Searle on the Is/Ought Dichotomy
May 4, 2014
Stuart W. Mirsky in Ethics, Ethics, John Searle, Moral Philosophy, Naturalism, Prescriptivism, R. M. Hare, Stuart Mirsky

In continuing to review what's left of my library post Sandy's flood in our region, I came across a small paperback, Theories of Ethics, edited by Philippa Foot. I did remember reading this one and found, as I paged through it, plenty of handwritten notes on the book's pages. I almost never write in book margins. It just seems wrong to me. But I obviously did so at that point, probably reflecting my effort to develop an ethics theory of my own which back in the seventies I was very keen on attempting. Never quite succeeded at it, of course, and nowadays I am leery of any sort of theory development in a field like this for Wittgensteinian reasons. But back then it's apparent I had fewer inhibitions in the matter. Anyway, the book consists of a series of articles gathered, and commented on in a foreword, by Foot who was then the pre-eminent exponent of naturalism in ethics and regarded as a major thinker in the field. One of the pieces she re-printed, it turns out, was an article by our old friend John Searle, How to Derive Ought from Is. It's followed by a piece by R. M. Hare (another major ethical thinker of the time) attacking Searle's position. Hare, of course, was defending his own view that the function of commending is radically different from describing and that one achieves moral claims by combining commendatory principles, to which one chooses to subscribe, with factual assertions to yield logically sound conclusions which serve as particular moral oughts. Searle had offered a different view.

Searle argues, in essence, that one can derive ought from is claims if one takes into account how language actually works, i.e., it consists not only of descriptive statements asserting what he describes as brute facts about the world but also descriptive statements asserting institutional facts. In this latter case, he points out, the assertions one makes often serve a constitutive role. In the example he gives, promising, one's stating that "I promise to do X" not only describes what one is doing but constitutes it. Thus by making that very statement, when the circumstances are right , of course (obviously the statement in cases of stage acting, say, or demonstrating something to another, etc., would not be the right circumstances), one has entered the institution of promising. Thus, as with playing a game of chess, for one to promise something by such a statement, one must follow through and do what one has promised or else one wasn't really playing the game one's words implied at all.

Searle compares this to a baseball player caught out on base and ruled out by the umpire and told that he ought now to leave the field. Could the ballplayer reply that you cannot derive an ought from an is and so there was no reason for him to leave the field? Of course, that would be to step outside the game's rules (and invite others to carry one out, as Searle reminds us). So on this Searlean view, ought claims do follow from is claims in certain cases and he proposes that this explains (or at least helps to explain) the moral claims we make. Hare responds by pointing out that there is still a moral question as to whether one should play the game in question and maintains that it is this sense of buying into the game that matters. Thus a person can say "I promise to do X" and change his mind, or be insincere in having made the original statement. One still has to, according to Hare, subscribe to the principle which says that when one undertakes to say things that amount to playing such a game, one ought to do so sincerely and for Hare we cannot get there absent a decision to adopt a principle that universalizes the rule that one ought to act sincerely in such cases (because, on Hare's view, if one doesn't, then others won't where we are concerned).

Hare is right here but you can see it without going through the highly technical account Hare rests his case on in the follow-up article, I believe. One has only to look at Searle's argument to see the problem with it though, admittedly, it is somewhat subtle. Searle offers the following:

1) Jones uttered the words "I hereby promise to pay you, Smith, five dollars." (Descriptive)

This leads to the statement:

2) Jones promised to pay Smith five dollars. (Descriptive)

which implies the further statement that:

5) Jones placed himself under (undertook) an obligation to pay smith five dollars.(Descriptive)

and the follow-up statement:

4) Jones is under the obligation to pay Smith five dollars. (Descriptive)

which, Searle proceeds to assert, leads to the following conclusion that:

5) Jones ought to pay Smith five dollars. (Evaluative)

After further analysis and amplification of the above by addition of ceteris paribus clauses and statements which elaborate the assumptions within the four core statements preceding the conclusion, Searle asserts that the apparently evaluative conclusion, statement #5, is implicit in the preceding descriptive premises. Therefore, he says, we have a case in which an ought is effectively derived from a series of "is" claims. Moral valuing claims, he adds, typically fit into cases like this.

But it seems to me Searle has made an important mistake here. That is, his final statement #5 can be read both descriptively and evaluatively and, supposing that the evaluative variant has been implied by the descriptive statements that preceded it, effectively elides the two different meanings. As he did in his Chinese Room Argument against the possibility of a computational account of minds, in which he introduces a premise with terms that can be read in two ways, he has once again slipped into his argument a subtle equivocation.

In the case of the Chinese Room Argument he offered the statement that "Syntax does not constitute and is not sufficient for semantics" as a key premise in the argument in order to support its conclusion that computational operations which he describes as "syntactical" could never be sufficient to account for the condition of getting meanings (so-called semantics) that we experience when we understand anything. From this he drew the conclusion that computers (which are "syntactical" in their operations) could never do what brains do, i.e., never produce understanding and, thus, couldn't produce minds. His mistake in that case was to fail to note that "does not constitute" and "is not sufficient for" (two key terms in the premise in question) can both be read in more than one way.

While it's true that, given a case of what Searle called "syntax" (formal operations), we would not have enough to say of it that we also had a case of semantics (instances of understanding), it doesn't follow that it is not a certain type of syntactical operations that combine to produce the semantical feature we call understanding. That is, while syntax alone certainly does not constitute (i.e., is not the same as, is not what we mean by) "semantics," it might still be the case that certain arrangements of syntactical phenomena do produce (constitute) what is needed for instances of understanding to occur. Searle's elision of what a term designates with what the thing(s) it designates may do (because of our tendency to use the same words in specifying the two cases) led him to a flawed conclusion. He has done exactly the same thing in his is/ought argument.

In that case he takes the final statement that "Jones ought to pay Smith five dollars" as an evaluative claim in the sense that he believes that to make it one must also accept the terms of the game of promising so that making that statement implies that Jones has already accepted those terms. But, as Hare notes, one can make promising statements falsely or come to conclude that one erred in the promise and simply choose at a later date to abrogate the obligation one had previously asserted. Where is the obligation then that making the original statement is thought to have implied? Even if Jones was sincere at the time, as Hare notes, there is a further issue here, i.e., should Jones keep his word and why? If he can simply cast aside the asserted obligation he explicitly undertook with no bad consequences to himself, then on what grounds is he obliged not to do so?

Searle argues that Jones has freely entered into the institution of promising and that alone constitutes his obligation, on pain of being drummed out of the game, carried off the field. But what if one were tagged out in that baseball game but the umpire (and the rest of the players) didn't see it? Would Jones, on pain of stepping outside of the game and so ceasing to play it, be obligated to report the "out" to the umpire and voluntarily remove himself from the field? If playing a game means always playing by the rules, then cheating could never occur. But cheating does occur in games all the time and sometimes the cheater gets away with it. It is precisely this question of whether one ought to cheat or not that is at issue!

Searle doesn't notice that statement #5, "Jones ought to pay Smith five dollars," can be read descriptively, as a statement of the rules under which the transfer of funds was initially committed to, and that this descriptive reading differs from the evaluative assertion that Jones is, in fact, under an obligation he cannot shirk. The descriptive interpretation of #5, that Jones ought to pay back the funds, still doesn't imply anything about whether or not Jones ought to play by the rules of the game he has entered into. Hare's attack on Searle's account is correct.

Article originally appeared on Ludwig (
See website for complete article licensing information.