The controversy over the interpretative issue-is Thomas Reid a perceptual direct realist?-has recently had channelled into it a host of imaginative ideas about what direct perception truly means. Paradoxically enough, it is the apparent contradiction at the heart of his view of perception which keeps teasing us to review our concepts: time and again, Reid stresses that the very idea of any mental intermediaries implies scepticism, yet, nevertheless insists that sensations are signs of objects. But if sensory signs are not mental intermediaries, what are they? Hasn't Reid merely swapped the common 'sensation' for the notorious 'idea', ending up with indirect realism? 1 Current imaginative strategies answer negatively: Reid's sensory sign does not contradict direct perception, and those who think otherwise merely fail to understand what it means.The first aim of this paper is to argue against compatabilist strategies that seek to redefine direct perception so that it suits Reid's theory of signs. I show that no matter how the idea of sensory signs is 'improved', as long as they play a cognitive role in perception, colour and hardness cannot be directly presented to us, and they are turned into theoretical properties.Why then does Reid include sensations at all in the process of perception? Is he an indirect realist? My second aim is to show that the reason for bringing them into his account of perception in the first place is his-to my mind erroneousmetaphysics of colour. Reid claims that colours (like all secondary qualities) are unknown physical qualities, and as such, cannot be perceived. By contrast, primary qualities are perceptible, and hence, in visual experience of them, Reid argues, there are no sensations. Thus, he restricts direct perception to primary qualities presented in vision. Given Reid's belief that a physicalist account of colours renders them imperceptible, a unified account of perception is not, for him, a viable option.But is not Reid's hybrid theory, 2 then, the worst of all worlds? With regard to colours, he is a clear representationalist; with regard to the visual primary qualities, he disavows any sensations, so that we are left with his definition of perception, which makes perception just an immediate belief, and thus loses the intuitive distinction between these two states. 3 Either way, the intuition that perception presents mind-independent objects is lost; secondary qualities are represented in experience, and primary qualities are believed to exist in our vicinity.