Closing the Question About the Open Question Argument or: What Geach Kant Do, and Much Much Moore!

Charles Johnson
10 January 2003

This essay is a draft of a work in progress; comments are welcome. The present draft is © 2003 by Charles Johnson, and is freed for credited reprinting and derivative works (but not, of course, for plagiarism) under the terms of the Creative Commons Attribution-ShareAlike 2.0 copyleft license. If you have any questions or comments about the essay, feel free to contact the author.

I. Introduction

In a remarkable passage of the Meno (71d-72d), Plato writes: #

… But Meno, by the gods, what do you yourself say that virtue is? Speak and do not begrudge us, so that I may have spoken a most unfortunate untruth when I said that I had never met anyone who knew….

It is not hard to tell you, Socrates. First, if you want the virtue of a man, it is easy to say that a man's virtue consists of being able to manage public affairs and in so doing to benefit his friends and harm his enemies and to be careful that no harm comes to himself; if you want the virtue of a woman, it is not difficult to describe: she must manage the home well, preserve its possessions, and be submissive to her husband; the virtue of a child, whether male or female, is different again, and so is that of an elderly man, if you want that, or if you want that, or if you want that of a free man or a slave. And there are many other virtues so that one is not at a loss to say what virtue is. There is virtue for every action and every age, for every task of ours and every one of us--and Socrates, the same is true for wickedness.

I seem to be in great luck, Meno; while I am looking for one virtue, I have found you to have a whole swarm of them.

Now, one defect of Meno's speech is that it is bad ethics. His account of men's and women's respective virtue, for example, reflects patriarchal norms that today we recognize as thoroughly unjust and exploitative. But Socrates draws our attention to another defect--a logical defect. For even if Meno were right, his speech would not have answered Socrates's question. In order to properly answer the question, we cannot simply list off myriad different virtues, for #

[The virtues do not differ in being virtues.] Even if they are many and various, all of them have one and the same form which makes them virtues, and it is right to look to this when one is asked to make clear what virtue is.

In their dialogue, Socrates and Meno reveal a central puzzle of meta-ethics. In most of our studies--whether we are inquiring after horses, or chairs, or triangles--we do not have much difficulty in describing what it is we are investigating. When we study horses, we have no trouble saying that what we are studying is a certain sort of animal--a hoofed quadruped with such-and-such characteristic features. But when we come to good or virtue, the subject-matter of ethics, the going is much more difficult. Like Meno, we often find ourselves in tension between the universal and the particular. When asked to explain goodness, we immediately find ourselves listing off a number of particular things and what makes for a good example of each sort of thing. But such answers somehow fail to live up to the question: we have, perhaps, said something about what is good, but we have not said anything about what good is. #

What Socrates wants from Meno is not a report on how it is with several good things, but rather, a description of the one feature that good things need to have in order to make them good. But is there any way to give him what he wants? In his landmark work Principia Ethica, G.E. Moore argues that Socrates got something importantly right, and something importantly wrong. He was right to challenge Meno's speech on virtue, because if it is intended as an account of goodness (rather than a list of good things) it commits what Moore calls the Naturalistic Fallacy. On the other hand, if Socrates expects Meno to give an account of good by a verbal definition, then he will be forever disappointed. Both conclusions, Moore believes, are established by one of his most famous arguments, the so-called Open Question Argument (OQA)--because the OQA establishes that the property good must name a simple, non-natural, indefinable property. Thus, while there is some one quality that all good things share, there is no possible conceptual analysis of it better than good is good, and that is the end of the matter (¶ 6 ¶ 2). #

There are, however, considerable difficulties remaining. In particular, it is unclear how well we can make sense of the notion of this non-natural property. Further, there is dispute over how to understand the exact interpretation and the logical force of the OQA. Finally, it is unclear whether Moore's answer can really give us a good explanation of the peculiar features of ethical concepts, including those features that put us in Meno's position. In Good and Evil, Peter Geach outlines several of these worries, and gives a basis for raising the rest, when he alleges that the Moorean view falls into error through an overly narrow view of language. On Geach's view, proper understanding of the logical status of good and bad will clear up the puzzle raised by Socrates without the need for any non-natural ethical properties. In turn, although Geach does not so apply it in his paper, his account can help us understand the logical force of the OQA without Moore's meta-ethical edifice. #

Careful attention to the OQA, and the crucial importance of normativity to its logical force, will show that like Socrates and Meno, Geach and Moore have each gotten something right, and gotten something wrong--and untangling it will require an understanding of normativity as such. A critical examination of Moore and Geach in light of Christine Korsgaard's work in The Sources of Normativity will provide us with the beginning of an understanding of how we can close the question about the Open Question Argument, and how we are to understand good, bad, and other terms of ethical evaluation. #

II. Hans, Fritz, and G.E. Moore

In order to see the difficulty that is raised in the investigations of Socrates and Moore, let us imagine an argument between two partisans of (particularly poor) ethical theories. Hans and Fritz are having an argument over whether someone should hook herself up to the Big Bright Green Pleasure Machine™, which will give her a lifetime of safe, healthy simulations of pleasant experiences without the need for any sweat or toil on her part. Hans thinks that any rational person ought to. Fritz argues that using the BBGPM would be terrible--it may be pleasant, but we desire truth and are repulsed by living a pleasant lie. Hans doesn't care whether or not we desire truth--such a desire is irrational masochism, if living a lie would give us more pleasure. Fritz insists that Hans is obviously mistaken: good just means what we desire. Hans rebuts Fritz: good just means pleasurable. Hans and Fritz may argue at some length over their theses; each may present several arguments for his own view and against his opponent's. Now, it seems (prima facie, anyway) that Hans and Fritz are having an argument about a serious ethical issue, indeed the very fundamental principle of ethics. Hans holds that the fundamental principle of ethics is to seek pleasure; Fritz holds that it is to seek what we desire. Given how they have phrased their positions, it seems like the dispute is that they are at odds over the meaning of the term good. What property does good name? On Hans's view, good designates the property pleasurable, and on Fritz's view, it designates the property desired. The debate between them will be closed by producing and validating a proper definition of good that resolves the question of meaning. #

So far, this seems to be a straightforward little dispute: Hans and Fritz say that they are right as to what good is and endeavour to prove that other people who say that it is something else, are wrong (§ 11 ¶ 1). But here we should begin to feel a bit puzzled. What sort of argument could possibly close the argument between Hans and Fritz? Each of them claims to have produced a definition which gives the meaning of the term good. How they are supposed to go about arguing will depend on what they mean when they claim to be giving a definition of the term. In § 8 of PE, Moore distinguishes three ways of giving a definition: #

  1. The arbitrary verbal definition: One may simply stipulate a meaning for a word, as when one says When I say horse, you are to understand that I am talking about a hoofed quadruped of the genus Equus.; #
  2. The verbal definition proper: One may state that a word is generally used in a specific way by speakers of the language, as when one says When most English people say horse, they mean a hoofed quadruped of the genus Equus. #
  3. Conceptual Analysis: One may, instead of giving a verbal definition, give an analysis, and mean that a certain object, which all of us know, is composed in a certain manner: that it has four legs, a head, a heart, a liver, etc. etc., all of them arranged in definite relation to one another. (§ 8 ¶ 1) #

With these three possibilities on the table, we may now try to capture what it is that Hans and Fritz are arguing over, based on what they meant when they offered definitions of good. #

If it is a matter of arbitrary verbal definition, however, then we should see that there is no dispute at all: Hans and Fritz are simply speaking at cross-purposes. Their dispute could be resolved without any argument at all, simply by distinguishing a good1 (= pleasurable) and a good2 (= desire) and resolving to specify which of the two they mean whenever they make an ethical statement. If we were to do such, then Hans would say that using the BBGPM would be good1 but not good2; and Fritz would no longer have anything to disagree with. If either tried to argue against the other's usage, then it would simply be missing the point: Fritz may be trying to prove that the object of desire is not pleasure (§ 11 ¶ 2). But while this argument may very well prove that pleasure cannot be good2, Hans never claimed that pleasure was good2 in the first place. By stipulation, he said just that pleasure is good1. #

A genuine dispute can be had by moving from an arbitrary verbal definition to a verbal definition proper. Since Hans and Fritz both take themselves to be speaking so much English, their dispute may be a dispute over the common English usage of the word good. Here, we can find a clear structure for the argument: When A says Good means pleasant and B says Good means desired, they may merely wish to assert that most people have used the word for what is pleasant and what is desired respectively (§ 11 ¶ 3). On this reading of Hans and Fritz's dispute, they cannot simply appeal to an arbitrary distinction: they must respect the way in which people have traditionally used the word, and empirical philology of English will determine whether one or the other is right to use the word good as he uses it. #

But let us say that Hans and Fritz understand their argument in this way; let us say that Hans defends his position by arguing that the word good has, as a matter of fact, been used to designate pleasure; but, as Moore asks, if this be all, where is his Ethics (§ 11 ¶ 2)? It will be quite important to see that reducing the dispute between Hans and Fritz to a dispute over English usage strips the argument of its ethical character. When Hans and Fritz began their argument, they had assumed that it was an argument about what ought to be done: should a person hook herself up to the BBGPM, or should she not? What are the reasons for acting one way or the other? Hans suggested that the maximization of pleasure was a reason to use the BBGPM, and Fritz suggested that the desire for truth was a reason to avoid it. But if Hans and Fritz defend their positions merely by appealing to an arbitrary verbal definition, or a fact about English usage, they have simply changed the subject. Thus Moore: #

Do, pray, act so, because the word good is generally used to denote actions of this nature: such, on this view, would be the substance of their teaching. … But how perfectly absurd is the reason they would give for it! … You are to say the thing which is not, because most people call it lying. That is an argument just as good! (§ 11 ¶ 3)2

Both sorts of verbal definitions say something about how the word good is to be understood, but neither sort provides a principle of ethics. For Hans and Fritz to be having an ethical dispute, they must be concerned not with how people have used a word, but rather with what we ought to do. The concern, then, is not with the sign good but rather with what is commonly symbolized by that sign. The definition of good to which Hans or Fritz appeals, then, cannot be a matter of lexicography: it must be a conceptual analysis of the subject-matter of ethics, as Moore calls it, of that feature a thing has that makes it normatively binding. #

III. The Open Question Argument

In order to definitionally close the dispute between Hans and Fritz, then, we will have to produce a conceptual analysis of that quality which we assert to belong to a thing, when we say that the thing is good (§ 10 ¶ 1). Moore contends that such an analysis is impossible, and he argues for this position with what has come to be called the Open Question Argument which he introduces in § 13 of PE: The hypothesis that disagreement about the meaning of good is disagreement with regard to the correct analysis of a given whole, may be most plainly seen to be incorrect by consideration of the fact that, whatever definition be offered, it may always be asked, with significance, of the complex so defined, whether it is itself good (§ 13 ¶ 2). #

Returning to the argument between Hans and Fritz, we may say that Fritz is offering an analysis of what it is for something to be good: that something's goodness is its desiredness. Hans, on the other hand, says that what it is for something to be good, is that it is pleasurable. Now we have a dispute between Hans and Fritz, and their dispute is an ethical one: the question is whether we ought to seek what we desire, or to seek what is pleasurable. The argument between them will be closed by showing that one or the other of their principles of ethics does not match up with what we pick out when we say good. #

But can either of their conceptual analyses close the argument in this way? Careful attention to the significant use of good should show that it cannot. Let us say that Fritz tells us the following: #

(A) Forgoing the Big Bright Green Pleasure Machine™ is good. #

If he has offered a correct conceptual analysis of the term good, then this substitution means the same thing: #

(A') We desire to forgo the BBGPM. #

Moore, however, insists that from A', we can carry the investigation further, and ask ourselves Is it good to [desire] A? and it should be apparent, on a little reflection, that this question is as intelligible, as the original question Is A good? (§ 13 ¶ 2). No matter what complex Hans or Fritz offers as the analysans of good, we may, so to speak, re-open the question: is it good that this particular thing has features XYZ? So, for example: #

(B) Is it good that we desire to forgo the BBGPM? #

But again, if Fritz claims to have defined good, then these two sentences mean the same thing as B: #

(B1) Do we desire that we desire to forgo BBGPM? #

(B2) Is it good that forgoing the BBGPM is good? #

But are either of these two questions what Hans meant to ask when he challenged Fritz's alleged definition? B1 seems like a question of empirical psychology, not an ethical question as in B. B2, on the other hand, is completely unintelligible. What would it be for it to be good that something is good? What would it be for it to be bad that something is good? B is an intelligible ethical question, whereas B1 is not an ethical question and B2 is not an intelligible question. Therefore, the putative definition Fritz has given us cannot properly capture the meaning of good. #

To sharpen the point, let us distinguish two kinds of properties: the property of goodness on the one hand (G), on the one hand, and non-moral, or natural (N) properties, on the other. Moore's argument seems to show that there is no possible correct analysis which defines good in terms of a complex of N properties. Any such analysis fails to pick out an important fact about good's significant use: we can always intelligibly ask whether it is good that a thing has a given set of natural properties--call this G-N predication. But we cannot intelligibly ask whether it is good that a thing has a given set of moral properties--call this G-G predication. G-G predication simply does not make sense. But if we analyze good in terms of any complex of natural properties, we will always face a situation such as that in B--it will be equivalent, by substitution, with a non-ethical question (N-N predication, as in B1), and also a question that is simply unintelligible (G-G predication, as in B2). There is, then, no conceptual analysis that can close the dispute between Hans and Fritz, because no conceptual analysis of good in non-ethical terms is possible. #

But surely Hans and Fritz were arguing over something! There is certainly an ethical issue at hand, when one says that pleasure is good and another says desiredness is good. How can we understand the argument between them so that it reflects the legitimate debates we have over principles of ethics? Moore proposes one way to do so in § 12 of PE: Hans and Fritz, if they were having a dispute over the fundamental principle of ethics, did not disagree about the meaning of good, but rather the identity of the things to which good applies. What they have given to close their dispute is not a definition, but rather a universal predication. With this distinction in mind, there is no difficulty in saying that pleasure is good and yet not meaning that pleasure is the same thing as good, that pleasure means good, and that good means pleasure (§ 12 ¶ 1). Rather, what is being said is that everything which is good is also pleasant, and everything which is pleasant is also good. To confuse this with a definition is to make the mistake of supposing that what is universally true is of such a nature that its negation would be self-contradictory (§ 13 ¶ 3). #

If Hans and Fritz make this mistake, then they may mistake the nature of their own dispute, and think that when they named those other properties that they think always accompany a predication of good, they were actually defining good; that these properties, in fact, were simply not other, but absolutely and entirely the same with goodness (§ 10 ¶ 3), and thus commit what Moore dubs the naturalistic fallacy. The error involved is simply that Hans or Fritz has attempted to give a definition of a term that is indefinable; and it is naturalistic because of the specific confusion of G properties with N properties. With the OQA in hand, however, Moore argues that we can see that the quality that is shared by all good things, which is picked out by the term good, is a simple, indefinable, non-natural property, and in particular, Socrates' quest was in vain, for good is one of those innumerable objects of thought which are themselves incapable of definition, because they are the ultimate terms of reference to which whatever is capable of definition must be defined (§ 10 ¶ 1). #

IV. Properties and Attributive Adjectives

With this forceful argument on the table, however, we must step back and see whether Moore's argument moves too quickly. Peter Geach, in particular, accuses Moore of a confusion over the logical role of good in the language. In developing the OQA, Moore had taken it for granted that there were only three options for the meaning of good: #

  1. Good names a complex of properties that good things have in common. #
  2. Good names a simple property that good things have in common. #
  3. Good means nothing at all. #

Having taken it for granted that (3) was false, and having eliminated (1) with the OQA, Moore contends that we ought to settle with (2) by process of elimination. But Moore's threefold division should already make us cautious. Does the word good have to name a particular property in order to be meaningful? Is the adjective here merely a concealed proper noun that names a universal? On Geach's view, the correct explication of good shows that: #

  1. [G]ood and bad are always attributive, not predicative, adjectives. Geach (1956) ¶ 3 #

And that a proper appreciation of this fact will show that, while the OQA is forceful, and the Naturalistic Fallacy is to be avoided, they do not actually tell us what Moore thinks that they do. #

To understand the argument, however, we will need to explicate Geach's distinction between predicative and attributive adjectives. Geach describes it thus: #

I shall say that in a phrase an A B (A being an adjective and B being a noun) A is a (logically) predicative adjective if the predication is an A B splits up logically into a pair of predication is a B and is A; otherwise I shall say that A is a (logically) attributive adjective. Geach (1956) ¶ 1

The difference may be glossed by saying that predicative adjectives, such as red or triangular, predicate a common property of anything they are applied to, the common quality that they all share that makes them red or triangular. Such properties may be definitionally simple, as with red, or a definable complex, as with triangular. Logically attributive adjectives, on the other hand, are quite different in their logical role. Attributive adjectives are adjectives such as big or forged, where the adjective leans on the noun to which it is applied, in such a way as to create a new description. #

As an example, consider the difference between a red flea and a big flea. Consider this argument, which clearly goes through: #

  1. x is a red flea #
  2. x is an animal #
  3. Therefore, x is a red animal #

And contrast it with this argument, which clearly does not: #

  1. x is a big flea #
  2. x is an animal #
  3. Therefore, x is a big animal #

A red flea is something that also has to be a red animal, because when I describe something as a red flea, I pick out the flea and I attribute to it an independent quality, redness. All red things share this quality, whether they are fleas or cows or cars. But when I describe something as a big flea, I do not attribute some independent quality, bigness, to that flea. There is no such quality to attribute. Big animals are creatures like elephants and whales; the biggest flea is still a tiny animal, and the largest animal is much smaller than the tiniest galaxy. #

If we ask, then, what quality all big things share, the answer is none at all. But big is by no means meaningless. Nor is it ambiguous: it means the same thing wherever it is used. A big flea, a big animal, and a big galaxy are one and all picked out as big according to exactly the same principle, based purely on the natural properties that they have--a spatial extension which is greater than the usual spatial extension of members of their category. We can now begin to see how Geach's sets his position apart from Moore's. On both Moore's and Geach's accounts, there is no one complex of natural properties that all good things have in common. For Moore, this means that--if we can call them all good--they must have some non-natural property in common. For Geach, on the other hand, there does not need to be any non-natural property added on top of the natural properties. Instead, Geach suggests, The traits for which a thing is called good are different according to the kind of thing in question; a knife is called good if it is UVW, a stomach if it is XYZ, and so on Geach (1956) ¶ 10, but no additional, special property is needed to reach up to goodness. What it is for a knife to be a good knife just is that it is sharp, durable, etc. And what it is for a horse to be a good horse just is that it is strong, swift, etc. As Geach argues elsewhere (1957), predicates are not things like objects: the assumption that horses must have some second property good in addition to their various properties rests upon a false Platonistic logic; an attribute is being thought of as an identifiable object (39). #

Why, then, should we accept Geach's analysis of ethical terms? Geach suggests three reasons why we might prefer it to the Moorean solution: #

(1) We may worry about the metaphysical status of the alleged non-natural property of good. Geach contends that nobody has ever given a coherent and understandable account of what it is for an attribute to be non-natural and, if we try to pin the position down it may seem as though Moore is just playing fast and loose with the term attribute Geach (1956) ¶ 5. Since he only appeals to a mysterious non-natural property, Moore's objectivism offers only the pretence of a way out of the Naturalistic Fallacy: it does not really give an account of how good differs in logic from other terms, but only darkens counsel by words without knowledge Geach (1956) ¶ 7.7 #

(2) Our established usage strongly suggests that bad is a logically attributive adjective, of the sort that medieval logicians called alienans. If X is a forged banknote, it does not follow that (1) X is forged and (2) X is a banknote. Forged leans on the noun banknote to produce a new description which no longer describes any actual banknote at all. Bad works in a similar way: we cannot safely predicate of a bad A what we predicate of an A … we cannot infer, e.g. that because food supports life, bad food supports life Geach (1956) ¶ 3. This does not immediately win the day for good, since whatever holds true of an A as such holds true of a good A Geach (1956) ¶ 3. But this does at least give us some reason to think that ethical terms, in general, are logically attributive in nature. #

(3) Finally, we may worry that the Moorean explanation leaves the relationship between moral and non-moral properties entirely mysterious. As Geach points out: #

I could ascertain that a distant object is a red car because I can see it I red and a keener-sighted but colour-blind friend can see it is a car; there is no such possibility of ascertaining that a thing is a good car by pooling independent information that it is good and that it is a car. Geach (1956) ¶ 3

Properties that are designated by predicative adjectives are logically separable from the other properties that a thing has: something which has a certain color property is red whether it is a car, or a van, or a cow. The property of redness does not depend on the non-color properties that a thing has. Goodness, however, depends entirely on the non-moral properties that a thing has. If goodness is a simple property apprehended through a sense of moral intuition then it seems odd that we cannot so intuit it except when we make a positive identification through our workaday natural senses. #

V. Attributive Adjectives and the OQA

With this understanding of ethical terms in hand, we may now return to the OQA and the Naturalistic Fallacy. If good is a logically attributive adjective, then we must always use it with reference to some category of things, of which the particular in question is a good instance. We may speak of good people, good marriages, good CDs, and so on. But we may not speak of a thing being good simpliciter, as there is no such thing as being just good or bad, there is only being a good or bad so-and-so Geach (1956) ¶ 3. If in ordinary speech we speak of a good or bad thing, either thing is a mere proxy for a more descriptive noun to be supplied from the context or else one is using a dummy noun to try to use good or bad predicatively Geach (1956) ¶ 3 which is, on Geach's thesis, wholly illegitimate. Just as with big, good things will all be picked out of their class according to the same principle, but the properties on the basis of which they are picked out will differ from one category to another. #

And here we should see that Moore has gotten something right, and something wrong. There is no one set of simple properties that will explain all the uses of good, and this is brought to the forefront by the OQA. Attempting to identify good with some set of simple properties will be a fallacy; and we can call it the Naturalistic Fallacy if we so please. But the force of the fallacy is not the confusion of a non-natural property with a natural property; rather, it is the confusion of a property with something that is not a predicated property at all. And similarly, the question that may be re-opened in ethical debates is not the definition of good, but rather the morally dominant category under which a thing is to be evaluated. #

Let us, at last, return to Hans and Fritz. If Hans says that using the BBGPM is pleasant, Hans has challenges him: Yes, but is it good that it is pleasant? This question is intelligible--not because good designates a different non-natural property, but rather because we step back into a different category, and now ask Yes, one gets pleasure from the BBGPM, but is that a good way to live? The G-G problem is thus handily dealt with. If pleasure is the good of feelings, then the G-G question now asks: Is getting a good feeling from the BBGPM a good way to live? And this is a perfectly intelligible question, which we can go about answering with a theory about human being and its relation to knowledge of the truth. G-G predication is a problem when we do not make a category change: it is unintelligible to ask, Is it a good feeling that we get a good feeling from the BBGPM?--because that is not a feeling at all, but rather a way to live. #

Moore, then, has failed to show us what happened with Meno and Socrates. If Meno's chauvinism were correct, then it would indeed turn out that the good of a man consists of being able to manage public affairs and in so doing to benefit his friends, that the good of a woman just is that she manage the home well, preserve its possessions, and be submissive to her husband, and so on. Since good and bad are attributive adjectives, the properties that constitute them differ, so that There is virtue for every action and every age, for every task of ours and every one of us--and Socrates, the same is true for wickedness. But there is also method to the madness, and we can join with Socrates in saying that all of them have one and the same form which makes them virtues, and it is right to look to this when one is asked to make clear what virtue is. Moore errs when he tries to make this one form a quality common to all the instances of virtue: instead, the one form is the common principle by which they are picked from their respective categories. The OQA shows us this: we can always intelligibly re-open the ethical question, but the question that is being re-opened is not the question of what it is for things within a category to be good; it is a question of what the morally dominant category in a given situation is. #

VI. Reasons for Action and What Geach Kant Do

Nevertheless, toward the end of Good and Evil, Geach confesses that he is aware that much of this discussion is unsatisfying Geach (1956) ¶ 22. Whatever success at logical analysis Geach has made, he has left a lot up in the air. He has explained how we can pick out some natural qualities as constitutive of a good X, without committing the naturalistic fallacy of identifying good itself with those qualities, and without invoking any mysterious non-natural quality as a solution. We are able to do this, he contends, because good identifies a principle that picks out different natural qualities according to the different categories of things it is applied to. But he has not told us what this principle is. We know what makes big fleas and big planets both big--that their spatial extension is greater than the spatial extension of most other fleas or planets, respectively. But what principle is it that picks out good people, good actions, good CD players, good horses, and so on? With Geach's corrections to Moore, we cannot any longer ask after the quality of goodness, but we can still ask to be let in on what it is that the adjective good does when it is applied to a noun. #

What sort of answer might Geach be able to give? He rightly recognizes that whatever he says good and evil are, good and bad must be normative terms for us: when we have icked out a thing as good or bad, it must belong to the ratio of want, choose, good, and bad, that, normally, and other things being equal, a man who wants an A will choose a good A and will not choose a bad A Geach (1956) ¶ 14. That is to say, when good and bad pick out qualities, those qualities have to matter to us; they have to give us reasons for action. The difficulty that Geach is confronted with by his account of goodness is well-described by Geach himself: #

Let us suppose that we have found a clear descriptive meaning for good human act and for bad human act, and have shown that adultery answers to the description bad human act. Why should this consideration deter an intending adulterer? By what logical step can we pass from the supposedly descriptive sentence adultery is a bad human act to the imperative you must not commit adultery? Geach (1956) ¶ 15

Geach's short answer is that even though calling a thing a good A or a bad A does not of itself work upon the hearer's desires, nevertheless it may be expected to do so if the hearer happens to be choosing an A. Since we are always faced with the decision of what we are going to do with ourselves, we are always choosing a manner of acting, and to call a manner of acting good or bad cannot but serve to guide action Geach (1956) ¶ 17. #

Geach, however, is not entirely happy with his own account of the normativity of the natural qualities that constitute the goodness of an A; he identifies it as one of the places where does not see clear Geach (1956) ¶ 22. While Geach has brought many useful observations to light in his analysis, this is a major gap in his argument, if we want to apply it as we have to Moorean concerns related to the OQA. We have already seen that the reason-giving force of moral terms is central to the force of the OQA and the error committed by the naturalistic fallacy. Moore ruled out verbal definitions precisely because such maneuvers do not actually give any reasons for the actions that are described as good. And the worries involved in the OQA itself were also crucially linked to normativity: N-N predication is different from G-N predication because G-N predication gives us a reason to care about the natural quality being called good. G-G predication is unintelligible because reasons apply to things, not to other reasons. And Geach's account is quite incomplete on all these features of moral discourse. It may very well be true that quidquid appetitur, appetitur sub specie boni is part of how it is with good, desire, choice, etc. But if it is so, Geach has given us no reason to think that it is so; certainly good is quite different from any other natural term in having this relation to desire and choice, and the analysis so far has given us no reason to see why this is. #

In order to fill this gap, we will have to turn to some theory of normativity as such, and see whether we can apply it to Geach's logical thesis about good and bad. In order to try to resolve this problem, I will examine the normative theory offered by Christine Korsgaard in The Sources of Normativity, and see how it might be applied to the worries raised by Moore and the corrections suggested by Geach. In so doing, we may be able to find a way to explain how a mere description of adultery can deter an intending adulterer, and to explain the binding force of moral evaluations, without resurrecting a non-natural moral quality. Nevertheless, Geach will also have to give up quite a bit. #

Korsgaard's examination of moral terms begins from the standpoint of normativity: the task that she takes is to show you where obligation comes from (495). The way that she will do this is by brashly and intentionally running together two questions which seem to be quite separate: #

  1. Why is it right? #
  2. Why do I care? #

Question 1 is a question about the truth-makers of moral claims, and belongs to the metaphysic of morals. Question 2 is a question about the conditions that bring me to regard a thing as something to be pursued, and belongs to moral psychology. But as we have already observed, an adequate theory of moral discourse must provide an explanation for the strong coincidence of the two: morality matters to us. What Korsgaard will suggest, in a deeply Kantian strategy, is that this fit is far from accidental: there is a deep conceptual relationship between question 1 and question 2, and without modification Geach's analysis--like many other analyses--errs by putting too much distance between motivation and morality. By developing a theory from the standpoint of the normativity of the moral, she hopes to correct this error and so allow us to both gain the world and keep our souls. #

In order to do this, our account of normativity will have to be closely tied with both moral terms, and also with a doctrine of the will. One immediate problem that confronts us, however, is that we can no longer speak of good as it is applied to horses, cars, computers, or marriages, and we now seem to be concerned exclusively with the goodness and badness the acts of a solitary rational agent. This anticipates a possible objection from Geach's camp, but we will sweep it under the rug for right now. Our task is to get clear on normativity, which certainly does bear a special relationship to the will of an agent. How this impacts good as it is used more broadly will be seen after the account is laid out. #

How does the question of the normative arise? On Korsgaard's account, we raise the question of reasons for action because the human mind is self-conscious in the sense that it is essentially reflective (495). We can turn our attention upon our perceptions and desires, and our capacity to turn our attention on our own mental activities is also a capacity to distance ourselves from them and to call them into question (495). Normative questions arise when--and because--we are called upon by ourselves, or by others, to account for why we did what we did. To take Plato's famous example from The Republic, I may decide not to give you back the weapon you've loaned me. Why? Because you've gone mad and you'll use it to kill an innocent person. Let us call what has just been given a because-story. We can generalize what is happening here with the following schema: #

(A)I __________________________ because__________________________.
(act) (motive)

Here, I fill in the first blank with what I did, the second with the story about why I did it--I kept the weapon you loaned to me because you'd use it to kill an innocent person, and I don't want that to happen. Now, it may seeem that simply by filling the schema in, we have finished the job. I told you the reason that I held onto your weapon: because you'd gone mad. #

But to make this move is to fail to appreciate what is special about human reflectivity. The reflective mind, Korsgaard writes, cannot settle for perception and desire, not just as such because reflectivity not only allows, but demands that we ask Is this desire really a reason to act? Any because-story that we give will be some report on internal motives and external facts about the world. Because-stories may sway me one way or another; because of them I may find myself with a powerful impulse to act (495). But do such because-stories count as reasons for action? Back to the A-schema and Plato: #

(A) I __________________________ because __________________________.
(act) (motive)
(A1) keep your weapon (k1) I want it (m1)
(A2) keep your weapon (k1) you've gone mad (m2)
(A3) keep my own weapon (k2) I want it (m1)

Now, for each because-story that is given for each act, we can legitimately ask ourselves whether the story is such that it gives us a reason to do what we did. In A1 and A3, I have succeeded in giving you a reason for refusing to give you a weapon. In A2 I have not. But if the normativity of an action just lies in the because-story that is given, then all of these instances should be on the same footing. Bad acts have psychological motives just as much as good ones do. We should begin to see that the question we are faced with here is the same question that we faced with the OQA, on both Moore and Geach's reading of it. Reflectivity is our capacity to re-open the ethical question, no matter how many facts about our own psychology or the world around us, are piled up in front of the mind's eye. We have adopted reflective success as the criterion for a reason, but when we ask So what brings reflection to an end? (496), the answer cannot be any report of this kind. Some facts count as reasons, and others do not, and whether or not a given set does is always an intelligible question. #

So what is going on in these instances of the A-schema? When we set these instances out, we are compelled to see that The goodness of a given action does not rest in the parts, but rather in the way the parts are combined and related; so the goodness does not rest in the matter, but rather in the form (499) of the relationship between the act and its because-story. Our will is accepting some instances of the schema as reasons and rejecting others. Some A-relations will bear on the will--be normative--and others will fail to: #

k1 connects to m2; k2 connects to m1 and also to m2

So, then, whenever the will is acting, it generates a set of acceptable instances of the A-schema, of all the arrows that you can draw between the act and the motive. And the principle by which this set is determined will be what relates acts to motives--accepting some A-relations as legitimate and condemning others. #

It is part of the nature of the will, then, that we will adopt some principle according to which certain because-stories count as reasons and others do not--but what principle will we adopt? How do we go about choosing how we will react when given an instance of the A-schema? Well, if we adopt a certain principle because of some set of facts about how it is with the world or what desires we have, then the adoption of a principle would be an action like any other--an action whose motives we could inquire after, and which we could again reflect upon to see whether or not the choice was justified. The question of what principle to adopt would merely be deferred, not answered, since if we desire to adopt a particular identity, the reflective mind must endorse the desire before it can act on it (496). And so, as Kant (qtd. in Korsgaard) says, We cannot conceive of a reason which consciously responds to a bidding from the outside with respect to its judgments (496). The will must be a law unto itself: it must have a rule to determine which because-stories are acceptable, and the rule must be an autonomous guiding principle. But here already we have a source of normativity: the will must adopt a law for itself, and with that law in place, any act which can be willed must be in conformity with that law--any which is not in conformity is, ipso facto, contra-normative. Call this requirement the Law of Freedom--or, what is the same, the first formulation of the Categorical Imperative. We must always act such that we can will our action according to a universal law, and the guiding principle we adopt is the source of all our reasons for action. #

We have normativity on the table, then. I have a reason to seek something if seeking it is part of the principle that I adopt for reason-giving force. And now, in fact, with normativity on the table, we may already speak of good and bad. For if we accept what has gone before in Geach, then we will not be looking for any qualitative description that good things must fulfill in order to count as good--rather, we will be looking for a principle of selection. And we are already speaking of principles of selection for the things that will count as reasons for action. The relationship between normativity and goodness will be that the right and the good, are names for problems--for the normative problems that spring from our reflective nature, specifically: Good names the problem of what we are to strive for, aim for, and care about in our lives. Right names the more specific problem of what we are to do (500). #

We are beginning to see some light at the end of the tunnel: when we understand the relationship between reasons for action and the free will, we already have in hand a connection between the natural facts that make a thing good, and the fact that it is normatively binding on us. But there is still a major problem left outstanding. We must have some principle which rules some because-stories as legitimate and others as illegitimate, but which principle should we adopt? I can draw the arrows, so to speak, as I do above; but any set of arrows that I draw will supply a principle as well as any other. Korsgaard frames the problem here in terms of a distinction (one which Kant failed to make) between the categorical imperative and the moral law. Thus Korsgaard: #

Any law is universal, but the argument doesn't settle the question of the domain over which the law of the free will must range. And there are various possibilities here. If the law is the law of acting on the desire of the moment, then the agent will treat each desire as it arises as a reason, and her conduct will be that of a wanton. If the law ranges over the interests of an agent's whole life, then the agent will be some sort of egoist. It is only if the law ranges over every rational being that the resulting law will be a moral law, the law of the Kingdom of Ends. (497)

The Law of Freedom is a law--the one that I adopt, I must conform to in all my actions--but it is of freedom--it is up to my will which one I will adopt. So we have, it may be said, shown where the goodness comes from when something is good, but we haven't given ourselves any basis to say what is good. #

In order to get a grip on the problem and begin to see our way towards a solution, we should look again to what the question is that we are faced with. Our problem is that the reflective structure of the mind is a source of self-consciousness because it forces us to have a conception of ourselves, but we have not yet determined what sort of conception it is we ought to have. The answer to this question will be an answer to the existential question: what sort of person am I going to be? The wanton, the egoist, the moral person, and many more--say a person who cares about maximizing the amount of pound cake in the world--will all seem to be up for grabs. But such a move will be premature. What you are faced with in adopting a law for the will is choosing a principle or law by which you determine your actions that you regard as being expressive of yourself (497)--you must adopt an identity that expresses what you do, and what you most deeply want, aim for, and need. But you are a human being--a being of a certain sort, that lives in the world according to a certain form of life. For Korsgaard, the Moral Law emerges from the interaction between the Law of Freedom and our humanity: we are such that the only identities we can adopt for ourselves will be identities that are concerned with justice, charity, and the other virtues. There is not space here to sketch out in any detail what sort of argument will be given for this conclusion; we will only gloss it by saying that, on Korsgaard's view, whenever we try to give reasons for action, we cannot help but (1) couch those reasons in terms of what is good for a human being as such--which gets us the virtues of respecting humanity, and (2) respect other human beings as having the ability to impose reasons upon us--which requires us to respect all rational agents (and perhaps also animals) as beings with reasons that are to be respected. The argument will not be an argument that some biological or social fact about human beings necessitates that we adopt a particular practical identity; this would be to abandon the autonomy of the will. Rather, Korsgaard argues that our reactions under elenctic pressure, when making claims of value, show that the only sorts of value we can make sense of adopting for ourselves are those that are in conformity with a certain form of life. #

VII. Consequences

What are Hans and Fritz arguing over when they argue over whether or not one should use the Big Bright Green Pleasure Machine™? What they will be looking for in the dispute is whether or not using the BBGPM can be described as part of the way a person ought to live--part of what Korsgaard speaks of as a description under which you find your life to be worth living and your actions to be worth undertaking (497). Fritz and Hans, when they advance their ethical principles, each are making a claim about what it's like to be a rational human being: Hans claims that our form of life is such that maximizing pleasure is what we ought to aim for, and Fritz claims that our form of life is such that satisfying desires is what we ought to aim for. #

Can Hans and Fritz close the ethical dispute between them? With Geach and Korsgaard thus understood, we can say that in one sense they can, but in another sense they cannot. In a sense, the question can be closed by pointing out the natural features of using the BBGPM that make it wrong. For Geach, we have fully and completely understood the badness of using the BBGPM once we have given a full description of it as a bad human act. The force of the OQA then would rest only in the fact that there are other categories under which we can re-open the question. With Korsgaard's account in place, however, there must be another sense in which the OQA has logical force. If we give a full description (as Hans and Fritz started to do) of what using the BBGPM is like, the that description tells us all we need to know (the because-story) about what makes it good or bad. The principle of the will has all that it needs to make a determination without adding in any intuition of a non-natural quality of goodness. But reflectivity demands not just that we have reasons, but that we can become aware of them: I will have a reason for action just in case my will could, on reflection, endorse a certain act. I must see that reason as a reason; I must be able to cotton onto the principle that picks out natural qualities as good-makers and bad-makers. Thus, Geach has underestimated the logical force of the OQA. #

On the other hand, what I am cottoning onto if I successfully see my reasons for action as reasons is not any quality possessed by the action itself. The endorsement of the will is not a thing that good acts have, but rather a function that relates acts to because-stories. Good is logically attributive--it picks out features of an act as good-makers or bad-makers according to the moral principle which is adopted by the free will. Thus, Moore has overestimated the metaphysical import of the OQA. #

VIII. Towards an account of natural goodness

With all this in place, there may be one final objection raised from Geach's camp, which we had alluded to earlier. Our discussion of normativity has focused entirely on the human will, and the principle that picks out good and bad human actions. But in developing his analysis of good as logically attributive, one of Geach's central concerns was to avoid the claim that there is a difference between the evaluations of things such as horses, cars, and CD players, on the one hand, and the moral evaluation of human conduct on the other. But doesn't our reading of good fundamentally limit the analysis of good we have presented to evaluations of human conduct? Doesn't this leave completely open the analysis of good and bad for non-human animals and for non-living things, and require that whatever analysis we come up with will dissolve good and bad into ambiguous terms, with different applications in moral and non-moral cases? #

I will close by suggesting a brief sketch of how one might expand this understanding of normativity to defuse this objection. Korsgaard herself expands her theory to provide an account of the good for non-human animals. Non-human animals, argues Korsgaard, have practical identities. There are things that are good for them (such as getting enough to eat) and things that are bad for them (such as being vivisected). They have reasons to be certain ways and to do certain things. The difference between them and us is that, because we are reflective, we know that we have reasons, whereas animals simply have them without knowing that they do. And now we can begin to see how we can speak of good horses and good dogs. A good horse is one that lives up to what a horse ought to be--it will be strong, swift, and so on. Being that way is good for the horse because it is in conformity with the practical identity that the horse has, even though the horse does not have the reflective capacity to try to work out for itself what that practical identity is. Further, the good of a horse is not just normative for the horse. It is normative for us, too, because our practical identity is so constituted that we care about the reasons of other beings. So now we can now see why it is with us that quidquid appetitur, appetitur sub specie boni: because the reasons that a horse has to be strong and swift are reasons for us as well as for the horse. The reasons that it has not to be subjected to cruel torture are the reasons we have not to cruelly torture it. So we may say, strength and swiftness are not just good for us in an instrumental sense--it is not just that we endorse the horse having those because it serves some purposes that are important to us--but we can also judge that a horse with speed and swiftness are good for the horse, and (all things being equal) what the horse ought to have. #

Many questions remain unanswered here. Could we also speak of non-living things as having this sort of autonomous good, or can we only speak of them being good for our purposes? What, exactly, is the relationship between our good and the autonomous good of an animal or another human being? What sort of considerations could override their reasons, and what sort of considerations would compel us to respect their reasons? All of these questions will have to remain open for the time being, but by offering one way to close the question about what the Open Question Argument does, we have at least set ourselves on track for finding an answer. #


  1. And God only knows what he would have said about the virtue of a child, or a slave (!).

  2. Certainly, appealing to an arbitrary verbal definition will be even weaker: You are to do this, because I have arbitrarily stipulated that a certain word denotes it.

  3. Some, of course, have thought that the good just is what we desire to desire, so B1 is actually equivalent to A. But this will only make the problem sharper. Cf. Moore: But it is also apparent that the meaning of this second question cannot be correctly analysed into Is the desire to desire A one of the things we desire to desire?: we have not before our minds anything so complicated as the question Do we desire to desire to desire to desire A? (§ 13 ¶ 2)

  4. We may, of course, be able to throw out some definition like good is what it is right to admire. But this is no help in illuminating the subject-matter of ethics; it merely relocates the problem to the analysis of right and admire.

  5. These words, it seems to me, give us a particular picture of the essence of human language. It is this: the individual words in a language name objects--sentences are combinations of such names.----In this picture of language we find the roots of the following idea: Every word has a meaning [Bedeutung]. This meaning is correlated with the word. It is the object for which the word stands. (Ludwig Wittgenstein, Philosophical Investigations § 1). It is the ancient Dragon, the philosophical conception of meaning, that hands over its own power and its throne and its immense authority to philosophical beasts risen from the sea.

  6. The first argument goes through because it can be expanded as follows:

    1. x is a red flea.
    2. x is an animal.
    3. x is red and x is a flea. (from 1)
    4. x is red. (Simp. 3)
    5. x is red and x is an animal. (Conj. 4, 2)
    6. x is a red animal. (from 5)

    The logical step from 1 to 3, and from 5 to 6, is justified because predicative adjectives split up as Geach described. The second argument fails because x is a big flea does not split up into x is big and x is a flea.

  7. As Geach puts it:

    In order to assimilate good to ordinary predicative adjectives like red and sweet they call goodness an attribute; to escape undesired consequences drawn from the assimilation, they can always protest, Oh no, not like that. Goodness isn't a natural attribute like redness and sweetness, it's a non-natural attribute. It is just as though somebody thought to escape the force of Frege's arguments that the number 7 is not a figure, by saying that it is a figure, only a non-natural figure, and that this is a possibility Frege failed to consider. Geach (1956) ¶ 5

  8. Quidquid appetitur, appetitur sub specie boni.

  9. Caesar's murder was a bad thing for a living organism, a good fate for a man who wanted divine worship for himself, and again a good or bad act on the part of his murderers. (Geach) Geach (1956) ¶ 19

  10. Thus Geach (1956/1967):

    The ordinary uses of good and bad are for Objectivists just a complex tangle of ambiguities. I read an article once by an Objectivist exposing these ambiguities and the baneful effects they have on philosophers not forewarned of them. One philosopher who was so misled was Aristotle; Aristotle, indeed, did not talk English, but by a remarkable coincidence ἀγαθός had ambiguities quite parallel to those of good. Such coincidences are, of course, possible; puns are sometimes translatable. But it is also possible that the uses of ἀγαθός and good run parallel because they express one and the same concept; that this is a philosophically important concept, in which Aristotle did well to be interested; and that the apparent dissolution of this concept into a mass of ambiguities results from trying to assimilate it to the concepts expressed by ordinary predicative adjectives. It is a mere prejudice to think that either all things called good must satisfy some one condition, or the term good is hopelessly ambiguous. Geach (1956) ¶ 5