves the final terms of analysis, the true constituents of the complex of manifold experience. (25)

According to the view presented here, the objects of our ordinary experience, such as tables and chairs, are considered to be conceptually constructed composite wholes that are made up of dharmas. They are considered to be wholes because they are reducible by analysis to more fundamental components, and they are understood to be conceptually constructed because they are conceived to be unified, stable, and persistent despite the fact that what appears to be one thing is really many dharmas and none of those fundamental components endure for more than a moment. Ordinary objects are only conventionally real because their natures depend on other things -- on the more fundamental dharmas and on the conceptualizing activity of the mind. For this reason these ordinary objects are not afforded the same ontological status as dharmas, which do not depend on anything else for their natures. As Williams (2000) notes, “Each dhamma has its own specific characteristic by which it is recognized” (91), its salakkhana (Skt. svalakṣaṇa), and it does not therefore depend on any other dhamma for being the kind of dhamma that it is. It is irreducible to any other element. Dhammas also do not depend on the conceptualizing activity of the mind in order to have the particular character that they do. Each exists as it is by its own nature -- its sabhāva (Skt. svabhāva).

Another similar example of the way in which post-canonical Abhidharma comes to understand the two truths in terms of two realities is provided by Vasubandhu’s Abhidharmakośabhāṣya, a text from the Sarvāstivāda tradition. After indicating that the Buddha declared that there are two truths: conventional truth (saṃvṛtisatya) and absolute truth (paramārthasatya), Vasubandhu (1989) presents the following verse: The idea of a jug ends when the jug is broken; the idea of water ends when, in the mind, one analyzes the water. The jug and the water, and all that resembles them, exist relatively. The rest exist absolutely. (6.4)

He explains this verse in his commentary in the following way:

If the idea of a thing disappears when this thing is broken into pieces, then this thing has relative existence (saṃvṛtisat); for example, a jug: the idea of a jug disappears when it is reduced to pieces. If the idea of a thing disappears when this thing is dissipated, or broken to pieces, by the mind, then this thing should be regarded as having relative existence; for example, water. If we grasp and remember the dharmas, such as color, etc., in the water, then the idea of water will disappear.

These things, -- jugs, clothes, etc., water, fire, etc., -- are given their different names from the relative point of view or conforming to conventional usage. Thus if one says, from the relative point of view, “There is a jug, there is water,” one is speaking truly, and one is not speaking falsely. Consequently this is relatively true.

That which differs is absolute truth. If, when a thing is broken to pieces or dissipated by the mind, the idea of this thing continues, then this thing has absolute existence (paramārthasat); for example, physical matter. (910-911)

This passage from Vasubandhu is very similar in spirit to the account given in the Theravāda text quoted earlier. Both traditions of Abhidharma maintain two different kinds of real entities, ultimately real entities and conventionally real ones, and both traditions understand the ultimately real entities to be those which withstand reductive analysis. Conventional realities, on the other hand, do not withstand analysis.

In this passage Vasubandhu makes it clear that assertions that make reference to conventionally real entities are true conventionally in virtue of being in conformity to conventional usage. Since the entities referred to lack ultimate ontological status, and the way in which they are conceived as enduring, stable, unified wholes cannot be predicated of the underlying dharmas that ultimately make them up, the truth of claims that make reference to ordinary objects must be accounted for not simply by appealing to the underlying reality but must also by appealing to linguistic convention.

In order to explain why linguistic convention is important Siderits (2003) uses the classic simile of the chariot, used, for example, by the monk Nāgasena in the Milindapañha, as an example of how human convention shapes the way in which objects are understood. He explains,

Consider the set of chariot parts: wheels, felly, axle, etc. Now first consider the set when its members bear to one another the set of relations we might collectively dub the ‘assembled chariot’ relation: wheels attached to felly, felly to axle, etc. Second, consider the set when its members bear to one another what we might call the ‘scattered across the battlefield’ relation: a wheel beneath this tree, axle ten meters to the southeast of the tree, etc. We have a name for the set when its members are in the ‘assembled chariot’ relation, but we have no name

for the set when its members are in the ‘scattered across the battlefield’ relation. It is obvious why this should be so: we have an institutionalized use – as a means of transportation – for the parts when assembled, but there is no institutionalized use for us of the parts when strewn across the battlefield that way. Now consider that we tend to readily conceive of the set as a single entity when its members stand in the first relation, but not when they stand in the second relation. This makes it clear that our ontological intuitions are being guided by our institutionally arranged interests… if we wish to know the ultimate nature of reality, we would do well not to allow our views to be shaped by conventions that reflect our interests. The ultimate nature of reality is how things objectively are – independent of our subjective wants, needs, and interests. (7-8)

Here Siderits illustrates the idea that it is in virtue of the kinds of creatures we are, with certain interests and needs, that we come to see and understand the world in the way that we do. Thus, one of the motivations of the Abhidharma for denying that composite objects are ultimately real is that their particular composition is something that rests in part on the activity of the mind. Thus for the Abhidharma, wholes borrow their nature not only from the dharmas that make them up, but also from the interests of the beings that take certain bundles of dharmas (but not others) to be unified wholes. Our conventional practices and pragmatic interests play a role in determining which bundles of dharmas we consider to be discrete objects and which we don’t. Since ultimately there are no such discrete objects, the positing of them must be due to the constructive activity of the mind. And whether or not our assertions

with respect to these kinds of conventional objects are true will thus depend in part on the way we use language conventionally to help carve up the world. It is important to note here that the mental construction involved in understanding the world as composed of unified, stable, and enduring objects is not something that the Abhidharma maintains that we do either consciously or deliberately. The vast majority of this kind of mental bundling happens at what we would call a subconscious or sub-personal level. However, insofar as this activity occurs, it distorts the actual nature of reality. Even instances of what we would call natural kinds are artificially constructed entities that derive their nature both from the elements that make them up and the constructive

activities of the human mind. The account presented above suggests a reductionist interpretation of the relationship between ultimate and conventional reality. Williams (2000) explains: The Abhidharma is characterized by some sort of reduction. Throughout this reductive process the search is driven by a quest for what factors, what elements, are actually there as the substratum upon which the forces of mental imputation and reification can form the everyday ‘life-world.’ An ‘[[Wikipedia:Absolute

(philosophy)|ultimate]] truth/reality’ is discovered as that which is resistant to attempted dissolution through reductive analysis. (93) This interpretation of Abhidharma ontology as reductive has been most rigorously argued for by Siderits, who characterizes the Abhidharma account as a form of ontological reductionism and then shows how such a view works to explain the Buddhist doctrine of the two truths.

Siderits (2003) defines ontological reductionism as the view that “holds that a certain sort of thing that is ordinarily thought to exist turns out to be reducible to certain other sorts of things that are in some sense ontologically more basic” (1). In order to help elucidate this particular ontological position he asks us to imagine a linguistic community that makes regular reference to things of kind K and then juxtaposes the reductionist position to the other two possible ontological positions one could take with respect to Ks, non-reductionism and eliminativism (1). A nonreductionist about Ks will claim that Ks are ultimately real and that they thus should be part of any complete theory about the nature of reality. Non-reductionists about Ks are realists about Ks and claim that Ks are among the class of entities that really exist.

The eliminativist about Ks, on the other hand, will claim that there are no such things as Ks. Eliminativists will argue that the positing of Ks rests on an adherence to a false theory and thus that all talk of Ks should be expunged from the language as a bunch of non-sense. The example that Siderits uses to illustrate this view is that of understanding human illness in terms of disease-causing demons (2-3). There was once a time when people explained human illness by claiming it was caused by possession by incorporeal demons. Siderits invites us to imagine a community of people who shared this theory of about human illness and who made regular reference to the various demons they took to be responsible for various illnesses. He explains that:

The eliminativist about demons would obviously claim that demons do not exist. For now we know that diseases are caused by microbial infection, not demon possession, and our only reason for supposing there to be demons had to do with the explanatory role they played in the now-discredited theory. (3) Such an eliminativist would be in favor of replacing demon talk with language that reflects the theory that has supplanted the old, false one. “The theory in which such talk is ensconced is simply incompatible with the theory we now accept. Demons should be eliminated” (3). Likewise, those who are eliminativist about Ks will say that we should replace talk of Ks with talk that is in congruence with the theory that has supplanted the theory of which Ks were a part. We should eliminate Ks from our ontology and our language.

With respect to the Abhidharma case, given that the Buddha himself, as well as subsequent generations of Buddhist teachers, employed terminology that they ultimately thought reflected an inaccurate description of the nature of reality, the Ābhidharmikas are not eliminativists about wholes. Insofar as such Buddhists wish to retain a place for such language in their philosophy, as we will see, their position is thus better understood as reductionist rather than eliminativist.

The reductionist about Ks will maintain, Siderits explains, that “while Ks may be said to exist (pace the eliminativist), their existence just consists in the existence of things of a more basic sort, things of which the Ks are composed, so that (pace the non-reductionist) Ks do not belong in our final ontology” (1). The reductionist, unlike the eliminativist, retains talk of Ks and thus is willing to grant a sense in which they exist, because unlike the demons of the example above, talk of Ks turns out to be a useful way to refer to what is found to be more ontologically basic and thus to be ultimately real. For the Ābhidharmikas, these ontologically basic and ultimately real entities are the dharmas.

if there is a relationship between the theory of which Ks are a part and the more foundational theory we now find to be more accurate and we can explain the phenomena referenced in Ks’ theory in terms of phenomena at the ultimate level, then it is possible to use talk of Ks as a method of talking about the real phenomena that are explanatorily more basic. This will be desirable when talk of Ks is pragmatically preferable to talk of the ontologically more basic entities. With respect to the Abhidharma case, just like organic chemistry talk, for example, is still useful despite the fact that we now know that quantum mechanics is more fundamental, so too, the Buddhist will argue, talk of tables and chairs is still useful despite the fact that we now know that they are really composed of dharmas that are more fundamental.

With reference to just such a reductionist account, Williams (2000) makes it clear that a reductive view about ordinary objects is not the same as denying their reality. He says,

Note (and this is important) that to be a conceptual existent (you, me, a chair, table, or a forest) is not thought in Sarvāstivāda to be the same as not existing at all. It is to bear a particular sort of existence, the existence of an entity that is quite correctly treated as a unity for pragmatic purposes but nothing more. It can be analyzed into a plurality of constituents which are thus to be taken as ontologically more fundamental. A conceptual existent is the result of a particular sort of causal process, a conceptual reification or unification out of a plurality. (94)

Conventional objects are real, on this view, but derivatively and thus not in the same sense in which dharmas are real. While the ordinary objects of our experience are considered to be ultimately unreal, they are taken to be conventionally real, and our talk about them conventionally true, because these ‘objects’ are made up of real dharmas and talk about these objects thus makes indirect reference to entities that are ultimately real. The truth at the conventional level can thus be accounted for in terms of truths at the ultimate level. Furthermore, by explaining conventional entities in terms of dharmas, the Ābhidharmikas are able to explain why it is that beings like us came to understand the world the way we mistakenly do. We do so because this mistaken worldview is based in something that is real, only misunderstood.

Two Truths in Mahāyāna

In the next chapter I will explain in detail various attempts made to explain the doctrine of the two truths in the Madhyamaka tradition of Mahāyāna Buddhism and the sense in which conventional truth is true in those accounts. Here I want to conclude this chapter by quickly giving a characterization of Mahāyāna Buddhism in general and then explaining why the doctrine of the two truths is particularly problematic for the Mahāyāna. This problem is made especially clear by contrasting the Mahāyāna position with that of the Abhidharma tradition just dealt with.

Scholars believe that the earliest Mahāyāna texts began showing up around the first century BCE (Gethin1998, 225). Evidence suggests that the origin of Mahāyāna was not sectarian but rather that it began simply as “esoteric teachings of interest to small groups of monks from various ancient schools” (Gethin 225). These teachings emphasized a group of related ideas that came to characterize the tradition. According to Peter Harvey (1990), the Mahāyāna tradition of Buddhism has three main characteristics: (1) “a wholehearted adoption of the Bodhisattva path,” (2) “a new cosmology arising from visualization practices devoutly directed at the Buddha as a glorified, transcendent being,” and (3) “a new perspective on Abhidharma, which derived from meditative insight into the deep ‘emptiness’ of phenomena, and led to a new philosophical outlook” (89-90). It is this last characteristic that is relevant to the discussion of the nature of the two truths.

The “new perspective” offered by the Mahāyāna is the perspective that understands all phenomena to be empty of svabhāva. The term svabhāva, or sabhāva in Pali, is used by the Abhidharma to characterize dharmas. According to the Abhidharma, dharmas are the ultimately existing foundation elements precisely because they have their own natures, svabhāva, and thus do not depend on any other thing for their nature -- not a conceptualizing mind or any other existent. In claiming that all phenomena are empty (śūnya) the Mahāyānist rejects the ultimate reality of the dharmas (Harvey 97).

Harvey explains the Mahāyāna attitude towards dharmas, as reflected in the Perfection of Wisdom (Prajñāpāramitā) literature, in the following way: The Perfection of Wisdom literature... regards all dharmas as like a dream or magical illusion (māyā). There is something there in experience, and one can describe it well in terms of dharmas, so it is wrong to deny these exist; yet they don’t have substantial existence either. What we experience does not exist in an absolute sense, but only in a relative way, as a passing phenomenon. (98)

In this passage Harvey explains that for the Mahāyānists, the dharmas are considered to be as conventional as ordinary objects. In a conventional sense it may be permissible to claim that they are real and exist, but the Mahāyāna deny the Abhidharma claim that dharmas are the ultimate real constituents of reality. Gethin (1998) explains that:

Central to the Abhidharma is the distinction between the conventional truth (that persons and selves exist) and the ultimate truth (that persons and selves are ultimately simply aggregates of evanescent dharmas -- physical and mental events). The main teaching of the Perfection of Wisdom is that from the perspective of perfect wisdom, even this account of the way things are is ultimately arbitrary. (235) According to the Mahāyāna the Abhidharma did not go far enough in their analysis and thus failed to realize that the dharmas also were only conventional entities that derive their nature from other (equally conventional) things and, importantly, from the conceptualizing activity of the mind.

The Mahāyānist not only deny the ultimacy of dharmas: they deny that any entity has such ultimacy. All things are characterized by emptiness (śūnyata). What this means is that nothing can serve as ultimate ontological foundation or ground. This means that the method of explaining truth at the conventional level in terms of a reductive ontological dependence of conventional referents to ultimately existing entities is precluded

for the Mahāyānist. Whether or not the Ābhidharmika can in the end be successful at giving a reductive account of the two truths, this reductive approach is at least available to her as an option given her ontological commitments. This is not the case for the Mahāyānist. While it is still clear that Mahāyānists maintain a distinction between the two truths (Harvey 2000, 98), they cannot account for truth at the conventional level by appealing to some ultimate level because Mahāyāna rejects that there is any ultimate ontological ground.

It is this ontological anti-realism that creates a problem for understanding truth in the Mahāyāna tradition of Buddhism. If there is no ultimate ground, no ultimately existing entities, then the question arises: how and in virtue of what are claims assessed for truth or falsity? There appears to be no objective standard to which the Mahāyāna can appeal. How then can the Mahāyāna retain a meaningful sense of truth in general, and between two senses of truth in particular? This is the problem that serves as the impetus for this project.

Giving an account of non-arbitrary standards according to which any claim can be assessed for truth is particularly important for the Mahāyāna because, as Gethin (1998) points out, according to the Mahāyāna even “the conceptual constructs of Buddhist theory are ultimately no less artificial and arbitrary entities than the conceptual constructs of the ordinary unawakened mind which sees really existing persons and selves” (236). The Buddhist path, including the insights of the Mahāyāna itself, are all on this view [[Wikipedia:Convention

(norm)|conventional]]. If the Mahāyāna wants to defend the Buddhist path as the conventionally right one, as opposed to the paths proposed by others in the religious marketplace, then there must be some non-arbitrary standard according to which to assess and compare rival positions. However, it is unclear what that standard is or could be for the Mahāyāna. In the next chapter I will explore some attempts made by others to explain the sense in which conventional claims are to be understood as true in the Mahāyāna before I proceed in the subsequent chapters to present my own account.

CHAPTER 3

TWO TRUTHS IN MADHYAMAKA BUDDHISM

Much has been written regarding how the relationship between the two truths is (or should be) interpreted by the Madhyamaka schools of Mahāyāna Buddhism -- in particular how issues surrounding this distinction have been taken up by the tradition in Tibet. Since issues that arise in the Madhyamaka context are similar for all Mahāyāna Buddhists, as I will explain, it will be fruitful to examine the ways in which this issue has been dealt with in the Madhyamaka context.

First I will explain the sense in which all Mahāyānists are anti-realist and why, in light of their anti-realist commitments, the Mādhyamika like all Mahāyānists must explain the standards that govern conventional truth in a way that does not rely on some ultimate ontology. I will then address a couple of the most promising ways to understand conventional truth in light of that anti-realist position.

The first view I will address attempts to understand Madhyamaka as a kind of fictionalism. While this view explains how truth at the conventional level could be insulated from truth at the ultimate level, I will argue that this position is in tension with Madhyamaka anti-realism. I will also raise the challenge that it lacks a nonarbitrary account of the norms that govern the fiction. If the Mahāyāna want to retain a robust notion of truth at the conventional level, then some non-arbitrary account must be given of what marks the difference between truth and falsity.

Another view that I will address is the view that interprets the distinction between ultimate and conventional truth in terms of Carnap’s distinction between claims that are external and those that are internal to a framework. This view has all the advantages of a fictionalist view but without having to adopt at the conventional level a pretense stance with respect to truth claims. This kind of a view also helps to temper the concern that conventions are arbitrary by incorporating pragmatic considerations to regulate the acceptance of conventional frameworks.

While the Carnapian approach is very promising as a general strategy, as it stands the view needs elaboration in order to make clear that it can be used to support an account of conventional truth. I will leave this issue undecided in this chapter but will revisit it in chapter five, where I will attempt to employ and elaborate this general strategy in giving Dharmakīrti’s account of meaning. There I will use resources from Dharmakīrti’s philosophy to flesh out the Carnapian account in a distinctively Buddhist way.

The last view I will discuss in this chapter is the deflationary account provided by Priest, Siderits, and Tillemans. I will explain the virtues of this view with respect to Madhyamaka ontological commitments and show that while promising, this view requires a supplementary account of meaning in order to be a successful account of conventional truth for Mahāyāna. In chapter five I will look to the philosophy of Dharmakīrti for the resources to provide a Buddhist account of meaning that could supplement a deflationary theory of truth.

Mahāyāna Anti-realism

In the last chapter I characterized Mahāyāna as anti-realist. Before proceeding to examine some attempts to understand conventional truth in Madhyamaka, I want to make clear the relevant sense in which all Mahāyāna Buddhism is anti-realist and thereby make clear the position that the Mādhyamikas share with their Mahāyāna co-religionists.

Characterizing Mahāyāna as anti-realist may seem inaccurate, given that much of Mahāyāna is often taken to accept that reality has an ultimate nature. While all Mahāyānists maintain that all phenomena are empty (śūnya) and reject the ultimacy of dharmas, some Mahāyāna schools such as the Yogācāra/Vijñānavāda, Pure Land, and Zen schools-- even the Madhyamaka by some interpretations -- still claim that there is an ultimate reality that underlies ordinary experience. For these Mahāyānists, the lesson to be drawn from the doctrine of emptiness is not that there is no way that reality absolutely and ultimately is, but rather that absolute reality as it exists in itself is ineffable and beyond conceptualization. Siderits (2003) explains that “the idea seems to be that if no real entity may be said to have its own essence, then the reals must lack determinate natures, and so must somehow transcend all efforts at conceptualization” (132).

For the Yogācāra/Vijñānavāda tradition of Mahāyāna, for example, we are told that the experienced world is nothing more than a continual flow of ideas (vijñapti-mātra) that when experienced in its perfected (niṣpanna) nature reveals the way things really are, a pure tathatā (‘thusness’ or ‘suchness’). The ordinary unenlightened mind, we are told, experiences the world as composed of distinct subjects and objects. According to the Yogācāra, this, however, is just an illusion.

All there is is just an ever-changing flow of ideas or information (vijñapti) that in itself is non-dual and non-conceptual. The mind superimposes conceptual distinctions and then mistakenly takes these to be real. However, ultimately they are not -- they are empty. Understood in a metaphysical, even an idealist, way, this view seems to posit an ultimate reality -- the ineffable ‘thusness’ of the flow of ideas. Another, related, Yogācāra position goes further to posit an even more fundamental Absolute. Wood (1994) explains

that the Vijñānavāda “Absolute” was not mind (vijñāna) -- not even absolute or non-dual mind -- but rather non-mind (acitta). (Mind, however, was its manifestation.) The Vijñānavādins identified this Absolute with emptiness, and its attainment was said to ensue upon the realization that the mind is empty (śūnya), in the sense that it is devoid of imagined and totally unreal external objects (i.e. matter). (xii)

According to Wood, the Vijñānavāda clearly endorses an underlying ultimate reality, rejects the mind-independent existence of physical objects, and gives ontological priority to the mind.

Pure Land and Zen traditions of Mahāyāna use and develop notions such as the tathāgatagarbha, the ‘womb’ or ‘embryo’ of the Buddha (Gethin 1998) or “buddha-nature” to refer to the intrinsic nature of all beings. This is the nature that is said to be realized upon becoming enlightened. On this view too, there appears to be some way that reality ultimately is. In this case it is Buddha-nature. Equating this buddha-nature with tathatā, Kalupahana (1976) goes so far as to describe this ultimate reality as a “transcendental monism” or a “noumenal reality,” emphasizing its inexpressibility and non-conceptuality (118, 122).

Since all these schools maintain that there is some way reality ultimately is, they would seem to qualify as realist views. Siderits (2003, 132), for example, categorizes these views as realist precisely for this reason. By this reckoning, a realist view is one that posits an ultimate nature for reality and an anti-realist view is one that denies that reality has such a nature. Only the Madhyamaka views that apply the doctrine of emptiness to absolutely everything would count as anti-realist. Only these views would reject the notion that there is some way reality is objectively, in and of itself.

While this is certainly a plausible way to mark the distinction between realism and anti-realism, it is not the way I would like to mark the distinction given the interests of this essay. Instead, I would like to follow those like David Chalmers (2009) who mark the distinction with respect to metaontological or metametaphysical considerations about the objectivity of ontological assertions. According to Chalmers, ontological realism and anti-realism are differentiated with respect to the distinct answers each gives to the metaontological question: does the basic question of ontology have objective answers? He explains, that “the basic question of ontology is ‘What exists?’ The basic question of metaontology is: are there objective answers to the basic questions of ontology? Here ontological realists say yes, and ontological anti-realists say no” (77).

The basic question of ontology, ‘what exists?’ has objective answers, according to the robust realist, because, according to her, the fundamental nature of reality determines an absolute domain that includes everything that fundamentally exists. Answers to the basic ontological question, when given in a specifically ontological or philosophical mode (as opposed to being part of ordinary discourse), are taken to quantify over this absolute domain, which determines the objective truth- value of the answer. Because the world is taken to come with this built-in, absolute domain, the basic question of ontology has objective answers.

There is a variety of reasons for an anti-realist to conclude that the basic question of ontology does not have objective answers. The reason why Mahāyāna Buddhists must so conclude is that they all must deny that the fundamental nature of reality determines an absolute domain that ontological existence assertions can quantify over. Even those Mahāyānists who maintain that reality does have a fundamental nature, still deny that ultimate reality has the kind of structure that could determine a domain of entities. According to Mahāyāna, ultimate reality is nonconceptual and ineffable. It lacks the object-property structure that would be required in

order for thought, which has a subject-predicate structure, to be objective. According to the Mahāyāna family of views, the conceptual structure of thought is an artificial superimposition on the non-conceptual and ineffable structure of fundamental reality. Another way to put this is to say that while the Mahāyāna may have a positive metaphysical position, this position does not amount to an ultimate ontology. Ultimate reality is not composed of distinguishable entities and their properties -- not even dharmas. This is the lesson to be learned from the doctrine of emptiness. And since there is no absolute domain, ontological existence assertions do not have objective truth-values. Thus, Mahāyāna is a form of ontological antirealism.

This characterization fits with Mark Siderits’ (2003) characterization of Buddhist anti-realism as a rejection of realism understood as the view that

(i) there is one true description of the world; (ii) with truth understood as the property that a statement possesses when the state of affairs that it expresses in fact obtains; and (iii) with the understanding as well that the states of affairs the obtaining of which makes up the world are independent of whatever relevant concepts we employ. (114)

In Chalmers’ terms, (i) would be an articulation of the metaontological position that asserts that the basic question of ontology has objective answers. The conjunction of all ontological existence assertions that are true would constitute the one true description of the world. Claim (ii) expresses the idea that the objectivity of those answers depends on the fundamental nature of reality, which determines the absolute domain that the ontological existence assertions quantify over. Claim (iii) makes clear that the fundamentality of

reality and the absoluteness of the domain preclude any concept dependence, contextualization or relative determination. The reality that is the subject of realist theories is the reality that is there anyway. Reality is taken to have the nature it has, and to exist as it exists independently of the way in which it happens to be cognized by beings like us. Siderits explains that for the realist,

To obtain truth we must let the world determine the nature of our representations. To be sure, our representations will make use of concepts, which we could agree are in some sense mental contents that do not exist apart from the minds that employ them. What we must insure, says the realist, is that it is the world that dictates the shape these concepts take... True statements reflect how the world is independent of the concepts we employ in describing it. (2003, 114)

Here Siderits emphasizes the importance for the realist that the direction of fit go from world to thought and not the other way around. The nature and structure of the world, as it is in itself, is taken to determine the structure of our thoughts when those thoughts are true, and it is in virtue of the fact that the world has a given nature that determines an absolute domain of entities that ontological existence assertions can be objectively true or false. The anti-realist, of course, rejects all this.

Chalmers notes that anti-realists have a difficult time accounting for the correctness conditions of ordinary existence assertions. For realists with revisionary metaphysics, it is possible to see how they might account for the correctness of ordinary existence assertions by making indirect appeal to the absolute domain (2009, 106). The Buddhist reductionist, for example, can take ordinary existence assertions to be correct as paraphrased sentences that indirectly make reference to the entities within the absolute domain. So, for example, the ordinary statement “There is a table” is correct if “there are dharmas arranged table-wise” is true (2009, 84).

However, Chalmers points out that

without an absolute domain, it is less obvious how an ontological anti-realist is in a position to analyze truth- or correctness-conditions for either ontological or ordinary existence assertions. In the case of ontological existence assertions, this might not be seen as too much of a cost, as the ontological anti-realist might deny that these assertions have truth-conditions or correctness-conditions at all. But there will remain the challenge of handling ordinary existence assertions, without the advantage of appealing to a built-in domain. (106)

This problem that Chalmers identifies is precisely the problem that I have identified the Mahāyāna as having. If we understand ordinary existence assertions as conventional, the problem all Mahāyānists have with respect to the notion of conventional truth is that of explaining what the truth or correctness conditions are for ordinary or conventional kinds of claims. This is a problem all Mahāyāna share, even if we grant that some schools retain the notion of ultimate reality. Given that all schools must deny the possibility of an absolute domain of entities, they all share this problem. In what follows I will examine some recent attempts to address this concern with respect to Madhyamaka Buddhism in particular.

Two Truths and Fictionalism

One way that the truth- or correctness-conditions of conventional claims (what I have, following Chalmers, been calling ordinary existence assertions) have been explained is in terms of fictionalism: roughly the view that assertions made within a certain discourse do not aim at literal truth but rather are circumscribed by the fictional domain to which they refer. Garfield (2006), for example, makes a case for a fictionalist reading of conventional truth in Madhyamaka and Tillemans (2011) reads the Mādhyamika Candrakīrti in particular as endorsing the same view. Their truth is not determined by objective facts but rather by the world of pretense facts that compose the fictional domain of discourse.

The paradigmatic cases of a fictional domain of discourse are, of course, works of fiction, like novels or plays. Garfield explains that Works of fiction are different from factual reports. The latter aim at getting it right about the actual world, and the former, for the most part, do not (though they nonetheless may rely on and comprise claims that are non-fictional). So, measured against reality, many of the claims in works of fiction are simply false, and nobody frets about that. (2006, 1)

So, the claim that Sherlock Homes lived at 221B Baker Street is true with respect to the fictional world that Sir Arthur Conan Doyle created, but is untrue with respect to the actual world where the person did not exist. As Garfield points out, no one finds this discrepancy between truth values problematic or mysterious. It is part of our understanding of what a fiction is that it is not aiming at presenting literal truth, so it is insulated from concerns about what the world is really like.

Advocates of fictionalism claim that, as in the case of literary fiction, it is possible to make unproblematic sense of claims being true within a domain of discourse while at the same time acknowledging that outside that domain, with respect to reality, they may actually be false. Likening a domain of discourse to a fiction allows those with revisionary metaphysics to allow for the continuation of discourse involving the entities that they deny really exist. By relegating discourse involving the suspect entities to a fiction, the revisionary metaphysician can continue to deny that those entities really exist while allowing that they do exist within the confines of the non-literal, ‘fictional’ domain of discourse to which they belong. And like in the case of fictional works, this should be unproblematic and unmysterious. If these discourses, like fiction, do not aim at literal truth, then they too can be insulated from real world ontological considerations.

Garfield (2006) maintains that fictionalism is a natural way to understand the Madhyamaka account of conventional truth (4). By adopting a fictionalist account of conventional truth, the Madhyamaka can allow for ordinary, conventional discourse to continue while at the same time maintaining their radically revolutionary (anti-)metaphysics that denies the ultimate existence of the entities ordinarily referred to in conventional discourse. Both Garfield (2006) and Tillemans (2011) find historical evidence for this kind of a view, especially in the works of Candrakīrti and Tsongkhapa.

According to this reading of Madhyamaka, the claims of conventional discourse are from the ultimate perspective wholly erroneous; yet from the conventional perspective, which is understood to amount to a fictional perspective, conventional claims can be assessed for conventional truth and falsity by the standards internal to that perspective.

As promising as this approach may seem, using fictionalism to understand the truth-conditions for ordinary, conventional discourse in Madhyamaka is problematic for several reasons. The first problem has to do with the intelligibility of the analogy with fiction when there is no contrasting non-fictional domain. As both Garfield and Tillemans point out, for the Madhyamaka, fictionalism would go “all the way down” (Garfield 2006, 6) and should be thus characterized as a “panfictionalism” (Tillemans 2006). But in the absence of a non-fictional reality, it is unclear in what sense one can say that ordinary discourse is fictional.

Usually the term fiction is used to indicate a non-literal, pretense context. To be a fiction is to be at variance with fact. To mark this distinction Garfield contrasts the conventional with the ultimate. He says that conventional truths when measured by the standpoint of ultimate reality are false (2). This seems in keeping with things that Candrakīrti in particular says to the effect that conventional truth is entirely false (2). However, if my characterization of the Madhyamaka position as a form of ontological anti-realism is correct, then there is no standpoint of ultimate reality to which ordinary discourse could be contrasted in order to deem it false, or merely pretense, or in some other way at variance with the facts. The point of Madhyamaka anti-realism is to deny that there are any objective facts to which conventional claims could be at variance. This seems to undermine the force of calling conventional discourse fictional. If everything is a fiction, then it seems that nothing is.

Garfield (2006) acknowledges that “for Madhyamaka, the very notion of there being an ultimate nature of reality is incoherent” and that for them there is no ultimate substratum behind the fiction (5). He cites this as the reason why the Mādhyamika Nagarjuna identifies the ultimate and conventional truths; the only truth to be had is whatever truth we can get from a conventional standpoint since, according to

Madhyamaka, there is no ultimate standpoint. Garfield quotes Tsongkhapa as saying that “without being posited through the force of convention, existence is not possible,” which makes the same point (5). There is no absolute domain of entities that existential claims can quantify over. The only entities that can be said to exist do so only with respect to some conventional

domain. Calling conventional domains fictions may be helpful for highlighting the fact that these domains are independent of ultimate ontological considerations, however, it is also misleading since it suggests that there is some ultimate frame of reference according to which the conventional can be judged to be false, which is precisely the position that Madhyamaka wants to avoid. I think this provides a good reason to avoid thinking of the Madhyamaka account of conventional truth in fictionalist terms. Another problem with the fictionalist account, which Tillemans (2011) highlights, is that this view threatens to amount to what Tillemans calls the “dismal slough,” the view that “one should accept ‘conventional truth,’ or truths for the world, as being only as the world accepts them” – a view that seems “to imply an extreme conservatism that nothing the world ever endorsed could be criticized or rejected and that,

on the conventional at least, a Mādhyamika’s principal epistemic task was just to passively acquiesce and duplicate” (152). Tillemans argues that Candrakīrti and his followers (the Prāsaṅgikas), as interpreted by Kamalaśīla, understood the dismal slough to follow from a fictionalist understanding of the conventional. Finnigan and Tanaka (2011) concur that for the

Prāsaṅgika, emptiness of all things is often thought to entail that a conventional truth is an unreflective endorsement of the world's beliefs. A consequence of this Prāsaṅgika view, as was pointed out by Kamalaśila and others, is that truth in conventional truth has no normative role and that, as a result, conventional epistemic practices lack any authority. This means that the truth of conventional truths is unexplained, and so no sophisticated analysis of anything can be given. (186)

The idea seems to be that in adopting the fictional stance one must play make-believe according to the rules of the adopted conventional fiction and one cannot substantially challenge the fiction on the grounds that what it claims is ultimately untrue. Since the fictions are insulated from external ontological considerations, there is no way to critique the fiction from the

outside. The best attempts one can make to reform the story must come from within, invoking concerns of coherence and consistency among the claims made in the fiction. However, Tillemans notes that still, “a considerable portion of the story needs to be taken as brutely given” and furthermore, with respect to the tradition historically, “the Prāsaṅgika seems to have perceived the desirability for reform to have been quite limited indeed and seems to have felt that because no account could ever be true, the world’s story should be accepted largely intact by default” (159).

In calling this kind of position the “dismal slough” Tillemans acknowledges that “[m]ost of us would agree that the potential flattening of the normative roles of truth and knowledge that such duplication [of what everyone else already accepts] brings is quite dismal” (152). Newland and Tillemans (2011) further claim that, the obvious danger is that if saṃvrtisatya [[[conventional truth]]] is not true at all but only mistakenly thought to be true, then it carries no normative force,

collapsing the important distinction between what is true and what is merely thought to be true. There would be little of interest to say about such “truth” beyond simply describing what people believe. (14)

These passages highlight some important concerns about the normativity of truth. If the norms of correctness are determined by the fiction and there is no perspective from outside the fiction from which those norms can be challenged, then the concern is that there is no way to challenge or test those norms.

Related to these concerns about truth as a normative concept, there is a further concern that the Cowherds (2011) do not address: that such an uncritical fictionalism denies the Buddhist the resources necessary to defend her normative position. In adopting the fictional stance, the typical Prāsaṅgika can identify and describe the norms of the conventional fiction she finds herself in, presumably established by her culture and its customs, but denies them any normative force independent of the conventions that institute them (Garfield

126-127). What this means is that outside cultural and social contexts that are already Buddhist, the Prāsaṅgika has no justification for asserting and conforming to Buddhist norms. In a context in which there are conventionally accepted claims that conflict with Buddhist claims and present alternative normative orientations, in adopting the Buddhist doctrine the Prāsaṅgika is put in the position of having to defend her choice. The fictionalist position, as presented by Tillemans, seems to leave her with no way to justify her choice at the conventional level if the conventions of her social and cultural context do not already accept Buddhist claims as true.

This is especially problematic when one considers the cultural context in which the Buddha taught or in which Nāgārjuna, the founder of Madhyamaka, wrote. In neither of these contexts were the Buddha’s teachings obvious or non-contentiously accepted by the society at large. The claims of Buddhism did not reflect the truths of the culture -- either its common-sense, custom, or law. In fact, the Buddha was in many respects a revolutionary. For example, he challenged traditional views about duty by leaving at a young age the householder life as father and husband to become a wandering ascetic; he challenged traditional

views about piety by rejecting the Vedas and the role of ritual and sacrifice in religious life; he challenged the social hierarchy of the caste system and traditional gender norms by accepting all people – men and women – into the Buddhist monastic community. And with regard to the Mādhyamika Nāgārjuna, the wider cultural context in which he wrote was not much

different than that of the Buddha, except that there were more fully developed orthodox philosophical positions to compete with in the religious marketplace. These deviations from the conventional norm cannot be explained or justified, it seems, from the fictionalist perspective. Nor can Buddhist truth claims be defended in a cultural or social context in which the truths of Buddhism are not in some sense already presupposed as part of that culture’s fiction.

Garfield (2006), Garfield (2011), and Thakchöe (2011) make clear that not all Prāsaṅgikas drew the conclusion that one ought simply to accept what the ordinary person accepts as true from the position that the only truth is conventional truth. Tsongkhapa, in particular, seems to have advocated the acceptance at the conventional level of epistemic instruments (pramāṇa) as authoritative. This, in effect, gives Madhyamaka an epistemology that provides a standard of correctness through which a distinction can be made between conventional truth and conventional falsity. This account is non-foundationalist because the method of arriving at truth, the pramāṇa, are not considered to be self-justifying or

foundational. Like everything else, the pramāṇas too are empty of inherent existence. They are not by nature instruments of knowledge, they are simply taken to be so conventionally. Garfield (2011) explains that on this view (which he attributes to Nāgārjuna, Candrakīrti, and Tsongkhapa) epistemic instruments depend for their authority on their epistemic objects, and the objects, in turn, depend for their actuality on the instruments in a coherentist spiral that defies grounding but characterizes epistemic practice in the only way we could ever hope to do so. (36)

The Mādhyamikas who endorse this kind of a view are keen to demonstrate that pramāṇas are as dependently originated as the objects they reveal. The

“coherentist spiral” that Garfield refers to here is intended to highlight the fact that the means of knowledge and the objects known co-determine and support one another in a way that aims at coherency and that lacks any ultimate, foundational support. This is intended to highlight the fact that the pramāṇas operate within the fictional realm of discourse, are determined by that realm of discourse, and act with respect to the fictional entities within that discourse. They are not to be understood as providing a perspective external to the fiction that can somehow validate it from outside.

This acceptance of pramāṇas at the conventional level allows someone like Tsongkhapa to claim that “there is a standard of correctness for conventional truth...And the standard for the truth of a judgment regarding conventional truth is that it is vouchsafed by the authority of conventional epistemic instruments and cannot be undermined by those instruments” (Garfield 2011, 34). Thus for him, the conventional truth is "that which is delivered by unimpaired

cognitive faculties when they are used properly" (35). For those Mādhyamikas who accept pramāṇas at the conventional level, the lesson to be learned from the realization that ultimately nothing has intrinsic nature is that there is no such thing as ultimate truth, meaning that there are no claims that are true in virtue of some ultimately existing things or state-of-affairs. What this means is that the only truth that is left is the conventional truth, which then can be understood through accounts of pramāṇa theory.

This account of the role of pramāṇas in conventional discourse appears to be addressing another concern that arises for Madhyamaka fictionalism, which Siderits (2006) raises in his comments to Garfield (2006). The term ‘fiction,’ he notes, has the connotation of something that is arbitrarily invented. Calling conventional reality and its related discourse fictional suggests that it too is arbitrarily determined (17). The fictional world of Sherlock Homes, for example, was arbitrarily created by Sir Arthur Conan Doyle and it serves as the standard according to which all claims made about that world are judged. So the question that Siderits asks is, how according to this analogy are we to understand where the conventional world fiction originates?

And is this origin arbitrary?

This concern with the possibly arbitrary origin of conventional reality on a fictionalist account seems to get at the heart of both Tillemans’ concern about the flattening of norms and the related concern about the ability to defend revolutionary ways of thinking. In the absence of an absolute standard, if conventional reality is arbitrarily determined, and that arbitrarily determined reality serves as the standard according to which claims made with respect to that reality are judged, is there any room for critical evaluation or must everything simply be accepted at face value?

The concern is that, as with literary fiction, on a fictionalist account the answer is no. With respect to pramāṇa theory one problem that arises is that there is not consensus among Indian philosophers about which pramāṇas are authoritative. There is not even consensus among the Mādhyamikas themselves about which pramāṇas are authoritative. Candrakīrti, for example, seems to have accepted the Nyāya account that allows four epistemic instruments: perception, inference, testimony, and analogy (Siderits 2011, 172). Other

Mādhyamikas, Tsongkhapa for instance, seem to follow Dignaga and Dharmakīrti in allowing only two: perception and inference. Which account is the conventionally correct one? Without an answer to Siderits’ question concerning the origin of the conventional fiction, it is unclear how such a dispute could be resolved non-arbitrarily or non-dogmatically.

Carnapian Approach to Conventional Truth

The advantage of a fictionalist account is that it presents a way of preserving truth in a context in which absolute ontological considerations are irrelevant. As Garfield (2006) points out, “fictions can constitute a world against which truth can be assessed” (1). These fictional worlds are “neither reducible to nor ontologically supervenient upon the actual world” and so they are entirely independent of any claims about how the world actually is (1). These virtues of fictionalism might, however, be better captured by a Carnapian distinction

between claims made internal and those made external to a framework. Thinking of the truth of claims as being dependent on something like a framework can insulate claims from absolute ontological considerations while still taking those claims to be earnest attempts to say what is literally true (as opposed to taking them to be a kind of pretense). Since there is no framework-independent context of assessment, literal truth on this kind of a view is framework-dependent truth. There is no need for a literal versus pretense distinction.

According to Carnap (1950), existence assertions are only meaningful with respect to the linguistic framework that establishes the structure, system, and rules for the forms of expression involving the type of entity that is the subject of the assertion. Existence assertions are, in light of this, answers to what Carnap calls

“internal questions,” which are questions that are internal to a framework. What Carnap calls “external questions” about the existence of entities are questions that purport to ask about the ontological status of their subjects independent of any framework. Such questions are meaningless or “non-cognitive”, according to Carnap, because the positing of the reality of the entity is inextricably tied to the framework of which it is a part. Thus, for Carnap, an external question about the reality of an entity really amounts to a question about the acceptability of the framework that contains it.

Accepting a framework, according to Carnap, is not a matter of checking it against reality since the notion of reality is for him a notion that is frameworkdependent. So for Carnap, the acceptability of a framework is not a factual matter but a pragmatic one. The question for him is whether or not the way of speaking that is systematized and structured by a framework, which involves taking as objects certain kinds of entities, is a useful way of speaking. He says,

The purposes for which the language is intended to be used, for instance, the purposes of communicating factual knowledge, will determine which factors are relevant for the decision [whether or not to accept the language]. The efficiency, fruitfulness, and simplicity of the use of the thing language [for example] may be among the decisive factors. (23)

Focusing here on the linguistic framework of what he calls “the world of things” or the “thing-language,” Carnap emphasizes that the purpose of constructing a system of language is to communicate with others. The acceptability of a linguistic framework, which is for him a matter of degree, is closely tied to the purposes to which we put the language. The thing-language -- “the spatio-temporally ordered system of observable things and events” -- for example, has proven to be very useful for the purposes of everyday life and is widely accepted on these grounds.

According to Carnap, then, when accepting a particular form of expression involving a particular kind of entity, one implicitly accepts the associated framework to which the entity is connected (31). However, he says that “the acceptance of a framework must not be regarded as implying a metaphysical doctrine concerning the reality of the entities in question” (32). Metaphysical doctrines positing the reality of entities independent of frameworks are confused, according to Carnap, and cannot be made meaningful.

The suggestion to use Carnap’s distinction between internal and external questions to understand the distinction between ultimate and conventional truth is offered by Finnigan and Tanaka (2011). In emphasizing the pragmatic aspect of Carnap’s account they explain how such an account can help the Mādhyamikas avoid Tillemans’ “dismal slough.”

Finnigan and Tanaka explain that the conclusion that some Prāsaṅgika Mādhyamikas draw, that common norms and beliefs should be unreflectively adopted, relies on the assumption that the norms of truth rely on the existence of ultimate truth-makers, which ontological anti-realism denies (186). The argument seems to go something like this:

(1) The doctrine of emptiness requires the denial of ultimate truth-makers. (2) Ultimate truth-makers are the only things that can ground the norms of truth and serve as the standard of assessment for what we say. (3) Therefore, there is no standard that can be appealed to to assess common views, (4) so we should accept those views uncritically.

According to Madhyamaka, (1) is correct; the doctrine of emptiness entails the rejection of the notion that there is some absolute domain of entities -- ultimate truth-makers -- that existence assertions can quantify over. However, the conclusion of the argument above can be avoided if, like Carnap, the Mādhyamika rejects (2), the presupposition that assessment of truth depends on ultimate truth-makers. Instead, the standards of truth can be understood to be already implicit in our everyday linguistic practices. On this view we should think of Buddhist conventional reality as being constituted by the frameworks implicit in our ordinary ways of talking. These frameworks will determine a non-absolute domain of entities that our ordinary existence assertions can quantify over. No ultimate truth-makers are necessary in order to have a standard of assessment.

Insofar as the linguistic practices and the frameworks that structure them are tied to our practical purposes, a Carnapian account also addresses the concern about arbitrariness that was raised with respect to fictionalism. According to Carnap the particular linguistic frameworks we adopt can be assessed with respect to whether or not they serve our pragmatic interests, the foremost of which is effective communication. Effective communication requires intersubjective norms of correctness, which the structure and rules of the

frameworks provide. Truth is then assessed with reference to the framework and its structure. Whether or not the framework as whole is acceptable will depend on its success in allowing us to communicate with others what we need to communicate in order to support our ways of life. This concern for the pragmatic utility of adopting a particular way of talking is also an important concern for the Buddhists, who often cite pragmatic utility as a rationale for speaking in conventional ways.

Employing the Carnapian distinction between internal and external modes of inquiry, might also help us understand the debate mentioned earlier between Mādhyamikas regarding which pramāṇas to adopt. The question of which epistemological framework one should adopt, for example between Nyāya and Dharmakīrtian accounts of praṃāna, could be considered an external question that is to be answered by taking into account pragmatic

considerations, for example, Carnap’s “efficacy, fruitfulness, and simplicity” (23). Thus, internally, whether or not a claim is true would depend on whether or not it is established by conventionally held pramāṇas. Which pramāṇas are accepted as part of the framework that decides these questions is an external matter that takes into consideration which pramāṇa system works better, given our interests. A critical account of pramāṇas on this view need not appeal to some absolute and foundational standard, but rather a pragmatic one.

The Carnapian approach to conventional truth is a more promising approach than the fictionalist approach but it is still not without problems. The main concern is about how the frameworks are determined and how they work to fix the domain of entities that are the referents of existence assertions. While the appeal to pragmatic utility helps to temper the charge that the frameworks are arbitrary, there is still a question about how exactly the

frameworks are formulated. In the case of introducing a new entity for some technical purpose, describing the method as one of stipulating the system of language that will refer to these entities and fix their domain seems appropriate. But the frameworks that lay out the system of language for the entities of our ordinary discourse, like the thing-language for example, surely were not generated in this way by stipulation. So how did they arise? In addition to questions surrounding the origin and function of frameworks, there is also the possibly related question of the revisability of frameworks. Are frameworks revised as our use of terms in the language changes and develops? Or must frameworks be adopted and rejected altogether?

The concern here is that if the use of a term entails an implicit acceptance of the linguistic framework in which the term is embedded and that framework is inflexible and static, then it would seem that there would be little room for growth and development of the use the term without abandoning the framework altogether. This all or nothing adoption of static frameworks does not seem to reflect the way in which language is commonly thought to develop. If frameworks allow for development and revision, the question then is about how this occurs in a manner that preserves the

normativity the structure of the framework is supposed to provide. As a general strategy the Carnapian approach is promising, but as these questions suggests, the devil is in the details. I will revisit this strategy in chapter five where I will attempt to incorporate it in giving an account of Dharmakīrti’s theory of meaning. There I will use resources from Dharmakīrti’s philosophy to flesh out the details of a Carnapian account in a distinctively Buddhist way. Conventional Truth as Deflationary Truth

Another promising and related attempt at giving an account of conventional truth in Madhyamaka is suggested by Priest, Siderits, and Tillemans (2011). They suggest that for the Madhyamaka school of Mahāyāna Buddhism, conventional truth might best be understood as deflationary truth. The deflationary theory of truth recommends itself because it retains a strong norm- supporting, non-relativistic distinction between what is true and what is false at the conventional level but without involving any

ultimate metaphysical commitments – two goals that are important for Mādhyamikas. Thus Priest, Siderits, and Tillemans propose the following as a possible understanding of Madhyamaka conventional truth: Retain correspondence as our understanding of the “truth” in “conventional truth” but go deflationary about correspondence. In that case, the absence of robust truthmakers to stand behind our acceptance of conventionally true statements need not be an embarrassment. For then when we are asked what makes it

true that there is a pot on the ground, we can simply reply that there is a pot on the ground… [A] deflationist theory, like that of Horwich, does not involve anything metaphysically charged. It might then seem that the deflationist’s version of truth, purely along the lines of

is true iff p and stripped of the excess baggage of truthmakers and ontology, would give an elegant reconstruction of Madhyamaka’s own oft-repeated principles. (143) The family of views commonly referred to as deflationary theories of truth, which include disquotational, redundancy, or minimalist theories, all share the view that truth lacks a substantive nature. The most carefully worked out version of a deflationary theory of truth is presented and defended by Paul Horwich. As Priest, Siderits, and Tillemans explain, this theory maintains that there is nothing more to the concept of truth than the equivalence schema: (E)

is true iff p where ‘

’ is short for ‘the proposition that p’, ‘p’ presents the conditions under which the proposition is true -- which are specified by the proposition itself -- and ‘iff’ is the logical operator ‘if and only if.’ The classic example of one of these equivalences is <snow is white> is true iff snow is white which is just one of the infinite number of equivalences that together constitute the axioms of the deflationary theory of truth. Deflationists, such as Horwich, argue that the equivalence schema is sufficient on its own to explain everything there is to know about truth. The deflationary theory of truth, Horwich maintains, is basic, maximally deflationary, and yet complete, leaving nothing necessary out of the account of truth. According to the deflationist, the property of truth is neither complex nor naturalistic and thus has no underlying nature, has no hidden structure that needs to be elucidated, and is not susceptible to conceptual or scientific analysis. There are no necessary and sufficient conditions generally for some proposition’s being true beyond those specified by each proposition individually. All the facts about truth are given in the equivalence schema or can be derived from the axioms of the theory (Horwich 1998, 12). Horwich explains that “the truth predicate exists solely for the sake of a certain logical need,” namely the ability to infer from

is true, p, and vice versa and “to enable the explicit formulation of schematic generalizations” (37). As Quine pointed out, certain generalizations, such as ‘Every sentence of the form “p or not p” is true,’ require the use of the truth predicate in order to indirectly assert every sentence of that form (Quine 1970, 11). This amounts to a “semantic ascent” whereby we talk about sentences instead of using them. Quine explains, This ascent to a linguistic plane of reference is only a momentary retreat from the world, for the utility of the truth predicate is precisely the cancellation of linguistic reference. The truth predicate is a reminder that, despite a technical ascent to talk of sentences, our eye is on the world. This cancellatory force of the truth predicate is explicitly in Tarski’s paradigm: ‘Snow is white’ is true if and only if snow is white. Quotation marks make all the difference between talking about words and talking about snow. The quotation is a name of a sentence that contains a name, namely ‘snow’, of snow. By calling the sentence true, we call snow white. The truth predicate is a device for disquotation. We may affirm the single sentence by just uttering it, unaided by quotation or by the truth predicate; but if we want to affirm some infinite lot of sentences that we can demarcate only by talking about the sentences, then the truth predicate has its use. (13) Here Quine highlights the disquotational aspect emphasized by some deflationary theories of truth but also points out an important instance in which we cannot do without the truth predicate; the one wherein we cannot simply use language but must make reference to it. While we need the truth predicate in these instances, it does not add anything conceptually to a sentence that the sentence itself does not already contain. Thus, Horwich maintains that the entire conceptual and theoretical role of truth may be explained on the basis of the equivalence schema (5). Everyone is prepared to infer that the proposition that snow is white is true from snow being white and vice versa. This use is captured by the equivalence schema, and beyond this there really is nothing more to say about the nature of truth. Priest, Siderits, and Tillemans are right to note that the deflationist theory “does not involve anything metaphysically charged” and that the theory is “stripped of the excess baggage of truth makers and ontology.” Horwich claims, for example, that “truth is metaphysically trivial – nothing more than a device of generalization” (146). This is indeed an attractive feature for a Mādhyamika. However, just because the theory of truth itself is metaphysically neutral does not mean that ontology is not still lurking in the background. In fact, the theory requires an independent, though supplementary, theory to explain the meaningfulness of propositions and their assertability conditions. Tarski (1944), for example, points out that, In fact, the semantic definition of truth implies nothing regarding the conditions under which a sentence like (1): (1) snow is white can be asserted. It implies only that, whenever we assert or reject this sentence, we must be ready to assert or reject the correlated sentence (2): (2) the sentence ‘snow is white’ is true. (140) Wittgenstein (1999) makes a similar point by saying, For what does a proposition’s ‘being true’ mean? ‘p’ is true = p. (That is the answer.) So we want to ask something like: under what circumstances do we assert a proposition? (112) and Really ‘The proposition is either true or false’ only means that it must be possible to decide for or against it. But this does not say what the ground for such a decision is like. (114) To clarify these points, it might be helpful to look at a possible inference: (1) <A pot is on the ground> is true iff a pot is on the ground. (By E) (2) A pot is on the ground. (?) (3) Therefore, <a pot is on the ground> is true. (By 1 & 2) Tarski and Wittgenstein point out that the equivalence schema, which is instantiated in the first premise, allows us to infer (3) from (2). However, they also point out that the schema itself says nothing about the assertability of the second premise, about whether or not the condition given in (1) is met. It is in virtue of this fact that the deflationary theory of truth can be neutral with respect to questions of ontology. The question remains, however, as to how the assertability of the second premise is to be established. What is the ground for a decision for or against it? An account of when the conditions given on the right-hand side of the equivalence schema, which amount to the propositions itself, are fulfilled is not given by the deflationary theory. The question concerning the assertability of the second premise of the above argument is important for the supposed normative role of the concept of truth because ordinarily we would be inclined to say that we ought to assert (2) or include it in our inference if it is true. That is to say, in addition to the norms that govern the use of the term ‘true’ in drawing inferences (e.g., inferring (3) from (2)), we are usually inclined to use the concept of truth normatively to give an account of when it would be appropriate to assert a sentence or include it as a premise in an argument. We ordinarily take this to be part of the normative aspect of truth. It is on these grounds that Crispin Wright (1999) criticizes deflationism saying, No deflationist has wanted, or ought to have wanted, to deny that believing and statement-making are normatively constrained activities – activities governed by standards, non-compliance with which opens a thinker to criticism. However, once that is accepted, the question has to be confronted of what the relevant standards are… What, for deflationism, are these norms? In order to answer this question, the deflationary theory must locate this normative role somewhere else, in a theory that is distinct from the theory of truth that it offers. In fact, this is exactly what Horwich does. He argues that this normativity is misplaced, attributed to truth when really it concerns something else. Norms of assertion and of belief are not, properly speaking, norms that concern truth. They are related to truth but part of a different theory. According to Horwich, an account of the assertability of (2) would have to be given by a separate theory. While Horwich takes his theory of truth to be complete, he does not maintain that the deflationary account alone will engender all the facts about truth. He states that in determining a deflationary or minimal theory of anything, Our goal… is to find a simple theory of X, which together with our theories of other matters, will engender all the facts… The virtue of minimalism, I claim, is that it provides a theory of truth that is a theory of nothing else, but which is sufficient, in combination with theories of other phenomena, to explain all the facts about truth. (24-25) Horwich thus assumes that the deflationary theory of truth will have to work in conjunction with theories about other matters in order to “explain all the facts about truth.” One of the things the deflationary theory of truth needs as a supplement is an account of meaning that explains the proposition on the right-hand side of the equivalence schema and thereby explains when the proposition is assertable. The deflationary theory is itself neutral with respect to which supplementary theory is to be advanced and in particular it is neutral with respect to the ontology such an account presupposes. While the deflationary theory of truth is neutral with respect to ontology, Horwich argues that it is nevertheless consistent with correspondence intuitions. In so arguing, Horwich maintains that the conditions on the right-hand side of the equivalence schema can be established through some kind of appeal to the nature of reality or the way the world is. He says that the deflationary theory of truth does not deny that truths do correspond – in some sense — to the facts; it acknowledges that statements owe their truth to the nature of reality; and it does not dispute the existence of relationships between truth, reference, and predicate satisfaction… It is indeed undeniable that whenever a proposition or an utterance is true, it is true because something in the world is a certain way – something typically external to the proposition or utterance. (104) In order to explain these statements he continues, For example, (1) <Snow is white>’s being true is explained by snow’s being white. That is to say, (2) <Snow is white> is true because snow is white. But these [correspondence] intuitions are perfectly consistent with minimalism [deflationism]. In mapping out the relations of explanatory dependence between phenomena, we naturally and properly grant ultimate explanatory priority to such things as the basic laws of nature and the initial conditions of the universe. From these facts we attempt to deduce, and thereby explain, why, for example, (3) Snow is white. And only then, invoking the minimal theory, do we deduce, and thereby explain, why (4) <Snow is white> is true. Therefore, from the minimalist point of view, (3) is indeed explanatorily prior to (4), and so (1) and (2) are fine. Thus we can be perfectly comfortable with the idea that truths are made true by elements of reality. Since this follows from the minimal theory (given certain further facts), it need not be an explicitly stated part of it. (104-105) In this passage Horwich makes reference to ‘facts,’ the ‘elements of reality,’ and the ‘nature of reality’ in order to account for (3), which then explains (4). This demonstrates that while the deflationary theory of truth does not itself explain truth in terms of correspondence, it does not necessarily dispense with the ontology that is usually associated with the correspondence theory of truth (which the Mādhyamikas are thought to be trying to avoid); it is perfectly consistent with such a view. Of course it does not require such a view either, which is why it is available as a possible interpretation of truth for Madhyamaka. Given the ontological neutrality of the deflationary theory of truth we do not need any “robust truthmakers to stand behind our acceptance of conventionally true statements,” however, we do need some account of the proposition given on the righthand side of the equivalence schema. We need an account of propositions as the content of our beliefs and the meaning of our statements. The deflationary theory of truth presupposes that it is possible to give such an account without appealing to the notion of truth. Therefore, if the Mahāyāna is to be interpreted as employing a deflationary theory of truth, then it must also be able to give an account of the content of our beliefs and the meaning of our statements that is independent of the notion of truth. This will be challenging for the Mahāyāna given that her anti-realist commitments make it difficult to account for aspects of meaning such as reference and predicate satisfaction. A further consequence of adopting a deflationary interpretation of truth in Madhyamaka is that any distinction made between conventional and ultimate truths cannot be explained as a distinction between two different kinds of truth. According to the deflationary theory, all propositions conform to the equivalence schema and all propositions are thus equally true when true and false when false. There is no room for degrees; there is only one truth according to the deflationary theory of truth. Given the character of Mahāyāna anti-realism, this is exactly the result we want since claims made with respect to an absolute domain of entities cannot have determinate truth values. This means that the only truth is conventional truth, which on this account would be deflationary. The deflationary account is the most promising account of conventional truth for Mahāyāna for the reasons Priest, Siderits, and Tillemans give. In chapter five I will argue that Dharmakīrti can provide the necessary supplementary account of meaning. This account is compatible with both the limitations of deflationism and the ontological commitments of the Mahāyāna. With this account in hand I conclude that the suggestion made by Priest, Siderits, and Tillemans to use the deflationary theory of truth to understand conventional truth in Madhyamaka is a good one for all Mahāyāna, when it is supplemented by Dharmakīrti’s account of meaning. CHAPTER 4 DHARMAKĪRTI AS A MAHĀYĀNIST In the last chapter I discussed some recent promising attempts to articulate the sense in which conventional truth is true for Madhyamaka Buddhism. In particular I focused on the suggestion made by Priest, Siderits, and Tillemans to understand conventional truth as deflationary truth. While this suggestion seems the most promising way to understand the way truth operates at the conventional level, it alone cannot generate, as Horwich puts it, “all the facts about truth.” Generating all the facts about truth requires in addition a supplementary theory of meaning. In the next chapter I argue that a Buddhist theory of meaning that is consistent with both Mahāyāna ontological anti-realism and with the deflationary theory of truth can be developed from the work of Dharmakīrti. However, before I proceed to present that account, I first want to make clear why Dharmakīrti himself should be understood to be a Mahāyāna anti-realist. The reason this is important is that the parts of his work that I hope to develop in answering the question of meaning are usually interpreted as presenting a Sautrāntika Abhidharma position. As a version of Abhidharma, this account would be a version of ontological realism. If Dharmakīrti is understood to be a realist of this kind, or to be presupposing such an account in giving his account of meaning, then his account could not help to explain the sense in which conventional truth is true for Mahāyāna since it could not address the issues that arise with respect to the anti-realism of that position. So before I present Dharmakīrti’s account of meaning, I first want to make clear the sense in which his account is anti-realist. Dharmakīrti as a Yogācārin It is common in the commentarial tradition, as well as in the work of modern Dharmakīrti scholars, to interpret Dharmakīrti’s ultimate philosophical position to be a Yogācāra one. Vācaspatimiśra, for example, in commenting on Dharmottara’s commentaries to Dharmakīrti’s texts, interprets Dharmakīrti as adopting both Sautrāntika and Yogācāra points of view at different places in his texts but taking the Yogācāra point of view to be the one ultimately defended. There he argues on behalf of the Yogācāra that external objects can be neither known through sensation nor through inference, and since these are the only two means of knowledge available, according to Dharmakīrti, the external object, even if it were to exist, could not be known. Similarly, Devendrabuddhi in his commentary on certain key verses in Dharmakīrti’s Pramāṇavārttika, and Śākyabuddhi in his commentary to Devendrabuddhi’s commentary, compare Dharmakīrti’s reasoning in these verses with those “presented by powerful thinkers such as Master Vasubandhu” and take him to be likewise concluding that “external objects do not exist,” a common Yogācāra position. Some of the key verses that Vācaspatimiśra, Dharmottara, Devendrabuddhi, and Śākyabuddhi comment on that lead them to interpret Dharmakīrti as a Yogācārin are verses 208-219 of the Pratyakṣa chapter of his Pramāṇavārttika. Leading up to these verses, in verse 194 Dharmakīrti begins a long discussion whereby he argues against the contention that objects can exist as wholes over and above their parts. The discussion begins with an objector challenging Dharmakīrti’s claim that perception is non-conceptual by appealing to the Abhidharma view that sensory cognition is the result of an aggregate (sañcita) of atoms. According to that view, only a group of atoms in close proximity to one another can give rise to sensory cognition. Isolated atoms alone are insufficient. Since an aggregate is a kind of sāṃānya, or universal, and is thereby a kind of conceptual entity, it follows that perception is conceptual, contrary to what Dharmakīrti claims. Dharmakīrti responds by clarifying the Abhidharma position. Even though Abhidharma might refer to the collection of atoms as an aggregate, Abhidharma does not consider the collection itself to be a distinct entity over and above the atoms that comprise it. The basis of the perceptual cognition is still nothing but the atoms, according to that view. While that collection of atoms might be called an aggregate, in the sense of just a plurality of atoms, this does not commit Abhidharma to consider the aggregation as a distinct entity. And since the basis of perception is not strictly speaking a conceptuality entailing aggregate, but rather discrete atoms, the objector’s argument fails to undermine Dharmakīrti’s claim that perception is non-conceptual. At verse 197 the objector persists by pointing out that the Abhidharma account is at odds with the commonly held view that perceptual cognitions only cognize one object at a time. According to that view, for example, the perception of a patch of yellow and a perception of a patch of blue do not happen at the same time but rather occur as distinct cognitions. It is only because cognitions happen rapidly in succession that it appears to us as if we see more than one object at the same time. If it is true that cognition only takes one object at a time, then it would not be possible for a perceptual cognition to cognize a bunch of atoms all at once. In reply, Dharmakīrti argues against the premise that cognition can only take one object at a time. That assumption, he argues, would make it impossible to explain the perceived difference between things occurring simultaneously and things occurring in succession. If all cognitions arise and disappear at the same rate, and it is in virtue of their rapid succession that we see things as occurring together, then why is it not the case that everything appears as if occurring simultaneously? The objector’s proposed account of perception would not be able to explain how it is possible that we experience some objects as occurring in succession and others as occurring simultaneously. Furthermore, at verse 200, Dharmakīrti argues that such a view cannot make sense of the experience of seeing an object that is multicolored. What could be the difference, on this view, between seeing a loosely assembled group of things of various colors, like a collection of jewels, and seeing one thing, like a butterfly, that has a plurality of colors? Dharmakīrti’s opponent wants to use the view that only one object is seen at a time to argue that whole, discrete, medium-sized, recognizable objects are perceptible -- a claim that Dharmakīrti denies. However, Dharmakīrti here points out that the account of perception that the opponent offers could not account for the difference between seeing one whole thing that has various colors and seeing a bunch of different things that each has its own color. And the attempt that some realists make to resolve the problem by referring to a distinct color they call “variegated color” (citra), Dharmakīrti rightly ridicules as ad hoc and ridiculous. Dharmakīrti points out that as in the case of a bunch of discrete objects, like a mass of jewels, in the case of a single object, like a butterfly, each of the separate colors is identifiable as distinct. Yet the appeal to one color identified as “variegated” as characterizing whole objects, like the butterfly, would preclude the ability to identify the various colors individually (under the assumption that distinct colors preclude one another), which goes contrary to experience. Calling the collection of colors by one name does not change the fact that they are distinct. As a result of his analysis, Dharmakīrti claims at verse 204 that “A manifoldness of entities (bhāvas) is due to the manifoldness of cognitions,” by which he means that whatever plurality of entities we posit must come from the plurality of objects we experience in cognition. There is no other way, he claims, for us to determine what kinds of entities there are. And since perceptual experience presents us with the same visual appearance, i.e. of distinct colors at certain locations, whether there is a loose collection of objects or there is what we take to be a single object, strictly speaking there are no wholes that are perceived. Rather what is perceived is a plurality of entities, in this case colors, that we only take to be whole objects in some cases and not in others. In verse 207 Dharmakīrti concludes “from that it follows that a single cognition may have various objects and that such a cognition is non-conceptual.” In reply to Dharmakīrti’s conclusion, in verse 208 the hypothetical opponent raises the following objection: “If oneness is not possible with respect to objects that appear variegated [such as a butterfly or a multicolored cloth], then how is the cognition that appears variegated one?” Here the opponent complains that, if Dharmakīrti’s reasoning is enough to establish the unreality of wholes, then it should also be enough to establish the unreality of the cognition over and above the plurality of objects that it takes as representations. The opponent is presumably hoping that, in accepting the possibility that cognition is a whole over and above its parts, the Buddhist will be forced to accept that this is possible also with respect to object-wholes. However Dharmakīrti surprisingly concedes to this objection by saying in verse 209 that, “this which the wise say is obtained by the force of reality itself, namely that the more things are reflected on, the more they fall apart.” Here begins what is taken to be the presentation of Dharmakīrti’s Yogācāra position. In the verses that follow this concession, Dharmakīrti explains that while it ordinarily appears as though cognition has two distinct aspects, one which is subjective and internal and another that is objective and external, thus giving the impression of there being a single, unified, subject that is distinct from the various external objects that it represents, in reality this duality is an illusion. He states in verses 212-215: Ascertainment appears as if located inside, this other portion i.e., the object form appears as if situated outside. For there are two distinct forms of a cognition which is really without difference, so the appearance is an error (212). In regard to that, by means of the non-existence of even one or the two, either object-form or subject-form, both of them are abandoned. Therefore, the very reality of that cognition, too, is the emptiness of duality (213). And this establishing of a difference of entities is based on their difference viz., of object and subject. And when that has the nature of error, the difference of those entities also is an error (214). And there is no characteristic other than object-or subject-form. Therefore, because they are empty of characteristics (lakṣana-śūnyatva), entities are shown to be without essence (niḥsvabhāva) (215). In his commentary on these verses, Śākyabuddhi explains Dharmakīrti’s position in the following way: Cognitive appearances such as blue seem to be external to awareness, but when one analyzes whether those appearances are singular or plural, they are unable to withstand that analysis; hence, they are not suchness. Therefore, there is ultimately no object that is distinct from awareness itself, and since that object does not exist, we say “the subject does not exist”; in saying this we mean the “subject” that occurs in expressions or concepts that are constructed in dependence on the [apparently external object], as in “this is the real entity that is the subject which apprehends that object, which is the real entity that it cognizes.” Since an agent and its patient are constructed in dependence upon each other, these two [i.e., subject and object] are posited in dependence on each other. The expression “subject” does not express mere reflexive awareness, which is the essential nature of cognition itself. The essential nature of cognition is not constructed in mutual dependence on something else because it arises as such from its own causes. The essential nature of cognition is established in mere reflexive awareness. Since it is devoid of the above described object and subject, it is said to be non-dual. In these comments Śākyabuddhi interprets Dharmakīrti to be endorsing a Yogācāra position. He explains that Dharmakīrti ultimately maintains that subject and object are non-different and that it follows from this ultimate unity that both the supposed objects external to consciousness and the supposed subject of the perceptual experience, that is to say both the perceived and the perceiver, are ultimately unreal. As verse 214 suggests, the only grounds we have for thinking that there are objects external to consciousness is that that is how our experience seems to us to be. However, our analysis shows that neither subjects nor objects can exist the way they seem to us to exist. The appearance of subject and object are illusions. All that remains is the true, non-dual nature of cognition, which is undifferentiated consciousness -- “suchness” or tathatā. However, since ordinary beings that suffer from ignorance (avidyā) cannot help experiencing consciousness as stratified between subject and object, in verse 219 Dharmakīrti explains, Thus by those who disregard the truth (upekṣitatattvārthaiḥ), imitating an elephant closing one eye, the thought of an external thing is displayed merely by means of the understanding of an ordinary person. Therefore, the Buddha (and Dharmakīrti too) talk as if there are objects external to consciousness in order to be able to engage with beings that due to their karma are not able to see past the illusion of subject and object. However, ultimately there are no such things. The view that the apparent distinction in consciousness between subject and object is illusory and that these two are “empty of characteristics” (lakṣana-śūnyatva) and ultimately “without essence” (niḥsvabhāva) is a view that is characteristic of Yogācāra. It is probably for this reason that the Yogācāra view is attributed to Dharmakīrti. Furthermore, that Dharmakīrti is ultimately read as a Yogācārin is not surprising given that Dignāga, the philosopher that Dharmakīrti comments on in his Pramāṇavārttika, is also read as a Yogācārin by the commentarial tradition. For example, Jinendrabuddhi’s commentary on Dignāga’s Pramāṇasamuccaya argues that, while Dignāga intended his work to be compatible with both the realists who maintain the existence of an external world and those idealists who deny it, ultimately Dignāga sides with the idealists by taking ultimate reality to be undifferentiated consciousness. Most modern scholars of Dharmakīrti’s work, in light of these commentaries, also take these passages in the Pramāṇavārttika to indicate that Dharmakīrti’s final position is that of the Yogācāra. Where these scholars tend to differ is with respect to how one ought to understand the nature of the Yogācāra position that they take Dharmakīrti to endorse. Among the scholars that take Dharmakīrti to be a Yogācārin, the Yogācāra view is most often identified as a kind of idealist position, specifically the kind of idealism that denies the existence of mind-independent, external objects. Georges Dreyfus (1997), for example, argues that Dharmakīrti’s argumentative strategy leads the reader up a hierarchy of positions to the Yogācāra view, which, as he characterizes it, denies the reality of the external world and maintains that objects exist only as reflections of consciousness. He explains this position in the following way: the view that external objects exist depends on the realist assumption that they exist as they appear to our mind. On this basis of objects appearing to our mind as existing independent from our consciousness, we decide that there are objects external to consciousness. Once this basis is questioned, the view that there are external objects is deprived of main support. We then understand the plausibility of the Yogācāra view that consciousness does not need any external support to perceive objects, not even that of infinitesimal atoms. The impression of extended external objects is not produced from external conditions but arises from innate propensities (vāsanā, bag chags) we have had since beginningless time. Under the power of these internal conditions, we constantly but mistakenly project the false impression that there are external objects existing independent of consciousness. (103) This position is taken by Dreyfus to be reflective of Dharmakīrti’s ultimate position, a deeper view of reality than the one reflected in the realist or Sautrāntika position that Dreyfus claims Dharmakīrti adopts only provisionally in other places in his work (104). John Dunne (2004) for the most part agrees with Dreyfus but is reluctant to label this ultimate position Yogācāra and opts instead to call this position “Epistemic Idealism,” in order to emphasize that Dharmakīrti’s arguments are from an epistemological (as opposed to a metaphysical) point of view. He says, “Dharmakīrti’s critique of extra-mental entities arises in the context of determining what it is we know in perception” (60). However, he seems to take the conclusions of the epistemological arguments to be just the conclusions that Dreyfus identifies in the passage above. Thus he characterizes “Epistemic Idealism” as the view that “all entities are mental” (59), indicating that he still takes the argument to be of ontological import. The relevant argument here, which as we saw above can be found in Dharmakīrti’s work, is something like the following: our only reason for believing that there are external objects is that there appears to be a distinction between internal subject and external object in our awareness. However, such a distinction is illusory. Since our belief in external objects is based on an error, we have no reason for believing that such things exist and thus we have no reason for believing that the objects as they appear in awareness are caused by anything external. So the claim that there are extra-mental objects is false. B. K. Matilal (2005) agrees that Dharmakīrti is a Yogācāra idealist, as characterized above, and provides some additional reasons for believing that such was indeed Dharmakīrti’s view. He points out that Dharmakīrti wrote a work, the Saṃtānāntarasiddhi, in which he attempts to refute the charge of solipsism -- a problem typical of idealist views. Matilal also notes that other non-Buddhist philosophical interlocutors, for example Uddyotakara, Kumārila, Jayanta, and Bhāsarvajña, who all tried to refute the Buddhist idealist position, understood Dharmakīrti to be among the Buddhists that defend such a view. An interpretation of Yogācāra’s idealism is given by A. K. Chatterjee (1962), who takes Dharmakīrti’s position, in so far as he accepts the vijñānavāda of Yogācāra position, to be specifically a form of absolute idealism. The ultimate reality on this view is pure consciousness free from the duality of subject and object. Chatterjee has this to say about idealism: “Idealism, in the strictest sense of the term, connotes three important things, viz. (a) knowledge is creative; (b) there is nothing given in knowledge; and (c) the creative knowledge is itself real” (x). By “knowledge” Chatterjee means discursive knowledge. By claiming that knowledge is creative he highlights that for the Yogācāra discursive knowledge involves construction by the mind. In saying that nothing is given in knowledge he highlights the anti-realist aspect of the Yogācāra position that denies that any kind of conceptualization can reflect absolute or objective reality. And (c) presents the Yogācāra idealist position that claims that all there is, what reality ultimately consists in, is just this constructive activity of mind. Chatterjee agrees with most scholars that Dignāga and Dharmakīrti accept the Sautrāntika position provisionally. For this reason Chatterjee distinguishes the Yogācāra of Dignāga and Dharmakīrti from the “pure idealism” of Maitreya, Asaṇga, and Vasubandhu. Chatterjee says of Dignāga and Dharmakīrti that “they essentially accept the doctrine of vijñaptimātratā and the unreality of the object,when they enter into logical discussions however they endorse the Sautrāntika standpoint of something being given in knowledge” (x). The Sautrāntika affirm that there is something ultimately real (i.e., dharmas) reflected in our discursive knowledge, which the Yogācāra denies. Chatterjee agrees that Dignāga and Dharmakīrti provisionally accept such a view, even if ultimately they reject it. While the Yogācāra interpretation is the most widely held reading among Dharmakīrti scholars, there remain several scholars who do not want to conclude that the Yogācāra view is Dharmakīrti’s ultimate position. For example Dan Arnold (2005) and Richard Hayes are of the opinion that Dharmakīrti is purposefully ontologically neutral. They acknowledge that Dharmakīrti’s work contains allusions to what they identify as Sautrāntika and Yogācāra philosophical positions, but they hold that Dharmakīrti does not take sides in the debate between them. Rather, they take Dharmakīrti as presenting his epistemology in such a way as to be consistent with either view. Arnold (2005), for example, characterizes Dharmakīrti’s epistemology as a kind of empirical foundationalism involving something like sense-data that “allows the possibility of bracketing the question of what might finally exist in the world” (23). Since both perspectives consider that “what we are immediately aware of – which is different from the ontological issue of what there is – is only things somehow intrinsic to cognition,” whether or not that mental content represents an external world is not important to his view – “the epistemology is the same either way” (5). Similarly, Richard Hayes is of the opinion that Dharmakīrti’s ultimate ontology cannot be decided on the basis of the available evidence. He argues that the notion that the argumentative method Dharmakīrti employs constitutes a hierarchy of views that places the Yogācāra at the top is a view that is presented by later Tibetan commentators and that such an interpretation is not warranted by the texts themselves. Mark Siderits (2007) agrees that Dharmakīrti and Dignāga were “careful to formulate their epistemology in a way that is compatible with a variety of metaphysical positions” (208). However, Siderits is of the opinion that Dharmakīrti was probably ultimately a Yogācāra idealist (208, fn1), a view he characterizes as “the metaphysical claim that nothing exists that is independent of the mind” (147, fn2). Another group of scholars interpret Dharmakīrti as a Yogācārin but understand that position to be a kind of phenomenology. Dan Lusthaus (2002), for example, reads Yogācāra in general as Buddhist phenomenology and includes Dharmakīrti in this Yogācāra camp. Lusthaus argues that Yogācāra should not be understood to be a form of idealism because idealism is a kind of ontology (only mind or consciousness is real) and Yogācāra is not offering another alternative ontology. Rather, their position is trying to overcome the propensity to ontologize altogether (535). While these scholars vary in their interpretation of what the Yogācāra position exactly is, they all agree that the claims made in Dharmakīrti’s Pramāṇavārttika present ideas that are characteristic of Yogācāra Buddhism. Thus they all agree that Dharmakīrti, even if only provisionally, accepts such a position. These scholars also tend to agree that the majority of the text is not presented from that Yogācāra perspective. The view he is taken to adopt, only provisionally according to most Dharmakīrti scholars, is a Sautrāntika position. In the remainder of this chapter I will present that view and then argue that there is an important difference, relevant to the question of conventional truth, between Dharmakīrti’s position and the Sautrāntika/ External Realist position: Dharmakīrti’s position, unlike the Sautrāntika, is an ontological anti-realist position. Dharmakīrti as Sautrāntika The Sautrāntika Abhidharma position developed as a critique of Sarvāstivāda-Vaibhāṣika Abhidharma. While the Sautrāntikas accepted most of the Abhidharma framework of the Sarvāstivādins, they differ from them in one important respect: they reject the claim that dharmas exist in the past, present, and future and argue instead that only present dharmas exist and that these are momentary, lasting only for a point-instant of time. This key difference gives rise to various other differences in the Sautrāntika account of the dharmas and how they operate. One of these differences concerns how the Sautrāntika account for the delayed effects of karma. Since they reject the notion that a dharma could be said to exist as a kind of potentiality before becoming actual, the Sautrāntika also have to reject the Sarvāstivādin claim that the delayed fruition of karma is explained by appeal to the existence of a special kind of “possessor” dharma that holds onto the karma from past actions and intentions as a potentiality to be actualized at a later date. Instead, the Sautrāntika claim that present dharmas are perfumed by present actions and intentions, and that they pass along this perfuming as metaphorical seeds to subsequent dharmas until the conditions are ripe for the karmic “seed” to come to fruition. Another important difference that the doctrine of momentariness has for the Sautrāntika position is that it requires that they give an indirect account of perception. The Sautrāntika argument for indirect perception is presented in a commentary to the Abhidharmadīpa (a Vaibhāṣika text) in the following way: The organs and the objects of the five sense-consciousnesses, being causes of the latter, belong to a past moment. When the object (rūpa) and the eye exist, the visual consciousness is non-existent. When visual consciousness exists, the eye and the object are not existing. In the absence of their duration (sthiti) there is no possibility of the cognition of the object... [therefore] all perceptions are indirect. (Kalupahana 1976, 103) According to this argument, because dharmas exist only for a moment and the cognition of an object is understood to be the result of the contact of object-dharmas with sense-organ-dharmas, the resultant cognition cannot exist in the same moment as the contact of the object and sense organ, since each of these only exists in the moment of contact. According to the Sautrāntikas there can be no simultaneous causation so the cognition of the object, as an effect, cannot exist at the same moment of its cause -- the contact of object and sense organ. This means that unlike the Sarvāstivādins who gave a direct account of perception, the Sautrāntikas must give an indirect account of perception. Siderits (2007) characterizes this indirect account of perception as “representationalism” (130). He explains this view as “the theory that in sense perception, the intentional object is not an external object but a representation” and then explains that a representation is “a mental image that is caused by contact between the sense faculty and an external object, and that resembles the external object” (131). This view is an indirect form of perception because the perceiver is only aware of the object indirectly through a representation. Siderits explains that even though perception is indirect in this way, “a representationalist does not deny that through our sensory experiences we become aware of things outside the mind. What they deny is that these are what we are directly aware of” (131). According to the representationalist Sautrāntika, then, what we are directly aware of is the representation of the object present in the cognition that arises as a result of the contact between the object and the sense organ, which occurs before the arising of the cognition. Dharmakīrti’s account of pramāṇa is taken to be a Sautrāntika type account since it appears to share characteristics of the Sautrāntika Abhidharma view. Dharmakīrti’s account of pramāṇa is given primarily in his Pramāṇavārttika, which purports to explain how it is that human beings come to have knowledge. The knowledge in question is understood by the Indian tradition to be a cognitive event -- a knowledge episode (pramā) -- and a pramāṇa is understood to be the means to that knowledge event. In the Pramāṇavārttika Dharmakīrti presents his distinctive account of pramāṇa, i.e., of how it is that knowledge episodes (pramā) are generated. In what is arguably the first chapter of the Pramāṇavārttika, the Pramāṇasiddhi chapter, Dharmakīrti gives a general account of what, on his view, constitutes a pramāṇa. In what follows I offer a verse-by-verse paraphrase of what he says in this section. His initial definition of pramāṇa is as follows: A pramāṇa is a non-contradictory cognition. Non-contradiction amounts to the sustainability, or non-frustration, of a goal driven activity. Or, a pramāṇa is that which reveals previously uncognized objects. (3a-c; 7c) In these verses we are told, first off, that a pramāṇa is a cognition. This is interesting because the way in which the cognizing process is usually described in most Indian pramāṇa theories, the cognition is considered to be the distinct result of a cognizing process and it is the process that is the candidate for being a pramāṇa. If the process is one that generates a cognition that is a knowledge episode (pramā), then the process is a means of knowledge (pramāṇa). According to some authors, in the case of visual perception, for example, which is widely accepted as a pramāṇa, it is the process whereby an object comes into contact with a well-functioning eye, under favorable conditions, that is the pramāṇa and the visual cognition of the object is the resulting pramā. Dharmakīrti rejects the typical account of pramāṇa by insisting instead that pramāṇas are cognitions. According to him, the visual cognition of the object is itself the pramāṇa. In verses 5 through 7 of the same chapter he explains that: It is cognition, or thought, that is by nature pramāṇa instead of the contact between the sense-organs and objects, etc. as the Hindu orthodox schools maintain because it is cognition that is the principal cause of activity with respect to things that are to be avoided or to be pursued. The cognition is pramāṇa also because differences in apprehension, or knowledge events, are due to a difference in the appearance of the objects in cognition, which in turn is due to the fact that the apprehension takes place only when the form of an object is present in cognition. From the cognition itself, its own form is understood. Whether or not a cognition is a pramāṇa, on the other hand, is determined in the course of ordinary life. (5b-7a) In these verses Dharmakīrti cites two reasons for identifying the cognition as pramāṇa. The first reason is that it is in virtue of the fact that objects appear in cognition and in virtue of how they appear to us in cognition that we are stimulated to action. Depending on the particular form the object takes in cognition, it appears either desirable or undesirable and we are stimulated to action accordingly. Only cognized objects, not objects simpliciter, stimulate such action and then the particular actions they stimulate will depend on how those objects appear to us. If pramāṇa is understood to be the element in the knowing process that presents us with objects we can then act upon, then it is the cognition of the object that serves this role and thus that should be identified as pramāṇa. The second reason Dharmakīrti gives for why the cognition is the pramāṇa is that knowledge events (pramā or adhigama) can only be differentiated by means of the differences in the appearance of the objects in cognition. This is so because in order for there to be an apprehension of an object in the first place, the object must be present in cognition. It is this presentation that is apprehended or known, and which then serves as the distinguishing feature of the apprehension, differentiating it from all other knowledge events. If pramāṇa is understood to be the element in the knowing process that determines the object apprehended and individuates knowledge events, then it is the cognition that serves this role and thus should be identified as the pramāṇa. Towards the end of this passage when Dharmakīrti writes that “from the cognition itself, its own form is understood” he tells us that the form the cognition takes, that is to say its intentional content, is understood from the cognition itself. This is basically an articulation of “content internalism,” which is the view that claims that intentional content is determined by properties intrinsic to the mental state, or in this case cognitive event, for which it is the content. In other words, the conditions for the content being the kind of content it is are given in the cognition itself. This view may be compared to that of Locke, who argues that we have an intuitive knowledge of the identity of our ideas, of the form each takes. This is an important point because it is in virtue of the fact that the content of a cognition is determined by and evident from the cognition itself that we can, and do, differentiate knowledge events by their content and act accordingly. Thus, the reasons that Dharmakīrti cites as evidence for pramāṇa being a cognition seem to require a content internalist account of intentional content. This account of why cognition and not the contact of the object with the sense organ is considered to be pramāṇa,, suggests that Dharmakīrti shares the indirect account of perception of the Sautrāntika Abhidharma. If all we have direct access to are representations present in cognition, then it stands to reason that it would be through those cognitions that we would come to have knowledge. Even if we grant that those cognitions arise due to the contact of object and sense organ, since all we have direct access to are the cognitions, even that process would have to become known to us indirectly through the cognitions they cause, for example, via a process of inference from cognition to cause. Attributing an indirect account of perception to Dharmakīrti also explains the important role that he takes confirmation “in the course of ordinary life” and “noncontradiction,” understood in terms of sustaining goal oriented activity, to play in his account of pramāṇa. As he says above, while we can know what our cognition is about from the cognition itself, we cannot know from the cognition itself whether or not it is a pramāṇa. In other words, we cannot know whether or not what is presented to us in cognition represents something real from the cognition itself. Since all we have direct access to is the representation, and it is possible for cognitions to misrepresent, the only way we can confirm whether or not the cognition is a pramāṇa is, again, indirectly by testing the cognition against experience. Because the cognition does not give us direct access to the object, we need to go beyond the cognition in order to confirm it. According to Dharmakīrti, it is those cognitions that can withstand the test of experience that are pramāṇas. But interestingly enough, it is only in virtue of the fact that cognitions present us with a certain form and generate certain kinds of expectations that we can even be in a position to test them. So again, the means of coming to know depends crucially on the nature of the original representational cognition. In verse 3 above, Dharmakīrti specifies that not all cognitions are pramāṇa. Only those that are “non-contradictory” and that do not present previously cognized objects are pramāṇas. The only two kinds of cognitions that fulfill these requirements are, according to him, perceptual cognitions (pratyakṣa) and inferential cognitions (anumāna). In the initial verses of the Pratyakṣa (Perception) chapter of the Pramāṇavārttika Dharmakīrti begins by distinguishing the two pramāṇas, perception and inference, on the basis of the distinct content that each pramāṇa has. These verses are as follows: Pramāṇa is of two kinds because its objects are of two kinds. The objects are of two kinds because the object either is or is not efficacious in bringing about an action that achieves its objective aim. Illusions such as floating hairs, etc., are not objective things because they are not apprehended to be such. The objects of pramāṇa are differentiated also for the following reasons: one kind shares a common nature with other objects while the other is totally unique in nature, one kind is expressible in language while the other is not, and cognition of one kind of object requires the existence of additional causes while cognition of the other requires only the presence of the object itself. That object which is capable of bringing about an action that achieves its objective aim is considered in this respect to be ultimately real. The other object, which lacks this capability, is conventionally real. These two are called the particular (svalakṣaṇa) and the universal (sāmānyalakṣaṇa). (1-3) In these verses we are introduced to the two kinds of objects that are cognized by the pramāṇas. In fact, in the above verses Dharmakīrti distinguishes the two pramāṇas by making reference to the kind of content that each cognizes. What Dharmakīrti above calls svalakṣaṇas are the objects of perceptual cognitions. These objects are inexpressible, unique particulars that importantly are efficacious in bringing about an action that achieves its objective aim. Inferential cognitions, on the other hand, cognize what Dharmakīrti calls sāmānyalakṣaṇas. These objects are broadly speaking conceptual. In a later verse Dharmakīrti cites numbers, universals, actions, and substances as the kinds of thing that are included in the class of sāmānyalakṣaṇa. This suggests that what Dharmakīrti has in mind are those objects that require a certain kind of construction by the mind. In addition to being particular, unique, ineffable, and efficacious, svalakṣaṇas are also understood to be simple and momentary (Dunne 2004, 80). While there is some debate about whether or not svalakṣaṇas also lack spatial extension, it is clear that like the Sautrāntikas, Dharmakīrti also understood svalakṣaṇas to be momentary. In verse three above we are told that the object of perception, the svalakṣaṇa, as that which is capable of being efficacious, is ultimately real (paramārthasat) while the sāmānyalakṣaṇa is only conventionally real (saṃvṛtisat). This identification of the svalakṣaṇa as that which is ultimately real, in contrast to the only conventionally real sāmāṇyalakṣaṇa, suggests that Dharmakīrti is adopting a framework that is like that of Abhidharma. Like the dharmas of the Sautrāntika, svalakṣaṇas are not the bearers of properties but are themselves property-particulars. Furthermore, like the Sautrāntikas, Dharmakīrti claims that these particulars are what are ultimately real and adopts a kind of nominalism about everything else, which he terms sāmānyalakṣaṇas. Katsura (2011) traces Dharmakīrti’s account of svalakṣaṇas and sāmānyalakṣaṇas back to the ontology and epistemology of the Sarvāstivādins and the Sautrāntikas. He highlights the similarities between the Abhidharma’s ambiguous use of the terms svabhāva and svalakṣaṇa and the distinction Dharmakīrti draws between svalakṣaṇas and sāmānyalakṣaṇas. Dharmakīrti and the Ābhidharmikas both use the term ‘svabhāva’ in two different ways to refer both to unique individuals and to members of a class. However, while the Ābhidharmikas use the terms svabhāva and svalakṣaṇa interchangeably, Dharmakīrti reserves the term svalakṣaṇa for the unique particulars and uses the term sāmānyalakṣaṇa to refer to members of a class. Despite this difference, Katsura argues that Dharmakīrti is still operating within the same framework as the Abhidharma and basically follows the Sautrāntika ontology (278). Similarly, Siderits (2007) explains that the only significant difference between Dharmakīrti’s views and the Sautrāntika view is that Dharmakīrti (and his predecessor Dignāga) was more consistent in his nominalism. Siderits explains that, Where Abhidharma sees its ultimate reals, dharmas, as falling into natural kinds, Yogācāra-Sautrāntika [i.e., Dignāga and Dharmakīrti] says its ultimate reals, the particulars, are each unique and thus indescribable. If the ultimate nature of reality is how the world is independently of all mental construction, and shared natures like redness are mental constructions, then no real entity can be ultimately like any other. The reals are just unique particulars. (213) Here Siderits alludes to the same key innovation made by Dignāga and Dharmakīrti that Katsura (2011) describes. The term svalakṣaṇa for the Abhidharma was used to refer to the distinguishing characteristic that all dharmas of a particular kind shared only with themselves and no other kind of dharma. Dignāga and Dharmakīrti limit this term and use it not to distinguish between classes of entities or to identify members of a class, but rather to distinguish every particular entity from every other. As Siderits notes, the analysis that lead the Ābhidharmikas to posit the dharmas as the foundational and ultimately real entities stopped at the point where each dharma is characterized by type, by what is referred to as its svalakṣaṇa. But understanding entities as tokens of a type requires the kind of distribution of one over many that the Ābhidharmikas argue is a mark of conceptual construction. In order to complete the analysis, therefore, Dharmakīrti uses the notion of svalakṣaṇa and radicalizes it by claiming that each ultimately real entity is its own unclassifiable thing. It is not, of itself, a token of any type. Dharmakīrti reserves the term sāmānyalakṣaṇa for referring to type-tokens, which he takes to be only conventionally real. Dharmakīrti’s version of the dharma, the svalakṣaṇa, is then understood to function in the same way that the dharmas are understood to operate in the Sautrāntika framework. For example, Vincent Eltschinger (2010) explains that Dharmakīrti’s account of pramāṇa presupposes the indirect perceptual process that the Sautrāntikas accept, with the only difference being that instead of dharmas it is svalakṣaṇas that serve as the causally effective entities that through their contact with the sense faculties give rise to cognition. Eltshinger explains that “this cognition reflects the undivided particular in its entirety, in all its aspects (sarvātmanā, sarvākāreṇa). The image it displays provides a vivid and isomorphic perceptual counterpart of the (causally efficient phase of the) real entity which it takes as its object” (402-403). The resultant perceptual cognition, which accurately represents the svalakṣaṇa, presents the cognizer with undistorted non-conceptual representation of ultimate reality. Problems with the Sautrāntika Interpretation of Dharmakīrti The Sautrāntika interpretation of Dharmakīrti’s theory of pramāṇa is an ontological realist position. More specifically, this position is understood to be a nonidealist ontological realist position that accepts the existence of objects outside the mind. Dharmakīrti’s adoption of the Yogācāra position at verse 208-219 of the Pratyakṣa chapter of the Pramāṇavārttika is interpreted as a rejection of that Sautrāntika position. In particular it is usually understood to present an alternative idealist ontology that is incompatible with the ontology presented by the Sautrāntika position. According to many scholars the result is that a unified account of Dharmakīrti’s philosophy is not possible. In order to reconcile this seeming incompatibility several scholars have argued that Dharmakīrti employs a hierarchy of views and a sliding scale of analysis. According to this view, Dharmakīrti employs more than one systematic description of reality, and it appears that these descriptions are mutually contradictory. Nevertheless, these descriptions are not in fact contradictory because they are not applied at the same level of discourse; instead, they fall into a hierarchy of discourses... where more accurate descriptions of what we perceive and think supersede less accurate ones. (Dunne 2004, 53) On this view, the Sautrāntika level of discourse, which Dunne calls the level of “External Realism,” is employed for a significant portion of Dharmakīrti’s work only for pedagogical reasons. In the spirit of the Buddhist notion of skillful means, Dharmakīrti is said to use this less accurate kind of discourse in order to help his audience, who may not be ready for the ultimate truth, to gradually arrive at it. Dunne explains that “a treatise that exceeds the abilities of its audience would not remove their confusion; hence, a composition that is superior in its analytical accuracy may be inferior soteriologically in relation to a particular audience” (54). For this reason, it is argued, Dharmakīrti employs multiple levels of discourse. The analytically inferior positions are eventually discarded and superseded by increasingly more accurate levels until the Yogācāra position is arrived at (Dunne refers to this position as “Epistemic Idealism”) (58-59). The interpretation of Dharmakīrti’s work as employing a hierarchy of views presupposes that the Sautrāntika position and the Yogācāra position are incompatible. This incompatibility is due to each position presenting a conflicting ontology. This presupposition in turn presupposes that both the Sautrāntika and the Yogācāra views are ontological realist views; it is only if both views are taken to present conflicting ontological existence assertions, that is to say, it is only if these views are taken to be giving different purportedly objective answers to the basic question of ontology, that it makes sense to say that they conflict with one another. However, as I argued in chapter 3, the Yogācārins are not ontological realists. They may be metaphysical realists, i.e. they may maintain that there is some way reality ultimately is, but the Yogācāra are not ontological realists because the way they understand the nature of ultimate reality as non-conceptual and ineffable precludes the possibility of there being objective answers to the basic question of ontology. Such a reality cannot determine an absolute domain of entities over which existence assertions can quantify. Yogācāra ontological anti-realism is therefore best understood as a metaontological position that only comes into conflict with the Sautrāntika position if the Sautrāntika position is understood to be making ontological existence assertions -- only if that position is understood to be an ontological realist position. While this is clearly how the Ābhidharmikas understand the position, there is evidence that this is not how Dharmakīrti understands it. In verse four of the Pratyakṣa chapter of the Pramāṇavārttika, Dharmakīrti makes it clear that he does not take the distinction he draws between ultimately real entities and conventional entities to be an absolute or objective distinction. Responding to an objection against the causal power of svalakṣaṇas, he says, If it be objected that nothing is efficacious, we say in response that this kind of effectiveness is seen when, for example, we see a sprout arise as a result of a seed. If it be objected that such effectiveness is only admitted by convention, then we answer, let it be so. (4) In this verse, in responding to an objection to the effect that the capacity to be efficacious (arthakriyāsamartha), which is the distinguishing mark of an ultimately real entity, is itself only conventionally admitted, Dharmakīrti responds with “astu yathā tathā” – let it be so. In so answering Dharmakīrti undermines a realist interpretation of the distinction between ultimate and conventional reality by relegating the very capacity that is supposed to be an indicator of an entity’s ultimate reality to a conventional designation. The understanding that the svalakṣaṇa is an entity that exists distinct from consciousness and that causes an image of itself to arise in consciousness, is a view that can only be accepted, if at all, at the conventional level because the causal process in general is itself only accepted conventionally. Relegating the distinction he draws between ultimate and conventional reality to the level of convention suggests that Dharmakīrti does not intend for the account he gives with respect to this distinction to constitute an attempt at giving an objective answer to the basic question of ontology. This account is not intended as an ontological realist account. And if this right, then there is no conflict between that account and the Yogācāra account. There is no need, therefore, to interpret Dharmakīrti as presenting a hierarchy of views since the two main positions he adopts are not in conflict. The compatibility of the so-called “Sautrāntika” position and the Yogācāra position is further suggested by the account that Dharmakīrti gives of the nature of the svalakṣaṇa. Given their ineffable and radically particular nature, it may be inappropriate to understand Dharmakīrti’s account of svalakṣaṇas as presenting even a conventional domain of entities. Examine, for example, the following comments made by Siderits (2007) about the particularity that Dignāga and Dharmakīrti ascribe to svalākṣaṇas. So here [with respect to the particularity of svalakṣaṇas] Yogācāra-Sautrāntika seems to be honoring the Yogācāra side of its heritage. Diṅnāga and his followers are also saying that the true nature of reality is inexpressible. This is not just because we humans can’t describe things carefully enough, or in fine enough detail. Any words that were used to describe reality would falsify it...If the real particulars are genuinely unique, they can never be described. We are aware of them in perception. But because they don’t have shared natures, we could never express the content of our awareness in words. (213) Here Siderits points out that because svalakṣaṇas are not tokens of any type, they are inexpressible and, as such, they cannot be the basis on which answers can be given to ontological questions, even at the conventional level. They cannot constitute a domain of entities in the sense relevant for answering ontological questions at any level. If this is true, then Dharmakīrti’s account of svalakṣaṇas precludes the possibility of attempting to reduce sāmānyalakṣaṇas to them the way that the Abhidharma attempts to reduce ordinary objects to dharmas. This suggests that the svalakṣaṇas are playing a very different kind of role in Dharmakīrti’s account of pramāṇa than the usual Sautrāntika interpretation claims. The preceding considerations lead me to conclude that Dharmakīrti is an ontological anti-realist and is in this respect to be considered a Mahāyānist. This means that his account of pramāṇa, and the associated account of meaning, are in a position to help the Mahāyāna answer the question about the sense in which conventional truth is true. In the next chapter I will show how Dharmakīrti’s work provides the resources to give an account of meaning that can supplement the deflationary theory of truth that Priest, Siderits and Tillemans propose as an account of conventional truth. CHAPTER 5 DHARMAKĪRTI’S THEORY OF MEANING In the last chapter I argued that Dharmakīrti should be understood as presenting an ontological anti-realist position and that for this reason he should be counted as follower of Mahāyāna. I argued that it is possible to read his account of pramāṇa as consistent with his overall anti-realist position when the ontology that that account presupposes is understood to be accepted only conventionally. Dharmakīrti does not, I argued, present that ontology as being objective or absolute in the way required of realist views. Therefore his account of pramāṇa, along with the commitments to the associated entities, need not be superseded or abandoned in favor of the Mahāyāna position because it is fully consistent with that view. In this chapter I will argue that insofar as Dharmakīrti can be read as presenting his account of pramāṇa as explicitly conventional, what he presents to us is in effect an account of the framework for what Carnap calls “the world of things.” This account, which is ultimately pragmatically justified, explains how the propositions concerning empirical objects -- things -- can be meaningful and referential despite the global anti-realist position that all Mahāyāna Buddhists share. This provides the account of meaning that is required to supplement the deflationary theory of truth that in chapter three I endorsed as the most promising account of truth for Mahāyāna Buddhism. I will begin this chapter by outlining the role a framework plays in an antirealist metametaphysics. Drawing on the work of Price and Horwich I will explain how functional differences in language use can be understood to mark boundaries between linguistic frameworks and how once set, frameworks allow for a deflationary account of reference. Since this account makes questions about language and meaning more fundamental than those of ontology, I then proceed to explain Dharmakīrti’s account of one particular use of language: describing our environment in terms of discrete and classifiable things and events, in other words, Carnap’s thing-language. I explain the unlike the realists that explain our use of such language by appealing to the independent nature of the world, Dharmakīrti instead gives an account of how the world of things is constructed out of non-conceptual and ontologically neutral perception. This process, which involves the distinctively Buddhist notion of apoha, shows how an engaged agent can come to understand the world in terms of discrete and classifiable empirical objects without presupposing an absolute domain of entities. Because the concepts developed by the this process are generated for the purpose of guiding behavior and are understood in functional terms, the result is a use theory of meaning. This use theory then serves as the foundation for drawing a natural boundary around the distinctive use of the thing-language. The empirical framework of that language, I argue, is precisely what Dharmakīrti’s account explains. Thus I understand Dharmakīrti’s account as providing a Carnapian, use theory of meaning that allows for a deflationary understanding of reference and truth, at least with respect to empirical claims. The Role Of A Framework In chapter three I introduced Carnap’s (1950) view that questions concerning the existence of entities can only be meaningfully posed with reference to the linguistic framework that talk of such entities presupposes. A framework, for Carnap, is a system of “rules for forming statements and testing, accepting, or rejecting them” (23). According to him, any statement or question must be made as part of a larger system of language whose use is governed by a certain structure and set of rules. A Wittgensteinian way of putting the point might be to say that making claims and asking questions are moves made within a larger language-game whose structure and rules make those moves meaningful as parts of the game and govern what counts as an appropriate and acceptable move. On this view, in the very act of asking a question, such as “Is three a prime number?”, we are invoking a larger language system that makes talk of numbers meaningful and provides rules according to which an answer to the question can be derived. Carnap’s point is that we cannot make sense of asking a question about the existence of an entity outside of a linguistic framework that provides the system and structure necessary to make the question meaningful and to provide the resources to answer it. And thus, when we ask basic questions such as “Do numbers exist?”, the answers end up being trivial insofar as any system of language that talks about numbers -- a system of language that must be employed even to ask the question -- presupposes their existence as part of its framework. The question only makes sense internal to that framework and from that perspective the answer is trivial. As I noted in chapter three, for Carnap the acceptance of a domain of entities in light of our acceptance of particular ways of talking is metaphysically noncommittal; it does not make one a metaphysical realist about those entities. The idea that accepting the existence of entities implicit in accepting a form of language does not commit one to a metaphysical realist position fits nicely with the deflationary account of truth that I endorsed in chapter three. Horwich (2005a) notes that insofar as the notions of truth and reference are inter-definable, reference also is deflated such that the following equivalence schema accounts for reference: “(y)[Singular concept N refers to y iff n = y]” (178). According to this schema, for example, the concept DOG refers to an individual if and only if that individual is a dog (77). Our understanding and use of the term ‘refers’ conforms to this equivalence schema and thus we are prepared to infer from the acceptance of the claim that some individual is a dog that the concept DOG applies to it and vice versa. From this simple equivalence it follows that in our employment of a concept, such as DOG, in our language we commit ourselves in a deflationary way to the existence of the referent of the concept, in this case to the existence of dogs. This simply follows from our accepting and using a language that expresses propositions that include the concept. And because the deflated notion of reference does not require any constitutive relation between the concept and the class of objects to which it refers, the use of the concept does not commit the user to any substantive metaphysical position that may be implicitly required of more substantive accounts of truth and reference. For both Horwich and Carnap it appears that questions of ontology are explanatorily less fundamental than question about language. Horwich (2010), for example, argues that the facts that we take to represent reality are, in terms of explanatory priority, dependent on propositions. He says, the schematic relation between p and the fact that p will be There exists such a thing as the fact that p because p rather than the other way around. For we explain why the fact that p exists, not by deducing it directly from laws and initial conditions. But by first deducing that p (thereby explaining why it is that p), and by then invoking the biconditional, ‘The fact that p exists iff p’, to deduce that the corresponding fact exists. Thus ‘the fact that p exists’ is always less fundamental in our explanatory deductive hierarchy than ‘p’ is. (311) It thus follows that from the deflationary perspective, the acceptance of the propositions that our language expresses is primary to there being any particular ontology or set of facts. Similarly, Carnap’s position entails that it is the structure and system of a linguistic framework that will determine the nature of the facts for any particular class of entities. For both these views, language must come first. Because of the central role that language and the associated frameworks play, it is critical that accounts like these be able to give an account of language and meaning that explains how we come to develop and use language without appealing to a pre-existing domain of entities to serve as the content of that language. I have already noted in chapter three that this is a necessary supplement for the deflationary theory of truth. Carnap’s account, while providing a promising general strategy for approaching questions of meaning and reference, is sparse on the details about how frameworks develop and work. This has opened him up to the criticism, famously from Quine, that no meaningful sense can be made of a framework, the boundaries it draws between internal and external, and thus how it operates to determine meaning. While scholars such as Bird (1995) and Price (2010) have made compelling cases that most of Quine’s criticisms of Carnap miss the mark, Price still acknowledges that Carnap does not provide a satisfactory defense of the charge that “there is no principled basis for Carnap’s distinction of language into frameworks” (329). Price explains that, Quine’s claim is that there are no purely internal issues, in Carnap’s sense. No issue is ever entirely insulated from pragmatic concerns about the possible effects of revisions of the framework itself. Pragmatic issues of this kind are always on the agenda, at least implicitly. In the last analysis, all judgments are pragmatic in nature. (326) Quine argues this way because he thinks that the distinction between frameworks is drawn by appealing to the analytic-synthetic distinction, which he argues is untenable. While there is reason to doubt that Carnap uses the analytic-synthetic distinction in the way Quine charges him of using it (see Bird 1995), even if Carnap did draw the distinction that way, it may still be possible to salvage his account by drawing the distinction in some other more plausible way. Price explains further that, The main effect of abandoning the analytic-synthetic distinction is that Carnap’s distinctions are no longer sharp -- there are no purely internal (nonpragmatic) issues, because linguistic rules are never absolute, and pragmatic restructuring is never entirely off the agenda. (326) It is possible to accept this consequence, argues Price, while still maintaining a Carnapian type view by explaining the distinction between frameworks in some other way. Price thinks that underlying Carnap’s distinction between frameworks is the presupposition that functional differences in language mark natural boundaries that generate category mistakes when talk about one kind of entity is confused with talk about another (330). On this view, it is because we use language to do different things, and the kinds of entities we talk about are bound to the language we use to talk about them, that there are different frameworks. Since, for example, talk about material objects, talk about mental states, and talk about numbers are used differently for different purposes, they are structured and operate differently, reflecting the uses to which each kind of talk is put. Price explains this view, which is a kind of marriage between Carnap and Ryle, in the following way: On this view, the subject-predicate form, and indeed the notion of an object itself, have a one-many functional character. In one sense, it is the same tool or set of tools we employ whenever we speak of objects, or whenever we use the subject-predicate from, or -- what seems part of the same package -- whenever we use the existential quantifier. However, there’s no further unitary notion of object, or substance, or metaphysical bearer of properties, but “only a subject position in an infinite web of discourses. ” Similarly, it is the same tool or set of tools we use whenever we speak of truth, whenever we make a judgment or an assertion. But in each case, the relevant tool or set of tools may have incommensurable uses, if there are important senses in which the bits of language they facilitate have different functions (in a way which doesn’t simply collapse into differences in the objects talked about). (332-333) Here Price explains that even though we employ the notions of existential quantification and truth in uniform ways, this does not entail that the concepts to which they are applied are uniform. On this view, by taking the differences between concepts as functional differences, we can explain why it seems strange and inappropriate to claim in the same breath that cats and numbers and Wednesdays exist. In each case the existential quantifier is the same, but because these concepts function differently in language, the resulting existential assertions can also be differentiated in functional terms. Likewise, while there is a basic sense in which we take cats and numbers and Wednesdays all to be objects insofar as they can occupy the subject position in a sentence, this does not entail that they are objects in the same sense. We can mark the difference between these kinds of objects by noting the functional difference between the concepts used to identify them, without needing to posit some deep metaphysical difference. Price thinks that Carnap presupposes such pragmatic and functional boundaries between languages in giving his account of frameworks; Carnap considers our choices of syntax to mark the already present functional boundaries, thereby creating a framework. The syntactical choices are therefore not arbitrary but rather reflect more fundamental, albeit more fuzzy, functional or pragmatic distinctions. The consequences of understanding framework boundaries in terms of functional differences between languages are that the boundaries between frameworks are not sharp or absolute and that frameworks are not static insofar as changes in usage can entail changes in framework. But these consequence need not undermine the intelligibility of the notion that frameworks can be identified and differentiated. Explaining the Thing-language According to the preceding account of frameworks, frameworks are demarcated by natural boundaries created by functional, pragmatic differences in language use. With respect to the Buddhist case Dharmakīrti is particularly concerned to give an account of one particular use of language and the framework it demarcates: our use of language to describe our environment as being composed of discrete and classifiable objects and events. Such an account is, in effect, an account of what Carnap (1950) calls the thing-language, which Carnap describes as the “spatio-temporally ordered system of observable things and events” and as the most basic language, the one we have all accepted early in our lives “as a matter of course” (22). The brahmanical realists of Dharmakīrti’s time explain our use of the thinglanguage by appealing to the real things that such language refers to. According to them, we talk about and describe our environment in terms of discrete and classifiable objects and events because, according to them, our environment is, in and of itself, composed of discrete and classifiable objects and events, to which we have access through perception. Through perception we cognize objects of various types and we formulate concepts and language to correspond to those objects we see. On this view we perceive, for example, type-token objects such as cows. We see a cow both as a discrete object and as an instantiation of cowness. So, according to the realists, we use the thing-language in order to reflect the structure and nature of the world, which is directly presented to us in perception. Dharmakīrti describes our use of the thing-language very differently, in a way that is consistent with his anti-realism. He does not deny that we cognize entities such as cows, however, he denies that these cognitions of type-tokens are perceptual. Consistent with his anti-realism, he claims that what is given in perception does not arrive already organized and categorized. According to Dharmakīrti, we do not experience our environment in the first instance as composed of type-tokens. If antirealism is true, this makes sense given that the world does not already, in and of itself, comprise an absolute domain of entities that could be revealed to us directly in perception. Instead, argues Dharmakīrti, in perception we are presented with the qualitative particulars that Dharmakīrti calls svalakṣaṇas. As I explained in the last chapter, these svalakṣaṇas do not themselves comprise a domain of entities and thus are ontologically insignificant. In light of this they are better thought of as something akin to sense-data rather than as proper objects. Because these svalakṣaṇas are ineffable, qualitative particulars, their representation of the environment is consistent with an anti-realist understanding of it. From this we can see that in Dharmakīrti’s view, there is a sense in which our environment is given to us in perception -- it is given to us as svalakṣaṇas. This givenness, however, which presents itself as of something external to the cognitive act itself (what we are here calling the environment), is, according to Dharmakīrti, built into the very structure of cognition itself, at least for the unenlightened. As was noted in the last chapter, according to Dharmakīrti, it is due to beginningless ignorance that we experience cognition as bifurcated between subject and object -- between a subject who cognizes and an object cognized. Ultimately, Dharmakīrti claims these two are really non-different. Just as the Buddhists generally deny that there is an independent subject that is the haver of cognition, Dharmakīrti also denies that there is an independent object that is represented in cognition. Both the subject and the object are, on this view, ultimately internal to the cognition itself. This means that the externality of the environment that is given to us in our perceptual experiences is a kind of illusion. A helpful way to think about the sense in which svalakṣaṇas represent the environment for Dharmakīrti is to use the distinction that Price (2011) draws between “external representation” (e-representation) and “internal representation” (irepresentation). The first notion, “e-representation,” “gives priority to the idea that the job of representation is to to co-vary with something else -- typically, some external factor, or environmental condition” (20). This notion of representation understands representation to be constituted by some kind of relation between the representation and some external representatum. A representation is thus taken to be about something that is external to that representation and that aboutness is accounted for by appealing to a relation between the representation and the thing represented. Understanding representational structures on this view entails understanding what the system-world link is that makes representation possible. For the brahmanical realists, this link is provided by perception, which links the perceptual cognition to the object that the cognition is then taken to represent. On this external understanding of representation, we understand what the representation is of and assess its accuracy by examining the kind of relation it has to the world. To illustrate how this is thought to go, Price uses the following helpful analogy: Imagine a child’s puzzle book, arranged like this: The left-hand page contains a large sheet of peel-off stickers, and the right-hand page shows a line drawing of a complex scene. For each sticker -- the koala, the boomerang, the Sydney Opera House, and so on -- the reader needs to find the unique outline in the drawing with the corresponding shape. The aim of the game is to place all the stickers in their correct locations, in this sense. Now think of the right-hand page as the world, and the stickers as the collection of all the statements we take to be true of the world. For each such statement, it seems natural to ask what makes it true -- what fact in the world has precisely the corresponding “shape.” Within the scope of this simple but intuitive analogy, matching true statements to the world seems a lot like matching stickers to the line drawing. (3) If we were to follow this analogy in understanding Dharmakīrti’s account of representation, we would have on the left-hand side two kinds of stickers -- perceptual and conceptual. On the right-hand side we would either have two different kinds of shapes, svalakṣaṇas and sāmānyalakṣaṇas, or, given that Dharmakīrti identifies the former as ultimately real and the latter as only conventionally real, we could adopt a reductionist approach that would have on the right-hand side only one kind of shape -- svalakṣaṇas -- but allow the placement of the sāmānyalakṣaṇa stickers by means of some process that maps them onto svalakṣaṇas. In both these cases, the world, represented by the right-hand page of the sticker book, would serve as the standard according to which the accuracy of the representations would be determined. As I explained in chapter four, this way of understanding Dharmakīrti’s account of representation is a common one. However, given Dharmakīrti’s ultimate rejection of the position that what is given in cognition is something external to it, it does not make sense for him to understand representation in this way. Accounts of representation that require some kind of constitutive relation between the representation and some distinct thing represented are not available to him given his anti-realism. To use Price’s analogy, according to the anti-realist there are no predefined demarcations on the right-hand page that can serve as the objective guide in placing the stickers. So, I argue, it is a mistake to understand Dharmakīrti as employing the notion of e-representation in giving his account of the representational capacities of cognition. Instead, I suggest that we think of Dharmakīrti’s understanding of representation in terms of what Price calls “internal representation” or “i-representation.” Price explains that the notion of “i-representation,” which is common in the fields of psychology and cognitive science, gives priority to the internal cognitive role of representation. Price explains that “a token counts as a representation, in this sense, in virtue of its position, or role, in some sort of cognitive or inferential architecture -- in virtue of its links, within a network, to other items of the same general kind” (20). With respect to the sticker analogy, Price says that this kind of representation offers “an account of what gives a sticker its propositional shape; what makes it the particular sticker that it is” in a way that makes no appeal to a relation between the sticker and already defined shapes on the right-hand page. Instead, the account explains representation in terms of the token’s functional role in our “cognitive economy” (22). With this kind of representation in mind, we can understand the stickers on the left-hand page as representing the environment on the right-hand page in a more deflationary way. The structure of the right-hand page is, on this account, determined by the stickers on the left by, say, tracing the outlines of the stickers onto the page. The right-hand page does not provide constraints but is rather structured by the stickers themselves. In order to help clarify how i-representations represent, Price offers the following alternative metaphor: Think of a data projector, projecting internal images onto an external screen. Even better, helping ourselves to one of tomorrow’s metaphors, think of a holographic data projector, projecting three-dimensional images in thin air. This isn’t projection onto an external, unembellished world. On the contrary, the entire image is free-standing, being simply the sum of all we take to be the case: a world of states of affairs, in all the ways that we take states of affairs to be. (28) Applying this metaphor to Dharmakīrti’s view, we can understand svalakṣaṇas as representing the environment in the trivial sense that they constitute a projection of that very environment. It seems to us as if the environment projected by cognition is external to it but upon reflection we can come to understand that this externality is merely an illusion. And according to the Yogācāra Buddhists, if we engage in a long and rigorous process of meditative training, we can eventually come to realize the illusory nature of such cognition directly for ourselves. However, so long as we remain in an ordinary, unenlightened state of mind, our experience remains as of erepresentation -- as of a subject representing in cognition objects external to her. It is clear that from this kind of account that we cannot explain our coming to use language to describe our environment in terms of discrete and identifiable objects and events by making reference to features that the environment already has. In particular, Dharmakīrti’s understanding of perception and its contents makes it clear that our coming to understand our environment in these terms is not the result of perception. This understanding is instead, according to Dharmakīrti, a subsequent constructive product of the mind. The process whereby we come to represent our environment in terms of discrete and identifiable objects and events is given in Dharmakīrti’s account of apoha. Apoha The theory of apoha, the first formulation of which was given by Dharmakīrti’s predecessor Dignāga, in its most basic articulation states that the meaning of kind terms is to be explained in terms of negation or exclusion (apoha) such that, for example, a term like “cow” or “dog” amounts to “not non-cow” or “not non-dog.” This understanding of the meaning of kind terms is intended to avoid commitment to the real existence of universals and is thus considered to be a form of nominalism. Chakrabarti and Siderits (2011) explain the theory generally in the following way: The apoha theory is first and foremost an approach to the problem of universals -- the problem of the one over the many. That problem is one of explaining how it is possible, when we see a pot, to think of it as a pot and call it by the name “pot,” a name that applies to many other particular pots. What is the one thing, being-a-pot, that this particular shares with many other particulars? Is there really such a thing in the world, over and above the individual pots, or is it just a mental construction of some sort? To hold the first alternative is to be a realist about universals, to hold the second is to be a nominalist. The apoha theory is a distinctive Buddhist approach to being a nominalist. (1) The idea behind this distinctively Buddhist approach seems to be, roughly, that the members of the complement class to the class of non-dogs, for example, need not share any common property and thus no universal need to be invoked in order to explain membership in the class. Ganeri (2011) explains this in terms of the absence of identity conditions. He points out there are no identity conditions for membership in the class of, for example, non-dogs; there is a wide variety of things that would be included such as lamps and cats and people. Because this is so, it would seem that there would also be no identity conditions for the members of the complement class, which is determined by negation of the original class. Thus the double negation alleviates the need to appeal to some common property -- a universal -- that is shared or instantiated in all the members of a class. What the apoha view exactly amounts to is a matter of much debate. Much ink has been spilled since Dignāga’s time by scholars clarifying, criticizing, and defending the idea that a double-negation can do the job of a universal. Most of this work approaches the issue from what Tillemans (2011) describes as a “top-down” approach (54). This approach, most often associated with Dignāga’s formulation, takes the function of the double-negation to be to stand-in for universals and to be what a word primarily expresses. However, there is an increasing body of work that seeks to make sense of the apoha theory using a “bottom-up” approach that is associated with Dharmakīrti’s articulation of the theory. This approach explains general concepts by giving an account of their formulation from particulars -- an account that makes no appeals to universals and thus is appropriately nominalist. It is this approach that will be the focus of the following account as it is the account that is presented by Dharmakīrti in his texts. In Dharmakīrti’s hands, apoha, or exclusion, operates as part of the process whereby conceptual cognitions are generated from non-conceptual, perceptual ones. This process begins with perception. As I have already explained, perceptual cognitions represent (i-represent) the environment in a non-conceptual way involving only svalakṣaṇas, which do not themselves constitute objects in the relevant sense. In order to better understand the nature of Dharmakīrti’s non-conceptual, representational, perceptual content and the distinction that Dharmakīrti makes between perceptual and conceptual cognitions, I think it is helpful to look at how the distinction between conceptual and non-conceptual forms of representation is drawn by Jerry Fodor (2007) and Richard Heck (2007). Fodor differentiates between two mutually exclusive kinds of representation: iconic representation and discursive representation. The main difference between these two kinds of representations is that they have different compositional structures and thus represent in very different ways. Fodor says that, “all the kinds of representations we are concerned with are compositional,” that is to say of a representation that its “syntactic structure and semantic content are both determined by the syntactic structure and semantic content of its parts” (107). However, how the parts compose to determine the content differs between discursive and iconic representation. Discursive representations, because they involve concepts, must be structured in such a way as to follow certain rules of combination and composition. For example, subjects and predicates contribute in distinct ways to the composition of a meaningful sentence. Elements like these must be arranged and related in the right kinds of way in order to be structured so as to represent a state of affairs. These various elements and relations contribute in a variety of ways to the composition of the whole, making the structure of discursive representations complicated. However, this complicated structure makes it possible for these representations to have “a galaxy of representational properties that icons don’t,” notably the ability to take logical form and thus play a role in inference (109). Iconic representations, on the other hand, do not have such a complicated structure. The way in which an iconic representation represents is in a manner analogous to how a picture or a map represents. Fodor explains that these kinds of representations have a compositional semantics that follows what he thus calls the “Picture Principle”: If P is a picture of X, then parts of P are pictures of parts of X. This principle entails that iconic representations have structurally homogeneous parts. No matter how you slice up the representation, each part contributes equally to the composition of the whole, much like each piece of a jigsaw puzzle contributes to the composition of the whole no matter how arbitrarily the pieces are cut. This means that iconic representations do not have a specific constituent structure -- there are no joints; there is no specific way that elements must be identified and related. Fodor explains that they are a “homogeneous kind of symbol from both the syntactic and the semantic point of view. They don’t have logical form, they don’t have a structure to make explicit” (109). All the parts of an iconic representation contribute equally to the interpretation and in the same way. Because of this, Fodor points out that iconic representations cannot do the same kinds of things that discursive representations can do. In particular they cannot enter into logical relations and cannot form part of an inference. Iconic representations are still semantically evaluable insofar as they are of something -- they are, after all, still representations -- but they are a different kind of representation in that they represent in a different way. In order to help explain how a perceptual mental state could be iconic, Heck compares it to “cognitive mapping.” He claims that There is strong empirical evidence that our ability to find our way around in the world depends upon our employment of what are known as “cognitive maps.” Each of us has a mental map of our surroundings that places locations we encounter relative to other, known locations. (125) These maps are representational yet they are non-conceptual. We can see this, Heck points out, by noticing that having such a mental map is not the same as having a collection of beliefs about one’s environment. In fact, because the structure of the map is non-discursive, Heck explains that there is no objective way to describe the map’s contents; “there is no unique structured proposition that gives the content of a map because there is no such structure in the map; a map lacks the syntactic structure present in a verbal description of what it represents” (126). Thus a complete description of the environment and a mental map of it are two different things, and one can have the latter without having the former. This is evident, Heck points out, in the phenomenon of “knowing how to get somewhere” without being able to give someone else directions (125). If we follow this way of characterizing the distinction between conceptual and non-conceptual representations and apply it to Dharmakīrti’s work, the representations (ākāras) of perceptual cognitions, which would be composed of qualitative particulars, can be thought of as iconic representations that are not linguistically expressible because they are not appropriately structured. Analogous to Heck’s cognitive mapping, we might think of these representations as qualitative maps of sensory space for each sense modality. A similar line of interpretation is offered by Ganeri (2011) who, drawing on the work of Austen Clark and Christopher Peacocke, argues that what Dharmakīrti’s perceptual representations present us with are “spatially located instances of phenomenal qualities” (229). On this view, perceptual cognitions present a phenomenal appearance of qualitative features, which for Dharmakīrti would be the svalakṣaṇas, and their apparent locations relative to each other and to the perceived subject (234). As was noted before, for Dharmakīrti this experience of qualities distributed in a relatively defined space that is oriented around a subject is part of the illusory structure of cognition for ordinary beings. By my interpretation, this illusory array constitutes an iconic representation -- a primitive sensory map -- that is momentary and confined to a particular sense modality. At this level of perception, there are no objects, no individuals, and no cross-modal connections. As Ganeri notes, this kind of feature placing does not involve the employment of concepts and “isn’t enough for the introduction of objects. For that we need sortal concepts, concepts that provide identity conditions for the things that fall under them” (235). The sortal concepts that are needed for the introduction of objects are Dharmakīrti’s sāmānyalakṣaṇas. The conceptual cognitions that are made up of sāmānyalakṣaṇas can be considered to be discursive representations that, unlike iconic representations, are structured in a way that is expressible (as Dharmakīrti puts it in his Pramāṇaviniścaya, in a way that “is fit to be associated with words” ) and that can thus play a role in inference. The goal of Dharmakīrti’s apoha theory is to explain the cognitive process whereby discursive representations are generated from iconic ones. According to Dharmakīrti, in ordinary, unenlightened circumstances a perceptual cognition initiates a cognitive process that generates a discursive representation on its basis. This process occurs mostly below the level of consciousness, sub-personally, wherein the cognitive system generates a subsequent conceptual cognition that represents the environment in terms of discrete and categorizable objects. It does this by bringing past experience to bear on the interpretation of the sensory information presented in an occurrent perceptual cognition. The process begins by a perceptual cognition triggering certain subconscious mental factors -- latencies, imprints, and capacities -- called vāsanās. In giving an account of the role of vāsanās in Dharmakīrti’s system, Dunne (2011) points out that “although he [[[Dharmakīrti]]] refers repeatedly to imprints [[[vāsanās]]], the precise mechanism of their operation receives no attention” (100). This means that giving a precise account of their operation is difficult. It is, however, clear from Dharmakīrti’s work that vāsanās play a very important role in explaining how the mind generates conceptual cognitions from perceptual ones. For example, in his own commentary to verses 68-70 of the “Svārthānumāna” chapter of his Pramāṇavārttika, Dharmakīrti states: A conceptual awareness arises in dependence on things which are excluded from what is other than those [that have the same cause or effect]. In conformity with the nature (prakṛti) of its imprints, that awareness conceals the distinct form of each of those things, and ascribes to them its own nondifferent appearance. Doing so, that conceptual awareness conflates those things and presents them as nondifferent. Those things conceptualized as nondifferent are excluded from others in that they have the same effects and causes; there is also a cognitive imprint (vāsanās) that induces one to conceptualize those things in that fashion. The nature (prakṛti) of those distinct things themselves and the nature of the imprint are such that the cognition that arises from those things and that imprint appears in this way [i.e., such that those things seem nondifferent]. (Dunne 2011, 339) Here we see that the svalakṣaṇas that are the elements of perceptual cognitions are not alone responsible for the generation of a subsequent conceptual cognition. It is the perceptual cognition in conjunction with vāsanās that “induces one to conceptualize those things in that fashion,” i.e., that leads the cognitive system to interpret the data presented in perception as presenting a describable object, one that can be taken to be similar to other objects. This kind of account is to be expected given the iconic nature of perceptual representation and its incommensurability with discursive representation. It is clear that some additional, intervening factor would have to be in operation in order to derive a conceptual cognition from a perceptual one. For Dharmakīrti, this role is played by vāsanās. In Dharmakīrti’s system, vāsanās thus seem to function to organize and process the information derived from the senses through perception. Dunne explains that generally in the Yogācāra tradition of which Dharmakīrti is a part, vāsanās are “a mechanism for the expression of karma,” that is to say, vāsanās are the way in which past experience gets encoded by the cognitive system (99). From Dharmakīrti’s discussion of vāsanās it is clear that he considers some of these vāsanās to be innate (anādi) and others acquired (āhita) (Dunne 100). The innate vāsanās include certain dispositions that a sentient creature has in virtue of being the particular type of creature that it is. In the Buddhist context this is commonly understood to be a product of experiences from past lives, which are said to determine the kind of form in rebirth a particular consciousness stream (commonly called a “being”) takes. Since this process is said to be beginningless, these vāsanās are also commonly described as beginningless. In more modern and secular parlance, we can understand innate vāsanās in terms of evolutionary acquired dispositions that a sentient being has in virtue of being the kind of being that it is. These are relatively stable dispositions that are not easily susceptible to change in the course of the being’s life. Acquired vāsanās, on the other hand, are formed through experience within a lifetime and develop in number and complexity as more experience is acquired. One important example is linguistic experience. For creatures with the capacity for language, past experiences of language learning and use leave vāsanā imprints that are then brought to bear in interpreting perceptual data from future perceptual cognitions. The generation of conceptual cognitions for such creatures capable of language is thus affected by the acquisition of language, a point that will be relevant in appreciating how linguistic frameworks contribute to determining cognitive content. When triggered, vāsanās operate automatically and unconsciously to transform perceptual experience into a tokening of a type that the system can then use to make the environment easier to negotiate and predict. At the most basic level, this involves associating present experience with past experience in order to draw functional similarities. This involves the operation of one of the innate imprints present in the subconscious of all sentient creatures: a disposition to judge perceived particulars as being the same as previously perceived particulars. Dunne describes this vāsanā as “part of a sentient being’s cognitive architecture” and as what serves as the foundation for the capacity to form even the simplest concepts (101). This vāsanā is a necessary part of the cognitive architecture for any being that represents the world conceptually because the objects of perception are particulars and thus do not themselves share any similarities. The perceived similarities are not read off perceptual experience but must be constructed by the mind. For this reason, Dunne points out, this vāsanā is, and must be, innate; it is not possible to learn through experience. He explains that, this disposition is not learned; indeed, on Dharmakīrti’s view it would be impossible to acquire it through experience because this would require an experience of two objects that are in fact the same, but for Dharmakīrti all perceptible objects are necessarily different in all ways. (101) In order to draw the similarities that are necessary for acts of recognition and concept formation, the minds of sentient beings thus have an innate vāsanā that is disposed to take the contents of present experience to be the same as the content of previous experience. When these vāsanās of past experiences are triggered, the mind is quick to take the present content to resemble the content of the past and thereby takes them all to be tokens of the same type. But as Dreyfus (2011) points out, for Dharmakīrti “the similarities that he posits are not objective but exist only in our experience or, rather, in the ways in which we spontaneously conceptualize our experiences” (213). And it is an innate vāsanā that allows for such spontaneous conceptualization. As the above account makes clear, this innate dispositional vāsanā works in conjunction with acquired vāsanās that encode past experience. Which of the vāsanās from past experience are triggered is determined not only by the initial perceptual cognition but also by what Dunne calls “subjective factors” (98). These subjective factors, which like vāsanās “may be thought of as involving conditioning (saṃskāraṇa)” serve as the background against which the cognitive process occurs (99). These factors include our aversions and desires, our purposes and interests, and the goal-oriented expectations we have in light of those purposes and interests. They basically constitute our engaged situation at the time of cognition and they create the subjective context that helps to bring certain features of the perceptual cognition into focus. These subjective factors, in conjunction with perceptual cognition, thus trigger the vāsanās of past experiences that bear on the situation of engagement and together they bring into focus certain features of iconic perceptual representations. This creates a demarcation between focus and field. What is in the focus range are features that are taken by the system to be functionally equivalent. These are then contrasted with those features that are not taken to be functionally equivalent. Dharmakīrti describes this as an exclusion. For example, in the passage from the Pramāṇavārttika included earlier in this section on pages 127 and 128, Dharmakīrti described the exclusionary manner in which similarities are derived as one in which “things are excluded from what is other than those.” This exclusion occurs when the features in the focus range, though particular, are taken to be non-different and are excluded from the remaining features that do not share the functional similarity. The focus range, which we can identify as X, can thus be described as being not non-X -- the classic articulation of apoha. The unification of features produced by the vāsanās and subjective factors at the simplest level involves simply lumping together features to form stable, uniform representations. For example, generating a representation of an enduring patch of color from a continually changing flux of color particulars involves unifying all the color particulars into the same color by taking them to be functionally equivalent. Taking a spatial expanse of color, say the color of a wall, to present uniformly the same color at the same time, as well as over time, requires formulating an abstract and general representation of the color, which all the particulars are mistakenly taken to share. More complex unifications involve lumping various focus features together, both within a modality and cross-modally, to form representations of individual objects and object types. To use a common Indian example, a water-jug representation would involve associating a wide variety of cross-modal features -- color, shape, heft, texture, solidity -- that all get lumped together into the same unified object. The multiplicity of features are taken to be non-different in so far as they are taken to be functionally equivalent, generating the same effect, and thus to constitute a unified object. In addition, when identified as specifically a water-jug type object, this constructed object is taken to be functionally similar to other collections of features that were also in the past taken to be objects of that type. Such a representation thus involves generalizations on several levels, which the subjective factors and the vāsanās are responsible for generating by constructing similarities. Thus, the way that the mind is able in this way to judge different particulars as the same is by ignoring irrelevant differences and focusing on mostly functional features that seem relevant to our interests, activities, and general interactions with the world. So, when current experience is compared with the past experience encoded in vāsanās, the assessments of similarity are made by excluding or ignoring the differences between the experiences in order to focus on their functional similarity. They are in effect judged to be non-different with respect to each other and different from all those things that cannot play the same functional role. In his own commentary to verses 137-142 of the “Svārthānumāna” chapter of the Pramāṇavārttika, Dharmakīrti uses the water-jug example to explain how apoha works: Those particulars that when conglomerated perform a single effect have no distinction from each other with regard to that effect. Therefore, it would be pointless to express any such distinction. For this reason, in order to refer (niyojana) to all of them at once, people apply one expression to them, such as “water-jug.” Those [i.e., the particulars that form a water-jug] are all equally different from their respective homologues and heterologues, but since they contribute to the accomplishment of that single purpose [such as containing water], they are distinguished from others that do not do so. Hence, they are cognized as non-distinct due to that nondifference. (Dunne 2004, 356-357) Here Dharmakīrti emphasizes that the conglomeration of particulars that we take to be a water-jug is taken to be the same in the sense of being a single, whole object in virtue of the fact that we take them of perform a single effect. In virtue of their function, the differences are glossed over and that conglomeration of features is excluded from all the other features present in perception that could not play the appropriate functional role. Thus they are judged not to be a non-water-jug. This identification will involve vāsanās and subjective factors that will bring together past experiences involving practical engagement with other not non-water-jug objects. This functional understanding, plus the associated sensory experience from the past, will then inform the present judgment of sameness between the present experience and those past experiences, which then will result in a conceptual cognition that will represent the general water-jug object. This is then mistakenly confused with the original perceptual experience. Chatterjee (2011) who likens Dharmakīrti’s account to accounts common in cognitive science, describes this exclusionary process, which is the hallmark of the apoha account, as one wherein the cognitive system “learns to conflate some differences and superimpose a single form on distinct representations” (251). She explains that what happens in the cognitive process is that the cognitive system uses vāsanās to determine that different stimuli are, though different, compatible, and through the power of overgeneralization comes to react to them in the same way. In particular, it comes to generate a general representation that obscures difference and presents the stimuli as the same. To illustrate this idea she uses the example of a frog being unable to discriminate between different kinds of moving black dots and reacting to them all in the same way (251). Whether what is presented in perception is a fly or a stone pellet, the frog overgeneralizes, ignores differences, and treats them as the same. The frog’s cognitive system in effect represents these different stimuli as the same in virtue of taking them to be functionally equivalent. Chatterjee understands Dharmakīrti’s account of human cognition to operate in much the same manner. Although we receive an infinite variety of stimuli through perception, we are disposed to overgeneralize for pragmatic reasons: in order to help us negotiate our environment and predict future events. Pragmatic Context, Use, and the Framework of Things The concepts generated by the cognitive system through the apoha process are heavily influenced by the engaged context in which the process occurs. This is because concepts are used by the system in guiding behavior. For example, Dunne notes that This appeal to experience and dispositions highlights the importance of mind dependency or “subjective factors” in the process of constructing exclusions. That is, Dharmakīrti maintains that when we construct exclusions, we do not do so haphazardly or out of some pernicious habit; rather, we have some purpose in mind, and that purpose provides expectations and interests that form the context of our concept formation. (98) The purpose of concept formation in this specifically empirical context is to aid the system in negotiating and predicting the environment, which is given to the system through perception. The concept forming process thus begins and ends with engaged activity. In the middle is the apoha process that constructs concepts based on the assessment of functional similarities between features given in perception. The concepts are then used by the system to guide behavior oriented towards the environment. In Dharmakīrti’s account this orientation towards the environment is, of course, really just an orientation towards future sensory stimulation, since the environment represented in perception is only seemingly external. Regardless, concepts are thus understood primarily in terms of the functional role they play in successfully guiding our actions. To be an X is to function as one. The concept forming process is a recursive process. The effectiveness of any given constructed concept, along with the associated sensory stimuli, subjective factors, and vāsanās that contributed to its construction are stored as vāsanās that then are available for use in future iterations of the process. This means that as a system acquires more experience, it acquires more resources for assessing functional similarities between perceptual stimuli. Unlike the frog’s system, perhaps, pragmatic success or failure can make a difference to the features that our system brings into focus and how it draws functional boundaries. This success or failure, which is itself experienced through perception, contributes to the vāsanās that the system brings to bear in generating future concepts. Insofar as the cognitive process depends heavily on the engaged situation of the system, the functional similarities that a cognitive system draws in order to formulate concepts are entirely contingent. Chatterjee (2011) explains that for Dharmakīrti and cognitive scientists alike, Ontologically speaking, all these abstracted elements, be they representations of properties, of class characters, or of objects, are on a par, since they are constructions of our sensory-motor system and cannot be found in the external world. This tacit categorization, however, has not been done once and for all. Depending on the background provided by the context, the system might opt for some different schemes and cut up the world differently. Moreover, there may be many correct categorizations for the very same set of representations, correctness being a function of success in everyday life. (252) Here Chatterjee highlights the important relationship between conceptualization and pragmatic success. The manner in which we categorize the world is closely tied to the uses to which we put those categories in guiding our actions. The innate capacity we have to draw similarities and to formulate concepts does itself determine how the concepts are formulated. That depends on our experience, which is an entirely contingent matter. Conceptualization thus does not carve the world up at its natural joints, for no joints are naturally occurring, but rather constructs joints in order to facilitate goal oriented action. Dunne explains the lack of ultimate ontological ground for our conventionally constructed world of things in the following way: In effect, he [[[Dharmakīrti]]] is saying that when we can call all fires “fire,” for example, it is not that they all instantiate the universal “fireness”; nor that they all possess some real, specifiable similarity; nor even that they all have the “same” effect in a way that we can ultimately specify in objective terms. Rather, all those things are just different from nonfire things, and the reason for their difference is simply that by their nature they appear that way to us when we attend to what we mean by “fire.” Even the seeming objectivity of this appeal to nature may disappoint some, for a thing’s “nature” (svabhāva) is also conceptually constructed through the apoha theory... in ultimate terms, there is no metaphysically defensible reason for the fact that we call them “fire.” Thus, if one is hoping for an ultimately defensible metaphysical reason, then Dharmakīrti’s answer to the problem of sameness is dissatisfying. (99) On my reading, the abandonment of the search for “an ultimately defensible metaphysical reason” to explain our use of concepts is the whole point of Dharmakīrti’s account of concept formation. For an anti-realist no such reason is to be had. Instead, Dharmakīrti argues that we carve up the world into non-fires and not non-fires for purely pragmatic reasons -- because doing so serves our interests. In this respect Dharmakīrti’s view closely resembles Carnap’s. As was noted in chapter three, Carnap considers the acceptability of a framework to be a pragmatic matter, not a factual one. This pragmatic utility is tied explicitly to the use to which a language is put. If a particular form of language serves some useful purpose that is not better served by employing a different form of language, then the language is worthy of acceptance. For example, with respect to the “thing-language” Carnap says that “to accept the thing world means nothing more than to accept a certain form of language, in other words, to accept rules for forming statements and for testing, accepting, or rejecting them” (1950, 23). And insofar as the thing-language has proven very useful for the purposes of everyday life, this efficacy justifies the acceptance of that language. In giving his account, Carnap focuses exclusively on language use. Dharmakīrti on the other hand grounds his account of language use in his contingent and use-governed account of concept formation. Dharmakīrti’s account of linguistic meaning is thus derived from a more fundamental account of intentional mental content, which as we have seen is the focus of his account of concept formation. On his view, the veridicality of a conceptual cognition is parasitic on the truth of its intentional content, p, such that a cognition of p is veridical if and only if

is true, and

is true if and only if p. This intentional content of conceptual cognitions is, as has already been noted, what Dharmakīrti takes to be fit to be associated with words. Thus Dharmakīrti’s account of intentional mental content provides an account of the meaningful content of our linguistic expressions As we have seen, Dharmakīrti’s account of intentionality presents two distinct kinds of intentional mental content. According to Dharmakīrti a cognition can have one of two kinds of ākāra, or ‘representation,’ which differ with respect to the kind of intentional objects that compose them. Perceptual, or sensory, cognitions have ākāras that are composed of svalakṣaṇas. In contrast to these cognitions there are conceptual cognitions that have ākāras that are composed of sāmānyalakṣaṇas. Sāmānyalakṣaṇas, which are the conceptual products of the apoha process, insofar as they are objects that are general and repeatable, are those that are capable of being linguistically expressed. Being linguistically expressible means that such representation has a discursive structure. Transforming iconic representations into discursive ones introduces a discursive structure that, as Fodor explains, gives the representation the ability to take logical form and play a role in inference. This ability involves the representation in a structured framework, what we might call a discursive framework, that when associated with language constitutes a linguistic framework. Thus we can take Dharmakīrti’s account of concept formation as in effect providing an account of the associated discursive or linguistic framework. The framework of things in particular is presented by Carnap as an example of an empirical framework, i.e. one that includes empirical evidence as part of its system. An empirical framework, as opposed to a purely logical one, includes as part of its system a role for empirical evidence such that the “results of observations are evaluated according to certain rules as confirming or disconfirming evidence” for the existence of entities within that framework (22). Carnap explains that, To recognize something as a real thing or event means to succeed in incorporating it into the framework of things at a particular space-time position so that it fits together with the other things recognized as real, according to the rules of the framework. (22) The rules of this kind of empirical framework require the incorporation of empirical evidence in deciding for or against the reality of a proposed entity within the system. The account Dharmakīrti gives of concept formation describes the generation of discursive representations that are empirical insofar as they are derived from perception, which constitutes an empirical element that would not be required, for example, in accounting for discursive representations of mathematical or logical objects. Thus the concepts, discursive representations, and associated framework described by Dharmakīrti in giving his apoha account basically amount to Carnap’s framework of things. Like Carnap’s framework, Dharmakīrti’s framework requires the incorporation of sensory information. This is what perceptual representations of the environment provide. Of course, for Dharmakīrti this information has to be understood in the internal sense described by Price in his account of i-representation. However, it remains empirical nonetheless insofar as it plays that functional role in our cognitive processing. It is a necessary element in the construction of discursive representations of our environment and, in conjunction with the subjective factors and vāsanās, accounts for our coming to describe our world in terms of discrete and categorizable things and events. If we think of what is given in perception as being ultimately real, as Dharmakīrti does, then the resulting conceptual cognitions will be in one sense erroneous -- they do not reflect what is given in the sensory cognitions from which they are derived. Dharmakīrti emphasizes this in various places. Here I think he calls attention to the anti-realist context in which his account of the conventional is situated. What is presented in conceptual cognition is not what is given in perception, which is itself ontologically neutral. Conceptual cognitions do not reflect reality as it is in and of itself because there is no way reality is in and of itself (ontologically speaking) and this is reflected even in the contents of perception, which are also devoid of ontological significance. By calling the contents of perception “ultimately real” he thus drives home the anti-realism of his account while leaving room for the conventional reality of the objects of our ordinary discourse -- the only kind of reality one can hope for, at least ontologically speaking. In Summary This essay began by asking in what sense conventional truth claims could be understood as true from a Mahāyāna perspective, given the ontological anti-realism that characterizes those positions. In chapter three I introduced the idea that adopting a Carnapian approach to the meaning of conventional claims, instead of a fictionalist one, allows one to take the attributions of truth to those claims to be literal but also deflated. The associated deflationary theory of truth, which is metaphysically neutral, is, as Priest, Siderits, and Tillemans point out, well suited to account for truth in an anti-realist context, when supplemented with an account of meaning like the one provided by Carnap’s frameworks. Employing the Carnapian approach as a general strategy, in this chapter I have shown how the work of Dharmakīrti can be understood as providing an account of at least one framework -- the framework Carnap refers to as the framework of things. Dharmakīrti’s account, unlike Carnap’s, explains how the most basic of linguistic frameworks is formed -- how it is we come to understand the world in those conceptual terms -- by explaining how our cognitive systems imaginatively transform non-conceptual, iconic representations of our environment into discursive ones. This process occurs within a context of practical engagement and relies heavily on that pragmatic context in generating the concepts that constitute the meaning of our words. Because these concepts are functionally defined in terms of the use to which they are put in guiding action, not in terms of the nature of the referents, the theory of meaning that emerges is a use theory of meaning. This use, as Price explains, then creates an associated framework, in this case the framework of things. On this view, the use to which a concept is put fixes its extension, which for Dharmakīrti can include a wide variety of things that need not share any features in common. In fact, on this view, the use to which concepts are put creates the conventional domain of entities of the associated framework. For Dharmakīrti, as for Carnap, Price, and Horwich, conceptualization and language come before ontology. Dharmakīrti has shown how this is possible by showing how it is possible to form concepts and acquire language without presupposing an absolute domain of entities. Thus Dharmakīrti provides an anti-realist account of meaning that can, in a deflationary way, explain the sense in which we take conventional truth to be true. This account thus recommends itself to all the Mahāyānists who share Dharmakīrti’s commitment to ontological anti-realism, but who would still like to retain a normatively constrained sense in which conventional claims can be assessed as meaningful and truth-apt. CHAPTER SIX CONCLUSION In this essay I have argued for a particular solution to the problem of giving an account of conventional truth in Mahāyāna Buddhism. In laying out the problem I identified two main criteria that any adequate account of such truth should meet: the account must be robust enough to support the objectivity and normativity of truth, as well as be consistent with Mahāyāna anti-realist commitments. As was made clear in chapter two, the latter requirement poses a special challenge for the Mahāyāna that is not shared by their Abhidharma co-religionists. All Buddhists maintain that our ordinary experience of a world populated by medium sized objects and enduring events does not reflect the world’s ultimate nature. It is on these grounds that it, and the claims made describing it, are considered to be merely conventional. However, unlike the Ābhidharmikas, the anti-realist Mahāyāna deny that the ultimate nature of the world is such that it can serve as an ontological foundation for the conventional world. While the Abhidharma maintains that the conventional world can be reduced in some way to entities called dharmas, the ultimate constituents of reality, the Mahāyāna is precluded from giving such an account due to their denial of the ultimacy of any entities. According to the Mahāyāna, all entities are empty of such inherent, ultimate existence. There is no absolute domain of entities that can serve as the basis of a reduction. In order to meet the challenge of giving an account of truth from an antirealist perspective, I have here offered the work of Dharmakīrti. Dharmakīrti’s work, as I interpret it, describes how the conventional world is constructed by the mind out of an ontological nothing. Without presupposing an absolute domain of entities, on the basis of nothing more than non-conceptualized and radically particular sensory information, the mind, through a process of overgeneralization and exclusion (apoha), creates the world of our experience. This constructed, conventional world then constitutes the world -- a framework -- against which truth can be assessed. Like Carnap’s frameworks, Dharmakīrti’s conventional world provides a discursive structure and system of rules for assessing thoughts and claims pertaining to that world. However, unlike Carnap, Dharmakīrti provides an account of the generation of the conventional world that explicitly makes reference to the role that pragmatic considerations play in fixing the functional meaning of our concepts and regulating their use. This account goes a long way to dispel the concerns that frameworks or conventional worlds are arbitrarily determined and thus not normatively constrained. With this account of the conventional world in hand, the Mahāyānist is at liberty to use the deflationary account of truth to explain conventional truth. The conventional world framework fixes the meaning of our concepts (in a truth independent way) and the norms of usage built into the framework provide assertability conditions for propositions employing those concepts. The deflationary account of truth then describes our use of the concept ‘true’ with respect to those propositions. In a ontologically and metaphysically neutral way, the deflationary account simply tells us that whenever a proposition is assertable, so is the claim that that proposition is true. On this view, when we think and make assertions with respect to a particular framework, such as the framework of things, we simply use the associated concepts and attempt to conform with the appropriate norms of usage. The question of truth does not really arise. Questions of truth in effect arise from within a different framework, a framework that uses concepts to think and talk about concepts, instead of, for example, things. Because the concept truth has a use within a different kind of functionally defined framework, it need not be presupposed as operative within the context of other frameworks, like the framework of things. One implication of this account is that the Buddhist notion of two truths is reduced to just one. The deflationary account of truth is a unitary account so there is just one kind of truth -- the deflationary, conventional kind. That there would be only one kind of truth is consistent with the Mahāyāna anti-realist position, which denies that there could be objective answers to the basic question of ontology. There is no way on this view to make truth-apt assertions about the ultimate nature of reality. Such assertions would be, in Carnapian terms, assertions made external to a framework and thus meaningless and not assessable for truth or falsity. This denial of the possibility of an ultimate truth is a position that is often associated specifically with the Madhyamaka tradition of Mahāyāna Buddhism. The account presented here extends that position to apply to all of Mahāyāna Buddhism. For all these schools, as Siderits puts it, the ultimate truth is that there is no ultimate truth. The conventional truth is all there is. Since the only kind of truth available to the Mahāyāna is conventional, the conventional takes on new level of importance. It no longer need be thought of as an inferior, second-class kind of truth. It becomes the kind of truth that it is important to get right. In particular, it is important to the Buddhists, as it was perhaps for Carnap, to appreciate truth’s conventionality. In other words, it is important to dispel the mistaken idea that our truths reflect an absolute reality. Much of Mahāyāna philosophical literature aims precisely at undermining that intuitive, realist position. However, this appreciation that truth lacks ultimate ground need not lead to the acceptance of the dismal slough or the abandonment of inquiry and rational discourse. This is what I take to be value of Dharmakīrti’s work. By highlighting the important pragmatic and functional role that concepts play in our ability to reach our goals, he brings them into focus as something that anti-realists should take seriously. This is particularly important for Buddhists insofar as the Buddhist soteriological project must be understood as being itself conventional. If the Buddhist path is to be taken seriously by anti-realist Buddhists, then the conventional in general must be taken seriously. Dharmakīrti’s work, as I read it, is an attempt by an anti-realist to take the conventional seriously. Dharmakīrti recognizes that truth for the rest of us turns out to be truth for all of us, and that this truth is important, despite its conventionality, for helping us to find our way out of a world of suffering. REFERENCES Access to Insight, ed. Aṅguttaranikāya. Access to Insight, 2012. http:// www.accesstoinsight.org/tipitaka/an/index.html. Anuruddha. A Comprehensive Manual of Abhidharma: The Abhidhammattha Sangha. Edited by Mahāthera Nāranda and Bhikkhu Bodhi. Seattle: BPS Pariyatti Editions, 1999. Arnold, Dan. Buddhas, Brains, and Believing: The Problem of Intentionality in Classical Buddhist and Cognitive-Scientific Philosophy of Mind. New York: Columbia University Press, 2012. ---. Buddhists, Brahmins, and Belief: Epistemology in South Asian Philosophy of Religion. New York, Columbia University Press, 2005. Bird, Graham H. “Carnap and Quine: Internal and External Questions.” Erkenntnis 42 (1995): 41-64. Bodhi, Bhikkhu, trans. The Connected Discourses of the Buddha: A Translation of the Saṃyutta Nikāya. Edited by Bhikkhu Bodhi. Boston: Wisdom Publications, 2000. Bodhi, Bhikkhu, and U Rewata Dhamma. “Introduction.” In Abhidhamattha Sangha: A Comprehensive Manual of Abhidharma, edited by Bhikkhu Bodhi. Seattle: BPS Pariyatti Editions, 2000. Buddhaghosa. Aṅguttaranikāya Aṭṭhakathā. Edited by M. Walleser. London: Pali Text Society, 1966-1979. Buddhaghosa. Saṁyuttanikaya Aṭṭakathā (Sāratthappakāsinī). 3 vols. Edited by Pali Text Society. London: Pali Text Society, 1977. Carnap, Rudolf. “Empiricism, Semantics, and Ontology.” Revue Internationale De Philosophie 4 (1950): 20-40. Chakrabarti, Arindam, and Mark Siderits. “Introduction.” In Apoha: Buddhist Nominalism and Human Cognition, edited by Arindam Chakrabarti, Mark Siderits, and Tom Tillemans. New York: Columbia University Press, 2011. Chakrabarti, Arindam, Mark Siderits, and Tom Tillemans, eds. Apoha: Buddhist Nominalism and Human Cognition. New York: Columbia University Press, 2011. Chalmers, David. “Ontological Anti-Realism.” In Metametaphysics, edited by David Chalmers, David Manley, and Ryan Wasserman. Oxford: Oxford University Press, 2009. Chatterjee, Amita. “Funes and Categorization in an Abstraction-Free World.” In Apoha: Buddhist Nominalism and Human Cognition, edited by Arindam Chakrabarti, Mark Siderits, and Tom Tillemans. New York: Columbia University Press, 2011. Conze, Edward. Buddhist Thought in India. Ann Arbor: University of Michigan Press, 1967. Cowherds. Moonshadows: Conventional Truth in Buddhist Philosophy. Oxford: Oxford University Press, 2011. de Silva, Lily, trans. Dighanikāya Aṭṭhakathā. 4 vols. London: Pali Text Society, 1970. Dharmakīrti. The Pramana Varttikam of Acarya Dharmakīrti: With Commentaries Svopajnavrtti of the Author and Pramanavarttikavrtti of Manorathanandin. Edited by Ram Chandra Pandeya. Delhi: Motilal Banarsidass, 1989. Dreyfus, Georges. “Apoha as a Naturalized Account of Concept Formation.” In Apoha: Buddhist Nominalism and Human Cognition, edited by Arindam Chakrabarti, Mark Siderits, and Tom Tillemans. New York: Columbia University Press, 2011. ---. Recognizing Reality: Dharmakīrti’s Philosophy and Its Tibetan Interpretations. Albany: SUNY Press, 1997. Dunne, John. Foundations of Dharmakīrti’s Philosophy. Boston: Wisdom Publications, 2004. ---. “Key Features of Dharmakīrti’s Apoha Theory.” In Apoha: Buddhist Nominalism and Human Cognition, edited by Arindam Chakrabarti, Mark Siderits, and Tom Tillemans. New York: Columbia University Press, 2011. Eltschinger, Vincent. “Dharmakīrti.” Revue Internationale De Philosophie 64 (2010); 397-440. Fodor, Jerry. “The Revenge of the Given.” In Contemporary Debates in Philosophy of Mind, edited by Brian P. McLaughlin and Jonathan Cohen. Oxford: Blackwell Publishing, 2007. Frauwallner, Erich. Studies in Abhidharma Literature and the Origins of Buddhist Philosophical Systems. Translated by Sophie Francis Kidd under the supervision of Ernst Steinkellner. Edited by Wilhelm Halbass. New York: SUNY Press, 1995. Ganeri, Jonardon. “Apoha, Feature-Placing, and Sensory Content.” In Apoha: Buddhist Nominalism and Human Cognition, edited by Arindam Chakrabarti, Mark Siderits, and Tom Tillemans. New York: Columbia University Press, 2011. Garfield, Jay L. Empty Words: Buddhist Philosophy and Cross-Cultural Interpretation. Oxford: Oxford University Press, 2002. ---. “Reductionism and Fictionalism: Comments on Siderits’s Personal Identity and Buddhist Philosophy.” APA Newsletter on Asian and Asian-American Philosophers and Philosophies 6, no. 1 (2006): 1-7. ---. “Taking Conventional Truth Seriously: Authority Regarding Deceptive Reality.” In Moonshadows: Conventional Truth in Buddhist Philosophy, edited by Cowherds. Oxford: Oxford University Press, 2011. Gethin, Rupert. The Foundations of Buddhism. Oxford: Oxford University Press, 1998. Harvey, Peter. An Introduction to Buddhism: Teachings, History, and Practices. Cambridge: Cambridge University Press, 1990. Heck, Richard G. Jr. “Are There Different Kinds of Content?” In Contemporary Debates in Philosophy of Mind, edited by Brian P. McLaughlin and Jonathan Cohen. Oxford: Blackwell Publishing, 2007. Horwich, Paul. “Reference.” In Deflationary Truth, edited by Bradley P. ArmourGarb and JC Beall. Chicago: Open Court, 2005. ---. “Truth.” In Deflationary Truth, edited by Bradley P. Armour-Garb and JC Beall. Chicago: Open Court, 2005. ---. Truth. 2nd Edition. Oxford: Clarendon Press, 1998. ---. Truth, Meaning, Reality. Oxford: Oxford University Press, 2010. Jayatilleke, K. N. Early Buddhist Theory of Knowledge. Delhi: Motilal Banarsidass, 1963. Kalupahana, David J. Buddhist Philosophy: A Historical Analysis. Honolulu: University of Hawaii Press, 1976. Karunadasa, Y. The Dhamma Theory: Philosophical Cornerstone of the Abhidhamma. Wheel Publication 412/413. Kandy: Buddhist Publication Society, 1996. ---. “Theravāda Version of the Two Truths.” Paper presented at the Korean Conference of Buddhist Studies at the Haein-sa Monastery in Gyeongsangdo, South Korea, April 2006. Kataoka, Kei. “Dharmottara’s Theory of Apoha.” Paper presented at the 17th Conference of the Association for the Study of the History of Indian Thought. Kyoto, 2010. Katsura, Shoryu. “From Abhidharma to Dharmakīrti -- with Special Reference to the Concept of Svabhāva.” In Religion and Logic in Buddhist Philosophical Analysis: Proceedings of the Fourth International Dharmakīrti Conference, edited by Helmut Krasser, Horst Lasic, Eli Franco and Birgit Kellner. Vienna: Österreichische Akademie der Wissenschaften, 2011. Kimura, T. “A New Chronology of Dharmakīrti.” In Dharmakīrti’s Thought and Its Impact on Indian and Tibetan Philosophy: Proceedings of the Third International Dharmakīrti Conference, edited by Shoryu Katsura. Vienna: Österreichische Akademie der Wissenschaften, 1999. Krasser, Helmut. “How to Teach a Buddhist Monk to Refute the Heretics? Bhāviveka’s Relationship to Kumārila and Dharmakīrti.” Journal of Rare Buddhists Texts Research Department 51 (2011): 49-76. Lindtner, Christian. “Apropos Dharmakīrti: Two New Works and a Date.” Acta Orientalia 41 (1980): 27-37. Lusthaus, Dan. Buddhist Phenomenology: A Philosophical Investigation of Yogācāra Buddhism and the Ch’eng Wei-Shih Lun. London: Routledge Curzon, 2002. Matilal, Bimal Krishna. Epistemology, Logic, and Grammar in Indian Philosophical Analysis. Oxford: Oxford University Press, 2005. McClintock, Sara L. Omniscience of the Rhetoric of Reason. Boston: Wisdom Publications, 2010. McCrea, Lawrence J., and Parimal G. Patil. Buddhist Philosophy of Language in India: Jñānaśrīmitra on Exclusion. New York: Columbia University Press, 2010. Nagatomi, M. “A Study of Dharmakīrit’s Pramanavarttika: An English Translation and Annotation of the Pramanavarttika, Book I.” PhD diss., Harvard University, 1957. Newland, Guy, and Tom J. F. Tillemans. “An Introduction to Conventional Truth.” In Moonshadows: Conventional Truth in Buddhist Philosophy, edited by Cowherds. Oxford: Oxford University Press, 2011. Price, Huw. “Metaphysics After Carnap: The Ghost Who Walks?” In Metametaphysics, edited by David Chalmers, David Manley, and Ryan Wasserman. Oxford: Oxford University Press, 2009. ---. Naturalism Without Mirrors. Oxford: Oxford University Press, 2011. Priest, Graham, Mark Siderits, and Tom J. F. Tillemans. “The (Two) Truths About Truth.” In Moonshadows: Conventional Truth in Buddhist Philosophy, edited by Cowherds. Oxford: Oxford University Press, 2011. Pruden, Leo M. “The Abhidharma: The Origins, Growth, and Development of a Literary Tradition.” In Abhidharmakośabhāṣyam byVasubandhu, translated by Louis de La Vallée Poussin and Leo Pruden, vol. 1. Berkeley: Asian Humanities Press, 1988. Quine, W. V. O. Philosophy of Logic Englewood Cliffs: Prentice Hall, 1970. Rhys Davids, T. W., trans. Milindapañha: The Questions of King Milinda. Vol. 35 of Sacred Books of the East. Mineola: Dover Publications, 1963. http:// www.sacred-texts.com/bud/sbe35/index.html. Ronkin, Noa. Early Buddhist Metaphysics: The Making of a Philosophical Tradition. London: Routledge Curzon, 2005. Siderits, Mark. Buddhism as Philosophy: An Introduction. Indianapolis: Hackett Publishing, 2007. ---. “Is Everything Connected to Everything Else? What the Gopīs Know.” In Moonshadows: Conventional Truth in Buddhist Philosophy, edited by Cowherds. Oxford: Oxford University Press, 2011. ---. Personal Identity and Buddhist Philosophy. Hampshire: Ashgate, 2003. ---. “Replies to Garfield, Taber, and Arnold.” APA Newsletter on Asian and AsianAmerican Philosophers and Philosophies 6, no. 1 (2006): 16-21. Stcherbatsky, Th. Buddhist Logic. 2 vols. New Delhi: Motilal Banarsidass, 1993. Steinkellner, Ernst. “Is Dharmakīrti a Mādhyamika?” In Earliest Buddhism and Madhyamaka, edited by D. Seyfort Ruegg and Lambert Schmithausen. Leiden: E.J. Brill, 1990. Taber, John. A Hindu Critique of Buddhist Epistemology. London: Routledge Curzon, 2005. Tarski, Alfred. “The Semantic Conception of Truth.” Philosophy and Phenomenological Research 4 (1944): 341-376. Tillemans, Tom J. F. “How Far Can a Mādhyamika Buddhist Reform Conventional Truth? Dismal Relativism, Fictionalism, Easy-Easy Truth, and the Alternatives.” In Moonshadows: Conventional Truth in Buddhist Philosophy, edited by Cowherds. Oxford: Oxford University Press, 2011. ---. “How to Talk About Ineffable Things: Dignāga and Dharmakīrti on Apoha.” In Apoha: Buddhist Nominalism and Human Cognition, edited by Arindam Chakrabarti, Mark Siderits, and Tom Tillemans. New York: Columbia University Press, 2011. In Moonshadows: Conventional Truth in Buddhist Philosophy, edited by Cowherds. Oxford: Oxford University Press, 2011. Vasubandhu. Abhidharmakośabhāṣyam. Translated by Louis de La Vallée Poussin and Leo M. Pruden, vol.3. Berkeley: Asian Humanities Press, 1989. Williams, Paul. Buddhist Thought: A Complete Introduction to the Indian Tradition. London: Routledge, 2000. Wood, Thomas E. Mind Only: A Philosophical and Doctrinal Analysis of the Vijñānavāda. Delhi: Motilal Banarsidass, 1994. Wright, Crispin. “Truth: A Traditional Debate Revisited.” In Truth, edited by Simon Blackburn and Keith Simmons. Oxford: Oxford University Press, 1999.

Source

Truth For The Rest Of Us: Conventional Truth In The Work Of Dharmakīrti

Source

Navigation menu

Personal tools

Namespaces

Variants

Views

More

Search

Navigation

Categories

Tools