Eric Moses Gurevitch - The Epistemology of Difference

JOURNAL OF SOUTH ASIAN INTELLECTUAL
HISTORY (2023) 1–35

brill.com/saih
The Epistemology of Difference:

Caste and the Question of Natural Kinds
in the Courts of Medieval India
Eric Moses Gurevitch | Orcid: 0000-0002-5677-9100
National Endowment for the Humanities Postdoctoral Fellow
Department of Asian Studies, Vanderbilt University, Nashville, TN, USA
[email protected]
Received 15 June 2022 | Accepted 2 December 2022 |

Published online 6 October 2023
Abstract
Philosophy mattered in medieval India, with philosophers serving as officials in

the political landscape of the Paramāra and Cāḷukya states. And caste mattered for
these philosophers. This article explores how three Jain philosophers writing in
Sanskrit—Anantavīrya, Vādirāja and Prabhācandra—raised the question of how a
convention created by humans could ultimately be considered real. These three epis-
temologists followed the Buddhist philosopher Dharmakīrti and his commentator
Prajñākaragupta in critiquing the mimāṃsā position that took language and class cat-
egories to be eternal and unchanging. They took caste to be like language: a human
product that is upheld by convention and history, but which has an immense impact
in the world. They interrogated the possibility of cognizing a general class, and social
class served as the foundation of their critiques. Turning to medieval disputes about
the epistemology of caste provides an opportunity to revisit contemporary debates
about the persistence of caste through changing evidentiary regimes in the colonial
and post-colonial periods.
Keywords
Sanskrit philosophy – caste studies – historical epistemology – medieval philosophy
…
Published with license by Koninklijke Brill NV | doi:10.1163/25425552-12340042
© Eric Moses Gurevitch, 2023 | ISSN: 2542-5544 (print) 2542-5552from
Downloaded (online)
Brill.com 10/10/2023 06:11:09PM
via Open Access. This is an open access article distributed under the terms
This is an open access article distributed under the terms of the CC BY 4.0 license. of the CC BY 4.0 license.
https://creativecommons.org/licenses/by/4.0/
2 10.1163/25425552-12340042 | Gurevitch
Scholars hold debates about class categories such as “Brahmin-ness,”

saying that the social order of the four castes is established merely
by linguistic convention. “Is it that classes such as ‘Brahmin’ exist as
do classes such as ‘being a cow,’ or do they exist differently?” This
sort of reflection on the nature of reality is irrelevant—what is its
purpose? The order of things, which is established among people in
the world by all the learned treatises, should not—in any way—be
critically examined.
Jayanta Bhaṭṭa, The Cacophony of Scriptures1
⸪
The Question of Caste
In May 1068, a reservoir in the Cāḷukya domains needed mending. This particu-
lar reservoir had proven to be a continual source of trouble for the agricultural
community at Bāḷguḷi that depended on the water collected and distributed
from it.2 Like many similar waterworks, the “great reservoir” at Bāḷguḷi was at
risk of silting and breaches. Periodic dredging and repairs were necessary, and
these were largely funded through local revenue sources approved by the cen-
tral Cāḷukya bureaucracy. At times these funds were sourced from local tolls.
At times they were redirected from local taxes demanded by the Cāḷukya court.
And in 1068, the funds were secured through a complex transaction initiated
by one Mahādevayya, a transaction that turned caste differences into a mon-
etary vehicle. Mahādevayya purchased the rights to revenue from the leases
of sharecropping, oppressed-caste Śūdras—this is the term the inscription
uses—from local, land-owning Brahmins. He then donated this revenue to the
upkeep of the reservoir. In exchange, the town elders granted Mahādevayya a
1 brāhmaṇatvādijātau ca vivadante vipaścitaḥ |

vadantaḥ śabdamātreṇa cāturvarṇyavyavasthitim ||
gotvādijātivad iyam prakaṭānyathā vā
viprādijātir iti tattvavicāraṇāiṣā |
aprastutā kim anayākhilaśāstraloka-
siddhā sthitis tu na kathaṃcana tarkaṇīyā || Jayanta Bhaṭṭa, Āgamaḍambara 4.78–9. All trans-
lations are my own.
2 This account of the reservoir comes from the inscription given in South Indian Inscriptions,
vol. 9, pt. 1, no. 132. See ll. 22–28 for the monetary transactions. For a reconstruction of life in
the agrahāra of Bāḷguḷi from the epigraphic record see G.S. Dikshit, Local Self-Government in
Mediaeval Karnataka, (Dharwar: Karnatak University, 1964).
Downloaded from Brill.com 10/10/2023 06:11:09PM

JOURNAL OF SOUTH ASIAN INTELLECTUAL HISTORY ( 2023) 1–35
of the CC BY 4.0 license.
The Epistemology of Difference | 10.1163/25425552-12340042 3
gift of the taxes placed on the oppressed-caste barber community. The steady
stream of revenue extracted by Brahmin landowners from Śūdra sharecrop-
pers was redirected by state apparatuses and kept the water flowing in the
inland Deccan plateau. Disputes over water could quickly turn violent.3 And
caste was one of the many ways in which resources and rights were allocated
and contested in medieval India.4 Caste—to borrow a phrase—mattered.5
Philosophy also mattered in western India. Starting in the 11th century,
philosophers came to serve important roles—both administrative and
intellectual—in the chancelleries supported by the Western Cāḷukya and
Paramāra courts. (fig. 1) Their names were inscribed in stone, and they became
the topic of donative inscriptions attached to public-works projects. And caste
mattered for these philosophers. They used their privileged court positions to
author long philosophical texts, and they invariably took up the topic of the
epistemology of caste.
This article follows the careers and arguments of three Jain philosophers:
Vādirāja, who was employed in the Western Cāḷukya court of King Jayasiṃha II
(r. 1015–1042); Prabhācandra, who was employed in the Paramāra courts of
King Bhoja (r. 1010–1055) and Jayasiṃhadeva (r. 1055–1070); and Anantavīrya,
a philosopher from a generation earlier, whom Vādirāja and Prabhācandra
followed. All three expanded on the interventions of Akalaṅka (c.770), a
scholar who rewrote Jain philosophy by putting it into conversation with other
schools of thought.6 These philosophers brought technical epistemology into
new contexts. “In the royal court,” a Kannada inscription from 1147 explains,
“Vādirāja’s pastime was achieving victory over the throngs of haughty people,
and Emperor Jayasiṃha’s pastime was composing records of victory to give to
3 See for instance Epigraphia Carnatica, vol. 5, pt. 1, Hassan Taluq, no. 34, which records a mur-
der over water-rights and subsequent arbitration also during the reign of Jayasiṃha II.
4 The complex way caste fit into other regional divisions of power is detailed in Cynthia Talbot,
“A Revised View of ‘Traditional’ India: Caste, Status, and Social Mobility in Medieval Andhra,”
South Asia 15, no. 1 (1992): 17–52. For a discussion of caste-disputes in inscriptions from this
period, see J. Dunncan M. Derrett, “Two Inscriptions Concerning the Status of Kammalas and
the Application of Dharmasastra,” in Professor K.A. Nilakanta Sastri Felicitation Volume, ed.
Saw Ganesan et al. (Madras: Prof. K.A. Nilakanta Sastri Felicitation Committee, 1971), 32–55
and more broadly Burton Stein, Peasant, State, and Society in Medieval South India (Delhi:
Oxford University Press, 1980).
5 See Surinder S. Jodhka, “The Problem: Introduction to Caste Matters, A Symposium on
Inequalities, Identities and Disintegrating Hierarchies in India,” Seminar, no. 633 (2012) and
now Suraj Yengde, Caste Matters (Gurgaon: India Viking, 2019).
6 On Akalaṅka, see Piotr Balcerowicz, “Siddhasena Mahāmati and Akalaṅka Bhaṭṭa: A Revo‑
lution in Jaina Epistemology,” Journal of Indian Philosophy 44 (2016): 993–1039 and Nagin
J. Shah, Akalaṅka’s Criticism of Dharmakīrti’s Philosophy (Ahmedabad: L.D. Institute of
Indology, 1967).

JOURNAL OF SOUTH ASIAN INTELLECTUAL HISTORY ( 2023) 1–35 of the CC BY 4.0 license.
4 10.1163/25425552-12340042 | Gurevitch
Figure 1 Political geography of 11th-century South Asia

Map by author
him.”7 King Bhoja, an inscription from 1115 says, bowed down to Prabhācandra,
thus anointing his feet with the red rays of light that shown from the gems
embedded in his crown.8 As these philosophers brought Jain epistemology to
the courtly stage, references to Akalaṅka and quotations from his work began
to be used in inscriptions appended to monuments, grants and charters across
the Deccan.9
That these philosophers chose to write about the epistemology of caste was
no accident. They wrote at a time when Jains claiming to be Brahmins were on
the rise.10 And there were intellectual precedents for the turn to caste in these
7 jayisuvudĕ binadam uddhatacayamaṃ

śrīvādirājasūrigĕ sabhĕyŏḷ |
jayasiṃhacakravarttigĕ
jayapatraṃbarĕdu kuḍutaṃ iruppudĕ binadaṃ || Epigraphia Carnatica, vol. 8, Nagar
Taluq, no. 37.
8 Epigraphia Carnatica, vol. 2, no. 55.
9 Eric Moses Gurevitch, “Everyday Sciences in Southwest India” (PhD Thesis, Chicago,
University of Chicago, 2022), 123–127.
10 For an overview, see Paul Dundas, “The Digambara Jain Warrior,” in The Assembly of
Listeners, ed. Michael Carrithers and Caroline Humphrey (Cambridge: Cambridge
University Press, 1991).

philosophers. Writing in the wake of the Buddhist philosopher Dharmakīrti

(c.660), these Jain scholars struggled with—sometimes refuting and some-
times accepting—the anti-essentialist arguments of Buddhist thinkers. Most
prominently, they interpreted the work of Dharmakīrti through the writings of
Prajñākaragupta (c.800), who, as Vincent Eltschinger has argued, “more than
any of his predecessors … made an effort to integrate the specific problem of
caste into the ontological discussion on universals.”11 Philosophers writing at
the major sites of political power in Western India in the eleventh century
deployed these arguments as part of a new project. They used Buddhist argu-
ments about class categories and conceptual constructions without accepting
the full extent of Buddhist conclusions to address problems that were pressing
in both Jain intellectual culture and medieval court society more broadly.
In the course of their arguments, Anantavīrya, Vādirāja and Prabhācandra
demonstrated that any formal definition of caste would fail due to the impos-
sibility of developing criteria to define caste as an object. But this does not
mean that caste does not exist. Navigating a space between full constructiv-
ism and essentialism, the medieval philosophers this article focuses on turned
to everyday practices to show that caste, while an imprecise and ultimately
unreal category, has a powerful effect in the world.
What is caste? If caste is not something you can force into an aerosol canis-
ter and spray—and yet it is something that you can feel all around—what sort
of thing is it?12 How is what caste is related to how it is known? And when was it
turned into a problem? These questions have been central to disputes over the
interpretation of South Asian history and society in the post-colonial period.13
11 Vincent Eltschinger, Caste and Buddhist Philosophy, trans. Raynald Prévéreau (Delhi:
Motilal Banarsidass Publishers, 2012), 127. Also see Prabal Kumar Sen, “Caste in Classical
Indian Philosophy: Some Ontological Problems,” in Classical Buddhism, Neo-Buddhism
and the Question of Caste, ed. Pradeep P. Gokhale (London: Routledge, 2020), 40–64.
12 I take this formulation from Ian Hacking, Representing and Intervening (Cambridge:
Cambridge University Press, 1983), 22–38.
13 The literature comprising these 20th-century disputes is massive. See especially, Arjun
Appadurai, “Is Homo Hierarchicus?,” American Ethnologist 13, no. 4 (1986): 745–61;
Bernard S. Cohn, “The Census, Social Structure and Objectification in South Asia,” in An
Anthropologist among the Historians and Other Essays (Delhi: Oxford University Press,
1990), 233–38; Ronald Inden, Imagining India (Bloomington: Indiana University Press,
1990); Nicholas B. Dirks, Castes of Mind: Colonialism and the Making of Modern India
(Princeton, NJ: Princeton University Press, 2001); Thomas R. Metcalf, Ideologies of the Raj
(Cambridge: Cambridge University Press, 1994); Susan Bayly, Caste, Society and Politics in
India from the Eighteenth Century to the Modern Age (Cambridge, England: Cambridge
University Press, 1999). Scholarly debates about the colonial construction of caste have
been put to both emancipatory as well as revanchist uses, being used to critique the colo-
nial order as well as to deny its continued reality into the post-colonial period. See Mosse,

6 10.1163/25425552-12340042 | Gurevitch
If caste was once taken to be the defining characteristic of a timeless South

Asian culture, it has since proven to be a more slippery thing. It is something
that has been shaped by colonial forms of knowledge. It is something that is
related to—yet distinct from—other modes of identification and exclusion.
And while it sometimes is presented as a purely ancient institution, caste—in
the words of Anand Teltumbde—persists, perhaps now more than ever.14
But these questions about the nature and knowability of caste were not
raised for the first time in the post-colonial period. If caste as we know it
today is in part a product of statistical, biometric and racialized ways of cat-
egorizing people, the possibility of different ways of knowing caste was itself
an important topic of dispute for epistemologists writing in Sanskrit in the
medieval period. While contemporary scholars can look to the pre-colonial
past for evidence of what something like caste looked like before the rise of
modern state forms, there is also ample evidence that that question of caste
was itself a question that occupied people prior to the advent of structuralism
or post-structuralism.
This article broaches a second-order problem related to scholarly back-
and-forth about the nature of the caste prior to colonial ways of knowing.
While current debates often focus on whether colonial forms of knowledge
invented or appropriated earlier categorizations, this article discusses a group
of pre-colonial scholars who themselves debated what it meant for caste to be
constructed, and who debated what the implications of such a claim might
be. These philosophers took the question of what caste is to be closely linked
to the question of how caste is known. By interrogating different evidentiary
regimes for knowing caste, they brought epistemological and ontological ques-
tions together. This article turns to the past not to look for caste, but rather to
look for the question of caste.
Dharmakīrti the Clown
It is a little-acknowledged fact in the history of Sanskrit philosophy that the

Buddhist philosopher Dharmakīrti was quite funny. Radically innovative, logi-
cally rigorous, yes. Humorous, less so. But many of Dharmakīrti’s arguments
hinged on comical asides and bitter barbs regarding the proponents of other
schools of thought. Dharmakīrti’s attacks on those he disagreed with often
David. “Outside Caste? The Enclosure of Caste and Claims to Castelessness in India and
the United Kingdom.” Comparative Studies in Society and History 62, no. 1 (2020): 4–34.
14 Anand Teltumbde, The Persistence of Caste (London: Zed Books, 2010).

traded on the argument that his opponents held patently absurd positions that
led to nonsense. He was impossible for later philosophers to ignore, in part
because of his reformulation of the fundamental building-blocks of epistemol-
ogy, and in part because he was able to get under the skin of those he disagreed
with. And Dharmakīrti’s jokes provide a useful entry point to understanding
how critiques of caste were connected with arguments about the Vedas and
the Sanskrit language.
These jokes would sometimes come at the expense of Jains. The Jain doc-
trine of ontological perspectivism provided Dharmakīrti with an opportunity
to demonstrate the ethical implications of his epistemological arguments. In
one of his most heated exchanges, Dharmakīrti mocked the “shameless” Jains
for espousing a perspectival theory of reality in which things can be under-
stood in multiple manners.15 This perspectival view of reality argues that we
can ascribe seemingly contradictory characteristics to the same phenom-
enon based on the circumstances and our point of view. But if things are as
the Jains say and—in some sense—both different from as well as identical
to each other, Dharmakīrti quipped, then what would stop someone who
is told to “eat yoghurt” from chasing after a camel?16 Akalaṅka and the Jain
epistemologists who followed him took umbrage at Dharmakīrti’s character-
ization of Jain perspectival ontology, arguing that the doctrine of ontological
perspectivism does not mean that anything goes. Commenting on Akalaṅka’s
rebuttal of Dharmakīrti, Vādirāja could hardly contain his disdain, writing that
“Dharmakīrti is here speaking about Jain philosophy without understanding
it, and so he is known as a clown. Because it is established that someone who
criticizes an opponent’s position without understanding it is a clown.”17
But the post-Akalaṅka Jain epistemologists would not shy away from using
Dharmakīrti’s jokes for themselves when it came to critiquing other schools
of thought. In particular, Anantavīrya, Vādirāja and Prabhācandra would all
appropriate a joke regarding dog-flesh that Dharmakīrti made at the expense
of Brahmins who venerated the Vedas. This critique of the Veda—together
15 Dharmakīrti, Pramāṇavārttika (svārthānumāna) 1.181–184 and the Auto-Commentary

thereon. This passage is discussed in detail in Piotr Balcerowicz, “Dharmakīrti’s Criticism
of the Jaina Doctrine of Multiplexity of Reality (Anekāntavāda),” in Religion and Logic
in Buddhist Philosophical Analysis, ed. Helmut. Krasser et al., (Wien: Verlag der Öster-
reichischen Akademie der Wissenschaften, 2011).
16 Dharmakīrti, Pramāṇavārttika (svārthānumāna) 1.182.
17 tataḥ syādvādimatam anavabuddhya tatredam ucyamānaṃ dharmakīrter vidūṣakatvam
āvedayati—pūrvapakṣam avijñāya dūṣako ’pi vidūṣakaḥ | Vādirāja, Nyāyaviniścayaviva‑
raṇa, vol. 2, p. 233. This follows a comment Akalaṅka made in Siddhiviniścaya 6.25.

8 10.1163/25425552-12340042 | Gurevitch
with Dharmakīrti’s joke about dog-flesh—was used as the basis of a critique

of caste.
The joke goes something like this: The Mīmāṃsā proponent of the Vedas
claims that the Vedas are a means of valid knowledge because they do not
have a human author. Not having a human author, the Veda is a special form of
textuality of which the meaning is “self-evident.” That is, it has one true mean-
ing that it communicates on its own.18 But, Dharmakīrti argued, the meaning
of scripture is actually underdetermined. When scripture states that, “a per-
son who desires heaven should perform the agnihotra sacrifice,” there is no
way of knowing whether the meaning of those words is not “one should eat
dog-flesh.”19 The words themselves, Dharmakīrti argued, cannot tell someone
what it means to say, “one should perform the agnihotra sacrifice.” It might just
entail eating dog-flesh. To exclude that possibility, we would need a human
interpreter and a tradition that taught us the meaning of those words.20
Vādirāja—writing in the court of the Western Cāḷukya king Jayasiṃha II
(r. 1015–1042)—repurposed this joke in an extended argument against the
Mīmāṃsā understanding of textuality in his Exposition of the Determination of
Logic (Nyāyaviniścayavivaraṇa). While explaining the purpose of philosophic
writing, Vādirāja turned the Mīmāṃsā argument about the unchanging nature
of the Veda against itself, sarcastically writing that:
It is not the case that the Veda, which is liable to be changed by other
people, loudly proclaims—slapping its knee—“This is my form! This is
my meaning!” If both the Veda and its meaning were determined in and
18 The term svataḥ prāmāṇyam is usually translated as “intrinsic validity,” a translation

which has much to recommend it in reconstructing Mīmāṃsā arguments. For nuanced
discussions, see Daniel Arnold, “Of Intrinsic Validity: A Study on the Relevance of Pūrva
Mīmāṃsā,” Philosophy East and West 51, no. 1 (2001): 26–53; John Taber, “What Did
Kumārila Bhaṭṭa Mean by Svataḥ Prāmāṇya?,” Journal of the American Oriental Society 112,
no. 2 (1992): 204–21; John Taber, “Dharmakīrti, Svataḥ Prāmāṇyam, and Awakening,”
Wiener Zeitschrift Für Die Kunde Südasiens 56/57 (2015): 77–98. The philosophers this arti-
cle discusses were less-than sympathetic to Mīmāṃsā arguments, and hence I translate it
as “self-evident” to preserve their critiques, which were at times less-than generous.
19 tenāgnihotraṃ juhuyāt svargakāma iti śrutau |
khādec chvamāṃsam ity eṣa nārtha ity atra kā pramā || Dharmakīrti, Pramāṇavārttika
(svārthānumāna) 1.318.
20 The argument is discussed in depth in Kei Kataoka, “Transmission of Scripture:
Exegetical Problems for Kumārila and Dharmakīrti,” in Scriptural Authority, Reason and
Action, ed. Vincent Eltschinger and Helmut. Krasser (Wien: Verlag der Österreichischen
Akademie der Wissenschaften, 2013), 319–74 and Vincent Eltschinger, Helmut Krasser,
and John Taber, Can the Veda Speak? (Wien: Verlag der Österreichischen Akademie der
Wissenschaften, 2012).

by themselves, then all the explanations of the Veda by Mīmāṃsā phi-

losophers as well as grammar would be useless.21
Vādirāja’s goal was to trap the Mīmāṃsā philosopher into admitting that
either the Veda is not eternal—since, if it were, it would be impossible for
a human being to distort it—or admitting that the entire Mīmāṃsā herme-
neutic project is pointless. The Mīmāṃsā philosopher Kumārila (c.660) was
wrong, Vādirāja argued, to say that “if there were any distortion of the Veda, it
would be removed by many people.”22 Such a statement failed on two counts.
First, it was flawed from an ontological viewpoint, since it implies that the
Veda could possibly be changed, since humans are only capable of changing
things that are impermanent. But Kumārila’s statement was also flawed from
an epistemological viewpoint. There is no way of assessing different claims to
unbroken traditions. After all, even the traditions of barbarians (mlecchādi)
could be considered authoritative on such lines, since they are accepted by
many people as being eternal.23
Vādirāja proceeded to bring the attack to the heart of the Mīmāṃsā proj-
ect by arguing that the Mīmāṃsā claim to an eternal meaning of the Veda is
disproven by the fact that different Mīmāṃsā philosophers disagree on the
meanings of Vedic passages. “Śabara considers the black magic of the Hawk
Sacrifice to be harmful, while Umbeka considers to be otherwise,” Vādirāja
argued, “How could that not be a contradiction in the meaning of the Veda?”24
Listing several disagreements among Mīmāṃsā philosophers, Vādirāja wrote
that “There are yet many other examples of differences in commentarial tradi-
tions. So how is it that one could know that a particular Vedic commentary
refers to true things, and not another?”25 The point was to bring into question
21 na hīdam eva me rūpam ayam evārtha ity api |

jānughātaṃ vadan vedaḥ śakyapracchādanaḥ paraiḥ ||
tat svato niścite vede vedārthe ca tadarthakam |
yad vyākaraṇamīmāṃsādy etat sarvam anarthakam || Vādirāja, Nyāyaviniścayavivaraṇa,
vol. 1, p. 32, vv. 137–8.
22 ajānan vedasāmarthyaṃ bhaṭṭas tad idam abravīt |
“anyathākaraṇe cāsya bahubhyaḥ syān nivāraṇam” || Vādirāja, Nyāyaviniścayavivaraṇa,
vol. 1, p. 29, v. 109; Quotation from Kumārila, Mīmāṃsāślokavārtikam 1.1.2.150.
23 Vādirāja, Nyāyaviniścayavivaraṇa, vol. 1, p. 30.
24 anarthetararūpatvaṃ śavarombekasammatam |
śyenasya yat sa vedārtho viruddho ’pi bhaven na kim?| Vādirāja, Nyāyaviniścayavivaraṇa,
vol. 1, p. 32, v. 129.
25 evamādiparo ’py asti tadvyākhyābhedavistaraḥ |
tatra na jñāyate kiṃ tad vyākhyānaṃ vastugocaram? || Vādirāja, Nyāyaviniścayavivaraṇa,
vol. 1, p. 31, v. 117.

10 10.1163/25425552-12340042 | Gurevitch
the Mīmāṃsā claims to exclusive interpretation of the Vedas by adducing the

diversity of the Mīmāṃsā intellectual tradition as evidence against itself. If
even Mīmāṃsakas disagree about the meaning of the Veda, there cannot be
any criteria to establish a true interpretation from a false one. “How is it,”
Vādirāja brought his argument to a head with a rhetorical flourish aimed as a
Mīmāṃsā philosopher, “that the eating of dog-flesh is not a possible meaning
of the Veda for you?”26
Vādirāja argued that the fact of there being multiple interpretations of the
Veda within the Mīmāṃsā school undermines the claims to hermeneutic cer-
titude made in Mīmāṃsā, namely that the Veda has a definite meaning that is
unchanging and that it can serve a means of valid knowledge as instantiated in
its interpretive tradition. And this set of arguments about tradition were used
in critiques of caste initiated by Anantavīrya, Vādirāja and Prabhācandra.
Between Tradition and Identity
The critiques written by Vādirāja and Prabhācandra expanded on arguments

made by Anantavīrya in his Commentary on the Determination on Accom-
plishment (Siddhiviniścayaṭīkā). And Anantavīrya linked questions about the
authority of the Vedic tradition with epistemological questions regarding
the cognition of caste.
Anantavīrya began with a familiar example, writing that, “the meaning of the
Veda is multiple. Otherwise, it would be possible that its single meaning could
be ‘Heaven can be obtained on account of the consumption of dog-flesh.’”27 To
prevent such a possible meaning for the Veda, his argument continued, the sin-
gular meaning of the Veda must be determined by an authoritative person. But
who might this authoritative person be? Anantavīrya’s interlocutor answered
that would be a Brahmin. But Anantavīrya argued that that Veda cannot itself
reveal who is a Brahmin—such a position would lead to a mutual interdepen-
dence in the argument—and so, just like the meaning of the Veda, who is a
Brahmin remained an epistemically under-determined question.28
It was at this point that Anantavīrya operationalized Dharmakīrti’s point
about dog-flesh in a novel manner, using it to address a set of questions about
26 śvamāṃsabhakṣaṇaṃ tasya vedārthatvaṃ kathaṃ na vaḥ? || Vādirāja, Nyāyaviniścayavi‑

varaṇa, vol. 1, p. 32, v. 131cd.
27 vedasya anekārtho ’rthatattvam, anyathā “sārameyamāṃsabhakṣaṇāt svargaḥ syāt” ekaḥ
tasyārtha iti | Anantavīrya, Siddhiviniścayaṭīkā, 513.
28 Anantavīrya, Siddhiviniścayaṭīkā, 514.

religious conversion and changing identities. He argued—in a sense taking

Dharmakīrti’s joke seriously—that “dog-flesh is consumed by a Brahmin who
has converted to Buddhism.”29 The situation of conversion presents several
problems relating to identity, and Anantavīrya exploited the epistemic ambi-
guities of shifting identities to make a set of arguments against the permanence
of class categories. If it takes an authority to explicate the Veda, and an author-
ity is defined as “someone who is endowed with Brahmin-ness,” then the
question remains whether a Brahmin who has converted to Buddhism remains
endowed with Brahmin-ness. If Brahmin-ness is—as the Mīmāṃsa philoso-
pher would have it—a permanent class category, then even if someone born a
Brahmin converted to Buddhism, they would still have authority in interpret-
ing the Veda. And this Buddhist convert, now being a strict Dharmakīrtian,
might say that the Veda told people to sacrifice with dog-flesh.
The question becomes then, to borrow a phrase, a question of what makes
people who they are.30 It becomes a question of identity—and self-identity—
over time. As Anantavīrya asked “Is it the case that such a person is a Brahmin
prior to converting to Buddhism? Indeed, it is not the case that a white horse,
which is devoid of a connection to the class category of cow-ness, can ever
become a cow.”31 If a white horse can never become a cow but a Brahmin can
become a Buddhist, then there must be a difference in the sorts of categories
these are. For Anantavīrya, the question of Brahmin-ness was closely linked to
the question of establishing a criterion to cognize Brahmin-ness. He used this
line of questioning to back his Mīmāṃsaka opponent into a corner:
Mīmāṃsā interlocutor: A man might be a Brahmin. But when he has con-

verted to Buddhism, he is opposed to being a Brahmin.
Author: Objection! How is it that Brahmin-ness is destroyed? If it is in that

manner, then a class category is not eternal, and a cow-without-horns
standing among white horses would not be considered to be a cow.
Interlocutor: Then maybe it is the case that Brahmin-ness is not destroyed.
29 saugatamatānusāriṇā dvijena śvapiśitāśanam | Anantavīrya, Siddhiviniścayaṭīkā, 514.

30 Rosalind O’Hanlon and Christopher Minkowski, “What Makes People Who They Are?
Pandit Networks and the Problem of Livelihoods in Early Modern Western India,” The
Indian Economic & Social History Review 45, no. 3 (September 1, 2008): 381–416,
31 kathaṃ tadanusaraṇāt pūrvaṃ sa brāhmaṇaḥ? na hi gotvasambandarahitaḥ karkaḥ
kadācanāpi gaur iti | Anantavīrya, Siddhiviniścayaṭīkā, 514.

12 10.1163/25425552-12340042 | Gurevitch
Author: In that case, even someone who has become a Buddhist is actu-
ally still a Brahmin.
Interlocutor: He is not, because he is not devoted to the Veda.
Author: Is it the case that Brahmin-ness simply is devotion to the Veda,

or is it something else? On the first view—namely, that Brahmin-ness
is devotion to the Veda—Brahmin-ness is not a class category, because
devotion is a particular state of mind. And thus, if someone killed a
Brahmin who was sleeping, they would not be committing the crime
of Brahminicide. On the second position—namely, that Brahmin-ness
is something else—a person would possess the condition of being a
Brahmin even when he didn’t have devotion for the Veda. And that is
a logical fault in the argument.
Interlocutor: Then a Brahmin is known by devotion to the Veda.
Author: That is not the case. Because devotion to the Vedas exists even
in Śūdras.32
Anantavīrya here argued that it cannot be a particular behavior that determines

a person’s caste. First of all, being a Brahmin cannot be a particular mental
disposition, because people’s minds change over time, and in everyday life we
act as if a sleeping Brahmin is in fact a Brahmin. Second, Anantavīrya made an
argument about behavior defining the class of Brahmin-ness. He described a
social world in which devotion to the Vedas is not limited to Brahmins. Even
Śūdras can be seen to express devotion towards the Vedas. Anantavīrya went
on to conclude the argument by saying that if family lineage were asserted to
be the basis of caste categorizations, that too would fail in a similar manner,
leading to a mutual inter-dependence in the argument. Ultimately, appeals to
lineage beg the question.
For Prabhācandra, the question of family lineage—that is, the question
of whether caste is cognizable through genealogy—was the central problem
32 tadāpy asau brāhmaṇaḥ tadanupraveśe tu viparītaḥ iti cet | nanu kiṃ brāhmaṇatvaṃ
naṣṭam? tathā cet; na nityaṃ sāmānyam iti karkamadhyam adhyāsīno muṇḍo na gauḥ
syāt | atha na naṣṭam; tadanupraveśe ’pi brāhmaṇa eva | vede bhaktyabhāvān neti cet | kiṃ
tadbhaktir eva brāhmaṇyam, anyad vā? prathamapakṣe na tat sāmānyam jñānaviśeṣatvāt
tasyāḥ | tathā ca suptādyavasthasya vyāpādane ’pi na brahmahatyādidoṣaḥ | dvitīyapakṣe
tadabhāve ’pi tadavasthaṃ tad iti sa eva doṣaḥ | atha tadbhaktyā sa eva jñāpyate | na; śudre
’pi tadbhāvāt | Anantavīrya, Siddhiviniścayaṭīkā, 515.

for the epistemology of caste. This terrain had earlier been delimited by the
Mīmāṃsā philosopher Kumārila.33 Kumārila rejected definitions of caste based
on physical appearance or behavior, but still argued that caste is cognizable
as a class category that is inherent in individuals. A class category—Kumārila
argued—can be recognized on account of a person’s lineage. He wrote:
Since people such as Brahmins are perceived in everyday observations to

be undifferentiated in terms of their appearances, which are similar
to people belonging to other castes in terms of their heads, their hands,
etc., there is a determination of caste differences only on account of a
recognition of the lineage of one’s mother and father. In a similar man-
ner, there is a cognition of the grammaticality of words. Thus, there is no
contradiction with everyday life.34
Prabhācandra took it as his role—once again following the Buddhist phi-

losopher Prajñākaragupta—to refute this final criterion for the cognition of
caste. And in order to do so, he used Akalaṅka’s scholarship to challenge both
Buddhist and Mīmāṃsā notions of class categories. This would have implica-
tions for the understanding of language as well as social class, each upheld in
his account by convention.
From Generic Class to Social Class
At several points in his Verse Commentary (Ślokavārttikam) on Mīmāṃsā,

Kumārila stated that he considered four Sanskrit words to be synonymous in
denoting class categories or universals: sāmānyam, ākṛti, jāti, and śakti.35 That
the word “jāti” is also commonly used to denote social class was not missed by
philosophers writing in Sanskrit. And much of the eleventh-century debates
about the epistemology of caste focused on defining class categories and how
they are perceived.
33 See the discussion on genealogy being the sole criterion of cognizing caste in Wilhelm.
Halbfass, Tradition and Reflection: Explorations in Indian Thought (Albany, NY: State
University of New York Press, 1991), 368–71.
34 yathaiva tulyaśiraḥpāṇyādyākāreṣv api saṃkīrṇalokadṛṣṭigrāhyeṣu brāhmaṇādiṣu mātā‑
pitṛsambandhasmaraṇād eva varṇavivekāvadhāraṇaṃ bhavati, tathā sādhuśabdāvadh‑
āraṇam apīti lokavirodhābhaḥ | Kumārila, Tantravārttikam on Mīmāṃsāsūtram 1.3.25, on
p. 229.
35 Kumārila, Ślokavārttika, (Ākṛtivāda) 13.18.

14 10.1163/25425552-12340042 | Gurevitch
Akalaṅka cut a path between the Scylla that is the Mīmāṃsā and Nyāya
notion of a class as a universal that pervades particular instances and the
Charybdis that is the Buddhist position that a class is a mental construct estab-
lished through the process of exclusion.36 For Akalaṅka and his followers, a
class category is not something that can be known through sampling individ-
ual manifestations, nor is it something that can be known through exclusion
of everything that is not included in the category. They denied the universal
nature of purported universals, arguing that a class category is not a singular,
pervasive and unchanging entity. And yet, they argued, the position developed
by the Buddhist philosophers Dignāga, Dharmakīrti and their followers over-
stated the case when they argued that a class category is nothing more than a
mental construction founded on exclusion.
In his summary of Akalaṅka’s philosophy, the Jain philosopher Māṇikya‑
nandin (c.950) built on Akalaṅka’s definition of class categories as similar
permutations and further divided them into “vertical” and “horizontal”—or
“diachronic” and “synchronic”—classes. Māṇikyanandin wrote that synchronic
classes are defined by being similar permutations, while diachronic classes are
a way of describing the substance that remains pervasive when something
undergoes successive changes.37 These distinctions would be utilized by both
Prabhācandra and Vādirāja in the eleventh century, and would also be repur-
posed to explicate the theory of exclusion by the Buddhist scholar Ratnakīrti
(c.1070), who was working at the Buddhist center of learning Vikramaśīla in
Bihar.38 These new tools for analyzing general classes provided epistemologists
with ways out of some of the problems that had plagued earlier theorists of
caste. But they would also present their own difficulties.
Prabhācandra’s critique of caste comes in both his Sun for the Blossoming
of the Lotuses of Realia (Prameyakamalamārtaṇḍa) and in his Moon for the
Blossoming of the Lily of Logic (Nyāyakumudacandra). Much of the argument
from the earlier Sun was repurposed in the Moon. And the argument—often
word-for-word—was quickly picked up by Vādideva, who wrote his Ocean
36 For a summary of Akalaṅka’s position, see Shah, Akalaṅka’s Criticism of Dharmakīrti’s

Philosophy, 76–154.
37 Māṇikyanandin, Parīkṣāmukha 4.1–6.
38 For a general discussion, see Piotr Balcerowicz, “How Could a Cow Be Both Synchronically
and Diachronically Homogenous, or On the Jaina Notions of Tiryak-Sāmānya and
Ūrdhvatā-Sāmānya,” in Approaches to Jaina Studies, ed. N.K. Wagle and Olle Qvarnström,
(Toronto: University of Toronto: Center for South Asian Studies, 1999), 211–37. For
an analysis of Ratnakīrti’s use of these categories, see Parimal G. Patil, “On What It Is
That Buddhists Think About: Apoha in the Ratnakīrti-Nibandhâvali,” Journal of Indian
Philosophy 31, no. 1 (2003): 229–56.

of Jain Philosophy (Syādvādaratnākara) in the Caulukya court of Jayasiṃha

Siddharāja (r. 1092–1142) at Aṇahillapura.39 Prabhācanda’s argument hinged
on the difficulty of establishing the conditions for a logical inference with the
terms agreed upon by philosophers writing in Sanskrit.40 For him, the ques-
tion of caste presented novel opportunities for these more general philosophic
problems.
Prabhācandra questioned the difficulty of establishing a logical relationship
to define social class that will hold in all instances. Arguing against the possi-
bility of a particular behavior serving as an epistemic criterion to determine a
class category like Brahmin-ness, he wrote that:
There is not a particular type of conduct by which one could perceive

Brahmin-ness. For this would have to be a particular mode of conduct
exclusive to Brahmins, such as conducting sacrifices, teaching, receiving
gifts or the like. But a particular mode of conduct cannot become a cri-
terion for the direct-perceptibility of Brahmin-ness—because both the
logical faults of under-determination and over-extension would follow.
There is a logical under-determination, because of the fallacy that there
is no everyday application of that criterion among those Brahmins who
do not do things such as sacrifice. And there is a logical over-extension,
because it erroneously follows that there would be Brahmin-ness even
among Śūdras, because all those behaviors, such as sponsoring sacrifices,
etc., are observed among even Śūdras.41
The point here is that direct perception prevents the foundation of the inference
of social class based on particular modes of conduct. The fallacy is twofold: On
the one hand, we can find people who would normally be defined as Brahmins
who do not engage in the particular modes of behavior supposedly associated
39 See the passages in Vādideva, Syādvādaratnākara, 958–962.

40 For discussions of Prabhācandra’s general critiques of inference, see Piotr Balcerowicz,
“Is ‘Inexplicability Otherwise’ (Anyathânupapatti) Otherwise Inexplicable?,” Journal of
Indian Philosophy 31, no. 1/3 (2003): 343–80 and Marie-Hélène Gorisse, “Is Inference a
Cognitive or a Linguistic Process? A Line of Divergence between Jain and Buddhist
Classifications,” in Puṣpikā: Tracing Ancient India, through Texts and Traditions, ed. Robert
Leach and Jessie Pons (Oxford: Oxbow Books, 2015), 1–23.
41 nāpi ācāraviśeṣaḥ | sa hi brāhmaṇasyāsādhāraṇo yājanādhyāpanapratigrahagrahādiḥ | sa
ca tatpratyakṣatānimittaṃ na bhavati avyāpter ativyāpteś cānuṣaṅgāt | yājanādirahiteṣu
hi brāhmaṇeṣv api tad-vyavahārābhāvaprasaṅgād avyāptiḥ | śūdreṣv api akhilasya
yājanādyācārasyopalabdhito brāhmaṇyānuṣaṅgāc cātivyāptiḥ | Prabhācandra, Nyāyaku‑
mudacandra, 773–4. See the parallels to this argument in Prajñākaragupta’s Pramāna‑
vārttikālaṅkāra, 23, and in Prabhācandra, Prameyakamalamārtaṇḍa, 485.

16 10.1163/25425552-12340042 | Gurevitch
with them. On the other hand, we also can find people who would normally
be defined as non-Brahmins who engage in the modes of conduct supposedly
associated with Brahmins.42 And hence, any inference that would be founded
on such criteria would be both under-determined as well as over-extended.
“Indeed,” Prabhācandra wrote, “even a Śūdra can travel to a foreign-land and
either study or teach the Veda.”43 But such a person would not be accepted by
the Mīmāṃsaka interlocutor as a Brahmin. Modes of conduct turn out to be
poor criteria for establishing a social class, since everyday life is much more
complicated than such a schema would allow.
Prabhācandra argued that the broad use of the term “Brahmin” in every-
day language should be taken as a warning about the nature—or rather, the
non-existence—of abstracted social class categories. Once he had shown
that direct perception and inference cannot be used to cognize a social class,
Prabhācandra considered the argument that “there is a teaching regarding
the three dominant castes that is not disputed by scripture, and when that
is cognized, it is a means of valid knowledge for really existent things.”44
Prabhācandra did not argue against the possibility of learning from teachers.
But a teaching should correspond to natural kinds, and as such cannot be con-
tradicted by direct perception. And the sorts of teachings that dominant-caste
scholars invoked in defining Brahmin-ness cannot be shown to cover all cat-
egories. We cannot rely on such people to serve as authorities, since, “we can
observe many people who are conventionally designated by those who talk of
the three dominant castes with the term ‘Brahmin’ without any dispute, and
yet who live in a contrary manner.”45
Prabhācandra argued that the conventional use of language can lead people
astray by introducing categories that do not exist as natural kinds. But phi-
losophy must also be able to account for ordinary language. Prabhācandra
concluded his discussions of social class in both the Sun and the Moon with
reflections on the everyday usage of terms such as “Brahmin.” Philosophy
cannot deny the fact that everyday language is used by people to talk about
Brahmins as people who act in certain ways. But Brahmin-ness is not a philo-
sophically rigorous category, and we should not depend on it to extrapolate a
42 Six modes of conduct (ṣaṭkarman) were associated with being a Brahmin in the classic
Sanskrit legal texts. The locus classicus is Manusmṛti 10.75.
43 śūdro ’pi hi kaścid deśāntaraṃ gatvā vedaṃ paṭhati pāṭhayati vā | na tāvatāsya brāhma‑
ṇatvaṃ bhavadbhir abhyupagamyata iti | Prabhācandra, Prameyakamalamārtaṇḍa, 487.
44 etenāvigānatas traivarṇikopadeśo ’tra vastuni pramāṇam | Prabhācandra, Prameyakama‑
lamārtaṇḍa, 486.
45 dṛśyante hi bahavas traivarṇikair avigānena brāhmaṇatvena vyavahriyamāṇā viparyaya‑
bhājaḥ | Prabhācandra, Prameyakamalamārtaṇḍa, 486.

coherent class category. Philosophy can be used to probe the positive and neg-
ative concomitance of certain predicates. And through such means, it can be
used to show that everyday appeals to caste are ultimately unreliable. So, while
everyday language allows for people to talk of “Brahmins,” rigorous analysis
should prevent people from talking of “Brahmin-ness” as a natural kind. This
conclusion came in a complex exchange with the hostile interlocutor:
Interlocutor: Objection! If you do not accept the class category of some-

thing like Brahmin-ness, then how could there be the social order of
castes and life-stages? Likewise, how could there be the everyday prac-
tices of performing austerities and giving gifts, which are founded on that
social order?
Author: That too is not a proper objection. Because the social order of
the castes and life-stages as well as the everyday practice of performing
austerities and giving gifts is established with regards to a particular indi-
vidual who is characterized by marks such as particular actions and being
endowed with the sacred thread. But Brahmin-ness, which has been con-
structed by you—my interlocutor—as an essence that is eternal, etc. is
not established from any means of valid knowledge. And so, the every-
day talk about Brahmins, etc., which is exclusively founded on particular
activities, is proper.46
After arguing that the social class of Brahmin-ness cannot be founded on a par-
ticular mode of conduct, Prabhācandra was forced to clarify his position. His
denial of the reality of Brahmin-ness as a class category should not be taken
as a denial of social distinctions. Brahmins exist. Caste can have real effects in
the world even if its ontological status is unreal. Prabhācandra did not deny
that people in the world go around talking about Brahmins, and when pressed,
define them through a set of activities. To deny these realties is not something
that philosophy can do. But philosophy can show that such usage is not rig-
orous, and that it does not refer to some eternal nature found in Brahmins.
Philosophy can show that doubts will inevitably arise with the everyday talk
46 nanu brāhmaṇatvādisāmānyānubhyupagame kathaṃ bhavatāṃ varṇāśramavyavasthā

tannibandhano vā tapodānādivyavahāraḥ syāt? ity apy acodyam | kriyāviśeṣayajñopav
ītādicihnopalakṣite vyaktiviśeṣe tadvyavasthāyāḥ tadvyavahārasya ca upapatteḥ | tan na
bhavatkalpitaṃ nityādisvabhāvaṃ brāhmaṇyaṃ kutaścid api pramāṇāt prasiddhyatīti
kriyāviśeṣanibandhana evāyaṃ brāhmaṇādivyavahāro yuktaḥ | Prabhācandra, Nyāyaku‑
mudacandra, 778–9. Compare with Prameyakamalamārtaṇḍa, 486.

18 10.1163/25425552-12340042 | Gurevitch
about social class. And philosophy can be used to reassess prevalent regimes
of evidence.
What Makes a Brahmin Different from a Cow?
Brahmin-ness was used in this argument to solve a problem that arose in the
wake of Akalaṅka. If a class cannot be known through sampling—that is, if
the repeated observation “this is a Brahmin, and this is a Brahmin” is insuf-
ficient to establish a class—it becomes difficult to determine what actually
that holds a class together. And if they rejected the Buddhist notion of exclu-
sion and the Mīmāṃsā notion of pervasion as definitional of a class category,
it remained a problem for Prabhācandra and other post-Akalaṅka thinkers
to demonstrate how similarity can be determined. Their definition of class
categories hinged on the notion of “similar permutations.” Similarity is not
quite identity, and it cannot be expressed as a shared quality that is found in
each individual. As such, it risks becoming a sort of black-box, something that
cannot be investigated any further. And there is the possibility of an infinite
regress, because one might be compelled to ask whether there is a similarity for
all similarities and how that could possibly be cognized. Prabhācandra denied
the basis of this question and argued that not all similarities are alike. It was at
this moment that the question of caste became epistemologically important,
and Prabhācandra used it to distance himself from the earlier Buddhist phi-
losophers he followed for much of his arguments.
If the Buddhist in Prabhācandra’s account opened by saying that all class
categories are unreal—like sky-lotuses or the horn of a donkey—Prabhācandra
went on to argue that class categories are really-existing things, albeit not eter-
nal essences distinct from particular instances. And yet some class categories
that are commonly taken to be real are not. The Buddhist argument that any-
thing expressible in language is a conceptual construction is too broad, and it
fails to distinguish between different sorts of classes for things in the world.
Brahmin-ness is one of those class categories that is as unreal as sky-lotuses.
But not all class categories are unreal. When asking questions about concep-
tual construction, it is most important to ask the question of what exactly is
being described as constructed. Asking “the construction of what?” will make
all the difference.47 Not all conceptual constructions are alike. Caste, follow-
ing Prabhācandra’s general discussion of classes, came to serve as a test-case
47 I take this phrasing from Ian Hacking, The Social Construction of What? (Cambridge, Mass:
Harvard University Press, 1999).

to probe the limits of the notion of similarity. The question of caste became a
question of identity in the literal sense—it became a question of what counts
as “the same” across a set of individuals who are irreducibly diverse.
In his discussion of the question of caste, the Buddhist philosopher
Dharmakīrti made a provisional distinction between natural and linguistic
differences that he later would nuance in making his larger arguments about
conventional and ultimate reality.48 In posing the question of the efficacy of
mantras, Dharmakīrti argued that “there is not a natural difference between
people, who may be given the titles ‘Śūdra’ or ‘Brahmin.’ Indeed, real objects,
which adhere to natural distinctions, do not become different because of
a nominal distinction of ordinary language, which depends on people’s
intentions.”49 On the other hand, a natural distinction, Dharmakīrti would go
on to argue, “can be observed when there is a difference in terms of form, quali-
ties, or capacity, as in the case of the difference between cows and horses. And
ordinary people cognize it without the assistance of a teaching.”50
It would seem here—and Dharmakīrti would elsewhere problematize this
distinction—that there are some sorts of class categories that are founded
on linguistic conventions and some that arise naturally in human cognition.
The necessity of instruction to convey a class category would, in this view,
serve as a criterion by which to distinguish linguistic from natural divisions.
Ultimately, for Dharmakīrti, all such class distinctions should be consid-
ered unreal, and the question of how to distinguish between different sorts
of non-existing entities would become a heated topic of debate among later
Buddhist philosophers—most prominently Prajñākaragupta.51 This discus-
sion was picked up by epistemologists with Jain commitments, who built
upon Buddhist arguments to deny many sorts of classes without accepting the
Buddhist conclusion that all classes are ultimately unreal.
48 See the recent debate over the “hierarchy of discourses” found in Dharmakīrti’s writing in
Dan Arnold, “Buddhist Idealism, Epistemic and Otherwise: Thoughts on the Alternating
Perspectives of Dharmakīrti,” Sophia 47, no. 1 (2008): 3–28.
49 śūdraviprābhidhānayoḥ puruṣayoḥ svabhāvābhedāt | na hi puruṣecchānuvidhāyino nāmav‑
yavahārabhedāt svabhāvabhedānubandhinām arthānām anyathātvam asti | Dharmakīrti,
Auto-commentary on Pramāṇavārttikam 1.295. See the summary of the full argument in
Eltschinger, Caste and Buddhist Philosophy, 103–107.
50 sa khalv ākṛtiguṇaśaktibhede dṛṣṭo gāvāśvavat | anupadeśaṃ cainaṃ lokaḥ pratipadyate |
Dharmakīrti, Auto-commentary on Pramāṇavārttikam 1.295.
51 On this, see Eli Franco, “How to Distinguish between Non-Existing Entities? Dharmakīrti
and Prajñākaragupta on Universals as Objects of Knowledge,” Rocznik Orientalistyczny 65,
no. 1 (2012): 51–62 and Lawrence McCrea, “Prajñākaragupta on the Pramāṇas and Their
Objects,” in Religion and Logic in Buddhist Philosophical Analysis, ed. Helmut. Krasser
et al. (Wien: Verlag der Österreichischen Akademie der Wissenschaften, 2011), 319–28.

20 10.1163/25425552-12340042 | Gurevitch
For Prabhācandra, ordinary language can be deceptive in distinguishing

objects in the world. But human conceptions—when properly analyzed—offer
the possibility of presenting the reality of the world as it really is. This is because
some conceptions are produced directly from things in the world, while other
sorts of conceptions will come about through a teaching, which is necessarily
linguistic and conventional. This second sort of conception, not being pro-
duced directly by an actually-existing object, is the type of conception that
can be dubious. Prabhācandra used the example of social class to investigate
how exactly we can distinguish between these different types of categories,
one real, the other not. As he argued:
It is certainly not the case that the social class of Brahmin-ness can be
perceived through direct perception in an ordinary person—call him
“Devadatta”—in the manner in which cow-ness is defined by a similar-
ity in things such as cows-with-broken-horns and cows-without-horns.
Otherwise, there would be no such doubt as “is this person a Brahmin,
or are they something different?” And so, the teaching of things such as
family lineage is useless for the removal of that doubt. Indeed, the deter-
mination as to whether “is this a cow or is this a person?” is not dependent
on the teaching of things such as family lineages.52
Prabhācandra argued that the ability to make distinctions between certain

types of natural kinds—here something like biological species—is inherent
in people. The doubt “is this a person or a cow?” is of a different form from
the doubt “is this a Brahmin or a Śūdra?” The cognition of a natural kind is
not dependent on instruction but is rather caused directly by the objects in
question. In particular, it is caused by the actually-existing similarity that
they possess. On the other hand, caste distinctions are entirely founded
upon instruction, and hence are not caused by an actually-existing similar-
ity between people in the same purported caste. Prabhācandra—who would
vehemently argue elsewhere against the possibility of women achieving
liberation—argued that caste is unlike both species and sex, which he took to
be natural categories.53 As he put it:
52 na khalu khaṇḍamuṇḍādiṣu sādṛśyalakṣaṇagotvavad devadattādau brāhmaṇyajātiḥ prat‑

yakṣataḥ pratīyate | anyathā “kim ayaṃ bṛāhmaṇo ’nyo vā?” iti saṃśayo na syāt | tathā
ca tannirāsāya gotrādyupadeśo vyarthaḥ | na hi “gaur ayaṃ manuṣyo vā?” iti niścayo
gotrādyupadeśam apekṣate | Prabhācandra, Prameyakamalamārtaṇḍa, 484.
53 For a translation of the essay on women’s liberation from Prabhācandra’s Nyāya‑
kumudacandra, see Padmanabh S. Jaini, Gender and Salvation (Berkeley: University of
California Press, 1991), 109–138.

There is no manifestation of a unitary Brahmin-ness that recurs in all

of its particular instances, in addition to the class categories such as
human-ness or masculinity within the individual instances of humans
who are visible before the eyes; as is the case with the class categories of
cow or horse in the individual instances of cows-with-broken-horns or
cows-without-horns or white horses that are visible before one’s eyes,
or as is the case with a quality such as whiteness.54
Unlike a class category such as humanness or a quality like whiteness, the cat-
egory of Brahmin-ness is not perceptible without an additional teaching, and
so it cannot be counted as a natural kind. The class category of humanness is
manifested by the form of the hands, feet and other factors that people pos-
sess, a form that is distinct from the particular forms possessed by cows and
horses.55 But no such criterion can be conceived for Brahmins. Prabhācandra
here argued that it is linguistic convention that produces conceptions of social
class categories that are additional to the natural class category of being a
human. Following this appeal to conceptions, Prabhācandra’s hostile interloc-
utor asked, “How is it that there is such a recurring cognition that is attached
to the social class that is Brahmin-ness anywhere?”56 To which Prabhācandra
replied:
It is on account of the power of convention. Indeed, it is just as there is

a single conception with the recurrent form “This is a cow, and this is
a cow,” for individuals in, say, a herd of cattle, which are different from
each other, even if there is no class category that has the form of a single
‘cow-ness.’ In a similar manner, there will be a single conception with the
recurrent form “this is a Brahmin, and this is a Brahmin,” for certain par-
ticular instantiations of ‘humanness,’ even though they are different from
each other. But, if it were caused by the capability of a real thing, there
would be the conception that “this is a Brahmin, and this is a Brahmin” for
particular instances only by means of a cognition of a mere thing itself.
And it would—in that case—be like the conception of a cow, which is
54 na ca visphāritākṣasya purovarttikhaṇḍamuṇḍakarkādivyaktiṣu gavāśvādijātivat śuklat‑

vādiguṇavad vā manuṣyavyaktiṣu manuṣyatvapuṃstvādyatiriktasya brāhmaṇyasyaikasya
akhilasvavyaktiṣv anugatasya pratibhāso ’sti | Prabhācandra, Nyāyakumudacandra, 770–
771. This argument is re-stated in Vādideva, Syādvādaratnākara, 958.
55 hastapādādyākāravyaṅgyamanuṣyatvam | Prabhācandra, Nyāyakumudacandra, 771.
56 katham evaṃ kvacid brāhmaṇatvānurakto ’nugatapratyayaḥ syād? iti cet | Prabhācandra,
Nyāyakumudacandra, 771.

22 10.1163/25425552-12340042 | Gurevitch
without exceptions—even when someone does not cognize the conven-

tion. And it is not like that.57
As already discussed, Prabhācandra did not deny the reality of conceptions of

Brahmins. These conceptions do exist and cannot simply be explained away.
In many respects, they are exactly the same type of conceptions as the concep-
tion of natural kinds such as cows or horses. But conceptions of social class are
different in an important respect: They are caused by convention. It is not the
things themselves that cause such conceptions. Rather, these conceptions con-
sist in the cognition of a convention, one defined by people in language and
that is changing over time.58 Such conventions map the world imperfectly. Not
all people called “Brahmins” will act according to the criteria specified by the
convention. Prabhācandra accepted the existence of cognitions of people as
Brahmins, which are founded on everyday language, but once again, he denied
the existence of Brahmin-ness—of anything additional to convention—in the
case of social classes. He was careful to make a distinction between how peo-
ple cognize cows and how people cognize caste.
Vādirāja too used a similar line of reasoning to Prabhācandra’s arguments to
bring the reality of caste into question. The presence of an additional cognitive
factor—here a teaching from another person—necessary for the perception
of caste means that social classes cannot produce cognitions on their own.59
And if they cannot produce cognitions on their own, then they are not natu-
ral kinds. “There is no manifestation of the class category of Brahmin-ness,”
Vādirāja wrote, pushing back against the epistemic validity of an additional fac-
tor assisting the cognition of a natural kind, “perceived in a particular instance
such as someone who is in the lineage of Kauṇḍinya. Otherwise, the determi-
nation of that would be possible without instruction—just as in the case of
cow-ness. And if a particular manifests that class category, it does so only with
the assistance of an additional teaching. If it were capable of manifesting the
57 saṅketavaśāt | yathaiva hi parasparavilakṣaṇeṣu gov[r]ajādiṣ[v] ekagotvarūpasāmānyāb‑

hāve ’pi “gauḥ gauḥ” ity anugatākāraikapratyayaḥ tathā anonyavilakṣaneṣv api manuṣyav‑
yaktiviśeṣeṣu “brāhmaṇo ’yaṃ brāhmano ’yam” iti anugatākāraikapratyayo bhaviṣyati |
vastusāmarthyaprabhavatve tu agṛhītasaṅketāsv api vyaktiṣu tanmātropalambhenaiva
avyabhicārigopratyayavat sa syāt | na caivam | Prabhācandra, Nyāyakumudacandra, 771.
58 For Dharmakīrti’s discussion of linguistic convention, see Dan Arnold, “On Semantics and
Saṃketa: Thoughts on a Neglected Problem with Buddhist Apoha Doctrine,” Journal of
Indian Philosophy 34, no. 5 (2006): 415–78.
59 This is a refutation of Kumārila, Ślokavarttika, 5.15 (Vanavāda), 26–29.

class category on its own, what would be the use of its being dependent on an
additional assisting factor?”60
Anantavīrya, Prabhācandra and Vādirāja—following Akalaṅka—all argued
that some class categories are real things that people can directly perceive. This
is because a component of any cognition is the cognition of similarity. And to
perceive similarity is to cognize a class category. This entails a class category
that is strikingly different from the universal of Mīmāṃsā or Nyāya. A class is
not an eternal essence that is separate from each particular and yet pervades
it. Rather, a class category is an aspect of any particular that is perceived. We
can know the reality of a class category through cognitions of similarity. But
this does not mean that a class is necessarily a mental projection. Some classes
are very real, for instance species and sex. Yet, some class categories are men-
tal projections. That is, some classes can only be cognized with the assistance
of an instruction. These classes are ultimately unreal. To distinguish between
these different types of conceptions is of utmost importance to a philosopher.
And the question of caste was the central site for Anantavīrya, Prabhācan‑
dra and Vādirāja to distinguish these different sorts of conceptions.
Deviating from Nature
It can often be difficult to distinguish what is natural from what is conven-

tional, and attempts to distinguish the two realms have proven to be important
cases for historians of science, gender and race.61 Medieval disputes over the
nature of caste have much to offer these discussions. The question of caste
allowed for philosophers in medieval South Asia to develop with specific cri-
teria for distinguishing certain class categories as natural from other ones as
conventional. Since Kumārila—who defended the possibility of cognizing
social classes—rejected all criteria for caste except for genealogy, much of the
critique of the eleventh-century epistemologists would take aim at that one
remaining criterion. In so doing, Jain philosophers employed in the courts of
60 kauṇḍinyāder na hi vyaktes tadvyaktir upalabhyate |

anyathānupadeśaḥ syān niścayas tatra gotvavat ||
upadeśasahāyaiva vyaktis tadvyañjikā yadi |
kevalaiva samarthā cet sahāyāpekṣaṇena kim || Vādirāja, Nyāyaviniścayavivaraṇa, vol. 1,
p. 500, vv. 1152–3.
61 For the medieval period, see especially Joan Cadden, Nothing Natural Is Shameful
(University of Pennsylvania Press, 2013) and Lorraine Daston and Katharine Park,
Wonders and the Order of Nature, 1150–1750 (Zone Books, 2001).

24 10.1163/25425552-12340042 | Gurevitch
western India once again extended arguments developed by Buddhist philoso-

phers but put them to new uses.
Prabhācandra’s attack proceeded by introducing an epistemic doubt into
genealogies. Here he sounded very much like Vādirāja and Anantavīrya. The
fact of an unbroken lineage, for Prabhācandra, cannot be established based on
direct perception, inference, or testimony. This is the case because of an obser-
vation he made about sexual reproduction. Prabhācandra wrote:
Certainly, there is no disparity—even in a dream—between a Brahmin

woman and her children whether they are produced from either a
Brahmin or a Śūdra, in the manner in which there is a disparity between
a mare and her offspring depending on whether they are produced from
either a donkey or a stallion.62
It is easy to perceive whether a mare has mated with a donkey or a stallion

based on whether she gives birth to a foal or a mule. Such distinctions are
directly perceptible and are, in Prabhācandra’s estimation, based on real differ-
ences and similarities. But no such distinction can be determined for people of
different social classes. Lineage is not something that we can directly perceive,
because we cannot directly perceive the fertilization and gestation process.
This means that we can never know whether a person comes from the unbro-
ken lineage that they claim. And if we cannot know lineage through direct
perception, then we will also have no grounds to make an inference regarding
it. As Prabhācandra wrote:
It is not, in the first instance, that that fact of one’s parents being
non-promiscuous is cognizable from direct perception, because it is
impossible for a normal person with limited vision to directly witness
an object of the form “this person was born from this mother and this
very father.” Nor is it cognizable from an inference. Because you—my
interlocutor—do not accept the possibility of an inference for something
that is not an object of direct perception. And an inference proceeds
from an inferential mark. And there isn’t any inferential mark for indicat-
ing the fact that one’s parents are non-promiscuous.63
62 na khalu vaḍavāyāṃ gardabhāśvaprabhavāpatyeṣv iva brāhmaṇyāṃ brāhmaṇaśūdra‑

prabhavāpatyeṣv api vailakṣaṇyaṃ svapne ’pi pratīyate | Prabhācandra, Nyāyakumuda‑
candra, 773.
63 na tāvat pratyakṣataḥ | “ayam etasmād eva etasyām utpannaḥ” ity evaṃrūpasyārthasya
arvāgdṛśā pratyakṣīkarttum aśakyatvāt | nāpy anumānāt | pratyakṣāviṣaye bhavatā
anumānānabhyupagamāt | liṅgāc ca anumānam udayam āsādayati | na ca pitraviplutatve
kiñcil liṅgam asti | Prabhācandra, Nyāyakumudacandra, 772.
For Prabhācandra—and here he followed the Buddhist philosopher

Prajñākaragupta—the possibility of multiple sexual partners was turned into
an epistemological problem, since any deviation will mean that an inference of
lineage is impossible. And this possibility was actually realized in the world. “It
is perceived that women are constantly deviating—even in this birth here on
earth right now—on account of their love-sickness,” Prabhācandra asserted,
before going on to say, “And so, how is it that even the god Brahmā could know
what it is that they do—and when it is that they do it—throughout time
immemorial?”64
In his rejection of the genealogical argument for caste, Prabhācandra moved
between making epistemological and ontological claims. On the one hand, he
rejected the genealogical argument because an unbroken lineage can never
be properly cognized. We can never be certain about the past, which renders
all caste-claims potentially dubious. This epistemological doubt is conditioned
by an ontological claim: We cannot be certain of a person’s lineage, because
different sorts of people produce children who look alike. Since Brahmin
women can reproduce with non-Brahmin men—and in so doing have chil-
dren who appear the same as the children they have with Brahmin men—we
must conclude that all people belong to a single species, and any further claims
to sub-divisions within that species will not be founded on direct perception.
The possibility of reproduction challenges claims to difference. These epis-
temological and ontological claims were closely linked in Prabhācandra’s
argument about natural kinds. He used conceptions as the grounds to test
the reality of class categories. If a class category is cognized in a certain man-
ner, Prabhācandra accepted it as ontically real. All other conceptions will not
be denied, but they are not caused by really existing objects, rather they are
caused by conventions.
The Problem of Caste
This article opened by arguing that caste mattered and that caste was a
problem.65 Caste was made into a problem in different ways in the medi-
eval India, and as such it became a matter of concern for different sorts of
individuals. In one sense, caste was a material problem, a problem about the
64 sadaiva avalānāṃ kāmāturatayā iha janmany api vyabhicāropalambhāt anādau kāle tāḥ
kadā kiṃ kurvantīti brahmaṇāpi jñātum aśakyam | Prabhācandra, Nyāyakumudacandra,
773. Compare Prabhācandra, Prameyakamalamārtaṇḍa, 482.
65 In thinking about the historical coming into being of caste as epistemological and socio-
logical problems, I follow Viswanath, Rupa. The Pariah Problem: Caste, Religion, and the
Social in Modern India. Cultures of History. New York: Columbia University Press, 2014.
26 10.1163/25425552-12340042 | Gurevitch
distribution of resources in society, with inter-caste disputes about the access

to resources recorded in inscriptions throughout the region. In addition to this,
the scholars discussed in this article turned caste into a philosophical problem.
These philosophers investigated the conditions of possibility upon which spe-
cific disputes over caste could be made.
As they acknowledged in their writing, it was important for these
eleventh-century scholars to carry out their intellectual projects under the
auspices and confidence of a royal court. They negotiated complex political
hierarchies as a part of their intellectual careers, and their scholarly output
was a means of negotiating such hierarchies. In discussing the uses of schol-
arly argument, Vādirāja argued that a powerful king was necessary to ensure
the limits of rational debate so that overly-partisan scholars do not go about
“squawking like crows.”66 And if institutions mattered for these scholars, these
scholars mattered to institutions. They were rewarded for their scholastic
efforts, they oversaw other scholarly projects within the court, and their names
found a place in stone across western India. This marks an important break
with philosophers of the era of Kumārila and Dharmakīrti. For the first time,
philosophy was brought into the political court.
How these philosophers used their positions in court society to affect every-
day experiences of caste is difficult to say. Did they merely interpret the world,
or did they seek to change it? There remained Jain intellectuals who modelled
their identities around those of Brahmins.67 And King Bhoja—who employed
Prabhācandra—provided support for caste distinctions across his own writ-
ings, prescribing, for instance, different sized houses for people of different
castes.68 But it is clear that caste was not simply one ready-to-hand example
among many for philosophers like Prabhācandra and Vādiraja, and in question-
ing its basis as an epistemological issue they attacked the heart of a number of
prevailing regimes of evidence.
Attempts to solve the epistemological problem of caste have been posed
by many different regimes of knowing and by many different political insti-
tutions. The attempt to know caste always results in a way of fixing caste, of
reifying it as an object to be known, and different knowledge-regimes have
resulted in different epistemic objects. Recent scholarship on the late-colonial
period has traced what we can call a shift from a bureaucratic evidentiary
66 Vādirāja, Nyāyaviniścayavivaraṇa, 50.

67 See Dundas “The Digambara Jain warrior,” although my treatment of Prabhācandra
diverges significantly.
68 Bhoja, Samarāṅgaṇasūtradhāra 19.15–20.

regime to a biometric one.69 This was a moment in which caste was turned
into a statistical object, wherein standard deviation could be understood, to
use P.C. Mahalanobis’s evocative phrase, to describe “caste distances.” Initially
using specialized anthropometric tools and later using techniques conceived
to be less prone to observer intervention, such as blood sera and genomic
markers, this biometric regime—never hegemonic—was constructed on the
analogy of race and caste and used multiple practices in an attempt to stabilize
caste as an epistemic object.70
This biometric regime built on top of—but did not replace—early-modern
evidentiary regimes, the practices and values of which have been exten-
sively mapped by recent scholarship. This was a regime in which written
documents mattered and were used to assess the validity of claims to caste.71
Understanding how these evidentiary regimes operated has helped to show
the effects of colonial and post-colonial state interventions into the question
of caste.72 Collectively, these studies have shown how, as Rosalind O’Hanlon,
Gergely Hidas and Csaba Kiss have put it,
Caste was not a changeless system until the colonial ‘reinvention’ about
which so many scholars have written. Rather, the colonial state inherited
an order of varṇa and jāti marked by acute and still developing tensions,
69 Projit Bihari Mukharji, “Profiling the Profiloscope: Facialization of Race Technologies and
the Rise of Biometric Nationalism in Inter-War British India,” History and Technology 31,
no. 4 (2015): 376–96.
70 Projit Bihari Mukharji, “From Serosocial to Sanguinary Identities: Caste, Transnational
Race Science and the Shifting Metonymies of Blood Group B, India c. 1918–1960,” The
Indian Economic & Social History Review 51, no. 2 (2014): 143–76.
71 See especially Dayal, Subah, “Making the ‘Mughal’ Soldier: Ethnicity, Identification,
and Documentary Culture in Southern India, c. 1600–1700,” Journal of the Economic and
Social History of the Orient 62, no. 5–6 (2019): 856–924; O’Hanlon, Rosalind, “Letters
Home: Banaras Pandits and the Maratha Regions in Early Modern India,” Modern Asian
Studies 44, no. 2 (March 2010): 201–40; Peabody, Norbert, “Cents, Sense, Census: Human
Inventories in Late Precolonial and Early Colonial India,” Comparative Studies in Society
and History 43, no. 4 (October 2001): 819–50; and Cherian, Divya. Merchants of Virtue:
Hindus, Muslims, and Untouchables in Eighteenth-Century South Asia, 2022.
72 See Rosalind O’Hanlon, “Contested Conjunctures: Brahman Communities and ‘Early
Modernity’ in India,” The American Historical Review 118, no. 3 (2013): 765–87 and
Rosalind O’Hanlon, “Caste and Its Histories in Colonial India: A Reappraisal,” Modern
Asian Studies 51, no. 2 (2017): 432–61.

28 10.1163/25425552-12340042 | Gurevitch
whose consequences continued to reverberate through many areas of

nineteenth- and twentieth-century society.73
A spate of recent studies has shown exactly what this early-modern order was,
what types of evidence counted, and how disputes about caste were adjudi-
cated and enforced.74
The quarry I am after is slightly different and is one degree removed from
this important scholarship. What I have tried to emphasize in this article is
that this fact—the fact that there were different regimes by which to know
caste—was itself an object of scholarly inquiry in medieval South Asia. The
fact that there were different regimes by which to know caste was turned
into an object of scholarship by philosophers who sought to investigate the
conditions of possibility for making claims about caste. And, in the eleventh
century, the epistemological problem of caste was interrogated alongside the
socio-legal problem of caste.
In their discussions of what constitutes a means of valid knowledge, Vādirāja
and Prabhācandra rejected what they called the everyday conception of evi-
dence. They both cited a commonly accepted statement of legal procedure
that say that “evidence is said to be of three types: documents, witnesses, and
usufruct.”75 The locus classicus for this juridical understanding of evidence was
given in the Brahmanical Law Code of Yājñavalkya (Yājñavalkyasmṛti).76 By the
medieval period, variations on this statement were included in general guides
to politics produced in the Deccan such as Somadeva’s Ambrosia of Political
Discourse (Nītivākyāmṛtam).77 Both Vādirāja and Prabhācandra refuted this
juridical notion of evidence, arguing that it did not capture with precision
73 Rosalind O’Hanlon, Gergely Hidas, and Csaba Kiss, “Discourses of Caste over the Longue
Durée: Gopīnātha and Social Classification in India, ca. 1400–1900,” South Asian History
and Culture 6, no. 1 (January 2, 2015): 102–29, on p. 118.
74 In addition to the scholarship cited above, see Ananya Vajpeyi, “Śūdra Dharma and Legal
Treatments of Caste,” in Hinduism and Law: An Introduction, ed. Donald Davis, Timothy
Lubin, and Jayanth Krishnan (New York: Cambridge University Press, 2010), 154–66;
Samuel Wright, “History in the Abstract: ‘Brahman-Ness’ and the Discipline of Nyāya in
Seventeenth-Century Vārāṇasī,” Journal of Indian Philosophy 44, no. 5 (2016): 1041–69;
Sumit Guha, Beyond Caste (Leiden: Brill, 2013); and Christian Lee Novetzke, “The Brahmin
Double: The Brahminical Construction of Anti-Brahminism and Anti-Caste Sentiment
in the Religious Cultures of Precolonial Maharashtra,” South Asian History and Culture 2,
no. 2 (2011): 232–52.
75 likhitaṃ sākṣino bhuktiḥ pramāṇaṃ trividhaṃ smṛtam | Quoted in Vādirāja, Nyāyaviniś‑
cayavivaraṇam, vol. 1, p. 57; and Prabhācandra, Prameyakamalamārtaṇḍa, 8.
76 pramāṇaṃ likhitaṃ bhuktiḥ sākṣiṇaś ceti kīrtitam | Yājñavalkyasmṛti 2.22ab.
77 bhuktiḥ sākṣī śāsanaṃ pramāṇam | Somadeva, Nītivākyāmṛtam 28.9.

the definition of evidence. Documents are not enough. They cannot speak for
themselves. The philosopher and the jurist have different criteria for assessing
evidence, and philosophers can demonstrate the limits of juridical reasoning.
Despite this critique, in the subsequent century, Yājñavalkya’s Law Code
would find a place at the heart of the Western Cāḷukya court, through Vijñāneś‑
vara’s Concise Commentary (Mitākṣara), which served to retrench caste
orthodoxy and a documentary order of evidence.78 Vijñāneśvara’s Concise
Commentary went on to be used as the basis for much colonial property law,
serving as a foundation for assessing inheritance and familial lineages outside
of Bengal.79 And in discussing mixed-castes, the Brahmin jurist Vijñāneśvara
invoked the argument about natural kinds that we have seen in a manner at
odds with that of Anantavīrya, Vādirāja and Prabhācandra. He wrote that “just
as a cow is produced in a female cow because of a male cow, a horse is pro-
duced in a mare on account of a stallion, in the same manner, a Brahmin child
is produced in a Brahmin woman because of a Brahmin man.”80 Vijñāneśvara
admitted a distinction between cows and Brahmins, but for him the differ-
ence between these two sorts of classes was insignificant. He wrote that “while
species is understood by means of direct perception, the social classes such
as Brahmins are defined by scriptures, in accordance with what has been
recollected.”81 Vijñāneśvara and the Brahmin legal theorists like him drew a
divide between epistemology and ontology: Even though it takes different
ways of knowing to distinguish Brahmins and cows, they are coherent catego-
ries of the same sort. And texts point to the reality of Brahmin-ness.
It was precisely this regime of evidence that Vādirāja and Prabhācandra
refused. They denied that documentary evidence could ever provide socially
constructed categories such as caste the kind of validity their sectarian and
intellectual opponents required. But they did not deny that such claims
78 Whitney Cox, “Law, Literature, and the Problem of Politics in Medieval India,” in
Hinduism and Law, ed. Donald Davis, Timothy Lubin, and Jayanth Krishnan (New York:
Cambridge University Press, 2010), 167–82. For an opposing interpretation of this text,
see Jason Schwartz, “The King Must Protect the Difference: The Juridical Foundations of
Tantric Knowledge,” Religions 9, no. 4 (2018). Donald Davis has shown how Vijñāneśvara
follows the 9th-century jurist Medhātithi in his arguments. See, Davis, Donald R. “Seeing
through the Law: A Debate on Caste in Medieval Dharmaśāstra.” Contributions to Indian
Sociology 56, no. 1 (February 1, 2022): 17–40, on. p. 31.
79 Nandini Bhattacharya-Panda, Appropriation and Invention of Tradition (New Delhi: OUP
India, 2007).
80 yathā gor gavi gauḥ, aśvād vaḍavāyām aśvaḥ tasmād brāhmaṇād brāhmaṇyām utpanno
brāhmaṇa iti na viruddham | Vijñāneśvara, Mitakṣara, 65.
81 yatra pratyakṣagamyā jātir bhavati tatra tathā | brāhmaṇādijātis tu smṛtilakṣaṇā
yathāsmaraṇaṃ bhavati | Vijñāneśvara, Mitakṣara, 65.

30 10.1163/25425552-12340042 | Gurevitch
were made and that they had very real effects in the world. Philosophy can-
not change how people normally talk. But it can change how people assess
evidence. Vādirāja and Prabhācandra—following Anantavīrya—used their
writing to bring different regimes of evidence into conflict with each other. In
so doing, they showed that there can never be a definitive way of fixing caste.
Acknowledgements
This paper would not have been possible without encouragement from
Pradeep Gokhale in philosophic matters. I thank Seema Chauhan, Whitney
Cox, Lorraine Daston, Pranathi Diwakar, Shireen Hamza, Anil Mundra,
Rosalind O’Hanlon, Andrew Ollett, Surabhi Pudasaini, Itamar Ramot, Samira
Sheikh, and Sarah Taylor for substantive discussions on the paper. Lawrence
McCrea provided corrections and suggestions on the final draft.
Bibliography
Primary Sources
Akalaṅka. Nyāyaviniścaya. In Akalaṅka, and Vādirājasūri. Nyāya Viniścaya Vivaraṇa of
Śrī Vādirāja Sūri the commentary on Bhaṭṭākalaṅkadeva’s Nyāya Viniścaya. Edited by
Mahendrakumāra Jaina. 2 vols. Benares: Bhāratīya Jñānpītha Kāshi, 1949.
Akalaṅka. Siddhiviniścaya. In Akalaṅka, and Anantavīrya. Siddhiviniścayaṭīkā. Edited
by Mahendrakumāra Jaina. 2 vols. Benares: Bhāratīya Jñānpītha Kāshi, 1959.
Anantavīrya. Siddhiviniścayaṭīkā. In Akalaṅka, and Anantavīrya. Siddhiviniścayaṭīkā.
Bhoja. Samarāṅgaṇasūtradhāra. Edited by T. Gaṇapati Śāstrī and Vasudeva Saran
Agrawala. Revised. Baroda: Oriental Institute of Baroda, 1966.
Dharmakīrti. The Pramāṇavārttikam of Dharmakīrti; the First Chapter with the Auto‑
commentary. Edited by Raniero Gnoli. Roma: Instituto italiano per il medio ed
estremo oriente, 1960.
Jayanta Bhaṭṭa. Āgamaḍambara otherwise called Ṣaṇmatanāṭaka. Edited by V. Ragha‑
van and Anantalal Thakur. Darbhanga: Mithila Institute, 1964.
Kumārila. Ślokavārttika. Edited by Dvārikādāsa Śāstri. Prāchyabhārati Series 10. Vara‑
nasi: Tara Publications, 1978.
Kumārila. Tantravārttika. Edited by Gaṅgādhara Śāstri. Benares Sanskrit Series.
Benares: The Benares Printing Press & the Vidya Vilas Press, 1903.
Māṇikyanandin. Parīkṣāmukha. In Prabhācandra. Prameyakamalamārtaṇḍa. Edited
by Mahendra Kumar Shastri. 2nd ed. Bombay: Satyabhamabai Pandurang, Nirnaya
Sagar Press, 1941.

Manusmṛti. Edited by Manmatha Nath. Dutt. Delhi: Chaukhamba Sanskrit Pratishthan,

1998.
Prabhācandra. Prameyakamalamārtaṇḍa. Edited by Mahendra Kumar Shastri. 2nd ed.
Bombay: Satyabhamabai Pandurang, Nirnaya Sagar Press, 1941.
Prabhācandra. Nyāyakumudacandra. Edited by Mahendra Kumar Shastri. 2 vols.
Bambaī: Maṇikacandra Di. Jaina Granthamālā, 1938.
Prajñākaragupta. Pramāṇavārttikabhāṣyam or Vārttikālaṅkāra. Edited by Rāhula
Sāṅkṛityāyana. Tibetan Sanskrit Works Series. Patna: Kashi Prasad Jayaswal
Research Institute, 1953.
Epigraphia Carnatica. Rice, B. Lewis, and R. Narasimhacarya, eds. Epigraphia Carnatica.
Bangalore: Mysore Government Central Press, 1886–1919.
Somadeva Sūri. Nītivākyāmṛtam. Edited by Pannālāl Sonī. Bombay: Māṇikacandra
Digambara Jaina Granthamālā Samiti, 1922.
Vādideva. Syādvādaratnākara. Edited by Motīlāla Lādhā. Osavāla. Punā: Osavāla, 1925.
Vādirāja. Nyāyaviniścayavivaraṇa. In Akalaṅka, and Vādirājasūri. Nyāya Viniścaya
Vivaraṇa of Śrī Vādirāja Sūri the commentary on Bhaṭṭākalaṅkadeva’s Nyāya
Viniścaya. Edited by Mahendra Kumar Shastri. Jaina. 2 vols. Benares: Bhāratīya
Jñānpītha Kāshi, 1949.
Vijñāneśvara. The Mitākṣara with Visvarūpa and Commentaries of Subōdhini and
Bālambhatti. Edited by S.S. Setlur. Madras: Brahmavadin Press, 1912.
Yājñavalkyasmṛti. In Vijñāneśvara. The Mitākṣara with Visvarūpa and Commentaries of
Subōdhini and Bālambhatti. Edited by S.S. Setlur. Madras: Brahmavadin Press, 1912.
South Indian Inscriptions. 24 vols. Madras: Archaeological Survey of India, 1890–1982.
Secondary Sources
Appadurai, Arjun. “Is Homo Hierarchicus?” American Ethnologist 13, no. 4 (1986):
745–61.
Arnold, Dan. “Buddhist Idealism, Epistemic and Otherwise: Thoughts on the Alterna‑
ting Perspectives of Dharmakīrti.” Sophia 47, no. 1 (2008): 3–28.
Arnold, Dan. “On Semantics and Saṃketa: Thoughts on a Neglected Problem with
Buddhist Apoha Doctrine.” Journal of Indian Philosophy 34, no. 5 (2006): 415–78.
Arnold, Daniel. “Of Intrinsic Validity: A Study on the Relevance of Pūrva Mīmāṃsā.”
Philosophy East and West 51, no. 1 (2001): 26–53.
Balcerowicz, Piotr. “How Could a Cow Be Both Synchronically and Diachronically
Homogenous, or On the Jaina Notions of Tiryak-Sāmānya and Ūrdhvatā-Sāmānya.”
In Approaches to Jaina Studies: Philosophy, Logic, Rituals and Symbols, edited by
N.K. Wagle and Olle Qvarnström, 211–37. Toronto: University of Toronto: Center for
South Asian Studies, 1999.
Balcerowicz, Piotr. “Is ‘Inexplicability Otherwise’ (Anyathânupapatti) Otherwise Inex-
plicable?” Journal of Indian Philosophy 31, no. 1/3 (2003): 343–80.

32 10.1163/25425552-12340042 | Gurevitch
Balcerowicz, Piotr. “Dharmakīrti’s Criticism of the Jaina Doctrine of Multiplexity

of Reality (Anekāntavāda).” In Religion and Logic in Buddhist Philosophical
Analysis: Proceedings of the Fourth International Dharmakirti Conference, Vienna,
August 23–27, 2005, edited by Helmut. Krasser, Horst Lasic, Eli Franco, and Birgit
Kellner. Wien: Verlag der Österreichischen Akademie der Wissenschaften, 2011.
Balcerowicz, Piotr. “Siddhasena Mahāmati and Akalaṅka Bhaṭṭa: A Revolution in Jaina
Epistemology.” Journal of Indian Philosophy 44 (2016): 993–1039.
Bayly, Susan. Caste, Society and Politics in India from the Eighteenth Century to the
Modern Age. Cambridge, England: Cambridge University Press, 1999.
Bhattacharya-Panda, Nandini. Appropriation and Invention of Tradition: The East India
Company and Hindu Law in Early Colonial Bengal. New Delhi: OUP India, 2007.
Cadden, Joan. Nothing Natural Is Shameful: Sodomy and Science in Late Medieval
Europe. University of Pennsylvania Press, 2013.
Cherian, Divya. Merchants of Virtue: Hindus, Muslims, and Untouchables in Eighteenth-
Century South Asia, 2022.
Cohn, Bernard S. “The Census, Social Structure and Objectification in South Asia.” In
An Anthropologist among the Historians and Other Essays, 224–54. Delhi: Oxford
University Press, 1987.
Cox, Whitney. “Law, Literature, and the Problem of Politics in Medieval India.” In
Hinduism and Law: An Introduction, edited by Donald Davis, Timothy Lubin, and
Jayanth Krishnan, 167–82. New York: Cambridge University Press, 2010.
Daston, Lorraine, and Katharine Park. Wonders and the Order of Nature, 1150–1750.
Zone Books, 2001.
Davis, Donald R. “Seeing through the Law: A Debate on Caste in Medieval Dharma‑
sastra.” Contributions to Indian Sociology 56, no. 1 (February 1, 2022): 17–40.
Dayal, Subah. “Making the ‘Mughal’ Soldier: Ethnicity, Identification, and Documentary
Culture in Southern India, c. 1600–1700.” Journal of the Economic and Social History
of the Orient 62, no. 5–6 (2019): 856–924.
Derrett, J. Dunncan M. “Two Inscriptions Concerning the Status of Kammalas and the
Application of Dharmasastra.” In Professor K.A. Nilakanta Sastri Felicitation Volume,
edited by Saw Ganesan, S. Rajam, N.S. Ramaswami, and M.D. Sampath, 32–55.
Madras: Prof. K.A. Nilakanta Sastri Felicitation Committee, 1971.
Dikshit, G.S. Local Self-Government in Mediaeval Karnataka. Dharwar: Karnatak
University, 1964.
Dirks, Nicholas B. Castes of Mind: Colonialism and the Making of Modern India.
Princeton, NJ: Princeton University Press, 2001.
Dundas, Paul. “The Digambara Jain Warrior.” In The Assembly of Listeners: Jains
in Society, edited by Michael Carrithers and Caroline Humphrey. Cambridge:
Cambridge University Press, 1991.

Eltschinger, Vincent. Caste and Buddhist Philosophy: Continuity of Some Buddhist

Arguments Against the Realist Interpretation of Social Denominations. Translated by
Raynald Prévéreau. Delhi: Motilal Banarsidass Publishers, 2012.
Eltschinger, Vincent, Helmut Krasser, and John Taber. Can the Veda Speak?: Dharmakīrti
against Mīmāṃsā Exegetics and Vedic Authority: An Annotated Translation of PVSV
164,24–176,16. Wien: Verlag der Österreichischen Akademie der Wissenschaften,
2012.
Franco, Eli. “How to Distinguish between Non-Existing Entities? Dharmakīrti and
Prajñākaragupta on Universals as Objects of Knowledge.” Rocznik Orientalistyczny
65, no. 1 (2012): 51–62.
Gorisse, Marie-Hélène. “Is Inference a Cognitive or a Linguistic Process? A Line of
Divergence between Jain and Buddhist Classifications.” In Puṣpikā: Tracing Ancient
India, through Texts and Traditions, edited by Robert Leach and Jessie Pons, 1–23.
Oxford: Oxbow Books, 2015.
Guha, Sumit. Beyond Caste: Identity and Power in South Asia, Past and Present. Beyond
Caste. Leiden: Brill, 2013.
Gurevitch, Eric Moses. “Everyday Sciences in Southwest India.” PhD Thesis, University
of Chicago, 2022.
Hacking, Ian. Representing and Intervening: Introductory Topics in the Philosophy of
Natural Science. Cambridge: Cambridge University Press, 1983.
Hacking, Ian. The Social Construction of What? Cambridge, Mass: Harvard University
Press, 1999.
Halbfass, Wilhelm. Tradition and Reflection: Explorations in Indian Thought. Albany,
NY: State University of New York Press, 1991.
Inden, Ronald. Imagining India. Bloomington, IN: Indiana University Press, 1990.
Jaini, Padmanabh S. Gender and Salvation: Jaina Debates on the Spiritual Liberation of
Women. Berkeley: University of California Press, 1991.
Jodhka, Surinder S. “The Problem: Introduction to Caste Matters, A Symposium on
Inequalities, Identities and Disintegrating Hierarchies in India.” Seminar, no. 633
(2012).
Kataoka, Kei. “Transmission of Scripture: Exegetical Problems for Kumārila and
Dharmakīrti.” In Scriptural Authority, Reason and Action: Proceedings of a Panel
at the 14th World Sanskrit Conference, Kyoto, September 1st–5th, 2009, edited by
Vincent Eltschinger and Helmut Krasser. 319–74. Wien: Verlag der Österreichischen
Akademie der Wissenschaften, 2013.
McCrea, Lawrence. “Prajñākaragupta on the Pramāṇas and Their Objects.” In Religion
and Logic in Buddhist Philosophical Analysis: Proceedings of the Fourth International
Dharmakirti Conference, Vienna, August 23–27, 2005, edited by Helmut. Krasser,
Horst Lasic, Eli Franco, and Birgit Kellner, 319–28. Wien: Verlag der Österreichischen
Akademie der Wissenschaften, 2011.

34 10.1163/25425552-12340042 | Gurevitch
Metcalf, Thomas R. Ideologies of the Raj. Cambridge: Cambridge University Press, 1994.
Mosse, David. “Outside Caste? The Enclosure of Caste and Claims to Castelessness in
India and the United Kingdom.” Comparative Studies in Society and History 62, no. 1
(2020): 4–34.
Mukharji, Projit Bihari. “From Serosocial to Sanguinary Identities: Caste, Transnational
Race Science and the Shifting Metonymies of Blood Group B, India c. 1918–1960.”
The Indian Economic & Social History Review 51, no. 2 (2014): 143–76.
Mukharji, Projit Bihari. “Profiling the Profiloscope: Facialization of Race Technologies
and the Rise of Biometric Nationalism in Inter-War British India.” History and
Technology 31, no. 4 (2015): 376–96.
Novetzke, Christian Lee. “The Brahmin Double: The Brahminical Construction of
Anti-Brahminism and Anti-Caste Sentiment in the Religious Cultures of Precolonial
Maharashtra.” South Asian History and Culture 2, no. 2 (2011): 232–52.
O’Hanlon, Rosalind. “Caste and Its Histories in Colonial India: A Reappraisal.” Modern
Asian Studies 51, no. 2 (2017): 432–61.
O’Hanlon, Rosalind. “Contested Conjunctures: Brahman Communities and ‘Early
Modernity’ in India.” The American Historical Review 118, no. 3 (2013): 765–87.
O’Hanlon, Rosalind. “Letters Home: Banaras Pandits and the Maratha Regions in Early
Modern India.” Modern Asian Studies 44, no. 2 (2010): 201–40.
O’Hanlon, Rosalind, Gergely Hidas, and Csaba Kiss. “Discourses of Caste over the
Longue Durée: Gopīnātha and Social Classification in India, ca. 1400–1900.” South
Asian History and Culture 6, no. 1 (2015): 102–29.
O’Hanlon, Rosalind, and Christopher Minkowski. “What Makes People Who They Are?
Pandit Networks and the Problem of Livelihoods in Early Modern Western India.”
The Indian Economic & Social History Review 45, no. 3 (2008): 381–416.
Patil, Parimal G. “On What It Is That Buddhists Think About: Apoha in the Ratnakīrti-
Nibandhâvali.” Journal of Indian Philosophy 31, no. 1 (2003): 229–56.
Peabody, Norbert. “Cents, Sense, Census: Human Inventories in Late Precolonial and
Early Colonial India.” Comparative Studies in Society and History 43, no. 4 (2001):
819–50.
Schwartz, Jason. “The King Must Protect the Difference: The Juridical Foundations of
Tantric Knowledge.” Religions 9, no. 4 (2018).
Sen, Prabal Kumar. “Caste in Classical Indian Philosophy: Some Ontological Prob‑
lems.” In Classical Buddhism, Neo-Buddhism and the Question of Caste, edited by
Pradeep P. Gokhale, 40–64. London: Routledge, 2020.
Shah, Nagin J. Akalaṅka’s Criticism of Dharmakīrti’s Philosophy. Ahmedabad: L.D.
Institute of Indology, 1967.
Taber, John. “Dharmakīrti, Svataḥ Prāmāṇyam, and Awakening.” Wiener Zeitschrift Für
Die Kunde Südasiens 56/57 (2015): 77–98.

Taber, John. “What Did Kumārila Bhaṭṭa Mean by Svataḥ Prāmāṇya?” Journal of the
American Oriental Society 112, no. 2 (1992): 204–21.
Talbot, Cynthia. “A Revised View of ‘Traditional’ India: Caste, Status, and Social Mobility
in Medieval Andhra.” South Asia: Journal of South Asian Studies 15, no. 1 (1992): 17–52.
Teltumbde, Anand. The Persistence of Caste: The Khairlanji Murders and India’s Hidden
Apartheid. London: Zed Books, 2010.
Vajpeyi, Ananya. “Śūdra Dharma and Legal Treatments of Caste.” In Hinduism and Law:
An Introduction, edited by Donald Davis, Timothy Lubin, and Jayanth Krishnan,
154–66. New York: Cambridge University Press, 2010.
Viswanath, Rupa. The Pariah Problem: Caste, Religion, and the Social in Modern India.
New York: Columbia University Press, 2014.
Wright, Samuel. “History in the Abstract: ‘Brahman-Ness’ and the Discipline of Nyāya
in Seventeenth-Century Vārāṇasī.” Journal of Indian Philosophy 44, no. 5 (2016):
1041–69.
Yengde, Suraj. Caste Matters. Gurgaon: India Viking, 2019.


Eric Moses Gurevitch - The Epistemology of Difference

Uploaded by

Copyright:

Available Formats

Eric Moses Gurevitch - The Epistemology of Difference

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Eric Moses Gurevitch - The Epistemology of Difference

Uploaded by

Copyright:

Available Formats

JOURNAL OF SOUTH ASIAN INTELLECTUAL

HISTORY (2023) 1–35

The Epistemology of Difference:

Received 15 June 2022 | Accepted 2 December 2022 |

Philosophy mattered in medieval India, with philosophers serving as officials in

Sanskrit philosophy – caste studies – historical epistemology – medieval philosophy

Scholars hold debates about class categories such as “Brahmin-ness,”

1 brāhmaṇatvādijātau ca vivadante vipaścitaḥ |

Downloaded from Brill.com 10/10/2023 06:11:09PM

Downloaded from Brill.com 10/10/2023 06:11:09PM

Figure 1 Political geography of 11th-century South Asia

7 jayisuvudĕ binadam uddhatacayamaṃ

Downloaded from Brill.com 10/10/2023 06:11:09PM

philosophers. Writing in the wake of the Buddhist philosopher Dharmakīrti

Downloaded from Brill.com 10/10/2023 06:11:09PM

If caste was once taken to be the defining characteristic of a timeless South

Dharmakīrti the Clown

It is a little-acknowledged fact in the history of Sanskrit philosophy that the

Downloaded from Brill.com 10/10/2023 06:11:09PM

15 Dharmakīrti, Pramāṇavārttika (svārthānumāna) 1.181–184 and the Auto-Commentary

Downloaded from Brill.com 10/10/2023 06:11:09PM

with Dharmakīrti’s joke about dog-flesh—was used as the basis of a critique

18 The term svataḥ prāmāṇyam is usually translated as “intrinsic validity,” a translation

Downloaded from Brill.com 10/10/2023 06:11:09PM

by themselves, then all the explanations of the Veda by Mīmāṃsā phi-

21 na hīdam eva me rūpam ayam evārtha ity api |

Downloaded from Brill.com 10/10/2023 06:11:09PM

the Mīmāṃsā claims to exclusive interpretation of the Vedas by adducing the

Between Tradition and Identity

The critiques written by Vādirāja and Prabhācandra expanded on arguments

26 śvamāṃsabhakṣaṇaṃ tasya vedārthatvaṃ kathaṃ na vaḥ? || Vādirāja, Nyāyaviniścayavi‑

Downloaded from Brill.com 10/10/2023 06:11:09PM

religious conversion and changing identities. He argued—in a sense taking

Mīmāṃsā interlocutor: A man might be a Brahmin. But when he has con-

Author: Objection! How is it that Brahmin-ness is destroyed? If it is in that

Interlocutor: Then maybe it is the case that Brahmin-ness is not destroyed.

29 saugatamatānusāriṇā dvijena śvapiśitāśanam | Anantavīrya, Siddhiviniścayaṭīkā, 514.

Downloaded from Brill.com 10/10/2023 06:11:09PM

Interlocutor: He is not, because he is not devoted to the Veda.

Author: Is it the case that Brahmin-ness simply is devotion to the Veda,

Interlocutor: Then a Brahmin is known by devotion to the Veda.

Anantavīrya here argued that it cannot be a particular behavior that determines

Downloaded from Brill.com 10/10/2023 06:11:09PM

Since people such as Brahmins are perceived in everyday observations to

Prabhācandra took it as his role—once again following the Buddhist phi-

From Generic Class to Social Class

At several points in his Verse Commentary (Ślokavārttikam) on Mīmāṃsā,

Downloaded from Brill.com 10/10/2023 06:11:09PM

36 For a summary of Akalaṅka’s position, see Shah, Akalaṅka’s Criticism of Dharmakīrti’s

Downloaded from Brill.com 10/10/2023 06:11:09PM

of Jain Philosophy (Syādvādaratnākara) in the Caulukya court of Jayasiṃha

There is not a particular type of conduct by which one could perceive

39 See the passages in Vādideva, Syādvādaratnākara, 958–962.

Downloaded from Brill.com 10/10/2023 06:11:09PM

Downloaded from Brill.com 10/10/2023 06:11:09PM

Interlocutor: Objection! If you do not accept the class category of some-

46 nanu brāhmaṇatvādisāmānyānubhyupagame kathaṃ bhavatāṃ varṇāśramavyavasthā

Downloaded from Brill.com 10/10/2023 06:11:09PM

What Makes a Brahmin Different from a Cow?