Chapter 13: BDI Theory and a General Theory of Value
I. Trapped
The University of Maryland had this annoying rule, that all universities seem to have, that its students must acquire a breadth of knowledge. In undergraduate school, they required that I take classes such as biology and literature before they would give me a philosophy degree. In graduate school, I was at least allowed to take philosophy classes, but they required that I take classes in philosophy of science and history of philosophy when I was there to study value theory.
Not only did I consider those courses outside of my major field interest a colossal waste of time, I suffered from a “small fish in a big pond” situation. Outside of my field, the classes were filled with students actually interested in those other subjects. These people rattled off research in this field the way most people talk about their favorite sports teams. I could not care less.
One of the courses I took to fulfill these requirements was a seminar in the Philosophy of Mind. I spent the first sessions struggling to stay awake as the professor talked about concepts I had little experience with. Then, in about the third lecture (I think), Dr. Geogres Rey started to say something interesting. Over the course of that hour, it was as if he had looked right at me and said, “Here, Alonzo, are the answers to the most central set of questions you have been contemplating.”
Even more surprising was the fact that those answers were not in some obscure fringe theory that, itself, required its own elaborate defense. I found them in the ‘default view’ in the philosophy of mind — the view that anybody else holding a competing theory thought that he would have to argue against.
II. BDI Psychology
At that point in the class, Dr. Rey started discussing a theory of mind called BDI Theory (which stood for Belief - Desire - Intention Theory). This theory holds that beliefs and desires are brain states, that these brain states combine to form intentions, and intentions then cause intentional actions.
I have a desire to get to work. I believe that the L-Bus will get me to within a block of where I work, and I believe that the bus will pick up passengers at a bus stop about 3 blocks from my home. I do some research and form the belief that the bus is scheduled to arrive at this bus stop at 7:00, from which I infer that there is a high probability that the bus will actually be there at 7:00 to pick up passengers. I believe also that I can walk the distance in 10 minutes. Given that I desire to be at the bus stop before the bus gets there, I intend to leave home at 6:45. Barring any change in my belief-desire construct (e.g., I trip as I walk down the stairs and break my arm), I actually do leave at 6:45.
This, then, is an explanation of my actions in terms of psychological states — states of belief, desire, and intention.
These entities lend themselves to the task of explaining and predicting a person’s behavior. “Why did he quit his job?” one co-worker may ask. Another answers, “Because he wanted to get away from his boss.” These are explanatory actions much like, “Why did the dinosaurs die off?”, “Because of a large meteor impact.”
“Why did the chicken cross the road?”
The answer comes to us in the form of a theory about the beliefs and desires the chicken had at the time.
Theories, that attempt to explain an action in terms of beliefs, desires, and intentions, may be mistaken. Further research may reveal that the person who quit his job had spent a great deal of time with the boss outside of work — fishing, going to movies, and entertaining each other’s families at their homes. Against this evidence, the second speaker may need to revise his theory — the employee who quit may not have the beliefs and desires originally attributed to him.
Again, this is a general truth about theories; evidence that there were a great many dinosaurs a million years after the meteorite impact would call into question the theory that the impact caused their extinction.
A. Propositional Attitudes
To say that beliefs, desires, and intentions explain behavior is a way of saying what they do, but it does not say what they are. So, what are these “beliefs”, “desires”, and “intentions”?
Here, I am not interested in what the agent believes; whether he believes in God or desires to teach the world to sing in perfect harmony. I am asking what a belief, or desire, or intention is, generally speaking.
Dr. Rey informed us that these types of states are called “propositional attitudes.”
Okay, that was a useless piece of information. What is a “propositional attitude”?
“A propositional attitude is a mental attitude toward a proposition.”
Fine. Now instead of one useless definition I had two.
Okay, let’s break this up into its parts.
A proposition is the meaning component of a sentence. Let’s just take a sample sentence, such as ‘1 + 1 = 2.’ This is just like saying “One plus one equals two.” They are two different sets of symbols, but they represent the same fact. These are two sentences, but they represent only one proposition.
Okay, I now know what a proposition is. A mental state is an attitude towards a proposition.
What does that mean?
B. Belief
Let’s take belief as an example. Let us say that ‘Agent believes that P’, where P is the proposition ‘one plus one equals two’. In the case of where an agent has a ‘belief that P’, the agent has the mental attitude that ‘P’ is true. That is to say, the agent who has a belief that ‘1 + 1 = 2’, has the mental attitude that the proposition ‘1 + 1 = 2’ is true.
This means that the agent will act as if P were true.
If Agent also has a belief that Q, where Q is the proposition ‘there is a red dragon outside my home waiting to gobble me up,’ he will behave in all relevant instances as if ‘there is a dragon outside my home waiting to gobble me up,” is true.
So, we look at how Agent behaves. We acquire a set of observations and, from this, we form a theory. The theory that we form for explaining Agent’s behavior (including his surprised question to any visitors, “How did you get past the dragon!?”) is best explained by theory that he has a belief that there is a dragon outside his home.
We may discover that Agent is wanted for some terrible crime. In this case, the ‘best theory’ may change. We may consider the theory that Agent does not desire to go to prison, and believes that if he acts like somebody who believes there is a dragon, people will attribute to him a belief that such a red dragon exists, and infer from this that the belief-forming function of his brain is faulty.
This takes us to the concept of ‘desire’.
C. Desires
Those philosophers who liked to write about propositional attitudes seemed almost exclusively interested in beliefs. When it came to desires, they said little more than, “By the way, desires work kinda the same way, almost.”
A belief, recall, is an attitude that a proposition P is true. A desire, on the other hand, is an attitude that a proposition is to be made or kept true.
This identifies the core difference between a belief and a desire. A belief is motivationally neutral; it is simply a record (potentially in error) of the world. A person who believes that there is a red dragon outside of his door holds that this proposition correctly describes the world.
Desires, on the other hand, are motivational. They identify a proposition — a target, an end — and say, “Make it so,” or, “Keep it so.” The person who wants a pet dragon outside of his door (to keep sales people away) holds that the proposition prescribes a possible world — one in which the proposition ‘there is a pet dragon outside of my door” is true.
If Agent has a desire that his child is healthy, and a belief that his child is healthy, will be motivated to a degree proportional to the strength of the desire, to keep P (which, in this case, is the proposition “My child is healthy” spoken by Agent) true. He will do this by getting the child immunized against disease, having the child eat his vegetables and brush his teeth, and ensuring that the kid knows to look both ways before crossing the street. If the parent does not do these things, we have reason to doubt that the agent really cares about the kid’s well-being. He lacks the desire that his child is healthy.
D. The Status of Desires
Somewhere in my studies, I do not remember where, I read of a distinction between fulfilling a desire and satisfying a desire, and of a parallel distinction between thwarting a desire and frustrating a desire.
Assume that an agent has a ‘desire that P’. This desire is ‘fulfilled’ if and only if P is true, and ‘thwarted’ if and only if P is false.
On the other hand, the term ‘satisfied’ refers to the positive emotion associated with the belief that a desire has been fulfilled, and ‘frustrated’ refers to the emotion associated with the belief that a desire is thwarted.
Action aims for the fulfillment of desires and seeks to avoid the thwarting of desires. Agents typically also have desires for satisfaction and an aversion to frustration, thus they also seek the fulfillment of these desires. However, the desire for satisfaction and the aversion to frustration are just two of the desires that exist.
E. Other Propositional Attitudes
There are no other basic attitudes. All other attitudes that people talk about can be reduced to some version of these two.
“Agent knows that P” means “Agent believes that P, P is true, and Agent has good reason to believe that P.” There may be some dispute as to the details of this account, but those details do not concern us here.
“Agent suspects that P” means that “Agent believes that there is a chance that P.”
On the desire side of the formula, “Agent hopes that P” and “Agent wants (that) P” are simply two alternative ways of saying, “Agent desires that P.” “Agent hates that P” means “Agent has a strong desire that not-P”, where “Agent fears that P” (as in, “I fear that the projector will break during my presentation”) means “Agent desires that not-P and believes that there is a chance that P”.
The point to be made from all of this is that, basically, we need only talk about beliefs and desires (the roots of all of our intentions).
“Agent intends that P” is the state that exists between belief and desire and actual action. A person who is paralyzed can form the intention to raise his arm, but breaks in the neurological system prevents him from actually moving his arm. The intent has all of the characteristics of an action except the actual muscle movement.
III. The Explanatory and Predictive Power of BDI Theory
BDI theory, where action aims at making the proposition that is the object of the desire true, very easily handles the imprisoned family and preferred world examples discussed in Chapter 2.
A. The Imprisoned Parent Case
Recall that the Imprisoned Parent case asked us to imagine a prisoner who has been given the following two options:
Option 1: Their child will live a secure and comfortable life, but the prisoner will be caused to believe that the child is being mercilessly tortured.
Option 2: Their child will be mercilessly tortured, but the parent will be caused to believe that the child is living a secure and comfortable life.
Hedonism and other direct-experience psychological theories say that Option 2 is the clear winner. The parent enjoys the greater pleasure with this option. Desire-satisfaction theories (as distinguished from desire-fulfillment theories above) suggest this same conclusion.
However, it seems unlikely that this prediction would hold up. Most people asked this question answer that they will select Option 1.
BDI theory correctly predicts this response. It says that if Agent has a desire “that my child is secure and comfortable” then he will be motivated by that desire to make the proposition “my child is secure and comfortable” true. BDI theory suggests that agents with this desire would pick Option 1.
Of course, we can expect the prisoner to also have a desire for satisfaction and an aversion to frustration. Therefore, the prisoner will feel some motivation to select Option 2. Given the possibility of a third option in which the child lives a secure and comfortable life and the prisoner knows this, the desire for satisfaction will prompt the parent to select this Option. In the absence of the third option (unless the parent really does not really desire that the child be secure and comfortable), Option 1 wins the contest.
B. Moore’s “Preferred World” Argument
The same is true when we apply BDI theory to Moore’s example of preferring that a beautiful world exist even where one was not around to experience it, to preferring that an ugly world exist. A person with a desire “that beautiful things exist” will pick the option in which “a beautiful world exists” is true. Its value will not depend on his actually experiencing that world. He will pick as Moore says he will pick. However, the option does not prove that there exists a basic, unanalyzable concept of ‘good’ that cannot be reduced to natural phenomena.
C. Addictive Rats
I have heard many people attempt to defend some form of hedonism after hearing of research where scientists attach electrodes to the pleasure centers of the brain. Rats who press a lever get a jolt to this area. The jolt is associated with great pleasure. The subjects of this experiment eventually do nothing but press the lever, to the exclusion of eating, sleeping, or any other activity.
Some argue that this provides a defense of hedonism, that the only things that humans do is perform those actions that experience tells them will give them another jolt to this part of the brain.
Of course, “aversion to pain” fits in here somewhere as well.
If somebody offers that “X is the sole criterion of choice,” (where X = a jolt to the pleasure center) we can scientifically look for ways in which this thesis can be falsified. We would do so as follows:
If the thesis “X is the sole criterion of choice” is true, then we need to look for instances where Agent believes that Option A has more X than Option B, but nonetheless the agent chooses Option B. Even one instance of such a case will prove that X is not the sole criterion of choice.
This is the way all science works. A scientist offers a hypothesis. From the hypothesis, he makes a prediction. He tests the prediction in an experiment. The experiment, then, either supports or falsifies the hypothesis.
So, let us test the hypothesis that X is the sole criterion for choice, where X = jolts to the pleasure center of the brain.
This would be true of the scientists doing the research as well, right? It is true of them, their co-workers, their lab assistants, many of their students.
These are also people who are aware of a relatively painless procedure that would provide them with far more efficient jolts to the pleasure centers of their brain than virtually any other option. They are aware of an “Option A” that contains more X (jolts to the pleasure center) than Option B.
However, they still choose against Option A. To the best of my knowledge, there is no rash of researchers in this area of brain science being discovered dead in their homes after having attached electrodes to this part of the brain. The thesis that individuals seek only to jolt this part of the brain and nothing more does not hold up to scrutiny.
The only option remaining is that these researchers look for something other than (or in addition to) “that which provides the greatest jolt to the pleasure centers” in their choices.
If somebody gave me the opportunity to hook myself up to a machine, where I could spend the rest of my life with impulses sent into the pleasure centers of my brain, I would refuse. I would see such a life as being empty. I find this option as appealing as suicide.
As an aside, even if it were true that people care for nothing but maximizing jolts to the pleasure centers of their brain, this is no threat to BDI theory. Rather, the thesis combines BDI theory with an additional proposition that the only “D” (desire) within BDI is the “D” of obtaining jolts to the pleasure center of the brain.
All attempts to reduce human choice to a single factor have fallen upon similar problems. Our desires seem quite resistant to these attempts.
I would not be surprised to discover that desires are like beliefs. People seem capable of believing just about anything. There seems no limit to the propositions that can take the role of P in different people’s “belief that P”. The huge diversity of human behavior suggests that there is also no limit to the propositions P that can become objects of desire. Anything that can be believed can be desired.
D. Reiteration
BDI theory handles the case of the imprisoned parent better than hedonism, adequately handles Moore’s “preferred world” test, and even handles alleged counter-evidence drawn from the behavior or rats. These are just three examples that illustrate how BDI theory has more explanatory and predictive power than psychological hedonism or psychological eudaemonistic theories have.
IV. A Preliminary Look of the Value of Desire
A. “Is/Ought” / “Belief/Desire” Symmetry
The part of the class that I found particularly interesting concerned the definitions of the attitudes “belief” and “desire” themselves. Beliefs are attitudes about how the world is, desires are attitudes about how the world should be. The “is/ought” distinction is perfectly replicated in the “belief/desire” distinction.
This provides us with a new perspective on Hume’s ‘is/ought’ distinction. Hume argued that no set of ‘is’ statements entails an ‘ought’ conclusion. We must include an ‘ought’ in our premises if our conclusion is to contain an ‘ought’.
Interestingly, no set of beliefs entails a desire — or, more precisely, a recommendation for or against some action. If the conclusion is going to contain some sort of recommendation, then the premises must contain at least one desire. The conclusion, then, is justified in terms of being ‘such as to fulfill the desires in question.’
In this, we capture the essence of Hume’s ‘is’/’ought’ distinction. However, in doing so, we are not postulating a split between one universe discoverable by science that consists of all things that are, and another universe of mysterious value type entities that somehow identify what should be. Everything fits comfortably into the world of ‘is’. Desires just happen to be that part of the world of “is” that prescribes states of affairs for the people who have them.
Desires are the entities that bridge the ‘is’/’ought’ chasm.
B. The Subjectivity and Objectivity of Desire
A desire is a propositional attitude. If there is no ‘desire that P’, then P cannot be “such as to fulfill the desires in question.”
This sounds like pure and simple subjectivism.
However, “Agent has a desire that P” is an element in a scientific theory, BDI Theory, that is supposed to obtain its merit from its ability to explain and predict behavior. It is a better theory (than, say, hedonism) in virtue of its greater power in explaining and predicting (as in the case of the imprisoned family example.) If it is scientific, it must be objective, right?
But it still sounds like pure and simple subjectivism.
Could the objectivists and subjectivists both be right?
(a) The Is and Ought of Desire
G.E. Moore raised a criticism that, when J.S. Mill used the word ‘desirable’, he confused the distinction between what people desire in fact, and what people ought to desire. To say that something is ‘desirable’ is not to say that it is something that people want, but that it is something that people should want. It is something that merits being wanted.
To handle this distinction, notice that there is a clear distinction between the proposition, “I desire that X”, and “I desire that Y”, where “X is not equal to Y”. There is no sense in which we expect these two statements to be equivalent.
Now, let us assume that X = “I am eating chocolate ice cream” and Y = “I desire that I eat chocolate ice cream.” There is quite a distinction between the two in that a person can very much desire to eat chocolate ice cream, and at the same time wish that he did not have the desire to eat chocolate ice cream.
So, we can easily account for the distinction that Moore worried about — the distinction between saying that chocolate ice cream is desired, and whether the desire for chocolate ice cream is desired. Moore gives us no reason to question the link between desire and value.
(b) Desires as Rules
As described in Chapter 7, when I was an undergraduate, I looked at rule utilitarianism and I wondered, “What if the rules were somehow written into the brain in such a way that they did not allow for exceptions?” If this were the case, then the principle of ‘ought’ implies ‘can’ will prevent rule utilitarianism from collapsing into act utilitarianism. There is no sense in which the act-utilitarian best act ‘ought to be done’ if it is not the case that a person with good desires could have performed that act.”
In this seminar, I was being introduced to the concept of a desire, which are nothing less than rules written into the brain in such a way that they do not allow for exceptions. Certainly, desires can conflict with each other, but the losing desire does not just vanish. It remains in force, causing a sense of frustration and motivating the agent to find a way in which he can ‘have his cake and eat it too’. This is not a case of the desire is “excepted”, it is just “outweighed”.
Now, apply this to the answer to Moore’s question about “What should I desire?” Rule utilitarianism says that the best rule is the rule with the best consequences. Desire utilitarianism says that the best desire is the desire with the best consequences.
What are the best consequences? Well, the best desire is the desire that best fulfills other desires.
As I said, desires do not allow for exceptions. They do, on the other hand, allow for different weightings. The aversion to obtain consent before using another person’s property does not allow for exceptions, but it may weigh less in a properly motivated person than the desire to save the life of a sick child. Therefore, the parent with a sick or injured child, whose only means available for getting the child to the hospital is a nearby car with the keys in the ignition, may take the car. Yet, the aversion to stealing will still gnaw at the person. It will cause him to hesitate just a moment to find another alternative, and it will make him truly regret a part of what he has done. He will say “I am sorry,” and he truly will be sorry. He will add, “But I did not have a choice.” The absence of a choice will weigh against the aversion to taking property, but it would not cancel it out.
Now, we return to our hapless sheriff who has an opportunity to prevent a riot by framing an innocent person. There is a reasonable argument to be presented that we want our law enforcement officials to have a strong aversion to framing innocent people. If he has this strong aversion, then he will not frame an innocent person even when it comes into conflict with other desires. Yes, he has a desire that the potential victims of the riot not be maimed and killed. Still, the aversion to framing an innocent prisoner will drive him to seek alternatives. He may stand himself before the crowd and say, ‘Over my dead body’ before he asserts that an innocent person is guilty.
V. A General Theory of Value
A. The General Theory
Value consists of a relationship between a state of affairs and a set of desires.
If the proposition, “Agent desires that P” is true for some P, then it is also true of P that it is “Such as to fulfill the desire of Agent.”
In more general terms, words like ‘good’, ‘should’, and ‘ought’ all describe relationships between possible states of affairs and desires. They pick out a set of desires and then claim, of whatever is ‘good’ or ‘should be done’ that it ‘is such as to fulfill the desires in question.” Correspondingly, ‘bad’, ‘should not’ and ‘ought not’ mean that the object is such as to thwart the desires in question.
Some important points to make explicit right away.
(1) The “desires in question” might be more than one desire. A person trying to decide if she should take that job in Atlanta weighs her options against a large number of desires.
(2) The “desires in question” need not be those of the person who is making the claim. Below, I will give examples of terms that refer to desires other than those of the evaluator. In the next chapter (Chapter 14), I will argue that moral value is a value of this type as well.
B. A Wide Range of Applications
A value theory should not just focus on one type of value. A good value theory should be able to look at all types of value and explain not only what they have in common, but what distinguishes one from another. It should be able to drill down and explain the differences between being ‘healthy’ and being ‘useful’; between ‘abuse’ and ‘beauty’; between luck and ‘harm’?
Above, I gave my general account of what all value claims have in common. They relate a possible state of affairs to a set of desires (not necessarily those of the speaker).
It follows that the way one value-laden term differs from another is that they describe different relationships between different objects of evaluation and different desires.
Specifically, a value-laden term has four elements:
(1) The possible state of affairs that provide the standard objects of evaluation for that term.
(2) A set of standard desires for the use of that term.
(3) Whether the state of affairs in (1) will fulfill (positive terms) or thwart (negative terms) the desires in question in (2).
(4) Whether the state of affairs in (1) fulfills or thwarts those desires in (2) directly or indirectly or both.
(a) Injury
Let me use the concept of ‘injury’ to explain this theory by looking at each of the four elements of value in turn.
(1) The state of affairs to be evaluated. “Injury” describes changes in bodily functions that are brought about through interaction with the inanimate external world, or are at least like those brought about in this way. Injuries are commonly held to be distinct from illnesses, though both describe cases where the body’s functioning is somehow impaired. The distinction between ‘illness’ and ‘injury’ can very well be made out in terms of whether our pre-scientific ancestors could easily perceive the cause of the state. If they could perceive the cause, it produced an injury. If they could not, it produced an illness.
(2) The relevant desires. The desires that are relevant for making this type of evaluation are those of the person whose body is being described as having the “injury”. Nobody else’s desires are relevant.
(3) Fulfill or thwart desires? Injuries are bad by definition. Injuries, therefore, thwart desires by definition. Whatever change the perceptible cause has made to the body, it is not called an ‘injury’ unless its effect on the body is something that thwarts the desires of the person whose body has suffered the impact.
(4) Direct or indirect thwarting? An ‘injury’ can thwart desires directly or indirectly. We have a direct aversion to pain (a desire not to be in pain), and any bodily blow that causes sufficient pain counts as an injury as a result. In addition, a broken bone, or a punctured eye, even if it happens to be painless, would still reduce or destroy the usefulness of the affected limb or organ. This too would qualify the change as an injury.
In summary, the word ‘injury’, like all value laden words, describes relationships between states of affairs and desires. Specifically, an injury is a change to the body brought out by a perceptible cause that thwarts the desires of the person whose body has suffered the effects.
No objection can be raised to this account by noting that a soldier, for example, can benefit from an injury or a student who has not studied for a test can consider himself better off if he dislocates his shoulder and cannot make it to school that day. In these cases, the soldier and the student would still rather have obtained the benefits of the injury without the costs. Injuries have costs to the people who have them, they thwart desires, even in those rare cases where their net effect is positive.
(b) Harm
In the writings of one of the leading figures in the Philosophy of Law, I found an analysis of harm that almost perfectly fits the model I use here.
In Harm to Others: The Moral Limits of the Criminal Law, Joel Feinberg defined “harm” as the thwarting of a strong and stable desire. One such strong and stable desire, for example, would be our strong and stable aversion to extreme pain. Take such a person and torture him, and that person is being harmed.
Feinberg distinguished “harm” from “hurt” in that “hurt” involves weaker and less stable desires. Criminal law, he argued, is not a sufficiently precise tool to be concerned with addressing matters of ‘hurt’. It would be like trying to do brain surgery with a chain saw. There is evidence, in fact, that one of the significant problems with the legal system in the United States is that it is commonly used to address matters of ‘hurt’ — issues where those involved should simply admit, “It’s not worth making such a big deal about.”
In reducing Feinberg’s account to the four-part model I am offering here, I get the following:
(1) Relevant objects: There are no limits to the type of objects that this term can be used to evaluate. A person can be harmed by having his arm broken, having the window of his house broken, being threatened by an anonymous caller, or having somebody make false statements about him.
(2) Relevant desires: Only the strong and stable desires of the person to whom harm is being attributed are relevant.
(3) Fulfill or thwart?: Harms are bad; they thwart the desires in question.
(4) Direct or indirect?: A harm can thwart strong and stable desires directly as with a personal injury, or indirectly as with having one’s car stolen so that one no longer has it available to do the things one would desire a car for.
(c) Useful
X is useful if and only if, and only to the degree that, X is helpful in bringing about something else, Y, that fulfills a desire.
(1) Relevant objects: Again, there is no limit to the things that might be useful.
(2) Relevant desires: Any desire can be relevant. In this case, one looks at the context for the word ‘useful’ to determine the desires that the speaker is referring to. A mechanic, disassembling an old engine, might say, “Save that distributor, it could still be useful.” He is likely referring to the desires that are typically served by a fully functioning car. The spy who copies a piece of information would refer to a different set of desires when talking about the usefulness of that information. There is no contradiction in saying that something might be ‘useful’ in one context, but ‘useless’ in another. The same mechanic may call a certain distributor ‘useless’ in the context of helping him to create a fully functioning car, yet still find it ‘useful’ in holding the shop door open on a hot but breezy summer day.
(3) Fulfill or thwart?: Useful things fulfill the desires in question. A useful distributor is one that fulfills the desires of the person potentially employing the distributor; useful information is one that helps to fulfill the desires of the person who receives it.
(4) Direct or indirect?: ‘Useful’ refers to things only in terms of their capacity to fulfill desires indirectly. Works of art, for example, are seldom known as being ‘useful’ because the value that they realize is direct. This is not to say that a work of art can’t be useful. This is to say that its usefulness would consist on indirect effects, such as in impressing a potential lover or members of a social group.
(d) Beauty
(1) Relevant objects: Typically, we use the word ‘beautiful’ when considering things that are seen or heard. It can apply to a sunset, or to a song. We can use it to refer to another person’s appearance, or to her voice. We tend not to use the word ‘beautiful’ to refer to how things taste, smell, or feel.
(2) Relevant desires: These are usually those of the speaker. A person who says, “that is a beautiful song” typically judges the song by listening to it and judging how it affects his or her own desires. It is sometimes the case that the term refers to “common desires” or “desires that are common within a certain community”. An art critic, for example, need not evaluate art according to his own likes or dislikes, but those of the community that he is attempting to serve.
(3) Fulfill or thwart?: Beauty fulfills desires. Ugly thwarts desires.
(4) Direct or indirect?: Beauty fulfills desires directly. To call something ‘beautiful’ it must be something whose value can be realized by direct experience. In this, what is ‘beautiful’ is distinct from what is ‘useful’.
(e) Knives and Tables
We often do not think of words like ‘knife’ and ‘table’ as value-laden terms, yet they do have value elements contained within their meanings.
Aristotle had an argument to the effect that a good knife is something that fulfills well the function of knives. Knives are meant for cutting, so a good knife cuts well. Tables are made for holding things at a convenient height, so a solid and stable table at a comfortable height is a good table. A sharper knife is better than a duller knife; a more stable table is better than a less stable table.
When humans invented the word ‘knife’ (or, at least, as its meaning evolved), it seems reasonable to believe that the inventors built the purpose for which knives were to be used into the meaning of the word. A knife is an instrument, or a particular type of instrument having certain physical qualities, that exists for the purpose of fulfilling a particular set of desires. These desires, then, become the ‘desires in question’ when one is concerned with whether a particular knife is a good knife.
‘Knife’ and ‘table’ are not value-laden terms in the sense that calling something a ‘knife’ automatically identifies it as having some sort of goodness the way that calling something a ‘blessing’ does. However, words like ‘knife’ and ‘table’ include as a part of their meaning some of the parameters that are to be used in determining the difference between a good knife and a bad knife, or a good table and a bad table. They include, as a part of their meaning “the desires in question.”
This explains how we fit these uses of the word ‘good’ into the overall model described above.
VI. Summary
These are examples. I do not mean these descriptions to offer a full-fledged linguistic analysis of any of these terms. I simply wish to illustrate the way in which this proposal offers both a description of what all value-laden terms have in common, and a way of specifying their individual differences. I will leave it to others to work out the specifics.
The one set of terms that I seek to look at in greater detail myself are moral terms. The whole of Chapter 14 will be devoted to this task.