Inscriptions as Artifacts: Precolonial South India and the Analysis of Texts Author(s): Kathleen D. Morrison and Mark T.

Lycett Source: Journal of Archaeological Method and Theory, Vol. 4, No. 3/4, New Approaches to Combining the Archaeological and Historical Records (Sep., 1997), pp. 215-237 Published by: Springer Stable URL: http://www.jstor.org/stable/20177366 . Accessed: 10/04/2011 10:57
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . http://www.jstor.org/action/showPublisher?publisherCode=springer. . Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Springer is collaborating with JSTOR to digitize, preserve and extend access to Journal of Archaeological Method and Theory.

http://www.jstor.org

Journal

of Archaeological

Method

and Theory,

Vol

4, Nos.

3/4,

1997

Inscriptions as Artifacts: Precolonial India and the Analysis of Texts

Kathleen D. Morrison1 and Mark T. Lycett1

South

This paper examines one assemblage of texts from southern India, stone inscriptions of the Vijayanagara period, and considers both how these texts our have been studied and how that history of research has structured texts the We these ask how be of past. understanding might interpreted of sampling and recovery, with a differently, (1) under different conditions specific focus on in-field locations of inscriptions, and (2) as sources of data. We suggest that traditional information combined with archaeological source-side criticism of texts might be profitably expanded routinely to include contextual analysis, such as archaeologists apply to studies of artifacts. KEYWORDS: texts; sampling; South Asia; context.

INTRODUCTION
than that, it is a culture. More This is a paper about text as material are consti discussion of how observations made on documentary materials tuted as data in the study of historical process. The growing importance and Co in anthropological of historical perspectives theory (e.g., Comaroff

maroff, 1992; Fabian, 1983; Ohnuki-Tierney, 1990; Sahlins, 1985;Wolf,

1982) forshadows the recent renewal school in explorations [evident, for example, structural history (Blintliff, 1991; Cobb, 1991; Hodder, 1987; Knapp, 1992)]. In a paper of limited scope, we cannot hope to trace the many uses and or concep of history as evidentiary source, subject matter, understandings tual framework in the archaeological literature (Deetz, 1987; Hodder, 1987; apartment of Anthropology, 1126 East 59th, University of Chicago, Chicago, Ilinois 60637. 215
O 1997 Plenum Publishing Corporation 1072-53?9/97,O9(XM)215$lL5aO

of archaeological of the Annaliste

interest

in history of French

216

Morrison

and Lycett

Lightfoot, 1995;Little, 1994; Schmidt and Patterson, 1995; Stahl, 1993;Tay

in order lor, 1948, pp. 25-44; Trigger, 1989; Young, 1987). Nevertheless, to be with these for archaeological under fruitful, history engagement and to critical must be evalu subjected unpacked, problematized, standings
ation.

to some aspects of how archaeologists limit our discussion make inferences about the past using observations made on objects: texts of texts from as well as artifacts. In this paper we discuss one assemblage India and consider both how these texts have been studied and southern of the past. how that history of research has structured our understanding Here we

We consider how these textsmight be interpreted differently (1) under dif

ferent tion combined of sampling and recovery conditions data. with archaeological and (2) as sources of informa

ARCHAEOLOGYAND HISTORY:METHOD AND THE PROCESSUAL/POSTPROCESSUAL DIVIDE

sources and their potential scope and limitations of documentary data are among many with archaeological or complementarity congruence use of texts. Often, written history issues in the archaeological important to archaeological relative a data, even when position occupies privileged textual accounts. This priority of docu the latter contradict or complement the material ments can be the product of archaeological systematics when or "known" historical to illustrate is simply employed record patterns or events. Of course, historical "knowing" is by no means straight simple literature on forward, as many scholars have pointed out. To the existing and as 1991; textual such topics (Galloway, interpretation multivocality bases on of the methodological a we consideration add 1995), Lightfoot, are drawn, and of archaeological inferences engagement which historical are the come to be history? What do documents with that process. How for available made and are texts in which contexts deposited, produced, a record as are their limitations What scholarly scrutiny? and, importantly, These are not obscure historiographie issues, but are of past experience? ma use of to the that documentary archaeological speak directly questions as evidence observations historic of inclusion for to the criteria terials and in archaeological arguments. observations of archaeological In a recent article on the constitution the contentious that notes as evidence, Wylie history of the despite (1992) a rather substantial exists there debate, actually processual/postprocessual is that common a common ground between largely methodo them, ground The

con Wylie begins (1992, pp. 271-272) by pointing out the inherent logical.

Inscriptions

as Artifacts

217

tradiction of some forms of anti-/postprocessualism that embrace a radical relativism of knowledge claims while at the same time denying that "just view the If to of is the latter is be any" sustained, past position acceptable. this would seem to demand that some account be given of "how, exactly, are to judge the relative credibility of evidential, as well as archaeologists of and explanatory claims" interpretive adds (1992, p. 273; final italics ours),
This

(Wylie,

1992, pp.

271-272).

She

a nuanced account of how archaeological will data?facts of the require as evidence, how they are "laden" with record?are constituted theory such that they can have a critical bearing on claims about the cultural past and can, in turn, sustain and Tilley and contingent what call a 'particular Shanks objectivity' or not this is properly termed a theory of 'testing,' or constitutes (1989:43). Whether an analysis of 'scientific' seems to me to be a semantic (or 'systematic') enquiry, a genuine It is here that I see convergence between the interests quibble, irrelevancy. and the antilpost-processualists. of processualists

Our goals in this paper are limited and, in the spirit of Wylie's iden We discuss how his tification of common ground, primarily methodological. torical knowledge, is influenced like archaeological knowledge, by the of data recovery, and we suggest that our disciplinary parameters experi ence of recovering, analyzing, and evaluating large data sets gives us a basis re for recovering, analyzing, and evaluating historical data. Documentary like their archaeological neither encode self-evident counterparts, the that about nor, past individually, wholly encompass past. Ar meaning on made in material culture consider observations routinely chaeologists cords, contexts?cultural, analytical?and freely light of multiple depositional, move back and forth between these in evaluating archaeological arguments. data must be similarly evaluated, both in terms of internal criteria Historical and perspective such as coherence, authenticity, credibility, (Galloway, 1991; Schiffer, 1990; Vansina, 1990, pp. 88-92) 1970, pp. 167-170; Wood, and as bodies of observations that are differentially collected, preserved, and analyzed. Such source-side evaluation is, certainly, a routine (if rarely discussed) part of the disciplinary practice of historians, although a number criticism for archaeologists of recent primers on source-side (Galloway, remain the point that archaeologists sometimes 1991; Wood, 1990) make to in innocent of this practice. We intend to expand this point, however, clude as source-side criticism not only such internal issues as author inten tion but
recovery.

also more

explicit

contextual

concerns

related

to sampling

and

Following discussions by Stahl (1993, pp. 245-250) and Lightfoot

criticism to the extension of source-side (1995, pp. 204-206), we advocate all forms of actualistic data used in archaeological analysis (cf. Haekel, as in data may observed Just 1970; Wylie, archaeological patterns 1985). be influenced by both "underlying" features of those data and parameters

218

Morrison

and

Lycett

in the literature on reflected of data recovery and analysis?a relationship screen size (e.g., Shaffer and Sanchez, issues such as 1994) and the use of and for statistics describing (e.g., Ly comparing assemblages appropriate are patterns structured in textual data multiply man, (Benne t, 1994)?so

1984). EMPIRE AND INSCRIPTIONAL THE VIJAYANAGARA DATA

we turn to a corpus of written texts: stone in By way of illustration its and environs. This of the Precolonial from Vijayanagara city scriptions a was the the of territorially expansive polity eponymous empire, city capital over much India between of southern the four that claimed hegemony A.D. centuries and teenth the sixteenth 1966; Stein, Sastri, (Nilakanta to which the imperial center actually realized material 1989). The degree control ex gain from the empire, and the nature and degree of political are and of contentious active ercised by imperial elites, however, topics debate among historians (Champakalakshmi, 1981; Palat, 1987; Stein, 1980, structure and control (and their spa regarding imperial 1995). Arguments of a large and varied tial and temporal variability) hinge on interpretations are the most stone which of of textual important. data, inscriptions corpus frontier of was situated near the northern The city of Vijayanagara India. One of the largest the empire in the semiarid interior of peninsular a was of monumental locus architecture, cities in South Asia, Vijayanagara and archaeological architectural both sacred and secular. The impressive remains of the city have been a focus of research for nearly a hundred the city, 1995). Outside 1985; for a review see Morrison, years (Michell, has conducted the Vijayanagara metropolitan regional survey survey project and test excavations 1995; Sinopoli and Morrison, 1995) in the (Morrison, the of record historical The the capital. Vijayana surrounding countryside 1973; Krishnaswami Ayy 1985a, b, 1990; Filliozat, gara period (e.g., Gopal, Sastri and Venkataramanayya, 1946) is neither angar 1919; Nilakanta nor narrative account but is, instead, a spatially and primitive ethnography transac the social and economic recording temporally variable database to in In addition interests. often competing tions of multiple, inscriptions are also we focus here, there inscriptions, gener portable stone, on which texts include literary, religious, and political ally on copper plates. Other treatises and accounts of foreign visitors. to a single form of textual evidence We restrict our discussion (stone in order inscriptions) facilitate quantitative to maximize treatment and between observations comparability concerns are there As always, significant

Inscriptions

as Artifacts

219

about chronology, Stone

authorship, and agenda in the analysis of these materials. are the most ubiquitous class of text, a situation no doubt inscriptions related both to the durability of the granite on which they are carved and an is The latter to their demonstrative aspect. important point, for stone

visible to literate and illit inscriptions were in a sense public documents, as of intent (Karashima, notices observers erate alike, serving 1996). public on large boulders, are carved also generally Stone inscriptions immobile, and of public structures, built into reservoir embankments, in basements are In carried off and defaced. so on. They cannot be conveniently rarely Sastri, 1955)] contrast, copper plate inscriptions [often forged (Nilakanta held by individuals or families and cut were private, portable documents than stone. into a (recyclable) matrix more expensive in 1871, epi the Archaeological When Survey of India was founded as a major branch of the survey (Chakrabarti, 1988), graphy was constituted and it remains an important area of historical Considering specialization. that there are approxi all time periods, Karashima (1996, p. 2) calculates in the Kannada Tamil in the 17,000 30,000 language, mately inscriptions Indian South the in and 10,000 languages), Telugu language (all language, while in Sanskrit and other north Indian lan the number of inscriptions at about he estimates Thus, South India, with its many 23,000. guages only a center of epigraphical has been and research, scripts, languages, dialects, and pub with more than a hundred years of work in collecting, preserving, lishing

inscriptions. (1996, Inscriptions vary a great deal in form and content. Karashima name of p. 2) notes, "Almost all inscriptions have in their initial part the the eulogy to also with the ruling king with his regnal year, sometimes have the Saka the king, and Vijayanagar year [a South Asian inscriptions name the regnal year." As and the with calendric king's system] together ismistaken we note below, even a careful scholar such as Karashima here; mention of kings are and omit short any actually very many inscriptions to or dates. These inscriptions have not, however, been of great interest as or discussed economic historians history and, reconstructing political one of temple pre below, are found in contexts other than the normative
cincts.

literature that all assume from the historical one might are in found in fact are with associated inscriptions temples, inscriptions a number of contexts. Temple precincts do contain a great many inscrip historical analyses of this pe tions; indeed, some of the most considered on studies of single riod are based containing complexes temple of inscriptions thousands Heitzman, 1987; 1976; (e.g., Breckenridge, shrines also Srinivasan and Reiniche, 1990a, b; Stein, 1980). More modest either on slabs set up in or near the temple, on may bear inscriptions, Although

220

Morrison

and Lycett

the various walls, columned of halls, paving stones, and other structures the temple complex itself, or on nearby natural features such as boulders. are demarcated and field boundaries sometimes by boundary Village or stones without are with inscriptions. 1978), Inscriptions (Kotraiah, on wells, on slabs or boulders within villages, roads, in alongside as even remote such in seemingly locations rock outcrops. In fields, and as "secular" architecture be with associated such walls, may scriptions found fortifications, are associated and stairways. with a significant number of inscriptions Finally, canals and reservoirs features, particularly agricultural

in context, also vary greatly in inscriptions one or are two "label of what from called inscriptions" length, usually words ge 1991) to lengthy texts containing (Patil and Balasubramanya, or historical accounts. It should be no surprise that the latter nealogical of than the former. Texts also vary in the skill are poorly of their language. There the elegance as as pol well and with executed errors, grammatical spelling inscriptions, could be said along these lines, but it may ished and erudite ones. More be enough to note that lithic inscriptions vary a great deal in form, context, and visibility. a large proportion refer to grants Inscription content also varies, but institutions. Dona or donations, to temples or other religious often made tions may take the form of outright gifts of cash or livestock (more common have received attention their execution and more

1995). (Morrison, to variation In addition

of rights to produce, in earlier periods) or alienations land, goods, or privi as of in that be also investments, may thought leges.2 Many of these gifts a for the do direct material created the transactions sometimes advantage nor as well as having religious and political 1978; (Appadurai, implications revenue ar kinds of various record also Breckenridge, 1985). Inscriptions for include the granting of tax exemptions A few examples rangements. of irrigation fa land or the construction the clearing of new agricultural or individuals; and payments for cilities; land sales by villages, assemblies, as construc or for specific tasks such of irrigation works the maintenance a donative aspect, and tion. Often these revenue-related inscriptions involve na in light of their demonstrative it is thus quite reasonable, particularly
from agricultural a king might 2For example, (tax revenue) assign his "share" of produce have been remitted that would to a temple. That lands in a specific village is, the produce and land might be purchased to the temple. to the king instead went outright Alternately, in some cases, of rights can be thought of a tax exemption This alienation then donated. record number of Vijayanagara not always being reassigned?a with the benefit inscriptions term of the funds. The for barbers, for example, with no further movement tax exemptions a transfer of property to mean is used thus in the legal sense alienation (cf. Breckenridge,

1985; Stein, 1980).

Inscriptions

as Artifacts

221

ture, to follow Appadurai (1978, 1981) and others in glossing these trans actions as prestations3 or simply gifts. Donors, along with their titles and father's names, are virtually always in donative inscriptions. Donors include not only kings and roy mentioned groups, alty, but also local elites of various types, royal officers, merchant groups of villagers (apparently dominant castes), temple offi agricultural cials or priests, and even individuals with no apparent political or religious and Lycett, 1994). While all these office (Karashima, 1984, 1992; Morrison individuals and groups must be thought of as elites, they constitute quite to a range of interests and levels of authority, with the record reflecting some extent their struggles for resources and power (Morrison and Lycett,

1994).
One feature of this historical record units of analysis. Individual inscriptions of text written at particular times for particular purposes. Thus, these texts and other artifacts is possible. comparison between it possible are also highly conventional and repetitive, making is that it is easily partible into as blocks clearly are differentiated quantitative Inscriptions to uniformly it is relatively

extract from them certain basic kinds of information. Further, historical record that narrative history is easy to see in this kind of mosaic out of the constructed building blocks of these small texts. Archae literally this type of historical world also encounter of the other in ologists parts texts to we mean in general and to to this discussion extend but record, are no texts less constructed all of historical that understandings suggest and no less subject to the kinds of data recovery and analysis issues we in dealing with inscrip describe here. Many of the problems we encounter to census data or probate in tions could apply equally well, for example, dilemmas ventories. Texts do not represent a way out of the methodological and landscapes, Like artifacts, structures, of archaeology. they are prob lematic rather than self-evident sources of information.

Sourcing The

the Sources:

Context are based

and Vijayanagara

Inscriptions

on analysis of a coded database of 1610 Vijayanagara-period inscriptions from the northern part of the Vijay and see Morrison, more information 1995; Morrison anagara empire (for all the includes Our inscrip Vijayanagara-period Lycett, 1994). compilation and unpublished, tions we know of in this area, published except for a few following sections
^e as introduced term prestation, by Mauss well developed but is particularly is, of course, (1967) in the theoretical well known on in economic South Asian

ritual economy and politics (e.g., Dumond, 1970; Parry, 1986;Raheja, 1988).

anthropology

literature

Morrison

and Lycett

Table

Inscriptions,with theMean Date (AD) for Each Attribute0

Number of Mean date inscriptions Language (A.D.)

L Languages

and Scripts

in the Database

of Vijayanagara

Sanskrit Kannada T?lugu Tamil Kannada & Sanskrit Kannada & T?lugu T?lugu & Sanskrit Persian Script Nagari Kannada T?lugu Tamil Tamil & Grantha Grantha, Tamil, &. Kannada Grantha Kannada & Nagari Persian Unknown

106 1290 87 49 63

1483 1495 1563 1380 1470 5 1548 7 1478 3 1647 1474 1496 1560 1 1542 1390 2 1345 1386 1 1417 3 1647 4 1456

146 1307 94 35 17

aNote

Total_1610_1494 in its own that a language need not be written omit undated calculations inscriptions.

script. Mean

date

our newly located ones that are still under study. Like all artifact samples, an How unknown is drawn from database (and unknowable) population. ever, just as we study ceramic distributions without knowing the total num at a site or in a region, we proceed ber of vessels or sherds ever produced stone inscriptions. with what is probably a very good sample of Vijayanagara the innovative work of a number At this point we must acknowledge of historians of South India who have treated inscriptional data in a quan and Shanmugam, 1988, 1989; Tal 1987; Karashima (Heitzman, data of using free from the practice inscriptional 1994), breaking anecdotally. As one might imagine with such a large and diverse body of using a single inscription. data, it is easy to support almost any position More thorough analysis of the body of texts, however, forces one to come structure (Talbot 1994), to terms with variability in many areas?political revenue and economies arrangements 1987), (Heitzman, temple focus on the Our for and 1988, example. 1989), Shanmugam, (Karashima recovery contexts (cf. Schiffer et al, 1978) that structure the quantitative titative way bot,

patterns in inscriptionsdescribed by these historians builds upon theirwork.

Inscriptions

as Artifacts

223

Observations. We can discern a number of interesting Assemblage-Level level. For example, the intensity patterns in this database at an assemblage of inscriptional activity, reflected by the temporal distribution of dated in a pattern that varies in is bimodal for the region as a whole, scriptions, not in but form within territorial smaller units (Morrison, 1995). amplitude This with and other activity associated pattern reflects the intensive economic the founding and initial rapid growth of the city in the fourteenth century, and a dramatic century, a period of relative stasis in the fifteenth of the and of intensification agricultural and other pro expansion empire duction

in the sixteenth century (Morrison, 1995, 1996). in the database. There are five languages and six scripts represented a shows both Their distribution trend, with Telugu outpacing temporal the present-day official Tamil and (by the end of the period) even Kannada, analyses region (Table I). Thus, monolingual language of the Vijayanagara of these texts would create a marked temporal skew in the data quite apart in the assem from the strong underlying pattern of temporal distribution em seem to been have different Further, differentially blage. languages as or in markers. in their role Inscriptions ployed quasi-public public are less likely to be associated with secure dates (10% are un Kannada as a whole than those in the database (where 6% are undated); dated) shorter, as we discuss below, and may inscriptions are frequently in languages less com local salience than longer notices have had more or communities. circles elite outside certain understood specific monly If we consider the subjects of the inscriptions (gifts) in the database undated of as a whole, is, of rights to specified proportions gifts of villages?that common most the from (Stein, 1980)?are villages produce particular = as nonagricultural 529), followed by what are lumped here (32.8%; N or "other" gifts (30.6%; N = 492); construction of temples or other build for perpetual lamps or religious offerings, tax remissions, ings, endowments so on. kinds of land grants (20.5%; JV= 330) of and Various office, grants in this record, trailed by grants here labeled agri also figure prominently cultural (4.4%; N = 70). The latter are grants specifically relating to the as canals and facilities such and maintenance of agricultural construction for this of reservoirs. These patterns establish broad parameters expectation re and historical record, but the database also shows significant temporal

gional variability (Morrison, 1995,Morrison and Lycett, 1994).

This variability points to the importance of inscriptional context. Con text can include the following dimensions: (1) temporal context, (2) regional context has obvious context, and (3) locational context. Temporal impor and change. The use of both language tance for issues of comparability and script, as noted above, shows temporal variation, and like all diachronic into units such as years, "time" may be differently partitioned analyses,

224

Morrison

and Lycett

dynasties, periods, and so on. Such partitions may reflect units thought to but also may reflect analytical necessity driven have empirical significance, for consideration of, sample size (Morrison and Lycett, 1994). example, by a large scale. This we mean context, spatial variation on By regional dimension of inscriptional variation has, surprisingly, been largely neglected of linguistic specialization. As history, perhaps because in South Asia number of the inscriptions largest (1996) notes, located in the Tamil country is in Tamil and these are mostly view of the the historiographie of the far south. For this reason, perhaps, Our Tamil-centric. is area, however, study remarkably empire Vijayanagara core region of the empire lies in northern Karnataka?the I, (cf. Table in the Vijayanagara the paucity of Tamil inscriptions region)? showing the Tamil country was an outlying, albeit important province of the while in South Indian Karashima as a whole in temporal and regional context Here and Lycett, 1995; Morrison 1994). (Morrison, previous publications we explore some aspects of locational context on both the content and the like of an inscription, of texts. The actual physical position interpretation the inten the intended audience, indicates and content, language, script, of the in and the ritual position resources, tions, the ability to mobilize to of land notations available were not walls all, scribers. Temple necessarily outlined a than those salience different on have the made boundaries spot and from the land in question, of kilometers in an inscription hundreds a satisfaction" the words of a devotee who had "achieved holy by visiting that place even when he or she had to could continue upon "gaze" place gone home. have discussed aspects of both empire. We

Taking Like

a Closer

Look: Does

Inscription

Context

Structure

Content?

database is subject to this inscriptional all artifact assemblages, bias imposed by the conditions of its recovery. We illustrate this point with two examples, one that relates to what can be called recovery context (see In the first Sullivan, 1978; Schiffer, 1987) and the other to field technique. in the of the empire, while northern we the use from data part example, and its immediate second we restrict the sample to the city of Vijayanagara hinterland District). by Bellary (represented survey and excavation may be One axiom of systematic archaeological survey strate that you cannot find what you are not looking for. Systematic as where they are, things are not as well gies allow us to consider where the context of an a truism that applies in the case of texts as well. Does itmatter of the inscriptional record? Does inscription influence the content

Inscriptions

as Artifacts

225

400 300 200 100 0 UNKNOWN ROYAL ELITEOFFICER GROUP OTHER UNKNOWN ROYAL ELITEOFFICER GROUP OTHER Temples Structures

ROYAL EUTE OFFICER UNKNOWN GROUP OTHER UNKNOWN ROYAL ELITE OFFICER GROUP OTHER Agricultural Villages Fig. 1. Donor category, controlled for context empire. of inscription, northern part of the

where

lead us to assume look? Acquaintance with Indian history might as are noted but with that inscriptions associated above, inscrip temples, focus tions also occur in a number of other contexts. Has the predominant on temple inscriptions of Vijayanagara slanted perceptions of historians In the following comparisons we employ a simplified typology prestation? of the most common inscriptional contexts: temples, other structures (non temple structures are found throughout the region and include a wide range of formal and informal architecture), agricultural features, and villages. In the category of village we include inscriptions found in the fields of par location was not contexts are those for which ticular villages. Unknown
recorded.

we

only the general category of donor me Categories. Considering in an inscription, a division by locational context (Fig. 1) shows = = = 1223.0, df 15, p < 0.000, Cramer's V significant (%2 to venues donors chose in which distinctions between particular 0.591) are were domi able to their and/or inscriptions. Temple inscriptions place sev nated by royal and "other" donors, a rather broad category comprising rare status. a number of unknown individuals eral donor types and of large is in terms of the proportional Another way to look at this relationship Donor morialized sharp and

226

Morrison

and Lycett

Table IL Crosstabulation of Donor Type (Columns) by Locational

Association (Rows), Northern Part Count Expected Standard, value resid. of the Empire

Agricultural
Features

Other Structures 12 38.4 -4.3 3 4.8 -.8 11 3.4 4.2 23 20.0 .7 30 18.3 2.7 10 4.1 2.9

Temples
39 150.7 -9.1 3 18.8 -3.6 26 13.2 3.5 62 78.4 -1.8

Villages
9 81.2 -8.0 1 10.1 -2.9 2 7.1 -1.9

Elite donor

444
233.6 13.8 56 29.2 5.0

Group

Royal

officer

5 20.4 -3.4

Other

donor

16 121.5 -9.6

161
42.2 18.3 7 38.7 -5.1 8 8.7 -.2

Royal
Unknown

13 111.3 -9.3 7 25 -3.6

190
71.8 14.0 29 16.1 3.2

standardized of donor types (Table II). Here, residuals representation (cf. and Lycett, 1994) make clear the proportional domination Morrison of both associated with royals and royal officers in temple inscriptions. Inscriptions to a that similar of structures follow pattern temples. nontemple in village contexts and on agricultural facilities Donor distributions the oveiwhelming show quite a different distribution. Local elites constitute and their deeds on canals, reservoirs, of donors who recorded majority as the is also proportional, this dominance features; agricultural standardized residuals in Table II indicate. Interestingly, inscriptions asso other ciated with villages and village lands are dominated donors, sug by as nayakas (local individuals not readily identifiable gesting that nonroyal recorded their texts in these village con elites) or as officers preferentially other texts. One also wonders if such donors were not able to place their inscrip refers to groups of merchants tions in temples. "Group" in this classification or village assemblies, in this corpus neither of which have a large presence of texts. the actual topic of the grant Grant Categories. If we further consider = = or gift (Fig. 2), significant differences 12, p < 0.0000, 74.46, df (x2 Cramer's V = 0.146) by context are apparent. Temples disproportionalry refer to grants of villages and to "other" gifts, particularly those that relate

Inscriptions

as Artifacts

227

UNKNOWN VILLAGE LAND AGRICULTOTHER UNKNOWN VILLAGE LAND AGRfCULTOTHER Temples Structures 250

UNKNOWN VILLAGE LANO AGRICULTOTHER UNKNOWN VILLAGE LAND AGRICULTOTHER Villages Agricultural Fig. 2. Gift category, controlled for context empire. of inscription, northern part of the

either ments

to the construction of the temple or to its functioning (e.g., endow or gifts for flowers, oil lamps, textiles, and other offerings). Propor high numbers tionally (Table III), temple inscriptions contain unexpectedly of gifts of villages [a common royal grant (Morrison and Lycett, 1994)] and on nontemple unknown gifts (these often are unspecified). Inscriptions structures such as stairways and mandapas (columned halls) do not follow the temple pattern as closely as they did for donor category. Grants of the association between villages are replaced by other gifts, highlighting one donation. and of Further, only nontemple royal villages, gifts temples, it clear that facility, making inscription refers to an agricultural a concerns. on set on limited structures of focus nontemple inscriptions those in villages and found features into bu?t agricultural Inscriptions concerns from those carved reflect different and fields, not surprisingly, to land and its produce into temples and other buildings. Here references other the record. dominate gifts are mentioned Although (nonagricultural) in inscriptions carved into agricultural facilities reasonably often, they occur less often (Table III) than might be expected given their overall number to the construction of canals and in the sample. There are more references are in the than there reservoirs on those features themselves numerically structural

228

Morrison

and Lycett

Table HL Crosstabulation of Gi? Type (Columns) by Locational Association (Rows), Northern Part of the Empire
Count Expected Standard Agricultural Features Agricultural value resid

Other
Structures 1 4.0 -1.5 14 19.6 -1.3 50 28.2 4.1 12 6.8 2.0 12 30.4 -3.3

Temples
11 15.6 -1.2 55 76.9 -2.5

Villages
10 8.4 .6 58 41.4 2.6 66 59.6 .8 10 14.3 -1.1 44 64.3 -2.5

30
24.1 1.2

gift
Land

130
119.1 1.0

Other

144
171,5 -2.1

110
110.7 -.1 34 26.6 1.4

Unknown

33 41.3 -1.3

Village

204
185.0 1.4

139 119.3
1.8

on temple inscriptions would larger temple sample. Thus, an exclusive focus a to suggest temples and a reduced role for land greater role for donations the context of an inscription and irrigation. Clearly, transactions strongly structures its likely content and the contexts of the sample of inscriptions of the past. we study will strongly influence our interpretations

Disciplinary

Traditions

of Analysis:

Field Technique

to interpretations of If, then, the context of the inscription matters are and recorded. found we how ask content, inscriptions legitimately might the sample of recorded inscriptions actually reflect the varied contexts Does in which inscriptions occur? Are some inscriptions more obtrusive than oth re ers (cf. Schiffer et al., 1978, p. 6)? Is there a size bias to inscriptional field called are be what issues of technique [discussed might covery? These as recovery theory by Sullivan (1978) and Schiffer et al. (1978)]. We dem do onstrate below that the association, length, and subject of inscriptions vary with fieldwork strategy. and its imme Here we narrow our sample to the city of Vijayanagara diate the Vijayanagara metropolitan Although (Bellary District). regional survey over a large por survey project has carried out systematic the number of inscriptions we have recorded tion of the city's hinterland, vicinity

Inscriptions

as Artifacts

229

TEMPLES STRUCTRS AGRICULT VILLAGES

TEMPLES STRUCTRS AGRICULT VILLAGES DAMOther

Fig.

3. Context

for recorder of inscriptions, controlled (DAM, Directorate and Museums, Bellary District. chaeology Karnataka),

of Ar

in and we use instead data gathered is too small to analyze convincingly and focused search for the course of a less systematic but more concerted and Mu State Directorate of Archaeology inscriptions by the Karnataka seums (DAM) Karnataka and Patil Balasubramanya, 1991). (Patil, 1991; the year, took the trouble resident in the area throughout archaeologists, and explore areas away to climb outcrops, examine all sides of boulders, in mapping from roads in order to find inscriptions. They also engaged at a time. This re and excavation that kept them in the field for months than the sur search took place in an area having many more inscriptions to size familiar problems. veyed hinterland, pointing sample interest in the city of Vijay As noted, archaeological and historical anagara has a long history, and we can consider first a sample of inscrip over the course of different researchers tions gathered by many are a In this hundred years. sample (Fig. 3, right), temples approximately the most common inscriptional locus, followed by agricultural features, vil lages, and other structures. This pattern is quite similar to that of the larger region, and it reflects much the same research agenda and recording strate gies. If inscriptions area by Karnataka brought to light by the recent thorough scouring of the are considered State (DAM) archaeologists apart from

230

Morrison

and Lycett

in locational recorded inscriptions, some interesting differences previously recorded far context emerge (Fig. 3, left). First, Karnataka archaeologists their and than did fewer predecessors (Patil inscriptions temple the of because many 1991, p. larger partly temple Balasubramanya, 19), recorded. had been the in however, Significantly, city already inscriptions of inscriptions also recovered a large proportion researchers the Karnataka and villages. The differences features, the close knowl in field strategy?both area their and to have of their study researchers willingness edge DAM record and study shorter inscriptions. Many of the inscriptions recorded by neither are label inscriptions and quite a few mention DAM archaeologists are recorded DAM nor the of In 64% dates. fact, by inscriptions kings lack others dates. of those 6% while However, undated, only gathered by same time period as the DAM sample does seem to be drawn from the no are there is dif when undated the larger sample; excluded, inscriptions the two samples of ference (t = 0.78, p = 0.436) in time period between from other in context structures, agricultural partly reflect differences
texts.

no include repetitive devotional inscriptions to sacred locales. Others, tices memorializing interestingly, pilgrimages in and near the city as being "watch in hills the rocky identify specific spots such as stairways donations small-scale (Patil and towers," or record are short inscriptions strikingly different 1991). These Balasubramanya, con the discussion above, the locational from longer ones, and following texts in the DAM types of gifts/topics sample lead us to expect different of concern as well as a different set of donors. We can explore the impact the of field technique on inscription content quite directly by comparing a versus other DAM in sig inscriptions, types of prestations memorialized = = = 4, p < 0.0000, Cramer's V nificant difference 0.316) 28.58, df (x2 or by comparing donor categories by recorder, again a significant difference = = 31.89, df = 4, 0.334). The "grain" of the p < 0.0000, Cramer's V (X2 in the structure fieldwork, then, seems to make a great deal of difference the of The to it data past we can put of the historical picture light. brings on longer inscriptions associated with temples?a a focus from together officers donate villages to temples?differs picture in which kings and royal less well-dated in more from a sample that includes shorter, fine-grained and individual local donors, organ In the latter sample, e?tes, scriptions. view of royal, official, ized groups seem more important. This differential and local elite activity should have a direct bearing on ongoing historical debates about the nature of Vijayanagara authority and control, political of elites fields of action for different the differential categories showing Clusters of these shorter

(Kulke, 1995; Sinopoli andMorrison, 1995).

Inscriptions

as Artifacts

231 Traditions of Analysis: Integrating and Archaeology Technique, Context, Field

Disciplinary

have

a few aspects of inscriptional context and have briefly considered in field technique can yield different kinds of shown how differences samples. Further, we have shown that locational context has implications We note, have to do with focused on Research structured our views (Morrison and Lycett, empire as whole. But

for inscriptional subject matter. These differences, we the goals, identities, and abilities of the inscribers. temple contexts, and the Tamil region, has powerfully of Vijayanagara history. We have already discussed how regional bias can affect our view of the 1994)

what happens when we seriously consider a different kind of sample, such as the DAM control is reduced, since many label corpus? Chronological on can be dated only (approximately) grounds. palaeographic inscriptions our view of textually recorded concerns also changes. We see, However, for example, more individual or perhaps personal memorials (I looked upon labels (this the sacred hill and achieved satisfaction), more military/security land use is the watchtower and more detail on agricultural of Hanuman), a on also in sometimes of recorded land appear temple grants (duplicates "on the ground" links be the landscape, providing a means of establishing tween places noted in texts and the material record). The activities of kings and their officers are less visible, while the actions of groups and individuals come into sharper focus. the important topic of archae Finally, we must bring to this discussion data. We have tried to bring with historical and data its integration ological of common ground, some in the spirit of Wylie's (1991) identification data that, we believe, of concerns textual in the analysis key methodological historians considered have been insufficiently by (who do not usually go criticism does source-side out and locate their own inscriptions). Historical contextual the not routinely include kind of rigorous analysis that archae We still have not, however, studies. to artifactual typically apply ologists of data into this discussion and, given constraints brought archaeological one more brief than example. space, cannot provide of and maintenance in the construction interested Archaeologists out, monumental architecture (Morrison and Lycett, 1994) or irrigation facilities

(Morrison, 1994, 1995) long have been concerned with establishing the so
such features. Inscrip cial role(s) and resources of those who commission social scale equates tional data can help us address the question of whether to facility scale in the construction of irrigation features (e.g., Sewell, 1982, of labor p. 162). That is, are larger facilities or facilities more demanding more research or financed by "larger" people? Archaeological powerful that agricultural facilities exhibit a much greater range shows immediately

232 of forms than the historical features

Morrison

and Lycett

record suggests (Morrison, 1995). Not only are such as terraces, gravel-mulched fields, check-dams, in inscriptions, but the majority of even large ag and wells not recorded in texts (Morrison, is not memorialized ricultural facilities 1997). Among recorded facilities, we can employ our simplified donor typology as a very of social "scale," with kings and royals at the top, royal rough measure smaller-scale and all others in the third position. and local elites in the middle, Is it the case that kings facilitate the building of canals, that nayakas (local and that farmers finance wells and terraces? subsidize reservoirs, elites) of the database provides qualified but support for this position Analysis also reveals a greater degree of diversity in strategies than this simple scalar officers suggest. in this analysis; thus our results sizes present some difficulties Sample references 10 the Of should be considered explicitly describing suggestive. of canals, kings and their officers donated 70% while con the construction stituting only about 30% of all donors. Still, three canals were said to be donated by people other than kings or their officers (one by a local elite, con follows a similar pattern. Reservoir two by others). Canal maintenance struction often was financed by local elites (48% of all reservoirs; N = 25). a king claims to have built one reservoir in this sample and both However, = = in officers 10) also have a significant presence 11) and others (N (N local tradition reservoir construction (and (40% of all reservoirs). Although see Sewell, 1900) associates one very large reservoir near the city of Vijay there is no explicit textual support for this. anagara with royal patronage, Thus, the simple correlation of facility scale with donor social scale appears It can be sustained only by considering kings and only roughly accurate. their officers as a group, ignoring their sometimes competitive positions; in the way that officers endowed this situation also is reflected reservoirs, archae and Lycett, 1994). Combining acting as local elites (cf. Morrison can point to the chasms in our understanding data historical and ological that result from reliance on either data source alone. Critical opposition of these data sources also can help us guard against potentially misguided in this case the (simplistic) conventional systematics from a single discipline, in the archae that social scale (or rank) is directly reflected interpretation scale. record through structures of varying ological correlation would

CONCLUSION: VIJAYANAGARAINSCRIPTIONS AS HISTORY

As India texts of Precolonial South the inscriptional a body of narrative(s), In of the nor are neither large spite complete. internally coherent

Inscriptions

as Artifacts

233

number of inscriptions published, many more are known only to a handful of scholars or lie unrecorded. Published inscriptions are not always com or translated with Even these limitations, transcribed. pletely however, there are many recorded inscriptions, creating a rich documentary record. more as reveal some about structure such the of Inscriptions topics, temple than they do about others, and in some cases whole groups of economies, people are virtually invisible in these texts (Morrison, 1996). In this, Vijay anagara inscriptions differ little, if at all, from most other documentary records. Further, information from inscriptions always is filtered through cultural categories, expectations, and arguments, reflecting a multiplicity of and others are all important for evalu These considerations perspectives. ating and using texts as sources of information about the past. However, it is no less important to consider the sources of the sources, or the bases of knowledge from which historical inferences are drawn. For many topics of study, such as political organization, the widely divergent in historical current the to in literature the in terpretations point difficulty transforming a unitary inscriptional data into narrative history (Kulke, 1995). Neither master from nor a proliferation of alternative voices arises naturally of these texts. Instead, they enter the construction of historical arguments only within the context of specific research programs, as well as theoretical on methodological programs crucially dependant an we tools. Thus, while do not suggest that past lies beneath unproblematic narrative examination and distributional the rhetorical conventions complexity of the inscriptional the contexts record (cf. Galloway, 1991, p. 467), it is important to consider as a record in which these texts were produced as well as their limitations of past experience. The contention that conditions of data recovery and analysis have im and archaeological for the content of historical interpretation plications we also seems this limited In controversial. however, making point, hardly are suggesting to archaeologists that we employ the broader analytical ex to the recovery, analysis, and interpretation of of our discipline perience self-evi textual material. Like other cultural objects, texts do not produce and cannot, in themselves, solve archaeology's central meth dent meaning and and historical odological challenges. opposing Integrating allows for particularly powerful analyses of past knowledge archaeological events and processes. As a discipline archaeologists have a role to play in as history the process of how texts, like artifacts, come to be constituted in Just as the justification of archaeological and, indeed, as archaeology. a concern that spans theoretical ferences demands a concern for method, so the use of documentary sources calls us to a fissures in the discipline, awareness shared methodological that both builds on and expands existing archaeological practice.

234

Morrison

and Lycett

ACKNOWLEDGMENTS
We thank the Government of India and the Archaeological Survey of to carry out the archaeological India for granting permission research de Institute scribed in this paper. This research is facilitated by the American of Indian Studies, to whom we extend our sincere thanks. We are particu and epigraphic to our archaeological the from colleagues larly indebted a and Museums and State Directorate of extend Karnataka Archaeology we and to assistance S. Patil his advice. for Dr. C. Finally, special thanks thank the reviewers of this paper.

REFERENCES CITED
Appadurai, A
(ed.), South

(1978). Kings, sects, and temples in South India, 1350-1800 AD. In Stein, B.
Indian Temples: An Analytical Reconsideration, Vikas, New Delhi, pp. 47-73.

Appadurai, A

(1981).Worship and Con?ct Under Colonial Rule, Cambridge University Press,

Cambridge. to the integration of archaeological of approach Text and context: Levels J. (1984). Bennet, Review from Cambridge and textual data in the Late Bronze Age Aegean. Archaeological 3: 63-75.

Bintliff, J. (ed.) (1991). The Annales School and Archaeology, Leicester University Press,
Leicester.

Breckenridge, C. A
Minaksi Wisconsin, 1350 Verlag,

(1976). Worship and Endowments in South India: The Case of the Sri
Temple, 1833-1922, Unpublished Ph.D. dissertation, University of

Sundaresvarar Madison.

Breckenridge, C A
to 1750.

(1985). Social storage and the extension of agriculture in South India

A. L. (ed.), Vijayanagara: City and Empire, Franz Steiner

Chakrabarti, D. K. (1988).A History of IndianArchaeology from theBeginning to 1947, Mun

shiram Manoharlal, article. Indian Delhi. and Social 18: 411-26. History Review in the prehistory of the mid Dur?e and the Longue reproduction (1991). and Postprocessual Archae In Preucel, R.W. United continental States, (ed), Processual Illinois University Press, Carbondale. the Past, Southern Ways of Knowing ologies: Multiple In Coma and the historical J. (1992). Ethnography J. and Comaroff, imagination. Comaroff, and the Historical Westview, J. (eds.), Ethnography Imagination, roff, J., and Comaroff, CO. Boulder, revisited. American Antiquity and archaeological J. (1987). History theory: Walter Taylor Deetz, 53: 13-22. An Essay on the Caste System, L (1970). Homo Hierarchicus: (trans., M. Sainsbury), Dumond, Economic Social Cobb, C. R. of Chicago University Fabian, J. (1983). Time and Press, New York. Press, Chicago. the Other How Anthropology Makes Its Object, Columbia University

In Dallapiccola, Wiesbaden, pp. 41-72.

Champakalakshmi, R. (1981). Peasant state and society in medieval South India:A review

d'Extr?me Fran?ais de debut a 1377, LEcole de Vijayanagar V (1973). Epigraphie F?liozat, Orient 91, Paris. D. H. In Thomas, narrative. of ethnohistorical R (1991). The archaeology (ed), Galloway, in Pan-American Bordedands Vol 3. The Spanish Columbian Perspective, Consequences, Institution Smithsonian DC, Press, Washington, pp. 454-470.

Inscriptions as Artifacts
B. R. (1985a). seums, Mysore. B. R. (1985b). Gopal, Gopal, seums, Mysore. B. R. (1990). Gopal, ums, Mysore. J. (1970). Haekel, 147-164. J. (1987). Heitzman, 791-826. Hodder, I. (1987).

235
Inscriptions, Inscriptions, Inscriptions, Vol Vol Vol 1, Directorate 2, Directorate 3, Directorate In Naroii, of Archaeology of Archaeology of Archaeology R., and Cohen, R. and Mu and Mu and Muse (eds.), A

Vijayanagara Vijayanagara Vijayanagara

Source

criticism

Handbook ofMethod in CulturalAnthropology,Columbia University Press, New York, pp.

Temple The urbanism of in medieval the South term. India. Journal In Hodder, pp. from of Asian Studies 46: as 850 Rule,

in anthropology.

contribution

long

Long-Term History, N. (1984). Karashima, Karashima, N. (1992).

Press, Cambridge, Cambridge University and Society: Studies South Indian History Toward a New Formation: South Indian

I. (ed.), Archaeology 1-8. Inscriptions, Under A.D.

1800, Oxford University Press, Delhi.

Society Vijayanagar

Oxford and IBH,Delhi. A new approach to their study. South Karashima, N. (1996). South Indian temple inscriptions:
Asia Karashima, 19: 1-12. Revealed Study Revealed the Study of as in Tamil Country Rule R (1988). Vijayanagar N., and Shanmugam, a Statistical Institute Terms in Inscriptions, for the Study of Revenue and Cultures of Asia and Africa, Tokyo. Languages as in Tamil Country Rule R (1989). Vijayanagar Karashima, N., and Shanm?gam, in Inscriptions, Part 2, Institute for Through a Statistical Study of Revenue Terms of Asia and Africa, and Cultures of Languages Tokyo. Through

Annales and Ethnohistory,Cambridge University Press, Knapp,A B. (ed.) (1992).Archaeology,

Cambridge. of Madras, S. (ed.) (1919). Sources of Vijayanagara History, University Krishnaswami Ayyangar, Madras. India. In Kulke, H. (ed.), The study of the state in premodern Introduction: Kulke, H. (1995).

The State in India,AD 1000-1700, Oxford University Press, Delhi, pp. 1-47.
Culture archaeology. contact studies: Redefining 60: American Antiquity the relationship 199-217.

K. G. (1995). Lightfoot, toric and historical

between

prehis

and Theory 1: 5-40. Method States. Journal of Archaeological American Antiquity in zooarchaeology. units and terminology Lyman, L L (1994). Quantitative 59: 36-71. in Archaic Societies (trans., I. Mauss, M. (1967). The Gift: Forms and Functions of Exchange New York. Norton, Cunnison), A. L G. A City (ed.), Vijayanagara Micheil, city. In Daliapiccola, (1985). A never forgotten Vol 1, Franz Steiner Verlag, Wiesbaden, and Empire: New Currents of Research, pp. 196 207. Con and the Course of Intensification, K. D. (1995). Fields of Victory: Vijayanagara Morrison,

Little, B. J. (1994). People with history:An update on historical archaeology in the United

tributions of theArchaeological Research Facility 53, Berkeley, CA Morrison, K. D. (1996). Typological schemes and agricultural change: Beyond Boserup in precolonial South India. CurrentAnthropology 37: 583-608. Morrison, K. D. (1997).Agriculture at the edges: Archaeology and history in theVijayanagara hinterland. In Allchin, B. (ed.), South Asian Archaeology 1995, Oxford and IBH, Delhi
(in press). centralized Ideo K. D., and Lycett, M. T (1994). Centralized power, authority? Morrison, 32: 327-350. patterns. Asian Perspectives logical claims and archaeological of Madras Nilakanta Press, Madras. Sastri, K. A (1955) The Colas, University Nilakanta Press, Sastri, K. A University of South India, 3rd ed., Oxford (1966). A History London. M. A and Venkataramanayya, Nilakanta Sastri, K. A, (eds.) (1946). Further Sources of Vijay of Madras, Madras. anagara History, University

236 Morrison

and Lycett

OhnukiTierney, E. (ed.) (1990). Culture Through Time: AnthropologicalApproaches, Stanford University Press, Stanford CA
Palat, of the agrarian order of medieval R. (1987). The Vijayanagara empire: Re-integration R (eds.), Early State Dynamics, H. J. M., and van der Velde, India. In Claessen, South E. J. Brill, Leiden, pp. 170-186.

Parry, J. (1986). The gift, the Indian gift and the "Indian gift."Man 21(3): 453-473.

D. V, and Patil, C. S. (eds.), Vijayanagara in Devaraj, studies, Patil, C. S. (1991). Epigraphical of Archaeology and Museums, 1987-88, Directorate pp. 15 Mysore, Progress of Research 34. In Devaraj, D. V, and Patil, studies. Patil, C. S., and Balasubramanya (1991). Epigraphical of Archaeology and C. S. (eds.), Vijayanagara 1984-87, Directorate Progress of Research Museums, pp. 19-70. Mysore, in a and the Dominant in the Gift: Ritual, Prestation, Caste Raheja, G. G. (1988). The Poison of Chicago North Indian Village, University Press, Chicago. of Chicago Press, Chicago. Sahlins, M. (1985). Islands of History, University of New Processes Record, University Schiffer, M. B. (1987). Formation of the Archaeological Mexico Press, Albuquerque. as Long-Term History. American 55: of Archaeology Schiffer, M. B. (1990). Review Antiquity 423-424. surveys. R, and Klinger, T G. (1978). The design of archaeological Schiffer, M. B., Sullivan, A 10: 1-28. World Archaeology Alternative Histories: The Practice R R., and Patterson, T C. (eds.) Schmidt, (1995). Making in Non-Western of American Research School and History Press, Settings, of Archaeology Santa Fe. New Services Reprint, Asian Educational Sewell, R. (1982). A Forgotten Empire (Vijayanagar), Delhi. of of 1/8" and 1/4" mesh J. L J. (1994). Comparison B. S., and Sanchez, recovery Shaffer, 59: 525-530. American mammals. controlled Antiquity samples of small-to-medium-sized on ar rather than answers: Shanks, M., and Tilley, G reply to comments (1989). Questions Review 22: 42-54. into 1990s. Norwegian Archaeological chaeology of imperial control: The Vijayanagara K. D. (1995). Dimensions C. M., and Morrison, Sinopoli, 97 (1): 83-96. Anthropologist capital. American A Saiva Sacred Complex M. (1990a). TiruvannamalaL R R., and Reiniche Srinivasan, of South de Pondichery, Institut Francais Pondicherry. India, 1.1 Inscriptions, A Saiva Sacred Complex M. (1990b). Tiruvannamalac P. R., and Reiniche, of South Srinivasan, de Pondichery, Institut Francais Pondicherry. India, 1.2 Inscriptions, in historical to analogical of time and approaches reasoning Stahl, A. B. (1993). Concepts Stein, 58: 235-260. American Antiquity perspective. in Medieval B. (1980). Peasant State and Society Delhi. South India, Oxford University Press,

The New CambridgeHistory of India, Vol 1, Part 2, Cambridge Stein, B. (1989). Vijayanagara,
Press, Cambridge. University in In Kulke, H. (ed.), The State state: Interim reflections. B. (1995). The segmentary Oxford University Press, Delhi, pp. 134-61. India, AD 1000-1700, of conceptual A discussion in archaeology: and evidence Inference prob Sullivan A. R (1978). and Theory, Vol 1, New Method in Archaeological lems. In Schiffer, M. B. (ed.), Advances Press, pp. 183-222. York, Academic Indian Economic 1175-1325. in Kakatiya intermediaries Andhra, Taibot, C. (1994). Political and Social History Review 31(3): 261-289. As of the American Memoirs Anthropological Taylor, W W. (1948). A Study of Archaeology, DC. sociation 69, Washington, In A critical analysis. American and contemporary archaeology: Trigger, B. (1989). History in America, G G (ed), Archaeological University Cambridge Thought Lamberg-Karlovsky, Press, Cambridge, pp. 19-34. Stein,

Vansina, J. (1970). Cultures through time. In Naroll, R., and Cohen, R, (eds.),A Handbook New York, pp. 165-179. of Method in CulturalAnthropology, Columbia University Press,

Inscriptions

as Artifacts

237

Wolf, E. (1982). Europe and the People Without History, University of California Press,
Berkeley. In Schiffer, M. B. (ed.), Archaeological R. (1990). Ethnohistory and historical method. Wood, and Theory, Vol 2, University Method of Arizona Press, Tucson, pp. 81-109. in Archaeologi In Schiffer, M. (ed.), Advances Wylie, A (1985). The reaction against analogy. cal Method and Theory, Vol 8, Academic Press, New York, pp. 63-111.

Wylie, A
and

(1992). Of "heavily decomposing red herrings": Scientific method in archaeology

theory, In Embree, been L (ed.), Meta-archaeology, concept? American Kluwer Antiquity has history a valid

the ladening of evidence with Dordrecht, Academic, pp. 269-288. Young, T G, Jr. (1987). Since Herodotus, 53: 7-12.