CLUSTER-ANALYSIS AND THE CLASSIFICATION OF MEDIEVAL DEMESNE-FARMING SYSTEMS

A methodology for classifying historic farming systems is developed with reference to the demesne sector of English agriculture during the period 1250-1349. Data on arable and pastoral production are combined from a sample of approximately 1000 manorial accounts representing some 388 individual demesnes. These demesnes are classified on the basis of the crops grown, livestock kept, and ratio of the latter to the former using a combination of cluster analysis and discriminant analysis. Cluster analysis is subject to a number of limitations and the typologies it produces may be no better than the techniques employed, The method adopted here, therefore, has been to establish a stable typology through the construction of stereotypes. These are defined by those cases which are similarly classified by different clustering techniques. The stability of these 'core' groups is then assessed through the reapplication of clustering techniques and the application of discriminant analysis. On both counts the 'core' groups emerge as clearly articulated. Function coefficients obtained from discriminant analysis are then used to assign each of the remaining cases to the 'core' group to which it has t;he greatest probability of belonging. The eight basic farming types thereby identified are evaluated and their character and distribution discussed. These farming types can be ranked on a scale from intensive to extensive. Overall, agriculture is shown to have been more extensive than intensive and consequently capable of considerable further development. The most developed systems are shown to have been concentrated in the east and south east where market demand, and the growth of concentrated urban demand in particular, was greatest.