KNOWLEDGE and the wallpaper are external to me, the after-image and the subvocal humming are interior events But Hume accepts the empiricist thesis that all we ever know are our own perceptions My hearing the bird sing is not a transaction between myself and the bird, but my encounter with a vivid bird-like sound For Hume, everyone’s life is just one introspection after another It has to be by introspection, then, that we tell the diVerence between our memories and our imaginings The diVerence between the two, one might think, could best be made out in terms of belief If I take myself to be remembering that p, then I believe that p; but I can imagine p’s being the case without any such belief As Hume himself says, we conceive many things that we not believe But his classiWcation of mental states makes it diYcult for him to Wnd a suitable place for belief The diVerence between merely having the thought that p and actually believing that p cannot be a diVerence of content As Hume puts it, belief cannot consist in the addition of an extra idea to the idea or ideas which constitute what is believed One argument for this is that we are free to add any ideas we like, but we cannot choose to believe whatever we please A more convincing reason would be that if belief consisted in an extra idea, someone who believes that Caesar died in his bed and someone who does not believe that Caesar died in his bed would not be in conXict with each other because they would not be considering the same proposition (T, 95) In the Enquiry, Hume says that belief is a conception ‘attended with a feeling or sentiment, diVerent from the loose reveries of the fancy’ But such a feeling would surely be an impression; and in an appendix to the Treatise, Hume argues forcefully that this would be directly contrary to experience—belief consists only of ideas But he still insists that ‘An idea assented to feels diVerent from a Wctitious idea,’ and he oVers various names to describe the feeling: ‘force, vivacity, solidity, Wrmness, steadiness’ He ends by confessing that ‘’tis impossible to explain perfectly this feeling or manner of conception’ (T, 629) But he urges us to accept his account on the implausible ground that history books (which we believe to be factual) are much more vivid to read than novels (which we are well aware are Wction) (T, 97) Some of the diYculties in Hume’s account of vivacity as a mark of belief are internal to his system We observe his embarrassment at discovering a perception that is neither quite an idea nor quite an impression We may 153