Preliminary remarks

It would be an understatement to say that the history of varṇa, rendered sometimes as caste and sometimes as social estate, is poorly understood, and to date there is no real working model for its development. There is no clear notion of how varṇa began, what institutions preceded it, and what historical factors lead to its creation. Treatments like Brian K. Smith’s Classifying the Universe project varṇa deep into Vedic prehistory as though in ancient times this institution resembled the exact same one that appears in the later Dharmasūtras and Dharmaśāstras.Footnote ¹ The net effect is a papering over of the complex conceptual history of this institution. What then is the early history of varṇa? When and under whose influence was it rearticulated into a system of social estates in the classical period? That question is far too large for this article alone. However, we will begin to answer it by interrogating the semantic gulf between the Vedic terms of art and those found in the later Dharmasūtras.

In this article, we will destabilize the conceptual history of varṇa by showing one place where Vedic understandings of the institution cannot be reconciled with modern or classical ones: the terms rājanyà and kṣatríya. Authors like Brian K. Smith readily translate texts that mention the rājanyà with ‘Kshatriya’, presupposing the two terms have identical semantic scopes because that is evidently so in the Dharmasūtras. I will demonstrate, however, that while kṣatríya is an archaic term for a certain position of social power and leadership, the term rājanyà qualifies a certain kind of ritual subjectivity, one that was not a term of self-reference or identity at all.

Provincializing brahmanism?

Before proceeding any further, it is worth evaluating the role this article plays in provincializing the brāhmaṇa, given that the focus of my article is on the kṣatríya. In my view there are two ways to strip historical brāhmaṇas of their imagined sovereignty and stability.Footnote ² One way is to examine how traditions that frame themselves as rejecting the brahmanism of the Dharmaśāstras reclaim the figure of the brāhmaṇa and rearticulate this character in ways particular to a given time, place, and tradition. In so doing, the scholar demonstrates there is no immutable essence to the brāhmaṇa sine qua non. Another way of provincializing the brāhmaṇa, however, is to briefly abandon him. We do this by investigating the voices outside of brahmanism, not for their theorization of the brāhmaṇa, but for their construction of a system by, for, and about themselves. Although the Dharmasūtras and Dharmaśāstras frame themselves as the inheritors of the Vedic tradition, this is just an appeal to authority. What I hope to show in this article is that the Vedic terms like brāhmaṇa and kṣatríya signified something radically different than those same terms do in later varṇāśramadharma, suggesting a sharp break and a reinvention, and not an immutable inheritance.

The Vedas, of course, were composed by lineages of poets, hierophants, and ritual adepts for social performance and for their posterity. Often, these figures styled themselves as brāhmaṇa. Consider, for example:

Here we are to imagine a council of plants, with their leader Soma coming to an agreement to help whoever is the patron of the brāhmaṇá. Thus, a brāhmaṇá is someone who performs rituals, who interacts with plants, especially the soma plant, and acts on behalf of his patron. In that capacity, there is no reading of the Vedas that is outside of a brahmanism. The point here, however, is that there are many brahmanisms, and interrogating Vedic brahmanism will reveal a very different brahmanism than that of, for example, the brahmanism of the Dharmaśāstras. Excavating the Vedas represents, in some ways, the strongest criticism of the Dharmaśāstras, as it provides an alternative way of seeing the world from a source that the Dharma tradition has already recognized as even more authoritative than itself. Indeed, my hope is that studies of this kind will support the contention that Vedic brahmanism was not monolithic either, but itself a succession and diversity of competing brahmanisms, and its historically posterior iterations were a product of accidence not essence. There was nothing inevitable about Classical varṇa.

So, who then was the kṣatríya before varṇa came to be a system of social estates? Who was the rājanyà? Consider Tim Lubin’s articulation of the status of brāhmaṇa as something that could be obtained by the ritual observance of brahmacarya:

Nathan McGovern, however, argues that even after the urbanization of the Gangetic watershed, during the time when the Buddha and Mahāvīra putatively lived, this notion of brāhmaṇa as a product of brahmacarya was still the primary sense of the word. According to McGovern, the Dharmasūtras are the first to articulate a system of inherited varṇa to the exclusion of any other conceptualization of the institution.

McGovern contends not only that brāhmaṇa was a term contested by ascetics and householders, but that this contest was not decided in the sixth century bce and was still undecided centuries later.

Thus, following Lubin and McGovern, who understand a brāhmaṇa to be someone who had completed their brahmacarya, I will argue that the rājanyà was conceptualized as a candidate for the rite of rājasūya, and it is the rājasūya that transformed a rājanyà into a rā́jan. Further, I will argue there were two types of rā́jan: the kṣatríya and maghávan. Each office exercises its tenure over different seasons, the winter and summer respectively. Vedic ritual life was highly seasonal: even a generic term for priest, ṛtvij, is a compound that has ṛtu ‘season’ as its first member.Footnote ⁶ F. B. J. Kuiper argued that the goddess Uṣas praised in Vedic hymns was not just any dawn, but the dawn of New Year’s day, released from the Vala byre by Indra.Footnote ⁷ Similarly, Indra’s defeat of the constrictor Vṛtra was re-enacted on midsummer day. After the Rigvedic period, this seasonality of the ritual calendar becomes much more muted, and the texts often depict the ritual as containing the entire year. Hans-Peter Schmidt has pointed out that liturgical material associated with the New Year seems to have been compressed into the morning sóma pressing of classical ritual. Similarly, material associated with Indra’s midsummer defeat of Vr̥tra seems to have been compressed into the midday pressing.Footnote ⁸

In my model, the kṣatríya represents a real historical position of power. The rājanyà, conversely, would not represent anyone’s social identity or term of self-reference but rather would have been a term of art used exclusively among ritual specialists for a particular type of ritual candidate. The fact that rājanyà appears in varṇa lists in the early Vedic texts and never kṣatríya suggests that varṇa itself was a system of ritual subjectivity used exclusively among ritual specialists and not anyone’s term of self-reference or identity.

My method of analysis in this article derives from serious consideration of the Vedas performative context. They should not be read as documentary accounts of a fixed social reality. Rather, their content and form always serve their teleology, which is to be used in public ceremonies of political legitimation. The Vedas are oral corpora composed by poets and priests, memorized and transmitted by hierophantic lineages, and organized into four discrete traditions: the Rigveda, the Sāmaveda, the Yajurveda, and the Atharvaveda. The intellectual property of a Veda begins with its oldest text, the Saṁhitā, of which there are often several recensions. Two priests might belong to the same Veda but use different recensions of the Saṁhitā, and therefore belong to different śākhās or ‘branches’ of that Veda. The Saṁhitās contain the mantras which were variously recited, sung, or muttered during ritual performance. These texts were chronologically followed by exegetical commentaries (from oldest to youngest: Brāhmaṇas, Āraṇyakas, and Upaniṣads). Composed over time and slowly canonized, these texts were not used as viva voce mantras during the ritual. Rather they do the work of justifying ritual substitutions, exceptions, and innovations, among other things.

Vedic texts were used in large public rituals of social promotion during the Vedic period, the most important promotion being from clan-lord to overlord of an alliance of clans. We might call this figure the Vedic king, although that fails to capture the many ways in which Vedic kingship is radically different than, say, inherited monarchies. A major focus of the Vedas is building political consensus and legitimacy for a social hierarchy that favours the clients that patronize these ritual activities.Footnote ⁹ While pegging Vedic texts to a particular century is notoriously difficult, it is possible to identify which texts are generally older or younger relative to one another based on their grammatical archaisms. The earliest phase of the Vedic period is the Rigvedic period, when the sūktas ‘true spoken (hymns)’ were composed. The Mantra period is the period of time when the Rigvedic sūktas were collected and organized into an anthology,Footnote ¹⁰ and the mantras of the Black Yajurvedic Saṃhitās were composed. Some material from the tenth maṇḍala of the Rigveda, much of the poetic material of the Atharvavedic Saṃhitās, and the Rigvedic Khilāni were also composed during this period. At the close of the period, exegetical prose passages were being composed and inserted into the body of mantras in Black Yajurvedic Saṃhitās. The reason I have provided this background on the performative context of Vedic ceremony is because they are often read without this context in ways that have led to unwarranted generalizations about Vedic social structure.

Declassifying the universe

In his monograph Classifying the Universe, Brian K. Smith argues that the later tradition of the Dharmasūtra and Dharmaśāstras continues a social project already prefigured in the Vedas:

The influence of the Dharma traditions on Smith’s analysis is palpable.Footnote ¹² In his translation of the only verse in the Rigveda that mentions the rājanyà,Footnote ¹³ he seamlessly replaces the term with kṣatríya without informing his reader. In fact, he does this with every citation of rājanyà. Are these two words really synonyms? They certainly are in the Dharma literature. Are they synonyms in the Vedas? If they were, why is it that only rājanyà appears in lists of varṇas in the earliest Vedic texts and never kṣatríya? The term kṣatríya is evidently of Indo-Iranian provenance,Footnote ¹⁴ and attested nine times in the Rigveda. So, while it stands to reason that it was available as a varṇa term of art, it is never used in that way in the early material. The term rājanyà appears once in the Rigveda in the same place that the terms śūdra and vaíśya make their debut—the enumeration of types in the famous Puruṣasūkta.Footnote ¹⁵ The only term that had history prior to the list of terms is brāhmaná. We must consider the possibility that rājanyà, vaiśya, and śūdra may have been coined by the poet of the Puruṣasūkta. In short, because kṣatríya is never identified as a varṇa in the Saṃhitās, while rājanya seems to begin its life as an item in a list, I argue that the two terms are neither a synonym nor do they even belong to the same conceptual spheres.

In his argument that ‘the caste system found in the Dharma Sūtras and Śāstras is nothing but a continuation and reworking of the earlier social theory inscribed in the Veda’,Footnote ¹⁶ Smith naively reads relations of social dependence from the texts. What this ‘dependence’ signifies for Smith, is a social class and a state of servitude. However, the passages Smith refers to do not support this reading when we take the ritual context into consideration: ‘There are three other sorts of men besides the Kshatriya—the Brahmin, Vaishya, and Shūdra. He thus makes them subordinate (anuka) to him.’Footnote ¹⁷ Smith reads this as the kṣatríya making these three figures dependent on him, the kṣatríya. He uses this passage to single out the śūdra as existing in a state of servitude, ignoring that whatever state the śūdra finds himself in, it is shared by brāhmaṇa and vaiśya as well. Reading this as describing a historical social relationship is one possibility but uncompelling. Another interpretation might come from asking ‘what does the Taittirīya Saṁhitā, a ritual text whose contents were performed on the ritual ground, accomplish by asserting this is so?’ To answer this question, the passage must be read in its broader context. First, I shall supply the default subject of the verb, which is usually the adhvaryu, the chief Yajurvedin, or the yajamāna, the patron of the whole affair. The yajamāna is really ‘the one being celebrated’, and it refers to the patron of the sacrifice only from his dīkṣā ‘consecration’ up until the avabhṛtha rite washes away his consecration and thus the ceremony. One is only a yajamāna for this temporary period, and it stands to reason that, if yajamāna is the implied subject, then we cannot assume that being anuka endures after the ceremony is over either.

Why should we understand that it is the yajamāna being tasked to recitation here? Because the previous passage ends with a bit of dialogue—the adhvaryu saying sādhú te yajamāna devátéty ‘(May the) divinity be successful for you, Yajamāna!’ who says in the next section ghr̥távatīm adhvaryo srúcam ā́syasvéty ‘Adhvaryu! Throw in a ghee-filled spoon(full)!’Footnote ¹⁸—it stands to reason, then, that this section occurs in the context of a verbal exchange between adhvaryu and yajamāna. Why should we take the yajamāna as the subject and not the adhvaryu? Precisely because of the verb ánu brūyād ‘he should say after [=repeat]’. Indeed, the speech of the adhvaryu is characterized by just the base verb brūyā́t ‘he should say’. This makes perfect sense, as the adhvaryu, as a ritual professional, would be in a better position to take the lead than the yajamāna.

Nowhere do I see a state of servant status depicted. So then, what does this passage really tell us? We cannot ignore the stylized repetition of anu ‘following, after, along’ in ánu brūyād ‘he should repeat’. Only after repeating, or ‘speaking after another’, do these three become ánukas ‘after-ones’ or followers. Thus, the image of mimed recitation, a speaking that follows a previous speaking, becomes the model for a person following a previous person. It is ritual speech, then, that becomes the basis for imagining relations of subordination. Further, we have a clear image of exactly what a rājanyà is—someone who is indriyakāma ‘desiring Indra-ness’—and the hint that this is somehow connected to making others follow your speech.

The key here is that someone who desires to be Indra is not an Indra yet, and we will explore what that means a little later on. This indriyakāma desire is important, as it helps us understand the rājanyà as not only someone passively worthy of becoming rā́jan but someone who will actively patronize the ritual because of that desire.Footnote ²⁰ After all, the importance of desire (kāma) is of perennial importance from the RigvedaFootnote ²¹ all the way to the Manusmṛti.Footnote ²²

A rājanyà is not Indra, the divine leader of the clans of the gods, but a potential leader, an aspiring leader, a pretender who desires Indra-ness. This simple gloss of rājanyà as someone who is indriyakāma and who might achieve Indra-ness through ritual recitation makes it clear that this is both a ritual subject and a situational one, one that would presumably precede the attainment of the office of rā́jan. The nuance of this reading is erased if we incuriously substitute kṣatríya for rājanyà.

Excavating the Rigvedic kṣatríya

How is the kṣatríya characterized in the Rigveda? Sometimes we see a relationship between Agni and the kṣatríya:

From this we can tell that kṣatríya are sponsors of Vedic sacrifice: they pour oblations in Agni, the god of fire from whom they receive gifts. We could say as much about any depiction of Vedic elites.Footnote ²⁶ Notice, however, the bitextuality of the second example above, Rigveda 10.66.8, for this could both refer to human patrons who overcome resistance to their ritual consecration and released the waters of their unction as well as offering oblations into the fire.Footnote ²⁷ Yet it is also true of the gods, who overcame the primordial constrictor Vr̥tra, releasing the pent-up waters for mankind, whose hotar priest is the god Agni. Thus, we do not see a clearcut picture of mortal kṣatríyas or divine ones; rather, the verse itself could be said to be mingling them together at the ceremonies (adhvarā́ṇām abhiśríyaḥ). In the following passages, however, we see the term kṣatríya explicitly connected to Mitra and Varuṇa.Footnote ²⁸

Whereas we previously saw the kṣatríya having a relationship with Agni, here we see something quite different. Mitra and Varuṇa are identified as kṣatríyas themselves, a status that seems exclusive to them among the gods. If kṣatríya represented a militaristic social estate, you would expect Indra and the Maruts to be dubbed kṣatríyas at least once, but this never happens in the Rigveda. Instead, we see that Indra may be the foe of the kṣatríya:

This presentation of Indra as scourge of the kṣatríya, when paired with the fact that he is never himself identified as a kṣatríya, is particularly interesting. Scholars have observed that Varuṇa and Indra seem to be associated with distinct forms of political power, saṃrāj and svarāj respectively.Footnote ³⁷ To this observation, I would argue that kṣatríya and maghávan denote two types of rā́jan, the former associated with Varuṇa and the latter with Indra. Further, I will argue that these terms were associated with seasonal leadership of the coalition of the Vedic clans.

In order to make sense of this claim, some comment on Vedic seasonality is appropriate here. As mentioned above, the hymns of the Rigveda were composed by semi-nomadic pastoral clans for use at events throughout their ritual calendar, which seems to have been heavily influenced by seasonal changes in their lifestyle and means of subsistence. In the summertime the clans dispersed to take advantage of pasturage in higher altitudes, where fields were fed by the myriad rivers that sprang to life from melting snow. Marianna Ferrara has persuasively argued that, in the context of a highly mobile society, the term yoga ‘yoking’ referred to the period in preparation for battle, and that kṣema ‘peace’ was the period of the time after battle, and the later semantic of yoga as ‘effort’ most likely derives from this sense.Footnote ³⁸ While I agree that these terms refer to periods of time, I follow Thomas Oberlies in noting that the activities of war and peace were highly seasonal.Footnote ³⁹

I argue that a yoga did not refer to any time before a battle, but to a period marked by increased mobility, isolation, and sporadic conflict, thus the season of yoga, ‘trekking (season)’, really referred to the summertime. In the winter, we find the exact opposite. The clans gathered for a more sedentary mode of subsistence at lower altitudes; they lived in closer proximity and most likely shared their resources during the lean times of little fodder and dried-up riverbeds. Rather than kṣema marking the moment of unyoking immediately after a battle, I argue that it was an epithet for the season marked by decreased mobility and lower inter-clan hostility. Thus, kṣema, the ‘resting (season)’, referred to winter.Footnote ⁴⁰ Let us consider the words of Indra himself as translated by Jamison and Brereton:

I hold up Jamison and Brereton as the scholarly standard in order to show clearly where my reading diverges. While I will not unpack every element in my own translation here, I will leave certain terms untranslated in order to discuss them in greater detail:

In both translations Indra claims to be the recipient of hospitality and promises to destroy the stingy. What I believe this article offers the translation is that it illustrates how the poet subverts the audience’s knowledge of seasonal modes of behaviour. If the maghávan is really a summertime rā́jan, the period of greater autonomy and mobility, we would expect unknown maghávans to be especially hostile, not hospitable. Next, if instead of reading kṣema as ‘peace’, we read it as ‘wintertime’, then we have the opposite expectation. Wintertime is a period of shared resources and communal habitation, yet Indra so despises stinginessFootnote ⁴⁶ that he would attack the wintertime leader, who, I am arguing, is the kṣatríya, even during this period associated with rest.Footnote ⁴⁷

That means that over the course of a year, the clans shifted from a mode of abundant resources and ample opportunity to go on cattle raids, to a far narrower lifestyle, when the consequences of inter-clan feuding might be acutely felt.Footnote ⁴⁸ For a segmentary society,Footnote ⁴⁹ Vedic ceremony may have been a tool to resolve these potential conflicts. The ritual is a public moment in which society could be reintegrated and subordinated under a leader, avoiding future hostilities between the allied clans through elaborate gift-exchange hospitality ceremonies. The patron of the Rigvedic soma pressing was called a sūri ‘impeller’. The preference for the term sūri begins to ebb in the later portions of the Rigveda, where it is ultimately replaced by the word yajamāna, although it remains important as a Jain honorific. Important here is that sūri and yajamāna were not social estates, they qualified someone as the dedicand of the ritual at hand and nothing more. Off the ritual ground, this patron was most likely a charismatic clan-patriarch (viśpati) who hoped to be ceremonially elected by the other clan-patriarchs of the alliance (in their ritual capacity as ratnins ‘prize-holders’) to the position of acting secretary-general of a coalition of the clans. Just how briefly this appointment lasted is a matter of much debate. Kuiper believed that the position of maghávan, ‘lord of the gift ceremony’, was not held for the full year,Footnote ⁵⁰ and indeed many communities in human history have operated with just these kinds of seasonal political arrangements.Footnote ⁵¹ Brereton, however, sees the tenure of the Rigvedic king as a multi-annual affair and potentially a hereditary office.Footnote ⁵² It is difficult to be certain, but it seems there were different kinds of political power operating simultaneously. I would speculate that while being a clan-patriarch (viśpati) was indeed a multi-year and, most likely, hereditary status, the higher position of rā́jan (and therefore both its summer office of maghávan and its winter office of kṣatríya) was not. The reason is that the divinized ideal of this figure, Indra, is largely free of mythological kinship relations. Indra’s mother is mentioned in passing in the Rigveda,Footnote ⁵³ we have only hints that he even has a father,Footnote ⁵⁴ and he has no children of his own. That would make sense if the human who occupied the position of maghávan or rā́jan neither inherited nor bequeathed his status. Instead, Indra gains his power through his cosmogonic deeds; for instance, releasing the Dawn from the Vala byre at New Year and vanquishing Vṛtra and releasing the waters at midsummer just as melting snow replenishes the desiccated riverbeds.

Were clan coalitions as unstable as they appear to be in the Rigveda? Did these alliances really need to be (re)negotiated annually? Or even seasonally? Or is this hyperbolic rhetoric from poets that admittedly adore agonistic, martial, and hypermasculine language? Are we getting the wrong historical impression because of the poetic style? I am open to the possibility, but I favour the model of seasonal kingship. Support for temporary rather than career suzerainty can be found in the rarity of royal panegyric that mentions a leader by name,Footnote ⁵⁵ which we might expect would have been produced in praise of leaders who ruled for decades. That is not to say that something looking like royal panegyric does not exist, but that rather than a human leader being glorified, the dedicand is Indra. We might imagine that poets praising Indra, as the model maghávan, might be indirectly praising any human maghávan, no matter how brief his tenure, who lives up to Indra’s standard at that moment, while also implicitly critiquing the also-rans who do not.Footnote ⁵⁶ There are exceptions to these norms, like Sudās and Trasadasyu, but there is little to support their historicity,Footnote ⁵⁷ so the situation is more likely a poet invoking the memory of these figures rather than producing encomia for them.

If we turn to Trasadasyu, however, we have an especially interesting case for our study. Rigveda 4.42 is recited from the perspective of Trasadasyu,Footnote ⁵⁸ and it is the only place in the Rigveda where a figure claims to be both a kṣatríya and a maghávan. This hymn seems to close a subcollection of hymns: Rigveda 4.38–40 laud Dadhikrā, the horse of Trasadasyu; Rigveda 4.41 is a praise-song soliciting the aid of Indra and Varuṇa; and, finally, Rigveda 4.42 constitutes Trasadasyu’s impersonation of Varuṇa (four verses) and Indra (four verses) and two verses that relate the legend of Trasadasyu in the third person.

It seems likely that this represents the earliest form of what would eventually become the aśvamedha or ‘horse sacrifice’ in the later period.Footnote ⁵⁹ The sacrificial horse would have been lauded as no ordinary horse, but as actually the horse of Trasadasyu.Footnote ⁶⁰ In so identifying this horse as Dadhikrā, the liturgy suggests that the patron of the sacrifice is actually Trasadasyu himself. This suggestion builds when Indra and Varuṇa are invoked to cooperatively aid the patron,Footnote ⁶¹ and then the climax of the performance is the poet impersonating Trasadasyu, who himself impersonates Varuṇa and Indra. The relevant scholarship on this hymn is aptly summarized by Jamison and Brereton:Footnote ⁶²

It is in the opening verse, in which Trasadasyu insists he is Varuṇa, that our term of art appears:

Notice, however, that when the speaker claims he is kṣatríya, it is while he is also claiming he is Varuṇa. Only when he begins his impersonation of Indra, in verse five, does he claim the title of maghávan as well:

I suggest that this is the wordplay of máma dvitā́ rāṣṭaráṃ ‘mine is twofold rule’. Unlike other rā́jans, Trasadasyu rules both as winter kṣatríya and summer maghvan.

I must stress that belonging to book four of the Rigveda suggests these hymns are very old and reflect very early Vedic ritual and poetic sensibilities.Footnote ⁶⁷ That may be a serious problem for my hypothesis, which conceives of leadership positions in the Vedic coalition as being seasonal to reflect the older state of affairs, with annual leadership coming later. It is certainly possible that seasonal leadership was merely a poetic ideal, and that annual and or even perennial leadership was the norm. To defend against that possibility, I want to point out that our scant hagiographical details suggest that Trasadasyu was anything but normal: the seven seers directly intervened to bring about his birth.Footnote ⁶⁸ It is also the case that horse sacrifices were atypical and rare among Vedic animal sacrifices, which usually involved the killing, cooking, and eating of goats or steer.

One would have performed the classical aśvamedha not in order to become leader of one’s coalition, which was a precondition of undertaking the rite, but to become something like an overlord (samrāj), accepting the submission and tribute of neighbouring coalitions. It is the ancestor of the digvijaya ‘conquest of the directions’, known from the Arthaśāstra, which presumably resulted in a king becoming a cakravartin ‘wheel-turner’, something akin to an emperor. Therefore, we should not lose sight of the fact that during the Vedic period, horse sacrifices were perhaps the most infrequently performed rite and that rarity may have contributed to its constant redevelopment.Footnote ⁶⁹ What I am suggesting is that in Trasadasyu’s sacrifice of Dadhikrā we see the exception not the rule. Perhaps annual and perennial leadership might have been possible to obtain early in the Vedic period, but it was not the norm and so it was not conceived of as a unitary office but rather as two successive seasonal offices and, in the case of this version of the horse sacrifice, kṣatríya and then maghávan. This interpretation is indirectly supported by the length of time of the journey of the horse in the classical aśvamedha, which lasted one full year and perhaps signified a rite that granted its patron an exceptional double term in office.

Excavating the Mantra period kṣatríya

We can only speculate as to how historical coalitions of the clans in the Mantra period might have become more stable, more populous, or more economically differentiated than their forebears in the Rigvedic period. Certainly, the hierophantic tradition did. We define the Mantra period as one in which the crucial reformation of the Rigvedic tradition into a triple Vedic tradition occurred. Ritual speech acts, the eponymous mantras, were divided into three types: rc ‘verses’, sāman ‘melody’, and yajus ‘sacrificial (dedication)’. We take the saṃhitās of the Black Yajurveda, comprising yajus and exegetical prose concerning the yajus, as definitive Mantra period collections.Footnote ⁷⁰ The responsibility for each type of speech was assigned exclusively to a specially trained professional who committed the corpus of just this type of mantra to memory. The specialization of ritual labour ultimately produced the Rigvedic, Sāmavedic, and Yajurvedic traditions, which were equipped to roll out a new ritual programme with ever more complicated and expensive royal consecration rituals as well as new lower-tier rituals of social promotion seen as preconditions to undertaking more elaborate kingship rituals. The patron comes to be identified with the totality of the year, and the three soma pressings, a special ritual drink imbibed by the patron and priests, seem to metonymically represent what we think were formerly three different ritual events that occurred in the annual ceremonial calendar.

In other words, a new programme emerged which compressed the totality of the old system into a single event. That event seems to have been taken as a default model and priests innovated on the basis of that default. It is hard, given the realities of the compression of the calendar, to imagine leadership as limited to a season in the Mantra period. The following passage from the rājasūya, or consecration rite, is an illustrative example of this, as the new leader must maintain a certain ritual practice, the agnihotra, for the entirety of a year.

Here the newly consecrated rā́jan sends forth news of his accession through Vāc, cognate with Latin vox ‘voice’, the goddess of speech. This is likely because it is both sacred speech that transforms him into the rā́jan but also profane speech in the form of the news of his accession disseminated by his messengers that makes those not in attendance aware of his new status. This subsection of this article is really all about the relationship between Vāc and the kṣatríya.

Now, despite this ‘reformation’ of the ritual as practised in the Rigvedic period, there is no evidence that ritual life had become subject to strict orthopraxy in the Mantra period. By its end, the nascent order of Yajurvedins had composed prose footnotes to their lists of yajus. These footnotes articulate the authenticity of a particular ritual practice, the need for which itself suggests their newness. Innovations require new charter myths, while universally accepted conventions do not.Footnote ⁷² In summary, we see both numerous continuities and innovations from the period when the poems of the Rigveda were composed into that time when the ritual system was reorganized into one that made use of three discrete Vedas. This context will inform my reading of the texts of the Mantra period.Footnote ⁷³

Perhaps the most archaic of the Black Yajurvedic collections, the Maitrāyaṇi Saṃhitā, invokes the figure of the kṣatríya as a winner of the kṣatra. Here we see this in the context of divine waters of consecration. Notice these two similar passages differentiated by the short prose commentary that accompanies the latter case after each yajus.

Indeed, very similar passages are found in the Kaṭha Saṃhitā Footnote ⁷⁶ and the Taittirīya Saṃhitā,Footnote ⁷⁷ suggesting they go back to a putative Ur-Yajurveda.Footnote ⁷⁸ Similarly, we see the term associated with the goddess Vāc, who is characterized as a kṣatríyā.

Again, we see roughly the same yajus in the sister texts of the Kaṭha Saṃhitā Footnote ⁸⁰ and the Taittirīya Saṃhitā.Footnote ⁸¹ Again this suggests that the notion of Vāc as a female kṣatríyā goes back to the earliest phase of the Yajurvedic tradition.

This makes perfect sense. As we saw above, in the rājasūya, it is Vāc who is called upon to empower the speech of patron who says rā́jabhūm ‘I am become rā́jan’. In this rite, Vāc is identified as the river Sarasvatī, no doubt to connect its waters to the waters of consecration.

Vāc as kingmaker and thus herself a queen is supported in Rigveda 10.125.3, where the poet, in impersonation of Vāc, asserts aháṃ rā́ṣṭrī ‘I am the rule’, and that she is prathamā́ yajñíyānām ‘first of those connected to the ritual’.Footnote ⁸² Now when, two verses later Vāc says, ‘Whomever I choose (kāmáye),Footnote ⁸³ him do I make mighty’, she is referring to choosing her brahmán not her kṣatríya.Footnote ⁸⁴ Nevertheless, the language is not so different than what we see in the previous hymn, Rigveda 10.124.5, where Indra addresses Varuṇa telling him that if he wishes to keep his powers, he should choose Indra as his leader and become a part of his political order.Footnote ⁸⁵ Indra uses the same verb that Vāc uses (kāmáya-). Now, this verb kāmáya- is not that which appears in the above Black Yajurvedin passage; that verb is juṣate. So, while I am not claiming that we see a repeated term of art here, I do think that we see Vāc’s choice as a poetic representation of status obtained or enacted by means of sacred speech, be it brahmán or kṣatríya, just as it is through Vāc that the patron becomes a rā́jan in the rājasūya.

What is meant by calling Vāc ‘both-headed Aditi’? Recall that scholars have long argued that Indra and Varuṇa are associated with two different modes of political power in the Rigveda: svarāj ‘autonomous rule’ and saṃrāj ‘collective rule’, respectively.Footnote ⁸⁶ If my hypothesis is correct, then Vāc was a vital component in the consecration of both the winter kṣatríya and the summer maghávan. I would speculate then that ‘two-headed Aditi’ is perhaps meant to present Vāc as politically bifarious, in so far as sacred speech is used to consecrate both types of leader.Footnote ⁸⁷ Indeed, in the Puruṣasūkta, Vāc is not explicitly mentioned, but the creation of the puruṣa depends on virāj, which, as the name of meter may be a cypher for Vāc.Footnote ⁸⁸ If the puruṣa is the cypher for the yajamāna, then we see a clear equation whereby the political leader and political speech are inter-dependent. If so, then Vāc, as virāj, may be capturing both svarāj and saṃrāj, thereby resolving the tension between seasonal forms of political power in so far as they are both enacted by one goddess.

When does kṣatríya become a varṇa?

I hope it has become clear at this point that the term varṇa and the term kṣatríya are both moving targets. Both institutions are changing, and it is difficult to pinpoint when first the notion of kṣatríya is united with the notion of varṇa as is the case in the Dharmasūtras. I have argued that kṣatríya referred to a political office, the winter king, as it were, an office obtained by both political and ritual means. I have argued that because the word kṣatríya never appears with the word varṇa or as a part of a list of varṇas in the early and middle Vedic periods, that kṣatríya was not considered a varṇa. I have not, in this article, produced an exhaustive study of the term rājanyà in the way I have kṣatríya. While that is a desideratum for a future publication, the premise of this article is that the total non-occurrence of kṣatríya in the varṇa list and the fact that it is never a substitute for rājanyà is sufficient for my argument. No parallels exist in the Saṁhitās where one text has rājanyà, the other kṣatríya, and the rest is identical. The evidence marshalled in this article should be sufficient to establish the two are distinct categories. While it remains to be determined exactly what varṇa is and when and how the term kṣatríya first came to replace rājanyà, I would like to proffer one later Vedic passage that is quite instructive. I would like to translate it here fully, as it gives us insight into the late Vedic institution of varṇa as a whole. This passage in the Aitareya Brāhmaṇa begins when Rāma Mārgaveya is explaining to the rā́jan Viśvaṃtara that:

Right away, this sounds strange, because drinking soma is presumably how one imitates Indra: it is his characteristic accessory alongside his vajra. Of course, clever Rāma Mārgaveya is creating a problem to which he happens to have the solution: by which drink, if not soma, is one able to secure kṣatra?Footnote ⁹⁰ It is worth looking at how Martin Haug translates this very same passage over 150 years ago:

For Haug, it seems that any reference to kṣatra entails the reality of a kṣatríya social estate. Furthermore, Haug sees Indra as the head of the kṣatríya social estate despite what we have just seen: that Indra is never called a kṣatríya. So what is Rāma Mārgaveya’s solution?

This is a curious passage, one that I hope to explore again in the future. So much here is at odds with what we have seen in earlier Vedic texts. For example, the śūdra, presumably the lowest tier in the list of varṇas, is associated with water. Yet, we have seen the important association of water and consecration. Here Rāma Mārgaveya has rejected the three most common fluids of Vedic ritual—water, milk, and soma—as inappropriate for the kṣatríya, suggesting instead juice of plants which he later says grow in the land of Kurukṣetra.Footnote ⁹⁴

Aside from revising the sacrificial menu, this text is the first that seems to provide definitions for three of the terms associated with the Classical varṇa system. This text tells us that the brāhmaṇa, vaiśya, and śūdra are characterized by their state of dependency. The brāhmaṇa gets gifts and invitations, but he can also be disinvited. In less glowing terms, the same is said of the vaiśya and śūdra. For a kṣatríya used to independence, this state of dependence is evidently bad. Is this patient zero for the injection of the kṣatríya into the varṇa list? Technically, we still do not have kṣatríya as an item in a list, but this is the closest juxtaposition of the term with the varṇas. We are still a far cry from the varṇa of the Dharmasūtras, for these labels are not inherited. Rather, they are produced by ritual and more specifically, as Rāma Mārgaveya tells us, by ritual error. Should a kṣatríya drink the wrong ritual drink, he would inculcate a desire in his progeny to become one of these types who will ultimately lack independence and autonomy. Rāma Mārgaveya is not telling Viśvaṃtara out of the goodness of his heart, but rather because he himself is dependent and needs the patronage that Viśvaṃtara can provide. This is his Vedic ‘job talk’, where he establishes how deep his understanding of the ritual and how essential his advice is. So, while we do not yet have a system of inherited and immutable social estates here, this seems to be the first place where the term kṣatríya, still a title meaning ‘possessor of the kṣatra’ and thus signifying the office of ruler, is juxtaposed with three types of ritual subjectivity, inviting the possibility that kṣatríya is a ritual subject as well.