In this paper I intend to describe some planned experimental work on communication networks in the context of work I have already completed. I will first describe the earlier work, then some computer simulations that lay the groundwork for experiments I plan to run next year.

The aim of this paper is to investigate how changes in ideology made it possible for the Dutch state to successfully adapt to the decolonization of Indonesia . We will analyze how ideological changes occurred within the Dutch administrative apparatus in this period. Our methodology derives from network theory. We posit that two distinct but interrelated networks operate within the policy making arena. The first is the formal, institutional network which coincides with the departmental structure of the state bureaucracy. The second consists of informal networks which crosscut the former. In the critical historical conjuncture of 1945-1949, informal networks played a dominant role in the changes of ideology. Throughout this paper we will demonstrate how informal networks were crucial in creating a new concept of control which guided the postwar reconstruction in the Netherlands.

The Soviet Union publishes over fifty some official publications that are publicly available throughout the world. Based on these publications, an on-line computer data base, SOVT on the server GECAM in Paris, provides detailed official biographies of leading figures of the Soviet Union.

Structuralists have often assumed, a priori, that our data are unlike those of more traditional quantitative sociology. Their data provide counts, counting the numbers of people with certain attributes. Our data describe relations, two people are friends or they are not. Their data have tables with rows and columns representing variables. Our data have tables with rows and columns representing unaggregated individuals. This analysis challenges the a priori assumption that such data are "different", requiring special treatment. It offers testable working hypotheses : The tests suggest some advantage to analyzing sociometric data as distributions of events and for analyzing structures in terms of interval scales.

These pages are inspired mainly by Muriel Hammer's (1981) contention that "the understanding of schizophrenia requires a largely unexplored 'purely' social dimension", term which designates in this author's view the need to study more thoroughly the structural variables usually investigated in network analysis carried out in normal populations. The argument for focusing on structural characteristics of networks is that only this approach would allow meaningful comparisons with data regarding network processes in free-of- pathology general population samples. The point I would like to emphasize is that prior to any discussion of the feasibility of the procedures proposed by Hammer for translating in structural terms some nonstructural network properties, we must try to answer the generally overlooked question of the specific conditions under which findings from network studies in general population can be generalized in order to explain the network processes of psychiatric patients, or the opposite.

The idea of abstracting and/or generalising the structure of a narrative (i.e. a set of weakly connected human actions) using a mapping rule which creates either a tolerance (a symmetric reflexive relationship) or an equivalence on the domain set is explored. The paper is an extension of earlier work by the author. A detailed example of the use of narrative analysis is provided.