In this paper I will argue that Professor Goodman was correct in thinking that there is a problem concerning counterfactual conditionals, but that it is somewhat different from the problem he thought it to be, and is one that is even more basic. I will also try to show that this problem is distinct from Hume's “problem” of induction, and that additional assumptions have to be made for counterfactual induction beyond those required for other kinds of induction.

Through an analysis of conditions under which the question of spatial anisotropy can be raised, the present paper brings out intimate conceptual relationships between the scientific concept of space and the concepts of entities, behavior, and explanation specified by scientific theories. Thus scientific departures from ordinary usage (or from usage in other scientific theories) of the term “space” entail corresponding shifts in the use of other terms not generally seen to be connected. As a case study of the relations between these terms, and of the refusal to allow the possibility of spatial anisotropy, Newtonian mechanics is examined.

It is common knowledge that the notion of the unconscious is an essential part of psychoanalytic theory. In recent years, however, Arthur Pap and A. C. MacIntyre have argued that Freud's theory of the unconscious is not explanatory. But a close examination of Pap's and MacIntyre's arguments reveals that they are invalid. If one wishes to show that the theory of the unconscious is unexplanatory, different arguments will be necessary.

A certain dilemma is inherent in relational accounts of space and time. If any objects endure through change, then temporal elements other than relations are required to describe them. If, on the other hand, no objects endure through change, no permanent reference system is available in terms of which to define the “same place” at different times. An argument which, by exploiting this latter difficulty, attempts to show that “objects with some endurance through time” must be accepted as fundamental is examined and found inconclusive. A sketch is then given of an alternative scheme which does allow the relevant spatial comparisons, but which does not countenance the reidentification of particulars. The discussion is intended to show that the relationist can, as indeed he must, deny the second horn of this delemma.