Key Ideas

Scholars often invoke state agency to explain the historical spread and establishment of languages. The political deployment of extant patterns of language use is also routinely invoked to understand the genesis of nationalism worldwide. Nation-states have indeed acted to establish, homogenize, and extirpate, favored and disfavored tongues and scripts. Sociolinguists describe such valorization and depreciation of languages as a “language ideology.” Finally, the idea that states are made up of coalitions of cliquish institutions has long been important in political sociology.Footnote ¹

This article develops a heuristic framework (or “model”) of the interaction of all of these processes as they unfolded across West and South Asia through 2,500 years. The largest and oldest Persian Empire—the Achaemenid (ca. 550–330 BCE)—did not deploy its founders’ ethnic speech in governance, but instead refined an older state language, Aramaic, for that purpose. That choice entrenched an extant interest group as scribal masters of the chancery. These endured through the fall of that empire and rule of the Hellenic one that overlay the region for two centuries. These bureaucrats were only displaced by the Arab conquest after 660 CE, but Persian revived as “New Persian” as the Caliphate waned. I therefore consider the cultural politics behind the efflorescence of “New Persian” in the eastern lands of the former Persian empire. After 1000 CE, new conquest elites emerging from Inner Asia carried this language and its scribes into their growing domains in the Indian subcontinent. In the fourteenth century, it was established at major imperial centers across India. But the fifteenth-century political fragmentation of the subcontinent shifted patronage from imperial towards regional languages. The next conquest empire, the Mughals, were non-Persians who adopted Persian as a state language from the 1550s. They were succeeded by the British who patronized the same language for a century and richly rewarded English officials who studied it. My article introduces new evidence to examine the extent of Persian literacy in the East India Company era and thereby also estimate its previous extent. It then explains the replacement of Persian in the Victorian era as a consequence of the modernization of imperial government. These processes generated a new politics in the colonial vernacular world that took shape by 1900. Thus, this article will examine the political life of imperial languages across the Indo-Iranian world, with a special focus on the period ca. 1500–1900.

Linguistic Change and Language Ideology

Linguistics has long used the concept of language ideology for a particular field of study, one that analyzes strongly held beliefs about language. Such beliefs can lead to the suppression of disfavored languages and promotion of favored ones.Footnote ² Language ideology is sociologically important because, to quote Raymond Williams, “a definition of language is always, implicitly or explicitly, a definition of human being in the world.” Kathryn Woolard adds that representations that “construe the intersections of language and human beings in a social world is what we mean by ‘language ideology.’” To be sure, she continues, almost any human act of signification reorganizes social relations in some way. Thinking and signification have what Karl Mannheim calls “social and activist roots.”Footnote ³ Some ideology of language is therefore found in every human society. I must emphasize that labeling these beliefs “ideologies” says nothing about the validity of the ideas held. It may be argued (for example) that German is the oldest living Indo-European language: that may or may not be a valid hypothesis. The claim nonetheless reflects a language ideology that valorizes both the language and antiquity. Anglophilia, Francophilia, or Persophilia are all forms of language ideology. A language is often valued as being “national,” or more rarely because it is “transnational.” Both evaluations are ideological. The former is favored by nationalists and the latter appeals to the globe-trotting academics or tax-minimizing corporations of our time. Several written languages have survived because they were identified in religious ideology as sacred or as a special marker of priestly clerical establishments. Because in many societies these religious specialists were often the only literati, religious belief might ensure the survival of fossilized forms of an old language such as Avestan, Vedic, or Church Syriac.

Traces of language evaluation are also embedded in old etymologies: Greeks called incomprehensible babblers “barbarians”; Arabs labeled the conquered Persians ‘ajam (stutterer).Footnote ⁴ Early Sanskrit focused deeply on exact utterance and developed a standard grammar and phonetics to ensure its correctness.Footnote ⁵ Pierre Bourdieu describes the power to select and use the “real” high language as a source of social distinction and social capital to its possessors. They would therefore be invested in preserving its exclusive status, and thereby their own.Footnote ⁶ Languages do not always exclude: they also serve to communicate. An early trade or contact language can stabilize into a pidgin and then a creole, perhaps replacing the parent languages. But power elites invested in linguistic purity resist amalgamation or creolization. This is a tension present in all complex societies.

The Herderian or nationalist ideology of language has been globally hegemonic for two centuries. It enters into most nationalisms. It has also been entwined with modern imperial projects. In mid-twentieth-century Iran, the Persian language became part of the parvenu Pahlavi dynasty’s state-building project. That Iranian national project also consciously promoted the modern study of Persian well beyond its own borders. Persian nationalists and their foreign epigones assume that the elegance and beauty of the language means that the world will naturally be drawn to it.Footnote ⁷ But so do advocates of every other major language.

Sociolinguistic Possibilities and Limits of Imperial Language: The Achaemenid Case

Nation-states and linguistic nationalism appeared after thousands of years when Eurasia was ruled by great empires. The latter, too, harbored language ideologies and nurtured languages, scripts, and clerical elites. Empires, however, largely accepted the existing multiplicity of languages and exploited them and their speakers as components of their power. Empires reached far across diverse speech communities. They had to sustain patterns of orderly reporting across distinct linguistic zones. Modern nationalists see the imposition of their own ethnic language as the imperial speech as the only authentic choice. This sometimes happened, since imperial administrative establishments sought to use a single language wherever feasible. But several impulses would compete in this setting. Languages do not speak themselves. Still less do they write themselves. Literacy was not common. Even among the literate, knowledge of specific scripts, increasingly cursive and specialized, served to create little scribal domains and exclude outsiders. Empires’ enduring character required they create at least rudimentary archives. State offices were where the great mass of steady scribal jobs endowed with arbitrary powers and regular douceurs was to be found. Some empires were ideologically driven to develop an ethnic tongue into a written imperial language. Others improvised by enlisting an extant scribal class. We can see the latter in one early empire, that of the Achaemenids who ruled Persia and beyond, ca. 550–330 BCE.

The Persian Empire became a great power across West Asia under Emperor Darius I (521–486 BCE) and his successor, Xerxes (486–465 BCE). Old Persian records multiplied under these two rulers. Darius I used cuneiform script to carve an elaborate trilingual inscription on the towering cliffs at Behistan. It was in Elamite, Akkadian, and Old Persian. Darius and his successors were proud of their Persian kingly heritage: “I am Darius the Great King … an Achaeminian, a Persian, son of a Persian, an Aryan, having an Aryan lineage.” But despite having conquered Elam and Babylonia, Darius recognized the reach of their older administrative and prestige languages and included them in his inscription.Footnote ⁸

Kent found that inscriptions in the emperors’ ancestral Old Persian were mainly created before 465 BCE, though some in a corrupted form of the language continued to the eve of Macedonian conquest.Footnote ⁹ Perhaps this signals a reduced ideological commitment to Old Persian. At any rate, the emperors governed the empire in conformity with extant scribal practice, in a bureaucratic Middle Eastern language, Aramaic. Achaemenid imperial satraps ruled from southern Egypt to northern India, but they all communicated in a standardized form of Aramaic rather than any ancestral Iranian tongue.Footnote ¹⁰ The probably ethnic Persian satrap in the Egyptian city of Memphis wrote to the priests at Elephantine in Aramaic: this was translated word for word by an Egyptian familiar with Demotic Egyptian script. The priests read it and replied in Demotic. As Briant writes, “the satrapal administration needed recourse to local scribes.”Footnote ¹¹ Language ideology was not powerful enough to override the convenience of using local scribal skills and records. The first and largest “Persian” Empire was therefore not Persianate.

I also cite the case of Aramaic to show that the sociolinguistic possibilities of a spoken lingua franca differ from those of a written one. The latter gives language control through the choice of scripts. A standardized common language written in a difficult script known only to the initiated can protect literati from the wrath of arbitrarily violent rulers. Such control of the archive permitted survival even under alien regimes as long as the records that scribes managed were still needed. Those owning a rare skill would not want to “cheapen” it by spreading it too widely. I have shown elsewhere that scribes would seek to be in at the founding of governments so as to mold the archive that any settled administration would later need to consult.Footnote ¹²

The inertial power of scribal guilds increased the longer a particular archive was permitted to accumulate. We may see this inertia at work even after the overthrow of the Seleucid successors of Alexander of Macedon and the founding of Persian-speaking empires in the late centuries BCE. Aramaic scribes nurtured through two centuries of Achaemenid rule had survived as an intermediate social stratum across the former Hellenic empire. Thus the Qandahar (Afghanistan) inscriptions of the Maurya emperor Ashoka (ruled ca. 270–238 BCE) were in Greek and Aramaic, not local Indo-Iranian languages. The emperor presumed that literate readers could only read these scripts but would orally transmit his ethical message to others.Footnote ¹³

Given their tenacious clerical presence, it is unsurprising that Aramaic scribes captured the chanceries of revived post-Hellenic imperial regimes. But the Parthian and Sassanian emperors increasingly created records from commands in their own language, Middle Persian. Scribes, however, continued to write in the Aramaic script. Skaervo notes that words were now to be sounded or uttered as their Persian meanings. It meant for example, that if the scribe wrote the Aramaic word for bread, it was to be pronounced as the Persian “nān.” Such terms are now called “heterograms” or “aramaeograms.” Thus, scribes were beginning to use more and more Persian when reading their records aloud even while writing Aramaic.Footnote ¹⁴ This layer of complexity obviously created a serious barrier to entry and to the legibility of the imperial archive. The guild of bureaucrats evidently did not wish to lay their records open to all comers even if their spoken Middle Persian language was comprehensible to (at least) people around the court. Writing creates new possibilities of privilege for those who control the technique, regulate its use, and serve as gate-keepers to the records in the archive. This privileged position only disappeared with the Arab conquest of Iran: the records, too, decayed. Installing a new courtly culture displaced the older literati, former arbiters of taste and producers of distinction. A scholarly Zoroastrian mourned the Arab conquest and the accompanying confiscations: “The world passes from us.”Footnote ¹⁵ The cultivation of one language and script has often depended on uprooting rivals, sometimes violently, sometimes by simple starvation and neglect.

The famous scholar Al-Biruni was an ethnic Persian who wrote in Arabic, the well-established high language of the Caliphate. He sought to develop a chronology of dynasties in northeast Iran. But he could find little information: Middle Persian priests and scribes had resisted the Arabs. So, “Kutâiba ben Muslim had extinguished and ruined in every possible way those who knew how to write and to read the Khwârizmî [Middle Persian] writing, who knew the history of the country and studied their sciences. In consequence, these things are involved in so much obscurity, that it is impossible to obtain an accurate knowledge of the history of the country since the time of Islâm.…”Footnote ¹⁶ Records might be lost but Iranian ethnic pride was reasserting itself. An example is the claim that Zaradust (Zoroaster) “produced his marvelous book in all the languages.…” This assertion was part of the movement labeled shu’biya (ethnicism) by Arab scholars of the day, a movement that “pitted non-Arab cultural goods against those of the Arabs.”Footnote ¹⁷

Following the disintegration of the Caliphate into regional satrapies in the tenth century CE, “New Persian” emerged in the fledgling kingdoms at eastern edge of the former Achaemenid empire. It arose around the short-lived courts of rival warlords. The Persian linguistic revival began through panegyric and poetic traditions.Footnote ¹⁸ This was, I argue, not accidental: such recitations were “portable,” not linked to an archive but part of a wandering scholar-minstrel’s stock-in-trade. Furthermore, the skill of the performer was publicly displayed for potential patrons to hear and, of course, was pleasing to the warlord addressed. And if he proved ungrateful, a few tweaks could redirect a praise-poem to a more generous host. Persian identity spurred an indirect cultural resistance to the Arabs through the preservation of ethnic speech and poetry. But the circumstances of the day made lives too short and regimes too unstable for anything more than a simple popular koine to emerge and spread.

This language was to become New Persian. John Perry classifies the emergent language as “Dari,” a widespread Eastern Iranian language, distinct from “Pārsik, Zoroastrian Middle Persian as written in Pahlavi script.…” He noted the relative lack of any ideology associated with the spoken Dari (nascent New Persian) of East-central Iran in the seventh to ninth centuries CE. It was not associated with the priesthoods of vanquished religions in the way that Pahlavi was associated with Zoroastrianism, Syriac with Christianity, and Hebrew with Judaism. It was plebeian common speech, a koine. It therefore did not have its own script and could only be written in the Arabic one. It was thus an unthreatening and convenient lingua franca.Footnote ¹⁹

Even after the completion of the Arab conquest, urbane urban Iran was to be periodically shattered by tsunamis of pastoral peoples from Inner Asia, a process that intensified with the Mongols after 1220 and lasted until the 1400s. Literati led insecure lives and moved frequently in search of safety, education, and employment. The recited panegyric supplemented the manual of revenue accounts in their métier. The Ghurid conquest of North India ca. 1200 CE came providentially for them: it was just two decades before the first invasions of Chinggis Khan. Refugees flooded into the new sultanates in the Indian subcontinent. These conquest states needed scribes and soldiers and the refugees supplied both. Religious and ethnic aliens in a new country, such men were more trustworthy than the remnants of the corresponding classes associated with just-conquered kingdoms. The thirteenth-century sultans of Delhi therefore made the same effort to develop an alternative immigrant bureaucracy working in new languages as the near-contemporary Yuan (Mongol) empire did in China.

The role of Persian in Mongol-ruled China was, however, not comparable to its place in the Indian subcontinent. Nile Green notes that it was in India, but not in China, that Persian acquired “the dominant partner sense that Hodgson intended.”Footnote ²⁰ The Mongols who founded the Yuan dynasty in China (1271–1368) were a conquest regime, like the Delhi sultans (1206–1398) in India. Both distrusted the indigenous intelligentsia and its anti-foreign ideologies. Both uprooted the institutions of older learning and confiscated their resources. But they differed in key respects. The core of the early Mongol army was the nomad in arms. Armies came with their flocks and families and converted large areas from arable to pasture to sustain them. The Turko-Afghan dynasties on the other hand, were created by a core of acculturated military slaves surrounded by mercenaries and auxiliaries. Key personnel in sultanate India were more culturally and linguistically malleable, while the Mongols were already sorted into units of social action under their own leaders. The Delhi sultanate was therefore more dependent than the Yuan on the immigration of trustworthy co-religionists. These were typically refugees from Mongol-ravaged Persianate lands. In Mongol-ruled China, the later Yuan dynasty succumbed to the need for indigenous bureaucrats. Nativist rebellions nonetheless forced the Mongols to withdraw north of the Great Wall by 1368.Footnote ²¹ The eastern Mongol lands saw the rise of the last nomad empire, that of Timur. It operated a system of dual offices in Persian and Turkish, with the latter written in Uighur script.Footnote ²² The Delhi Sultanate was shattered by Timur’s invasion in 1397–1398, with its survivors dispersed into the regional kingdoms within South Asia. What followed was a waning of Persian scribe-craft with a corresponding shift toward vernacular registers. An emerging creole was the default speech of the short-lived courts and small bureaucracies of little kingdoms, just as New Persian had been in eastern Afghanistan.

Power Structures and the Fifteenth-Century Emergence of Hindavi Creoles

Why was a version of North Indian or Hindavi creole—let us call it “Hindustani”—the default speech? Great households of West Asian origin would contain vast numbers of slaves and servants of diverse, but largely Indic origins.Footnote ²³ Even the children of immigrant lords would be exposed to wet-nurses and domestics’ speech during their most linguistically malleable years. Unschooled dependents developed various creoles. The default linguistic state in such a milieu would then be a more or less Persianized Hindustani. It would have a greater or lesser admixture of Perso-Arabic lexifiers depending on gender and rank. This was an orally transmitted language used in the women’s quarters, courtyard, army encampment, and bazaar.Footnote ²⁴ The end result was, of course, the widespread family of dialects out of which the dominant Urdu of colonial North India ultimately emerged. It is not surprising, then, that the short-lived fifteenth-century kingdoms of subcontinental India would see a decay in Persian linguistic and scribal skills even among the nobility. The potentate Dawlat Khan, for example, could not understand Persian. But the Mughal king Babur (1526–1530) wanted to send him a stern message, so he called “someone who knew Hindustani” and had that man translate Babur’s utterances word by word. This has usually (and correctly) been argued to signal that Persian was not widely spoken among the North Indian aristocracy of the time. Dawlat Khan shared this trait with the great majority of people in the lands that Babur had conquered. Babur remarked that his men had to “deal with an unfamiliar people whose language we did not know and who did not know ours.”Footnote ²⁵ Persian was indeed in retreat in the fifteenth-century Indian subcontinent.

Consequently, Indic vernaculars began to be patronized at the regional courts of various states. These would be common tongues known to the polyglot assemblies present at each little court. Persianate scribal bureaucracies were thinning out: they had to be administratively supplemented by Hindavi-reading clerks.Footnote ²⁶ Passive Persophonia may have persisted for some, but Persographia could no longer be assumed even among state functionaries. That is why biscriptual and even bilingual documents were issued by the Lodi rulers of Delhi (ca. 1450–1526). Sometimes, the Persian text was followed by a transcription in Kaithi script, an Indic cursive used by record-keepers. “This Afghan practice of biscriptual documents continued under Sher Shah (r. 1540–1545) and is discernible in some documents of the early part of the reign of Akbar.”Footnote ²⁷ It is naïve to consider the practice a gesture of comity and inclusiveness: it was functional. Such duplication means that one party to the exchange could not read Persian and had only an aural knowledge of its bureaucratic forms. The latter would be a common accomplishment among aspirants to the lower bureaucracy. Courts and offices were always crowded with petitioners and hangers-on listening to letters and messages being read to the persons in authority and often translated viva voce, too.Footnote ²⁸

The Second Coming of Persian: The Mughal Era

The Mughal Empire under Akbar (r. 1556–1605) is best viewed as another iteration of the conquest state, rapidly assembled out of pre-existing fragments under the direction of a foreign elite. Akbar’s grandfather Babur had allotted speculatively valued domains for his followers to collect what tribute they could exact in order to pay the troops who did the exacting.Footnote ²⁹ The second emperor, Humayun (enthroned 1530, d. 1556), was undermined by his brothers and betrayed by warlords who had previously submitted to his father. He finally fled to Iran where Safavid zealots required his conversion to their sect as the price of aid.Footnote ³⁰ Humayun therefore returned to Delhi in 1555 with a new entourage of Persians and a core of older Turkic followers. The insubordinate Turkic retainers soon irked Humayun’s son Akbar, who decisively repudiated them. His centralized empire depended on ending the autonomy of the nobles and thus breaking with Chaghatai Turkish tradition. Pliable “new men,” both Persians and Persian-knowing Indians, staffed a powerful central establishment he founded ca. 1560–1580.Footnote ³¹ The Mughal imperial household was by this time well acquainted with Persian, though Chaghatai Turkish remained its domestic language. Indian languages were little known.Footnote ³² That ignorance would rule them out as languages of imperial governance since their use would undermine the supervisory power of the imperial household. Ancestral Chaghatai, on the other hand, was little known in India and also associated with the subversion of royal power by ambitious relatives. Persian was accessible to the emperor, and pliable personnel could be easily found within and without the nascent empire.

Persian was therefore chosen as the dominant language of the emergent empire. Persian-writing personnel formed the basis of Akbar’s reorganization of upper-level administration as he broke free from the tutelage of Babur’s old retinue in the 1560s. Imperial efforts certainly standardized reporting in Persian across the core provinces of the empire. It became an intermediary language between native speakers of other tongues, both at the court and in major provincial centers. Around 1594, Akbar’s empire comprised 107 major districts divided into 2,737 subdivisions.Footnote ³³ Many of the subdivisional centers were probably the lowest level of any real Persian-literate administration. Muzaffar Alam certainly exaggerates when he declares that “Persian, then, had practically become the first language of North India.”Footnote ³⁴ Alam is correct, though, that founding emperors, especially Akbar, were not content to exercise merely a paramount authority, but aspired to “evolve a political culture….” They then centered their courtly enactment of it on Persian.Footnote ³⁵ He does not however interrogate why any of the Mughal emperors were in a position to make such decisions, what gave them that power and what its limits were.

The Mughals were a Timurid conquest state quite unlike the Safavid millenarians. The practices that created and maintained their empire were drawn from Timurid tradition, one still inflected with Chinggisid ideals and practices.Footnote ³⁶ That inheritance included endowing the king with a sacred power overriding earthly theologians of all sects. Resistance to his will deserved fearsome punishment. The Indian Mughals were the most ecumenical of the empires of their era. Important early emperors, notably Akbar and Jahangir, patronized several schools of literature and broke new ground in sponsoring translations from Sanskrit into Persian.Footnote ³⁷ But they had to first build an empire. Resources for cultural innovation came from the exaction of heavy taxes enforced by the sacking and enslavement of entire villages that resisted imperial demands.Footnote ³⁸

The Mughal emperors additionally created a more organized administrative apparatus than any previous imperial regime in North India had done. The large influx of New World bullion energized the cash economy and enabled the building of an expat-heavy centralized apparatus. But while upper-level administration was run in Persian, village records, where they existed, were still kept in regional Indian languages by hereditary village accountants known in North India as paṭwārī. Across the central Gangetic plain, records were written in Kaithi, a caste script impenetrable to outsiders. Akbar’s edict (if it ever existed) about using Persian probably never reached deeper than the collector’s office where a few multilingual clerks would translate—often orally—between Persian and various regional records.Footnote ³⁹ Potemkin villages, and entire fictitious districts, sometimes existed in Persian records but not on the ground.Footnote ⁴⁰ Yet Persian script was widely used across the empire in major towns, the Mughal court, and in diplomatic correspondence. But it had its limits and large areas easily lapsed back into regional vernacular forms. When the British took over administrative charge of the Mughal province of Bengal in the 1760s, even the revenue records of large districts were submitted in the regional language, Bengali. The highest administrative body in Calcutta reported to London that they had received the “very voluminous” accounts, but these had first to be translated into Persian in Calcutta and then translated into English for dispatch to London. And this could not be done before the last ship sailed. In another letter, it emerges that the council had only one Anglophone translator from Persian.Footnote ⁴¹ This shows the implausibility of Alam’s claim that “in Bengal, administrative papers prepared and issued in the name of local Hindu intermediaries were in Persian.”Footnote ⁴²

Spooner and Hanaway claim that Persianate officials “constituted a professional class with high social status, political involvements, literary talents, and wealth.”Footnote ⁴³ They are clearly thinking of a tiny handful of men, equivalent to chair professors or football coaches at major universities today. But even the highest of scribes might still be executed at the whim of a suspicious minister.Footnote ⁴⁴ Syed Ahmed Khan, writing under the British, described how respectable men were “addressed harshly and often abused by their [English] superiors when reading out documents to them.…”Footnote ⁴⁵ Below these worthies were those bearing the real burden of sustaining Persian as an administrative language. They were a scribal proletariat, analogous to the graduate assistants and adjunct lecturers who form the submerged workforce of the corporatized university today. They, too, were subject to cut-throat competition for patronage and emolument.

Creolization: The Maratha Case

The Maratha courts that replaced Mughal rule over much of South Asia in the eighteenth century used a Persianized register of their own ethnic language. O’Hanlon writes, “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 Persianized Devanagari, rather than using Perso-Arabic script.”Footnote ⁴⁶ There is no good reason to believe that such a choice was available to them. When the English envoy Elphinstone was at the Court of Nagpur, he wanted to speak privately with the Raja’s minister Sridhar Pandit and asked him to come to another tent. The minister said they would converse in Persian, which no one else present understood. At another time, during a direct negotiation with Raghoji Bhosle, the king only spoke in Marathi which Elphinstone did not know. Sridhar Pandit interpreted for both, probably through Hindustani.Footnote ⁴⁷ This mirrors a court setting in which even letters were not read directly by recipients, but rather read out to them, sometimes being translated extempore.Footnote ⁴⁸ Such practice is what caused the political intelligence from Pune sent to the Nizam in Hyderabad to be written in Marathi. These were obviously read out and perhaps translated by Marathi-reading clerks in Hyderabad. I have argued that the practice of reading aloud would produce a large number of courtiers who broadly understood and used Perso-Arabic words or phrases without being able to write them in the original script.Footnote ⁴⁹ Persian, however, had great cultural cachet and would quickly infiltrate Marathi, usually via spoken Urdu/Hindustani. Edward Moor, who spent many years as a junior officer in the Bombay Native Infantry, remarked that in “conversation, if Hindvi [Hindustani] be the language, which it generally is in our communications with these people, and a Mahratta hears a new word that pleases him, he will enquire its meaning, and perhaps in half an hour introduce it into his own discourse with another Mahratta….” But all across the Indian peninsula, Moor rarely met anyone who could speak Persian. Though his journeys covered several thousand miles, “we rarely, although they were sought, met with opportunities of conversing in it.” Footnote ⁵⁰ The routine garbling of Persian titles in Marathi records also suggests that this was the case. The common speech in even the central chancery of an important Persianate kingdom in South India is revealed by the lexicographer Patwardhan’s observation that Marathi nouns crop up in their Persian official edicts.Footnote ⁵¹

The administrative use of local language is even more strikingly revealed by treaties between the Nizamshahi regent, Malik Ambar, and the Portuguese. The treaty of 1615 regulated their mutual boundary around the fortress of Cheul. It was “signed” by Agi Acute Istamboly and Habascan (Haji Yaqut of Istanbul and Habas Khan) on behalf of the Sultan and Manoel D’Azevedo and Luis D’Almada for the Viceroy. The treaty was only written in “letra Hindou” (probably Modi) and Portuguese.Footnote ⁵² This would indicate that no Persian scribe was available at the time. Almost two centuries later, Hyder Ali, ruler of Mysore addressed the official Portuguese emissary-translator, Sadasiva Camotim Vaga, in simple Hindustani: “sab meri bat : tumaré Cavanda cum lighó: Samjanám.”Footnote ⁵³

Since English officials such as Elphinstone who studied “Oriental languages” had often begun with Persian, communication continued to be mediated by extempore oral translators, much as in the seventeenth century. We get a glimpse of this process at work in a diplomatic dispatch sent to the British governor-general from C. W. Malet, the English envoy to the Maratha court in Pune. It contained an English translation of a document written in Persian by the ambassador’s confidential Persian secretary. The Persian text had been composed extempore from the oral presentation by the Peshwa’s emissary of a Marathi letter whose script was probably unintelligible to anyone in Malet’s entourage. That letter was a reply to Malet’s earlier “verbal representations” to the Maratha ruler.Footnote ⁵⁴

In other words, a document composed in Marathi was read out to Malet’s Persian writer, who wrote the gist of it in Persian from which it was translated into English. Several intermediaries clearly found employment in this situation and each one had an obvious interest in maintaining the division of labor that brought them their daily bread. Nor was this a result of the active destruction of the Persianate intelligentsia. The Marathas had displaced the Mughals in large parts of their empire but did not remove the hereditary Islamic judges and other patrimonial literati as long as they cooperated with the new regime. Persian-learning networks could still break down as the descendants of literati who inherited sinecure office as a patrimony had no incentive for study. Dhulia (modern Dhule) was an important town on the Agra-Pune road and was selected as the headquarters of an early British district. Like many other areas governed by the Marathas, it had a hereditary Qazi (Islamic judge) and a Sanskrit-knowing Shastri. John Briggs summoned them to attend the court when he heard cases. They both appeared and participated in his proceedings. At some point, the scholarly orientalist Briggs asked the Qazi for a written fatwa, only to discover that this worthy could neither read nor write! The Shastri could read but had only a single law book from which he claimed to draw his jurisprudence.Footnote ⁵⁵ Further east, in the sizable kingdom of Nagpur, the Marathas had also left hereditary judges undisturbed. But this did not ensure the continuance of any tradition of learning. By 1825, Jenkins found that Mullahs and Qazis were “quite uneducated” and only a few could even read or write the Persian language (not to mention Arabic).Footnote ⁵⁶ This example certainly confirms Samsam al Daula’s biting denunciation of the hereditary Qazis of his day (ca. 1780?) as ignorant (and, he adds corrupt, too).Footnote ⁵⁷

Sectarian Displacement and Academic Decay

Elsewhere, many of the Persianate but Sunni literati established in North India during Mughal rule were, three generations later, dispossessed by the Persian immigrant governor Burhan al-Mulk in Awadh. He first rejected the emperor’s efforts to transfer him out of the province. Then he strengthened his regional power through a drive to establish his Persianate Shīʻa dependents in control of formerly Sunni endowments with consequent harm to Sunni scholarship. These immigrants created a network of support for their fellow Shīʻa nawabs and also introduced a Shīʻa-Sunni hostility that lasted into the twentieth century in cities like Lakhnau (Lucknow). Juan Cole draws on contemporary sources to describe how Burhan al-Mulk confiscated the stipends and land grants of both old established and new families. As a result, financial worries compelled many students to forsake their studies for immediate employment. The decline in learning and in state subventions to Persian education continued under Burhan al-Mulk’s successors, whose confiscations ruined many Muslim institutions of higher learning. There were a few exceptions to the new ruling family’s policy of disendowing Sunni Muslim schools. These occurred where a powerful religious basis and past imperial sponsorship protected the personnel and property of key literati against the new power elite.Footnote ⁵⁸

But the new Nawabs never re-gained the power that a real Mughal governor would have possessed. Only British protection prevented their destruction by the Rohillas or the Marathas. That protection was purchased by large tributes paid to the Company and its officials. Across the countryside, parvenu tributary lords with their own private armies warred with rivals and bargained with the Nawab’s officers at gunpoint. By the mid-nineteenth century, Persian at the court competed with the increasingly cultivated creole, Urdu.

The British Empire and the Second Life of Indo-Persian

We should explain why, of all the European empires of the nineteenth century, only the British in India retained Persian to a significant degree. The early East India Company grew up under the shadow of the Mughal Empire. Its functionaries developed a deep faith in Mughal administrative records and sought to secure their own claims within them. The English East India Company seized administrative control of eastern India between 1757 and 1765, theoretically while subordinate to Mughal suzerainty. Embodying the original vision of a universal empire, Mughal normative texts contained claims that supported the Company’s efforts to extract larger taxes from the countries that it ruled under the legal fiction of Mughal consent. In 1803, Company armies finally defeated the Maratha rulers of North India and took over the old imperial cities of Delhi and Agra.

The Company’s officers tried to gain an understanding of the preceding empire as a guide to their own administration. In the late eighteenth century, the ambitious James Grant secured a large body of Mughal tax assessment records for the empire’s eastern provinces and produced translations of them. They were part of his effort to show that Bengal and Bihar under the East India Company were seriously under-assessed according to his reading of the Mughal standard. Grant’s assessment records supposedly went back to the reign of Akbar (d. 1605) and successively traced additions and changes made down to the Company regime in 1765. But Grant’s starting point was that an Asian sovereign was “the sole virtual proprietor of the soil” who was therefore entitled to the entire “rental” or economic surplus of the land. He presented these records to Governor-General Cornwallis as demonstrating the sovereign had shared the gross produce of the land only with the cultivating peasant and parceled it out down to the smallest measured division of land. The powerful landed magnates who collected and remitted taxes were therefore removable functionaries who received a commission for their work. This was a Mughal ideological claim especially pleasing to a new administration that wanted to create a tabula rasa for its own projects. So even Grant’s caustic critic John Shore did not controvert the argument that authentic knowledge of Indian revenue resources could only be found in Persian records generated by the previous regime. The abundance of papers and correspondence from the Mughal period seemed to confirm this idea.Footnote ⁵⁹

The Mughals had also established Persian as the main language of diplomacy across South Asia. The grants and concessions that the early Company avidly sought came written in that language. Later, much British correspondence with other states was in Persian and a special office was established to handle that correspondence. If not only the declining but still prestigious Mughals, but also the rising British corresponded in Persian, then regional kingdoms and local magnates would need to maintain at least a small chancery able to correspond in the language.

The Company soon attracted members of the Persian-knowing scribal classes seeking to impress the new regime with the value of their Persian skills. Michael Fisher describes how they “sought to perpetuate the types of Persian-based learning that they embodied.” Since the seat of power was in London, several Persian scribes from India and Iran found their way there. They initially secured significant positions at high salaries. But neither in Britain nor in India were they allowed to colonize the British administration as they had the Mughal one. In Britain and then British India, they were marginalized within a few decades by Persian-knowing Englishmen.Footnote ⁶⁰ Charles Stewart, Persian professor at the Company’s college, wrote in 1816 that “such is the prejudice of Young Men against the Tuition of a Native of India, that only the few steady ones derive any benefit from his Lectures.”Footnote ⁶¹

Since learning Persian as a foreign language took several years of study and considerable effort, “the few steady ones” who invested in the language were committed to it. They would naturally vaunt the value of their skills. Persian-knowing secretaries (munshis) working off-stage remained necessary to many British officers once they got to India. These secretaries’ dominance of diplomatic correspondence had network effects in maintaining Persian scholarship. But meanwhile, the far more numerous body of poorly paid teachers continued to work as tutors, supplementing their earnings by copying manuscripts or writing school texts that they sold to their students.

Late eighteenth-century debates around Company policy, as well as Cornwallis’ reforms (1786–1793), were driven by a serious crisis in East India Company affairs. Peculation and ineptitude had turned the envisioned riches of Bengal into an empty dream, both for the Company treasury and the British Exchequer. There was an evident need for administrative reform based on authentic knowledge. Yet, the legend of a bygone system that had produced boundless wealth for the Bengal Nawabs appeared uncontestable. And the potentially unemployed functionaries of the previous Mughal administration hastened to claim that the necessary knowledge existed in Persian and they would find it if only the English would employ them. But problems persisted. Many early administrative failures were blamed on the misdeeds of “native agents” or “banians” whose skills had made them indispensable to their English masters. The latter therefore needed to learn local languages to surveil their doings. In the early years of Company rule, the famous William Jones warned that using Indian translators exposed confidential transactions to men “on whose fidelity they could not depend.…”Footnote ⁶²

It followed, therefore, that men like the Persian-knowing James Grant possessed an essential skill. The early East India Company continued and perhaps even intensified the use of Persian in the provinces that it acquired across northern and central India. But it also encouraged or required its English employees to learn it.Footnote ⁶³ Finally, the Company centralized political and administrative power to an unprecedented degree. It may actually have sustained and increased the study of Persian by those who looked forward to becoming a tentacle of the powerful new administration or to making money in its newly energized and powerful law courts. Men who sought employment in the corresponding head offices (pidgin “cutcherries”) of great landowners thus had an additional incentive to learn the language.

At the same time, the immense wealth thought available to Company officials spurred assiduous competition for these appointments. With Lord Cornwallis’ reforms (1786–1795) came a more bureaucratic, better regulated, and now highly paid “Covenanted” Service. Their enormous salary bill constantly threatened the Company’s financial stability and so it sought to limit the number of posts in that Service. Entry into it was regulated by patronage carefully allocated among the Company’s governors in Britain and responsive only to the political pressure of the London power elite. Proficiency in Persian still offered an advantage to trustworthy (meaning English) but “uncovenanted” outsiders seeking better paid posts in the government. Such lateral careers were a recognized “perk” of the Company Army’s officer corps, which became the main source of language scholars. It also dwarfed the Covenanted Civil Service in size: there were six thousand army officers in 1854 versus eight hundred “civilians.”Footnote ⁶⁴ By 1798, the Army had adopted promotion by seniority, creating a frustrated surplus of subalterns who were shut out of the most lucrative posts. Yet they already had to learn a local language—usually a simple version of Hindavi or Urdu (often called “Moors” or “Hindostanee”)—in order to function in their units. Hiring a tutor to upgrade what was often a Persianate pidgin to standard Persian allowed a proficient officer to move to more lucrative jobs while preserving his military seniority. He could then work as an “Oriental Translator” or join the ranks of British officers foisted as Residents or Residency staff on dependent princes like the King of Awadh or Nizam of Hyderabad. Service in diplomatic or intelligence missions offered another path to better earnings and more interesting work than the routine of parade ground and officers’ mess.

Anyone acquainted with nineteenth-century British records from India will have observed how many military officers served in non-military departments. Regular boards of examiners existed to certify the competence of officers seeking extra pay or better appointments. In 1826, for example, Lt. William Hunter was examined in Hindi/Urdu and Persian in two scripts and successfully translated into and out of both. Having passed, he became eligible for executive office outside his regiment.Footnote ⁶⁵ The Company preferred its own army officers who would not be removed to other parts of the empire. Indeed, one Royal Army officer with decades of imperial service cited the lack of special appointments for Persian scholars with the Queen’s Commission as a drawback for their Service. He gave the example of Colonel Longden, who was “a first-rate Persian scholar, a first-rate colloquial scholar and a man of the highest intellect” but could never get anything better than overseership of a canal or an appointment on the roads.Footnote ⁶⁶ Meanwhile, though administrative emphasis on the language came just as senior governmental appointments were removed out of Indians’ reach, the prospect of maintaining some status sustained its cultivation among Indians. But such pragmatic study by the shabby genteel would not lead to any great proficiency.

The extent and level of competence achieved may be ascertained from William Adam’s careful and protracted inquiry in the 1830s, when Persian was still the official language. He found that in the districts neighboring the British capital of Calcutta, Persian was taught to small groups of students in domestic settings. But wherever substantial Muslim landowners existed in interior districts of Bengal, Persian was also sustained for both its pragmatic value and as a legacy of past Islamic rule. These landowners saw it as “the language of the former conquerors and rulers of Hindustan from whom they have directly or indirectly sprung, and the memory both of a proud ancestry and of a past dominion—the loyalty which attaches itself rather to religion and to race than to country—attract them to its cultivation.” Its study was thus motivated by a language ideology that identified Persian as an Islamic conquest language. Adam believed that there were a few learned men who did not need to collect fees who taught the language out of ideological zeal. They imparted the learning of “their faith and race” (i.e., Arabic and Persian instruction) freely as being “productive of moral and religious benefit to themselves and their fellow creatures.”Footnote ⁶⁷ Despite all this, Persian scholars were not numerous even in the 1830s when it was still the major administrative language. The population of Hugli (Hooghly), a district close to Calcutta, was estimated at one million. The Mohsin Trust taught Arabic and Persian. In 1831, the former had sixteen students and the latter only seven; English on the other hand, enrolled sixty students. Adam also found a solitary government grant-supported school with twenty-five students of Persian. Demand for English arose from the opportunities that proximity to the new masters now offered. In Calcutta, it had called forth a number of English schools conducted by Indians. The Armenian school sustained by that community also taught both Armenian and English: the ethnic sacred and functional secular respectively.Footnote ⁶⁸

Adam also heard reports of desultory instruction offered in the households of some wealthy men. In general, students were not numerous and those studying under a single instructor rarely exceeded six. They sometimes lived under the instructor’s roof, ran errands, and provided menial services. Any disagreement would terminate a course of study. Scholars learned the alphabet from one, a simple “wisdom text” from a second, and so on “till they are able to write a tolerable letter and think they have learned enough to assume the title of Munshi, when they look out for some permanent means of subsistence as hangers-on at the Company’s Courts.”Footnote ⁶⁹

Persian was still sought by some parents simply because the British used it in the law courts and police administration. “Some Hindu landholders and other respectable natives have expressed to me a desire to have Persian instruction for their children, but they apparently had no other object than to qualify them to engage in the business of life, which unhappily in their case is for the most part identical with the business of the courts.” Even so, there was an oversupply of teachers. The earnings of those proficient enough to teach the language were scanty, even in Bengal and Bihar in the 1830s when Persian was still an important official language. We may consider a major center like Natore (in modern Rajshahi district). There were four Persian schools with a total of twenty-three students. Teachers were paid small monthly stipends with food and small perquisites. The aggregate in cash and kind ranged from four to ten rupees a month, the average being seven.Footnote ⁷⁰ We may get some idea of how small this pay was if we consider that around 1850, a newly arrived British Army officer was expected to maintain at least six servants, ranging from a sweeper at four rupees monthly to a bearer or footman at seven.Footnote ⁷¹

We also have a comprehensive survey of indigenous schooling from North India (mainly today’s U.P.) around 1850, when official Persian was only gradually being replaced by a Persianized Urdu in the same script. All schools in eight districts were surveyed. There were in total 3,137 teachers of all kinds, of whom 460 offered free instruction. The 2,677 who were paid received a total of 10,785 rupees per month, or just under 4 rupees each. Many, however, also received either cooked or uncooked food. The Persian teachers often fared “more luxuriously,” getting two meals with 2 pounds of wheat bread, 4 ounces pulse and 1 ounce of ghee. The value of free food was guessed to be from 1/2 to 1 anna daily, or approximately 1.25 rupees monthly. Nonetheless, 694 Arabic and Persian teachers still earned less than 5 rupees a month in total: 223 received more, and eleven earned from 12 to 20 rupees.Footnote ⁷² Most teachers, therefore, earned less than an English officer’s bearer.

There were 27,736 students in all, out of a population of approximately twenty million. Of the students, less than one-third (8,507) were studying Persian. So only about one out of perhaps two thousand persons in the heartland of the old Mughal Empire was seeking to learn the language.Footnote ⁷³ The difficulty of sustaining a foreign tongue may be gauged by data from nineteenth-century North India, where vernacular Urdu was studded with Persian loan-words and phrases. Even so, it was reckoned in 1850 that it took nine years of study for a boy to attain competence in Persian. But 60 percent of 8,507 students of Persian had been enrolled for less than three years and only 9 percent more than seven. Not surprisingly, barely one-fifth (1,746) were able to write, read, and̄ comprehend what they read “if the book be an easy one.” The remainder could at best “only read, being unable to give the sense of any author that they had not previously encountered.”Footnote ⁷⁴ For comparison, we should note that a decade later, a single print run of Shiv Pershad’s vernacular history textbook for colonial schools was ten thousand copies.Footnote ⁷⁵

Valorization of Persian was accompanied by a contempt for vernacular Urdu in parts of North India. Around 1850, the new government schools (“Tahsili schools”) in the Shahjahanpur and Bareli districts attracted unusually few students because they taught Devanagari (“Hindi”) and Urdu scripts. Many preferred the cursive Kaithi script to Devanagari because of its use in village records. But the scribal caste of Kayasthas who controlled village accounts still valued additional Persian study for their sons as opening doors to higher appointment. And among the Rohilla Muslim gentry of Shahjahanpur and Bareli regions, the “study of Persian [was] popular and Urdū [was] proportionately looked down upon.…”Footnote ⁷⁶ Writing in 1858, Syed Ahmed Khan described the spread of Urdu and English as being a special source of religious grievance to Muslims that (alongside other complaints) drove them to join the great uprising that year.Footnote ⁷⁷

The valuable report of 1852 was also used by Tariq Rahman in his article on colonial-era Persian. I cover some of the same ground but disagree with Rahman. His essay is mainly focused on North India, which he evidently deems equivalent to British India as whole. His citations are selective, and sometimes actively misleading. For example, a single official’s explanation about poor enrollment in new government schools in two districts in their first year of existence is attributed to “officials” (plural) and presented as describing the entire “Urdu-Hindi belt.”Footnote ⁷⁸

Bureaucratic Capture and Language Choice in Punjab

The Sikh rulers who followed the Mughals perpetuated the positions and perquisites of religious worthies of all faiths. The intensely cellular but hierarchical structure of South Asian society has historically acted to preserve multilingualism alongside language hierarchy.Footnote ⁷⁹ The British annexed the whole Punjab in 1849. Sufi centers, “Hindu” monastic seats and temples, as well as custodians of Sikh gurdwaras were all found to have grants from the period of Sikh rule (ca. the 1760s to 1849). Colonial education officers thus found three languages and four scripts represented. These included schools teaching Landa (a cursive merchant’s script related to Devanagari) to boys of the business classes. Sufi religious centers usually had schools open to all communities.Footnote ⁸⁰ The results of the Sikhs’ impartial distribution of the agrarian surplus among religious and educational centers were evident even in 1880, when at least four different languages and scripts were taught across the greater Punjab region. But most students studied only a single script and in general education that script was the Persian (Urdu) character.Footnote ⁸¹

The Persian language retained some prestige. Punjab’s Director of Education T. W. Arnold wrote that in the province the great object of Persian education was to teach a boy to read the poet Saʻdī’s Gulistan and Bustan (two thirteenth-century works). The lad who could read a page of either in a fluent sing-song had received an education that fully satisfied both teacher and parent.Footnote ⁸² The new rulers faced an administrative dilemma: varieties of Punjabi were the area’s common language, but few of the British officers or the numerous clerks their government needed knew it. A large and loyal class of subordinates was trained in Urdu in adjoining British provinces and Urdu therefore spread as New Persian had done. British and indigenous Urdu writers from the older territories found fresh pastures in the British Punjab. They naturally deprecated the regional language and vaunted the one they knew.Footnote ⁸³ Nonetheless, a spoken and increasingly printed Punjabi bubbled up despite official indifference. Mir traces how literary variants of the indigenous language grew in literature and society despite colonial patronage of Urdu.Footnote ⁸⁴

Meanwhile, a decade after the British annexation of Punjab, Sikh gentry were seeking English tutors for their sons, and thus attracting Bengalis who had a smattering of the language to Lahore.Footnote ⁸⁵ Such changes were unsurprising if we consider the effort and expense needed to sustain a foreign second language. Rather than Persian, English was a logical choice for those able to afford either. Urdu was more easily acquired and therefore learned by seekers of subordinate government employment.

Controlling the Archive and the Refusal of Print

In an earlier section, I suggested that Aramaic clerks created linguistic complexity in order to control the archive. A similar drive explains the steady refusal of print technology in South Asia up to 1800. Widespread adoption of printing would have endangered the livelihood and social status of castes and lineages of copyists. These were powerful cliques, especially when lodged in state offices. An anonymous seventeenth-century author spent several pages denouncing the extortions and oppressions of “Hindu clerks and drowsy writers…” He asked rhetorically if any man had ever seen tyranny equal to the demand that each letter of a soldier’s identification marks be “written by a [different] clerk?”Footnote ⁸⁶ Obviously, this arrangement ensured that each of them got a share in whatever could be extracted from the client. Even paying off a Mughal governor, the East India Company discovered, could not ensure compliance by his clerks. The Company paid 5,000 rupees and a document was duly prepared in the governor’s office, but then the clerks took hold of it and refused to hand it over, demanding 800 rupees for themselves, and finally settling for 350.Footnote ⁸⁷

Scribes fortunate enough to be in at the founding of a lasting government would therefore seek to capture the chancery and archive and turn office into patrimony. Karen Leonard has shown how a few Kayasth families came from North India just as the initially insecure regional governorship acquired by Asaf Jah I (c.1748) stabilized under British protection. They quickly took charge of key offices and cemented their control by forming a small intermarrying caste observing bilateral inheritance. They preserved their power by maintaining charge of the archive of administrative records. Even the “reforming minister” Salar Jang I could not dislodge them. They held the key to understanding records of tax-arrears, tax-farms, and many other state matters. But, supported by the new colonial power, Salar Jang was able to bypass them by creating a treasury and records office under his own control. Gradually, therefore, an administrative record accumulated outside of hereditary officials’ control and their archive was thereby rendered obsolete. The new offices were increasingly operated by an Urdu-speaking and writing service gentry drawn from British territory, especially Awadh.Footnote ⁸⁸

The East India Company was gradually being transformed from a great corporation to a modern bureaucratic state. Printing was enthusiastically embraced for proclamations and official documents. It was an obvious solution for replicating the massive volume of reports and returns sent to Presidency capitals in India, transmitted to its headquarters and on to Parliament in London. School textbooks and missionary tracts began to roll off presses in British Indian towns. The invention of lithography made mass printing available for Persian and for other languages that used cursive scripts. The nineteenth-century adoption of printing, whether lithographic or with movable type, drastically changed the social role of writing. Nile Green is one of few historians to consider how printing changed markets and society in South Asia and around the Indian Ocean.Footnote ⁸⁹ Green’s important observation ties in with the long rejection of printing in South and Southwest Asia. It was only when mass-produced tracts by missionaries began to pose a serious threat to the indigenous intelligentsia that they energetically deployed printed tracts in reply.Footnote ⁹⁰ By contrast, mid-nineteenth-century teachers in North India occasionally earned a few rupees by copying texts for their employers. They also safeguarded their own status by using manuscripts (“nostrums”) that few others possessed. This, I argue, was why in 1850 there were 224 separate texts read in the 1,257 North Indian schools that taught Persian. Of those, 105 were read in one school only. Each represented a few rupees in extra income to a poor scrivener as well as a small bulwark against competitors.Footnote ⁹¹

The economics of manuscripts were very different from those of print. Printed books depended on widening markets to reduce marginal cost per copy, while the tenth copy of a manuscript cost just as much as the first. As is well known, the first print technology in Asia was the Chinese wood-cut print. Wood-cut block printing was used in India for textiles but not texts. Movable type was introduced in the Portuguese colony of Goa in the 1550s. Apart from the Roman font, Church workmen also created a “Malabar” font, but as Qaisar notes, there was no local interest in adopting this innovation. Qaisar strains to avoid explaining this as being due to resistance from calligraphers (and scribes), even though he finds no other plausible explanation. Indeed, he unreasonably rejects the English traveler Ovington’s direct and contemporary statement that this was indeed the cause.Footnote ⁹²

Only with colonial rule and the missionary threat was scribal resistance broken and mass print replaced expensive manuscripts that could only be purchased by wealthy cognoscenti. But printed works had to be in the new rising vernaculars. The world of print was not “the thinly spread, networked geography of Persian as lingua franca….”Footnote ⁹³ Print media was now necessarily addressed to and priced for a larger public. Delhi College textbooks of 1840 were Urdu translations from English. Theological studies of Arabic were debated and taught through Urdu in the new Islamic center of Deoband in North India.Footnote ⁹⁴ Courts and bureaucracies in British India began to operate in Persianized Urdu in the north and other regional languages elsewhere. At the uppermost levels, government proceedings were largely in English.

Persian’s Administrative Failure and the Rise of Bureaucratic Vernacular Languages

The new print-adapted bureaucracy grew as the East India Company discovered how little ground-level information actually resided in Persian records. The disjuncture between reports and realities now revealed originated in the long-standing insulation of vernacular village recorders from Persian-writing administrative offices. Even in the Mughal heartland of North India, village records were seemingly almost always kept in a special scribal script, Kaithi, one little known to Persianate literati. Awadh was a deeply Persianate Mughal successor state dominated by an immigrant-heavy Persian-speaking Shi̔i elite ruling from Lucknow (Lakhnau).Footnote ⁹⁵ In 1867–1868, a decade after Awadh was annexed, the colonial government attempted a provincial census there. The Census Commissioner discovered that village statistics could only be compiled through village accountants, who knew only the Kaithi script. When it was learned that he was hiring clerks to process these returns, “150 to 200 respectable Mohamedans, well dressed and tolerably educated” applied for these short-term posts at 15 rupees a month. But none of them could read Kaithi and therefore process the village census returns that—three hundred years after Akbar supposedly decreed that they be in Persian—were only available in that script.Footnote ⁹⁶

The push for a combination of English and accessible regional vernaculars in administration was thus generated by the slow realization that the old administration had little idea of “facts on the ground.” The new, modernizing British administration needed a new class of Indian subordinates who were educated to a uniform standard and equipped to interface with a small, bilingual colonial elite that included engineers and surveyors. Printed forms of statistics and reports began to be distributed across entire provinces in new, standard languages such as print Bengali, Marathi, Gujarati, Devanagari Hindi, Urdu, and others. The new school system was oriented to fit these systems. Village and district record-keepers no longer inherited their positions and were now required to attend training programs and pass examinations. By 1884–1885, even including late-modernizing NWP and Oudh (modern UP) districts, there were training schools for village accountants in every district except two Himalayan ones.Footnote ⁹⁷

The requirements of the new administration thus gradually created a new standardized school system and a matching bureaucracy. These modernizing processes, in turn, pushed the question of a “ground-level” choice of language and script to the forefront. Across much of North India, the already established and widely spoken Hindi/Urdu was the functional default option. But selection of the script that would render sounds into signs became a bone of political contention between advocates of a modified Perso-Arabic script and those of a print-shaped Devanagari.Footnote ⁹⁸ That polemic has had deep political ramifications down to the present but falls outside the ambit of this article. However, Urdu in the Perso-Arabic script was valorized as the language of the Muslims only when Persian had become irrelevant. Formerly viewed suspiciously as a bearer of Christianity, Urdu quickly displaced Persian as the language identified with the heritage of South Asian Islam.Footnote ⁹⁹ Census commissioners often found it labeled “Mussalmani” in “parental tongue” reports.Footnote ¹⁰⁰

We can see the tussle of Urdu pragmatics and Persianate ideology in Hyderabad state. It was, like Awadh, a state that survived under British protection. It claimed to sustain the Mughal tradition of administrative Persian until 1883, even though the vast majority of its subjects spoke Telugu or Marathi. The chief minister Salar Jang I (in office 1853–1883) ideologically championed the language as a symbol of Islamic power. He sedulously adopted colonial administrative “reforms” and imported many officials from British territory. But he still valued Persian as representing the conquering past of Islam. When the idea of replacing Persian with Urdu was raised before him, he exploded: “You Hindustani (North Indian) people have little practice in Persian speech and writing. Persian language is the symbol of the victory of the Muslims. We have conquered this land with the sword. Having destroyed this symbol in your own country [North India], you people now want darkness here too. Persian shall remain here and flourish so long as I am alive.”Footnote ¹⁰¹

But the Urdu-writing bureaucracy, often of North Indian origin and connected to the colonial government, ultimately prevailed. Persian was dropped even in Hyderabad state, a process made inevitable by Salar Jang’s own need to create a “modern” colonial-style bureaucracy bypassing the masters of the older archive. “New men” across British India were embracing Urdu as their “ancient” heritage. In the Hyderabad census of 1891, almost 1.2 million reported that they spoke Urdu, while seventy-eight thousand reported “Hindi.” The Census commissioner suggested the two could be deemed a single spoken tongue used by about 10 percent of the Nizam’s subjects. The waning of Persian language ideology had been swift: only 815 persons reported speaking Persian, of whom 186 were born in Persia. Arabic had preserved its prestige, being reported as a parental tongue by 12,869 though only 4,810 were born in Arabia.Footnote ¹⁰²