SPECIALISATION IN no region is Herbert Spencer's statement, that development consists in change from an indefinite incoherent homogeneity to a definite coherent heterogeneity, truer than in that of industry. The change takes place by a process of specialism under the direction of considerations of economy. But increasing specialism is not the only striking aspect of industrial development. A certain amount of de-specialisation has taken place, and the causes of this, as well as the causes of the reverse process, must be investigated.
The best way to understand the conditions under which specialism takes place in manufacture will be to consider the course of development within a particular factory. The process which begins merely in arranging work among a group of people united for production, so as to avoid waste of time and enlist the economy of habit, leads eventually to the intense specialism of industries, businesses in industries and commercial processes, which characterises the productive system of the most advanced communities to-day. Adam Smith minutely described the initial stages in his chapter on the division of labour, but it was apparently beyond the stretch of his imagination to envisage the ultimate outcome. The first step to specialism is the discovery (if so it may be termed) that production by a group operating as an organic whole is more economical than system of duplicat ing tasks according to the size of the producing group. By the "group working as an organic whole" is meant that each does a part only of the complete operation involved in turning out a certain commodity. The immediate gains are self-evident; less time is wasted in going from task to task, changing tools, and settlin4 down. Speed is gained through habit; and an action can only become habitual if it is comparatively simple and is constantly repeated. Moreover, habit simplifies a process; it is only the task which is performed again and again which gets to be carried out by the shortest cuts. Further, there is a gain in that much duplication of tools is avoided, and another gain in that adaptation of tasks to tastes becomes possible. The secondary results of division of labour are perhaps even more important. Improvement of tools must follow and the develop ment of machinery. It is exceedingly difficult to think out, al) initio, a simple way of reaching a complex end, and it will be found, therefore, that the path of invention has almost invariably been tracked out by division of labour. By division of labour a complex operation is broken up, and by
repetition of a part the shortest process is discovered. The operation is finally reduced to a few simple movements, and at this stage the appearance of the machine cannot be long delayed.
The most cursory glance at the history of any manufacture will show that it has passed over to the large-scale system at a certain stage in its organic development. It reaches a certain level of complexity, and one or two important machines are introduced, and at this stage enterprise is likely to foresee large results from applying to it organisation on a grander scale. No doubt the utilisation of water- and steam-power had much to do with the development of the factory system, but this was by no means the sole influence at work.
Now it is evident that the larger the producing group the wider is the scope for specialism within it ; but, as every manufacturer knows, it by no means follows that, when the factory is made bigger, the cost of production falls, apart even from difficulties attending sales. For, as a factory expands, the intensity of control over each part weakens, and the cost of production is a function of this as well as of specialism. It is easy to see how the limit to growth is determined. It continues so long as the anticipated gain from enlarged scope for specialism is greater than the loss due to weakening management in respect of a given hint in production, and stops when the two equate. Hence, for:these reasons, a certain typical magnitude tends to be attained by each industry. This typical magnitude is a function of the internal complexity and quality of the industry, and we shall see later that it is determined also by other conditions. Of course a tendency only can be alleged. Large variations are actually witnessed—so commonly, indeed, that diversity is more apparent than uniformity in the magnitude of the business units of an industry—because a business is a slow growth and businesses rise and disappear like forest trees, and because industrial administrators are not equally endowed with capacity. But, though in a forest of pines, say, there are trees of many sizes, it is correct to say that a pine tends to reach a certain normal size.