7* PROBLEMS IN MEANING
and the like; or, to express the matter otherwise, the concept of truth as artistic insight is perhaps misleading, since it does not sufficiently take into account the foregoing linguistic distinction. A specific illustration will make this clear.
In an essay, "On the Semantics of Poetry," 135 Philip Wheelwright opposes Richards' distinction between referential and emotive discourse. He grants to begin with, however, that some distinction should be made between what he calls 'literal and poetic statement/' between uses of language which are "purely referential" and those which are "to some degree reflexive and evocative." He then asks and answers an important question: do the poetic statements "involve any sort of truth and falsity at all? I think they do, and it is important to see in what sense this is possible; in what sense there can be such a thing as poetic truth, distinguishable from and unexchangeable with any strictly logical truth." 136 His argument then urges that to consider poetry in the light of Richards' distinction between literal statements, which are primarily referential and verifiable, and pseudo or poetic ones, which are primarily emotive and unverifiable, is a mistake. He suggests, rather, that the differences between the two kinds of language "are differences of what may be called as-sertive weight. A literal statement asserts heavily; it can do so because its terms are solid. A poetic statement, on the other hand . . . has no such solid foundation, and affirms with varying degrees of lightness." 137 Now this distinction, which Wheelwright elaborates most interestingly, may be138 useful and important for criticism; but I would ask: does it in any
135. The Kenyon Review^ Summer, 1940, pp. 263-283.
136. Ibid,, pp. 270-271.
138. "May be/' since I believe that Wheelwright's distinction requires further explanation and considerable qualification. Can all dramatic poetry be said to "assert lightly"? Is the content of many Shakesperean soliloquies in any sense "fragile"?