Inscriptions as Historical Source Material

My paper is an exception at this colloquium, in that it discusses the study of ancient and medieval villages by examining contemporary inscriptions, unlike other papers, which study modern and contemporary villages by examining government statistical records or data from researchers’ field surveys. It is well known that India lacks history books compiled in the ancient and medieval periods on Hindu dynasties, other than a few like Rājataraṅgiṇī. In compensation, however, there remain a large number of inscriptions from the remote past that enable us to reconstruct the history of ancient and medieval Hindu dynasties, although there is a difference between north India and south India in respect of the number of inscriptions that have survived from the past.

First, let us look at the distribution of the inscriptions by language and period. The distribution by language is as follows:

Table 1 shows that the Dravidian-language inscriptions remaining in the south far surpass, in number, the Aryan-language inscriptions which survive largely in the north, including the northern part of the Deccan.² Inscriptions are mostly engraved on the stone walls of Hindu temples, and the fact that a far larger number of ancient and medieval Hindu temples remain in the south than in the north explains this difference. To some extent, the same reason may explain the difference in the distribution of inscriptions between the three Dravidian languages.

The chronological distribution of inscriptions, if we take up Tamil inscriptions for example, is as follows.

Table 2 shows that we have a large number of Tamil inscriptions from the 10th century to the 16th century, within which fall the periods of the Chola, Pandya and Vijayanagara states. More or less the same tendency will be found in the chronological distribution of inscriptions in other languages too.

Village Studies Based on Inscriptions

Inscriptions usually refer to the reigning king with his regnal year or the year of an era, such as the Saka era, on the basis of which we can reconstruct the political history of the period to a certain extent. Moreover, as the contents of inscriptions, especially of Tamil inscriptions, are various, with many different matters being recorded, we can also learn about the socio-economic conditions of the period from them. Inscriptions engraved on temple walls mostly record grants of money, cows, land, state revenue as tax, etc., made to that temple for conducting daily worship, festivals and repairs therein. Matters are often recorded in full length; if we consider, for example, land grants, the inscriptions often go into such detail as to inform us how the granter acquired the land, who should cultivate the land, what should be cultivated, what sort of taxes the temple should pay or be exempted from, etc., in addition to the extent and location of the land as defined by four boundaries.

Inscriptions also record other matters, such as government orders to the locality; decisions made by local assemblies such as an ūr (assembly of an ordinary village called by that name), sabhā (that of a Brahmana village), nagaram (that of a town) and nāḍu (that of a local unit called by that name); the solidarity pact of a particular community; revolt resolutions by some oppressed social groups; disputes among people and villages; and political compacts between local chiefs. Therefore, with regard to the Tamil country, we can also undertake village studies based on inscriptions for the period from the 10th to the 16th centuries – as I have done.³

In past studies, efforts have been made also to elucidate the conditions of ancient and medieval villages by examining Tamil inscriptions. The best example of this is Nilakanta Sastri’s Studies in Cola History and Administration,⁴ in which he clarified the functioning of the sabhā formed by Brahmana landholders in Brahmana village called brahmadēya or chaturvedimangalam. The sabhā managed village affairs, including cultivation, through various committees (vāriyam) formed under it, including those for garden land (tōṭṭam), tanks (ēri), wet fields (kaḻani), taxation (pañchavāram) and accounts (kaṇakku). The method of electing members to the sabhā and vāriyam was also clarified by Sastri. About the same time as Sastri conducted these studies, A. Appadorai studied the agrarian society of medieval south India by examining inscriptions in Tamil and other languages, in his Economic Conditions in Southern India (1000–1500 AD).⁵ Landholdings in villages were also discussed in this book.

Although these studies are excellent pioneer works, they have their own problems. One problem concerns their nationalistic bias, related to the national movement of the 1930s. Sastri and others tried to prove the existence of democratic local government in ancient and medieval India, and gave much emphasis to the democratic way of electing sabhā members in Brahmana villages. Other aspects of village study were left out without being examined.

Another problem concerns the way the studies treat inscriptions. In most cases, Appadorai picked up only one or two inscriptions to support his argument although there were many relevant inscriptions to be analysed, including ones that suggested contrary arguments. Moreover, Appadorai depended, for his argument, not on the original inscriptional texts, but in many instances, on the brief gist of the inscriptions published in English in the Annual Report on Epigraphy. In the case of general works such as his, especially in a pioneer work that is not based on more specific work conducted previously, this deficiency may be permitted to some extent. Unfortunately, however, this attitude is still prevalent in recent works. That is why I introduced the statistical method in the study of inscriptions -- dealing with a related corpus of inscriptions in place of single inscriptions, in order to avoid the arbitrariness that is conspicuous in previous studies.  

Village Communities Revealed by
Thanjavur and Other Inscriptions

Now, let us look at village studies based on inscriptions taking up some of my past studies as examples. First, I introduce my study of village community based on Thanjavur inscriptions, leaving an explanation of the statistical method to a later section. There are three long and continuous inscriptions in the Brihadisvara temple in Thanjavur (SII, ii, 4, 5 and 92), which record revenue grants from more than 56 villages in south India and Sri Lanka. The inscriptions start with a preamble that runs as follows:

Hail! Prosperity! There was engraved on stone, as orally settled, the revenue in paddy, – which has to be measured by the measure (marakkāl) called (after) Adavallan, which is equal to a rājakēsari measure, – and the gold and money, which has to be paid from the land paying taxes; and there was also engraved on stone the land free from taxes, – including the village-site, the sacred temples, the ponds, the channels passing through the villages, the quarter for Paraiyas, the quarter for Kammalas, and the burning ground, – in the villages, which the lord Sri Rajarajadeva had given in the Chola country (maṇḍalam⁶), and in other countries as divine gifts for the expenses required by the supreme lord of the sacred stone-temple, called Sri Rajarajesvara, – which the lord Sri Rajarajadeva has caused to be built at Tanjavur….

After this edict, full details are enumerated for 40 villages in Chola-mandalam and abridged records for 16 villages in other maṇḍalams, including Sri Lanka. For each of the 40 Chola-mandalam villages, therefore, we are able to know the total extent of the village, the extent of taxable and tax-free land, the amount of tax imposed on the taxable land, the categories of tax-free land, and the categories of residential quarters, temples, etc. From these we also learn about the existence of threshing grounds and granaries as tax-free land, and of residential quarters for the Ilavar too, though they are not mentioned in the preamble. From the amount of tax on paddy and the extent of taxable land we can calculate the rate of taxation, which was more or less 100 kalam of paddy per vēli  of land – though a similar inscription in Gangaikondacholapuram (SII, iv, 524) gives us a somewhat different ratio, ranging from 16 to 92 kalam per vēli.⁷ Chart 1⁸ in the appendix shows the result of analysis of these matters for the 40 Chola-mandalam villages of the Thanjavur inscription and some 7 selected villages of the Gangaikondacholapuram inscription.

In respect of residential quarters for various communities, another type of inscription – which record royal grants of villages to Brahmanas for establishing a brahmadēya – reveals the inclusion of a variety of village-servant groups in villages. The land in the village to be assigned to Brahmanas and others, including village-servant groups and temples, is usually expressed in terms of paṅgu (share) in these inscriptions, and most of the Brahmanas are given one paṅgu, while village-servant(s) and temples are given somewhat different shares. In the Tandantottam plate inscription of the Pallava king, Nandivarman II (SII, ii, 99: 789 CE), the assignments are as follows.

A Chidambaram inscription of Vira Pandya (ARE 1959-309: 13th century) reveals the following distribution.

From these inscriptions we are able to know about the existence and composition of village-servant groups in Brahmana villages. More or less similar groups, except for vēda/sūtra teachers, seem to have existed in non-Brahmana villages too.⁹

Two Different Types of Landholdings
in Two Different Types of Villages

Next, I give an example of my study of landholdings.¹⁰ My intention here is to show that there was a change from the practice of common landholding as prevailed in the early Chola period (roughly, the 10th and 11th centuries) to individual landholding in the later Chola period (roughly, the 12th and 13th centuries). For that, we shall first examine the existence of two different types of villages in which different types of landholding were seen in the early Chola period. 

From the early-period inscriptions of Allur and Tiruchchendurai, both close to Tiruchirappalli on the southern bank of the Kaveri river, we are able to ascertain the common landholdings that prevailed in ūr-type villages, in contrast to the individual landholdings seen in brahmadēya villages.¹¹ ūr-type villages were traditional villages that had existed from earlier periods, and were inhabited by agricultural and herding castes such as the Vellalas and Manradis. Brahmadēya villages, quite different from these, were created in and after the Pallava period by royal grants of villages to Brahmanas who had been invited from the north. As stated earlier, villagers in ūr-type villages formed an assembly called ūr, while Brahmanas in brahmadēya formed the assembly called sabhā.

In Allur, which was an ūr-type village, there remain nine inscriptions recording the donation and sale of land to a temple, and the remission of taxes on the temple’s land. Of the 11 land transactions, 7 were made by ūrār (members of the ūr assembly), 2 by talaivāchchāṉ (regulators of water supply at the main sluice), 1 by a kōyilār (temple priest?), and the remaining 1 by the temple itself. This shows that in Allur the relationship between land and individuals was very weak, with most of the land being held by the village community, a group of people or an institution. In Tiruchendurai temple there are 21 inscriptions recording similar transactions concerning land in Isanamangalam, a brahmadēya village. Out of the 21 pieces of land in Isanamangalam, 3 were transacted by the sabhā (village assembly), 4 by a paruḍai (group of sabhā members), 1 by a kiḻavar (landholder in a brahmadēya), 2 by an individual and his followers (uḷḷiṭṭār), and 13 by individual Brahmanas. This shows that, in contrast to the situation in Allur, there was a stronger relationship between land and individuals in Isanamangalam.

In both villages most of the lands transferred are described in the inscriptions in terms of their four boundaries, such as water channels, roads, some other land, etc. If we examine closely the nature of the land appearing as a boundary, we are able to determine in some cases the owner of the land. In Allur, out of the 9 pieces of boundary land, 6 seem to have been owned by dancers, a musician, an accountant of the ūr and an astrologer. These persons, however, seem to have come into possession of these lands by receiving them in the form of emoluments given to them as village servants who performed or served specific functions/professions.¹² In contrast to this, out of the 21 boundary lands of Isanamangalam, 18 pieces seem to have been owned by persons individually, and this confirms the difference between the two villages concerning landholding.

Another important finding of this comparative study was that in Allur the landholders, who made up the ūr, were themselves cultivators of the land held by them.¹³ In contrast, the landholders in Isanamangalam, who were Brahmanas, obviously did not cultivate the land by themselves; instead, they either rented it to others or engaged cultivators for wages. Therefore the villagers in brahmadēya-type villages were stratified into at least two distinctive strata, landholders and cultivators, while the villagers in ūr-type villages were basically not stratified, though the land held by astrologers, dancers, etc., in Allur may have been rented out.

Importance of Statistical Study of Inscriptions

Although landholding practices in these two villages were quite different and contrastive in character, namely, common landholding by cultivating landholders in the ūr-type village and individual holdings by landholders separated from cultivators in the brahmadēya village, since this was a case study, it does not allow us to generalize the findings. However, some other studies, including two statistical studies conducted later, warrant our doing so to a certain extent. The following table shows the results of a statistical study made by Subbarayalu of 260 land sales recorded in the published Chola inscriptions.¹⁴  He examined the chronological distribution of the people who sold or granted land by dividing the Chola period into four sub-periods, and the transactors into seven categories.

A remarkable trend revealed by this table is the decrease, with the passage of time, in the Brahmana assemblies and individuals transacting land (except for Period III), and, in their place, the striking increase in non-Brahmana individuals who transacted land. We shall discuss this point later, but the important thing to note here is the contrast between Brahmana individuals (27.8%) and non-Brahmana individuals (1.5%), and also between non-Brahmana assemblies (12%) and non-Brahmana individuals (1.5%) in Period I, which seems to indicate that non-Brahmanas, who lived in ūr-type villages, did not possess land individually in this period. Though not stated earlier, in some of the brahmadēya villages the land was held by the sabhā in common and not individually. Even in brahmadēya villages where most of the land was held individually, a part of the land was held in common, and there were many instances of such common land being disposed of by the sabhā as charity. This accounts for the high percentage of transactions by Brahmana assemblies. At any rate, this analysis endorses as well as allows us to generalize the findings of the comparative study given above. Herein lies the importance of statistical study of inscriptions.¹⁵

Individual Landholdings in the Later Chola Period

Next, I shall briefly examine the late-Chola period inscriptions to point to changes in the landholding system. On the wall of the Tiruvanaikka temple in Jambukesvaram, close to Tiruchirappalli, many inscriptions have survived from the time of Rajaraja III and Rajendra III in the 13th century, which record the sale/donation of land to this islet temple.¹⁶ The lands were scattered across several villages not far from the Kaveri river on its northern bank, and were sold or donated by people who had kāṇi right¹⁷ to the village land.

Four inscriptions record four such sales of dry land at Isanaikkurai village to the Tiruvanaikka temple, one by an individual (uḍaiyāṉ) and his brother, two by two separate individuals (uḍaiyāṉ/araiyaṉ), and one by the ūr of the village. Though some of the land in this village was owned by the ūr, other lands were held and transacted by individuals whose names are given in the inscriptions along with their titles, such as uḍaiyāṉ (literally, possessor) and araiyaṉ (literally, king); in contrast, the transactions recorded in the early-period inscriptions were made by the ūr and have no reference to individual names.

Five inscriptions record the sale of five pieces of land in Rajaraja-kurrangudi village. These five pieces of land, which constituted the whole village, were sold by 39 persons whose names were known; the records also reveal that all these pieces of land were once owned by an uḍaiyāṉ title-holder and his younger brother, from whom the 39 individuals had purchased the land they sold later. One inscription records the sale of Sembiyan-kurrur village by an individual who was an uḍaiyāṉ, but it is also known that this village was once owned by four kiḻavaṉ title-holders¹⁸ and their three brothers. Another inscription records the sale of Sembiyan-nallur village by an individual who held the title of uḍaiyāṉ/arayaṉ. This seller also appears in another inscription as a seller of three vēli of land in Tattanallur village, and in yet another inscription as a donor of seven-and-odd vēli of land in two pieces in Tandangurai village. He had inherited all these lands from one of his ancestors, who in turn is stated to have purchased the land from people who had obtained it at a government auction (peruvilai).¹⁹

The inscriptions clearly show that landholding practices changed drastically between the 10th century and the 13th century. In the 13th-century villages examined above, which were all of the ūr type, the land was held individually by persons who were described mostly as uḍaiyāṉ, and, in some cases, had grand titles such as araiyaṉ or nāḍāḷvāṉ(he who rules the nāḍu). The records also clearly show that frequent land sales took place between individuals living in ūr-type villages, as did government auctions of land. Though the villages examined above are restricted to a small area of the lower Kaveri valley, if we examine the inscriptions of the 12th and 13th centuries, we can readily see that a similar situation prevailed in many localities in the Tamil country under Chola rule. The figure of 37 per cent for sales by non-Brahmana individuals in Period IV in Table 5 explains the situation well. Individual landholdings, which had been seen mostly in brahmadēya villages in the early period, became prevalent in ūr-type villages too, in the late period.

What brought about this change? Answering this question at length is outside the scope of this paper, so all I can say here is that the change seems to have been caused by the frequent grants of land to Brahmanas and state officials by the kings of the middle Chola period. For more details on this issue, please see the Introduction and Chapter 2 in Karashima (2009).

Village, ‘nāḍu’ and the Formation of ‘jātis’

We shall now proceed to examine the relation between the village and the locality (nāḍu), which includes a number of villages. Royal orders to a locality were usually addressed to nāṭṭār(representatives of a nāḍu), to kiḻavar of brahmadēyas (leading landholders of brahmadēyas), to ūrgaḷilār (representatives of ūr-type villages) and to nagaraṅgaḷilār (representatives of towns), from which we can understand the importance of the locality called nāḍu in state administration. Nāḍu was the basic areal (supra-village-level) unit where agrarian production and reproduction were carried out in ancient and medieval Tamil Nadu,²⁰ hence it had importance in administration. There are various inscriptions which attest to its importance. Here, I quote a Pandyan inscription as an example.

An Agattiyampalli inscription (SII, 17, 549: Tj, 1299 CE) records a land grant to a temple and its tax remission made by nāṭṭavar (same as nāṭṭār) of Kunrur-nadu for the health of the king. Taxes consisting of kaḍamai and kuḍimai, including nel-kaḍamai, kāśu-kaḍamai, kuḍimai, mēr-pāḍikāval, āyavargam, nāṭṭuvari and ūrvari, were to be borne by the nāṭṭavar, who seem to have tried to show their fidelity to the new Pandyan ruler²¹ by this charitable deed.

Nobody can deny the importance of nāḍu in the agrarian society of the ancient and medieval periods, but I do not accept the understanding of nāḍu as composing a segment of the so-called segmentary state, which was proclaimed by Burton Stein.²² Stein considers nāḍu to have continually constituted a segment from the Pallava period to the Vijayanagara period without change, but it actually changed its character around the 13th century.²³ I am against Stein’s segmentary state theory itself and have criticised it elsewhere in more detail.²⁴

Lastly, let me explain the study of the issue of jāti based on inscriptions. Of course the area of a jāti group exceeds a village, and in this connection, I refer to an inscription which defines the area of the Palli people who frequently appear in the 12th and 13th century inscriptions in Tiruchirapalli, South and North Arcot, and Chingleput districts. An Aduturai inscription (ARE 1913-35: Tp, 1315? CE) records a resolution made by the Palli people in their assembly as follows (in an abridged form):

The pannāṭṭar (also called paḷḷi-nāṭṭavar) from the nāḍu and nagaram of all maṇḍalams met at the garden called Periyanāṭṭan-kā in a large assembly and decided to collect one paṇam (a coin) per bow held by members, etc., for worship in the local temple. The decision was made to revive an old arrangement made by their ancestors and recorded in an inscription of Vikramachola (1122 CE). According to that inscription a large assembly of the paḷḷi-nāṭṭavar, including all the Pallis living within the area bounded by the Pachchai hills in the west, the tank Viranarayana-pereri in the east, the Pennai river in the north, and the Kaveri river in the south, had decided to contribute 50 kāsu and one kuṟuṇi of rice from each family to the temple at Iraiyanpunchai Kurangadu[turai] on the happy occasion of the reconsecration of images recovered from Dorasamudram, the Hoysala capital, where they had been taken during a Hoysala invasion. At that time the king also permitted them to carry their banner with the words pannāṭṭār tampirān (the god of pannāṭṭār) on festival processions.

The Palli people described here composed the bowmen (archery) regiment of the Chola army, and this regiment seems to have recovered the images by attacking the Hoysala capital under the command of Vikramachola. The area of their habitation as defined in this inscription covered a hilly and dry area extending roughly 100 kilometres from north to south and 80 kilometres from east to west, in Tiruchirapalli and South Arcot districts. During the 13th century many of the ex-hill tribes seem to have descended to the plains and became agriculturists, acquiring land. We have many more inscriptions recording the activities of some of these ex-hill tribes, including Pallis, Surudimans and Malaiyamans, who increased their strength during the 12th and 13th centuries. Some members of their families grew into local chieftains, such as the Kadavarayas in South Arcot district and the Sambuvarayas in Chingleput district.

Without going into a detailed discussion of the issue here, it may also be noted that there are many inscriptions recording the activities of supra-local assemblies called chitramēḻi-periyanāḍu of agriculturists, valaṅgai (right hand) and iḍaṅgai (left hand) of lower jāti people composed of artisans and others, ainūṟṟuvar of merchants, etc., and revealing the names of their composing groups.²⁵ For village studies, such pieces of information on a locality like nāḍu, which functioned as the areal production unit, and on jātis, which increased their number during the 13th and 14th centuries, are indispensable. Tamil inscriptions, which afford such information, await more intensive study.