300. As to relations in time and space, we have already traced out the inconsistencies which attend Hume’s attempt to represent them as compound ideas. The statement at the beginning of Part III., that they are relations not dependent on the nature of compared ideas, is itself a confession that such representation is erroneous. If the difficulty about the synthesis of successive feelings in a consciousness that consists merely of the succession could be overcome, we might admit that the putting together of ideas might constitute such an idea of relation as depends on the nature of the combined ideas. But no combination of ideas can yield a relation which remains the same while the ideas change, and changes while they remain the same. Thus, when Hume tells us that ‘in none of the observations we may make concerning relations of time and place can the mind go beyond what is immediately present to the senses, to discover the relations of objects’ [1] the statement contradicts itself. Either we can make no observation concerning relation in time and place at all, or in making it we already ‘go beyond what is immediately present to the senses,’ since we observe what is neither a feeling nor several feelings put together. If then Hume had succeeded in his reduction of reasoning from cause or effect to observation of this kind, as modified in a certain way by habit, the purpose for which the reduction is attempted would not have been attained. The separation between perception and inference, between ‘intuition’ and ‘discourse,’ would have been got rid of, but inference and discourse would not therefore have been brought nearer to the mere succession of feelings, for the separation between feeling and perception would remain complete; and that being so, the question would inevitably recur--If the ‘observation’ of objects as related in space and time already involves a transition from the felt to the unfelt, what greater difficulty is there about the interpretation of a feeling as a change to be accounted for (which is what is meant by inference to a cause), that we should do violence to the sciences by reducing it to repeated observation lest it should seem that in it we ‘go beyond’ present feeling? [1] P. 376. [Book I, part III., sec. II.] As also does the ‘observation concerning identity,’ which the comparison involves.