Pakistan's first major social forestry project,(2) the joint Pakistan/U.S. Forestry Planning and Development Project, begun in 1985, immediately experienced problems with its field component, which was designed to assist small to plant trees on their farms. Early in 1986 a two-day, countrywide meeting was held to address some of these problems. On the first day, as project anthropologist, I addressed what I perceived (perhaps naively) as a misidentification of project clientele by emphasizing that the project was designed for small and that they should be the principal target of the field staff. A Pakistani forester present countered this advice with the assertion that there are no small in Pakistan. On the second day of the meeting I took another tack, emphasizing that the project was not designed to meet the special needs of large farmers, and so the field staff should avoid them. This suggestion also was contested; there are no large in Pakistan. The problematic existential status of the Pakistani farmer was reaffirmed to me in a subsequent meeting with a divisional forest officer who, in response to my request to meet with in the project area, assured me that he personally knew every farmer in the district, and what is more, could have them all in his office in less than one hour. The officer proved to be as good as his word: he made a flurry of telephone calls, and within the hour one-half dozen drove up to his office. These foresters' views of Pakistani were at variance with the government's own data. For example, government statistics then put the average farm size in Pakistan at just 4.7 hectares, with 74 per cent of (approximately three million farm families) working less than 5 hectares (Government of Pakistan 1988:80). Of the total of four million-plus farms in Pakistan, 55 per cent were worked by their owners, 26 per cent by tenants, and the remaining 19 per cent by tenants who also owned some land of their own (Government of Pakistan 1988:80). If the small farmer was not an endangered species in Pakistan, neither was the large farmer. Although national law ostensibly limits landholdings to a maximum of 40 acres, I met who owned, in every real sense of the word, tens of thousands of acres. The district with only a half-dozen farmers - all with access to cars and telephones - contained in 1986 approximately 1.5 million people (Mahmud et al. 1980:20),of whom fewer than one out of 326 Owned a motor vehicle, and fewer than one in 100 possessed a telephone connection (Government of Pakistan 1989:238,241). These statistics, and their variance from the government foresters' perceptions, do not just mean that the latter are in error. Rather, they mean that the disagreement over the nature of Pakistan's was part of a broad disagreement over representations of rural social reality, involving the project foresters,(3) the project designers and advisors, and the farmers. The focus of the disagreement is not the existence of small versus large farmers, but the official recognition of small versus large as suitable clients for government services. This disagreement, and its obfuscation as another sort (e.g., representation of the lack of desire to provide services to small as the lack of small farmers) is ubiquitous in development planning in the developing world, and it is the source of many if not most development impasses. The object of the current study is to deconstruct such disagreements by revealing the differing social representations or constructions that give rise to it. The emphasis is less on what the facts are, than on what the different parties involved want the facts to be. The object is to show how differing constructions of social reality affect development projects; and this refers not just to the constructions of farmers, whom we are used to treating as subjective actors, but to the constructions of officials, whom we are not used to treating in this fashion. …