Rethinking Terms: Dohā, Vajra-, and Caryāgīti

Abstract

Dohās, vajragīti, and caryāgīti are key terms associated with the poetic writings of the Mahāsiddhas. This study focuses on Apabhraṃśa dohās, their commentaries, Tibetan translations, and collections containing them, shedding light on previously neglected aspects of this text type. By investigating the historical and original contexts of these three terms and comparing them to their later applications in traditional contexts and academia, this paper argues against the prevailing notion that they are genetically distinct and that this text type is primarily defined by orality and spontaneity. Consequently, it challenges the romanticized myth of certain origin narratives, such as student–teacher encounters. Instead, this brief presentation demonstrates that the often-repeated stereotypical definitions of these terms should be largely rejected, as they are merely different labels for the same text type with blurred and ill-defined subcategories. The analysis of primary sources reveals that various facets, e.g., compilation (an important but neglected aspect), go beyond the strongly emphasized oral component of this text type, thereby leading to the inaccurate definitions of the terms. In conclusion, intertextuality, compilation, and assigned authorship are crucial yet overlooked elements in defining the text type and understanding its function.

1. Introduction

This paper explores three generic terms—dohā, vajragīti, and caryāgīti—and their significance within the Indian tantric song-writings of the “Great Accomplished Ones” or mahāsiddhas. It sheds light on crucial aspects of this text type that have not been given much emphasis in previous academic explorations. This research draws from selected Apabhraṃśa works, their Sanskrit commentaries, and Tibetan translations, including the Mekhalāṭīkā and Dohākośaṭīkā (both commentaries to Kṛṣṇacaryāpāda’s poem), the Sārārthapañjikā (commentary to Tilopāda’s poem), the Dohākoṣapañjikā (commentary to Saraha’s poem), and the *Caryākoṣagītivṛtti (a commentary on a collection of “short songs” by Munidatta). Additionally, later Tibetan writings and collections, such as the Phyag rgya chen po rgya gzhung, are also utilized in this research.

The terms dohā, vajragīti, and caryāgīti have been widely used in both the Indo-Tibetan traditions and academic literature. However, there have been no prior dedicated studies to systematically explore and define these terms. This paper aims to challenge the prevailing notion that these three terms are generically distinct from each other, as previous research has often suggested. When investigating these terms and their use and classification, many of the classifications found in scholarly literature likely originated outside of the Indian subcontinent, such as in the Tibetan or Newar cultural spheres. Even within academic and scholastic traditions, these classifications may have resulted from terminological and contextual confusion, stemming from a limited number of primary sources available for examination.

This paper traces the development of the terms dohā, vajragīti, and caryāgīti by examining their original context within Indian usage and comparing them to their later applications outside India. By analysing primary sources, the paper argues that these seemingly distinct generic terms are actually sublabels within a single text type. While the oral component of this text type has been emphasized, the paper highlights other essential literary aspects that have been given less emphasis. Through a closer examination of primary sources and their contexts, this paper reveals more commonalities and standardised views and sources than expected for a text type often associated with strong individualism and spontaneity. The observation of shared literary themes and contexts suggests a religious function beyond orality and spontaneity as a mere means of the self-expression of the author, raising crucial and previously only little explored questions that are vital for a more comprehensive understanding of this text type.

2. Results

There is compelling evidence to consider the term dohā, which refers to “Buddhist tantric rhyming couplets”, as a hypernym encompassing both vajragīti, often translated as “diamond or adamantine songs”, and caryāgīti, commonly translated as “performance songs”. Additionally, there are other songs with similar contents (see note 24) that have not been categorized under any of these three terms. Depending on the context, this classification may also include indigenous Tibetan labels like nyams mgur (“experience songs”), which context-wise can be related to Indian dohā traditions.

Although secondary sources frequently discuss these terms together due to their close relationship, they are often treated as though they belong to different genres, which is a misconception that will be demonstrated below. Contrary to what is commonly asserted in the secondary literature, three main aspects support the idea of considering vajragīti and caryāgīti as subcategories or alternative labels for the text-type designation dohā, rather than viewing them as separate designations for different text types. The reasons for doing so are as follows:

(1)

There is limited evidence to suggest that vajra- and caryāgīti were used as separate generic terms alongside the dohā in primary Indian literature. While the term caryāgīti does not appear to be attested at all in primary Indian literature but only in Tibetan (spyod pa’i glu), the term vajragīti is found in the Indian literary context as a specific song-mode within tantra and sādhana. However, it cannot be used interchangeably with the label dohā.

(2)

In contrast to the Indian context, particularly in Tibet, all three terms (dohā, vajra-, and *caryāgīti and other combinations of these) are attested and can be used somewhat interchangeably. This means that different terms are employed to label similar works and collections and vice versa. Within the Tibetan tradition, the terms rdo rje’i glu and spyod pa’i glu, which correspond to vajra- and *caryāgīti, respectively, are both considered equivalent to the Indian term dohā.

On the other hand, nyams mgur is a native Tibetan label that expresses something comparable to but not identical to a dohā. While nyams mgur shares many features associated with Indian dohās, it also exhibits distinct characteristics that set it apart from them.

(3)

Indeed, despite variations in the structure and form of different song-poems, their contents and contexts often remain similar. Considering this, it appears more suitable to classify vajra- and caryāgīti as subcategories or alternative labels for the broader term dohā. This distinction allows for the acknowledgment of the similarities of vajra- and caryāgīti within the larger context of the dohā lore.

(4)

The aspect of spontaneity is insufficiently addressed in view of the fact that song-poems are largely defined by the use of the Apabhraṃśa language, in view of the very limited resources presently available for their study and as the important aspects of compilation and attributed authorship are largely ignored. Further investigation is necessary to understand the extent of spontaneity within this linguistic framework and how it impacts the overall composition and meaning of the texts.

3. Discussion: Defining the Terms

3.1. Caryā and the Aspect of “Performance”

The term caryā, which is translated as spyod pa in Tibetan, primarily represents an idea or concept associated with how salvation is presented in the dohās. This notion extends to tantric Buddhism in general, as evidenced in various sources, including early and well-known exegetical texts like Padmavajra’s Guhyasiddhi:

vrajanty anuttaraṃ sthānaṃ kālāvadhivivarjitam |

prāṇātipātinaḥ krūrāḥ krūrakarmaratāś ca ye ||1.13||

mithyāvāgvādajālena mohayitvāpi ye narāḥ |

narā jīvanti te ‘py āśu yatra sidhyanti caryayā ||1.14||

paradārābhigantāraḥ paravittāpahāriḥaḥ |

jugupsāhīnakarmāṇi kurvanto ‘pi nirantaram ||1.15||

yāṃ caritvā vrajanty āśu kāmadhātūrdhvataḥ param |

pracchannavratarūpeṇa sādhakānāṃ bravīmy aham ||1.16||

guhyacaryāṃ parāṃ divyāṃ sopāyāṃ sarvasiddhidām |

prāpyate janmanīhaiva vajrasattvapadaṃ yayā ||1.17||1

“Those2 who proceed to the unsurpassed state that is beyond the limits of time, may take lives and be ferocious, taking pleasure in cruel deeds. Such men may also cause confusion through nets of lies.3 All those men who live [like this] get accomplished swiftly through caryā.4 They may make love to another’s wife and steal another’s wealth and even they are constantly performing the deeds of the disliked and inferior. Proceeding as such, they swiftly reach the other shore which lies beyond the desire realm. I teach for sādhakas in a manner of secret observances (vrata), the supreme and divine secret caryā (tantric conduct) granting all accomplishments and by which the state of Vajrasattva is obtained in this very lifetime.”

One may observe that this general notion is closely linked to the concept of unmattavrata, often rendered the “conduct of a mad-man”. It signifies a practitioner’s deliberate departure from societal norms and conventions to demonstrate their transcendence beyond those limitations. This notion forms a major conceptual framework for the activities and the religious practices of the siddhas, who are the authors of the dohās. The dohās, in turn, represent one of the literary expressions of this tantric milieu and culture.5

In this context, the term caryā (as evident from the above passage) primarily reflects tantric conduct and practices displayed in the dohās rather than referring to the mode or manner of performance. Therefore, the term caryā can be equally applied to all types of texts used to express this idea, encompassing the various labels discussed previously. Moreover, it may even serve as a unifying concept that subsumes the diverse expressions of tantric notions and practices found in the dohās and related texts. As a sidenote, it should be also kept in mind that song-poems, regardless of how these are classified, are composed and transmitted in Apabhraṃśa,6 which presupposes some mode of performance by default since Apabhraṃśa was never used as a prose language, a fact to which I will return later.7

Indeed, translating the term caryā as “performance” might be misleading as it does not necessarily capture the strong soteriological connotation of the term, which refers to a “mode of being”. It is important to clarify the precise meaning of “performance” when using the term caryā, distinguishing between “the way or manner of performing something” and “a conduct or mode that is performed”. While I advocate for the latter interpretation, there is no consistent use of these labels, making it challenging to draw definitive conclusions. Regarding the differentiation between caryāgīti and vajragīti, the former sub-genre may emphasize the act of performance more, while the latter could emphasize the content, a distinction proposed in various secondary sources.8 However, this approach can be problematic due to the lack of consistency in the use of these labels.

Taking caryā as “the manner or mode of performance” unnecessarily narrows down the broader notion of the term, leaving aside its significant soteriological component, which is characteristic of the text type. The caryāgītis frequently refer to tantric codes of conduct (caryā) in combination with various means (upāya), e.g., psycho-physic techniques and experiences (*sūkṣmayoga), similar to what is found in vajragītis (when used in the Tibetan context) or dohās. The only potential difference lies in the consistent style observed in the songs contained in the *Caryākoṣagītivṛtti.9

In this context, it is essential to note that the most popular and, in fact, the only collection of caryāgīti known to me in Indian languages is the *Caryākoṣagītivṛtti, which, as indicated by the asterisk, is a back-translation from the Tibetan sPyod pa’i glu’i mjod [sic mdzod] kyi ‘grel ba.10 As such, there is currently no direct attestation of the term caryāgīti in primary Indian sources, although it is of course possible that the term existed and will be attested for in the future.

However, the term caryā does appear in the opening stanza of Munidatta’s commentary. As correctly pointed out by Kvaerne, in that context, it serves a descriptive function, referring to the nature of the song-poems rather than directly implying the title of the composition. Nevertheless, it is plausible that the development of the term caryāgīti as a genre label might have evolved from earlier titles given to the collection11 and that the opening section cited in the following played a role in the emergence of the term caryāgīti, which eventually came to denote a specific text type:

namaḥ śrīvajrayoginyai ||12

śrīmatsadguruvaktrapaṅkajarasāsvādasphuraddhīdayo13

natvā śrīkuliśeśam advayadhiyaṃ14 śraddhāprasannānanaḥ |

śrīluyīcaraṇādisiddharacite ’py15 āścaryacaryācaye16

sadvartmāvagamāya17 nirmalagiraṃ18 ṭīkāṃ19 vidhāsye sphuṭaṃ20 ||

“Homage to the great Vajrayoginī—I, after having—with compassion (°dayo) shining forth in [my] mind (°sphuraddhī°) through tasting the nectar of the lotus (°paṅkajarasāsvāda°) of (coming from) the mouth of the glorious true Guru (śrīmatsadguruvaktra°) [and] a clear face with faith (śraddhāprasannānanaḥ)—bowed down (natvā) to the glorious Vajra-Lord (śrīkulīśeśvarmam), who has a non-dual cognition (advayadhiyaṃ), will compose (vidhāsye) a clear (sphuṭaṃ) commentary (ṭīkāṃ) of pure words (nirmalagiraṃ) on the collection of amazing deeds (āścaryacaryācaye), which were performed by the siddhas (°siddharacite), such as the glorious venerable Luyī (śrīluyīcaraṇādi°) for [, i.e., so that others may] understand the[ir] excellent path (sadvartmāvagamāya).”

Consequently, regardless of the attestation in Indian sources and the potential scriptural influence on the emergence of the genre term, one should reconsider translating the Sanskrit word caryā, i.e., spyod pa in Tibetan, as “performance” to alleviate some of the confusion outlined above. Instead, given the actual content of the songs in question, a more suitable translation for caryā or spyod pa could be “tantric conduct”, aligning with Munidatta’s apparent intention in using this term. Therefore, if one chooses to translate the term as a text type label (although the use of such translation may be limited), a more accurate rendering would be “songs expressing tantric conduct”, encompassing the content, function, and context of this genre term within the broader Buddhist framework.

Further evidence, if more was needed, that reinforces the notion that “performance songs” may not represent a distinct independent genre and that the labelling is debatable, is the fact that the boundaries of this term (and others as well) when used as a genre label are not well-defined and are, in fact, quite fluid. This fluidity in usage has resulted in ambiguous and overlapping applications of these labels (refer to notes 24 and 26–27).

In the title of his commentary on the songs of the *Caryākoṣagītivṛtti attributed to Kṛṣṇacaryā, Tāranātha uses the expression dohā thor bu, which simply means various dohās.21 This indicates that, in terms of content, there was no generic difference between the different song writings in Apabhraṃśa for the renowned 16th-century mKhas grub pa of the Jo nang school. And indeed, it is evident that he was well acquainted with Indian dohā writings.22 Additionally, while the images and lyric pictures found in the *Caryākoṣagītivṛtti may have a distinct “crypto folkloristic” style and exceptionally esoteric elements, their overall content is not fundamentally different from those addressed in the dohās of Tilopāda or Kṛṣṇacaryā. This is emphasized by the fact that Munidatta quotes about two-thirds of the latter’s Dohā in his commentary.23

Hence, I am inclined to follow Tāranātha, whose title suggests not to regard the *Caryākoṣagītivṛtti as a collection of songs constituting an independent genre, as implied, for example, by Roger Jackson, who states that: “performance songs and diamond songs, […] differ generically from Dohās because of their different context and function” (Jackson 2004, p. 6).

The fluidity in definitions and the interchangeability of how similar poetic forms are referred to becomes evident when examining various instances where terminology appears to have multiple layers of potentially confusing use. This can be illustrated through David Templeman’s translation of Tāranātha’s Life of Kṛṣṇācārya/Kāṇha, wherein Tāranātha makes the following valid statement: “although the Vajragīti came from the collections of songs in the tantras themselves, the true siddhas expressed their experiences in the form of dohās, and doubtless of more conducive circumstances they later appeared as forms of Vajragīti” (Templeman 1989, p. 46).

This clearly conflicts with the above given statement of Roger Jackson. More of such terminological conflicts, that seem to be based on earlier usages of “genre terms”, can be found in various places. For the purpose of this paper, however, a few examples may suffice. Lara Braitstein stated that: “the caryāgīti are not transmitted or performed as a genre of spontaneous song [.Vajragītis …] are performed by Tāntrikas at Tantric gatherings,” (Braitstein 2004, p. 133). This statement clearly contradicts the previous assertion made by Roger Jackson, as well as Tāranātha’s statement—all of which are in disagreement with each other—and introduces another problematic concept, namely, the idea of spontaneity (see Section 3.3).

In Per Kvaerne, without mentioning a certain performative setting or proposing a manner of origination, we read that:

“The term caryāgīti does not apply to any particular collection of songs. It is a general term used to designate a genre of spiritual songs which at one time must have been composed in great numbers. The fifty songs collected and commented on by Munidatta represent only a small selection of what must originally have been a considerable body of texts.”

(Kvaerne 1977, p. 7)

Despite the apparent conflicting information and ambiguities in the use of genre labels, the idea of generic differences continues to be reiterated without—I feel unfortunately obliged to say—sufficient hard evidence and explanations for the origin of these ideas and statements. Many of these claims seem to be based on unsubstantiated assumptions, such as the belief that these songs “must originally have been many.” However, the few available testimonies, although their absence is not proof of their non-existence, actually suggest the opposite.

A recent example of such idealized and commonly upheld definitions can be found, to my knowledge, in Karl Brunnhölzl’s translation of the miscellaneous Dohās found in the rGya gzhung, where he states:

“In sum, practically speaking, vajragītis and caryāgītis differ from dohās mainly in terms of their different context and function. Dohās are spontaneous spiritual aphorisms expressed in the form of rhyming couplets. Vajragītis are songs sung in the specific context of a gaṇacakra. Caryāgītis came to be individual or group performances (often of a cycle of such songs), typically elaborately choreographed, that can either be presented in the context of a gaṇacakra or on other occasions. However, just like the songs themselves do not follow a strict pattern, the distinctions between these three “genres” are far from being hard and fast. For example, dohās can also be sung at a gaṇacakra and vajragītis outside of a gaṇacakra. Also, any of them can be in the dohā meter or other meters, can include more sophisticated prosodic elements, and may or may not be accompanied by music and dancing.”

(Brunnhölzl 2019, p. 16)

Amidst the apparent contradictions found in the aforementioned definitions and acknowledging the scarcity of evidence to substantiate these definitions beyond our own ideas, assumptions, and assertions, there exists enough evidence to prompt a reconsideration of the concept of the performance song both as a label and in terms of its established translation. Additionally, it has become evident that the precise relationship among the three terms discussed, while somewhat closely related, remains far from clear.

Despite this lack of clarity, the distinction between the performance song and other forms like dohās and vajragītis has been repeatedly emphasized, seemingly based primarily on the paratextual element of a given title. However, this emphasis, based on historical or contextual evidence, may be questionable. The very idea of “performance”, moreover, as will be discussed below, could easily be applied to other song-poems (including dohās and vajragītis) than those that are similar in form, structure, and content to those contained in the collection named *Caryākoṣagītivṛtti or those found among the so-called performance songs in the rGya gzhung. Moreover, the notion of “performance”, especially when accompanied by music and dance in a theatrical or comparable setting, deserves attention in relation to literary language. Exploring this connection could prove valuable in future investigations concerning the label caryāgīti and its role within the broader dohā tradition.

Another overlooked point worth mentioning is that when classifying dohās and their related sublabels as song-poetry, it may encompass a number of texts not explicitly labelled as dohās. This fact appears to have been largely unaddressed in current discussions. It is, however, crucial to consider these various aspects and explore further into their implications to gain a more comprehensive understanding of the subject matter.24

3.2. Vajragīti and the Dohās, Same–Same but Different

As rightly stated by Tāranātha in his above-cited translation regarding the genre terms dohā and vajragīti, the main difference lies in the fact that these two terms have different usages. They are applied to different forms of songs within the distinct cultural spheres of Indian and Tibetan tantric Buddhism.25

This important distinction of terms, which are used across Indian and Tibetan cultures in more than one context, can be nicely exemplified by comparing the two following quotes that both refer to the same text, namely, the famous cycle of Saraha’s songs: “the final consummation […] This is the content of which the Dohā-verses sing.” (Snellgrove 1959, p. 37) and “[vajragīti] are easily identified on sight by the fact that usually a title will have ‘adamantine song’ worked into it, but in fact, the ‘vajra’ of ‘vajra songs’ refers to the meaning of the songs, not the form.” (Braitstein 2004, p. 134).

In the light of Tāranātha’s observation and considering that different labels are used within Indian and Tibetan literary sources to denote similar text types (and actual texts), it becomes evident that both of the above statements are correct when taken in their respective contexts. Moreover, it demonstrates that the notion of these terms describing distinct text types must be clearly refuted.26 This becomes even more apparent when we consider that the same textual collections have been labelled as both dohā and vajragīti, as is the case with Saraha’s song-poems. Both of the above-cited statements by David Llewellyn Snellgrove and Lara Braitstein pertain to the context and description of Saraha’s song-poems.

The confusion, however, lies with the use of the term vajragīti, which, originally (as noted already by Tāranātha), is the name given to Apabhraṃśa lines that are used within tantras, such as, among others, in Hevajratantra II.iv.6–8, and ritual practice texts (sādhana), for instance, Ratnākaraśānti’s Bhramaharanāma Hevajrasādhana, wherein so-called vajragīti are used in reply to elucidate the nature of song and dance or sung by yoginīs so as to awaken a deity from its state of non-physical samādhi. Such verses, quite certainly, are indeed (generically) different from dohās, but also from both caryāgīti and vajragīti when used to refer to collections of song-poetry, such as those found in the rGya gzhung, i.e., denoting the subordinates of the Indian and Tibetan dohā traditions. In non-Indian contexts, all three terms, dohās, caryāgīti, and vajragīti, can constitute stand-alone songs (or parts of song collections of a similar nature). In the Indian language context, vajragīti appear, according to my information, only within the abovementioned specific context of a tantra or sādhana. Finally, it has to be emphasised, however, that the stand-alone song-poems available in Indian language all use the terms dohā or that, due to damaged textual witnesses, no generic terms are found within these.27

3.3. Spontaneous Performance in the Light of Apabhraṃśa, the Aspect of Form and Compilation, and Quantitative Limitations

It should be kept in mind that all of the song-poems and their subordinates could have been performed or spontaneously uttered under various circumstances. This somewhat romanticized notion of the songs’ origins, akin to the discussed genre-terms, has also been reiterated on multiple occasions.

Among the first scholars to study the dohās using a more Western academic approach was Herbert V. Guenther, who even incorporated the concept of spontaneity in the title of his renowned book about Saraha’s poetic writings, Ecstatic Spontaneity, published in 1993 (Guenther 1993). This idea of spontaneity is echoed by subsequent generations of scholars as well. For instance, Ann Waldman proposes the notion of student–teacher interactions as the possible origin of such songs:

“the doha, a song of realization that acknowledges an encounter with a master teacher, traditionally a guru or lama, and explores a particular wisdom or teaching transmitted through a kind of call-and-response duet format”.

(Waldman 1996, p. 264)

Miranda Shaw also utilizes the somewhat elusive label “songs of realization”—a formulation loaded with meaning that has been applied to encompass a broad range of poetic literature and alludes to sahaja, another significant notion in relation to song-poetry in Apabhraṃśa.28 However, in formulating her statement, she incorporates the idea of spontaneity, but in this case, it pertains to the caryāgīti:

“The feast culminates in the performance of tantric dances and music that must never be disclosed to outsiders. The revelers may also improvise “songs of realization” (caryagiti) to express their heightened clarity and blissful raptures in spontaneous verse.”

(Shaw 1994, p. 81)

The idea of spontaneity, as evident from the previously cited passages, appears to be a recurring element in the various definitions found within the secondary literature. Consequently, it is appropriate to highlight another significant yet somewhat overlooked aspect concerning Apabhraṃśa and the kind of spontaneity implied by its use, leading to the question of how spontaneity can be meaningfully understood in the context of tantric song-poetry.

In the case of Apabhraṃśa as a literary language, regardless of contextual notions, titles, or other (con-)textual matters, a certain “artificiality” is inherent in all the song-poems that employ the Apabhraṃśa language and its features, such as phonetic plays (saṃdhyabhāṣa).29 The reason for ascribing a certain artificiality to Apabhraṃśa, and thus to the song-poems in which this language is used, stems from the fact that Apabhraṃśa is exclusively employed in poetic literature in combination with and based on other vernaculars, Prākṛts, and classical Sanskrit. Unlike Sanskrit and other Prākṛts, it is neither used as a prose language nor spoken in everyday communication.30 As Vít Buberník states:

“Classical Sanskrit, Prākrits, (stage Prākrits in Sanskrit plays, Māhārāṣṭrī in epic and lyric poetry) and Apabhraṃśa […] are not […] three different languages in the sense of the word; rather we are dealing with ‘triglossia’ definable as the simultaneous use of three functional varieties of the same language for literary purposes.”

(Buberník 1998, p. 16)

Indeed, the fact that Apabhraṃśa was not spoken but masterfully utilized in poetry alongside other commonly used languages implies that composing skilful and deliberate works using the linguistic features of Apabhraṃśa requires some level of preparation and training. Contrary to the notion of spontaneous creation, the process of creating song-poems in Apabhraṃśa would therefore likely involve a conscious and purposeful effort by the poet to artfully employ the unique linguistic qualities and phonetic plays of the language. This implies that the use of Apabhraṃśa in poetry demanded a certain level of expertise and familiarity with the language’s particularities, making it a more structured form of expression rather than an impromptu act of creation.

In this context, the term “spontaneity” should be understood as a form of outpouring of creative inspiration resulting from the poet’s (siddha’s) contemplation, which, when applied to Apabhraṃśa song-poetry, may denote a fusion of inspired expression and the intentional use of the language’s poetic potential.

Consequently, in phrases like “Dohās are spontaneous spiritual aphorisms” (Brunnhölzl 2019, p. 16), the concept of spontaneity referred to is likely not the product of “random impulses” but rather the result of rigorous training and familiarity, reflecting a “mastery” of the content that is spontaneously uttered. It is a form of effortlessness that arises from a deep understanding and command over the subject matter.31

Yet, discussing this notion in depth requires a separate occasion, involving poets and scholars who specialize in its study, raising further intriguing questions.32 Even if we accept that the song-poems originated as spontaneous expressions emerging from the mastery of tantric practice, it remains challenging to ascertain which parts of any given dohā truly stem from such spontaneous origins and which elements were added over the centuries through scriptural Indian and Tibetan compositional and editorial practices. As Roger Jackson has already observed:

“if a dohā is a self-contained (…) couplet, probably oral in its initial transmission, then the very idea of a ‘treasury’ of dohās is fraught with difficulties”.

(Jackson 2004, p. 10)

The accuracy of this observation becomes evident when examining Tilopā’s Dohā as transmitted within the Sārārthapañjikā. This text shows significant intertextual relations with other sources, indicating that various verses and ideas have been adopted or “loaned” from other texts. This suggests that Tilopā’s Dohā (and other texts of this genre to some extent) is not an individual’s independent composition but rather contains shared, copied, loaned, and adapted elements. As a result, establishing a particular authorship for the entire text, as well as some of its parts, becomes challenging.

The same can be observed in the case of some songs contained in the Dohā mDzod brgyad, a compilation of eight song-poems. These compositions combine older Indian elements with additional Tibetan materials, resulting in only portions of the songs available in Indian languages being retained, incorporated, or merged into a new text. Some of the songs within this compilation appear to have been written in a spirit different from the Indian dohās and closer to later song-poems claiming to be of Indian origin but more likely of Tibetan origin. One such example is the famous Mahāmudropadeśa, which does not even bear the label dohā but instead alludes to a “romanticized” narrative of its origin—a song orally received at the Ganges River by the advanced student Nāropāda from his realised teacher Tilopāda. This allusion aligns with the archetypal portrayal of the songs’ origins as a “student-teacher dialogue”.33 These cases demonstrate the complexities of tracing the origins and authorship of song-poems, underlining how intertextuality has shaped composition and development of the dohās over time.

Also, upon closer examination of the main stylistic elements that differentiate the so-called “performance songs” from other types of song-poems, doubts arise regarding the notion of a “spontaneously uttered unity.” Ironically, the most esoteric and cryptic “spontaneous utterances” follow a very strict format, consisting of five couplets, with the first being a refrain verse (dhruvapada) and the last being the bhaṇa-line (Kvaerne 1977, p. 73). It is noteworthy that all other three dohās for which we have surviving Indian language witnesses, especially the one attributed to Saraha, contain bhaṇa-lines similar in style to the songs found in Munidatta’s collection.34 This observation, combined with the aspect of compilation, where it becomes unclear what constitutes “a single dohā” as the result of an individual’s composition, weakens the significance of the number of lines paired with the presence of a bhaṇa-line in attributing a song to a particular genre or manner of origination.

The strict adherence to a specific form observed in these so-called “performance songs” is yet another aspect challenging the idea of pure spontaneity, as it—similarly to the deliberate use of linguistic features—suggests a deliberate adherence to a predefined structure. The interplay between compilation, format, and the elusive nature of authorship makes it hard to confidently ascribe such song-poems to a specific origin or genre based on their structural elements.

The lack of historical knowledge regarding the original forms of the treasuries and collections of songs, paired with the few primary textual witnesses, hinders our understanding of how performance songs were arranged, vajra songs were altered, and dohā utterances were collected until they reached their present form; we simply do not possess information on how these song-poems evolved over time and what modifications they underwent.

It is worth reiterating that only one collection of so-called performance songs, Munidatta’s collection, exists in its original language. Similarly, there are only three dohās available in their original language, found in the Sārārthapañjikā, Advayavajra’s (deutero) Viṣamapadapañjikā, Amṛtavajra’s Dohākośaṭīkā and Mekhalāṭīkā. Similarly, due to the scarcity of relevant texts, it is challenging to draw definitive conclusions or establish genre-specific definitions about the collections and treasuries accessible to us.

The historical challenge of acquiring information about the mode and manner of performances, combined with the limited number of original texts, precludes us from making claims about the origin and development of these song-poems and its building blocks, whether they were spontaneous or not. As a result, when attempting to deduce genre-specific definitions for the available collections, considerations of form and labelling should probably be excluded, or, at least, less emphasized. Instead, further exploration is necessary to understand the phenomena of compilation and intertextuality, and thus the literary function of the text type song-poetry, more comprehensively.

4. Conclusions

The uncritical adoption and reproduction of definitions and terminology, without thoughtful reflection about their origin and use, can foster confusion rather than clarity. This often results in unsubstantiated claims and superimpositions, leading to erroneous conclusions. A striking example of such terminological confusion can be found, unsurprisingly, on Wikipedia, in an article titled “Songs of Realization”, wherein it is stated:

“Songs of realization, or Songs of Experience (Tibetan: ཉམས་མགུར, Wylie: nyams mgur; Devanāgarī: दोहा; Romanized Sanskrit: Dohā; Oriya: ପଦ), are sung poetry forms characteristic of the tantric movement in both Vajrayana Buddhism and in Hinduism.”35

The initial statement, despite its conceivable reasons, clearly mixes up English, Tibetan, and Sanskrit terms, while also introducing another Indo-Aryan language and its respective terminology. This statement implies that the origins, functions, and contexts of these terms are, at least substantially, similar, bluntly ignoring significant contextual, historical, and religious differences. The article concludes with another astonishing claim, once again mixing up terms from different contexts and time periods:

“An example of a Doha song available in English translation, is by Rangjung Dorje (1284–1339). The Doha song is entitled Distinguishing Consciousness from Wisdom (Wylie: rnam shes ye shes ‘byed [sic!] pa).”.

(Ibid)

To avoid such significant confusion, I propose resolving terminological issues caused by the unwanted overlapping of terms by using indigenous terms only. For example, we can use dohā, spyod pa’i glu, rdo rje’i glu, or whatever the text type label used in relation to a song-poem, instead of backtranslations that are either unattested (i.e., caryāgīti) or used in different and/or multiple contexts (i.e., vajragīti).

Secondly, I suggest further reflection on the concept of what “performance” truly means and refers to. We should discuss the extent of evidence that can be derived from Munidatta’s collection, given that it appears to be a unique example (assuming, for now, the accuracy of the Tibetan translation). Additionally, we should consider whether other translations might better capture the broader contextual implications of the word caryā.

Thirdly, I suggest defining what “spontaneity” means in the context of Apabhraṃśa and how the genre of song-poetry, to a large extent, is the result of processes of compilation. We need to consider our lack of knowledge about the actual mode and manner in which songs were performed in their original historical settings and how they were arranged into compositions. Understanding the notion of “spontaneity” within the Apabhraṃśa tradition and its relation to the compilation process can shed light on the intricate nature of these song-poems and their development over time.

Finally, based on the evidence presented here, I invite other scholars to reflect on the question of “to what extent (or whether at all) text titles and labels can or should be seen as implying different genres or text types?” For instance, consider *Caryākoṣagītivṛtti 4.4, which reads:

sāsu ghare ghāli koñca tāla | cānda suja beṇi pakhā phāla || (“[She] is leading to the house of wind [and] the key is in the lock; sun and moon having been brought together, the parts are bound.”)

and comparing this to Kṛṣṇacaryā‘s Dohākoṣa 22, attesting:

jahi maṇa pabaṇagaaṇa duāren diṛha tāla bi dijjaī jaī tasu ghore andhareṃ maṇi dibaho kijjaī | (“When the door of the moving wind that is the mind has been locked firmly, when the mind has been made a lamp in the terrifying darkness, […].”)

We encounter a recurring motif, that of the lock (tāla) closing the door (duāren), serving as a skillful metaphor to refer to tantric breathing practice and visualization related to kumbhaka or vase breathing. The apparent similarity between these works highlights the challenges associated with the idea of spontaneous composition. Moreover, the fact that approximately two-thirds of Kṛṣṇacaryā’s Dohākoṣa are cited within the *Caryākoṣagītivṛtti to elucidate its content (see note 23) emphasizes that these two writings can hardly be considered generically different from each other. The observation of shared motifs between, as an example, Kṛṣṇacaryā’s Dohākoṣa and *Caryākoṣagītivṛtti 4 suggests some interrelation and, possibly, a shared thematic framework. This reinforces the notion that strict generic distinctions between these works might not be appropriate, and it urges the further exploration of their interconnectedness and literary function.

In conclusion, I propose abandoning the label translations and the attempt to define them as distinct “genre terms.” Instead, I suggest treating all related labels as nearly synonymous and closely related subcategories of the genre of tantric (song-)poetry. In its entirety, this genre can be seen as belonging to the text-type dohā or, in Tibetan, mGur, respectively. I recommend investigating the significance of the “myth of spontaneity” and suspect that this examination could significantly contribute to our understanding of the function of this text type within the Buddhist traditions.

Furthermore, exploring and analysing common themes and motifs, and the use of compilation techniques are crucial aspects in our endeavour to better understand the context of song-poetry, their evolution, and origins. These open questions need to be addressed before the text type and its possible subordinates can be meaningfully defined across centuries and cultures.