When one speaks of a sense-impression, the term is perhaps sufficiently neutral to rule out a special act of a subject on the one hand and an object, in the sense of something known, on the other. The word can surely just stand for a mental occurrence of a given character, a mental process, an event, and it is so used here. In the simpler and early forms of mental life there occur these sense-impressions, and though they form part of the life stream of an individual, they are not in virtue of this fact to be regarded as caused by the individual in whose life they occur. They occur in relation to stimulation or are given', and may in this sense be termed objective '. As processes sense-impressions are cognitive processes, but they are not in themselves cognitions. They subserve cognition, but for this function to be possible it seems necessary, as a minimum, that likenesses and differences between processes should emerge. The process in itself is never what is known' when it occurs. (Reservation must be made as regards introspective knowledge, to be discussed. later.) Even in the case of a simple sense-quality, such as 'redness', 'coldness', awareness of 'red', 'cold' must be regarded as a 'construction'. In respect of function the process occurring has been assimilated to, differentiated from, other processes. It has a background. If it subserves cognition, it means more than it itself is. At once come the questions, who constructs? who assimilates? who differentiates ? and the conception of the individual in whose life stream the process occurs, rises as an answer. But then the individual is not something over and above his mental life. Knowing is the function of the cognitive processes, and we cannot divorce it from the processes as being the function of something else, x, the unknown. Nevertheless we may not simply identify it with these processes considered as events in mental life.

We have said that there is continuity of function with continuity and repetition of process. We have now further to maintain that a cognitive process can be representative'.

It is so whenever it means more than is its function in its own right. For, notice, it is in respect of its function that it is representative, not in virtue of its nature as an occurrence. The constructions of sense-perception are the outcome of ' representation', and it is just at this point, if at no other, that the recognition of conative processes seems essential to the story of the development of cognition.

We can only form a hypothesis as to the nature of primitive cognition. Processes of the cognitive type which do not actually subserve cognition never occur in human life. We speak of 'sentience', but even here the processes have some vague' awareness of' as their function. We need some such hypothesis as that of Prof. Ward's psychological individual to portray the origin of 'meaning'. We can only study its development. In mental life as we know it, apparently in every species of animal, certain predetermined sense-impressions will be selected, i. e. stand out amid the mass of impressions and arouse conative processes, wanting, striving, accompanied by affective processes. This in itself is no reason for limiting conation to these situations; on the contrary, the fact that the groups of 'predetermined' senseimpressions vary in different species, would seem to point to conation as something more fundamental than a process evoked by a limited number of sense-impressions. It may well be that conation can arise with any sense-impression whatever. But confining ourselves to the selections which are predetermined for the individual, the instinctive selections, we find that there will ensue in relation to these a series of conative processes, which however much we may distinguish them as moments or phases are a continuity. There is one continuous endeavour working itself out through them until it is completed. The completion is a moment with an affective value of satisfaction, over and above any other affective process which may be present, whether pleasant or painful. The series of conative processes has its beginning and its

end and a continuity between those points which is independent of continuity in time. The endeavour is throughout accompanied by cognitive processes. In so far as there is bodily activity there will be the cognitive processes of muscular and organic sensations, and there will also be the changing impressions of sight, sound, contact, taste, smell. The continuity of the conative series gives continuity to the cognitive processes, which are thereby brought into relation with one another. This is not to be understood as any loose relation of parallel chains of events, but as an organization of interdependent members. As an illustration take the experiences of a child trying to open a box which, as we say, has excited curiosity. The series of endeavours which starts with the sight of the box, and ends with, or is satisfied by, its opening and the discovery of emptiness, is a continuous whole. Throughout striving and knowing are interrelated; the changing contacts, the resistance here, the yielding there, the changing sight of the box from this angle, from that angle, of the lifting lid, of the lid falling back again, all these sensuous experiences go with the progress of the endeavour. Now it is thwarted, now it is going forward, is intensified, culminates and is fulfilled or complete. Hereafter the box is abandoned, unless a new conation is aroused in connexion with it.

But to return to meaning in sense-perception. Where such organizations as we have described have been formed, any one cognitive process that has entered into the organization can' mean' the function of any other cognitive process which has entered into the same organization. The construction of knowing in such a case might be termed 'active' in contrast to the passive construction due to continuity, similarity, or repetition of process. We may illustrate by a simple construction. For such an animal as a dog smell may be a 'predetermined' selection. On its occurrence there ensues striving, seeking; there is nosing round, contact with this and that,

eating, the taste which satisfies the striving. The smell and taste which were not in any temporal continuity are now continuous within the conative whole. They are in organic unity with one another. Hereafter that smell, as part of the same organic unity, can mean that taste; an active construction. What is all important for the development of such perceptual 'active' constructions is the control which the individual has over bodily movements. Control of the limbs means control of cognitive link processes whereby various senseimpressions are woven into the endeavour process. Handling, looking, moving to and fro in working out an endeavour, brings this into relation with that as forwarding or arresting progress. Thus muscular sense-impressions acquire their significance side by side with the impressions of the special

senses.

The integrations which are the outcome of conation, and the localization of sense-impressions, which is the outcome of both reflex and conative activity, form a long chapter in the history of the development of the individual. The point for our purpose here is that as a result of these active constructions we have retentiveness manifested in a new form. The object perceived on the occasion of sense-perception is a construction. It is real in that it is 'given' by the senseimpression occurring, but what is 'given' only partially determines what is perceived. The object known is not an event in the life history of the individual, whereas the 'given' impression is. This claim is essential for what follows as to imagery. In the active constructions of sense-perception there may be no imagery. When for the dog a given smell can mean a taste, and on the occurrence of the smell there ensues nosing round, hunting, there may be no imagery of taste. The sense impression now initiates the nosing round differently from the first occasion; we may say the progress of the endeavour towards fulfilment is more directly sustained than before. To adapt language from Alice in Wonderland,

the situation is labelled more plainly than before, 'Seek me'. Yet even in these simpler forms of what is described as 'impulsive action', conation may be sustained by an image representative of fulfilment. It seems impossible to draw a line between the cases in which this occurs and the cases in which it does not occur. Adult experience would lead one to cite cases where there is difficulty in the fulfilment as the cases where imagery is present.

How is imagery here and imagery as it occurs in genuine cases of memory to be regarded? Imagery in the cases. considered is the supplement to sense-impressions, and in memory proper it is the substitute for them. When we regarded a sense-impression of a given character as a continuation of a previous sense-impression, because it carried on the same function, we said this continuity did not imply the persistence of the sense-impression in any literal sense. It is not continued throughout the interval of time as a ghost process in any realm of sub-consciousness. As already pointed out there is no spatial structure within which to postulate continuity of function. If we use the expression structure at all with reference to mind, it can only be in a non-spatial sense. We may use it to express the organization to which processes contribute in respect of work or performance. Viewing knowledge or conduct as a work or performance, we can conceive of the organization of knowledge and of conduct as a structure. To seek to predicate of such a structure 'existence', as it is predicated of a substance in space, or 'persistence', as it is predicated of an event in time, is meaningless. The structure is neither a substance in space nor an event in time, but a framework in which these are projected. We can only say of this structure that it holds for, or is relevant to, each moment of mental life. The 'function' of a process is its office or performance with reference to the work of knowing or doing or appraising. We can speak of continuity of function without the implication of any temporal continuity