Kautilya's Arthashastra Reached Print in 1909 From a Single Manuscript Handed to a Mysore Librarian, and Scholars Still Date Its 15 Books Anywhere From the Fourth Century BCE to the Third Century CE

The Arthashastra counts itself in its first chapter: 15 books, 150 chapters, 180 sections and 6,000 slokas, or verses. R. P. Kangle, whose critical edition (first published in 1960) is the base for Patrick Olivelle’s English translation of 2013, counted the real text and found mostly prose with about 380 true slokas. He wrote that the claim “hardly appears to come from the author himself.”

A verse in the same chapter credits Kautilya, and the last verse of the work names Vishnugupta, whom later Indian tradition identifies with Kautilya and with Chanakya, the Brahman minister who overthrew the Nanda dynasty and put Chandragupta Maurya on the throne. Whether he wrote the text around 321 BCE, or a compiler put it together in the third century CE, has been argued since the Sanskrit text first appeared in 1909.

A pandit, a librarian and one manuscript

A pandit from the Tanjore district handed a manuscript of the text to R. Shamasastry, librarian of the Mysore Government Oriental Library. J. F. Fleet, who introduced Shamasastry’s English translation in a note dated 20 November 1914, says it came with a commentary by Bhattasvamin on a small part of the work. Shamasastry began printing a tentative translation in the Indian Antiquary in 1905, and Trautmann says installments ran there and in the Mysore Review until 1909, the year the Sanskrit text came out as volume 37 of the Bibliotheca Sanskrita of Mysore. Richard Davis, writing in 2015, says the work was widely cited until the early eighth century CE, fell out of use from the ninth, and “resurfaced in about 1905.”

Shamasastry had one manuscript, in Grantha script, and the partial commentary. Fleet called the text “laconic and difficult to a degree” and noted that two or three other manuscripts had by then been traced. The full English translation appeared in 1915. Julius Jolly and Richard Schmidt edited the text in 1923 and 1924 from a copy of a Malayalam-script manuscript held in Munich, and Muni Jina Vijay published a Devanagari manuscript from Patan in Gujarat, then the only known northern one, in 1959. Kangle’s edition, which collated them all, followed in 1960.

Fifteen books, seven parts of a state, twelve kings

Book I, Concerning Discipline, covers the training of the king, the testing of ministers, spies and the king’s daily timetable. Book II lists the state’s offices and workshops one superintendent at a time, from the treasury and mines to commerce, tolls, weaving, agriculture, liquor and ships. Law fills Book III, criminals and calamities Book IV, courtiers and the treasury Book V. After the parts of a state and the six forms of foreign policy in Books VI and VII come calamities, the invader, war, corporations, a weak king facing a strong one, the capture of a fortress and secret means, with the plan of the treatise closing Book XV.

Book VI opens with the elements of sovereignty. In Shamasastry’s translation they are the king, the minister, the country, the fort, the treasury, the army and the friend, which the text calls the limb-like elements, with the enemy listed alongside and set apart. It then draws the map of foreign policy, the circle of states, or rajamandala. The king bent on expansion, the conqueror, sits at the centre. The king whose land touches his is the enemy, the king beyond the enemy is the friend, and the pattern repeats in front and behind until Shamasastry counts four circles of states, twelve primary kings, 60 elements of sovereignty and 72 elements of states.

Book VII sets out six policies: peace, war, neutrality, marching, seeking another’s protection, and peace with one king while fighting another. Vatavyadhi, the text says, holds that there are only two, peace and war, but Kautilya holds that conditions differ and the forms are six. A king weaker than his rival makes peace, a stronger one makes war, and one who cannot defend himself seeks protection. Kangle argued in 1965 that the circle of states is realistic rather than pedantic, since the state next door tends to be hostile and the one beyond it friendly.

Spies on a salary table

Book I, chapters 11 and 12, builds a spy service in two groups. Five types, the fraudulent disciple, recluse, householder, merchant and ascetic, form what the translation calls institutes of espionage. The householder and merchant are farmers and traders who have fallen from their trades, the text says, and they work land the state allots them. The wandering spies are classmates, who are orphans the state feeds and trains in palmistry and reading omens, fiery spies, who are men that fight elephants or tigers for money, poisoners, and mendicant women. The king sends them to watch his own ministers, priest, army commander, heir apparent and heads of department.

Loyalty is tested as well. In one test a priest dismissed for refusing an improper order sends spies posing as classmates to ask each minister to join a plot to replace the king. A minister who refuses is judged pure.

The pay scale is in Book V. The sacrificial priest, the teacher, the minister, the army commander, the heir apparent, the king’s mother and the queen get 48,000 panas each, the pana being the text’s silver coin, and the text says that sum will keep them from temptation and discontent. Each of the five institute spies gets 1,000. Fiery spies, poisoners and mendicant women get 500, the same as trained soldiers and the staff of accountants and writers. Servants in charge of animals get 60.

Taxes, tolls and a cap on profit

The superintendent of the store-house tracks a sixth share of produce, a term its commentary says can cover larger shares. In Book I a spy is told to answer critics of kingship by saying that people suffering anarchy, where the big fish swallows the small, chose Manu as king and agreed to pay a sixth of their grain and a tenth of their merchandise for protection. Shamasastry’s footnote reads this as a social-contract idea.

In an emergency, Book V lets the king ask more. In a rain-fed region rich in grain he may demand a third or a fourth of the crop. He may not demand it from poorer land, from people on the border, from people who build roads and forts or work mines, or from settlers of waste land, whom he is told to supply with grain and cattle. The text says “Such demands shall be made only once and never twice.” The same chapter is franker about methods. It lets the king gather temple property into the treasury and raise money by pointing to an evil omen.

Tolls are policed at the city gate. Goods without the state’s seal mark pay double, a forged seal costs eight times the toll, and a merchant who passes the toll-house flag without paying is fined eight times what he owed. The superintendent of commerce fixes a profit of five per cent over the fixed price of local goods and ten per cent on foreign goods, and a trader who takes more is fined. In famine, Book IV tells the king to supply seed and provisions, to hand out his own stores or the hoarded income of the rich, or to seek help from friendly kings.

War, with the ambush written in

Book VII names three forms of battle: open, treacherous and silent, the last defined as killing an enemy through spies when nobody has spoken of war. Book X says a king with a strong army, a successful intrigue and favourable ground may fight openly, and otherwise by treachery. It lists what counts as ambush: deserts, marshes, mountains, valleys, mist and night. For the open fight it supplies a speech, in which the king names the place and time to his assembled army and tells them “I am a paid servant like yourselves.”

Invaders, in Book XII, come in three kinds in Shamasastry’s rendering. The just conqueror is satisfied with obeisance, the greedy one with what he can safely gain in land or money, and the demon-like one takes land, treasure, wives and sons and the lives of the conquered too. A weak king should shelter with the first, pay off the second and keep the third at a distance with gifts.

Book XIII, chapter 5, covers the day after a conquest. The conqueror should remit taxes, adopt the people’s dress, language and customs, join their festivals, release prisoners, help the sick and helpless, and forbid animal slaughter for half a month in the Chaturmasya season. It also tells him to move officials who conspired with the enemy to separate places along the border.

The argument over who wrote it

Fleet and Shamasastry took the text at its word. Shamasastry cited the Vishnu Purana, which says Kautilya will kill the Nandas and install Chandragupta, and Kamandaka’s Nitisara and Dandin’s Dasakumaracarita, which credit the science of polity to Vishnugupta. He concluded the work was written between 321 and 300 BCE. Fleet allowed that the surviving text might not be word for word what Kautilya wrote.

Doubters answered quickly. Kangle’s 1965 survey names J. Jolly, M. Winternitz, A. B. Keith and O. Stein among them. Keith’s objection, as Kangle gives it, is that the text never hints at the fall of the Nandas, Chandragupta’s wars, the cession by Seleucus, or even Pataliputra. Kangle replied that the Arthashastra is a treatise and not memoirs, that it names almost no places except as sources of goods, and that Sanskrit treatises seldom mention events of their day. He found no convincing reason not to treat it as Kautilya’s. The old Malayalam commentary, he reports, lacks both the statement on 6,000 slokas and the verse that credits Kautilya, although the Patan manuscript has them.

Trautmann’s 1971 book used a different tool. Counting how often small words such as eva, evam and va appear in each book, and measuring sentence and compound lengths, he concluded that the separate authorship of Books 2, 3 and 7 was well established. His supervisor A. L. Basham’s preface sums that up as the work of at least three hands. Trautmann read the text’s opening line as saying it is a compendium of earlier treatises, and held that Kautilya cannot have written the whole. Book 2 he dated from content. Sanskrit inscriptions effectively begin about 150 CE, and the book names Chinese silk, coral from Alexandria and Sri Lanka as Parasamudra, where Ashoka’s edicts say Tambapanni. That put Book 2 at about 150 CE and the compilation at about 250 CE, though the science itself, in his view, is much older. Kangle’s survey had met the silk argument earlier, noting that K. P. Jayaswal read Cina as the Shina people of Gilgit and that the line could be a later addition.

Olivelle’s translation, as Davis summarises it, finds three phases: an expert text of the mid first century BCE, a Kautilya recension of 50 to 125 CE, and a Shastric recension of 175 to 300 CE, the version in the manuscripts. Following Trautmann, Olivelle accepts that Chanakya was a real person but suggests the author was identified with him only in Gupta times, when the playwright Vishakhadatta had the final version before him. Mark McClish dates the surviving text to a third century CE redaction of a treatise written a few centuries earlier, and the summary of his Cambridge chapter says it rules out production at one time by a single author. Romila Thapar writes that the date is uncertain, that the present form is the work of a Vishnugupta of about the third century CE, and that sections such as Book II may be Mauryan. The Chandragupta piece sets out Olivelle’s phases in more detail.

The Megasthenes test

Fleet’s note said the text endorses and extends much of what Megasthenes reports. Stein’s Megasthenes und Kautilya, as Kangle summarises it, found the opposite. Megasthenes describes a wooden wall at Pataliputra, whereas the rampart passage in Shamasastry’s translation says the chariot road along the top must be made of palm trunks or stone slabs, “never of wood, as fire finds a happy abode in it.” For the city he reports six boards of five men, and the army has six more, one for each arm. In the Arthashastra, one city officer has assistants and single superintendents run infantry, cavalry, chariots and elephants.

Kangle’s reply has three parts. First, the Arthashastra is a theoretical treatise addressed to kings in general, so it need not describe Pataliputra. Second, Stein himself doubted Megasthenes on several points, among them the width of the moat, a claim that Indians kept no slaves, and a claim that only the king owned horses and elephants. Third, L. A. Waddell’s excavation report of 1903 found timber at scattered points in Patna, which Kangle doubted formed a rampart round the whole city, reading some of it as flood barriers. Thapar writes that even a Mauryan section like Book II should not be read as description, since a theoretical treatise sets out an ideal.

How it is read now

Max Weber, in “Politics as a Vocation,” published in 1919, named the Arthashastra as the classic Indian case of radical Machiavellianism and said that beside it Machiavelli’s The Prince is harmless. Davis notes that Kautilya is often called the Indian Machiavelli, but argues the comparison undersells a text that gives far more detail on the running of an absolute prince’s state than Machiavelli did. Olivelle’s 753-page translation, in his view, supplants those by Kangle in 1977 and L. N. Rangarajan in 1992.

On 18 October 2012, at a workshop on Kautilya held by the Institute for Defence Studies and Analyses in Delhi, the National Security Advisor, Shiv Shankar Menon, said that reading the Arthashastra would broaden India’s vision on strategy. He called it a “text of its time and place, Mauryan to Gupta administration.”