EXPLAINING REVISITED

Returning to the concept of an illocutionary act of explaining, we are now in a position to formulate some further conditions. So far no restric­tions have been imposed on the indirect question q being explained, or on u, what is uttered by the explainer.

Consider once again the question

(1) How high is the Matterhorn?

I claimed that someone could not be said to understand how high the Matterhorn is simply in virtue of the fact that he knows of the proposition

(2) The height of the Matterhorn is 14,700 feet

that it is a correct answer to (I). Now I suggest that an analogous claim can be made for explaining: By uttering (2) a speaker S is not explaining how high the Matterhorn is. Unless (1) is being construed as elliptical for something else, it is not the right sort of question to fill the q-position in “S explains q by uttering u.” Still, intentions being what they are, it seems possible for some misguided speaker S to utter (2) with the intention de­scribed in condition (3) for explaining given in section 1: that of rendering q understandable (Q = (1)) by getting others to know of the proposition expressed by (2) that it is a correct answer to Q.

Let us call Q a content-question if and only if (Kp)(p is a complete content­giving proposition with respect to Q). A content-question has a complete answer form whose blank can be filled with content-giving expressions that will transform the result into a sentence that expresses a content-giving proposition. A question such as “Why did Nero fiddle?” is a content-question, but one such as “How high is the Matterhorn?” is not. Now I suggest that the q-position in “S explains q by uttering u” can be filled only by interrog- atives expressing content-questions, or by interrogatives that in “S explains q by uttering u” are elliptical for ones expressing content-questions. Since

(1) is not a content-question, if, by uttering (2), S is said to be explaining how high the Matterhorn is, what is said is either false or else elliptical for something in which the explanatory question is a content-question.

I turn next to a proposal for a restriction on u, what is uttered in an act of explaining q.

u-restriction for Q: what S utters is, or in the context is transformable into, a sentence expressing a complete content-giving proposition with respect to Q.

Suppose that S explains why Nero fiddled, by uttering

u: Nero was happy.

In the context of S's utterance u is transformable into (what S said in that context can also be expressed by) the sentence

The reason that Nero fiddled is that he was happy,

which expresses a complete content-giving proposition with respect to the question “Why did Nero fiddle?”

The u-restriction will allow us to exclude a number of kinds of cases. For example, with it S cannot be explaining how Jones escaped from prison, by uttering

The reason that Nero fiddled is that he was happy

(unless, in the context of utterance, this is to be understood as expressing some different proposition). Nor can S be explaining why Nero fiddled,

by uttering “The reason that Nero fiddled is difficult to grasp,” or nonsense words like “glip glop.”

However, the u-restriction still allows a broad range of cases.

The reason that atoms emit discrete radiation is that God is love, provided, of course, that the other conditions on explaining are satisfied (in particular, e.g., that S believes that this sentence expresses a correct answer to the question). Will the u-restriction countenance explaining acts involving utterances even more bizarre than this? That depends upon which u's we take as expressing propositions.

I construe propositions broadly to be expressible by that-clauses following a range of psychological verbs (such as “believe,” “fear,” and “hope”) and illocutionary verbs (such as “say,” “state,” “propose,” and “suggest”). Such verbs can be followed by that-clauses that are gram­matically or semantically deviant, even though the resulting sentence is neither, and indeed is true, e.g.,

Heidegger believed that the nothing noths,

John said that numbers speak silently.

I shall say that the emphasized words in these sentences express proposi­tions. More generally, u can be said to express a proposition if and only if it can appear in (many)⁵ contexts of the form

(3) Subject term + psychological or illocutionary verb + that + u.

(Here the “that” is to be constructed as associated with the verb, not as mod­ifying u.) Although by this criterion syntactically and semantically deviant utterances can express propositions, not all utterances can. Thus, since

John said (believes, hopes) that glip glop,

John said (believes, hopes) that go table chair,

are deviant, the emphasized words do not express a proposition. (“John said ‘glip glop') (or ‘go table chair')” is not deviant, but this is not the relevant context.)

To adopt this criterion is not to be committed to the view that every true sentence of form (3) requires the subject to have de re knowledge (by acquaintance) of the proposition expressed by u. If Sam, who knows no physics, is told by a reputable physicist that strangeness is conserved in strong interactions, then it may be true to say that

(4) Sam believes that strangeness is conserved in strong interactions,

5. Obviously, some such contexts will be inappropriate, e.g., “S predicts that yesterday it snowed.”

which is a sentence of form (3).

(4) true. If Sam understood physics, what could make (4) true is his belief, of the proposition in question, that it is true—which involves acquain­tance with the proposition itself. Accordingly, even though propositions are expressible by that-clauses in sentences reporting beliefs (etc.), this does not necessitate an analysis of belief sentences that requires the believer to be acquainted with some proposition.

Nevertheless, if sentences such as “The nothing noths” and “Numbers speak silently” do express propositions, then one ought to be able to have de re knowledge by acquaintance with the propositions they express. But how is this possible if these sentences are meaningless? In section 2 (in ma­terial omitted here), I suggested that to be acquainted with a proposition p is to be in a certain mental state that can be functionally characterized by typical causes and effects. Typically, such a state is caused by having seen or heard certain sentences that express p whose meaning one knows, and it results in the ability to produce the same or equivalent sentences, of which one knows the meaning and knows that they express p. Now, just as “proposition” is being construed broadly, so is “meaning.” Not every syntactically or semantically deviant sentence is utterly without meaning; there are degrees of deviation. A sympathetic reader of Heidegger knows something about the sentence “The nothing noths” that the nonreader does not. Within the Heideggerian corpus it has some meaning (I am inclined to suppose) despite its deviance, and the sympathetic reader knows what this meaning is.

With the present understanding of what can express a proposition, the u-restriction for Q allows S to be explaining by making deviant utterances. Heidegger can be explaining something by uttering a sentence of the form

The reason that____ is that the nothing noths.

And John can be explaining why numbers cannot be heard, by uttering

The reason that numbers cannot be heard is that numbers speak silently.

However, John cannot be explaining why numbers cannot be heard, by uttering

The reason that numbers cannot be heard is that glip glop.

Thus, although the u-restriction (in conjunction with a criterion per­mitting deviant sentences to express propositions) does exclude various utterances from explaining acts, it is sufficiently broad to allow some that violate rules of syntax or semantics. But this is as it should be. At least in certain cases we do, I think, want to describe the philosopher, the scientist, or the nonspecialist as engaging in an act of explaining, even if we criticize the act, the explanation, or both for insufficient intelligibility. Some will insist that we have allowed too broad a class of explaining acts. Perhaps, but while this may be somewhat arbitrary, I am inclined to stop at this point and treat further restrictions as proposals for evaluating explaining acts and their products.

We may now formulate the resulting conditions for explaining, as follows:

“S explains q by uttering u” is true if and only if either

(a) Q is a content-question, the u-restriction for Q is satisfied, and condition (3) of section I holds (i.e., S utters u with the intention that his utterance of u render q understandable by producing the knowledge, of the proposition expressed by u (in that context), that it is a correct answer to Q): or

(b) “S explains q by uttering u” is elliptical for “S explains q' by uttering u,” Q' is a content-question, the u-restriction for Q' is satisfied, and condition (3) obtains with respect to Q'.

The parenthesized words “in that context” in (a) allow S to explain q by uttering words that do not normally express a proposition, but do so in the explaining context. Thus, S might explain why Othello killed Desdemona by uttering simply “uncontrollable jealousy”—which, in the context, can be taken to express the proposition that Othello was uncon­trollably jealous.

Just as the u-restriction permits a wide range of explanatory utterances, though not every concatenation of words, so the restriction of the inter­rogative to one expressing a content-question also permits a wide range of interrogatives, but not just any. It permits interrogatives with false

presuppositions. (Mary can explain why helium is the lightest element.) Indeed, it permits ones with presuppositions expressed by sentences that are grammatically or semantically deviant. (Heidegger can explain why the nothing noths.) But not every interrogative is possible. As noted, one cannot explain how high the Matterhorn is (unless this is elliptical for something else). Nor can one explain why go table chair, since nothing of the form “the reason that go table chair is that ” expresses a propo­

sition, which is required for q to express a content-question.^[58]

6.