INSTRUCTIONS

In section 3 the following condition was proposed as a necessary one for understanding: A understands q only if

(1) (0p)(A knows of p that it is a correct answer to Q, and p is a complete content-giving proposition with respect to Q).

Is this also sufficient for understanding?

What complicates the issues is that often a question can be correctly answered in different ways by providing various kinds and amounts of information. A person might be said to understand q in one way but not another. This idea can be explicated by introducing the concept of instruc­tions for a question.

Consider the question

(2) What caused Smith's death?

Each of the following, let us assume, is a correct answer to (2):

(a) The cause of Smith's death was his contracting a disease;

(b) The cause of Smith's death was his contracting a disease involving a bacterial infection;

(c) The cause of Smith's death was his contracting legionnaire's disease.

Someone who replies to (2) in one of these ways may be following certain instructions pertaining to (2), e.g.,

la: Say in a very general way what caused Smith's death, e.g., whether it was caused by contracting a disease, or by some accident that befell him, or by an act of suicide;

Ib: Follow la, and if a disease is cited indicate something about what it involves, e.g., whether it is bacterial or viral;

Ic: Follow la, and if a disease is cited identify it by using a common name for it.

Instructions are rules imposing conditions on answers to a question. They govern the proposition that is the answer, and the act of answering only insofar as they do this. (“Answer in a low voice” imposes a condition on the act of answering, but not on the proposition that is the answer.) Various instructions are generally possible for one and the same question. Some instructions will be vague, some precise.

(2), but does not know that (b) and (c) are, can be said to understand q in a way that satisfies instructions la but not Ib or Ic.

More generally, utilizing (1), we can say that A understands q in a way that satisfies instructions I, or, more briefly, A understands q_l, only if

(3) (0p)(p is an answer to Q that satisfies instructions I, and A knows of p that it is a correct answer to Q, and p is a complete content-giving proposition with respect to Q).

For example, A has fulfilled this condition with respect to question (2) and instructions Ia, if A knows of proposition

(a) The cause of Smith's death was his contracting a disease

that it is a correct answer to

(2) What caused Smith's death?

An answer to question (2) will be said to satisfy instructions Ia if and only if that answer says (and does not merely purport to say) in a very general way what really did cause Smith's death, i.e., if it cites a true cause. In such a case, if the instructions have been satisfied then question (2) has been correctly answered. But consider the following instructions for (2):

Id: Give an answer that George accepts.

An answer to (2) satisfies Id if and only if it is an answer to (2) that George accepts. The satisfaction of these instructions, unlike the satisfaction of la, does not guarantee that (2) has been correctly answered. In general, some, but not all, questions and instructions are such that if the instructions are satisfied the questions will have been correctly answered. Of course, because of false assumptions, inconsistencies, excessive vagueness, or just sheer irrelevance, some instructions cannot be satisfied at all with respect to a given question. In such a case, understanding q in a way that satisfies I will be impossible.

In section 2, I said that A is in a knowledge-state with respect to Q if ($p)(A knows of p that it is a correct answer to Q). If condition (3) is satisfied—i.e., if A knows of some complete content-giving proposition with respect to Q which satisfies I that it is a correct answer to Q—I shall say that A is in a complete knowledge-state with respect to Q. Being in such a state, I am claiming, is at least a necessary condition for understanding q_r By this criterion, if A knows of (a) (which is a complete content-giving proposition with respect to (2) that satisfies Ia) that it is a correct answer to (2), then A fulfills a necessary condition for understanding what caused Smith's death, in a way that satisfies instructions Ia.

Can a claim of the form

(4) A understands q,

where there is no explicit appeal to any instructions, be construed in terms of understanding q1? Three possibilities suggest themselves. First, it might be that (4) is true if and only if A understands q in a way that satisfies some instructions or other, i.e.,

is a set of instructions for Q, and A understands q,).

But this would render claims about understanding very weak, since their truth would then require only a complete knowledge-state with respect to an answer satisfying the weakest instructions for Q. My knowing that Smith's death was caused by his contracting a disease would always suf­fice to say that I understand what caused his death. But it seems doubtful that we would want to say that such knowledge is always sufficient for understanding.

This leads to the second proposal, with is that when a speaker utters a sentence of form (4) his claim is always elliptical for

A understands q,

where I is some contextually implicit set of instructions intended by the speaker, and I may vary from one context of utterance to another. Sentences of form (4) are, I think, sometimes used in this way. We may say that some­one understands something, meaning that he understands it in a way that we have in mind.

There is, however, another more likely possibility. Suppose that I hear speaker S assert (4). I may not know what instructions, if any, the speaker S intended, and this may not be clear from the context of his utterance. Moreover, I may have no idea what particular instructions A's under­standing satisfies. Still it seems possible for me to assert (4), because I believe that A understands q in some way that is appropriate—even if I do not know what this is.

This leads to the third suggestion. When a speaker utters (4) his claim is elliptical for

(5) ($I)(A understands q_p and I is a set of appropriate instructions for Q).

Let me call this the “appropriate-instructions” use of (4). It is, I suggest, the most typical use of (4). Various views are possible about how to de­cide whether I is a set of appropriate instructions for Q. On one, there are universal standards of appropriateness, at least in science. On another, the standards of appropriateness can vary, depending on contextual features of A's situation. (See chapter 7 for an account of the latter.)

Recall now that A is said to be in a complete knowledge-state with respect to Q1 provided that

(3) ($p)(p is an answer to Q that satisfies instructions I, and A knows of p that it is a correct answer to Q, and p is a complete content-giving proposition with respect to Q).

If we use (5)—the appropriate-instructions use of “understand”—then on the basis of (3) we may conclude that

A understands q only if is a set of appropriate instructions for Q, and A is in a complete knowledge-state with respect to Q₁).

On the appropriate-instructions use, then, knowing of a complete con­tent-giving proposition that it is a correct answer to Q is not sufficient for (nonrelativized) understanding; the proposition known must also satisfy appropriate instructions for Q.