At one point in his discussion of dharma , Āpastamba, author of the oldest extant Dharmaśāstra, appears to show both exasperation at every Tom, Dick, and Harry trying to teach dharma and an awareness of the difficulties surrounding the epistemology of dharma when he exclaims: “Dharma and adharma do not go around saying: ‘Here we are!’ Nor do gods and Gandharvas, or the ancestors declare: ‘This is dharma . This is adharma ’” (ĀpDh 1.20.6). He tells his readers not to become “vexed or easily deceived by the pronouncements of hypocrites, crooks, infidels, and fools” (ĀpDh 1.10.5). Who were these hypocrites, crooks, infidels, and fools? And what were they saying about dharma ? We will never know, although I will present below some educated guesses; but at least this much is clear: dharma was a concept and term over which there were strong debates and disputes and thus a site of contention. This was the reason, I think, why all the Dharmaśāstras begin with a section on the epistemology of dharma : What is dharma ? And how do we come to know it?
The explicit discussion of how we come to know dharma , about the sources of dharma , is a unique and unprecedented feature of Dharmaśāstras; no text of parallel expert traditions deals with this core issue. 1 The ritual texts—Śrautasūtras and Gṛhyasūtras, belonging to the same textual corpus of Kalpasūtras as some of the texts on dharma —have no similar discussion of their epistemic sources. Even in later times, the most that is offered is the mythical origin of a particular discipline such as medicine or drama. These unique epistemological discussions on dharma provide us with valuable clues regarding the sociological and theological underpinnings of the term dharma and its application to various legal sectors.
The epistemological discussions within the Dharmaśāstra tradition fit neatly into broader concerns in the philosophy of law. In his pioneering and influential work The Concept of Law , H. L. A. Hart proposed a significant classification of law into primary and secondary rules. Most of the rules in the Dharmaśāstras fall into the category of primary rules or substantive law, that is, norms that govern individual and group activities. Hart’s category of secondary rules, that is, rules about primary rules, encompasses three groups: rules of recognition, change, and adjudication:
While primary rules are concerned with the actions that individuals must or must not do, these secondary rules are concerned with the primary rules themselves. They specify the ways in which the primary rules may be conclusively ascertained, introduced, eliminated, varied, and the fact of their violation conclusively determined. (Hart 2012 : 94)
Epistemology deals principally with the rules of recognition: how do we come to know the laws that apply to us? In the case of the Dharmaśāstras, how do we come to know dharma ? It also deals with the ways in which laws, once in force, may be changed or abrogated. These are the two issues dealt with in this chapter. Rules of adjudication, on the other hand, provide criteria for determining whether a primary rule has been violated, identify individuals who are competent to adjudicate, confer judicial powers on them, and provide legal procedures to be followed in adjudicating cases in a court of law. We will deal with this aspect of dharma in Chapter 22 .
Gautama is the first to present in unambiguous terms the party line of the Dharmaśāstra tradition and of its companion school Mīmāṃsā with regard to the epistemology of dharma when he opens his treatise with the thesis: “Veda is the root of dharma ” (vedo dharmamūlam, GDh 1.1). Making “Veda” the very first word of his treatise, Gautama clearly demonstrates the unrivaled position of the Veda as the root or epistemic source of dharma ; dharma is essentially Vedic. This epistemological position, in Wezler’s (2004 ) felicitous expression, is the vedamūlatva ideology. Although Gautama does not appear to take this final step, later scholars of Dharmaśāstra will present the Veda not just as one, but as the sole epistemic source of dharma . If there are other sources, as most authors will acknowledge, they are only secondary and proximate and must go back to and be based on the Veda from which alone they derive their authority. This position is clearly articulated at the very beginning of the Mīmāṃsā Sūtra (1.1.2): “Dharma is something beneficial disclosed by a Vedic injunction” (codanālakṣaṇo ‘rtho dharmaḥ ), complemented by the exclusionary provision given later (PMS 1.3.1): “Because Vedic texts are the foundation of dharma , anything lying outside the Vedic texts should be disregarded” (dharmasya śabdamūltvād aśabdam anapekṣaṃ syāt ).
This theological position, however, was not original or unchallenged, and two of the earliest writers on the subject, Āpastamba and the second-century bce grammarian Patañjali, offer quite different explanations of the nature and the major epistemic sources of dharma . Patañjali’s views are especially significant and interesting because it comes from a scholar not directly attached to the expert tradition of Dharmaśāstra. As we have already seen, Āpastamba, in stating explicitly that “gods and Gandharvas, or the ancestors do not declare: ‘This is dharma . This is adharma ’” (ĀpDh 1.20.6), appears to dismiss any kind of divine revelation of dharma , something we find later in the opening scene of Manu’s text.
Āpastamba addresses the epistemological issue indirectly in the opening sentence, defining the kind of dharma he will explore in his treatise: “Now, then, we shall explain the dharmas derived from agreed-upon normative practice. The authority is the agreement of those who know dharma ; and the Vedas”—athātaḥ sāmayācārikān dharmān vyākhyāsyāmaḥ | dharmajñasamayaḥ pramāṇam | vedāś ca | (1.1.1–3). Āpastamba uses the technical term pramāṇa , also employed by philosophers dealing with logic and epistemology, in this context with the meaning of “means of knowing” or epistemic source, as well as of authority, especially in the context of scriptural sources recognized in Indian logic as verbal authority (śabdapramāṇa ). This usage will be continued by later authors. Āpastamba gives a twofold answer to this epistemological question. Dharma rests first on agreed-upon normative practice and second on the authority of the Veda. The two words in this expression, agreement and normative practice , give us an insight into what Āpastamba had in mind when he characterized dharma in this way. In Āpastamba’s vocabulary the term ācāra refers specifically to normative practice that becomes a source of knowledge with respect to dharma : one can learn dharma by observing the practice of those who know and follow dharma just as one can learn good Sanskrit—so Patañjali would argue—by observing the speech patterns of particular individuals and communities. The term’s usage in the grammatical texts shows that it refers to behavior patterns characteristic of a particular group of people, behavior patterns that become models for others to follow. 2
The expression that qualifies dharma in the opening sentence of Āpastamba, then, indicates that the epistemic source of dharma is the normative behavior patterns that are generally accepted. But accepted by whom? According to Āpastamba, by those who know dharma (dharmajña ). This is a somewhat circular argument: how would one know that some people know dharma , when it is through their conduct that one comes to know dharma in the first place? Or, were the dharmajña a demographically identifiable group similar to śiṣṭa of Patañjali? This is a problem similar to that confronted by Patañjali (Ch. 1.II: 2) in defining the category “cultured elite” (śiṣṭa ). Patañjali resorts to the notion of a cultural or sacred geography (āryāvarta ), the region where the cultured elite live, while Āpastamba does not directly address the issue.
The second way to know whether a particular practice is authoritative is to see whether it is enjoined in the Vedas. Note, however, that the Vedas are given here only as an external check regarding the validity of a particular practice of a particular community; it is not given as the direct source of dharma and, unlike the prominence given to it by Gautama, it is here tucked away at the very end of the passage. That for Āpastamba the Veda is not the single source of dharma is clear also from his statement at the very end of his treatise (ĀpDh 2.29.11–2.29.15) that dharma can be known from women and Śūdras. This position is astonishing, given that in the mainstream of Brahmanical theology it would have been inconceivable to present Śūdras and women as having access to Vedic knowledge or as models of correct behavior.
The second-century bce grammarian Patañjali draws an interesting and significant parallel between correct Sanskrit and proper dharma . For him, there are two distinct and parallel domains of correct Sanskrit, the “Vedic” (vaidika ) and the “worldly” (laukika ). 3 The first is found in the extant Vedic texts, and the second, correct contemporary Sanskrit, in the speech of a special group of Brahmans whom he identifies as “cultured elite” (śiṣṭa ). What is significant for our discussion is that these two categories also comprehend the Vedic and the worldly realms of law. The dual domains of dharma of Āpastamba parallel the two domains of Sanskrit in Patañjali.
The examples of worldly speech given by Patañjali are not common everyday expressions but, significantly, are all derived from Dharmaśāstric statements. What is significant here is that for grammarians both the Veda and the “world,” the two domains of Sanskrit, are authoritative with respect to correct Sanskrit. This authoritative nature of the “world” is carried over into the framework of dharma when Patañjali cites worldly injunctions. Clearly, not everything that is said or done in the world is so authoritative. Thus world for Patañjali referred to Dharmaśāstra. We have confirmation of this conclusion. The two examples of not killing Brahmans and not drinking liquor that Patañjali (on Pāṇini 6.1.84; III: 57–8) takes to be worldly are cited by him again in his comments on Kātyāyana’s statement (Vārttika 39 on Pāṇini 1.2.64), which reads: “And so also the Dharmaśāstra.” And as an example of such a statement Patañjali gives the two injunctions: “One should not kill a Brahman. One should not drink liquor.” Clearly, for Patañjali world and Dharmaśāstra are, if not synonyms, at least equivalents with respect to authoritative speech. So for Patañjali, just as for Āpastamba, large areas of the Dharmaśāstra, the rules regarding proper conduct, are not derived from the Vedas but from normative practices in the world. For Patañjali, moreover, these practices come encoded in injunctions that are part of texts, and it is these textual forms, that is, the Dharmaśāstra, that invest them with authority.
Two authors of Dharmasūtras who come after Gautama, namely, Baudhāyana and Vasiṣṭha, advance the discussion of epistemology only marginally. Both authors take the category of “recollection” (smṛti ), introduced as a root of dharma by Gautama, as not simply orally articulated recollections but actual texts that record the authoritative recollections of the Brahmanical elite, which is the general meaning ascribed to this important category in later Indian literature. Yet neither Baudhāyana nor Vasiṣṭha explains what precisely these “texts of recollection” are. It seems unlikely that the category is self-referential, which would be tautological: the dharma that they are expounding in their texts cannot have as its authoritative source the very texts they are composing. We will have to wait until Manu to have this question answered.
Whereas Gautama presents the practice of “those who know the Vedas” as an epistemic source of dharma , both Baudhāyana and Vasiṣṭha use instead the expression “cultured elite,” who are presented as the standard for correct dharma as they are for correct Sanskrit in Patañjali. But we see the authors still groping for a proper technical term to use with regard to the behavior patterns of these individuals that provide the basis for dharma . Baudhāyana uses the term āgama with a meaning something like traditions that have come down from generation to generation. Elsewhere in his text, Baudhāyana uses the expression śiṣṭa-smṛti , that is, the recollection of the cultured elite. Vasiṣṭha, on the other hand, uses the term ācāra , normative practice, already employed by Āpastamba and Patañjali, and this term will become the standard in later Dharmaśāstras for the third epistemic source of dharma .
One other significant innovation introduced by Vasiṣṭha is the term śruti in place of Veda in discussing the epistemic sources of dharma . The term literally means “hearing or what is heard,” and it emphasizes the aural nature of the Veda; you can actually hear it being recited at any given point in time. And Vedic recitation is a central duty of every Brahman. One can find out the exact textual form of a Vedic passage from this hearing, and not its gist or meaning. The term is probably related to the pedagogy of Vedic instruction; the students recite exactly what they hear from the mouth of the teacher. Vasiṣṭha’s text represents the first use of this important term in the discussion of legal epistemology, and its use by him in the coordinative compound śruti-smṛti (scripture and text of recollection) referring to the dual textual sources of dharma , will become standard in later legal literature.
The three major treatises of the first millennium ce , those ascribed to Manu, Yājñavalkya, and Viṣṇu, present new elements and raise new issues into the discourse on the epistemology of dharma . As we have seen in Chapter 1 , Manu introduces several literary innovations including a frame story presenting his text as the work of the creator god himself, and on the issue of the sources of dharma Manu gives not three but five: “The root of dharma is the entire Veda, as also the recollection and conduct of those who know it; likewise the practice of good people, and satisfaction of oneself” (MDh 2.6). The first part of this statement is almost identical to Gautama’s, but Manu then appends two other sources: practice (ācāra ) of good people and satisfaction of oneself. 4 The latter is repeated by Yājñavalkya, but it had little impact on later discussions. The former, on the other hand, in the handy expression sadācāra (conduct of the good) becomes the standard third source of dharma ; Yājñavalkya, for example, gives this while dropping the category practice of those who know the Veda . By substituting good people for people (i.e., Brahmans) who know the Veda , Manu has broadened the authoritative community. This was a smart move, because broad swaths of dharma , such as the dharma of villages, families, and corporations, cannot be located among just people who know the Veda, and good people connects it to the deeply moral connotation of dharma in Manu’s understanding of the term.
Another significant development concerns the ambiguous term recollection (smṛti ): is it simply live memory or memory fixed in texts, or both? If it consists of texts, what are they? Manu clears up this ambiguity with the straightforward statement: “Scripture” (śruti ) should be recognized as the Veda and “recollection” (smṛti ) as Dharmaśāstra (MDh 2.10). In the early texts, recollection is presented as the source of dharma and thus external to the texts that the authors are engaged in composing. Manu, on the other hand, identifies recollection with these very texts on dharma and specifically with his own composition. Recollection that remained ambiguous in the zone between living recollection and textualized recollection is now firmly and unambiguously presented as Dharmaśāstric texts.
The second phase in the epistemology of dharma is represented by the major commentaries on Manu and Yājñavalkya composed between the fifth and ninth centuries ce . These commentators take as their basis the epistemology of dharma presented in the Dharmaśāstras that we have discussed above. In particular, they all assume as a fundamental principle that Veda is the sole foundation of dharma , even though some, like Medhātithi, think that some kinds of dharma may be extra-Vedic. We find this basic principle articulated at the very beginning of the foundational text of Mīmāṃsā: “Dharma is something beneficial disclosed by a Vedic injunction”—codanālakṣaṇo ‘rtho dharmaḥ (PMS 1.1.2). They also take for granted that smṛtis are also a legitimate epistemic source of dharma . Even though they also accept proper conduct as a third epistemic source, their discussion focuses primarily on smṛtis and their relationship to the Veda.
The two statements—Veda is the sole epistemic source of dharma , and smṛtis constitute a valid epistemic source of dharma —create a serious theoretical and theological problem for our authors. The two, prima facie, appear to be contradictory. The easy solution is to ditch the second proposition and affirm unambiguously that the Veda is not only the primary but also the sole epistemic source of dharma ; what is outside of the Veda, what is not found in the Veda, is not and cannot be dharma . This extreme conservative position is held by a hypothetical opponent (pūrvapakṣa ) whose arguments are presented in PMS 1.3.1 and by all our commentators. He could possibly be just a straw man, a foil used by our authors to present and then rebut his arguments and thereby buttress their own positions. Yet, I think this hypothetical opponent probably represented the real views of a segment of the thinkers in the Mīmāṃsā tradition and, perhaps, even in Dharmaśāstra itself. I base this, among others, on Viśvarūpa’s commentary on Yājñavalkya (1.7), which reproduces an almost excessively long argument of the opponent that, in the printed text of the original Sanskrit, occupies over eight pages, including objections raised by the opponent’s opponents. However, this extreme position never established itself in the mainstream of either Mīmāṃsā or Dharmaśāstra.
So we are left with both horns of the dilemma: how can one hold on to both the Veda and smṛti as valid epistemic sources of dharma ? Two solutions, neither without serious problems and undesirable consequences, are offered: (I) The first, already proposed by Śabara (fifth century ce ), the commentator of the Mīmāsā Sūtras, posits that the Vedic texts extant today do not comprise the entire Veda; many texts have been irretrievably lost. It is the memory of the basic injunctions contained in these lost texts—not their exact verbal form but the gist of their content—that is preserved in the smṛtis . So, these latter texts are actually rooted in the Veda: vedamūla . This theory, therefore, posits two kinds of Vedic texts available to us. The first consist of texts actually recited in contemporary Vedic schools, and they are referred to by the technical tem “perceived Vedic texts” (pratyakṣaśruti ). The second, on the other hand, are Vedic texts whose existence must be inferred on the basis of injunctions given in smṛtis (and by extension in normative practice or ācāra ), and they are referred to as “inferred Vedic texts” (anumitaśruti ). 5 (II) Kumārila, writing a subcommentary on Śabara a couple of centuries later, is quite troubled by the implications of this theory. If the smṛtis are based on the Veda and thus authoritative, what is to prevent Buddhists and others from claiming a similar status to their own heterodox scriptures? To obviate such drawbacks inherent in the theory of a lost Veda, Kumārila proposes a novel solution to the problem of the connection between smṛtis and the Veda. He says that all smṛtis are based not on some hypothetical lost Vedas but on currently available Vedic texts. Then why can we not find them? Because, Kumārila contends, the Vedic branches that preserve their respective Vedic texts are scattered (viprakīrṇaśruti ) across the vast country of India and no single individual is able to collect and study them all at any given moment. To help people find out the entirety of the Vedic dharma contained in the texts of all these branches, the authors of smṛtis presented the content of those texts not verbatim but in their own words and in a logical order. Thus, Kumārila presents the novel proposition that there cannot be any contradiction between Vedic provisions and those of smṛtis. Thus, all injunctions found in smṛtis have an authority equal to that of explicit Vedic texts, and when two smṛtis or a smṛti and a Vedic text contradict each other, there is an option or, more likely, according to Kumārila, a simple inability on our part to understand the specific scope of each injunction.
There is a divide, however, between the scholars representing Mīmāṃsā and those belonging to the Dharmaśāstra tradition. Although the two scholarly enterprises are joined at the hip, their focus is different. Mīmāṃsā has a narrow focus in its preoccupation with the interpretation and correct performance of Vedic ritual. Dharmaśāstric scholars, on the other hand, have broader perspectives and interests: they have to deal with the real life situations of individuals and social groups, with differing customs and norms of different regions and groups, with court procedures and the resolution of disputes, and with the civil and criminal laws governing societies. Can one expect to find all these diverse laws in the Veda, which by definition is supra- historical and cannot be seen to engage in temporally or geographically specific issues?
This epistemological conundrum is often left without direct engagement or resolution, but Medhātithi, possibly the greatest jurist of ancient India, provides a forthright answer: not all of dharma is based on the Veda. In his comments on the duties of a king (MDh 7.1), he acknowledges that “the dharmas explained here have their roots in various epistemic sources; not all of them have the Veda as their root”—pramāṇāntaramūlā hy atra dharmā ucyante na sarve vedamūlāḥ . He presents five possible answers to the relationship of texts of recollection to the Veda and finds them all wanting. So, then, what solution does the great jurist present to this dilemma? Being one of the most refreshingly frank and honest scholars of the period, he is able to throw up his hands and admit defeat. His basic answer is that there must be some sort of a connection between the texts of recollection and the Veda, but we have no idea what that connection might be! He concludes enigmatically: “There is no authority, however, to specify the particulars, nor is it useful.” But after reviewing the five alternatives, he presents a clearer conclusion:
Therefore, there certainly exists a connection between the Veda and Manu and others with respect to this issue (dharma ). It is, however, impossible to determine the specific nature of that connection. When people who know the Veda have the doggedly resolute conviction that something must be carried out, then it is appropriate to assume that it is, indeed, rooted in the Veda and not rooted in something else, such as an error. In this way, an assumption comes to be made with respect to the cause that is in keeping with that conviction. 6
The third phase of the epistemology of dharma is represented by commentaries and legal digests produced in the second millennium ce . With rare exceptions, the deep interest in this topic exhibited by the scholars of the earlier period is absent among those of the second millennium. Most frequently they simply give the epistemic sources of dharma found in the Dharmaśāstras with minimal comment and do not engage seriously with the many theoretical problems raised by them. The intellectual milieu was probably different, and new issues probably came to dominate the conversation. That dharma was based on the Veda and smṛtis was taken as a given. Only Aparārka (commenting on YDh 1.7) takes seriously the epistemological issues in the context of the rising importance of Hindu sects, both Vaiṣṇava and Śaiva, and the prominence of their respective sacred texts. Buddhists and Buddhist texts that loomed large in the discussions of the scholars of the second phase are, for the most part, absent.
We noted earlier the threefold division of Hart’s secondary rules encompassing rules of recognition, change, and adjudication. I have dealt above with the first under the rubric of epistemology of law. Rules of change identify the legitimate ways in which existing laws can be modified and annulled, or new laws enacted. It is to this aspect of law and dharma that I now turn. Rules of change are, furthermore, integral to the epistemology of law; an individual or institution with the authority to enact laws will also have the authority to change and annul existing laws and to enact new ones. Without such an explicit agency for enacting laws, Indian epistemology of law needed to come up with hermeneutical principles to account for change.
Given the theory that law as dharma is derived from the Veda and that the Veda is eternal, it is theoretically impossible to (i) change any dharma , (ii) annul any dharma , or (iii) enact any new dharma . All change is theoretically foreclosed. But human societies and their customs and mores inevitably change, whatever the theoreticians of dharma may say. Changes in customary laws within various communities reflecting these societal changes are imperceptibly introduced when those laws are unwritten. As laws and customs change, the older ones are simply forgotten by new generations, creating an appearance of immemorial custom. The dharma articulated in smṛtis , on the contrary, was written down, and these treatises were studied, commented on, transcribed in new manuscripts, and handed down from one generation to another. How do you introduce change into such immutable and inscribed laws? Or better, how do you theoretically manage and justify any change that invariably occurs? This is the challenge that faced the jurists and their scholastic techniques.
An early strategy is employed by Āpastamba. He is concerned about the seeming immoral acts performed by ancient seers recorded in the Veda. In general, the practices of such great sages, just like the behavior of contemporary elite, would be a source of dharma and something to be emulated. How do you abrogate such a dharma and prevent people from following those ancient practices? Āpastamba, of course, cannot abrogate such exemplary activities of ancient sages, but he does the next best thing; he makes such examples inapplicable to his own time. “Transgression of dharma is seen, as also violence, among men of ancient times. They incurred no sin on account of their extraordinary power. A man of later times who, observing that, does the same, perishes.” 7 Implicit in Āpastamba’s argument is that times change and with it the capacities and strength of human beings. 8 At least by the time of Manu, that is the middle of the second century ce , the general argument of Āpastamba became incorporated into the doctrine of the four world ages (yuga ) that came to be applied to the functioning of dharma in society. As the lifespan, strength, and virtue of human beings decline in each subsequent world age, the dharma that govern their lives changes correspondingly. Manu enunciates this doctrine: “There is one set of dharmas in the Kṛta Age, another in the Tretā, still another in the Dvāpara, and a different set in Kali, in keeping with the progressive shortening taking place in each Age” (MDh 1.85). The Parāśara Smṛti (1.24), a text composed in the second half of the first millennium ce , goes so far as to limit the authority of different smṛtis to specific ages, Parasara’s own composition being the one most appropriate for the current age: “In the Kṛta Age the dharmas proclaimed by Manu are said to be operative, in Tretā those of Gautama, in Dvāpara those of Śaṅkha-Likhita, and in Kali those of Parāśara” (Par 1.24). The hermeneutical strategy based on the world ages permits jurists to relegate rules of dharma that they found objectionable to a previous world age and to make them inapplicable to contemporary times. For all intents and purposes, therefore, these rules are abrogated and rendered null and void.
The legal fictions created by jurists, both ancient and medieval, to introduce novelty and change into the de jure unchangeable and eternal dharma are instructive with respect to the scholastic enterprise within the Science of Dharma. They also demonstrate the singular importance of customary laws, mostly unwritten, within the edifice of Indian jurisprudence that is theoretically supposed to be derived from and based on the immutable Veda. As Lariviere (1997 ) and Wezler (2004 ) have argued, the idea of vedamūlatva of dharma was a theological construction. The historical reality at the beginnings of Dharmaśāstric composition, as during the medieval period when the Nibandhas were written, was that dharma of the Dharmaśāstras was very much anchored in the actual customary laws of various geographically and temporally dispersed communities.