Explaining Behavior 2: Styles of Explanation
What does an explanation look like?
In the last installment, we distinguished explanations from other types of idea that also get called “theory,” especially in sociology and related fields.
An explanatory theory, one that tells us why things are as they are, requires some statement about how one thing is empirically related to another. For instance, to explain why crime increased, we might posit that it is because poverty rates have gone up, and more poverty leads to more crime. Or we might propose that as police presence on the streets goes down, crime goes up, and explain the spike with a decline in policing. Either way, we explain by connecting crime to some other thing in the world.
This sort of theoretical idea is often called a proposition, or a principle. If we’re not very confident in it, or if it’s a specific idea derived from a more general proposition, we might call it a hypothesis. If we’re very confident in it we might call it a law.
An explanatory theory has at least one proposition in it. Since little of what gets called theory has these sort of statements, and not all theorists are savvy enough to highlight them, let’s consider what they might look like.
Forms of Proposition
In the hard sciences, such as physics, propositions take the form of equations. Newton’s theory of gravity (also called his laws of universal gravitation) is expressed as:
f = gm1m2 / r2
If we parse this in ordinary language, it decomposes into two propositions: The force of gravity between two objects increases with the product of their masses, and the force of gravity decreases with the square of their distance from one another.
Sometimes social scientists use equations as well, but it often has the flavor of a cargo cult. This is because the variables involved aren’t measured with anything resembling mathematical precision. It’s just not clear what mathematical expression adds when the predictions are, at best, ordinal — a statement that something ought to be more common in Sample A than in Sample B.
Incidentally, I’ve seen a few economists call their regression equations, complete with error terms, a “theory.”
More commonly, though, propositions in sociology and related fields are expressed in words. These might be statements about causality: Famines cause revolutions; downward mobility causes suicide; because he drank too much, he got in a car accident.
The cautious will express them in terms of probability or likelihood: Economic depressions make civil unrest more likely; low self-control makes one more likely to commit crime.
Or they might be coy about the direction of influence and simply state the nature of the correlation between the two: Suicide increases with social isolation; altruism varies inversely with social distance.
Or they might invoke conditional, if-then language: If strain is acute and an outgroup member is easily accessible, scapegoating will occur.
Or perhaps they use the language of “when” or “must”: When people feel frustrated, they are more likely to turn violent; For a revolution to happen, the following three things must occur….
Or they might talk about motive and reason: People who want to get ahead are likely to adopt the language and beliefs of their social superiors; Many people drink heavily because they want to dull emotional pain.
The better class of theorists will draw attention to their propositions — number them, indent them, italicize them, or what have you. For instance, throughout Donald Black’s The Behavior of Law we see the main propositions blockquoted and italicized somewhat like this:
The relationship between law and relational distance is curvilinear.
And the text then goes on to clarify exactly what this means and give examples of the patterns it is meant to explain. An adequately bright reader will have little doubt that is a main idea, even if they’re unclear about how exactly it explains anything (which we’ll get to below).
Others writers bury their propositions in the text with no fanfare, and so demand a little more attention from the reader.
Still others aren’t good at summing up their main idea, but give a rather lengthy argument that one can boil down to a single empirical relationship. Though he never outright says it, one could fairly parse Robert Merton’s theory of crime into a few propositions such as: “To the extent someone lives in a culture that emphasizes achieving goals more than it emphasizes using legitimate means, and to the extent that person lacks access to legitimate means, they are more likely to commit crime.”
Styles of Explanation
Another complication that comes in recognizing the core idea of an explanatory theory is that explanations sometimes operate with different standards. Philosophers of science spill a lot of ink arguing about exactly what makes a proposition count as an explanation.
Some, like Nancy Cartwright, argue that it specifically has to invoke causality, including time-order and direction of influence, and that anything less doesn’t really count. By that standard, things like Newton’s laws don’t count as explanatory theories, which allows philosophers like Cartwright to make provocative statements about how the laws of physics don’t explain anything.
Physicists tend to disagree. One physicist, Stephen Weinberg, also notes a different reason for charging that his field doesn’t explain anything: Physical theory doesn’t tell us the reasons why the universe is like it is. What’s the purpose of it all? It appears some people don’t consider something a satisfying explanation unless it gives us a goal or purpose for the thing being explained.
Sociologist Donald Black, in his course “Sociological Explanation” at the University of Virginia, bypassed these arguments by recognizing that explanation comes in different styles. He identified three styles: teleological, causal, and deductive. Each has a slightly different logic, positing a different sort of relationship between the thing explained and the thing we explain with. And so each has a different standard for what makes a proposition a satisfying or complete explanation.
The Teleological Style
In the teleological style, the propositions state a goal or motive for the thing being explained. That is, they posit a mean-end relationship between two or more things. Why do people commit suicide? Because they want to end their pain. Why do people commit theft? Because they desire material goods. Why did humans develop the ability to brew beer? Because God wants us to be happy.
While some philosophers of social science argue that motive or purpose is functionally the same as a cause of human behavior, others point out that it gives us additional information. Knowing the goal of a behavior might allow you to make inferences about what other sorts of behaviors the actor will adopt if his current course of action becomes impossible.
For instance, if we know someone drinks to numb emotional pain, we might predict that, if barred from drinking, he would turn to some other common coping mechanism, including other sorts of drugs. Or if we know the cat is running in order to catch the mouse, we might predict other actions toward the same goal, such as if the mouse ducks into a hole, the cat will stick its paw in after.
Teleological explanation is probably the most common style in everyday life, especially when we address the behavior of living things. Philosopher Charles Taylor observed that teleology is even baked into our language for describing the activities of animals, such that words like hunting or hiding or fleeing all imply goal-directed behavior.
There’s even evidence that teleological explanation is a kind of human default, as if we’re hardwired for looking for functions and motives. For instance, there’s research arguing that children give teleological explanations for pretty much everything, including things that a Western adult would explain differently. “Why is the mountain top pointy?” “To keep people from sitting on it.”
In earlier periods in history, even sophisticated thinkers used a teleological physics. The natural philosophy of Aristotle, widely accepted in Western Europe from ancient times up until the seventeenth century, was built around teleological ideas. According to Aristotle, all objects in reality desire to be in their natural state, and so strive to achieve that state. For example, heavier objects had a stronger desire to be on the ground, and so therefore they fell faster than lighter ones. About the ubiquity of teleological physics, Donald Black writes, “the paths of the planets were explained with the pursuit of their own peculiar destinies, for example, rain was explained with its contribution to the growth of crops beneficial to humanity, and both were part of God’s larger plan.”
Teleology largely receded from modern physics and chemistry, and beyond the descriptive level has far less place in biology. But, Black argues, it still permeates the social sciences. Some explanations are purely teleological, stopping and calling it a day when they identify the motive or purpose behind something. This includes some of the lazier examples of functionalism (“societies evolve schools in order to socialize the young”) or Marxism (“religion exists in order to keep the proles from revolting”). In other cases, teleological ideas are only part of a large body of propositions that include other ideas, like causal explanations.
The Causal Style
As Black defines it, causal explanation involves explaining something with prior events. According to physicist David Bohm, causal explanation is based on the fundamental principle that “everything comes from other things and gives rise to other things.” Or, as philosopher David Hume put it back in 1748, one thing is the cause of another when “if the first object had not been, the second never had existed.”
The cause of a behavior can be distinct from its motive. Consider two explanations for the same behavior:
Bill jumped off a building because he wanted to end his suffering
Bill jumped off a building because his wife left him.
Both might be accurate, though one points to the motive or reason for his behavior, and the other to the event that sparked it.
A causal relationship is one of asymmetrical influence, in which a cause produces the effect, rather than the effect producing the cause (though there can be feedback loops in which two variables influence each on in turn). Time-order is thus a key feature of causal relationships: Causes must, in principle, precede effects in time, while the reverse can never be true. The stimulus can only cause the response if it comes before the response.
But not all antecedents can qualify as a cause. In his book Moral Time, Black argues that an event — a war, a killing, an argument — is something dynamic, and can never be explained by something static — that is, a background condition that has not changed. If Joe was poor the day he shot his old lady, but was equally poor for the past five years during which he didn’t shoot anybody, it doesn’t make a lot of sense to say being poor caused him to shoot his old lady. After all, it didn’t cause it yesterday or the day before. To adapt a phrase from anthropologist Gregory Bateson, the cause is the “difference that makes the difference” — such as if Joe just found old his old lady was running around.
Along these lines, philosophers and scientists often make a similar distinction between background conditions and the immediate or proximate causes of an event.
Suppose that after shooting his old lady, Joe jumps from a building and dies. True, he would not have died were it not for his being on the surface of a planet with a substantial gravitational pull. So we might say gravity was a necessary condition for his dying. But it was a background condition, not an immediate cause. After all, the gravitational pull had been there, unchanging, all his life, and have never before led to ill effect. The proximate or immediate cause of his death was his jumping off the building. It was the difference that made the difference, the action that produced the reaction.
Because of the centrality of time-order, casual explanation is the dominant style for explaining historical events, from particular events like the extinction of the dinosaurs to general classes of events like political revolutions. It is also central to explaining what happens in controlled experiments, where a stimulus (such as a dose of drug) is understood to cause a response (such as recovering from an illness). And the controlled experiment is the gold standard for trying to measure a causal relationship.
The Deductive Style
Newton’s laws of gravity certainly aren’t teleological — they don’t invoke a purpose of any kind. But critics like Nancy Cartwright are correct that they don’t contain any statement of causality, either. There’s nothing about time order between events. Instead there’s a correlation, a quantitative relationship stated in mathematical terms. So what does it mean to say that Newton’s laws explain this or that? What and how are they explaining anything?
Philosophers of science such as Carl Hempel understand the kind of explanation that happens in modern physics as a process of logical deduction. This is sometimes called the deductive-nomological or covering law model of explanation. Or, as Black calls it, the deductive style.
In deductive explanation, something is successfully explained when it can be logically deduced from a general proposition. The propositions all take the form of some sort of law-like regularity in the universe. We apply this this proposition or law to some particular conditions. We can then deduce what particular pattern ought to obtain under those conditions. Consider some simple examples:
Gravity increases with mass. The moon is less massive than the Earth. Therefore, gravity is weaker on the moon than on the Earth.
Suicide increases with social isolation. There’s more socially isolated people in Town A than in Town B. Therefore, suicide rates are higher in Town A.
Price inflation increases with money supply. The money supply is greater in Zimbabwe than in Kenya. Therefore, price inflation is greater in Zimbabwe than in Kenya.
In this style of explanation, we’ve successfully explained the particular pattern when we show how it’s a logical consequence of the proposition operating under the given conditions.
So in this style, when poor Joe jumps to his death, we consider his jumping off the building a mere initial condition. We explain why he falls by appealing to propositions about gravity, which tell us that unsupported objects on the surface of the earth will accelerate toward its center at 2.2 feet per second per second. Joe happened to be an unsupported object on the surface of the earth, and so his falling is a logical outcome of our theories of gravity and motion. In this sense they explain his motion.
Of course, part of what makes such an explanation useful in the case of physics is that the laws of physics allow one to calculate the trajectory and velocity of a falling object with great precision. Given the height of the building, we can predict how fast Joe was going when he hit pavement. But the deductive style operates with the same logic even in less precise fields.
One example is Black’s work, such as the proposition from The Behavior of Law quoted above. The proposition is about how the quantity of law — defined as the amount of governmental authority brought to bear against a deviant — varies across different social conditions. It asserts that, just as the force of gravity has a predictable relationship to physical distance, so too the force of law has a predictable relationship to relational distance. As Black elaborates in the text, when both parties to a crime or conflict are from the same society, the more distant they are, the more likely legal intervention and the greater that intervention is likely to be.
We can apply this general proposition to different sets of conditions. Compare, say, a sample of women raped by an acquaintance to a sample raped by a stranger, and you would predict there would be more legal intervention in the second set of cases, because strangers are more distant than acquaintances. We would thus say the proposition explains why women raped by strangers call the police at higher rates; it’s an instance of a broader regularity in how much government intervention a crime produces. We see the same regularity in cases where people convicted of killing intimates get lighter sentences than those convicted of killing strangers, and so and so forth.
The deductive model of explanation generates a lot of controversy among philosophers of science. Just as there are those who say it’s not really an explanation without invoking causality or teleology, there will be those who claim that every explanation at least implicitly assumes a general proposition from which one can deduce the thing being explained.
I don’t think that is necessarily the case, and that it is useful to think of the deductive style as a particular kind of explanation alongside causal and teleological explanation. A source of confusion, perhaps, is that to apply a general theory to a particular example we must always use some sort of deductive logic: It is the only way to proceed from general proposition to specific case. But not all general theories are purely deductive in style: The general propositions may assert causal or teleological relationships. But because deductive reasoning is so crucial for using and testing a general theory, the deductive model of explanation deserves more attention. Thus the next installment of this series will delve further into the deductive model and ask what exactly a good explanation can do.
Next installment: The Deductive Model
Further Reading:
Bedau, Mark. 1992. “Where’s the good in teleology?” Philosophy and Phenomenological Research 52: 781-806.
Black, Donald. 1976. The Behavior of Law. New York: Academic Press.
Black, Donald. 1995. "The epistemology of pure sociology." Law & Social Inquiry 20: 829-870.
Black, Donald. 2000. “Dreams of pure sociology.” Sociological Theory 18: 343-367.
Black, Donald. 2004. “Contemporary Sociological Theory.” Graduate seminar taught at the University of Virginia. Charlottesville, Virginia: August-December.
Black, Donald. 2011. Moral Time. New York: Oxford University Press.
Bohm, David. 1959. Causality and Chance in Modern Physics. Philadelphia: University of Pennsylvania Press.
Braithwaite, Richard Bevan. 1953. Scientific Explanation: A Study of the Function of Theory, Probability and Law in Science. New York: Free Press.
Cartwright, Nancy. 1983. How the Laws of Physics Lie. New York: Oxford University Press.
Clagett, Marshall. 1948. “Some general aspects of physics in the middle ages.” Isis 39: 29-44.
Hempel, Carl G. 1965. "Aspects of scientific explanation." Pages 331–496 in Aspects of Scientific Explanation and Other Essays in the Philosophy of Science. New York: Free Press.
Homans, George. 1967. The Nature of Social Science. New York: Harcourt, Brace and World.
Hume, David. [1748] 1955. An Inquiry Concerning Human Understanding, edited by Charles W. Hendel. Indianapolis: Bobbs-Merril Company, Inc.
Kaplan, Abraham. 1964. The Conduct of Inquiry: Methodology for Behavioral Sciences. San Francisco: Chandler Publishing Company.
Kelemen, Deborah. 1999. “Why are rocks pointy? Children’s preference for teleological explanations of the natural world.” Developmental Psychology 35: 1440-1453.
Kelemen, Deborah; Callanan, Maureen; Casler, Krista and Deanne R. Perez-Granados. 2005. “Why things happen: teleological explanation in parent-child conversations.” Developmental Psychology 41: 251-264.
Koyre, Alexandre. 1943. “Galileo and the scientific revolution of the seventeenth century.” The Philosophical Review 52: 333-348.
Merton, Robert K. 1938. "Social Structure and Anomie." American Sociological Review 3: 672–682.
Miller, Richard. 1987. Fact and Method: Expanation, Confirmation, and Realityin the Natural and Social Sciences. Princeton: Princeton University Press.
Popper, Karl R. [1934] 2002. The Logic of Scientific Discovery. New York: Harper and Row.
Shweder, Richard A. and Donald W. Fiske, editors. 1986. Metatheory in Social Science: Pluralisms and Subjectivities. Chicago: University of Chicago Press.
Taylor, Charles. 1964. The Explanation of Behaviour. London: Routledge and Kegan Paul.
Turner, Jonathan H. 1979/80. “Toward a social physics: reducing sociology’s theoretical inhibitions.” Humboldt Journal of Social Relations 7(1):140-155.
Turner, Jonathan H. 1992. Classical Sociological Theory: A Positivist Perspective. Chicago: Nelson-Hall.
Weinberg, Steven. 2001. "Can science explain everything? Anything?" New York Review of Books 48 (Number 9, May 31): 47-50.
Wilson, Fred. 1969. “Explanation in Aristotle, Newton, and Toulmin: Part I.” Philosophy of Science 36: 291-310.