Hume Studies

236 REMEMBERING THE PAST In his recent article "'Lively' Memory and 'Past' Memory," Oliver Johnson argues that the remarks in Treatise I. iii. 5 entail that Hume rejects all mnemonic knowledge of the past and that "we must drop the concept of the past, as it is ordinarily understood as something that once existed, from our account of memory." While I shall grant that Hume could claim no knowledge of the past, I do not believe Johnson has shown that Hume rejects the past from his account of memory. In this paper I argue that Johnson misconstrues the relationship between Hume's two 'criteria' for distinguishing between ideas of memory and ideas of the imagination. I also show that if Johnson's account were correct, much of Hume's philosophy would be incoherent. Johnson begins by claiming that in Treatise I.i.3, Hume introduces two criteria for distinguishing between ideas of the memory and ideas of the imagination , namely, a criterion based on the greater force and liveliness of the idea of memory vis-a-vis the idea of imagination ('lively' memory) and a criterion based on the correspondence of the idea of memory to the impression from which it was derived ('past' memory). After commenting that these criteria follow one's common sense beliefs regarding memory (LMPM 345-346), Johnson suggests that Hume came to realize that there is a problem with his initial characterization of the distinction. He sets the problem in the form of a question: How do we go about distinguishing a memory-idea from an imagination idea in practice? Or, to sharpen the problem. How are we able to tell when we are really remembering some past 237 event instead of only imagining that we remember it? The answer to this question would seem to be simple. Since we have at our disposal two marks by which to distinguish memoryideas from imagination-ideas, all that we need to do is to determine whether the idea in question satisfies one of them. If it satisfies either, it is a memory-idea; if it satisfies neither, it must be an imaginationidea . Furthermore, since every memory-idea satisfies both criteria, it should be immaterial to which we appeal in making our test. (LMPM 346) Johnson goes on to show that in Treatise I. iii. 5 Hume claims that the criterion for past memory "is not sufficient to distinguish them [ideas of the memory from ideas of the imagination] in their operation, or make us show the one from the other; it being impossible to recai the past impressions, in order to compare them with our present ideas, and see whether their arrangement be exactly similar," which he takes to mark Hume's rejection of the criterion of past memory (LMPM 347). If a criterion is a principle that can be used to divide a class of objects into two subclasses without remainder, then, as Johnson presents them, both of Hume's 'criteria' are inadequate because each is too broad. The 'criterion' for past memory is concerned with the truth of an idea (cf. T 448 and 458), and, as such, it is equally applicable to ideas of memory and the imagination. For example, a historian 's reconstruction of a past event might be true, but the reconstruction is certainly an idea of the imagination, not an idea of the memory. Similarly, the 'criterion' for lively memory ('force and vivacity') is used to distinguish impressions from ideas (T 1) and beliefs regarding events of any temp- 238 oral location from ideas that are not believed (T 96) as well as ideas of the memory from ideas of the imagination. Thus, neither 'criterion' can singularly mark the distinction between ideas of the memory and ideas of the imagination. But if there is a problem with inadequate criteria, it is not clear that the problem is Hume's, for Hume did not claim that the 'criteria' can operate independently of one another. In Treatise I.i.3, Hume first discusses the 'criterion' for lively memory (T 8-9), and then he indicates that "There is another difference betwixt these two kinds of ideas, which is no less evident," viz. . that...