Entrepreneurs in diplomacy: Maratha expansion in the age of the vakil

In eighteenth-century South Asia, ‘political’ vakils are familiar to us principally as diplomats, active in the inter-state negotiations of the period. They were unlike their predecessors, the īlchī and hejib of earlier centuries, who were associated with the service of courts and states. Maratha political vakils, like others, worked rather more as the mobile agents of individual rulers. Their activities extended far beyond the diplomatic arena. Since revenue rights were central to many inter-state negotiations, vakils often oversaw arrangements for local-level revenue collection. Frequently acting on behalf of several employers, they also had key roles in the remittance of cash, to meet the costs of their own establishments, to participate in the gift economy of the court, to pay the costs of local mercenaries, and to make down-payments for revenue farms on behalf of their employers. Drawing on support of their own extended families, for whom vakil service was often a profession that extended over several generations and regions, many political vakils combined mobility with deep connections to local economies and societies, sharing some characteristics of the ‘portfolio capitalism’ of the eighteenth century. What distinguished them, though, was their access to subcontinent-wide networks of political intelligence, and their expertise in the ‘soft skills’ of negotiation and persuasion, which further enabled them to exploit local social networks and political institutions. Colonial reforms of the late eighteenth century broke this flexible and entrepreneurial service role apart, dissipating it within the lower levels of colonial bureaucracy. The old figure of the political vakil disappeared, to be replaced by the semi-professional ‘native pleader’ in courts of law, and by ‘munshi’ assistants and translators to the Residents of the princely states within the uncovenanted civil service.


Introduction
Writing in 1784, the Hyderabad chief secretary Lala Mansaram recalled the arrangements of Nizam ul Mulk (r. 1724-48) for the vakils or agents of other chiefs as they arrived at the Hyderabad court to represent the interests of their masters: The Nizam ordered: 'The vakils of many notable chiefs arrive at the camp individually. Many of them are put to great distress. It is better if all these remain together. A locality should be assigned to them. It should be known as the vakilpura. It would be convenient for the vakils to stay there'. Devidas was a vakil on behalf of Sadatulla Khan and the Afghans of Kapada and Karnool. He was a rich man and had elephants for his conveyance. He used to maintain a free kitchen in his camp and feed the poor. The vakils of other chiefs started settling near the camp of Devidas. 1 Lala Mansaram's recollections offer a vivid example of the rapidly expanding numbers of those serving as diplomatic agents at the courts of eighteenth-century India, as well as of the sharp contrasts between the wealth and prestige of some of them, and the straitened circumstances of others. As regional states and smaller political players of all kinds came to terms with the stronger multilateral relationships that developed across the subcontinent in the wake of Mughal decline, a vakil's presence at other courts assumed new importance. Deriving from the Arabic wakāla, 'one entrusted', the title referred in fact to any person with delegated authority to act on behalf of another. 2 In practice, this could cover many political, juridical, commercial and administrative functions. For any historian of the eighteenth century, however, it is the 'political' vakil that stands out most prominently, sent to other courts to conclude agreements, oversee intelligence sent back home, and negotiate the complex diplomatic protocols of the court to maintain his employer's prestige.
Consequently, we know much about vakils as part of the information and diplomatic order of eighteenth-century India, as well as the apparatus of intelligence gathering and reporting that vakils usually oversaw. Yet we know rather less about what it actually meant to do the work of political vakils: their social and career aspirations, their understandings of service to individual employers as opposed to service to the state, the portfolio of property and rights they might accumulate and the kinds of influence their position and skills enabled them to wield, and the role of family in their peripatetic lives. This essay will explore these questions in the lives of some Marathi-speaking political vakils, whose association with the Maratha empire took them to every part of the subcontinent.

Diplomacy as a Portfolio Profession
Like the 'new diplomatic history' elsewhere, studies of diplomacy in pre-colonial India now extend well beyond communiques and treaties to explore diplomacy as a multifaceted dimension of wider power relationships. 3 Christopher Bayly has examined the intelligence gathering operations of eighteenth-century states, and the apparatus of news-writers, posts, runners and spies which underpinned vakils' reporting. 4 Michael Fisher has explored the way in which the 'residency' system of the East India Company gradually squeezed Indian vakils out of the flows of information between Indian courts. 5 In important new work, Dominic Vendell has drawn our attention to the language of political ethics that vakils, ministers and others employed in political negotiations. Vendell notes the emphasis placed on jab-sal, the conversational to-and-fro of negotiation, as a means to establish trust, alongside other material ways of affirming friendship in oaths, gifts, and ritual exchanges of religious tokens. Jab-sal here constitutes a distinctive form of politics as communicative action, concerned not with abstract notions of sovereignty, but rather with the processes through which political alliances or agreements could be shaped and carried forward. 6 Alongside their work in the inter-state culture of diplomacy, however, political vakils also became deeply involved in many areas of state and military finance. In this era of intense competition for control of agrarian revenues, a key dimension of diplomacy lay in the conduct of agreements about revenue rights. The widespread practice of revenue farming involved political vakils in the supply and disbursement of cash. They ensured that cash gathered in by local collectors was moved on to where it was required, to meet the needs of troops, to pay court officials for a wide range of services, to provide gifts on important diplomatic occasions, and to meet the significant costs of their own establishments. It was also the task of vakils to make the down-payments on revenue farms for employers in this lucrative line of business. If merchant and banking families supplied the credit, it was very often vakils who cashed the hundis and physically carried the cash to where it was needed.
These other dimensions of a political vakil's work emerge clearly in studies of regional courts. In Jaipur, vakils stood surety for their masters when they made down-payments for revenue farms, paid court officials to procure estates and titles for their masters and to ensure attention for their masters' business and used cash to pay their own retinues of news-writers and runners. But cash shortages were also a perennial challenge, with salaries unpaid and bankers reluctant to honour their credit. 7 At the Maratha court, formal procedures governed vakils' arrival and departure and the complex mix of salary, court allowances and one-off rewards in cash, land or hereditary office through which they were paid. 8 This necessary engagement with local financial as well as political networks meant that vakils often developed their own agents and clients, and portfolios of interests in banking and revenue farming alongside privileged land tenures. 9 Their local embeddedness enabled many of them, quite openly and routinely, to take on work for several employers at a time. 10 The joint family assumed major importance in managing these extended interests. Many vakil families developed into dynasties, involving several generations and branches of the same family. In these practices, vakil families shared aspects of the culture of 'portfolio capitalism' that Bayly and others have identified in the precolonial centuries, when family and mercantile corporations accumulated interests in a wide range of activities, from commerce to revenue farming, office-holding to privileged land tenures. 11 In addition, though, successful vakils had access to subcontinent-wide intelligence networks, expertise in the language of persuasion and skill in the management of relationships large and small. This unique assemblage of skills, experience and resources made them one of the most numerous and influential service communities of eighteenth century, whose importance we have not sufficiently appreciated.

Envoys and Ambassadors in Indo-Persian Scribal Culture
How did the service communities that we identify as the diplomats of eighteenthcentury India come to be so deeply engaged in these other roles? In pre-eighteenthcentury India, and throughout much of the Islamic world, a range of titles other than 'vakil' commonly described envoys of different kinds. Most commonly used were the terms safīr, 'envoy', rasūl or messenger, and most commonly, the term elči, a Turkish word for an ambassador, which also entered Persian as īlchī. 12 The term elči, from el or il, people, country or state, with the occupational suffix či, translates literally as one who serves or works for his country. 13 Īlchī seems therefore to denote a person serving from one court or state representing its interests at another.
In Mughal India, an īlchī described both an envoy sent to courts outside India, and those at other Indian courts. Abū al-Faz̤ l's Akbarnāma reports that a Portuguese messenger claiming the status of īlchī petitioned Akbar in 1573 to help Christians caught up in the siege of Surat. 14 Ibrāhīm 'Ādilshāh of Bijapur sent his īlchī, Shah Khalīlullah, to Shāh 'Abbās of Iran to offer an alliance against the Mughals. 15 Courts within the subcontinent also exchanged īlchī: the envoy, Isma'il Mita, sent from the ruler of Bengal Nusrat Shah to Babur in April 1529 was an īlchī. 16 Perhaps because of their closer links with the Indian ocean world, the term most commonly used in the Deccan states was hejib, from the Arabic al-ḥājib. Connected with ḥījāb, signifying the curtain screening the monarch from the gaze of courtiers, the term has applied to a variety of roles over its long history, all suggesting a connection to a royal court. 17 The Deccan states used the term equally for envoys sent to the Mughal court, to other Indian states, to the European companies, and to courts abroad. Sherwani notes that the Qutb Shahis used three terms: rasūl or adhoc envoy for a particular purpose, Hājib-i Muqīmī, denoting a resident envoy, and Hājib-i Maṣliḥatī, an envoy on a special mission. 18 In practice, however, the term was used loosely. The Nīzām Shāhī embassy of Chand Bibi to Akbar's son Murad in 1596 was termed a ḥījāb. 19 Following Persian diplomatic usages more closely, the Bijapur court of the 'Ādil Shāhīs termed their ambassadors īlchī, safīr, rasūl and ḥājib. 20 At the Bijapur court of Shahaji Bhosle, the poet Jayarama Pindye composed the Rādhāmādhavavilāsacampū, which described Shahaji sending a hejib to the Mughal court. 21 The Maratha court of Shivaji employed the term widely. 22 Its usage suggests that a hejib should be a dignified person, able to oversee the exchange of solemn pledges. The late seventeenth century chronicler Sabhasad described how two hejib, Krishnaji Bhaskar and Gopinath Bokil, otherwise known as Pantiji Pant, were sent to negotiate guarantees of safety before the 1659 meeting between the Mughal general Afzal Khan and Shivaji. As Afzal Khan's hejib, Krishnaji sought assurances for his master. The Khan in turn offered a hastapanjāri, an oath confirmed with a palm-print placed on the written document, itself clearly part of the array of bodily and ritual exchanges that Vendell has described. Shivaji's hejib Pantaji Pant stood as guarantor, his solemn oath delivered before witnesses and his status as a holy Brahman serving to reassure the Khan. 23 Some uses of the term suggest that a hejib's appointment might also extend for some period. Sabhasad described how Pralhad Pant, son of Niraji Pant and a man of great wisdom, was placed as hejib at the court of Bhaganagar. 24 The Maratha term is ṭhevale, 'placed', suggesting more than a specific short term mission. The chronicler Dattaji Vaknis uses the same term, describing how, in the wake of Shivaji's raid on the Keladi chief Shivappa Nayak's city of Basnur, Shivaji sent a hejib to restore friendly relations. After the raid, 'Uma Pandit was always placed, nehami thevale, in that location as hejib. From that day, there was friendship with Shivappa Nayak'. 25

From īlchī and hejib to Vakil
Alongside the roles of īlchī and hejib went a different kind of agency, that of vakil, associated with authority delegated directly from one person to another. Great ministers of the Bahmani kings such as Mahmud Gawan and his successor Malik Hasan enjoyed the title of vakil-us saltanat. 26 The Mughals likewise employed great plenipotentiaries of the kind that Abū al-Faz̤ l described, 'the emperor's lieutenant in all matters connected with the realm and the household'. 27 Vakils as lesser agents were ubiquitous figures in the houses of the powerful, as in the case of Gobind Das, vakil of Raja Suraj Singh Rathor, described in the memoirs of Jahangir as having murdered the Raja's nephew during a private quarrel, but whose services as a vakil were so valued that he was reluctant to punish him. 28 At a more modest level, Mughal officers maintained their own agents at the imperial court. 29 Lesser lords and men of substance in provincial societies equally needed agents to conduct their business at the local level, in the offices of state officials, army camps and the houses of merchants and bankers. 30 Seventeenth century contemporaries made a clear distinction between these kinds of delegated authority. Sabhasad describes how Raghunath Pant Korde had been serving as hejib at the Mughal court. To prepare the ground for Shivaji's visit in 1666, the Rajput Jaisingh sent his vakil, along with Korde, now serving as Shivaji's vakil. 31 The Jedhe family history also demonstrates a clear sense of the distinction between the roles of hejib and vakil. The entry for May 1672 tells us that Shivaji's hejib Niraji Pant went to Bhaganagar and negotiated an alliance in return for a lakh of gold hons. In November of that year, the peace treaty with the 24 ibid., p. 104. 25 Vakaskar,Śrī Śivachhatrapatīcī 91 Kalamī Bakhar,no. 65,p. 46. 26 Anwar,Mughals and the Deccan, Abū al-Faz̤ l ibn Mubārak, Ā'īn-i Akbarī, Vol. 1, p. 4. 28 Beveridge, Memoirs of Jahāngīr, pp. 291-92. 29 Fisher, 'The Resident in Court Ritual', pp. 425-26. 30 Calkins, 'A Note on Lawyers'. 31 Heravadkar,Sabhāsad Bakhar,p. 55,61. 'Ādil Shāhī court broke down, and so Babaji Nayak Punde, who had been serving as vakil there, returned home to Shivaji's court. 32 Envoys between states in the India of the eighteenth century, however, came everywhere to be understood and titled not as īlchī or hejib, but as vakil. How did a term associated with personal agency also come to signify a political vakil's role in representing the interests of states? Christopher Werner's observations about the complex nature of delegated authority in eighteenth-century Iran, with its own multi-layered communities of Persian-trained administrators paralleling those of India, is helpful here. Werner suggests that three overlapping semantic fields were present in the wide application of the title of 'vakil' to many different kinds of delegated agency. These were the juridical, with the vakil as a legal agent or proxy; the honorary, in which the vakil represented the ruler or some connected aspect of state power; and the administrative or commercial, in which the vakil conducted business on behalf of his master in many mundane local settings. Despite this breadth of application, the term 'vakil' retained its core meaning of an authorised agent or proxy acting on behalf of another person. 33 What Werner suggests for Iran is helpful in the Indian context. A political vakil could flexibly encompass the roles of īlchī and hejib, but at the same time import them into the familiar world of personal connection, as well as agency in financial transactions, which still lay at the heart of so many eighteenth-century political institutions. This very interdependence of the financial with the political meant that questions of trust, honour and ethical conduct were more, rather than less important. Vendell's conception of the significance of jab-sal, the public affirmation of trust, is particularly helpful here. As we will see, however, vakils deeply embedded in these plural roles developed not only concepts of ethical negotiation as practical politics in action, but also a language of the well-being of the state itself.
These multiple roles help us to understand the perplexities of officers of the East India Company when they cast around for an appropriate term for their own envoys to Indian states. As Fisher has described, they sought the opinion of a local expert on diplomatic practice, who told them 'I understand Vakeel, but what is the meaning of Public Minister, I do not know, Vakeel is one thing and Elchei is another'. 34 Significantly, one of the few eighteenth-century occasions in which īlchī continued to be used was when an envoy was sent to a court outside India. When the Mysore ruler Tipu Sultan sent his envoys to the court of Louis XVI in August 1788, the medals struck in their honour celebrated their visit as the īlchī from Mysore. 35 Portuguese diplomatic usages from Goa also reflected these plural roles. The Estado da India issued clear instructions regarding the hierarchy of different diplomatic visitors and the reception to be accorded to each: a sumptuous reception 32 Kulkarni, Jedhe Śakāvalī-Karīnā, p. 69. 33 Werner, 'Ambiguity in Meaning', pp. 317-25. 34 Fisher,Indirect Rule in India,p. 49. 35 Lafont,175. I thank Shailen Bhandare for this reference.
for the embaixador of the Mughal emperor, and a more modest welcome for the enviado of lesser and local courts. 36 But the Portuguese used a further term, procurador, which shares something of the multivalent meanings of the term 'vakil', to denote broadly agents deputed to act on their behalf. As we will see, Bahiropant Mehendale, the peshwa's vakil to the British Resident at the Pune court from the mid-1780s, was also appointed as procurador to the Portuguese, responsible for reporting intelligence from Pune to Goa, and resolving trade-related disputes between the two courts. 37

Political Vakils in the Maratha State
Whilst they shared much in common with political vakils elsewhere in India, the Marathi-speaking vakil community also revealed distinctive features. After his accession to the throne in Satara in 1708, the Maratha Raja Shahu revived his grandfather Shivaji's claims to chauth and other levies in the six Mughal provinces of the Deccan. Granted after the visit of the first peshwa Balaji Vishvanath Bhat to Delhi in 1719, the claims brought the Marathas into direct conflict with the emerging Nizam Shahi state. Their political rivalry shaped the politics of central and southern India for the rest of the century, exacerbated by the further right, given under the grant, of Marathi revenue collectors themselves to be stationed in Hyderabad's territories. 38 Chauth claims became a general legal lever as individual Maratha sardars sought to establish domains of their own elsewhere in the subcontinent. 39 The drive to expansion itself, and the legal claims accompanying it, meant that Maratha scribal professionals quickly found service as political vakils at the leading edge of Maratha expansion.
These claims, and the growing needs for war finance that they serviced, meant that many vakils were directly involved in negotiations over chauth tribute. They also played a key role in transmitting local intelligence about how far chauth was being successfully collected on the ground by the kamavisdars who followed Maratha raiding parties, taking contracts for revenue collection and transforming military levies into regular civilian taxes. 40 The role of kamavisdar was itself a means of upward mobility, offering the chance not only of substantial reward in salary and expenses, but also a more direct connection with the peshwa, since, like political vakils, kamavisdars also worked directly for him. 41 There were other parallels between the roles. 42 Kamavisdars' experience of geographical mobility, 36 Melo,'Respect and Superiority', Sen,Studies in Indian History,. 38  collecting local intelligence, negotiating rights to chauth, managing a small team of clerical assistants and overseeing the transmission of large amounts of cash transferred readily to the role of a political vakil. Both had to raise credit and make loans on behalf of their employers. Political vakils were often remunerated through estates assigned as ijarah revenue farms, which again drew them close to the operations of revenue collection. Many who started out as kamavisdars went on to become political vakils. Maratha vakils were also almost exclusively Brahmans, as opposed to the heterogeneous population of political vakils. 43 There were exceptions: the north Indian Kayastha Lala Sevakram and his extended family served in many vakil roles in Satara, Jaipur and Calcutta. 44 Beniram Pandit and his brother Vishvambhar worked first for the Bhosles of Nagpur, and then as agents of Warren Hastings in Banaras. 45 The great majority of vakils, however, both those employed by the peshwas and by the Maratha sardars, were Marathi-speaking Brahmans. Some were family relations of the peshwas, and many others members of the same Chitpavan Brahman caste community, appointed for their administrative skills, their prestige and their universally recognisable identities as Brahmans. 46 The documentary sources for this essay reflect this social make-up of those serving as Maratha vakils. As agents in the diplomatic world, their work took place in a multilingual environment, which required familiarity with Persian as the lingua franca of the subcontinent's diplomatic world, with the vernacular language of the local court where they served, and well as with their own Marathi language. 47 As Michael Fisher has described, the courts of the eighteenth century usually despatched professional news-writers to work alongside political vakils, their reports accompanying and providing contexts for the letters of political vakils themselves. News-writers' letters were usually in Persian, and the Pune court routinely employed Persian-acculturated Kayastha or Khatri families for the role. 48 However, the distinguished Parasnis family of Pune, the hereditary Persian language secretaries to the Pune court, were Deshastha Brahmans, who had acquired their proficiency in Persian during a family member's period of service at the Adil Shahi court in Bijapur. 49 In the later part of the eighteenth century, the peshwas also received Persian news-letters from what seems to have been a commercial syndicated news service operating 43  across a number of north Indian courts, whose news-letters bore the distinctive seal 'Khemkaran Mansaram' on them. 50 For the most part, however, there seems to have been a clear division of labour between Maratha vakils, who sent back home letters and reports written in Marathi in the Modi cursive script, and the Persian scribal secretaries responsible for translating incoming Persian news-letters and diplomatic communications into Marathi, and translating the outgoing Marathi responses from the peshwa's and other Maratha courts into appropriately ornate diplomatic Persian. As we will see, there is some evidence, indeed, that even when Maratha Brahman vakils were clearly fluent in Persian, and writing on their own account to Persian-speaking correspondents, they preferred to write in a heavily Persianised Devanagari, rather than using Perso-Arabic script.
The great bulk of Maratha vakils' letters that survive, and that form the basis for this essay, are therefore in Marathi. They follow a common format. They open with greetings to the recipient, confirm the date and place of reception of the letter, and summarise its questions and instructions, before launching into the body of the vakil's report. As we will see, the tone of the letters is varied, depending on the relationship between the vakil and his employer. Some letters are deferential in the extreme. Others, from vakils whose families were relatives of the peshwas, or members of Brahman families long in the service of the peshwas, spoke to their employers very much on terms of equality, even upbraiding them periodically for failing to follow advice or neglecting to send the funds essential for a vakil's effective work. In these cases, the Marathi of the letters has a domestic and familial tone to it, written very much in a Sanskritised Brahman idiom, and with very few of the Persian words that occur in so many Marathi administrative documents of this period. 51

Diplomacy and Finance: Vakili as a Family Enterprise
While the joint household offered an invaluable resource for many scribal and service families, it offered particular advantages in a political vakil's work. 52 As we will see, vakils clearly learned much of their trade by working alongside their fathers and brothers. Fathers, sons and brothers offered a pool of talent that could be circulated around different courts, contributing to better coordination in the reporting of intelligence. Trust between family members was important in managing flows of cash between countryside, court and army camp. Family connections offered reassurance in the conduct of complex and long-distance political relationships. 50  Families diversified their sources of employment and improved their prospects by moving from kamavisdar to vakil work.
Best known of the Maratha vakil families from this era were the Hingnes, originally priests of the peshwas at the shrine town of Nasik. Mahadevbhat Hingne accompanied the first peshwa Balaji Vishvanath on his mission to Delhi in 1719. During the 1720s, he acted as a mobile vakil for the peshwa, moving between Delhi and the Rajput courts. From the 1730s, as links developed between the courts of Jaisingh and the northern Maratha sardars, Mahadev served as the peshwa's vakil at Jaipur. As Jaisingh took an ever-larger part in Delhi politics, so Mahadev began to spend more time in Delhi, and his eldest son Bapuji became peshwa's vakil in Jaipur. 53 By 1744, when he was killed in a brawl between Maratha and Jaipuri troops, Mahadevbhat was both peshwa's vakil at the imperial court, and also vakil for the new Raja of Jaipur, Ishwarsingh. 54 At this point, Bapuji focussed on affairs in Delhi, while his brothers Damodar, Purushottam and Devarao worked as mobile vakils at other north Indian courts. A deed of grant issued in 1757 by the peshwa Nana Saheb, after a temporary confiscation of the family's offices and estates, reveals an extraordinary array of vakil offices, at the courts of Delhi and Jaipur, with Shuja ud Daula in Awadh, with the Rohilla and Pathan courts of Ahmad Khan, Sadulla Khan, Dunde Khan, Hafiz Rehmat Khan and Najib Khan. In an early attempt to channel intelligence through his own vakils, Nana Saheb's grant stipulated that the northern sardars Holkar and Shinde and the Maratha general Visaji Krishna should utilise the services of the Hingnes, and the Hingnes alone, for vakil work in North India. 55 With the Maratha expedition of 1754 to Delhi under the command of Nana Saheb's less competent younger brother Raghunathrao, we get some sense of the struggle that the next generation of the Hingnes undertook to extract the huge cash subsidies for the Maratha armies promised by Alamgir II's wazir Imad-ul-Mulk. In October 1754, Mahadev's son Bapuji described their role: We are constantly pressing the wazir for what he still owes us. But the Emperor has no cash, there is none in the wazir's house, and the soldiers are perishing of hunger. Their blades and daggers lie around in heaps. We are pressing them for money in the midst of all this. There are no ornaments and no cash. The wazir has told us that there are estates near Delhi which have been set aside to feed the himself and the emperor, and we should take an assignment on those revenues for the thirteen and a half lakhs we are owed. 56 After failing to extract cash from the wealthier citizens, Bapu Hingne and the Emperor's minister Nagar Mal together attempted to impose a levy on the traders 53 Khare and Athavale,Hingne Daftar,Vol. 1, Purandare, Purandare Daftar, Vol. 1, no. 154. 55 Kibe, 'Some original Marathi Documents', pp. 1-5. 56 Sardesai, Selections from the Peshwa Daftar (hereafter SPD), Vol. 27, no. 90. and market people, resulting in closure of the bazaars and major riots. By November, Nagar Mal had taken refuge at the Hingne mansion in Jaisinghpura. 57 So extensive was the Hingnes' portfolio of vakil offices that they delegated their work in Jaipur to another Maratha Brahman family, the Galgalekars, who also combined diplomacy and reporting on tribute collection. 58 The Kuntes also combined diplomacy with reporting on tribute collection. Serving at the Nizam's court from the late 1720s, they were at the sharp end of Maratha negotiations over chauth in the Nizam's territories. Ganesh Ballal Kunte served as the peshwa's vakil to the Nizam from the late 1720s, reporting from the Nizam's camp the all-too-familiar state of local-level war between forces loyal to the Nizam and Maratha revenue collectors. 59 His younger brother Sadashiv was at the same time keeping him in touch with the latest intelligence from Aurangabad city. 60 By 1735, Sadashiv was serving as peshwa's vakil in Udaipur, describing the grand diplomatic progress of the peshwa's mother across North India, as well as Bajirao's northern tour the following year to secure written confirmation of his chauth claims. 61 By this time, evidently, Sadashiv had managed to secure a personal jagir, since he slipped in a plea for its exemption amid the rest of his reports. 62 The Hingnes were part of the negotiations, and Sadashiv sent his own advice to Bapu Hingne as to the best means of conducting them. 63 While Sadashiv shuttled between Udaipur and Jaipur, Ganesh Ballal and his son Raghunath Ganesh remained at the Nizam's court, the latter beginning service as vakil in December 1745. 64 A further area of advantage in the jointness of vakil family enterprise lay in the upwardly mobile path from kamavisdar to political vakil. In some cases, such as that of the Gulgules, the Shindes' vakil in Kota, whole families started out as kamavisdars and then moved into roles as political vakils. 65 In other cases, families already successful in vakil service often put out their sons into work as kamavisdars, in the hope that they would progress to vakil service.

Between Personal Service and Service to the State
Running through all of the challenges that political vakils encountered was the tension at the heart of the role, between service to an individual employer, and larger obligations to the state. This tension surfaced in a familiar problem for every vakil, that of managing a personal relationship with their employer, often at a long distance, alongside the larger strategic calculations they needed to make as agents of the state in the field. Mahdevbhat Hingne's long relationship with Bajirao I exemplifies this difficulty. As noted above, 1736 was the pivotal year for negotiation of Maratha claims to chauth in North India. Mahadevbhat, along with Rajmal the Diwan of Jaisingh, Rajput ruler of Jaipur, brought down a draft treaty from the emperor to Udaipur, and the Mir Bakshi Nawab Khan Dauran had promised many additional gifts and grants. In January 1736 the terms with Jaisingh were arranged through Mahadevbhat, accompanied by much ceremony. 71 At this key juncture, it was learned that a body of Bajirao's troops had run amok and captured the fort of Ujjain. On 24 June 1736, Mahadevbhat wrote to Bajirao, describing Khan Dauran's bitter reproaches and withdrawal of the treaty offer. Here we see the language of trust and confidence in negotiations that Vendell has described, now tested and broken when a vakil's employer failed to follow his advice.
'It is true that people of the Deccan can be trusted. I have not betrayed your confidence. But if you are going to go behaving like this, what is the point of your presence here? What am I to tell the emperor? You have proved the enemies who speak against you to be right'. Thus speaking, the Nawab withdrew all of the promised inam, elephants, horses, jewellery and grants, and recalled the hundis of two lakhs he had only just instructed Rajmal to issue, and recalled his army and people.
Because of Bajirao's failure to control his troops, Mahadev continued, 'the whole mould of our negotiations has been broken'. There was only one course of action now, he impressed on Bajirao: You should withdraw your troops from Ujjain and canton them a few miles away. Get Savai Ji to send some influential man and put him in the fort. Write a letter to the Nawab in your own name. Write in that letter that we did not know about this work, it was done by some small man, and we have caught and punished him.
Mahadev further urged an appeal to Mughal vanity. 'Write the letter very humbly, and make yourself inferior to them. This is a Mughal court, and they will be very happy if you write about their greatness'. He closed the letter, barely able to contain his exasperation.
Well, go ahead and write whatever letter you want, taking my advice or not. I'm here in the emperor's city, bearing enormous expenses, and with troubles on every side. Write to me quickly. Why on earth did you allow this affair to happen? God only knows. 72 The corporate interests of the family sometimes ran up against the wider needs of state service, particularly when revenue rights and flows of cash were concerned. Some of the Hingne brothers were notorious for their unprincipled financial dealing, as well as for their long-running personal feud with Antaji Mankeshwar, commander of Maratha forces in Delhi during the 1750s. Damodar Hingne accused Antaji of making false returns of the number of troops on his muster rolls. Antaji in turn complained bitterly that Damodar maligned him to the emperor Alamgir II and to Nana Saheb back in Pune, and alleged Damodar himself took bribes in the course of his work. 73 In the crisis year of June 1754, as Jats, Gujars, Baluches and Rohillas fought for control of Delhi, and Maratha hopes of wealth from the imperial court went up in flames, Nana Saheb admonished Damodar for his failed management of Maratha interests in the city, compared to his father Mahadevbhat. 'Your father was intelligent, honest and worthy of his post, and because of that, it was agreed to make you vakil. For a while, you served obediently, and so we looked after you'. But these days, every sardar complained about his behaviour. His allegations to the emperor about Antaji had caused great harm to the peshwa's strategy, and he wavered constantly in his political advice, now counselling alliance with one party 72 Khare and Athavale, Hingne Daftar, Vol. 1, no. 5. 73 Sardesai,SPD,Vol. 27,nos. 80 and 95. at court, now another. 74 These complaints culminated in the temporary confiscation of the Hingnes' property in 1757. 75 The sometimes uncertain nature of delegated authority made itself felt in uses of the formal insignia of a vakil's office, the sanad or deed of grant stamped with the seal of his employer confirming him in his role. Ganesh Ballal's son Raghunath Ganesh, peshwa's vakil to the Nizam in the 1740s, found himself in difficulty in late 1746, as the drought and crop failures of that year began to make themselves felt. Nana Saheb peshwa had been highly displeased to learn that Raghunath had taken liberties with the seals of office. Raghunath had made his own personal seal out of wax, and given it out as a dastak or official pass of exemption from transit dues. In November 1746, he wrote in great contrition to explain. There had been a great shortage of grain in Aurangabad. The Nizam's son, Nasir Jung, ordered the city's vakils to issue passes to the grain carriers, as they were allowed to do by local custom, to ensure that supplies were not held up. This placed Raghunath in a quandary because he did not possess a seal with his own name on it. So, he explained, he had made a seal, but had written only his own name, and the name of the god Sadashiv on it, and not the name of Nana Saheb himself. He concluded with further apology and explanation.
Your servant is certainly guilty. If you give the order, then a seal can be made with your own name on it, and used to issue official passes. If you do not want to do this, then I will plainly tell the Nawab that I cannot issue official passes without an order from my lord. 76 The honour of a vakil in the field was another area in which the sometimes uncertain nature of delegated authority presented challenges. The honour of a vakil was a local and personal matter, yet inextricably bound up with the prestige of his employer, and almost invariably placed under strain by shortages of cash. A protégé of Baburao Barve, Shamji Govind Takle, served as Bajirao I's vakil with the Nizam between the late 1730s and early 1760s. 77 In June 1738, he wrote to Bajirao I's brother Chimaji Appa, reminding him of matters still needing his attention, including Shamji's own financial embarrassments. The issue of Sayyed Jamal Khan's military estate held on revenue farm had yet to be resolved. He himself was owed money by powerful ministers at court, and he pleaded Chimaji to send letters to them by fast runner instructing them to pay up. Shamji himself and his associate Jotipant were responsible for paying the Arab and Siddi mercenary soldiers at the court maintained there by the peshwa. Remittances had fallen behind and the soldiers were becoming belligerent. Nasir Jung and the city's Kotwal had been ordered to assist, but refused to act. Shamji begged Chimaji to act speedily. 'If you delay my request, then there will be nothing left of my honour. I have tried every trick and device I know, but I am dealing here with absolutely idiotic people'. 78 The vakil Raghunath Ganesh was even more explicit about the shame of his financial embarrassments. He wrote to Nana Saheb in April 1747 from his post at the camp of Nasir Jang. The rains were coming, his team was discontented, and his accoutrements were worn out.
Your servant is completely without the proper equipment. The horses of the troopers are not in a good state. The saddles are all worn out. I have a macebearer, a pair of runners and two lads to help me, but they are all foreigners. On top of that they are all anxious because the rains are coming and keep asking your servant for permission to leave.
Nana Saheb had sent him a draft for 300 rupees, but that was now three months ago. On top of his own expenses, there are the sundry expenses for Nasir Jung's servants, and on top of that the costs of the runners and the mace-bearer. Prices are very high in Nasir Jung's army. It would not be fitting for me to ask Nasir Jung for money. I feel very awkward doing that. 79 On other occasions, vakils abandoned the focus on honour, and simply looked to their personal needs. Many hoped to secure a jagir estate, as a mark of prestige and a reliable path to security in retirement. Towards the end of his career, Shamji Takle was offered a jagir by Nizam Ali Khan, during negotiations with Haidar Ali about a possible alliance against the peshwa Madhavrao I. 80 We saw above that the Udaipur vakil Sadashiv Ballal had secured one for himself. But his nephew Raghunath Ganesh seems to have been less successful. He wrote rather miserably to Nana Saheb in 1748, reminding him of his past promises of a Mamledar post on the Konkan coast. Now, your servant's request is that you look kindly on me and issue orders for the Mamledar post to be given to me. Then my family will be established there. We still haven't built a house, and are just eking out our days in a makeshift residence. Your servant is poor, and can't run around anymore, because of age and infirmity. That is why I am requesting you. 81 Raghunath's pleas evidently fell on deaf ears. Through 1753 and 1754, he was despatched as vakil to the camp of the French commander Bussy, as part of Nana Saheb's attempts to secure a French alliance. 82 From these earlier eighteenth-century perspectives, it is possible to see the deep social and financial hinterland of a political vakil's life, which is sometimes obscured by the drama and colour of courtly diplomatic exchange. The expansion of the older and more capacious agential role of vakil to encompass both diplomacy and financial management made the vakils of the eighteenth century much more flexible instruments of the state than their predecessors in the īlchī and hejib of earlier centuries. With their extended family networks, their ability to diversify risk across several professions, and their skill in facilitating the business of several employers at once, their presence was as vital to eighteenth-century states as that of merchants, bankers and revenue contractors. Uniquely, though, political vakils were often mobile across subcontinent-wide networks. Their roles as diplomats gave them skills of persuasion that were useful in building their own personal networks and shaping local institutions and relationships around their own interests. At the same time, vakils as well as their employers struggled with the tension between the personal and the state-focussed aspects of vakil service, and with the place of the extended family in state service, its independent social and financial imperatives making it sometimes an asset, and sometimes a liability.

Mid-century Transitions
What were the implications for Marathi-speaking vakil communities as the European trading companies expanded their political operations from the middle of the century? As Fisher and others have described, Indian states tried various means to counter European attempts to control the flow of political information, including developing their own independent networks of vakils. For the Maratha state, this was the trademark of the prominent minister Nana Phadnis (1742-1800), whose grandfather Balaji Bhanu came up from the Konkan with the first peshwa Balaji Visvanath. 83 Nana Phadnis came to the fore after the Maratha defeat at Panipat in 1761, and rapidly rose to power as leader of the Barbhais council which presided over Maratha affairs after the murder of the young peshwa Narayanrao in 1773. During the 1770s and 1780s, he developed an extraordinary network of vakils across the subcontinent, moved around as political circumstances and challenges, particularly from the English, demanded. 84 The other major challenge to Nana's position in the field of intelligence lay in the rise to north Indian dominance of the great sardar and over-mighty subject of the peshwas, Mahadji Shinde (1730-94). First coming to prominence after the Maratha defeat at Panipat in 1761, Mahadji was principally responsible for the defeat of the English in the First Anglo Maratha War of 1779-83. His military dominance in North India enabled him in 1784 to recover the Maratha position in Delhi, when the emperor Shah Alam II conferred on him the title of Vakil-i-Mutlaq, the direct deputy of the emperor, to the great annoyance of Nana Phadnis himself. 85 As we saw above, the peshwa Nana Saheb had in 1757 been able to decree that Shinde and Holkar should use the Hingnes for all vakil work in North India. It was a measure of the challenge to the Pune court's control over these networks of political vakils that from 1776 Mahadji began to insist that all Maratha communication with the Company should flow through the Resident at his own court in Gwalior. 86 Maratha vakil families flourished in this new milieu. Some represented the peshwa's interests in Bombay, Madras and Calcutta, principal among them the Ranchod family in Bombay, the long-serving North Indian Kayastha Lala Sevakram and his extended family in Calcutta, and Janardan Shivram and his brother Nago in 1780s Madras. 87 Janardan's experience offers us a good insight into a vakil's own sense of his career, in a landscape where Indian polities small and large flourished alongside the East India Company as sources of employment, as well as of informal advantage. While serving as peshwa's vakil to the Company in Madras, Janardan had been given three villages in inam by the Nayaka chief Venkapatah at Kalahasti, as thanks for Janardan's having persuaded Hyder Ali not to plunder his lands. Once secure from Hyder, Venkapatah took his villages back. Clearly fluent in English from his Madras experience, Janardan wrote to the Governor of Madras in February 1785, emphasising the personal factotum aspects of a vakil's role: 'I am a Vakeel and whose Duty is to render Service to those that are desirous and receive from them favors in return'. Possession of a jagir or inam lands such as Venkatapah had given him, he explained, was so important to vakils like himself that 'they will wait for twelve years to accomplish their Aims'. Moreover, he explained to the Governor, he was not some fly-by-night vakil, but a settled and substantial servant of the peshwa in Madras, whose personal plans depended on recovering the villages: I am fixt to the Office of Vakeel here on behalf of the Mahrattas, and not same as others that remain here to-day liable to be removed the next day. I have therefore rejected the Offer of Money and accepted the village from Colastry Zemindar thinking myself that I shall always have country provisions while I remain at Madras. 88 Other vakils posted to the Company pursued their own family interests alongside those of the peshwa in other ways. Between 1786 and 1797, Bahiropant Mehendale 85 Pandey,Mahadji Shinde,Indirect Rule in India, Disalkar,'Maratha Vakils,. 88 Disalkar,'An English Letter', of the prominent Mehendale political family represented the peshwa with the British Resident Charles Malet in Pune. 89 His younger brother, Bachhaji Raghunath, was at the same time peshwa's vakil to Governor of Madras William Medows, and was present with Appa Balwant Mehendale at the surrender of Tipu Sultan's sons to Cornwallis in February 1792. 90 Like many others, Bahiropant had early in his career worked as a kamavisdar in Surat, when he may have gained the language skills that enabled him to work with the British Resident. 91 As noted above, the Portuguese in Goa also retained Bahiropant's services as procurador at the Pune court, an appointment noted quite matter-of-factly by local Maratha state officials. 92 96 The perspectives emphasised here, of the plural roles and local embeddedness of political vakils, helps us to appreciate their individual resilience in the face of European efforts at control. Their ability to deploy family alongside political networks and to transfer skills learned in one role to another, often brought them personal advantage, and could sometimes be useful to their employers. But the plural involvements of vakils' lives could also be a disadvantage to their employers, blurring the lines of political loyalty in an age where political advantage lay in tightening them.

Messages from Nana: A Vakil Goes South
Yet the drive to control intelligence was only part of the story for Maratha vakils and their employers as the Company's power expanded. How far did its presence, its consciousness of its dignity and honour as a political player, its legalistic 89  approach to political relationships, and its adroit exploitation of regional rivalries and political factionalism, shape the understandings of statecraft and political ethics that Vendell has described? Let us follow the journey south of Nana Phadnis's vakil Krishnarao Joshi, despatched to Seringapatam in October 1779. His mission was to negotiate terms for a sudden combined attack on the Company by the Nizam Sikander Jah, Mudhoji Bhosle of Nagpur, Hyder Ali and the Pune court. 97 Travelling with Krishnarao was Hyder's vakil to Pune, Narsinghrao Verulkar, whose brother Anandrao Narsi was then acting as Mahadji Shinde's vakil with Shah Alam I in Delhi. 98 Also travelling were two clerks, Tryambakrao Appaji for Mahadji Shinde, and Govindrao on behalf of Raste, the two sardars who were to stand guarantee for the agreement with Hyder Ali. 99 This journey certainly reveals the closer direction that Nana Phadnis exerted over his vakils' work. A veritable stream of letters followed Krishnarao and his party south. His formal preparations began on 22 October 1779, with the issue of a travel pass to ease his journey.
To kamavisdars, chowkidars and certain people, to mokadams, to boatmen, to thokrekaris. The Sarkar has sent Krishnarao Narayan Joshi to Hyder Khan. With him have been sent palanqeens, horses, camels, guards, troopers and bearers. Let there not be any obstruction to his coming and going. At night a watch should be posted for them, and help given them to cross rivers and nullahs. 100 The vakils were instructed to rendezvous at the Mahuli river confluence to the south, then travel in long marches for maximum speed. Nana repeatedly emphasised that Krishnarao and the two clerks should keep to a common line of argument. 'On everything, you should speak with one voice. Once you have arrived, do not in your speech or your behaviour let any differences between you show outwardly'. 101 A further stream of anxious letters followed Krishnarao through December into January, demanding news and urging him to make haste. On 20 January, Krishnarao wrote to say that he had arrived in Seringapatam, and been given an excellent reception. Two days later, Hyder received them in great state, and negotiations began.
A key obstacle lay in competing claims to the Adoni district in Andhra, which Hyder Ali hoped to secure. Eventually, a draft agreement was prepared. But when it came back from Seringapatam to Pune, Hyder was found to have inserted additional clauses in his favour. On 23 February, the commander Hari Ballal sent a query to Krishnarao: 97  When the document arrived from Patan, it was as per our agreement. But there was one further item about Adoni. How was it that this new draft was produced? There is a great difference between the first agreement and the present draft. It is very damaging to us. But it is very well known now that we have established a friendship with the Nawab. It would not now be fitting for us to take offence. 102 On 26 February, Nana sent Krishnarao a further missive on the proper conduct of state policy, where a stronger assertion of the ethical basis of political negotiation for great states is apparent. 'The Sarkar's custom is this, that once discussions have happened and a promise been made, it is not right to turn round and make changes to it. In great states, solemn promises and assurances given are the very treasures of the kingdom'. 103 There was an additional reason that Hari Ballal adduced, as to why the Pune court should not attempt to undo the agreement. It was that friendship with Mysore, and that too in an alliance against the English, was a new thing for the Marathas. It was no small thing, he said, that the Nawab and the Sarkar should have become friends in order to defeat the English. 'Such a friendship and alliance did not happen during the rule of the late peshwa. This union has been brought about in our own time'. 104 The spring passed and the monsoon approached, but despite all Nana's urgings via Krishnarao, Hyder remained in his camp. On 15 May, Nana pressed Krishnarao still more vehemently. This is a time for speed. Perhaps the Nawab Bahadur will say, the season is over, the rains are coming, so what does it matter? We will go into battle in the rains. Tell him that fighting has already started in Gujarat. Men are coming from Seringapatam by sea to help them. If the Nawab acts to check the English now, help can reach them.
Hyder Ali's troops did eventually move out to engage the English in the Karnatak. Hastings was able to draw the Nizam and Mudhoji Bhosle away from the alliance, and the treaty of Salbye in 1782 recognised the compromises to be made on both sides. 105 The episode suggests reveals Nana's tight control over his vakils in the field. It also suggests that the process of political negotiation itself, Vendell's jab-sal, acted as a focus around which vakils and ministers alike developed more explicit understandings of the marks of a dignified state and what might count as its major achievements amidst the complex political relationships of the subcontinent. Krishnarao was shortly after handsomely rewarded for his services, by the grant of three villages in Sholapur district, worth Rs 10,000 a year. 106

Govindrao Kale: A Vakil at the End of Empire
The themes outlined here emerge with particular clarity in the history of Pune's most powerful vakil family in the Deccan. Between the 1760s and the first decade of the nineteenth century, three generations of Kales, Krishnarao, Govindrao and Govindrao's sons, were peshwa's vakils at the Nizam's court. With its French military force maintained there, Hyderabad remained a key diplomatic arena for Pune. The year 1776 saw a temporary agreement about chauth when Sikander Jah accepted Maratha contentions that the assessment was now a serious underestimate, given the growth in Hyderabad's revenues over the previous half-century. 107 By 1781, the agreement had broken down again. During these years of relative stability in Maratha-Nizam relations, many Marathi-speaking speaking military men and scribal communities had entered the Nizam's service, adding to the Marathi revenue collectors already working there and further complicating relations between the two states. 108 It was not surprising, then, that Krishnarao Kale, a leading political figure in Pune, should have been made Madhavrao peshwa's vakil to the Nizam's court in the 1760s. He had been present with Raghunathrao's army that carried the Maratha flag across the Chenab river into Punjab in 1758, defeating the Durrani forces to take territories around Attock. 109 He remained behind as the peshwa's vakil at the Mughal court, and by 1764 was acting as the peshwa's vakil with the Nizam. 110 Thereafter he combined periodic spells in Hyderabad with a career at the forefront of Maratha politics. He was a member of the Barbhais council, and a key liaison between the council and Raghunathrao, thwarted aspirant to the peshwaship. 111 In 1776, he played a major part in the negotiations about a Maratha-French alliance with the French envoy St Lubin. 112 In August 1781, he was once again in Hyderabad, to renew Maratha chauth claims, armed with copies from the Satara archives of the treaties made under Madhavrao peshwa. 113 His son Govindrao learned his craft as a political vakil working alongside Krishnarao. He was with his father in Holkar's camp in Gohud in May 1766, helped to welcome the French envoy St Lubin in July 1776, deputed for his father while Krishnarao was in Hyderabad, and succeeded him as vakil after Krishnarao's death in September 1786. 114 Govindrao possessed a remarkable sense of the long historical trajectory of the Maratha state. Part of this derived from his work alongside his father, with his own substantial north Indian service, and part from the long association with Hyderabad, so central in Maratha political calculations throughout the century. Govindrao also possessed substantial south Indian diplomatic experience. He represented Maratha interests in the negotiations with Tipu Sultan after the fall of Serinapatam in early 1792, and was present, along with Appa Balwant and Bachhaji Mehendale, at Tipu's surrender and the hand-over of his sons as hostages in February 1792. 115 Govindrao expressed this sense of Maratha history in a lyrical letter to Nana Phadnis shortly afterwards. Written in June 1792, the letter also reflects his skill, gained over long decades in his work as a vakil, in crafting Marathi prose. He had, he wrote, put pen to paper immediately after receiving news of the accord established between Nana and Mahadji Shinde, now just returned with his armies from North India to Pune. The news of the accord had sent a thrill through his whole body. Between the two of them, they had made the land from the Attock to the southern sea 'no longer Turkestan, but now Hindustan'. Begun under Shivaji, this expansive state had been the joint accomplishment of Brahman wisdom in leadership and the martial valour of the Maratha sardars, now returning home in triumph. 'Now, because of the Shrimant's mighty virtues, and the Patilbuva's wisdom and prowess with the sword, all have come home'. It was very much a Hindu patriot's view of the state, which had achieved 'not just territories and rule, but protection of the Vedas and Shastras, the setting up of Dharma, the protection of cows and Brahmans'. The gathering to celebrate this wonderful achievement should rightly, he said, have been taking place in the public square of Lahore. 116 Given the apprehensions actually surrounding Mahadji Shinde's return to Pune, and the sensitivities of Brahman and Maratha roles in the state, this was actually a very ingenious diplomat's letter. Like the marks of a dignified polity which emerged out of Nana Phadnis's correspondence with the vakil Krishnarao Joshi, Govindrao's skilful and selective evocation shows how important a milieu the conversations of vakils and ministers were for shaping a sense of the Maratha state as arising essentially out of collaborative political relationships developed over its long history.
This sense of the importance of cooperative relations also emerged in Govindrao's negotiations with Hyderabad. The negotiations between 1791 and 1795 swung between acrimony, insult and protestations of friendship, concluding in a summary defeat of the Nizam's troops at Kharda in March 1795. Throughout, Govindrao's priority seems to have been to deflect Maratha demands, prolong the negotiations and to plead for relations of amity that would benefit both states.
To understand this, we must appreciate Govindrao's remarkable position of accumulated influence in Hyderabad. As Pune's vakil, he had a direct line to Nana Phadnis in Pune, reinforced through his assistant vakil in Pune, Govindrao Pingle. Kale also had two local assistants in Hyderabad, Ethal Pundit and Kunal Pandit, who remained behind with the Nizam when Govindrao himself visited Pune. 117 Writing in January 1794, the Pune Resident Charles Malet, evidently much less tolerant of Govindrao's multiple political involvements, lost no time in educating the incoming Hyderabad Resident James Kirkpatrick about the malign influence of the 'two Govindraos'. The Nizam's affairs were in their present sorry state because they were 'carryd on through Govindrow Kishun, alias Bappo, alias Kalla, this Court's delegate at Hyderabad, whose looseness of conduct has frequently been notorious. His (Govinrow Kishen's) delegate, Govinrow Pingla, alias Mamma, here acts as far as it meets the views of this Court in concert with Ragotim Rou, the Nazim's own agent here. 118 Kirkpatrick quickly came to the view that through his 'court', Kale's aim ultimately was to 'compel His Highness to model his Ministry conformably to his wishes'. 119 Govindrao's private correspondence during the early 1790s also reveals his vakil's involvement in a wide range of agrarian, commercial and judicial matters. As Pune's representative, his authority in Hyderabad's agrarian affairs derived from Maratha revenue claims and from the practical presence of Maratha kamavisdars and troops on the ground. Kale sent a stream of letters to Pingle, from pleas for relief on revenue demands on account of crop failure, to disputes about property between kamavisdars and local people, to the issue of passes exempting trade goods from charges, to the troubles of debtors harassed for repayment. In each case, he indicated what letters of instruction Nana was to send in favour of one party or the other. 120 Govindrao also maintained a voluminous direct correspondence with provincial and district level officials and holders of jagirdar estates, the great majority Marathas. He sent instructions on many different matters, from shortfalls in revenue payments, the levy of payments on villages properly exempt, illicit seizure of lands, recompense after theft and the recovery of debt. In most cases, his role seems to have been to chase up orders already given to these Marathi officials, expressing surprise that the officials concerned had not yet acted on them, disappointment at local disorders, and, sometimes, the satisfaction that it would give the Nizam's own high officers to know that a particular dispute had been resolved. 121 This was very much an informal authority, arising as much from the fact that most of those he was writing to were Marathi-speakers aware of his influence in Pune, as from his position as a prominent vakil and adviser to the Nizam.
Here, interestingly, Govindrao adjusted his Marathi when addressing Persephone correspondents of rank. He wrote in November 1792 to one Abdul Samad Khan, 'honourable friend the Nawab Sahib of exalted fortunes', who was visiting the Hyderabad court, to apologise for the delay in bringing Abdul Samad's business to the attention of the Nizam. Throughout the letter, Govindrao adopted an extremely ornate Persianised style, every other word in the letter a Persian term rather than a Marathi one. Rather than simply choosing to write to this Persian speaker in Persian, however, the basic language and grammar of the letter were Marathi, and the script Devanagari. We can only speculate as to why this might have been. Perhaps Govindrao did not feel sufficiently confident in his own Persian epistolary powers to be able to address a high-ranking visitor whose command of literary Persian was likely to have been superior to his own. Perhaps also, in this particular context, Marathi, rather than Persian, may have seemed to be the language of power, a reminder of the close political and familial ties that ran direct from Govindrao's establishment at the heart of the Hyderabad court, to the court of the Maratha peshwas in Pune. 122 Did Govindrao's shielding of the Nizam's state arise from a desire to protect this powerful position? This may be part of the explanation, but not all of it. He was quite capable of lodging vehement protests when the honour of his own master was impugned. The Nizam's minister Mushir ul Mulk harboured a notorious hostility towards Nana Phadnis. In the public space of the Nizam's camp, he was heard to threaten that Hyderabad would give the Marathas no peace 'until they had despatched the Peishwa to Benaras, with a cloth about his loins, and a pot of water in his hand, to mutter incantations on the banks of the Ganges'. 123 Undeterred, Nana sent over a detailed statement of revenues owed dating back to 1774, showing debts of some 26 m rupees. 124 'Warm discussions' then took place between Govindrao and Mushir ul Mulk, when at last the former was told, in public durbar, that Nana Furnuwees must himself attend at the court of Hyderabad, in order to afford an explanation of the different items of their intricate claims. The envoy replied, 'Nana Furnuwees is much engaged, how can he come?'. 'How can he come?' re-echoed Musheer Ool Moolk, 'I will soon show how he can be brought to the presence'. 125 Mushir ul Mulk then arranged a pantomime of just this event, with men dressed up as Nana Phadnis and other Maratha dignitaries parodying them before the whole court. Deeply insulted, the two Govindraos rose to leave. As a contemporary of the last peshwa later told the story, The rest of the company asked why they were leaving, and the pair replied that it was not seemly to sit through this tamasha of their Master. The Nawab called Govindrao over and said, there is no need to go, it is just a tamasha. What, is there a law against it? Govdindrao answered him, 'This is a royal court, and a mockery of our Lord has been staged here. It gives us no pleasure to sit further'.
Issuing their own threats to take the Nizam's minister and grandees to Pune and parade them bodily from door to door, Kale and Pingle made obeisance to the Nizam and returned to Pune. 126 After the prolonged stasis in the negotiations engineered by Govindrao, these insults provided the catalyst for the Maratha attack at Kharda that followed in March 1795. 127 Even on the eve of battle, however, he demonstrated a remarkable desire to protect the Hyderabad state. Kirkpatrick reported that Govindrao had suggested that the two parties should pause, withdraw to the banks of the Godaveri where they could wait conveniently, and make another attempt at negotiation. 128 After the defeat, he urged conciliation. When the Nizam's son Ali Jah fled the court in June 1795, Govindrao urged Nana to send out troops to bring the rebel to heel. He should reassure the Nawab and set his mind at rest, that you only have your common wellbeing at heart. There is no bar to friends being as one heart in their affections. The Nawab's and the Sarkar's prosperity are one and the same. It is the same with their prestige. 129 He impressed the same message on a deeply suspicious Nizam and his ministers. 'Step by step and day by day, our two states should together sweep away the dark grudges between us, and our relations become open and clear, with a comprehensive agreement and treaty that we can keep for good'. 130 In a further letter to Nana Phadnis, he returned to the same theme.
The Nawab may understand this, or he may not. But whatever is good for both the Nawab's and the Swami's state, should be done. Even the Nawab will come to understand this. The Swami should give his attention to protecting the Nawab. His state should be preserved. 131 In this way, Govindrao made clear his sense that states were very much more than their rulers, and that the point of politics was ultimately to preserve states through collaborative relations between them. His position reflected an understanding of the inter-state relationships of the eighteenth century that complemented the culture of jab-sal as Vendell has described it. Part of this understanding grew out of the longstanding respect, shared across the legal cultures of the subcontinent, for established precedent in many fields of social practice, particularly those that related to rights of various kinds, and to the long-established structures of political authority from which many rights flowed. 132 This principle was nowhere better enunciated than in the instruction that the Raja Shahu issued to Bajirao I in his dash to Delhi in May 1739. Warning Bajirao against any notion that he might displace the Mughal emperor rather than support his power, the Raja emphasised that 'he considers it a higher merit to renovate an old dilapidated edifice than to build a new one'. 133 The same principle may have shaped Govindrao's own approach to the Hyderabad state. In this, he was again following a longer established Maratha policy. As many historians have observed, for all the long history of hostilities between the Maratha state and the Nizam, the Marathas never set out completely to destroy his state. In part, this may have reflected Maratha regard for Hyderabad's own inheritance of legitimacy from the Mughals, and in part a recognition that in the contested politics of the Deccan, today's enemy might well be tomorrow's ally. 134 Govindrao's own concern to shield Hyderabad from Pune's demands may also have arisen from his own consciousness of the deep interpenetration of territories, rights and populations shared between the two states.
Either way, this willingness to live with rivals represented a political ethic fast outliving its currency. Govindrao's personal situation underwent dramatic changes in the months after Kharda, as factional conflict and Hyderabad's failing finances seemed to open up further space for his talents. In July, he reported to Nana that 'Everybody here is quarrelling with everyone else. Everyone thinks that I should take over the management of the Nawab's affairs'. But, he said, it was a burden few could carry, given dissensions at the court, the huge sums owed to the peshwa, and the refusal of the Nawab to trust anyone. 135 But his fortunes shifted rapidly after the sudden death of the peshwa Madhavrao in October 1795. Govindrao moved back into Pune and, like the peshwa's vakil Bahiropant Mehendale, was caught up in the turmoil over the succession. Also, like Bahiropant, Govindrao now transferred his primary loyalty away from Nana, and to Bajirao II and his party. But the timing of these swings was against Bahiropant, who died in Raigad fort in August 1797, amidst the severities of his detention. 136 With Nana's own brief imprisonment in December 1797, Govindrao himself was elevated to de facto leadership of the Maratha state. 137 In August 1798, Malet's successor in Pune, William Palmer, reported to the newly arrived Governor General Lord Mornington that Govindrao was doing his best to persuade Bajirao against a subsidiary alliance with the Company, and had written to Tipu Sultan with offers of a compact. 138 Within months, however, Nana had returned to favour, and Govindrao was himself imprisoned in Satara fort in January 1799. By then, the Nizam's own state had been drawn into a subsidiary alliance with the Company. In July 1800, Palmer reported that Govindrao had been released through the intercession of his sons, now serving as vakils in Hyderabad. But, he warned, 'to whatever cause this man owes his liberation, his determined and constant opposition to British interests and influence may be expected'. 139 Palmer anticipated rightly. Govindrao was still present in Bajirao's circle in 1817, with other advisors of the old regime pressed back into service in the months of Bajirao's final rebellion and defeat. 140 The historian James Grant Duff reported that Govindrao was still alive when he left India in January 1823. 141

Conclusion
Although greatly amplified by the Maratha presence in Hyderabad and by the failures in authority within the Nizam's state, Govindrao's career nonetheless illustrates the great potential for individual and family advancement that lay in service as a political vakil. Vakils certainly had to negotiate between the competing demands of loyalty to family, obedience to individual employers, and the public performance of their role as representatives of the interests of the state. Those able to do this were uniquely well placed to use their skills in diplomacy, their expertise in the management of personal and political relationships, and their familiarity with the movement of cash and credit, to become exceptionally influential players in eighteenth-century politics. In their dialogues with ministers and others, they also articulated contemporary understandings of the ethical ends of statecraft, as well as its practical means. That they did so as vakils, rather than as īlchī or hejib, was very much a reflection of contemporary conditions, in which when the skills of intermediation across different arenas were at such a premium, and the extended family was such a critical locus of power and resources.
As Robert Travers has described, the Company's government took a very different view of the elastic roles of the political vakil. Cornwallis's principal experience of vakils had been in the law courts of Bengal. He felt that they were a poorly qualified group, who lacked the legal knowledge needed to represent their employers' interests properly, and whose loyalties were suspect, 'often bribed by the opposite party to betray the cause of their constituent'. His regulations for fundamental reform of Bengal's judicial system drafted in February 1793 envisaged their replacement by a professional cadre of trained and state-approved 'native pleaders' on the model of professional pleaders within the British courts. 142 Political vakils underwent a second transformation elsewhere in the lower reaches of the colonial bureaucracy. They found a further new role within the 'uncovenanted civil service' that supplied the Residents of the Princely States. The title of 'vakil' disappeared, to be replaced by 'munshi', signifying a subordinate Indian clerk and translator to the Resident. 143 In this role, as in the new class of the native pleader, uncovenanted civil servants were increasingly required to pass statecertifying examinations, and were rewarded with individual salaries and pensions, rather than with family estates granted in perpetuity. 144 What had been a composite line of work, encompassing many skills and arenas for action, and allowing scope for family as well as individual entrepreneurship, broke apart into distinct, and relatively inconspicuous roles at the lower levels of colonial administration.