Chapter 13: Beyond Sanskrit: Regional Languages, Persian and the Teaching of Arithmetic¶
The history told in most of this volume is a history of Sanskrit texts. That is a fair reflection of what survives and of where the most advanced mathematics was written, but it leaves out much of the mathematics that people actually used. From about the eleventh or twelfth century onward, arithmetic was written in Kannada, Telugu, Tamil, Malayalam, the Prakrit-derived languages of the North and Persian, and it was taught in thousands of schools to children who never learned Sanskrit. Chapter 1 listed the main languages briefly. This chapter looks at the texts themselves, at the counting practices they reflect, and at the way arithmetic was taught, from the Sanskrit teacher’s household to the village school and its transformation under colonial rule. Much of this literature is unedited, and some of what follows rests on a small number of studies; the uncertainties are noted as they arise (Sarma 2011; Plofker 2009).
13.1 Why the Regional Literature Matters¶
The regional and Persian texts matter for three reasons. First, they were closer to practice. Their authors were often officials, accountants or poets writing for merchants and clerks, and they included material—local units, current coins, crop yields, building estimates—that the Sanskrit books, with their stable syllabus and conventional examples, seldom recorded. Second, they show how the Sanskrit tradition was received and adapted: most of them translate or rework Sanskrit models, and the choices they made reveal what their readers needed. Third, they widen the social range of the history. Jain merchants and officials, Kāyastha and Khatrī clerks, village accountants and Muslim revenue officers all appear among their authors or their readers.
The study of this literature has hardly begun. Lists of texts can be compiled from library catalogues and from literary histories of each language, but few have been edited critically or translated, and dates are often uncertain. S. R. Sarma’s survey of mathematical literature in the regional languages gives an overview and a warning: much of what is said about these texts in secondary works repeats unverified statements (Sarma 2011). Table 13.1 lists the works discussed in this chapter; the dates of several are uncertain.
| Language | Author and work | Date | Remarks |
|---|---|---|---|
| Apabhraṃśa | Ṭhakkura Pherū, Gaṇitasārakaumudī | c. 1318 | Sanskrit syllabus with Sultanate units |
| Kannada | Rājāditya, Vyavahāragaṇita and other works | early 12th c. | Jain author; practical arithmetic |
| Telugu | Pāvulūri Mallana, Sārasaṅgrahagaṇitamu | 11th c. (traditional) or later | Adaptation of Mahāvīra |
| Tamil | Kāri Nāyaṉār (attributed), Kaṇakkatikāram | uncertain | Practical arithmetic; named fractions |
| Malayalam | Jyeṣṭhadeva, Yuktibhāṣā | c. 1530 | Rationales of the Kerala series |
| Bengali | Śubhaṅkara (attributed), Śubhaṅkarī | uncertain | Rules for mental calculation |
| Persian | Fayḍī, translation of the Līlāvatī | 1587 | For Akbar |
| Persian | ʿAṭāʾ Allāh Rushdī, translation of the Bījagaṇita | 1634–35 | For Shāh Jahān |
13.2 Prakrit and Apabhraṃśa: Ṭhakkura Pherū¶
13.2.1 The author¶
Ṭhakkura Pherū was a Jain of the Śrīmāla community from Kannāṇa (probably the modern Kanina in Haryana) who served the Khaljī sultans of Delhi, in the treasury and the mint, in the early fourteenth century. He wrote a series of short technical works in Apabhraṃśa, the late literary Prakrit of northern and western India, drawing on his professional experience: on gems (Ratnaparīkṣā), on coins and their assay (Dravyaparīkṣā), on metals (Dhātūtpatti), on architecture (Vāstusāra), on astrology (Jyotiṣasāra) and on arithmetic, the Gaṇitasārakaumudī (“Moonlight of the essence of calculation”). His dated works fall between the 1290s and about 1318. His writings on gems, coins and metals are discussed in Volumes 14 and 17, and his Vāstusāra in Volume 16 (SaKHYa 2009).
13.2.2 The Gaṇitasārakaumudī¶
The Gaṇitasārakaumudī was edited, translated and given a mathematical commentary in 2009 by a group of scholars who published under the joint name SaKHYa (S. R. Sarma, T. Kusuba, T. Hayashi and M. Yano). Their edition shows a work with two layers. The first follows the syllabus of the Sanskrit arithmetic books, especially Śrīdhara’s, with rules for the operations, fractions, proportion, interest, mixtures, series and mensuration, in Apabhraṃśa verse. The second layer is Pherū’s own. It contains material drawn from the administration and markets of the Delhi Sultanate: the units of weight, measure and currency in use in Delhi, which mix Indian and Persian terms; estimates of the yields of crops on a given area of land; rules for estimating the materials of buildings; and a section on magic squares and other recreational topics (SaKHYa 2009).
The result is the earliest known arithmetic book in a language of northern India other than Sanskrit, and a rare document of the mathematics of a Persianate state as practised by its Indian officials. It shows a Jain mint-official of the fourteenth century keeping the Sanskrit tradition of pāṭīgaṇita while adapting it to a new political and economic order: the units are Sultanate units and the problems those of a revenue and market administration. Its crop-yield figures have attracted the attention of economic historians as well as historians of mathematics, although, like all the numbers in such texts, they need to be compared with other sources before they can be used as data (SaKHYa 2009; Sarma 2011).
13.3 Kannada and Telugu¶
13.3.1 Rājāditya¶
The earliest known author of mathematical works in Kannada is Rājāditya, a Jain who lived in the early twelfth century and is associated with the court of the Hoysaḷa king Viṣṇuvardhana. Kannada literary tradition attributes to him several works in Kannada verse, among them the Vyavahāragaṇita (practical arithmetic), the Kṣetragaṇita (mensuration), the Vyavahāraratna and a Līlāvati that is distinct from, and earlier than, Bhāskara II’s book of the same name. Their contents, so far as they have been studied, are the arithmetic of the pāṭī books—proportion, interest, mixtures and the mensuration of fields and buildings—adapted to Kannada usage. The choice of Kannada rather than Sanskrit is significant. The Jain community of Karnataka had a long tradition of writing in Kannada, and Mahāvīra’s Gaṇitasārasaṅgraha, written in Karnataka three centuries earlier, was part of its learning (Chapter 10). Rājāditya wrote for readers, many of them in trade, who wanted the content of that learning in their own language (Sarma 2011).
13.3.2 Pāvulūri Mallana¶
In Telugu, Pāvulūri Mallana rendered Mahāvīra’s Gaṇitasārasaṅgraha into Telugu verse as the Sārasaṅgrahagaṇitamu. Telugu tradition places him in the eleventh century, at the court of the Eastern Cālukya king Rājarāja Narendra of Rājamahendravaram, who is said to have rewarded him with a village; other scholars place him in the twelfth century or later, and the question is not settled. The work keeps Mahāvīra’s plan but adapts the units, the examples and much of the presentation, and it became the basis of later Telugu arithmetic. It is the clearest case of a Sanskrit arithmetic book becoming, through translation, a vernacular textbook (Sarma 2011).
13.4 Tamil: Arithmetic and Fractions¶
Tamil has its own tradition of arithmetic manuals in verse. The best known is the Kaṇakkatikāram (“Authority on calculation”), attributed to Kāri Nāyaṉār; its date is uncertain, and the surviving versions are late medieval. No critical edition or translation of it was available for this volume, and what follows rests on secondary descriptions, which should be checked against the text: the work is described as covering the operations, measures of land, grain and money, and problems of the practical kind, including rules of thumb for estimating quantities such as the number of fruit or seeds from simple measurements. Other Tamil texts include tables for teaching, collectively called Eṇcuvaṭi (“number book”), which pupils learned by heart.
The most distinctive feature of Tamil arithmetic is its system of named fractions. Instead of writing an arbitrary numerator and denominator, Tamil reckoning built quantities from a fixed set of fractions, each with its own name and written sign, related to one another mostly by halving and by division by four and five. Table 13.2 lists the commonest. Values smaller than the smallest named fraction were expressed by a second, “lower” (kīḻ) series, in which the same names were applied to a unit of \(\tfrac{1}{320}\); the lower muntiri is thus \(\tfrac{1}{320^2} = \tfrac{1}{102400}\).
| Name | Value | Name | Value |
|---|---|---|---|
| mukkāl | \(\tfrac34\) | mā | \(\tfrac{1}{20}\) |
| arai | \(\tfrac12\) | araimā | \(\tfrac{1}{40}\) |
| kāl | \(\tfrac14\) | kāṇi | \(\tfrac{1}{80}\) |
| araikkāl | \(\tfrac18\) | araikkāṇi | \(\tfrac{1}{160}\) |
| mākāṇi | \(\tfrac{1}{16}\) | muntiri | \(\tfrac{1}{320}\) |
Worked example: calculating with named fractions
Composition. The name mākāṇi (\(\tfrac1{16}\)) can be read as mā plus kāṇi, and indeed \(\tfrac{1}{20} + \tfrac{1}{80} = \tfrac{4 + 1}{80} = \tfrac{1}{16}\).
A sum. Kāl, araikkāl and mākāṇi together make \(\tfrac14 + \tfrac18 + \tfrac1{16} = \tfrac{7}{16}\), which is how \(\tfrac{7}{16}\) would be said.
A product. Three-quarters (mukkāl) of a mā is \(\tfrac34 \times \tfrac1{20} = \tfrac{3}{80}\), that is, three kāṇi.
A system of this kind suited a world of physical measures, in which a quantity of grain or land was divided into standard parts, and it had to be learned thoroughly, with tables of the products of the named fractions. Tamil numerals for whole numbers were not positional until modern times (Chapter 7), and the fraction signs, like the numerals, were used in accounts and documents well into the colonial period.
13.5 Malayalam¶
The most famous mathematical text in a regional language is the Malayalam Yuktibhāṣā (c. 1530), whose explanations of the Kerala series were the subject of Chapter 6. It was probably written in Malayalam because its author wished to explain reasons to readers who might not follow a Sanskrit verse text, and it shows that a regional language could carry the most demanding mathematics of its day. Beside it stood a more modest literature of practical arithmetic and astronomy in Malayalam, used in the village schools of Kerala, and many Kerala astronomical manuals mix Malayalam prose with Sanskrit verse.
13.6 Bengali, Marathi and Hindi: Rhymes and Tables¶
In eastern and northern India the vernacular arithmetic that survives is mostly the arithmetic of the schools: verse rules, mnemonic rhymes and memorised tables. In Bengal the best-known body of such rules is the Śubhaṅkarī, verses for mental calculation attributed to an otherwise unknown Śubhaṅkara, whose date is uncertain. The verses give short methods for the calculations of daily life: the area of a plot of land in the local units, the price of a quantity of goods from the price of one unit, wages for a number of days, and interest. They were taught in the Bengali village schools and were still remembered in the twentieth century.
In Hindi-speaking regions and in Maharashtra, pupils learned multiplication tables (Hindi pahāṛā, Marathi pāḍhe) far beyond the ten-by-ten table of modern schools: tables up to twenty or more, and tables for fractional multipliers such as \(1\tfrac14\), \(1\tfrac12\) and \(2\tfrac12\) (in Hindi savāyā, ḍeoṛhā and ḍhaiyā). These were needed because prices, weights and measures were quoted in quarters and halves, and a shopkeeper who knew the “one-and-a-quarter” table could price \(1\tfrac14\) ser of any commodity at once.
Worked example: the one-and-a-quarter table
The table of \(1\tfrac14\) runs \(1\tfrac14,\ 2\tfrac12,\ 3\tfrac34,\ 5,\ 6\tfrac14,\ \ldots\). A pupil who knew it by heart could say at once that \(1\tfrac14 \times 13 = 16\tfrac14\): thirteen units at a price of one and a quarter cost sixteen and a quarter. In the money of account of many regions, where a rupee was divided into sixteen ānās, the same table gives \(16\tfrac14\) rupees as 16 rupees and 4 ānās.
Counting itself was often done in groups other than tens. In Bengal, cowrie shells, which served as small change until the nineteenth century, were counted in fours: four cowries made a gaṇḍā, five gaṇḍās a buḍi of twenty cowries, four buḍis a paṇ of eighty, and sixteen paṇs a kāhan of 1280 cowries, as nineteenth-century accounts of the Bengal currency record. The upper part of this scale is the same as the money table at the beginning of the Līlāvatī (Chapter 1), where twenty cowries make a kākiṇī, four kākiṇīs a paṇa and sixteen paṇas a dramma: the paṇ of eighty cowries and the kāhan of 1280 correspond exactly to Bhāskara II’s paṇa and dramma. Whatever the history behind the correspondence, it shows how long such systems of account could last, and why the school arithmetic of Bengal included tables for reckoning in fours, twenties and sixteens.
Worked example: reckoning in cowries
How many cowries are there in 3 kāhan, 5 paṇ and 7 gaṇḍā?
\(3 \times 1280 + 5 \times 80 + 7 \times 4 = 3840 + 400 + 28 = 4268\) cowries. Conversely, a price of 4268 cowries is written, by repeated division by 1280, 80 and 4, as 3 kāhan, 5 paṇ and 7 gaṇḍā—the same procedure as for the Līlāvatī’s units.
13.7 Persian¶
13.7.1 Translations from Sanskrit¶
The Mughal court took an interest in Sanskrit learning, and two of the classical mathematical books were translated into Persian. The poet Abū’l-Fayḍ Fayḍī translated the Līlāvatī for Akbar in 1587; his preface tells the story of Bhāskara’s daughter that has attached itself to the book ever since (Chapter 5). ʿAṭāʾ Allāh Rushdī translated the Bījagaṇita for Shāh Jahān in 1634–35; he is identified in the sources as a son of the architect Aḥmad Lāhaurī, who is associated with the building of the Taj Mahal. The translations made Bhāskara II’s arithmetic and algebra available to readers of Persian, and the Persian Līlāvatī was itself copied and studied. They also had a second life: the first English version of the Bījagaṇita, by Edward Strachey (1813), was made from the Persian, before Colebrooke translated the Sanskrit in 1817. They were part of a wider programme of translation from Sanskrit at the Mughal court that included the epics and works on astronomy, while Islamic astronomy and mathematics travelled the other way into Sanskrit (Chapter 12) (Plofker 2009; Colebrooke 1817).
13.7.2 Persian arithmetic and accounts¶
Persian was the language of administration across much of India from the thirteenth century to the nineteenth, and the calculations of revenue and accounts were done in it. The madrasa curriculum included Arabic and Persian arithmetic, notably the Khulāṣat al-ḥisāb of Bahāʾ al-Dīn al-ʿĀmilī (Chapter 12). Revenue clerks had their own notation. Their siyāq numerals are a set of special signs derived from abbreviated Arabic number words, written in a fixed order of magnitude, which made accounts compact and hard to falsify. They were used in Mughal revenue documents and survived in some regions, notably the Hyderabad state, into the twentieth century.
Under Akbar the revenue of much of northern India was assessed by the system known as ẓabṭ: the area of land under each crop was measured, and the revenue was computed from schedules of cash rates per unit of area for each crop. The unit of area was the bīgha, a square of sixty by sixty gaz (Akbar’s gaz was roughly a yard), and Akbar’s administration standardised the gaz. The system required measurement and calculation on a vast scale—area times rate, summed over crops, villages and districts—and created a demand for clerks trained in Persian arithmetic and accounts. Many of these clerks came from Hindu scribal communities such as the Kāyasthas and Khatrīs, who learned Persian for the purpose; they carried the methods of Persian accountancy into the administrations of the regional states of the eighteenth century and into the early East India Company. The institutions are described in Volumes 18 and 19.
13.7.3 The meeting of traditions¶
In the Persianate world of India, then, three traditions of calculation met: the Sanskrit pāṭīgaṇita, read in the original and in translation; the Arabic and Persian arithmetic of the madrasa, itself built on Indian numerals and methods received seven centuries earlier (Chapter 12); and the vernacular practice of merchants and village accountants. Pherū’s Gaṇitasārakaumudī, with its Indian syllabus and Sultanate units, stands at the beginning of this meeting, and the Persian translations of Bhāskara II and the Sanskrit translations of Euclid sponsored by Jai Singh stand near its end.
13.8 How Arithmetic Was Taught¶
13.8.1 In the Sanskrit tradition¶
The Sanskrit mathematical texts were learned as other Sanskrit texts were: by memorising the verses and hearing them explained. The pupil learned the rules by heart, often before understanding them; the teacher then explained them, worked the examples on the board, and set further problems. The commentaries preserve something of this teaching. Agathe Keller’s study of Bhāskara I’s commentary on the Āryabhaṭīya shows a teacher glossing each word of a rule, setting out examples with their data, working them and verifying the results, and occasionally arguing for a rule’s correctness (Chapter 3) (Keller 2006). The examples themselves often address the reader directly as a student to be tested: “if you know the method, tell me quickly”. Learning was concentrated in families of astronomers and in the households of teachers, and access to it, as Chapter 1 explained, was limited by caste and gender.
13.8.2 In the village schools¶
Most people who learned arithmetic did so in village schools, where the medium was the regional language: the pāṭhaśālās of the North, the pyal schools of the South, held on the verandah of a house, and the maktabs for Muslim children. The best sources for them are the surveys made by colonial officials in the Madras Presidency in the 1820s and William Adam’s reports on education in Bengal and Bihar (1835–38); the Madras material and other early accounts were collected by Dharampal (Dharampal 1983; Basu 1941). These accounts allow the teaching of arithmetic to be described in some detail.
-
Materials. Pupils began by writing letters and numerals with a finger or a stick in sand spread on the floor or on a board, then moved to palm leaves (in the South) or plantain leaves (in Bengal), and only at the end to paper. The sequence spared costly materials and matched the erasable “dust-work” of the old arithmetic.
-
Tables and rhymes. Pupils chanted multiplication tables, including the fractional tables described above, and tables of money, weight and land measures, until they could recall them instantly.
-
Mental arithmetic. The aim was quick and reliable calculation, much of it in the head: prices, wages, interest and the area of land. Rules such as those of the Śubhaṅkarī were memorised for this purpose.
-
Accounts. The most advanced pupils learned the forms of commercial and agricultural accounts, the writing of letters and documents, and the keeping of books, which fitted them for work as shopkeepers, clerks or village accountants. The merchants’ account books themselves are described in Volume 19.
-
Older pupils as teachers. A single teacher could manage a large school because older pupils taught the younger ones. The Scottish clergyman Andrew Bell, who saw this practice in Madras in the 1790s, made it the basis of the “monitorial” or “Madras” system, which was widely adopted in British elementary schools in the early nineteenth century.
The schools were not open to all. The surveys show pupils from many castes in some districts, but very few girls and, in most places, few or no children of the castes then regarded as untouchable. Their arithmetic was narrowly practical and did not include the algebra or geometry of the Sanskrit books. But it was effective at what it set out to do, and it reached far more people than the Sanskrit tradition ever did.
13.8.3 The colonial transition¶
In the nineteenth century the village schools were gradually displaced by the government and missionary schools, which taught European arithmetic with printed textbooks, and with them went a great deal of practical mathematical knowledge that had never been written down. Bodies such as the Calcutta School-Book Society, founded in 1817, published arithmetics in Bengali and other languages, combining European methods with Indian units. The change was not only one of content but also of medium and method: printed books replaced memorised verse, written working on paper or slate replaced mental calculation, and examinations replaced the teacher’s judgement.
A few efforts were made to build on the Indian tradition instead of replacing it. Lancelot Wilkinson, a British political agent in central India, established a school at Sehore in the 1830s in which astronomy was taught from Bhāskara II’s Siddhāntaśiromaṇi, in the hope of introducing modern astronomy through the reasoning of an Indian classic. His pupil Bāpūdeva Śāstrī went on to teach mathematics at the Sanskrit College in Banaras, where he wrote textbooks of arithmetic, algebra and trigonometry in Hindi and Sanskrit that combined the two traditions (Chapter 14) (Plofker 2009). Such efforts remained marginal. By the end of the century the mathematics taught in Indian schools and universities was the mathematics of the European textbooks, and the Sanskrit and regional traditions survived mainly in the households of traditional scholars, in the calendars of astrologers, and in the mental arithmetic of the bazaar.
13.9 Assessment¶
The regional and Persian literature of mathematics was not, on present evidence, the source of new mathematical results; its authors adapted the Sanskrit books and the Islamic arithmetic to new readers. Its importance lies elsewhere. It records the units, coins and practices of particular times and places; it shows how mathematical knowledge was shared across languages, religions and communities—Jain officials in the service of Muslim rulers, Persian translations of Sanskrit classics, Telugu and Kannada versions of a Jain arithmetic; and, together with the village schools, it carried arithmetic to people whom the Sanskrit tradition never reached. Much of it remains to be edited and studied, and its history will change as it is.
Exercises¶
-
Express \(\tfrac{11}{16}\) as a sum of named Tamil fractions from Table 13.2.
-
How many muntiri make one mā? How many kāṇi make one kāl?
-
Using the table of \(2\tfrac12\), find the price of 17 units at \(2\tfrac12\) each.
-
A bīgha is 60 by 60 gaz. How many bīghas are there in a rectangular field of 240 by 150 gaz?
-
Under a schedule of cash rates (the rates here are hypothetical), wheat pays \(2\tfrac12\) rupees a bīgha and cotton \(3\tfrac14\). A village has 24 bīghas under wheat and 16 under cotton. What revenue is assessed?
Answers. (1) \(\tfrac{11}{16} = \tfrac12 + \tfrac18 + \tfrac1{16}\): arai, araikkāl and mākāṇi. (2) \(320/20 = 16\) muntiri; \(80/4 = 20\) kāṇi. (3) \(2\tfrac12 \times 17 = 42\tfrac12\). (4) \(240 \times 150 = 36000\) square gaz, which is \(36000/3600 = 10\) bīghas. (5) \(24 \times 2\tfrac12 + 16 \times 3\tfrac14 = 60 + 52 = 112\) rupees.
Scholarly Note: Further reading
Sarma (2011) surveys mathematical literature in the regional languages; SaKHYa (2009) edits and translates Pherū’s Gaṇitasārakaumudī, with a full introduction; Plofker (2009), chapters 8–9, covers the Persianate and early modern periods; Dharampal (1983) and Basu’s edition of Adam’s reports (1941) are the main printed sources for the village schools; Keller (2006) describes teaching through commentary (Sarma 2011; SaKHYa 2009; Plofker 2009; Dharampal 1983; Basu 1941; Keller 2006).