In connexion with this theory a curious charge has been brought forward by Coleridge, who says, "In consulting the excellent commentary of St. Thomas Aquinas, on the Parva Naturalia of Aristotle, I was struck at once with its close resemblance to Hume's essay on association. The main thoughts were the same in both. The _order_ of the thoughts was the same, and even the illustrations differed only by Hume's occasional substitution of more modern examples. I mentioned the circumstance to several of my literary acquaintances, who admitted the closeness of the resemblance, and that it seemed too great to be explained by mere coincidence; but they thought it improbable that Hume should have held the pages of the angelic doctor worth turning over. But some time after, Mr. Payne of the King's Mews, showed Sir James Mackintosh some odd volumes of St. Thomas Aquinas, partly perhaps from having heard that Sir James, (then Mr. Mackintosh,) had in his lectures passed a high encomium on this canonized philosopher, but chiefly from the fact that the volumes had belonged to Mr. Hume, and had here and there marginal marks and notes of reference in his own handwriting. Among these volumes was that which contains the Parva Naturalia, in the old Latin version, swathed and swaddled in the commentary aforementioned." On this, Sir James Macintosh says, that "the manuscript of a part of Aquinas, which I bought many years ago, (on the faith of a bookseller's catalogue,) as being written by Mr. Hume, was not a copy of the commentary on the _Parva Naturalia_, but of Aquinas's own _Secunda Secundæ_; and that, on examination, it proves not to be the handwriting of Mr. Hume, and to contain nothing written by him."[287:1] So much for the external evidence of plagiarism. With regard to the internal evidence, the passage of Aquinas particularly referred to, which will be found below,[287:2] refers to memory not imagination; to the recall of images in the relation to each other in which they have once had a place in the mind, not to the formation of new associations, or aggregates of ideas there; nor will it bring the theories to an identity, that, according to Hume's doctrine, nothing can be recalled in the mind unless its elements have already been deposited there in the form of ideas, because the observations of Aquinas apply altogether to the _reminiscence_ of aggregate objects. But the classification is different: for Hume's embodies cause and effect, but not contrariety; while that of Aquinas has contrariety, but not cause and effect. In a division into three elements, this discrepancy is material; and, without entering on any lengthened reasoning, it may simply be observed, that the merit of Hume's classification is, that it is exhaustive, and neither contains any superfluous element, nor omits any principle under which an act of association can be classed.