On the Rightness of Rendering
A profound theme in the work of Nelson Goodman is the attempt to go beyond a too-narrow conception of truth to a criterion of rightness of rendering as the standard for our descriptions of an unapproachable reality. Beginning his contrast of truth, validity, and rightness with the well understood case of deductive logic, Goodman notes: (in chapter 7 of Ways of World Making)
Among the most explicit and clearcut standards of rightness we have anywhere are those for validity of a deductive argument; and validity is of course distinct from truth in that the premisses and conclusions of a valid argument may be false…. A deductive argument is right in a fuller sense only if the premisses are true and the inferences valid. Thus rightness of deductive argument, while involving validity, is still closely allied with truth.
He then proceeds to the more difficult issue of induction.
Now consider inductive validity. Here again neither truth of premises nor truth of conclusions is required; and inductive like deductive validity consists of conformity with principles that codify practice. But inductive validity is one step further removed from truth than is deductive validity; for valid inductive inference from true premisses need not yield a true conclusion. On the other hand, while inductive RIGHTNESS like deductive rightness does require truth of the premisses as well as validity, it also requires something more. To begin with, a right inductive argument must be based not only on true premisses but upon all the available genuine evidence…. No parallel requirement is imposed upon a deductive argument…
Still, inductive rightness is not fully characterized as inductive validity plus use of all examined instances…. Even if all examined emeralds have been grue [note 1], still inductive argument to the hypothesis that all emeralds are grue is wrong. Inductive rightness requires evidence statements and the hypothesis to be in terms of “genuine” or “natural” kinds — in my terminology, to be in terms of projectible predicates like “green” and “blue” rather than in terms of non-projectible predicates like “grue” and “bleen.” Without such a restriction, right inductive arguments could always be found to yield countless conflicting conclusions….
In sum, then, inductive rightness requires that the argument proceed from premisses consisting of all such true reports on examined instances as are in terms of projectible predicates. Thus inductive rightness, while still demanding truth of predicates, makes severe additional demands. And although we hope by means of inductive argument to arrive at truth, inductive rightness unlike deductive rightness does not guarantee truth…. Any feasible justification of induction must consist rather of showing that the rules of inference codify inductive practice — that is, of effecting a mutual adjustment between rules and practice — and of distinguishing projectible predicates or inductively right categories from others.
This brings us, then, to the question what are inductively right categories, and so to a third kind of rightness in general: rightness of categorization. Such rightness is one step further removed from truth; for while deductive and inductive rightness still have to do with statements, which have a truth value, rightness of categorization attaches to categories or predicates — or systems thereof — which have no truth-value….
In such a context, I am not so much stating a belief or advancing a thesis or a doctrine as proposing a categorization or scheme of organization, calling attention to a way of setting our nets to capture what may be significant likenesses and differences. ARGUMENT FOR THE CATEGORIZATION, the scheme suggested, COULD NOT BE FOR ITS TRUTH, SINCE IT HAS NO TRUTH-VALUE, BUT FOR ITS EFFICACY IN WORLD MAKING AND UNDERSTANDING. An argument would consist rather of calling attention to important parallels between pictorial representation and verbal denotation, of pointing out obscurities and confusions that are clarified by this association, of showing how this organization works with other aspects of the theory of symbols. FOR A CATEGORICAL SYSTEM, WHAT NEEDS TO BE SHOWN IS NOT THAT IT IS TRUE BUT WHAT IT CAN DO. Put crassly, what is called for in such cases is less like arguing than selling.
Our descriptions of reality are an imposition WE make. Our investigations may disturb whatever might have been there otherwise. As evolved creatures, we are limited by the experience of our species to ideas at a range of sensation and conception which is effective but very partial… We can only hope our descriptions go with the grain of an unapproachable reality. I try to offer a fair sample of empirical detail to show how my vision of mind applies to behavior and how it can be applied powerfully to the comprehension of human learning. The question is less whether my description of mind is true or false than whether or not I make it work on important problems.