Thus the colours red and blue and the odour of a rose, are simple impressions; while the ideas of blue, of red, and of rose-odour are simple copies of these impressions. But a red rose gives us a complex impression, capable of resolution into the simple impressions of red colour, rose-scent, and numerous others; and we may have a complex idea, which is an accurate, though faint, copy of this complex impression. Once in possession of the ideas of a red rose and of the colour blue, we may, in imagination, substitute blue for red; and thus obtain a complex idea of a blue rose, which is not an actual copy of any complex impression, though all its elements are such copies. Hume has been criticised for making the distinction of impressions and ideas to depend upon their relative strength or vivacity. Yet it would be hard to point out any other character by which the things signified can be distinguished. Any one who has paid attention to the curious subject of what are called "subjective sensations" will be familiar with examples of the extreme difficulty which sometimes attends the discrimination of ideas of sensation from impressions of sensation, when the ideas are very vivid, or the impressions are faint. Who has not "fancied" he heard a noise; or has not explained inattention to a real sound by saying, "I thought it was nothing but my fancy"? Even healthy persons are much more liable to both visual and auditory spectra--that is, ideas of vision and sound so vivid that they are taken for new impressions--than is commonly supposed; and, in some diseased states, ideas of sensible objects may assume all the vividness of reality. If ideas are nothing but copies of impressions, arranged, either in the same order as that of the impressions from which they are derived, or in a different order, it follows that the ultimate analysis of the contents of the mind turns upon that of the impressions. According to Hume, these are of two kinds: either they are impressions of sensation, or they are impressions of reflection. The former are those afforded by the five senses, together with pleasure and pain. The latter are the passions or the emotions (which Hume employs as equivalent terms). Thus the elementary states of consciousness, the raw materials of knowledge, so to speak, are either sensations or emotions; and whatever we discover in the mind, beyond these elementary states of consciousness, results from the combinations and the metamorphoses which they undergo. It is not a little strange that a thinker of Hume's capacity should have been satisfied with the results of a psychological analysis which regards some obvious compounds as elements, while it omits altogether a most important class of elementary states.