Kāmarūpañčāšikā | Monography or Translations of known period | Survey | Perso-Indica

Monography or Translations of known period

The oldest Persian treatise on yoga and divination by breath (svara) is the anonymous text known by the Hindi name, Kāmarūpañčāšikā or “the fifty verses of Kamarupa.” While the compiler remains unknown, the text was composed prior to 1353; that was the death date of Šams al-Dīn Muḥammad ibn Maḥmūd Āmulī, who quoted excerpts from the Kāmarūpañčāšikā in his encyclopedia of the sciences, Nafā’is al-funūn (Ernst 2011). While the Kāmarūpañčāšikā was unquestionably written in India, by an author quite familiar with Indian sources, it is known also through manuscripts circulating in Iran, from the time of Āmulī (who died in Shiraz in service to the Mongols) up to the present day. While the complete text only survives in a single manuscript, an abridged version in six chapters has continued to be popular among Shi`i scholars (e.g., in the anthology Ḫazā’in of Mullā Aḥmad Narāqī, see Narāqī 1376/1997).

The task of describing the structure and principal topics of the Kāmarūpañčāšikā is complicated by the lack of coherence in the organization of the text as it has come down to us, which is compounded by obscure features of the manuscript, such as the use of numerical ciphers to conceal certain occult practices, flawed attempts to copy Sanskrit letters, and illegible transcription of Hindi verses and mantras. Nevertheless, the preservation of versions of this text, either in abridgements or as partially incorporated into later works such as the Arabic and Persian versions of the Amṛtakuṇḍa, testifies to the ongoing engagement with yogic materials in Persianate circles over centuries. The selective nature of the appropriation of the text by Persian intellectuals also indicates some important self-imposed limitations on the translation process.

The sole surviving complete copy of the Kāmarūpañčāšikā was acquired by the Italian traveler Pietro della Valle in Persia in 1622, and it was copied for him by a group of independent thinkers (possibly Nuqṭawīs) in the southern Iranian city of Lar (Ernst 2009). The Hindi title appears five times in the text (Ms. Vatican City, Vatican Library, Persian 20, ff. 2a, 14b, 27a, 28a, 56a), and it is also cited in several manuscripts of the Arabic version of the Amṛtakuṇḍa; but it has been spelled inconsistently, so that the underlying Hindi phrase has not been easy to discern.  Della Valle himself was uncertain about the title, observing that “since it is an Indian word, the Persian copyist who wrote my original, not understanding it, has not known how to read or write it well, neither for the vowels nor the consonants, so it may be that I too write this name wrong” (Della Valle 1843, pp. 561-562). The most convincing reading of the title, proposed by Kazuyo Sakaki, is Kāmarūpañčāšikā or “the fifty verses of Kamarupa” (Sakaki 2005, p. 139). Kamarupa of course refers to the semi-mythical region now associated with Assam, which is indeed the location of several important narratives of the text; it is described in the text, using the spelling Kāmarū, as “a great city in the region at the frontier of India, and it is in the middle of the China Sea” (following the chapter divisions of Ernst forthcoming, chapter 18, paragraph 9, cited henceforth as 18.9). Less clear is why the title should refer to fifty verses, when in several places it refers to its source as a work of either 32 or 84 (or 85) verses (Ernst forthcoming, 17.1, 17.25, 8.4, 16.5). The source text does not survive as an independent work, but there is considerable overlap with the Śivasvarodaya in the chapters on divination by breath (see Danielou 1982).

The Persian text itself may be divided into twenty chapters (faṣl) plus a brief colophon, based on clear indications of section divisions, most but not all announced with descriptive titles. The majority of the chapters are relatively short, occupying two to four pages in a large script; but the last four chapters, which consist of lengthy lists of spells, range from eight pages for chapters 17 and 19, to thirteen pages for chapter 18, and thirty pages for chapter 20, amounting to over a quarter of the entire text. The contents are as follows (including folio numbers of the Vatican copy): (i) preface, 1b; (ii) [the science of the breath], 2b; (iii) on the rule of the question of the questioner, 4a; (iv) section on reading the mind, 5b; (v) section on recognizing death, 6b; (vi) section on love and hate, 8b; (vii) the beginning of the book, 10b ; (viii) section on the incantation of deeds, 11b; (ix) section on the knowledge of breath, 12b; (x) the book of imagination, written by the sages of India, 14b; (xi) section on the beginning of the book of breath and imagination, 16a; (xii) description of the šakti, 17b; (xiii) description of the places of the šakti and its stations, 19b; (xiv) description of another imagination [rhīn], 20b; (xv) description of the imagination of the soul [haṃsā], 23a; (xvi) imagining the haṃsā in the center of a flower, 24b; (xvii) [verses of Kāmak], 26a; (xviii) section [on the soul], 30a; (xix) [theurgy and incantations], 37b; (xx) [theurgy and spells], 41b; (xxi) [colophon], 56b.

The overall literary structure of the Kāmarūpañčāšikā is obscure, and the general effect is that of an unfinished work. The preface is conventional, beginning with praise of God including two Qur’anic quotations, followed by blessings on the Prophet. Then comes the description of the text: “Now, so says the translator of the book, that in India I have seen many books, on every subject with total benefit; most of their books are in verse, because one can memorize verse better, and one’s nature is more inclined toward it. I found a book that they call Kāmarūpañčāšikā, and it is one of their most precious books. They are extremely devoted to it, and in it there are two kinds of knowledge (1.3).”

These two sciences are described as “the science of imagination and asceticism” and the science of observing the breath, which is known by the Indic term svaroda. Oddly enough, the seventh chapter is then entitled “The Beginning of the Book,” giving it the appearance of the introduction of a separate treatise on the left and right breaths. In addition, the tenth chapter is entitled, “The Book of Imaginations, Written by the Sages of India,” introduced with the following words: “Know that they call this book Kāmarūpañčāšikā, and they have turned it from the Indian language into Persian.” Here too one has the impression that different writings have been imperfectly stitched together. In a similar fashion, the eleventh chapter is called “the Beginning of the Book of Breath and Imagination.” Likewise, the commentary on the Hindi verses of Kāmak Devī in chapter 17 is described as “the beginning of the book of Indian poetry and Indian verses.” Similarly, chapter 18 on yoginīs begins thus: “You should know that this book is one of the books of the Indians that they have transmitted, and somebody among their Brahmins and scholars wrote it, and it was conveyed from the Indian language to Persian.” Occasionally, the translator indicates that some of his sources are oral: “I learned the spell from a great Brahmin, and I chose it and wrote it from their books […] These are the spells and the theurgy of the Indians, and the summoning, which we have gathered together in this [book]” (20.7).

These multiple statements about the character of the book underline the impression of its composite character. The translator gives evidence of a certain frustration in dealing with these diverse materials. Periodically he cuts short his explanations with the apology that he does not wish to make the work overly long (15.4, 15.7, 18.16). Cross-references are mixed up; in 4.1, the translator refers the reader back to the introduction for an explanation of the five breaths, a discussion that actually occurs later, in 7.2. The treatise ends with the abrupt announcement, “This is the book Kāmarūpañčāšikā ,” but then one additional spell is included, as an afterthought.

Indications of authorship are vague to nonexistent. The translator does provide some personal data, offering an eyewitness account of events at the shrine of Kāmak Dev (6.10), and he records a successful cure of poisoning (11.6, 18.4) and the summoning of a yoginī (19.8). He relates his discussion of mantras with Hari the Brahmin (19.8, cf. 20.7). The translator abruptly announces his presence in the text for no reason, introducing a new topic with, “I say” (13.2, 14.1) or “I will tell” (19.1, 19.6); towards the end of a list of spells in the final section, he states, “Now we resume the translation of this book” (20.61). One assumes a single translator (1.3), but occasionally multiple translators are assumed: “Know that they call this book Kāmarūpañčāšikā, and they have turned it from the Indian language into Persian” (10.1). Sometimes passive constructions leave the matter ambiguous: “it has been explained in Persian” (17.1); “somebody among their Brahmins and scholars wrote it, and it was conveyed from the Indian language to Persian” (18.1). Even when single authorship is claimed (6.10), he receives editorial assistance: “Then I turned it from the Indian language to the Persian language, taking great pains. It was read with a group of brahmins and scholars, and it was compared, corrected, and clarified” (10.4). But at the same time he is contemptuous of them: “I presented these verses to a group of the scholars of India, the Brahmins, and the jogis. They could not explain the commentary, and were incapable of understanding that, because the words are strange and difficult” (17.13). So the relationship of the translator to the text remains ambiguous, particularly with regard to the local experts who were his informants.

As the table of contents indicates, there is much diverse content to be found in the Kāmarūpañčāšikā. The chief topics, spread out disconnectedly in the twenty chapters, include divinatory qualities of the breath for success in various activities, mind-reading, predicting death, maleficent magic, gaining support of the powerful, visualizing mantras, the nine cakras, summoning the yoginī goddesses by tanjīm (astral magic or theurgy), Kamarupa, etc. The yoginīs are presented as 64 female immortals capable of bestowing miraculous favors, including the traditional 8 siddhis or paranormal powers (20.58). Mantras (including some attributed to Islamic prophets) are provided with detailed instructions for visualization and preparation of magical paraphernalia, for purposes ranging from healing to black magic. The text devotes considerable attention to the apex cakra, normally sahasvara (thousand petalled lotus, a variant form of sahasrara) but here almost always spelled sfara (28 times, once as svara).

Six Hindi verses are presented in Persian script in chapter 18 with Persian translation and commentary (17.2-17.12). These are identified as extracts from a 32-verse poem by Kāmak, on divination by breath. Kāmak is in effect treated as the author of the text, and is repeatedly quoted. She is described (10.2, 18.9) as a sorceress, dwelling in an enormous cave in Kamarupa where she receives offerings, accompanied by jogis, magicians, and the 64 yoginīs. There is little said to connect Kāmak to the contemporary Kāmakhya temple in Assam, aside from a brief reference to a stone from which white water flows, perhaps an oblique acknowledgement of the red spring identified as the goddess’s menstrual blood in Puranic legend.

The presentation of Indic material in the Kāmarūpañčāšikā is obscured to some extent by the use of codes or ciphers to conceal topics evidently considered to be controversial or otherwise worth hiding from the unappreciative. In one case (6.4), a single word is rendered twice by figures resembling the occult alphabets familiar from Arabic writings such as the Šams al-ma`ārif attributed to al-Būnī (d. 622/1225). Comparison with a parallel passage in an abridged version of the Kāmarūpañčāšikā (Ms. Cambridge, University Library V.21, fol. 62b) indicates that the concealed word is “destroyed” (halāk), employed in a murderous ritual casting ingredients into a fire with the hope of killing an enemy. In an addendum at the end of the book (20.66), a charm for healing is accompanied by other enigmatic figures.

More frequently, the text employs a numerical substitution code to conceal certain phrases, a practice that was widespread in premodern Persian (Bosworth 1992). On examination, this code turns out to be a fairly simple one, based on the standard abjad system of numerical equivalents of the Arabic alphabet. Most of the instances of this use of numerical code can be satisfactorily solved in this manner. In another murderous curse (14.5), the practitioner visualizes the mantra hrīm in black to destroy someone, with the word “destroy” written in code. In numerous locations, the phrases “he dies” (six times, 14.10, 18.37-41), “he never dies” (five times, 17.20-24), and “he does not die” (three times 14.10, 15.2, 17.31,) are conveyed in code, although in some places the scribe omitted the extra dots denoting tens and hundreds, as if in haste. There are also three longer statements written in code, mostly decipherable. One describes the near-immortality of the “spiritual beings” (i.e., the yoginīs): “they live for a thousand and ten thousand years” (15.3). Another (12.3) has to do with attaining the invulnerability that yoginīs possess, so that neither fire nor water affects one. A third promises that, like them, one “becomes long-lived and [lives] many years” (15.3).

The repeated use of such a simple code raises the question of why the scribe employed a masking technique that in most cases could be easily figured out by an energetic reader. Given that all these phrases have to do with death, murder, and immortality, one may speculate that the scribe was not seriously attempting to make the text inaccessible, but that he may have felt a desire to insulate himself, as it were, from potent and in some cases malevolent materials, that were literally matters of life and death. In any case, in the translation of a text of occult power from Sanskrit to Persian, the presence of such deliberate esotericism indicates that there were certain subjects that aroused discomfort and hesitation among at least some readers. The use of these ciphers may be an idiosyncratic reflex of the particular scribe who copied this manuscript, but it does represent one possible reader response to the material. Determining the limits of what can be translated is a problem that can be illuminated by examples like these.

Alongside these coded alphabets, there is another feature of the Kāmarūpañčāšikā which must have remained a mystery to a majority of Persian readers: the attempts to copy various mantras in an Indic script, probably a form of devanagari. The popular mantra hrīm (invariably spelled rhīn in Persian script) is employed over twenty times in the text; it is also presented in graphic form (14.1) along with a number of remarkable claims: “This word, which in the Indian language is [hrīm], is the source of all magic and the root of all spells. This word is the great name of God (great is his majesty), and the meaning of this word is “merciful” (raḥīm), but written in our script it is rhīn.”

It seems clear that the Persian scribe could not attempt to copy the Sanskrit mantra without drawing letter shapes strongly reminiscent of the Perso-Arabic alphabet. Looking at this grapheme without any further description, one would be hard pressed to see here a genuine representation of a Sanskrit word. Indeed, the graphic representation of the mantra accomplishes an Islamization parallel to its description as the name of God, not to mention the pun on raḥīm – claims that are repeated later on in the text (20.26). Other mantras, such as ā’īn (Sanskrit aiṃ) and hansā (Sanskrit haṃsa) are also depicted. One passage (19.6) pulls together five mantras considered to be fundamental: “Now I will tell the principle of all the spells. You should know that, though all the Indian spells are numerous, yet the principle of all is these five words: 1) ā’īn, 2) rhīn, 3) srīn, 4) pahn [phat?], 5) hasūm[haṃsa?].” With some repetition and a confusing layout, four of these five mantras are presented both in Persian transcription and in a vague approximation of the devanagari script. Whether any readers were able to appreciate the depiction of Sanskrit mantras by using the shapes of Persian letters is difficult to guess.

Although this is a unique surviving complete copy, there are several anonymous abridged versions of the Kāmarūpañčāšikā that all focus on the first quarter of the text, usually in undated manuscripts in six chapters but with slightly different selections. By omitting the last three-quarters of the text, they leave aside the extensive practices relating to šakti, mantras, and yoginīs found in later chapters. These versions may be distinguished as follows: extracts in encyclopedias, such as Amulī’s Nafā’is al-funūn, Samarqandī’s Jawāhir al-`ulūm-i Humāyūnī (16th century), and the Qajar-era Kašf al-ṣanāyi`; a six-chapter version used by Sufis, typically entitled Risāla-yi `ilm-i dam wa wahm or Risāla-yi `ilm-i żamīr; a similar five-chapter version circulated among Shi‘i clerics in Persia who were interested in occult sciences (Narāqī 1376/1997); a Qajar-era version acquired by E. G. Browne (Ms. Cambridge, University Library V.21/3, dated 1223/1808); and also an Arabic version called Ḫawāṣṣ al-anfās (Properties of Breaths). It is clear that the Kāmarūpañčāšikā was consulted by the translator of the Arabic version of the Amṛtakuṇḍa, Mir’āt al-ma`ānī (16th century), particularly for the ninth chapter on summoning yoginīs. And its recipes for divination by breath coincide with practices found in chapters 2 and 3 of the Baḥr al-ḥayāt of Muḥammad Ġawṯ (d. 970/1549). In short, the Kāmarūpañčāšikā was an important source for the dissemination of yogic practices in Persian texts, both in India and beyond.

i) Place of copying; ii) Period of copying; iii) Copyist; iv) Commissioner;
v) Information on colophon; vi) Description of miniatures/illustrations; vii) Other remarks; viii) Information on catalogue(s)

Vatican City, Vatican Library, Persian 20, ff. 56

, i) Lar, Iran, ii)

25 rajab 1031/5 June 1622

, iii) Mullā Jamāl ibn Muḥammad Bahrām Nayrīzī, iv) Pietro Della Valle, viii)

Rossi 1948, pp. 47-49


English translation: Ernst, Carl W.  —, D'Silva, Patrick J., 2023, Breathtaking Revelations: The Science of Breath, from the Fifty Kamarupa Verses to Hazrat Inayat Khan, Richmond, Virginia, Suluk Press. 


Bosworth, C. E., 1992, “Codes,” Encyclopaedia Iranica, vol. 5, pp. 883-885, available online at http://www.iranicaonline.org/articles/codes-romuz-sg. Daniélou, Alain, trans., 1982, Le Shiva-Svarodaya: La naissance du Souffle de Vie révélé par le dieu Shiva, Ancien Traité de Présages et Prémonitions d'après le souffle vital, Milan, Archè. Della Valle, Pietro, 1843, Viaggi di Pietro della Valle il pellegrino, descritti da lui medesimo in lettere familiari all'erudito suo amico Mario Schifano, divisi in tre parti, la Turchia, la Persia e l' India, colla vita e ritratto dell'autore, Brighton, G. Gancia, vol. 2. Ernst, Carl W., 2009, "Being Careful with the Goddess: Yoginīs in Persian and Arabic Texts.” In: Performing Ecstasy: The Poetics and Politics of Religion in India, Pallabi Chakrabarty – Scott Kugle, eds., Delhi, Manohar, pp. 189-203. Ernst, Carl W., 2011. “A Fourteenth-Century Persian Account of Breath Control and Meditation.” In: Yoga in Practice, David Gordon White, ed., Princeton Readings in Religions, Princeton, Princeton University Press, pp. 133-39. Narāqi, Aḥmad, 1376/1997, Ḫazā’in, Ḥasan Ḥasan-zāda Āmulī, ed., Qum, Našr-i Qiyām. Rossi, Ettore, 1948, Elenco dei manoscritti persiani della Biblioteca Vaticana: Vaticani, Berberiniani, Borgiani, Rossiani, Studi e Testi 136, Vatican City, Biblioteca Apostolica Vaticana. Sakaki, Kazuyo, 2005, “Yogico-tantric Traditions in the Ḥawḍ al-Ḥayāt,” Joumal of the Japanese Association for South Asian Studies, 17, pp. 135-156.

Carl W. Ernst

Originally published: 20 July 2019
How to quote this article:

Ernst, Carl W., 2019, "Kāmarūpañčāšikā", Perso-Indica. An Analytical Survey of Persian Works on Indian Learned Traditions, F. Speziale - C. W. Ernst, eds.

available at http://www.perso-indica.net/work/kamarupancasika.

Download article
Main Persian Title: Kāmarūpañčāšikā
English Translation of Main Persian Title: The Fifty Verses of Kāmārūpa
Approximate period of composition: 1300-1350
Place: Northern India
Other Languages of Work: Old Hindi/Hindavi
Local Informants: Brahmins,  Yogis
Later texts quoting this Work:

Nafā’is al-funūn   

Jawāhir al-`ulūm-i Humāyūnī

Kašf al-ṣanāyi’   


Baḥr al-ḥayāt