EXPLANATION AND CONTENT
I begin with a category of content-giving sentences for a content-noun. Here are some examples:
(1) The reason John's symptoms were relieved is that he took medicine M.
The excuse was that Sam got sick.
The danger in climbing the Matterhorn in the afternoon is that the ice will melt.
The penalty for trespassing is that the person convicted will be subject to a $1,000 fine.
These sentences contain content-nouns (ones in italics) together with that-clauses, or, more generally, nominals, that give content to the noun. In sentence (1) “reason” is a content-noun and “that he took medicine M” is a nominal that gives the content of the reason. Content-nouns are abstract nouns (“reason,” “explanation,” “excuse,” “danger”) rather than nouns for physical objects, properties, or events. They can be used to form content-giving sentences, such as ones above, with this form:
The + content-noun + {prepositional phrase} + form of to be + nominal {that-phrase}
The nominals are noun phrases with a verb or verb derivative. They include that-clauses, infinitive phrases (“the purpose of the flag is to warn drivers of danger”), and many others.
Content-giving sentences are to be contrasted with ones that contain a content-noun but do not give a content to that noun. For example,
The reason John's symptoms were relieved is difficult to grasp.
The excuse was unacceptable.
The penalty for trespassing is severe.
These sentences do not say what the reason, excuse, or penalty is.
A content-giving sentence such as (1) above may be equivalent in meaning to another, such as,
(2) The reason John's symptoms were relieved is his taking medicine M.
I shall say that sentences (1) and (2) express the same proposition, and that this proposition is a content-giving proposition for a concept expressed by the noun “reason.”
Sentences (1) and (2) contain an answer to the question “Why were John's symptoms relieved?” That answer is given by the content of the reason (“that he took medicine M”).
The proposition expressed by sentences (1) and (2) above will be said to be a content-giving proposition with respect to that question.A question such as “Why were John's symptoms relieved?” presupposes various propositions, for example,
John had symptoms.
His symptoms were relieved.
(3) John's symptoms were relieved for some reason.
A complete presupposition of a question is a proposition that entails all and only the presuppositions of that question. Of the three propositions just given, only (3) is a complete presupposition of the question “Why were John's symptoms relieved?” Sentence (3) can be transformed into a complete answer form for the question “Why were John's symptoms relieved?”:
(4) The reason that John's symptoms were relieved is____
by dropping “for some reason” in (3) and putting the expression “the reason that” at the beginning and “is” followed by a blank at the end to yield (4). (For more details and a generalization of this to various types of questions, see my The Nature of Explanation, 28ff.)
We can now say that
p is a complete content-giving proposition with respect to question Q if and only if (a) p is a content-giving proposition (for a concept expressed by some noun N); (b) p is expressible by a sentence obtained from a complete answer form for Q (containing N) by filling in the blank; (c) p is not a presupposition of Q.
Finally, Q will be said to be a content-question if and only if there is proposition that is a complete content-giving proposition with respect to Q. By these definitions, (1) and (2) express complete content-giving propositions with respect to the content-question “Why were John's symptoms relieved?”
If Q is a content-question (whose indirect form is q), and p is a complete content-giving proposition with respect to Q, then p provides an explanation of q. Thus the proposition expressed by
(1) The reason John's symptoms were relieved is that he took medicine M provides an explanation of why John's symptoms were relieved.
Any explanation that is offered is, or can be transformed into, one that answers some content-question by supplying a complete content-giving proposition with respect to that question. Thus, an explanation of the relief of John's symptoms that appeals to his taking medicine M can be construed as answering the content-question “Why were John's symptoms relieved?” by furnishing a complete content-giving proposition with respect to that question, namely, (1). The latter provides an explanation of the relief of John's symptoms.The following simple condition holds for correctness of such explanations:
(5) If p is a complete content-giving proposition with respect to Q, then p provides a correct explanation of q if and only if p is true.
So, for example, since (1) expresses a complete content-giving proposition with respect to the content-question
Q: Why were John's symptoms relieved?
Proposition (1), assuming it is true, provides a correct explanation of why John's symptoms were relieved. Condition (5) for correctness of explanations has considerably more bite than it might seem. For example, compare (1) with
(6) John took medicine M.
Even if (6) is true, that will not suffice to guarantee that (6) provides a correct explanation of why John's symptoms were relieved. (His symptoms might not have been relieved because he took M.) (6) is not a complete content-giving proposition with respect to Q. By contrast, if (1) is true, then, since it does express a complete content-giving proposition with respect to Q, it provides a correct explanation of q.
A correct explanation may or may not be a (particularly) good one. Nor must a good explanation be correct. Goodness in explanations is a broader concept than correctness, and unlike the latter is context-depen- dent.[22] Whether (1) provides a good explanation of why John's symptoms were relieved depends on the “appropriateness” of the answer it provides, which is determined by the knowledge and interests of those for whom the answer is provided.
If the context is one in which the intended audience already knows that medicine M is the agent that produced relief but not how it did so, then although (1) provides a correct explanation of why John's symptoms were relieved, it is not a good one. Moreover, there are contexts of evaluation in which correctness is not required. Explanations such as the Ptolemaic or Newtonian ones of the observed motions of the planets may be good ones in virtue of their comprehensiveness, precision, or predictive qualities, even though they are incorrect. What is needed for the concept of (potential and veridical) evidence is not a contextual concept of good explanation but a noncontextual one of correct explanation.In section 8 I indicated three criteria that need to be satisfied by any concept of correct explanation employed for potential evidence. First, it must be objective rather than subjective. Whether h correctly explains e cannot depend upon what anyone knows or believes. Second, it must be noncontextual. It cannot depend upon standards of goodness appropriate in one context of evaluation but not another. Third, the concept must not itself be understood in terms of evidence. I maintain that the concept of correct explanation given in (5) satisfies these conditions.
10.