Separation by Fact and Value

We now come to our first example of a potentially insightful Separation Principle. (It seems to be endorsed by no less an intellect than Einstein, in his "Science and Religion.")

Notice that this proposal does not claim that religion alone tells us about values; it allows for the existence of other sources. It does say that science is different from religion in this respect.

Second, we have to understand the word "ought" (or "value") correctly here: there are several different uses of it, only one of which is relevant. One kind of "ought" is an "ought" of practical advice, a prudential ought: "if you want to rob a bank, then you ought to use Uzi's," "if you want the house to be well-insulated, then use 2 x 6's, and not 2 x 4's." We mean simply to give a bit of practical advice about the best means for achieving a certain goal, and intend no evaluation of the goal itself. Another kind of ought is an etiquette ought, or the ought of Emily Post: "You ought to place the fork on the left of the dinner plate," "You ought not belch loudly after meals." This kind of "ought" tells you how to conform to local customs. (I'm told that in some locales, belching loudly after meals is taken as a compliment to the host. I know some people who should consider moving there.) Maybe it is a kind of prudential "ought". There is also a thought "ought" - an "ought" of 'right thinking': "you ought to rid your (arithmetic) beliefs of inconsistencies", "if you have no independent evidence for a belief, then you ought not hold it". These three contrast with the moral "ought": "You ought not torture innocent infants", "You ought to do unto others as you would have others do unto you (unless you're a sadomasochist)", "One ought to treat people with respect and dignity", "You ought to keep your promises, but if you break a promise, you ought to apologize".

I assume that there are objective truths of morality, that it is not all 'merely a matter of opinion'. One very bad reason sometimes given for thinking that 'it's all relative', is that there is a lot of (sometimes very persistent) disagreement about moral issues. But that doesn't show that there are no truths of morality. Disagreement can come about for two other sorts of reasons, consistent with there being objective moral truths: it may not be clear what the truth is, though there is one there to be discovered; and there may be agreement on fundamental principles, but disagreement about how to apply them. One good reason for thinking that there are truths about what we ought to do and about what we ought not to do, is that we know some of them, a fact that we forget when we focus, quite understandably, on the unresolved cases. But we can all think of clear examples, where it's not 'just a matter of opinion': kindness to a sick child that aids his recovery and harms no one, is a tolerably clear example of a right act; the Nazi extermination by torture of over six million human beings, including many thousands of infants and children, was clearly wrong. I trust that you are not inclined to assert, "Although I don't go in for infant skull smashing myself, Hitler and his henchmen thought it was right - it fit in with their 'value system' - so it was what they ought to have done. OK for them, not-OK for me. But we're all OK." (More on "moral relativism")

So I assume both that there are truths about morality, and that we sometimes know them. This is not, of course, to say that we get it right most of the time, that it is easy to get it right, or that we tend to answer the important moral questions correctly.

These two assumptions, which ought to be uncontroversial, create a problem for the separation principle. A statement is true if it accurately describes a fact. So, if there are (knowable) truths about morality, then there are facts about morality; and if there are facts about morality, then they are parts of how things actually are. A list of facts about the world might contain statements like these:

But the separation principle says that science attempts to discover facts, and now we are reminded that among the facts are moral facts.

There's an obvious response to this obvious problem: what's intended is a distinction between two kinds of facts, moral and non-moral, and science is concerned with the latter alone. This is a good first response, but it has to be treated carefully here. We don't want to render the proposal empty by simply stipulating

so we need some independent way to characterize the overall difference between moral and non-moral. Let's look at two contrasting approaches to the analysis of the meaning of moral terms. Then we'll be able to see the main weakness for Separation by Fact and Value.

Two Approaches to Analyzing Moral Terms (after Feldman)

It will be convenient if we focus on some one moral term. It's grammatically simpler to talk about "good" than to talk about "ought", so that's what I'll do. The two approaches are Ethical Naturalism and Ethical Anti-Naturalism. Oddly, it's easiest to present the "anti" view first.

Let's begin by distinguishing between two different sorts of characteristics. Suppose I ask you to list some properties that this book has. You're likely to mention its shape, weight, color, perhaps something about the materials in it (paper, ink, etc.). If you want to be very thorough, you might do some chemical tests on the ink, or even check for radioactive isotopes. In doing all this, you are using your five senses, sometimes by extending their reach with scientific equipment, to discern what you can about the object. Call something a natural property if it's a characteristic that you can detect by using your five senses, possibly augmented by extra equipment.

According to Ethical Anti-Naturalism, goodness is a real, objective property of objects and actions, but there is no scientific or observational method that can even so much as help us to determine whether a particular thing or action has it. Goodness is not a natural property. Nor is it analyzable as some cluster of natural properties. According to this view, if two individuals disagree about whether or not a certain thing or action is good, they may be unable to come to any agreement, but exactly one of them is right, though there is simply no way to determine who it is. According to Ethical Anti-Naturalism, goodness is indefinable, unanalyzable, and undetectable.

Ethical Naturalists find this view unacceptable because it makes moral knowledge impossible in two ways. We want our children to acquire the concept of goodness. The psychology of learning and concept acquisition is a big, complicated subject, but in many cases, we begin with examples. If you want your daughter to know what "red" means, you'll show her lots of red things (and tell her they're red) and you'll show her lots of nonred things (and tell her they're not red), and, with amazing speed, she'll pick up on the relevant difference (if you choose your examples judiciously; obviously, if every red thing you show her is also round, you're going to confuse her). Nothing like this can work if goodness is a nonnatural property, since nothing your daughter can do with her senses (even if she drags in equipment from the lab) will allow her to detect the goodness in the fifty thousand examples you show her. So, if Ethical Anti-Naturalism were correct, no one could learn what goodness is.

I've sometimes been offered the Revelation Defense of Ethical Anti-Naturalism: No one can know any moral truths if they are not revealed, directly or indirectly, by God. I have two questions for anyone who makes this claim. Isn't it possible that an isolated community, deprived of contact with the Bible or other scriptures, could develop some appreciation of moral truths; e.g., mightn't they come to see that the wanton torturing of infants is wrong? Second, supposing that some fundamental moral principles are revealed or instilled (in all human beings?) by God, we will need to be able to apply them to individual cases. Won't that require an ability to perceive goodness in particular things? Unless a special mental power of moral intuition is built into us by God, it seems that we must be able to perceive goodness in things and acts. If there is such a power of moral intuition, how does it work, why does it sometimes fail so miserably, and why is there so much disagreement about moral matters?

So, a second way in which Ethical Anti-Naturalism makes moral knowledge impossible is this: if, by some miracle, you could get the concept of goodness into your mind, there'd be no way for you to apply it to particular cases to tell whether it was present or absent.

Naturalists claim instead that goodness can, after all, be tied down to what we can perceive or feel. The Ethical Naturalists claim that goodness can be explained by the use of terms that each express an ordinary, observable property. Different Naturalists have different proposals about how exactly to define or analyze goodness, but they all agree that it can be analyzed, and analyzed purely in terms of natural properties.

There are many good books on ethics, and I leave it to them to delve into the many Naturalist analyses. For our limited purposes, it will be enough to examine just one suggestion about how science might be a source for knowledge of moral truths.

Science, Normal Function and the Well-Being of Human Beings

Given two highly plausible premises, we can argue for the moral relevance of science.

(i) What is good for an individual, what constitutes her well-being, what she ought to do, and what she ought to be, all depend directly upon her nature. It is almost a commonplace that what is good for human beings depends on human nature, that their well-being is determined by what kind of being they are. For example, what kind of government and society human beings ought to have depends upon what their natures are. Or, to take a more obvious case, torturing people is not good because it harms them; and it harms them because of the nature they have. We can put this same homey truth in other words: there are certain conditions and states it is normal for a human being to be in, states and conditions conducive to their well-being; some departures from normal function are not good, and ought to be avoided if well-being is to be maintained. (Let's put aside questions about supernormal function.)

(ii) It is possible to discover, by means and methods familiar in science, what is and what is not normal for a human being. How an organism functions normally is something that biologists are often quite successful in discovering. There seems to be no reason to believe that medical doctors and psychologists have not or can not discover a great deal about normal human function, or human well-being. I must stress here that I am not talking simply about average function. There is a difference between normal and average: even if a whole population of organisms suffered from some sort of harmful disease, (e.g., an unusually bad flu epidemic) we would say that, on the average, their function was then below normal. It is normalcy in this optimum, well-being sense, as well as statistical averages, that science can discover. Based on these two plausible assumptions, we can give an argument for the moral relevance of science:

In the second premise, some form of Ethical Naturalism is assumed to be correct. It is a substantive assumption. But it certainly seems to be true. It may even be difficult to imagine how human well-being could otherwise be discovered. Nevertheless, I would mislead you if I didn't tell you that this argument, and the Ethical Naturalism it is based on, are highly controversial. So I won't pretend to have refuted Separation by Fact and Value. But I think that I have raised some serious questions about its adequacy.

What of Einstein's proposal? (Do I think I'm smarter than he was?!) He took the rather strong position that conflict between religion and science is impossible; here's his reason:

He does, however, see the need for "strong reciprocal relationships and dependencies":

What is Einstein proposing here?: he is proposing a redefinition of religion and science, how they should be seen, and not primarily how they are or have been seen (-as he implies in his historical remarks, 109). He wants to redraw the map. His proposal is not without its odd consequences. On his view, Marxism, with its emphasis on higher goals, would be at least part religion. He also proposes a redistribution of labor and resources: to remove from religion much of its traditional explanatory role, and to give that job to science alone. So, at best, Einstein is offering a personal vision of science and religion, and proposals for changes in, and reconciliation between, the two. (His vision is not unique to him, as he notes. A much more sophisticated version of it is defended at length by Robert John Ackermann, who accepts Marxism as a religion.) There is nevertheless reason to treat his proposal with some caution: there may be more fact in function (goal and purpose) than he allows. And even if the job of explaining is divided between the two, religions will still want to say a great deal about "facts and relationships between facts", on any reasonable interpretation of "fact".

So I think that we need to keep looking for a separation principle. Next, everyone's favorite: religion is based on faith, but science is not.

Additional Sources

2000 David F. Austin

Go to next section Separation by Foundation