But we need not stop here. Attention may be fully directed to the subject of smells. Smells may become the object of knowledge. They may take, _pro tempore_,[14] the place which the rose formerly occupied. One may, that is, observe the cases in which odors mean other things than just roses, may voluntarily produce new cases for the sake of further inspection, and thus account for the cases where meanings had been falsified in the issue; discriminate more carefully the peculiarities of those meanings which the event verified, and thus safeguard and bulwark to some extent the employing of similar meanings in the future. Superficially, it may then seem as if odors were treated after the fashion of Locke’s simple ideas, or Hume’s “distinct ideas which are separate existences.” Smells apparently assume an independent, isolated status during this period of investigation. “Sensations,” as the laboratory psychologist and the analytic psychologist generally studies them, are examples of just such detached things. But egregious error results if we forget that this seeming isolation and detachment is the outcome of a deliberate scientific device--that it is simply a part of the scientific technique of an inquiry directed upon securing _tested_ conclusions. Just and only because odors (or any group of qualities) are parts of a connected world are they signs of things beyond themselves; and only because they are signs is it profitable and necessary to study them _as if_ they were complete, self-enclosed entities.