EPW    Special Articles July 10-16, 1999

Statistical Concoctions and Everyday Lives

Queries from Gujarat Resettlement Sites

J Whitehead

?There exist procedures in the social sciences which constitute a break in social consciousness between a knowledge located in specific times and places and in the experiences of individuals and a formalised impersonal mode of knowing articula ted to (and indeed a part of) an apparatus of ruling. This disjuncture in the social consciousness is the focus of our inquiry

- Smith 1981

THE INDIAN EXPRESS of May 25, 1997, summarised the 1996-97 survey findings of the Centre for Social Studies in Surat on Gujarat-based project-affected peoples ?rehabilitated? from the Sardar Sarovar submergence zone. The headline: ?Oustees of Narma da Project Find Life Better in New Sites?. The newspaper article stated that the CSS study reported fewer deaths per year in the resettlement sites and lower infant mortality rates compared to the national average. Health care in general was deemed better in resettlement sites than in submerging villages, with all villages reported to have access to primary health care centres within a 10 km radius. Almost all major sons were deemed to have received the allotted five acres of land. Primary, secondary an d high schools were also reported to be more accessible than in the hill areas where the bhils? original settlements were located. Finally, the CSS Reports confidently asserted that the living standards of oustees were better in the resettlement sites. Th is finding was based on the statistical ?fact? that all resettlees experienced an ?increase in consumption income? following their shift into the plains of Bharuch and Vadodara districts.

Yet this rosy picture jars with my own lived experiences arising from visits to several resettlement sites and an original, submerging village during the summers of 1997-1998. I have therefore been led to question several procedures by which the objects o f knowledge, methodology, and even the ?data? itself have been worked up and produced in the CSS Reports. By ?produced? I do not mean to imply that CSS Reports purposefully and consciously distorted the ?facts?. Rather I am referring to the types of socia l scientific abstractions, practices and the philosophies supporting them which erase ?subjectivity, transforming local and particular experiences into general facts, terms, and categories by freeing them of their experiential grounding in lived social re lations? [Smith 1983]. Such procedures are not limited to CSS Monitoring and Evaluation Reports, particularly where ?development indicators? are studied. Indeed, the work through which relations between persons become independent of particular individuals and expressed as abstractions is common to many objectifying procedures found in the reasoning, information-collecting, and decision-making practices. This type of work is crucial to all formal organisations mediated by texts and computer technologies, b eing a part of the managerial process [Smith 1997].

?Facts?, after all, are also produced, and do not just exist out there. They are constructed through diverse types of social science methods which convert local, individual and concrete experiences into objects of knowledge validated, apparently, by their generalisability across a range of particular or individual experiences. Indeed, in ?high? social science theory, e g, in the ?ideal types? of Weber, these generalisations are often universalised. Statistical abstractions of the type characterising the q uantitative analyses found in CSS Reports start from an a priori base of generalising procedures to produce facts, which become the objects of knowledge. The latter are then worked up through additional mathematical procedures comparing relationships betw een groups of facts which usually make no further references to lived, concrete experiences of individuals.

The standpoint method I offer as a contrast to the qualitative one pursued in the CSS Reports, often loosely glossed as qualitative or ethno-methological, does not take these initial steps of objectification for granted. Rather it questions the objectifyi ng procedures through which the construction of ?facticity? and statistical relations are socially organised through a world of (bureaucratic) action mediated by texts [Smith 1997]. Too often, the reality that all social scientists occupy a particular soc ial or discursive space is ignored, thought to be erased through training. Such training in objectivity, however, because of its claims to universal validity, erases who the real political subjects and actors might be. If such subjects were named, the act ual particularism of their knowledge and its relation to political projects would be revealed. Hence, the standpoint reflected in ?objective? accounts often reflects unnamed relations of ruling, rather than the relations and experiential horizons of thos e studied. This problem of a priori biases is particularly acute when the subjects of research cannot represent themselves - either due to lack of literacy and training or to their powerlessness - to provide a dialogic counter-check on their representatio n. By relations of ruling, I refer to the social relations, technologies and expertise involved in the production and replication of texts, both written, and increasingly, electronic. These relations are the orgnisers and regulators of an increasingly glo balised managerial world.

The standpoint taken as a point of departure in this paper is the experienced social relations of bhils in a resettlement site in Dabhoi taluka, Vadodara district, as these relations were co-ordinated, maintained, and/or changed from one day to the next, and from one round of production to the next.1 This analysis examines the economic, social and nutritional reproduction of households both daily and through the summer agricultural cycle. It does so by examining extent of landownership, access to credit, access to and use of marketed inputs, costs of production and consumption items and total output on a daily basis, in particular focusing on those social relations required for the co-ordination of sequences of work. Here social relation s of work refers not to a fixed division of labour, but to an ongoing co-ordination of courses of action in which what people do becomes materialised [Smith 1987:185]. This qualitative household ethnography will then be compared to the objectifying proced ures in CSS Reports to reveal the contradictions and differences between the two methods.

The most obvious way in which the objectifying procedures of CSS Reports erase subjectivity and lived experience relates to the use of statistical averages across 120 different resettlement sites. These statistical averages are applied to crucial variable s such as caloric intake, yields per acre, amounts of irrigated land, and extent of agricultural labour. Such factors inevitably differ from site to site, sometimes quite drastically. Hence, a statistical average begins from an unfounded - indeed even une mpirical - assumption of similarity between sites. There is already inserted in the reports an objectifying procedure which falsely generalises categories away from local lived experiences and relations. Hence, the relationship between yields per acre, so il fertility, extent of agricultural labour, and caloric intake, are homogenised, when their inter-relationships really differ due to the influence of one or several variable factors. Almost anyone who has done more intensive, qualitative, rural fieldwork can testify to the reality that soil fertility, soil type, access to irrigation, use of commercial inputs, local weather patterns, cropping techniques, and even varieties of crops, and the inter-relationships of the above can profoundly affect the output and yield between villages in a taluka or a district.

To take a rather drastic example of the falsely homogenising effects of this form of statistical abstraction, the 1996-97 CSS Reports state that caloric intake, on average, is much higher in the resettlement sites than it was in the original villages. Cal oric intake was, on average, 2,418 in submerging villages, then rose about 100 calories on average to 2,541 in 1992-93, and fell to 2,237 calories in 1996-97.2 The report goes on to state that about 42 per cent of households in resettlem ent sites reported lower than the minimum recommended 2,000 calorie intake per adult in the rabi season, 31 per cent were lower from October to December, while 23 per cent reported consumption less than 2,000 calories per person per day from January to Ap ril 1997. However, as the tables from the CSS Reports themselves show, these figures are differentially distributed across resettlment sites, and vary according to the relative fertility of the soil, access to stable irrigation sources and markets, govern ment credit, etc. In other words, in resettlment sites with low yields due to lack of irrigation or bad land, the figures of lower caloric intake were probably much lower than the recommended minimum of 2,000 calories, and could have indicated conditions of scarcity, distress and hunger. This is where the use of statistical averages hides more than it reveals and where qualitative time-intensive fieldwork can provide a more accurate picture.

The second assumption underlying the implied and actual comparisons between resettlement sites and original villages is that ?tribal? societies are constituted by a lack of ?modernity? and ?development?, the latter of which seems to be assumed to be self- evidently normative and valuable.3 This is because monetisation and cash flow by themselves are construed as indices of ?development? and ?improvement? in these reports. This bias is built into the reports because the objectifying proced ures of the reports have not provided any way of calculating non-monetised inputs of production and consumption in the original villages, where they were quite substantial. Hence they disappear, only to reappear as a lack, an indice of underdevelopment. T hey re-enter the concluding section as an indice of lack of cash income, while market-determined monetary values constitute the unconscious ?norm? of development against which the original villages are measured.4 Hence, the substantial items of both production and consumption acquired either through monsoon-fed agriculture or the forest in the submerging villages are simply erased.

This particular form of objectifying procedure may reflect the familiar opposition between primitivity and modernity, and an unconscious bias in favour of the latter. Surely, we can acknowledge that an economy can be socially organised through relations o f subsistence, both between people and nature, without falling into the danger of ?romanticising? subsistence agriculture or ?tribal? societies. The reports seem to fall into the anthropological error of viewing ?tribal? societies as a bygone stage of s ocial evolution, placed backwards in time although the societies in question occupy the same real time frame as ourselves. Hence, in CSS Reports a dispersal in space between original villages and resettlement sites has been constructed as a sequence in ab stract time [Fabian 1983]. The widespread, if sometimes unconscious, use of the metaphors of ?underdevelopment? and ?backwardness? is evident in these reports. Not coincidentally, it was also used to justify fairly brutal processes of land alienation of t ribal peoples in the cause of capitalist modernisation and has already been well documented for Amerindian native peoples. These objectifying, temporal metaphors now seem to be applied to the bhills of the Narmada region.5

The following questions and criticisms to the objectifying procedures of these reports cannot claim to be comprehensive. Due to time constraints of university teaching and administration, the preliminary study from which I draw my questions has certain li mitations. First, it only covers one crop season, the rabi seasons of 1997-98, and the survey itself covers one resettlement site and one partially-submerged original village. However, information generated from an in-depth house-to-house survey during a bad crop season that followed a prior heavy monsoon in 1996, as well as day-long observations of other sites, can provide indicators of more permanent trends. This is because economic processes set in motion during repeated periods of crop failure, such a s distress sales, out-migration for wage labour, death and/or sale of livestock, are often irreversible. In addition, a common complaint amongst resettlees is that their landholdings are relatively unproductive - infertile or vulnerable to waterlogging, c ompared to the district and taluka averages. They believe that nearby farmers sold only their most economically marginal holdings to the SSN. The erosion of previous cropping mechanisms because of repeated crop failures are cumulative and possibly permane nt trends.

Data for this paper was generated through household surveys and budgets of costs of production and consumption as well as incoming income that covered 25 households in a resettlement site comprising 76 households. There was also two weeks of fieldwork car ried out in a resettlement site, which included participant observation and more in-depth discussions with focus groups. These data indicate that pauperisation and the transformation of a formerly self-sufficient adivasi community into casualised agricult ural wage workers may become a permanent consequence of resettlement, at least in this site. Second-hand reports from other sites indicate that this process is probably occurring elsewhere as well.

The average household size in the resettlement site was 5.7, with an increasing tendency to form stem and joint households being remarked upon, as compared to the nucleated settlements reported for the original villages.6 The original vi llage was Gadher, a large submerging village of 2,500 people [Joshi 1987:44] consisting of 17 ?faliyas? or hamlets often related by descent, which had been split up and resettled in 120 different sites. the date of original resettlement was 1990. Although a number of families tried to return to Gadher in 1996, they had been prevented by police form entering their home village and were forcibly returned to their resettlement site. The people in this resettlement site all belong to the Tadvi group and to on e faliya. The faliyas, in fact, seem to be both residential and descent groups, as a number of households possessed ghar jamais and the site studied was now said to constitute one faliya even though brothers had been placed in nearby resettlement s ites.7

As the Gujarat resettlement policy decrees that five acres of land is to be awarded to household heads (defined as male) and all major sons, a majority of families did possess landholdings, although 40 per cent reported less than the decreed five acres an d there were also a number of complaints that major sons had not been awarded any landholdings at all, 12 to be exact. This figure of non-allotment of land may be greater, as a number of families are still residing in the partially submerged riverside set tlment of Gadher. There was also a widow who had possessed two acres of land in Gadher but was not compensated for her loss of landholding in the resettlement site. There was also no irrigation in the resettlement site, although the construction of a near by canal had contributed to waterlogging and greater soil salinity, according to residents. Everyone I spoke to in the village complained that their land was relatively unproductive in comparison to neighbouring patels? landholdings even during a good mon soon. In a heavy monsoon, like the summers of 1996 and 1997, there was almost total crop failure.

The major finding from the household budgets and surveys was that incoming income from all sources, including agricultural wage labour, was much less than the average necessary costs even of food for a family of 5.7 during an average month, which was repo rted to be Rs 2,000. This figure was corroborated through household budgets. In addition, incoming income was entirely from daily wage work on farms located at a distance of from 5 to 15 kms from the resettlement site. The major reason why almost all fami lies were compelled into casual wage work was because of the total crop failure caused by severe waterlogging of land in the rabi seasons of both 1996 and 1997. Indeed, the total lack of yield from monsoon cultivation was immediately visible in a number o f resettlement sites as one surveyed the fields which had only a few stubs of ?makkai?, ?tuer?, or rice. All households surveyed reported that at least one member was engaged in casual agricultural labour, while a majority (72 per cent) reported wage work for more than one person in the household, including women and teenage sons. Workers were often picked up by trucks owned by farm owners, who paid contractors for an entire truckload, with the average daily wage being Rs 15, and the average days of work per month being 20. People reported that they sometimes had to return home from the roadside 5 kms distant, as there was no work available. One teenage boy and a woman also reported some harassment over unpaid salaries in the time I stayed in the resettle ment site. Four households also reported that one or two members had left the resettlement site to seek construction or casual labour work in Surat or Baroda.

All households reported a loss of livestock through the resettlement process, livestock which had died through hunger and disease, and particularly from the lack of grazing land in the resettlement site. Although the most severely hit were goats, there we re also a large number of families who reported a significant loss of cows and bullocks. Two households now possessed only one bullock, an obviously insufficient number for farming. Another five households reported keeping alive only two bullocks, with al l goats, chickens and cows having died. The chickens had apparently died from a viral disease. The meagre livestock holdings in the resettlement site compared unfavourably with the rather large herds of animals reported for the Gadher, where cattle-raisin g among Tadvis was considered an economic speciality and a sign of social prestige. Ten individuals from separate households also stated that they used to acquire quite a large income from selling cows and bullocks.

In contrast to the original village, the oustees in the resettlement site purchased all inputs and sold all output they produced through market mechanisms. Hence in statistical terms they would appear to be more ?developed?. However, this is where statist ical surveys, such as that of CSS break down, since they provide no way of calculating the ?value? of non-monetised inputs, and see monetisation itself as an index of better living conditions and increasing ?real? wages. Yet costs of agricultural producti on in the resettlement sites were and are much higher than in the original villages. This was not only due to the fact that most inputs in the submerging villages were home-produced and therefore non-monetised, but also due to the more capital-intensive n ature of inputs required in the new locations. In the resettlements sites, ?green revolution? packages of seeds, fertilisers and pesticides were considered necessary but were not so in the original villages.

Due to a decrease in diversification of crops, semi-annual output is also less secure. The only major crops attempted to be grown in the rabi season were rice and makkai, plus a bit of tuer. The majority of households surveyed attempted to cultivate their land in the 1997 rabi period using high-yielding varieties of rice and makkai, and also applied DAP and urea at an average of 0.8 bags per acre of each. The costs of DAP and urea were Rs 415 and Rs 150 per bag, respectively. Four households also reporte d using pesticides, at a cost of Rs 350 per litre, and an average of one litre per acre seeds were also purchased from the market, with an average cost per household being Rs 91. Hence average costs, not including the pesticides, for circulating capital l ast rabi season were Rs 3,660 for an average of four acres cultivated this season per household, an amount totally lost through the monsoon crop failure. This aggregate cost does not include the amount that had to be spent on bullocks, ploughs and the mai ntenance thereof. If we assume that three members of a household on average are working as agricultural workers, then incoming income was, on average about Rs 900, producing a huge shortfall of income in relation to basic costs of production alone. Consum ption costs were also reported to be much higher than in the original village, where a great deal was produced for subsistence and did not have to be purchased from the market. Despite this increasing impoverishment, only one person reported being able to take a loan from a shopkeeper/restaurantowner five kms away, and that happened last year, for which he washed dishes in repayment. No one else in the village was able to contract a loan this year from any source: bank, co-operative society, or private mo neylender. In fact, there was no bank or co-operative society in the resettlement site. The sensitivity of the debt issue and its relation to honour and self-respect may have clouded responses to this question and require a greater level of trust between researcher and interviewees to reveal [Hardiman 1996]. However, a near 100 per cent negative response is striking and possibly indicates that the poor quality of adivasi land in this resettlement site excluded them from participation in the credit market in any form. This phenomenon has been noted for landless and marginal households in Uttar Pradesh and elsewhere, and may indicate that neighbouring merchants and large farmers themselves consider the land allotted to resettlees to be too worthless to coun t as security [Whitehead 1991]. In fact, those resettlement sites with unproductive landholdings may constitute a quite novel phenomenon: middle peasant households on paper who are, at least in the monsoon season, landless agricultural labourers.

The extent to which resettlees overall were given unproductive land under the Gujarat NWDT cannot be gauged from preliminary fieldwork of two weeks which involved one in-depth village study. However, I was taken to six resettlement sites in Dhaboi taluka which is one of the most favourably irrigated of all talukas in which the project affected peoples (PAPs) have been resettled. These resettlement sites reported crop losses of between 50 and 100 per cent in the monsoon season, and sightings of resettlemen ts sites from highway also indicated substantial crop failure due to waterlogging elsewhere.

Under these conditions, viz, a drastic shortfall between incoming and outgoing income, the response of Tadvi households during the summers of 1996 and 1997 was to increase self-exploitation. In plain English, they consumed less, and during the week I live d in the resettlement site, this meant reducing consumption to one meal per day. This compared unfavourably with two meals a day plus morning chai in their original village. It may be argued that the 1996 and 1997 seasons were particularly bad due to a he avy monsoon that led to waterlogging of land. However, bad weather conditions have not been infrequent since they resettled. The occurrence of bad harvest seasons in 1991-92, 1995-96, 1996-97 and 1997-98 has pushed people in this site to a sense of despai r and makes their desire to return to their original village more than understandable even from a purely economic point of view. Human beings, as well as animal life, after all, need replenishment not just in three out of seven years, but during the enitr e period of resettlement and after. Many of the coping mechanisms reported for the drought of 1991-92 in the CSS Reports,8 such as help from relatives, sale of animals and silver, seem to have been depleted. The latter two options are, i n any case, one-shot affairs, while this resettlement site reported the death, rather than the sale, of livestock due to sickness and the difficulty in acquiring fodder. Fodder collection was now done by women rather than through grazing and hence, added to their daily household chores.

Given an overall response of unproductive land, waterlogging, crop failure, and households engaging in daily agricultural wage work to make ends meet, the CSS Reports? statement that only 1 per cent of all oustees in 1996-97 were involved in agricultural wage labour seems almost fantastic, since there were more than 30 agricultural labourers from itself for the rabi season, which would constitute more than 1 per cent of their global population of 2,981 resettlees [Sah nd]. Examining the tables more closel y, and the breakdown by sites, it is clear that there are some sites, e g, Dhamadra, Piparvati, Sandhia and Jemalgadh, where the percentagesare over 10 per cent while in Vadaj-2, the proportion is 41 per cent, and in Kali-Talavadi, it is 8 per cent. Hence , the probably very distressing situations in Vadaj-2 and Piparvati, have been minimised by including them with 25 other villages which apparently had less, or negligible proportions of adult men engaged in agricultural labour. What this calculation accor ding to average percentages omits are the inter-relationships between landholding, quality of land, irrigation, yield, access to credit, and options for agricultural wage labour in each site. Hence, if soil fertility, irrigation and other facilities are f avourable in 33 per cent of sites, marginally favourable in another 33 per cent of sites, and unfavourable in the remaining third, then the overall percentages will fall in the middle range.9 In addition, by multiplying percentages, low er percentages always emerge. Hence, if 55 per cent of sites reported some agricultural labour, and the average percentage was 3 per cent, then the total multiplied is close to 1 per cent, the global percentage given in CSS Reports. Strange as it may seem , my finding for this resettlement site that all households reported at least one member working for daily agricultural wages can be constructed to correspond with an overall statistical average for all resettlement sites of less than 1 per cent, masking and minimising the real hardships faced by the 77 households in the resettlement site studied at least during the summers of 1996 and 1997, if not in other periods.

It is indeed a pity that the same method of calculating averages has not been applied to the tables of male and female members engaged in their own cultivation. Surely, if all adult working members were granted five acres of land, one would not expect to find much less than 100 per cent of male working members engaged in ?own cultvation?. Yet percentages of 42 per cent in Ambavadi, 48 per cent in Chandanagar, 49 per cent in Chicadia, 24 per cent in Dhamadra, 50 per cent in Khandupura, 46 per cent in Navga m, 48 per cent in Pani-Mahuda, 47 per cent in Sanoli, 15 per cent in Vadaj-2, and 36 per cent in Vaviala - all below 50 per cent for the rabi season. This surely indicates that there were and are continuing flaws in the implementation of the resettlement policy and that there have been everyday hardships associated with that process that have been glossed over through statistical means. In fact, the percentages are not much higher for the kharif season for these resettlement sites as well, yet noticeably decrease when averaged against resettlement sites with better land.

There were also many complaints about the infrastructure at the resettlement site, which are somehow erased through the averaging procedures of statistical analysis. One of the most serious concerned the lack of medical care in or near the site itself. Th e nearest doctor was in the taluka headquarters, Naswadi, at a distance of 15 kms. Having accompanied someone there, I can attest to the difficulties of reaching a doctor: first there is a five km walk of Keladia, then a wait and finally a 20 minute ride on an overcrowded bus. Indeed, one might speculate that a seriously ill person would have expired through the attempt to reach medical care. Sarcasm aside, the impossibility of acquiring adequate allopathic care during emergencies raises concerns about th e CSS Reports that there were better levels of medical care in or near resettlement sites as compared with submerging villages.

This statement can only be construed as ?factual? by the averaging procedures previously mentioned and/or by assuming that there was no medical care in submerging villages. In other words, if five resettlement sites possess primary health care centres wit hin a five km distance, while five more have to travel 15 kms to reach one, then on average, they would all fall within the stipulated distance of 10 kms.10 Granted that the major form of medical care in submerging villages was through ? desidvai? practised by traditional healers or Bhuwa. It is only by discounting the Bhuwa as medical practitioners that CSS Reports can assert that there are better forms of medical care available in the plains than in the hill settlements.11 In addition, there were often no allopathic practitioners in the hill villages, hence any allopathic medical facility, however distant, can be reported as a statistical improvement, although it does not necessarily reveal a real increase in well-bein g. Also, epidemiological patterns are different in the original villages as compared with resettlement sites, a factor that seems to have been ignored in the CSS comparisons of health care ?before and after?. More concentrated settlement patterns characte ristic of the resettlement areas have been historically accompanied by higher levels of transmission of infectious diseases, as the original villages often contained hamlets of one faliya each separated by a distance of 50 metres or more. Although I do no t yet have statistics relating to morbidity, most people in the resettlment site thought that there were higher levels of infectious diseases in the resettlement site and two deaths of infants from diarrhoeal infection were reported in the last three mont hs. Waterlogging around the village pumps and the nearby canal seemed to provide an ideal breeding ground for mosquitoes, of which there were many.

The amount of compensation paid for house-building was a frequent complaint as Rs 10,000 was enough only to lay the cement foundations of houses and I saw at least five house-plots that consisted of foundations alone. Besides two pucca houses, one belongi ng to a cement contractor and another to a salaried machine-operator, the rest of the dwellings were the original tin sheds which had been provided seven years ago. These dwellings were said to be extremely hot during summer and cold during the winter mon ths. The cement contractor was said to have ?eaten? (?khayo?) four-fifths of the money allotted to construction for wells, pump-sites and house construction. Complaints to the SSPA about this problem were apparently ignored until 1998.

Access to educational facilities was also a major problem, with potentially long-range consequences. A primary school had been constructed, but a caved-in roof since the start of the monsoon season made it non-functional. The distance to secondary and hig h schools of 12 kms, combined with the economic distress of households has meant that many adolescents have been withdrawn from school. Instead, they are needed to help their parents through wage work or housework. Given the fact that Gadher possessed thr ee schools, and the Tadvi have had relatively high levels of literacy and education [Joshi 1987:56], this lack of educational facilities is acutely felt. Some parents, particularly those from ?bhagat? families, worry about an emerging drinking problem amo ng male adolescents. The water supply is also a major complaint as none of the water taps in the village are working, while the well has only monsoon rain water which is stagnant and unusable. Hand pumps are difficult to use, and dwellings are regularly i nundated during the monsoons. Water collection by women is thus a more time-consuming affair than it was in the home village. People were also not given any money for initial land clearance, and had to spend from Rs 7-8,000 clearing the land six to seven years ago. It is not an exaggeration to say that the resettlement site resembles a rural slum.

In terms of other facilities that were promised in the resettlement package, a weekly market existed 30 kms away, a grocery shop did exist in the resettlement site, there was no rehabilitation office, the police ?chowki? was 10-15 kms away, no temple or m osque, no threshing ground, a flour mill at a distance of 5 kms, a long and muddy approach road, no tailor, cloth shop or barber, and no homeopathic, ayurvedic, unani or bhuwa practitioners or nurse, or bone settlers in the village. For most ailments, peo ple tended to prefer a bhuwa, who used ayurvedic remedies and who lived at a distance of three kms from the village. One man who appeared to have lost a great deal of weight through a chronic ailment of three years, reported no satisfaction from the gover nment hospital in Naswadi, but subsequent improvement after treatment for three months by a bhuwa.

On a more qualitative note, everyone in the village, although attempting market-driven agriculture, yearned for the relative security of a more subsistence-oriented economy in the original village. Striking comparisons between the high costs of consumptio n and production in the resettlement site and their negligibility in the home village were a frequent topic of conversation. Everything there, they said, was provided by the jungle and their own cultivation. Extra land for cultivation could also be acquir ed in the forest, a pattern of use of rights which has been administratively ?read? as encroached land and thereby legally exempt from compensation. In the home village, only salt and clothing was exchanged in the market for ?bidis?, there were ?free? veg etables from the forest, cool mountain breezes in the hot season, and people grew enough in just one cropping season to sustain them throughout the entire year. Today, by contrast, everything, including people, was for sale. They also asserted that in the ir home village, women had Rs 200 in their pockets, while in the resettlement site, even men had hardly five rupees to spend.

Women also complained that the distance to their natal families was now much greater and that they had a lot more work to do in collecting fodder and fuel. Many women were also working as agricultural day labourers. The women in the resettlement site, how ever, did not seem to take an active part in political affairs associated with resettlement issues.

These qualitative statements stand in contrast to the ?perceptions of oustees? reported in various CSS Reports. For the 1996-97 period, apparently 90 per cent of PAPs in the rabi season reported that their economic condition had either improved or remaine d the same compared to in the submerging villages [Sah nd:9].

Only in Vadaj-2 have PAPs reported significant deterioration in their economic condition; here only 33 per cent reported that it had improved. The situation in Ambavadi, Chandanagar and Navgam is also not very encouraging, for about 25 per cen t PAPs in these locations were unsatisfied with their economic condition. In these sites...either farm production was low, e g Chandanagar and Ambavadi, or land had not been received by PAPs by the time this survey was conducted [Sah nd:10].

CSS Reports #24 does not provide any quantitative (or qualitative) information on the economic conditions of people in the submerging village. It continues by minimising reports of negative perceptions:

The findings relating to perceptions should be interpreted with caution...As had been mentioned in earlier reports, these findings do not always reflect the comparison of economic conditions in new sites with those of submerging villages. PAPs observe their surroundings and compare their situation with peer groups as well as host villages. Moreover, over the years PAPs? exposure to various groups may have heightened their expectations from R and R. It is also likely that some PAPs may be using the survey results as a platform to pressurise the policy-makers to relocate them owing to reasons different than their dissatisfaction relating to economic well-being in their new sites. Since it is not possible to decipher genuine from spurious respons es, the findings may need further verification...Despite the operation of these forces, the proportion of PAPs responding favourably is quite high [Sah nd:10].

It is hard to understand why the most negative possible spin is given to findings which might contradict the Gujarat government?s resettlement goals. In fact, social scientists generally agree that if there is distortion in answers to survey questionnaire s, it more often reflects an attempt by respondents to provide the answers desired by the questioners, especially as the investigators are often urban-based, literate, and enter the villages with the trapping of institutional authority. Having basically a ccused the respondents of consciously manipulating the interviewers, CSS Reports #24 then offers them a modicum of reason by noting a correlation between negative perceptions of before and after conditions, and those PAPs having low yields, poor land, cro p failure, low irrigation, higher dependence on market for food consumption and higher dependence on agricultural labour.

These findings reveal that farm production of households who are dissatisfied is significantly lower, and so is their capability to feed their members from their own production. Those households who had to resort to agricultural labouring or c asual labouring are also those who perceived that their economic condition has deteriorated [Sah nd].

The report then discounts potential rationality of oustees by stating that ?perceptions are (sic) not necessarily reflect a comparison between pre and post situations but are formed by contemporary experiences of PAPs in the new sites?, [Sah nd]. < P> However, despite CSS Reports goal of ?hard? statistical and quantitative analysis, there are no comparative data on the submerging villages given in this report by which readers could evaluate their reasons for negative perceptions. Nor is there any other justification for this statement, which seems to directly contradict the previous paragraph, and seems to assume that the PAPs are, at best, simpletons, and at worst, liars.

In talking to people and living with them in the resettlement site over a two-week period, the major reasons for the apparent distortions in the CSS Reports became much clearer. Their statistical calculations basically compare incomparable entities, since their baseline is a series of quantitative before and after calculations that assume value is only associated with the market and its exchanges. In fact, there are little or no ?before? statistics, as there has been no attempt made to assign quantitative values to non-marketed inputs and output in the submerging villages. Since non-marketed seeds, fertilisers, rain-fed irrigation, and home-made implements represent the bulk of agricultural inputs in submerging villages, these reports are misleading at be st. The report is structured around a bias which assumes that capitalist modernisation and development is necessarily better, while subsistence-oriented production and traditional health care is either non-existent or represents a negative blockage to be overcome. Since only market-oriented agriculture is assigned a quantitative value, and subsistence-oriented agriculture enters only as a lack. Similarly, allopathic, modern health care is taken as the ?norm? and anything different, such as traditional, ay urvedic care, does not statistically exist, hence it enters the monitoring only as a lack.

This association of non-monetised agriculture and non-allpathic medical care with primitivity and lack is, indeed, a bias which has been evident in dealings with Gujarat adivasis since colonial times: various resettlement schemes in the 19th and 20th cent uries have attempted to destroy shifting cultivation and introduce market-driven agricultural ?reforms? to induce bhils to become intensive agriculturalists producing a stable revenue. In addition, the 1865 and 1878 forestry acts criminalised slash-and-bu rn cultivation, and decreed any use of forest products to be illegal. Not surprisingly, discourses of primitivity, wildness and nature accompanied this passage of the bhils into the civilising process of colonial resettlement and reform [Skaria 1997], a p rocess which the bhils seem to be re-experiencing through their current resettlement process. Behind these ?modernisation? and ?betterment? schemes lies a drive to extend private property rights and ?develop? a market in land to provide a stable revenue b ase for the state. This involved not only the spread of private property rights in land, but also the extinguishing of customary, use rights in the soil, with the latter being termed ?encroached? land. In addition, private property was extended to forest products themselves through legal limitations placed on common use of forest products, which often became international commodities. Prior to the forestry acts, the adivasis had been able to shift their cultivation easily since land was relatively abundan t; the forestry acts, however, withdrew between 28 per cent and 85 per cent of land from shifting cultivation in some districts in Gujarat and thus created land as a ?scarce? resource. At the same time, the ?primitiveness? of adivasis was highlighted in o rder to justify the harshness of this ?modernisation? process.

Support for my criticisms of CSS Reports and for the views of resettlees that their real standards of living and well-being were higher in the original villages are supported by my 1998 observations in a partially-submerged bhil village in Nandod taluka, just behind the Sardar Sarovar dam, which arise from a stay of six weeks during the summer of 1998. Qualitatively, the difference in the partially-submerged hill village and the resettlement sites is obvious from a quick glance. The village is situated in a beautiful site, close to what used to consist of the Shurpaneshwar sanctuary and consists of relatively dispersed hamlets scattered on hillsides broken by small, cultivated fields and forests. The housing provides perhaps the most vivid visual contrast to the resettlement sites: compared to the tin sheds which are still predominant in resettlement sites, the hill sites boast of spacious, one-room houses constructed of mud and thatch with teak doors and beams. It is easy to see how a tendency to romanti cise adivasi culture can develop, as the contrast between the highlands and the plains is a stark one, even to a cynical academic!

Yet this partially-submerged village is no authentic, isolated primeval habitat: about 45 per cent of the families consist of people who have returned from various resettlement sites, particularly Ambavadi and Piparvati, either due to rocky land or to hos tility from host villagers. The village therefore consists of people who have chosen a subsistence-orientation over market-driven agriculture due to the inability to secure a livelihood from the land distributed through the resettlement programme. The vil lage can therefore be termed a subsistence-oriented or even resubsistence-oriented village not because of its isolation from markets, but due to the fact that agricultural production is not dependent on the market for inputs such as irrigation, fertiliser s, pesticides, and seeds. Nor is it dependent on markets for the selling of agricultural produce, and is hence buffered from the competitive pressures which such regional and national markets impose on commercially-oriented agricultural producers. Subsist ence agriculture versus commercial farming are analytically distinct, not because of the relative isolation of the former from markets, but because the underlying conditions of production remain outside of competitive circuits of capital, inputs, and sell ing of output which provide the defining feature of commercialised farming [Friedmann 1981]. Indeed, many households possessed a number of market-purchased consumer items such as bicycles, clocks, clothing, one tape recorder, soap, watches, etc. This can be interpreted as an index of the relative security of subsistence cultivation and the availability of other sources of incomes in the partially-submerged village.

I formally interviewed 67 heads of households out of a total of 72 households in the village, as well as documenting the varieties of fruit trees, vegetables, medicinal plants, and crops and their uses that are available to the Tadvi and Vasawa households who have remained behind in the partially-submerged hamlets or have returned from resettlement sites in the past two years. Thirty of the 67 households interviewed consisted of people returning from various resettlement sites, and they cited poor land pl us the high cost of pesticides and fertilisers as the major reasons for their return. Other reasons included death of cattle, bullocks and goats, lack of medical facilities and/or irrigation and schools. A few respondents reported that they had been cheat ed by the government, since the land they were originally shown was of higher quality than the land they finally received. Several also reported the harassment of Tadvi women and theft of personal belongings in the resettlement sites. Of the 47 households who were not returnees from the resettlement sites, 12 had not yet been allotted land, two households are still planning to shift, 10 householders reported that they did not want to shift as the allotted land was infertile or problematic, five did not le ave due to other reasons, such as lack of medical, irrigation or educational facilities, and 10 households had one or more members, often the eldest son, working the resettlement land while the rest of the family stayed behind.

For the purposes of this paper, however, the most significant finding to emerge from this study was a computation of household income and expenditures, which showed a great difference with the resettlement site previously reported. These findings also con tradict the economic advantages claimed by the CSS Reports for the resettlement programme. Sources of income in the partially-submerged village were much more diversified than in the resettlement site I surveyed. They included the sale of goats and chicke ns, and sale of agricultural produce, of bidis, medicinal plants, and of fish, work in the forestry department, government jobs, and labour on the dam project itself. Even in terms of cash income, 60 out of 67 households reported a monthly surplus, while four households broke even, and only two showed a minor deficit.

Yet the index of cash income alone does not tell the whole story, since it precludes those household inputs which are non-monetised and produced on their land or acquired in the forests. This includes one crop per year, mainly of jowar, tuer, urad, banti and bhaidon, and which they state lasts them for the entire year.12 Hence neither double-cropping nor irrigation is necessary. These crops do not cost much to produce, since most inputs, including seeds, dung fertiliser, irrigation, and ploughs are non-monetised and produced by households themselves. Indeed, carpentry seemed to be a skill known by most men in the village. Other non-monetised inputs include vegetables, tree fruits, fish, buttermilk, and medicinal plants for consumption, t imber for firewood, house construction, ploughs, dung for fertiliser, ?datun? for teeth-cleaning and other medicinal purposes, and ?khair? for house-making. In total, I documented 66 types of trees, 20 types of medicinal plants and 42 types of fruits that were used by residents. The importance of forest resources for the household economies of this partially-submerged village is indicated by the common statement that even if agriculture failed for three to four years, people could survive on forest produc e alone. Such statements contrast with responses from the resettlement site, which revealed that their winter crops of wheat and jowar only lasted two to three months since their land was poor, and their summer crops were destroyed completely in 1996 and 1997.


The different assumptions underlying qualitative fieldwork-based information reporting versus quantitative, statistical analysis that have yielded such different results raise questions about the reliability of averaging procedures used in the CSS Reports . The averaging procedures characteristic of the CSS tables and conclusions minimise the stress and difficulty faced by unfavourably endowed resettlement sites and households. This is most starkly evident in questions of caloric consumption, in which the average of low-calorie with high-calorie households and sites can hide real cutbacks and even hunger experienced by badly endowed households and sites during periods of crop failure, eg, in the resettlement sites of Dabhoi taluka, Vadodara district in the summers of 1996 and 1997.13 In addition, where statistical methods include biased assumptions as a starting point, e g, that non-monetised inputs are a sign of backwardness and should not therefore be counted, statistical correlations a nd coefficients only lend an air of sophistication to an already ideological procedure.14 However, if we start from peoples? everyday lives and experiences, and the social relations through which households co-ordinate their activities i n response to externally-generalised market pressures, rather than prior assumptions about what constitutes ?development?, then a very different picture emerges. In the case of the resettlement site in Dabhoi taluka, what I witnessed was the immiserisatio n of formerly self-sufficient adivasi societies, the creation of new labour nomads of south and central Gujarat.15

The provisions contained in the NWDT for resettlees in Gujarat are probably the most generous awarded by any state or central government in India to peoples displaced by dam and/or other development projects. The pain and trauma involved in the implementa tion of even this resettlement process raises questions about whether there is such a thing as a beneficial or equitable resettlement policy. Development, after all, like other terms such as ?modernity?, ?tradition?, ?progress?, and gross domestic product , is a vague abstraction which has diverse meanings for different people. Yet it carries considerable emotive weight. As a vague yet semantically loaded abstraction, it often hides the disparities that accompany capitalist forms of change, as well as smoo thing over the top-down and non-participatory character of its managerial regulation [Esteva 1992:7-8]. Originating first as a term denoting the unfolding growth of an organism, it became linked through 19th century social Darwinism to the idea of social transformation towards ever-more perfect forms. In this hierarchical reincarnation, it becomes almost impossible to question its prerogatives, and the ascendancy of homo economicus which is presupposed in its theoretical unconscious. In the case of India, the semantic associations linked with the term ?development? carried the added weight of the Nehruvian versus Gandhian legacies of the different visions of india?s post-independence future. Yet framing the question of the resettlement of the bhils in terms of an inevitable cost or side-effect of the development process forecloses many questions relating to the general character of capitalist forms of change in which there are a few winners and many losers. The ?development? of bhil society over th e past century implies that they were formerly ?underdeveloped? as slash-and-burn forest dwellers, a term that carries additional pejorative connotations of primitivity, ?jungli? backwardness, inefficient resource use, and closeness to nature. It diverts our attention from the reality that they most probably enjoyed a much more secure subsistence base and even perhaps better health than they have experienced through the various development projects imposed upon them without their input. It is perhaps wise to remember that for the bhils, as well as three-fourths of the world?s former slash-and-burn cultivators and hunter-gatherers,16 ?development? is a threat that has already been carried out.


[I would like to thank the Social Sciences and Humanities Research Foundation of Canada for providing the funding which made this preliminary study, and subsequent ones possible.]

1 By standpoint methodology, I do not mean solely that I will develop my concepts and present my information from the bhil?s point of view. Rather, I mean my concepts will be developed close to the ground, as it were; they will take the lived social rela tions, rather than a priori abstractions about them, as the point from which to question and re-question specific objectifying and generalising procedures contained in CSS Reports.

2 I am not sure why one 100 calorie increase between the submerging village and 1992-93 in the resettlement site constitutes a major increase in caloric consumption.

3 This association of ?tribal? societies with primitivity and underdevelopment, closer to nature than to culture, is not, of course, limited to the CSS Reports, but is a pervasive metaphor in both western and post-colonial south Asian discourses.

4 I will deal with the ways in which this conceptual move undervalues the quality of life of original villages later in this paper.

5 And not only to them, of course, as similar current images of primitivity abound in the literature on contemporary Amazonian ?tribes?.

6 The name of the resettlement site has not been given due to necessities of anonymity.

7 This observation was confirmed by visits to sites near Vagodhia, where a particular ?faliya? was referred to as ?petrol-pump faliya?.

8 CSS M and E Reports, Surat, Gujarat, #21.

9 As previously mentioned, there is no way of actually calculating from the CSS Reports how many villages might be favourably endowed, and how many might have partially or totally unproductive landholding. If one assumes that the 31 villages out of 120 v illages were selected according to random sampling, then it would appear that at least 6, or over 1/5th possessed unproductive landholding, and were forced into distress conditions even during relatively good monsoons. For the four years in which ?abnorma l? weather conditions have prevailed, the number of villages which suffered crop failure would have been much higher.

10 By such averaging procedures, the UNDP has declared Canada to be either first or second in ?development indicators? since they started monitoring ?national progress?. When these same yardsticks are applied just to Native Peoples of Canada, the country drops to 67!

11 There is now a substantial body of literature on the relative merits of allopathic and ?traditional? forms of medical care in medical anthropology (Biomedicine Re-examined, M Locke and D Gordon, 1995, Body of Power, Spirit of Resistance, Comaroff and Comaroff, 1985, M Taussig, Colonialism, Shamanism and the Wild Man, 1992, T Csordas, Embodiment and Experience, 1994, etc). The fact that CSS Reports ignore the criticisms of the mind/body duality underlying allopathic healing a nd the advantages of socially constituting disease agents reflects their bias in assuming that modernisation and allopathic care are inherently superior in all treating all forms of un-wellness. Traditional healers, ipso facto, become an absence, a lack. In fact, many medical anthropologists now see allopathic care as superior mainly in infectious diseases and surgery, while chronic complaints, i e, those more frequently encountered in the hill environment, are more adequately treated through ?trad itional? health care.

12 People in the partially-submerged village claim that ?banti?, ?bhaidon? and ?kodra? will last for 30 years, hence providing a fall-back for subsistence in lean seasons or bad monsoons.

13 I had been told that Dabhoi taluka boasted of the best resettlement sites with the most adequate irrigation. Unfortunately, most seemed to be under water in the monsoon season of 1997.

14 I am using the term ?ideology? here in the sense that Smith (1996) defines it, as a knowledge procedure that erases subjects? life experiences and the specificities of local communities in the process of producing quantitative information which is appa rently generalisable across communities and individuals, but which omits significant and important local and/or individual information. This is very different from Mannheim?s definition of ideology which is often confused with that of Marx.

15 This parallels J Bremen?s findings for the earlier, Ukai dam. See Bremen (1994).

16 These groups are often referred to as ?indigenous? peoples, a term coined by the UN and arising from the historical experience of North American native people.


Breman, J (1994): Wage Hunters and Gatherers: Search for Work in the Urban and Rural Economy of South Gujarat, OUP, Delhi.

Esteva, G (1992): ?Development? in W Sachs (ed), The Development Dictionary: A Guide to Knowledge as Power, Zed Books, London.

Fabian, J (1983): Time and the Other: How Anthropology Creates Its Objects, Columbia University Press, New York.

Friedmann, H (1981): ?Peasants and Simple Commodity Producers: Some Analytical Distinctions?, Journal of Peasant Studies, Vol 8, No 3.

Hardiman, D (1996): Feeding the Baniya, OUP, Delhi.

Joshi, V (1987): Submerging Villages, Ajanta, New Delhi.

Sah, D C (nd): M and E Report, Vol 24, Centre for Social Studies, Surat.

Skaria (1997): ?Shades of Wildness: Tribe, Caste and Gender in Western India?, Journal of Asian Studies, Vol 56, No 3, August.

Smith, D (1993): ?No One Commits Suicide: Textual Analysis of Ideological Practices?, Human Studies, Vol 6, pp 309-59.

- (1987): The Everyday World as Problematic, University of Toronto Press, Toronto.

- (1997): ?From the Margins: Women?s Standpoint as a Method of Inquiring in the Social Sciences?, Gender, Technology and Development, Vol 1, No 1, pp 113-35.

Whitehead, J (1991): ?The State and Kabir Panth in Central Uttar Pradesh?, Social Scientist, Vol 11, Nos 11-12.

Current Issue    Previous Issue    Archives    Subscription
About the EPW    Sameeksha Trust    EPW Research Foundation
Advertising    Announcements    epw@vsnl.com    Home    The Site