Long-Range Forecasts of Society and Culture: Four Quantitative Methods from Cultural Anthropology

Article excerpt

Abstract: Cultural Anthropology occupies a position of strategic importance for those wishing to forecast the future. Hundreds of long-range trends in culture are known to exist going back thousands of years. These trends, along with some theoretical models fitted to past societies, seem intuitively to be extrapolatable into the future. Worldwide archaeological chronologies are increasingly available to facilitate construction of time series. Four long-range quantitative forecast methods are developed for extrapolation of: (1) universals of culture; (2) atheoretical long-range time series models; (3) directional long-range trends; and (4) theoretical models. Examples are given.

Resume: L'anthropologie culturelle occupe une position strategique pour ceux et celles qui veulent predire l'avenir. On reconnait que des centaines de tendances de longue portee remontent a des milliers d'annees. Ces tendances, associees a des modeles theoriques pertinents pour les societes anterieures, paraissent pouvoir etre intuitivement extrapolees dans le futur. Des chronologies archeologiques a l'etendue du globe sont de plus en plus accessibles et permettent la construction de sequences temporelles. Quatre methodes quantitatives pour les predictions a long terme sont developpees dans cette perspective: 1) les univesels des cultures; 2) des modeles non-theoriques de series chronologiques etendues; 3) des tendances a long terme orientees dans une direction donnee; 4) des modeles theoriques. On presente des exemples.

Introduction and Preamble

Cultural Anthropology occupies a position of strategic importance for those wishing to forecast the future. Hundreds of long-range trends in culture are known to exist going back thousands of years. These trends, along with some theoretical models fitted to past societies, seem intuitively to be extrapolatable into the future. Worldwide archaeological chronologies are increasingly available to facilitate construction of long-range time series. Four quantitative methods are developed for making long-range anthropological forecasts of culture and society. The four methods are extrapolation of: (1) universals of culture; (2) atheoretical long-range time series models generated by the ongoing, underlying process of social/cultural differentiation; (3) directional long-range trends; and (4) causal models. While there may be more ways to forecast than these, the four forecast methods are widely applicable.

This paper is intended to be "reader" friendly to all readers whether they use quantitative methods or not. Throughout the paper, ideas are developed in plain English understandable to all. Where, at times, it is possible to distill results into mathematical form such mathematics are set off in rectangles or shunted to footnotes. Readers who so wish may skip over these rectangles or footnotes without loss of continuity.

Before we start, we need to consider several matters which set the stage for the forecast methods, and the forecasts themselves, which follow. A "long-range" forecast is defined here to be a forecast to a point in time 10 years or more ahead. Such a definition is admittedly arbitrary. Long-range forecast methods may also at times be used to make short-range forecasts. Our forecasts will be for times up to 2050 AD inclusive. Reasons for this will be stated below.

Forecasts are defined here to be unconditional predictions of what will happen in the future. Extrapolations are defined to be conditional predictions of what will happen in the future, provided certain assumptions are made. In the case of some extrapolations, for example the United Nations (1998) population extrapolations to 2150 AD, the extrapolator leaves unanswered the question of which of several different extrapolations will actually occur. This will not be the case here. In the present paper we will make extrapolations contingent on only a few premises which seem justifiable to the writer. These premises are:

A. …