"It never fails!"
"Won't I ever learn?"
"You say that every time!"
"You always do things backwards!"
That manner of speech is one of life's major annoyances. It's part of the "all or nothing" style, which is "always" criticized, but "never" understood. It should have a name, and be examined closely, because it can help to show the essence of what is called induction.
Let's call it a mock universal. "That always happens," really means, "That happens so much that it seems universal." It compares to a true universal: "All men are mortal." The true universal specifies a group, then says something about it. The mock universal specifies all or nothing, but is careless about the rest.
"You always do that," is really an emphatic way of complaining that you do that more than I like. It gives emphasis by expanding one action into a universal action. In a psychological context, this expansion is used as an excuse:
- "I know, I always do that; I can't help it."
- "It's the way I always react; it doesn't mean anything."
- "It's just the way I am."
Thus, the same form can be both accusation and defense:
- "You always do that!"
- "I know. I can't help it. That's just the way I am."
The implied excuse is that I did it not from personal choice, but from universal compulsion. It was not a subjective decision, but an objective necessity. It did not come from in here, but from out there.
The mock universal intensifies not just by exaggeration, but by pretending to move from "in here" to "out there"—from this truth to the truth. It is the real version of this movement from particular to universal that differentiates an induction from a guess.
Induction is the widest term for a process of deriving the general from the specific. It is what you are doing when you decide that one lie makes a liar. What some call the "problem of induction" is a question of method. It's how you move from "That's what happened," to "That's what happens." The subjectivist simply declares that it is so, making no distinction between a mock universal and the real thing. Why? Because the subjective mental arrangement does not contain any way to make such a distinction. It just says: "He always lies."
Here's how it looks to the subjectivist:
lie è ME è liar è ME è him
If I relate the person and the lie primarily through their effect on me, the word "liar" names an association rather than a fact. While the lie is fresh in my mind, whenever I think of the person, I think of lying. "He lied," "He lies," and "He's a liar," all have the same meaning to me, because my focus is on the relationship to me, rather than on the relationship between lying and him. If I say, "He always lies," then I don't know whether that is a mock universal or a real one: it's just what's there in my mind. Later, when the dotted lines have faded from memory, I may have other associations, and forget this one. I may be genuinely surprised to hear this: "Aren't you ever going to catch on? How many times are you going to fall for his lies?"
To stop being a sucker, I need to figure out the truth. I need to make a judgment which stands by itself, independent of me—an objective judgment.
How can I make my judgment stand independent of me? Well, I can consult you. I can ask you what your judgment is. But that's not what I really want; I want my judgment to stand on its own feet, independent of us both. I want to stop saying, "I guess," or "It seems to me." I want to say, "It is." I want to consult reality.
This consideration makes me see what is wrong with my mental arrangement. It leaves out the nature of things. I need to take myself out of it, and instead put in why. I can see the fatal flaw right there in the diagram: what is labeled "me" should be labeled "causality."
The law of causality, says Ayn Rand, is the law of identity in action. To find out how things will interact, you investigate their nature. To see why, you find out what.
The diagram becomes a set of two questions:
- Why did he lie this time?
- What in his nature causes him to lie?
Since human consciousness regulates human actions, and also regulates itself, the second question answers the first, and reduces the problem to one question:
- Why does he choose to lie?
We have used the conceptual method to simplify the problem. All the parts of the diagram have been telescoped together—integrated into one. It is a question of examining the nature of the person.
Now we have a limit. We have defined the area to investigate. We can try the inductive leap. Having defined a limit, we can start at that limit, and see if our induction fits. That is, in this case, we can try saying: "It is his nature to choose the lie instead of the truth." Or, in other words, "He always lies."
Are we back where we started? Not at all: that statement is no longer a mock universal, or a guess. It is now an easily tested hypothesis. No guessing is necessary. We have only to observe.
What we observe is that he does not always lie.
Have we failed? Not at all: we now have two limits firmly established. He does choose to lie, but not always. We defined an outer limit, began testing at that limit, and can now proceed to narrow down the induction, confident that we are not missing the essence.
Remember our method: it is the conceptual method viewed as a search for truth. Accordingly, we did not look at his lie as an isolated action; we looked at it in the context of chosen actions, then narrowed it down to the context of his chosen actions. This puts us in position to ask: how does lying integrate with the rest of his choices? How does it fit in with his ideas?
We now have things sorted out to the point where we know what action to take to reach the goal. To find out if he is, in objective truth, a liar, we have to find out what he thinks of lying. Does he believe in the efficacy of lying? Does he regard it as a viable policy? If so, then he will choose it in the future. If so, he's a liar.
To me as a subjectivist, the important thing is what I think of his lying. To find objective truth, the important thing is what he thinks of it. To find truth in general, the important thing is not what my opinion of things is, but what the nature of things is.
Note that we did not focus on his lie and try to work outward, one lie at a time, until we reached a point where we felt justified in making a generalization. That is the quasi-perceptual method—the method in imitation of a snail. In using such a non-method—often called enumeration—we would be ignoring, among other things:
In the process of making the induction, we lost all necessity to count how many times he lied. It ceased to matter whether he lied one time or a thousand times. What matters to being a liar is not the instance, but the reason. A common error about induction is that it consists of building up instances—one, and then another, and then another—until a point is reached where a general conclusion can be defended. Induction proceeds in the other direction: it starts at the outer limit and gets narrowed down to a defensible position.
To establish that living things die, we observe death; but we do not start counting deaths, and decide that some large number will amount to "all." We zoom in to investigate the cause, and we zoom out to look at the context. Machines wear out. Things deteriorate. Living things heal themselves, but less well as they get older. Organisms produce replicas of themselves, possibly for a reason.
The method is integration: fitting everything together, with emphasis on the nature of things in interaction—on causality.
The key to understanding induction is to remember that when you sort reality out, you are sorting all of it out. Within the context of an arrangement of all reality, you wonder if an observation of one thing in a class can be applied to the whole class. You extrapolate, or infer. It's true here; so it's true everywhere within these limits—as long as the same cause operates everywhere within these limits. To see if that's so, you look for exceptions within the limits, and connections outside the limits.
To see how the process of induction is implied by a child sorting shapes into piles, imagine that the child finds a round shape in the pile of square shapes. "No! This is not supposed to be here! It belongs in that pile." That is, the shape goes in that pile not from personal preference, but because of the nature of things. Induction is a process of formalizing a sorting-out of reality, so it becomes a principle.
The purpose of sorting things out is to deal with them. Even a bad sorting-out is better than none, because units must be reduced to a number your mind can manage. A good sorting-out can help not only you, but everyone, everywhere, forever. As a principle, it can be tested, refined, communicated, and incorporated in anyone's mental arrangement who is willing to understand it rather than just memorize it. A good principle can get so universally used that it becomes a law.
We all take gravity for granted. We see an apple fall in a mental context relating this fall to all other falls, and to weight, and to vast areas of reality. The first man to see an apple fall in this context was not Adam, but Isaac Newton. Imagine yourself for a moment as an Etruscan, centuries before Newton. When you saw an apple fall, it was not obeying natural law, but proclaiming its readiness to be eaten, or pleasing the gods, or displaying its appleness. Its fall had no relation in your mind to a bird's flying or a feather's floating. Without this universal relationship we call the Law of Gravity, you would handle things mentally in a more laborious and inefficient way.
One might test an induction's quality of lawfulness by trying to do without it. Gravity, Earthly Rotation, and Thermodynamics would pass the test; nobody now would even know how to do without them. The Flat Earth Theory would fail. One might go a step further and say that an induction is done when it cannot be undone. First, it cannot be undone by finding exceptions. Then, it cannot be undone without sacrifice, since it enables you to handle things better. One thing it enables is deduction.
Imagine that you are setting out to justify the proposition: "All men are mortal." Using the conceptual method, you keep things sorted out. You keep track of where you are in the totality of things. You start by checking the context. You widen out the focus and see if there is a larger limit to your mental pile of humans. Well, men are living things. All living things are mortal. You form a syllogism: all living things are mortal; all men are living things; therefore all men are mortal.
This demonstrates the advantage of principles arrived at by induction. If we can establish the truth of the principle that all living things are mortal, then we don't have to establish separately that all trees are mortal, all dogs are mortal, all men are mortal. We can establish those by deduction—showing that they are part of the induction.
Deduction is the conceptual method viewed not as searching for truth, but as applying truth. It is applying what you know to the question at hand. It is especially popular with arm-chair experts, because it does not involve investigating the nature of things, but applying what is known already. A syllogism is a statement which says: "This thing I wonder about it is part of this thing I know about, so I can stop wondering."
When Sherlock investigates a crime scene and makes deductions, he is not investigating the nature of things, but the disposition of things. He finds some cigar ash. Having written a monograph on cigar ashes, he is in a position to apply his knowledge: this ash comes from a cigar only sold in one shop in....
A deductive syllogism is often accepted as final, when all it does is include something in an unexplained induction. "It's raining here, so it'll be raining downtown," has a satisfying, deductive ring to it. But what it really says is, "If the rain here means that it is raining all over, then that includes downtown." The real statement of interest is the hidden induction: "Since it is raining here, it is raining all over." The next time you figure that since it's nice here it'll be nice there—and get wet—remind yourself that what you thought was a deduction was in fact a mock induction: a guess.
Deduction is often useful for solving problems; but it is not, as some pretend, the main method for solving problems. It is not another name for analysis. To understand problem solving, we should analyze analysis.
|Next Chapter||Previous Chapter||Contents||Home|