This paper attempts to present a theory of economic growth. In Section I it discusses the experience of growth modelling over the past 40 years and argues that it fails to capture the most important features of institutional and technical change. Nevertheless as a method for ordering concepts it can be a useful complement to historical research. The problem with history is the almost infinite multitude of events, which have to be classified, described and analysed. A simplifying theoretical framework is essential and inevitable. Section II tentatively presents such a simplifying classificatory framework. It argues that five historical processes or sub-systems of society have been shown by historical research to be relatively autonomous although interacting major influences on the process of economic growth. These five overlapping sub-systems are science, technology, economy, politics and general culture. Each of these is briefly defined. Humans share with other animals the natural which can also powerfully and reciprocally influence economic growth. The other five historical processes each have their own partly autonomous environment and are uniquely human, which is one reason why biological evolutionary analogies have limited value. Although each of the five has its own distinctive features and relative autonomy, it is their interdependence and interaction which provides major insights into the processes of ahead, catching up and falling behind in economic growth. Positive congruence and interaction between them provides the most fertile soil for growth, while lack of congruence may prevent growth altogether, or slow it down. Although a satisfactory theory of economic growth should help us to understand the evolution of the world economy much better, the limits of forecasting and prediction in the social sciences should be clearly recognised. Popper was surely right in maintaining that the most important historical changes are qualitative and non-repetitive. The fact that we can predict eclipses does not mean that we can predict revolutions. Section III discusses the problem of non-recurrence for the social sciences. Section IV takes a major example to illustrate the theory which has been tentatively advanced -- the archetypal example of forging ahead in the British Industrial Revolution in the late 18th Century. It briefly discusses a dozen or so major features of this revolution as identified by historians and suggests that together they justify the notions of confluence and congruence between science, technology, economy, politics and culture as a plausible explanation of the leap ahead in economic growth then achieved for the first time in world history. Section V then discusses British falling behind in the late 19th Century and 20th Century and suggests that this can probably be explained in terms of loss of congruence between the five sub-systems of British society. The rise of new increasingly science-based technologies and of specialised professional management in large corporations fitted ill with some of the older now traditional British political and social institutions. After a brief discussion of the more deliberate processes of catching up the paper concludes by pointing out that the theory put forward here resembles many earlier explanations of economic growth. Marx's materialist conception of history stressed the tensions between forces of production, relations of production and superstructure as a source of social and political change or of stagnation in economic growth. Many other historians and economists (e.g. Veblen, Mokyr, von Tunzelmann, Galbraith, Perez) have stressed in particular the inter-action between technical change and organisational change within firms, as well as political and institutional change at other levels in society. This paper differs from most of them and from Marx's theory in two respects. First, it attaches greater importance to science and to general culture. In this it resembles the theories of Needham and Bernal. Secondly, it does not attempt to assign primacy in causal relationships to any one of the five spheres, whereas most other theories assign primacy to technology or the economy or both. It emphasises rather the relative autonomy of each of the five spheres, based on the division of labour and each with its own selection environment. It is this co-evolution which generates the possibility of mis-match between them and periodically of radical institutional innovations which attempt to restore harmonious development. Such harmony however is not necessarily favourable to economic growth, which is not the only objective pursued by human beings. Congruence which is favourable to economic growth must be distinguished from other types of congruence.