Monday, October 7, 2013

Ethical Statements are Ordinary Scientific Statements


1. Introduction
Ayer writes in Language, Truth and Logic, “in so far as statements of value are significant, they are ordinary ‘scientific’ statements; and in so far as they are not scientific, they are not in the literal sense significant, but are simply expressions of emotion which can be neither true nor false.”[1] Before attempting to demonstrate the nonsensical nature of non-scientific statements of value, Ayer attempts to rule out the former suggestion that such statements are scientific.
Ordinary scientific statements Ayer characterizes as both synthetic and verifiable. These criteria immediately exclude the “absolutist” view of ethics in which statements of value are held to be discoverable through some mysterious form of “intellectual intuition” rather than empirical observation. Such statements would be synthetic but not verifiable – and therefore not scientific.[2] The “naturalistic” view, as represented by subjectivist and utilitarian models, Ayer thinks is beset with difficulties all its own. The first objection he brings against such views is of the general form: since it not self-contradictory to say that some interest-gratifying thing, x [i.e., some pleasant thing, desired thing, approved of thing, etc.] is not good, or that some bad things are interest-gratifying [i.e., pleasant, desired, approved of, etc.], it cannot be the case that the sentence “x is good” is equivalent to “x is interest-gratifying.”[3] While I concede that this is the case for a large number of interest-based models, I deny that it is true of every possible model. My project in this paper, then, is to work out at least a sketch of how one might provide a reductive analysis of ethical statements that is as satisfactory as Ayer’s analysis of scientific statements, proceeding on the hypothesis that both classes of statements are synthetic and verifiable to the same degree.
The general account of ethical statements I will be working with goes as follows: To say that some act x is good, or better than some other act y, is to assert one’s belief that act x, when all relevant evidence has been taken into account, will prove to have been most conducive toward producing for the agent some intrinsically desirable outcome, e.g., eudaimonia or human happiness. With respect to this analysis of the ethical term “good,” I will simply push back against Ayer’s objection and affirm that it is, in fact, self-contradictory to call that action wrong that is most conducive toward producing the eudaimonic state. As I’ve already noted, if my account of ethical statements is accurate, then such statements would fall under Ayer’s broad definition of “scientific.” The bulk of this paper, then, aims to address the apparent differences between these and other scientific statements by providing alternative analyses of troublesome ethical statements.
But this is not quite yet sufficient to satisfy the logical positivist’s challenge. A further difficulty facing those who would seek to vindicate ethical statements as meaningful is to demonstrate not merely the possibility of a purely scientific analysis of ethical statements, but to show that the model that it suggests actually does underlie our existing ethical notions.[4] As Ayer writes, “we reject utilitarianism and subjectivism, not as proposals to replace our existing ethical notions by new ones, but as analyses of our existing ethical notions.”[5] What the logical positivist requires from us, therefore, is an analysis of ethical notions as they are actually employed in everyday language. Focusing our attention exclusively on the term “good,” what criteria might we employ to determine whether our analysis is properly connected with the “existing ethical notion”? Only once we are sure our analysis has the proper target in view can we attempt to show how such a notion might derive from our interactions with the empirical world.

2. A Reductive Analysis of Goodness
According to Stevenson, a satisfactory analysis of the term “good” must capture its “vital sense.” While this vital sense itself Stevenson has difficulty explicating, he takes it to involve several basic requirements – requirements that “appeal strongly to our common sense.” These he enumerates as follows: first, goodness must be a topic for intelligent disagreement; second, It must be “magnetic,” i.e., it must explain the connection between goodness and normativity; and third, it must not be discoverable solely through the scientific method.[6] We will now address each of these requirements in turn.

2.1 Goodness must be a topic for intelligent disagreement
In order to satisfy our common sense notion of the term, it is argued, we must be able to sensibly disagree about whether something is “good.”[7] One apparent weakness of interest-based ethical models, if you recall from Ayer’s first objection, is that two individuals arguing whether x is good, where “good” is taken to be synonymous with [some variation of] “interest-gratifying,” are not contradicting one another when they assert opposite judgments. For example, A’s assertion, “I desire x,” and B’s assertion, “I don’t desire x,” are not incompatible assertions, while A’s assertion, “x is good,” and B’s assertion, “x is not good,” are incompatible assertions. For two individuals’ disagreement to be a sensible one, argues Ayer, the locus of disagreement must be in the beliefs of each with respect to some common empirical fact(s) and not merely in their different attitudes concerning that fact or set of facts. The subject of ethical disagreement will be picked up again in section 3.3. There I will demonstrate with respect to differing ethical judgments that all apparent disagreement in interest is ultimately reducible to disagreement in belief.

2.2 It must be “magnetic.”
A second requirement to which our analysis is expected to conform if it is to satisfy common sense is that it be able to account for the normative force of ethical ascriptions. “A person who recognizes x to be ‘good’ must ipso facto acquire a stronger tendency to act in its favor than he otherwise would have had.”[8] With respect to interest-based theories that analyze goodness relative the native interests of the speaker, this requirement imposes no special difficulty.[9] While my account is of this general class, it will nonetheless require some argumentation to demonstrate its immunity to this type of objection.
According to my account, the term “good” may be used variably to mean either instrumentally good or intrinsically good. The intrinsically good will be given by the somewhat obscure title, eudaimonia. This designation is intended generally to capture a particular ideal state of being – a state of human “flourishing,” if you will – and is in principle assessable by entirely empirical criteria. Only people can be “good” in the intrinsic sense. The instrumentally good, then, is that which contributes toward the actualization of the eudaimonic state. Actions, objects, experiences and mental states (e.g., beliefs, desires, attitudes) might be called “good” in this second sense.
Generally speaking, then, all statements of the form “x is good” entail a corresponding normative imperative appropriately related to x. Since the vast majority of these statements concern things that are “good” in the instrumental sense, we will focus our analysis on instances of this type. Proceeding in this fashion, “x is good” roughly entails the following set of normative imperatives relative to the nature of x:
1)    Where x is some action, “x is good” roughly entails “A ought to perform x
2)    Where x is some object or experience, “x is good” roughly entails “A ought to desire to have or experience x
3)    Where x is some mental state, “x is good” entails “A ought to [believe x, desire x, or take such and such attitude toward x],” etc.

From here we may provide a complete analysis of one of the above to demonstrate our general technique. In (1) above, “A ought to perform action x,” is understood to mean, given A’s interest in attaining the eudaimonic state Z (which, again, is an objective matter: A’s interest is empirically given by Z), and provided the causal datum “if x, then Z,” then A is rationally obliged with respect to Z to perform action x. The one point I hope is apparent from this account is that normativity is at bottom a function of human rationality, not emotional caprice. 
Note also that this way of speaking is not unique to ethical contexts. Normative terms are also employed, I think in equivalent fashion, in situations in which a state of affairs is supposed to obtain, i.e., in virtue of its following from certain facts previously verified, when we are no longer in a position to verify those supporting facts as still obtaining. For example, A’s wife might ask him, “Are the keys in the car?” to which A might reply, “They should be.” What A means by his use of the term “should” in this context is that the keys being in the car follows from the evidence that is presently in his possession, but that he cannot now be certain that new facts which bear on his conclusion haven’t disrupted the integrity of his judgment (e.g., A’s son may have removed the keys upon returning from school). Normative terms, understood in this way, are simply affirmations of rational or logical entailment spoken in non-ideal epistemic situations, and are employed equivalently in both ethical and ordinary empirical contexts. 

2.3 It must not be discoverable solely through the scientific method.
A final requirement with which Stevenson thinks our analysis must comply to satisfy common sense is that “the ‘goodness’ of anything must not be verifiable solely by use of the scientific method.”[10] My address to this criterion will be slightly different than those previously dealt with. In the first place, I deny its primary assumption: my account recognizes no fundamental difference between ethical and ordinary scientific statements. Rather than address the requirement directly, therefore, I will address instead the argument Stevenson gives to justify it. The argument attacks interest theories, the strongest form of which Stevenson believes himself to be representing below:
Consider, for example, the definition: “x is good” means most people would approve of x if they knew its nature and consequences. How, according to this definition, could we prove that a certain x was good? We should have to find out, empirically, just what x was like, and what its consequences would be. To this extent the empirical method, as required by the definition, seems beyond intelligent objection. But what remains? We should next have to discover whether most people would approve of the sort of thing we had discovered x to be. This couldn’t be determined by popular vote – but only because it would be too difficult to explain to the voters, beforehand, what the nature and consequences of x really were.[11]

The first of Stevenson’s objections is that in defining the good as that which most people would approve of x if they knew its nature and consequences, we create an epistemic barrier to individuals forming accurate ethical judgments. But how serious of a problem is this for interest theories? After all, there’s no evidence to suggest that normal moral functioning requires epistemic perfection. It seems the vast majority of us are quite capable at forming at least inductively strong ethical judgments on the relatively small number of facts presently in our possession. [PC4] Furthermore, doesn’t the common sense notion that we are each capable of maturing and improving in our ability to make sound ethical judgments suggest some meaningful relation between the quality of our ethical judgments and the number of relevant facts in our possession?
Epistemic considerations aside, Stevenson objects that even if the majority came to approve of x on the basis of all the empirical facts pertaining to it, it would nonetheless remain an open question whether any individual who found out about this would be rationally required to conform to the majority view. As the following Moorean argument contends,
No matter what set of scientifically knowable properties a thing may have (says Moore, in effect), you will find, on careful introspection, that it is an open question to ask whether anything having these properties is good. It is difficult to believe that this recurrent question is a totally confused one, or that it seems open only because of the ambiguity of “good.” Rather, we must be using some sense of “good” which actually is not definable, relevantly, in terms of anything scientifically knowable. That is, the scientific method is not sufficient for ethics.[12]

            Such an objection, however, fails to address my view, for the account I’m proposing does not rely on a majority’s approval, but on an objective factual relation obtaining between each individual, a particular course of action[PC5] , and his or her ideal state. In the case that all empirical data were provided with respect to the constitution of our ideal state and the practical means by which to attain it, I deny that any one of us would question our ethical responsibility with respect to these facts. At worst, some individual A might fail to perform the action that she acknowledges is good, but then the “open question” would concern the quality of A’s character, or perhaps her sensibility, not the objective ethical value of x.

3. Ethical Statements as Ordinary Scientific Statements
Recall that according to Ayer, “a statement is held to be literally meaningful if and only if it is either analytic or empirically verifiable.”[13] Ordinary scientific statements fall within this second class of meaningful statements. My purpose in this section is to demonstrate that ethical statements, too, belong to this class of meaningful synthetic propositions. I will do this by first defining my obligations with respect to each criterion and then demonstrating through a discussion of ethical disagreement how both are satisfied.

3.1 Verifiability
            Upon discovering his first (strong) iteration to have excluded even the statements of empirical science, Ayer offered the following emendation to his verification criterion of meaning: “[a proposition] is verifiable, in the weak sense, if it is possible for experience to render it probable.”[14] That such an emendation was necessary to vindicate scientific statements as meaningful is evident from a cursory consideration of empirical propositions. As Ayer notes, “Empirical propositions are one and all hypotheses, which may be confirmed or discredited in actual sense-experience. And the propositions in which we record the observations that verify these hypotheses are themselves hypotheses which are subject to test of further sense-experience. Thus there are no final propositions.”[15] He goes on later to describe hypotheses generally as “rules which govern our expectation of future experience.”[16] In order to vindicate ethical statements by the verification criterion, therefore, we need only demonstrate that they, too, function as hypotheses in this capacity.

3.2 Syntheticity
            Ordinary scientific statements, furthermore, are synthetic, i.e. they communicate facts about empirical objects. Their truth or falsity, therefore, is conditioned upon the actual features of the empirical world as provided through phenomenal experience. For ethical statements to be included within this larger class, we must be able to prove their common relation to the empirical world. In order to do this, we need merely to show that all ethical terms are ultimately reducible to non-ethical terms, i.e., terms the referents of which are phenomenal in nature. It is precisely the accomplishment of this task that the logical positivist denies is possible.[17]

3.3 Ethical Disagreement
            One way in which logical positivists often seek to demonstrate the irreducibility of ethical terms to non-ethical terms is to provide an analysis of ethical disagreement. According to the most standard forms of the argument, there are two potential sources of disagreement: disagreement in belief and disagreement in interest. Disagreement in belief, typical of the sciences, occurs when A believes p and B disbelieves p. Disagreement in interest, on the other hand, “occurs when A has a favorable interest in x, when B has an unfavorable interest in x, and when neither is content to let the other’s interest remain unchanged.”[18] With regard to this distinction, Stevenson remarks that, “empirical knowledge resolves disagreement in interest only to the extent to which such disagreement is rooted in disagreement in belief.” The problem, however, arises in the apparent fact that not all disagreement in interest is of this sort. Rather, as Stevenson points out, “The disagreement in interest may arise not from limited factual knowledge, but simply from A’s sympathy and B’s coldness.”[19] If ethical statements really were like ordinary scientific statements, Stevenson is essentially arguing, all disagreement in interest would ultimately be rooted in disagreement in belief, and then would be resolvable, at least in principle, by appeal to empirical facts alone. But cases like that in which it seems every empirical reason has been provided and yet the disagreement persists seem to provide a counterexample to the generalization we might otherwise wish to draw. Before investigating what might be happening in these latter cases, let’s take a more in-depth look at those cases in which disagreement in interest is clearly reducible to disagreement in belief.
In Stevenson’s example, A and B can’t agree on whether it is better to go to the cinema or a symphony. A has an interest in cinema; B has an interest in the symphony. Each is trying to modify the interests of the other to achieve a consensus.
Clearly, [each] would give reasons to support their imperatives. A might say, “But you know, Garbo is at the Bijou.” His hope is that B, who admires Garbo, will acquire a desire to go to the cinema when he knows what play will be there. B may counter, “But Toscanini is guest conductor tonight, in an all-Beethoven program.” And so on. Each supports his imperative (“Let’s do so and so”) by reasons which may be empirically established.[20]

In such a situation, if any of A’s reasons were to succeed in swaying the desires of B so that she shifted her interest from the one activity to the other, then we would have established that the apparent disagreement in interest was actually rooted in a more fundamental disagreement in belief, namely a belief over which activity best served the common desire of both parties.
In characterizing the disagreement above an “apparent” disagreement in interest, I am already departing slightly from Stevenson’s analysis. According to Stevenson, the introduction of new facts works to resolve disagreements in interest by affecting a change in one person’s interests to match those of the other. But it should be clear in such cases that the introduction of new facts can be efficacious in changing a person’s interests only to the extent that those interests are conditioned upon the former or latter set of facts. But then it is not right to call these genuine interests at all, but rather opinions or beliefs about one’s genuine interests. For this reason, rather than call these interests in the primitive sense, I propose we think of them as the derived conclusions of practical deliberations. Such conclusions are read off as practical imperatives, with all the reasons supporting them serving as their inductive ground.
In making this latter suggestion I have parted ways completely with Stevenson, who argues, “empirical facts are not the inductive grounds from which the ethical judgment problematically follows…If someone said, ‘Close the door,’ and added the reason, ‘We’ll catch cold,’ the latter would scarcely be called the inductive ground of the former.[21] With respect to this assertion, I simply disagree. I can only speculate that the presentation of the two statements (and the obvious enthymematic quality of the argument) is obscuring for Stevenson their logical relation to one another. Slightly adapted for sake of clarity, the argument can be interpreted as having the following implicit logical structure:
1.     If A closes the door, then we will not catch cold (reason)
2.     If we don’t catch cold, we each individually will be closer to attaining our respective eudaimonic states (empirical hypothesis)
3.     Therefore, if A closes the door, then we each individually will be closer to attaining our respective eudaimonic states (by 1 and 2)
4.     If the performance of a particular action brings an individual closer to attaining his/her eudaimonic state, then that action is good for that individual (definition of instrumental good)
6.     Therefore, A ought to close the door (imperative, by 5)


But how does our analysis handle those apparent cases in which the disagreement in interest arises not from limited factual knowledge, but from A’s sympathy and B’s coldness? Simply by taking A’s sympathy and B’s coldness themselves to be the products of rational calculations. The fact of the matter is that attitudes often change – in response to new facts, or, perhaps more commonly, new experiences that teach us to see the old facts in a new way. The plasticity of attitudes in response to empirical data suggests that such attitudes, too, are analyzable by the same methods as employed above (i.e., via empirical demonstration and the exchange of reasons).
To the extent that such methods still fail to achieve consensus, we may resolve the tension in either of two ways: first, we may simply conclude that one of the individuals is, as a matter of fact, wrong (though whether or not we are able to determine with certainty which individual is wrong, or why, remains an epistemic concern); or, second, we may reason that the two individuals, despite appearances, are engaged in two arguments rather than one, and that both individuals’ judgments are right (or wrong, as it may be) relative to the distinct set of facts relevant to each case. This is what I propose is happening in Stevenson’s second example. Here, A and B are arguing about whether a public dole would be good. All the consequences of the dole have been made known to both parties. Now Stevenson has us suppose a situation in which A is poor and unemployed, and B is rich. In such a case, he argues, “the disagreement might not be due to different factual knowledge [but] to the different social positions of the men, together with their predominant self-interest.”[22] In such a case as this, it is ambiguous just what the matter of disagreement is precisely. Is the question whether the public dole would be good with reference to all empirical facts pertaining to A exclusively? to B exclusively? to both A and B as part of an integrated community? Different sets of empirical data become relevant to the ethical judgment as we shift from one context to another. Relative to any one context, however, there remains just one right answer with respect to its relevant data set. If the arguments I have provided are sound, then there exist at least three legitimate methods of resolving ethical disagreements without recourse to anything non-empirical, and without offending our common sense notions of ethical terms.

BIBLIOGRAPHY

Ayer, Alfred Jules. Language, Truth and Logic. New York: Dover Publications, Inc., 1952.

Stevenson, C.L. “The Emotive Meaning of Ethical Terms.” Logical Positivism. Edited by A.J. Ayer. New York: The Free Press, 1959.


[1] Alfred Jules Ayer, Language, Truth and Logic, (New York: Dover Publications, Inc., 1952), 102-103.
[2] Ibid., 106.
[3] Ibid., 105.
[4] Ibid.
[5] Ibid.
[6] C.L. Stevenson, “The Emotive Meaning of Ethical Terms,” in Logical Positivism, edited by A.J. Ayers (New York: The Free Press, 1959), 266.
[7] Ibid.
[8] Ibid.
[9] Ibid., 277.
[10] Logical Positivism, 267.
[11] Ibid., 267-68.
[12] Ibid., 268.
[13] Language, Truth and Logic, 9.
[14] Ibid.
[15] Ibid., 93-94.
[16] Ibid., 97.
[17] Logical Positivism, 275-276.
[18] Ibid., 277.
[19] Ibid., 279.
[20] Ibid., 278.
[21] Ibid., 278-79.
[22] Ibid., 279.

No comments:

Post a Comment