II. Books and Performances, Books for Performance

5. Note to Self: What Marathi Kirtankars’ Notebooks Suggest about Literacy, Performance, and the Travelling Performer in Pre-Colonial Maharashtra

Christian Lee Novetzke

In modern scholarship on non-modern, and especially non-Western, subjects, we tend to treat the dialectic between orality and literacy as if all theoretical and historical questions in this regard have been answered. After several decades of reconstruction around the problems posed by postcolonialism and postmodernism, we still have a tenacious hold on the idea that whatever is literate or literary is more highly valued than whatever is (“simply”) oral. If a thing is written down, it must be important; and everywhere it is the important and powerful (kings, for example) who first write things down. It is perhaps one of the unfortunate side-effects of the interventions of postmodernism, particularly of Foucauldian thought, that we see power as the essential category of all critique: find the powerful and you will find the essence of a thing or time. However, I have argued elsewhere that we should abandon the distinction between orality and literacy and instead explore the relationship between performance and permanence—and in the process try to avoid the power relations and modern conceits tied up with the idea of literacy.1 In this essay I make a different argument. I show how a particular logic of practice within an ostensibly written tradition tends to emphasise orality and especially performance, and how this practice institutes a set of formal conventions that resist the static encroachment of literacy, conventions that invite the creativity of live performance. This is an essay that explores how to read a text that resists its own literacy by emphasising its secondary position in relation to performance.

It is the paradox of using texts to study practices that were essentially oral that traces of writing endure where collective memory fades, and we are often left with the impression that cultures which merely used literacy were dominated by that usage. This is particularly true of the plethora of regional, public, performative religious expressive modes that have come under the rubric of bhakti or “devotionalism” more generally. In these contexts we often trace textual remnants to refashion a literary corpus out of what must have been a vaster oral context. This essay shows how the Varkari religious practice of Maharashtra prompts us to develop a critical engagement with literacy and orality mediated through a theory of performance in order to establish a theory of literacy that is subjected to the demands of performance. This prompt is inchoate in the very written texts that form the literary archive of the Marathi Varkari sants that guide us in this direction, in particular the mode of writing used by a specific kind of religious public performer in Maharashtra, a kirtankar. Through this mode of writing I’ll contend that literacy here is subservient to performance, and that writing is seen as a tool, a carefully contained tool, in the practice of performance.

Public memory in India recalls that in the thirteenth century a precocious teenager, the Marathi sant Jnaneshwar, placed his final touches on his paraphrase and commentary in Marathi on the Bhāgavadgītā. Indeed, the main text attributed to him, popularly known as the Jñāneśvarī, is so important that it earned its own stamp issued by the Government of India seven years before its author’s biographical stamp was issued—the text precedes the author, a testament to the modern power of the written word (see Figure 5.1). In this time, which witnessed a plethora of sants, most of them are remembered to have been literate. In the sixteenth century, we have the sant Eknath who is remembered as not only literate but a textual critic who culled together all available versions of the Jñāneśvarī and created its first critical edition. Perhaps Maharashtra’s most famous sant, Tukaram, of the seventeenth century, is said to have written his own songs down by hand, a manuscript under glass in the town of Dehu near Pune. He also had a duplicate of his manuscript produced by his tabla player, Jaganade, a meticulous back-up copy. Observing this legacy, one might conclude that literacy, and indeed a proto-philology of a sort, was of primary importance to the Varkari tradition in Maharashtra for well over four hundred years.

Fig. 5.1 (L) Jnaneshwari Stamp, issued in 1990 to commemorate the 700th anniversary of the composition of the Jñāneśvarī; (R) “Saint Dnyaneshwar” stamp, issued in 1997 in memory of Jnaneswhar/Jnandev. Public Domain.

In fact, very little of the dissemination of Varkari materials is through writing. The most common way to consume the songs of the sants of Maharashtra is through a live performance, called kirtan. Though kirtan is a pan-Indian performance art, it as different in its many articulations as the traditions that use it. In Maharashtra, a kirtan is essentially a didactic public performance. This can be religious or not—a common kirtan subject is family planning and AIDS prevention, for example. But the focus of this paper is in the religious kind of kirtan in which the performer chooses a devotional song attributed to a sant and provides a good deal of exposition. This kind of kirtan is sort of a religious variety show, usually with a clear moral or political theme. There is often a story from the life of a sant that goes along with the song, and usually other songs or texts are brought in that can range from Sanskritic philosophy to sufi mysticism to the wisdom of political leaders and popular adages and sayings of unknown provenance in any language, including English. Usually a core text, an abhang or another type of devotional song, is interlaced with other scripted material, like the songs of other sants, as well as non-scripted, improvised material, and this all alternates between song, didactic narrative, and general storytelling and humour, with musical accompaniment. There is almost always only one lead performer, a kirtankar, and sometimes musicians are attached to the kirtankar forming a kind of troupe.

Depending on the kind of kirtan being performed, the kirtankar stands and the audience sits (e.g. Eknathi-kirtan, Rashtriya-kirtan, Naradiya-kirtan), usually segregated by gender. In some cases everyone stands and dances (e.g. Varkari-kirtan). Most often, however, kirtan is a seated affair for all but the kirtankar, and generally this marks what I would call a didactic rather than ecstatic kind of kirtan performance. An audience participates in many ways, singing along with the songs, finishing well-known verses along with the kirtankar, sometimes interacting with the kirtankar, and so on. Indeed, the audience is there to be won, to be engaged, and this is readily apparent when, at the end of a “sit-down” kirtan, the audience will present the kirtankar with gifts, often of cash (though the kirtan performance itself is either done for free or for a fee paid by a patron). Though kirtankars can be of any caste,2 they follow in the footsteps of a shudra performer, Namdev, a sant who is said to have flourished in the fourteenth century.

The origins of Marathi kirtan are traced to Namdev, who is remembered to have been born in Pandharpur to a low caste tailor or shimpi of the shudra varna, though his varna designation is often disputed, especially in northern India. Though born in Maharashtra and speaking Marathi as a mother tongue, he is remembered as having travelled all over India performing kirtan, and thus we find his songs in the Dadupanth, Kabir Panth, Sikhism, and so on. His role as the epicentre of kirtan in Marathi contexts is constantly reinforced. An illustration from a nineteenth-century publication of the Marathi hagiographer Mahipati’s eighteenth-century text shows Namdev performing a kirtan accompanied by Krishna (see Figure 5.2). While Narada, the maverick sage, is considered the first kirtankar, his quasi-divine status makes Namdev its first fully mortal practitioner, and of course the first to perform kirtan in Marathi. But Namdev’s legacy as a kirtankar is primarily within Maharashtra, and the kind of kirtan he is said to have performed is different from Sikh kirtan or the kirtan of most other regions and religions in India.3

Fig. 5.2 Namdev Performing a Kirtan, folio from a nineteenth-century publication of Mahipati’s eighteenth-century biography [1890]. Public Domain.

Namdev’s association with kirtan must be juxtaposed to his attributed position on literacy. As mentioned above, almost all the other Marathi Varkari sants of his ilk could write, and in typical iconographic depictions they are usually displayed seated before manuscripts. Namdev, by contrast, is never seated before a book. Indeed, he is almost never depicted as seated in Marathi contexts—he’s always dancing, singing, and in thrall to the moment of kirtan. In other words, he is always performing publicly, never writing privately. Indeed, various songs attributed to him concur that he disdained books and writing, and considered only the public performance of kirtan to be an appropriate way to express the sentiments of bhakti. Interestingly, most references to Namdev performing kirtan occur during the death ceremonies, the samadhis, of his friends and family, as if in the face of death Namdev is insisting on the immediacy and survivability of performance.

As far as I know, Namdev’s songs never describe him as writing anything, nor do the songs of his companions suggest that he wrote his verses down. When the verb “to write“ appears in songs attributed to Namdev it usually refers to the “writing of fate”. Even when a contemporary of Namdev is linked to writing, Namdev is neatly set apart, as in this song by Janabai, his so-called “maid”, who is attributed the verse, “Nivritti wrote the words that Vitthal placed on his tongue […] and Namdev expanded the practice of kirtan”.4 What is far more common in the songs of Namdev and his contemporaries is speaking and listening to the speech of others. As for books, Namdev suggests that “the tongue makes a good book (pothi)”, and states in a song’s refrain:

I don’t know Veda or Purana.
I’ve ignored paper and books.
Without verses like pearl-white clouds
I have only the spotless gem of my mind.5

As a non-literate performer, Namdev is an apparent anomaly in the context of Varkari sants. But these two aspects of his remembrance—being the Marathi originator of kirtan and presenting a disdain for writing as a bhakti mode of expression—establish Namdev as the archetypical kirtan performer in Maharashtra. Indeed, he comes to embody the ideal conception of what a kirtankar does—he provides, in a sense, the kirtankar’s theory of practice. This is a theory that not only places orality at the centre of bhakti practice, but requires that orality be public, something open to all in the kirtan. And this theory casts some suspicion on writing, insisting instead that the antidote to death and decay—to the kinds of loss to which the written archive also addresses its preservative powers—is not appropriate for bhakti. In Namdev’s theory of practice, bhakti is on display, through the body and voice, and never on paper. And as the archetypical kirtankar, this theory of practice has been inherited, for many centuries, by those who revere him as the originator of their art. If this contention is correct, that despite the rather impressive rate of literacy among the Marathi sants the logic of the practice of Varkari religious expression has been oral, or rather performative, and not written, is there a way in which we can see it inscribed in the written record, and if so, how?

The oldest written records of Namdev’s songs in Marathi date from the early sixteenth century, though most material is from the seventeenth and early eighteenth centuries. These literary remnants appear not in organised compendia of any sort, but in small notebooks, called badas or vahis, that were compiled and carried by kirtankars (see Figure 5.3). Over the years, those who have handled these notebooks have often referred to them as “kirtan akhyane” or the “moral stories for performing kirtan”. In other words, they are remembered as resources for the composition of kirtan, objects that serve performance; they were not meant to displace the human voice and body. These notebooks are distinct in terms of orthography, content, etc., from manuscripts as such, which are often called pothi in Marathi, signifying a different class of document altogether. Badas, by contrast, are truly notebooks, loosely organised and often hastily constructed with lots of margin corrections, lines crossed out, and other emendations. The material used to make these notebooks was not meant to weather the ages, but to hold for a short time the notes and jottings of kirtankars. The use and status of badas is far different from the carefully copied manuscripts of Sanskrit and Persian literature, or even those manuscripts associated with most of the other Marathi sants. Major archival institutions—such as university libraries, research centres, and other institutes—amid their many collections of pothis do not generally collect badas. Even the concern with which these notebooks are preserved today differs from the careful attention given to manuscripts—badas are alien to the teak-wood glass case, rarely have an index or catalogue citations, and generally rest in haphazard stacks or even piles in the closets of institutions and private collectors.

Fig. 5.3 Four typical badas or “notebooks” in the collection of the Bhandarkar Oriental Research Institute. Author’s photograph, CC BY.

Most notebooks are a jumble of information far beyond the enumeration of songs attributed to famous sant-singers. Although some badas—perhaps one tenth of the ones I examined—contain only songs, the majority are filled with an array of information, and this varying content in the badas speaks to the professional lives of their former owners. The badas contain all kinds of mundane information: astrological charts; notes on crop prices, geography, weather conditions; surveys of general news. I have found in badas love potions and bawdy songs called lavani. These materials indicate the diverse applications of the kirtan profession, which required a kirtankar to be a peripatetic, multi-tasking jack-of-all-trades, carrying information about neighbouring villages or entirely different regions of the subcontinent. Kirtankars kept track of births and deaths, changes in commerce, politics, and the lives of famous personalities. A kirtankar was part journalist, part foreign correspondent, part actor, part scholar, and part religious commentator, all in the context of the kirtan performance. Records of royal patronage to kirtankars make clear that monetary rewards reflected the entertainment value of a performance—the kirtankar had to sing for his supper. The best records of the economy of this and other kinds of live performance are from the Peshwa period, which obsessively documented the daily life of the Peshwa in Pune and in other areas of the Maratha Confederacy throughout the eighteenth century. Here it is clear that the performers of kirtan and the folk variety show, often bawdy in nature, called tamasha, were the same people. There was no line between “religious” and “secular” performance—a market economy of public reception governed the strategies of performers. However, for the professional performer appearances at the Pune darbar were few. The regular terrain of the kirtankar consisted of the many village centres, pilgrimage networks, and holy sites that dotted the Deccan. In these locations they performed the sacred stories, biographies, and songs of those figures that had ascended from ordinary life to hagiographical stardom, and they mixed in political parody, racy love songs, and the buffoonery of the vidushak, often portrayed as a bumbling, Sanskrit-mumbling Brahmin.

Clearly, this is a lot to scribble down in a small notebook as you’re moving from town to town in the tumult of the late medieval Deccan. So the written record of this heterogeneous performance tradition, the bada, is itself a peculiar kind of writing that had, as a matter of necessity, to serve orality and performance. No one wanted to read a kirtan or an abhang, after all—they wanted to see it, hear it, and experience it displayed before them. The kirtankar, summoning the collective memory of a sant and his or her verses, could mediate a public experience of bhakti for devotees. Still, kirtankars kept records, and those records collectively give us the literary archive of the Marathi Varkari sants (and many outside the Varkari fold).

When I first started investigating these old Marathi notebooks to search out Namdev songs, I noticed a peculiar pattern. I would find songs grouped together by theme or purported author, which in itself was not peculiar. But the songs within such a cluster would be intersected by the songs of another sant. For example, notebook number 5 in the collection of the Bhandarkar Oriental Research Institute, dating probably from the late seventeenth century, begins with a few verses by Namdev about the death of Jnaneshwar, part of a series of songs called the Samādhi, a eulogy for Jnaneshwar. As mentioned, such moments around death were usually celebrated with a kirtan, and here we have an abhang by Namdev that is to be understood as being performed by Namdev at this moment diagetically in the Samadhi narrative sequence—so it is a song within a song, so to speak. The kirtan song, which in the printed editions of Namdev’s verses is ten lines long, stops here after two lines. Two new songs are inserted, one by Janabai, Namdev’s “maid” mentioned earlier, and another by Tukaram, both of which are about the glories of listening to kirtan. Then the Namdev song is restarted from where it was interrupted, from its third line until the end. Another bada, from a private collection in Dhulia and probably finished around the eighteenth century, gives only songs by Namdev and Tukaram, but they are, again, sliced, edited, and intermingled, as if offering only first and second lines as reminders, but positioned together to bear out their thematic interaction. I found this pattern to be quite common in manuscript collections of the late seventeenth and eighteenth centuries, as several more notebooks revealed the same alternating pattern.

Let me highlight another example. Bada number 52 in the collection of the Bhandarkar Oriental Research Institute contains a section that tells us it will relate “the beginning of the story of Uddhav’s arrival” (Uddhav āgamana prārambha), i.e. the arrival of Krishna’s messenger to give word to the gopis as to whether Krishna has survived his encounter with the evil king Kamsa in Mathura (see Figure 5.4). The text begins to divide songs in Hindi, called sakhis, and Namdev’s Marathi abhangs. This goes on for ten pages, moving through four Namdev Marathi abhangs interspersed with Hindi sakhis. While this may seem a strange way to record songs, it is perfectly normal from the point of view of the performance of kirtan. Most kirtankars interweave songs in exactly this way. They select a single, exemplar abhang; then they give one couplet, or sometimes simply a single line, of this abhang; and follow with a short portion from another sant’s songs, or a brief story, or an aphorism, or a Sanskrit shloka, etc. In the context of performance, these interwoven items can come from any source—and indeed Namdev is not the most commonly invoked. But in most texts I’ve seen it seems that Namdev is the one who allows for the interlinear commentary to be enacted—his songs appear most frequently in such intermingled contexts. The format that can be seen in the notebook is, I believe, a transcription, or perhaps a plan, for a kirtan. It is as close as the text can come to replicating the oral performative context.

Fig. 5.4 Transcript of a kirtan from a Marathi bada, c. eighteenth century. Bhandarkar Oriental Research Institute. Author’s photograph, CC BY.

Yet this peculiar form is more than just transcription. One sees this format abundantly in the kirtan primer handbooks that have been published over the last hundred years or so. The Kirtan mārgadarśikā (Exposition on the Kirtan Tradition) is a series of kirtan primers produced by Krishnadas Lohiya, a famous kirtankar of Pandharpur (see Figure 5.5). The layout of his primers follows almost the same pattern as we saw in the badas above: a “kirtanace abhaṅg” is chosen, which provides the exemplar and theme for the kirtan, and it is delivered in lines and couplets interspersed with songs from other sants. Lohiya also gives his own commentary among the interspersed lines, but one presumes that this voice, the voice of the kirtankar (as opposed to the voice of the sants whose songs are being used) would change according to the performer and the performance context. If one removes the kirtankar’s voice, the format is exactly the same as one finds in bada 52.

Fig. 5.5 A representative page from Lohiya, K. 1997. Kirtan mārga darśikā (Pune: Sharada Sahitya), pp. 194-95. All rights reserved.

This is a format that privileges performance. Text here is submitted to the demands of performance. The text is a tool, a means and not an end. Furthermore, its role is not preservation—it may function as an archive, but its composition is not intended to ossify a text and convey it into the future as a fixed form. Instead, the text is dynamic, meant to trigger and prompt a performance, existing as a kind of outline of a kirtan, but containing no narrative, typological, or historical logic independent of performance. In other words, it cannot stand alone, as can a pothi or a fixed, complete composition. The sort of format seen here is always a “work in progress” because it is a means not an end. Other scholars have noted the echoes of performance in manuscript collections. Winand Callaewart and Mukund Lath, for example, in their wonderful work on Namdev in Hindi contexts have found that Namdev’s songs were eventually standardised according to raga, much as the songs of the Ādigranth and other collections in North India were standardised.6 Yet I would argue that this process reveals not a secondary status vis-à-vis performance. Over time, raga associations become fixed to particular compositions in their entirety—words and music are joined—and a new kind of text emerges. The categorisation by raga suggests an origin in performance, but becomes its own literary organisational strategy. In time, raga categorisation became a fixed system, a literary logic to manage a growing corpus of songs. But in the case of bada 52, for example, the textual format resists distillation into a new literary type (like the word-music association of pada and raga) because it remains self-consciously incomplete. Such badas record fragments of songs arranged for performative effect—hints and indications to a performer. But taken alone, these are fractured compositions unified only when plugged into the music-narrative-embodiment matrix of a kirtan performance.

These examples concur to show that performance, not literacy, dominated the pre-modern public sphere of performative devotion in Maharashtra. Literacy was everywhere, there is no doubt. And it was not just a technology of the elite, but clearly part of the daily work of performers who were not in any real sense elites. Literacy did not have the patina of mystery we often associate with the “literary” classes, high caste priests, scribes, and authors, nor was it central to the conveyance of devotion. Kirtan, embodied in the roving kirtankar, was the nexus of the public culture of bhakti, and of public entertainment in general. In this context, a kirtankar might be considered an intermediary between text and orality, though his bread and butter, as it were, was the oral public kirtan.

The logic of performance has remained within the written record of the kirtankar up to the present, even though most archived badas were composed before the twentieth century. Kirtankars still carry notebooks and use them in the same way badas were used, only now, of course, they also have many more printed aids—the kirtan handbooks produced by people like Lohiya and the many editions of the songs and hagiographies of the sants. So perhaps the last example of the kind of logic of writing before the advent of generous printed resources I can point towards can be found in the monumental works of Mahipati, who lived throughout the eighteenth century, dying in 1790. He was a Deshastha Brahmin kulkarni or village accountant of Taharabad, but he is more famous now as a kirtankar who specialised in the stories of the lives of the sants. Around the middle of the eighteenth century, Mahipati is said to have written his most famous composition, the Bhaktavijay. In this work, Mahipati clearly relies upon other materials, especially earlier hagiographies and biographical or autobiographical songs attributed to various sants, the most important of which, for the earliest strata of Varkari lore, is Namdev. If we take, for example, Mahipati’s retelling of a purported journey to North India undertaken by Namdev and his companion, Jnandev, we find that Mahipati follows an earlier version of this story, attributed to Namdev, almost word for word (see Figure 5.6). In the illustration given, on one side you see the record of this composition attributed to Namdev from a manuscript dated 1631 CE. On the other side you see Mahipati’s retelling dated 1762 CE. The bolded portions are common to both versions, whereas the unbolded text in Mahipati’s version represents Mahipati’s inventions in, I would argue, the style of a typical kirtankar. We can see that Mahipati adds many interlinear comments to Namdev’s song in a manner very similar to the format of bada 52 and other manuscripts. He combines lines 2 and 3 from Namdev’s song to make his line 87, fills in his own presumably improvised materials from the performance of a kirtan, and then returns to lines 4 and 6 from Namdev for his line 93. The format is again that of a kirtan. Though Mahipati self-consciously refers to his works as “pothis” and “granthas” (literally, book), throughout his hagiographies he calls upon his audience to “listen” and to “hear” of the lives of the sants and always refers to himself as a kirtankar. Indeed, in his lifetime, he was famous for his kirtans. Only after his lifetime did the written texts he left behind come to flesh out his fame in religious culture.

Fig. 5.6 (L) section is taken from the Śrī Nāmdev gāthā (1970), p. 343; (R) is taken from Abbott and Godbole’s translation of the Bhaktavijay 1996 [1933], pp. 164-65.
Image by the author, CC BY.

At the conclusion of his Bhaktavijay, Mahipati makes his ability to connect oral performance and literacy clear. He writes:

I have written every single letter in this book
Just as Rukmini’s husband has commanded.
Like the puff of breath blown by a musician,
I am the wind that sounds the flute.7

Mahipati’s peculiar metaphor, mixing literacy and the sound of the flute, is clearly meant to invoke a performance, not the preservation of literacy—he is not the ink of the pen, for example. Yet he is explicit that he has “written every single letter” just as each note of an instrument is sounded. Mahipati concludes his Bhaktavijay this way:

To some you have given knowledge of the soul.
Some have begged to dwell in the union of self and universe.
My heart’s desire is that I will sing [varṇīn]
About the character [guna] of Hari.
Some sit on beds of nails,
Some sit with Vishnu in heaven.
For me, in the kirtan of your servants
I have become lost in supreme love.8

Mahipati is both producer and consumer of the kirtan of the sants. Indeed, as we have seen, he takes their voices and adopts them into his own in a format that I’ve referred to elsewhere as corporate authorship, following the work of Jack Hawley and others on authorship in bhakti contexts.9 Mahipati does not tell us when he is quoting Namdev and when we are hearing his original voice. But Mahipati is no plagiarist either—he is a performer, whose performance is crystallised, by his own hand, into a text, i.e. a permanent, non-performative form. Mahipati closes what is probably his most famous hagiography by invoking the oral context of performance, set down on the written page. Within fifty years of his death, this text, the Bhaktavijay, would enter the world of print as one of the first and most popular devotional texts published in the Marathi colonial public sphere.

In this essay I have tried to demonstrate how one can read the written archive of Marathi bhakti materials in a way that reveals the logic of kirtan and oral performance and emphasises the telling rather than the text. In other words, certain texts themselves suggest that sometimes the telling is the text, fossilised in literacy, imperfectly cemented in time. The written materials I have reviewed here are all the product of a performance economy, in a sense, a literary space that served the performance of kirtan before the modern period. Clearly, orality and literacy existed in a symbiotic relationship. In his study of the practices of “everyday life” Michel de Certeau has suggested that the modern idea of writing and authorship occurred because writing had lost its connection to orality when progressive developments in writing and the accumulation, especially through print, of vast bibliographic genealogies began to refer to earlier written sources, rather than to “original” oral ones.10 At this point writing took on a character all its own, made up its own genealogy resting on replication and consumption, and the symbiotic connection between orality and literacy was lost. In our case we could say that when Mahipati the oral performer became Mahipati the literary and printed author, his fame as a kirtankar faded in lieu of his fame as a writer of books about the bhakti past. A similar historical-teleological notion is readily apparent in the work of Goody and Watt, for example: although they clearly state that literacy is always afloat in a sea of orality, the dominance of literacy is also the advent of history and of sustainable rational thought, the foundations of modernity.11 This shift to a self-referential writing system can also be seen to mark a loss of cultural integrity, or a schism that takes on the shape of class differences. Indeed, as I have pointed out, the “oral” texts of the bada are of a different class than the “literary” texts of the pothi or grantha.

Very few concrete examples exist anywhere in the world of a clear transition from an oral society to a literate one—this distinction is one of the key problems of the orality-literacy debate, and has been rejuvenated in the “cyber” world where text, image, video, and live human interaction can now take place in rapid and simultaneous ways as if technology were constantly trying to replicate the experience of “being there” at a live performance. In the public performance of kirtan in the pre-colonial era and later we see a systematic interweaving of orality and literacy that nevertheless privileges orality. I have argued in this essay that the logic of practice at the heart of the public performance of bhakti among Varkaris in Maharashtra not only coexisted with forms of literacy but recognised the “threat” of literacy, as well as its necessity. The tradition, through figures like Namdev and centuries of performers, innovated ways to integrate writing and performance, but was always careful to maintain the necessary privilege of performance over writing, to control the power of literacy. I have tried to demonstrate this logic of practice for kirtankars by noting how writing functioned in a system of performance that sought to guard against the loss of the immediacy of cultural memory, physical display, and devotional interaction. For the Marathi kirtankar, who remains the lifeblood of the devotional tradition of Varkaris, the tongue would always make the best book.

---