Intelligence and the Neurosciences: Second Installment
In the first installment of these two blogs on Intelligence and the Neurosciences I recounted some of the important progress that has been made establishing the genetic, neurophysiological, and neuroanatomical underpinning of individual differences in cognition, i.e. of intelligence. The findings are steps toward establishing a theory of the biological basis of intelligence. In this blog I am going to turn the argument around, and argue that for many aspects of psychology a purely biological approach is not all that useful.
In the end all of psychology reduces to biology. (The only alternative explanation would be the influence of good and bad angels, and that is a topic quite outside of science!). However there are situations where it does not pay to think about behavior in biological terms. This argument stems from two things; a view of theory that may owe more to engineering than psychology, and a specialization of the concept of Brunswikian symmetry, about which a bit more later.
First an anecdote. A number of years (quite a number of years) ago I was in a meeting of department chairs, where our new Executive Vice President was extolling the virtues of goal-based management of academic departments. After telling us about the wonders of having clearly stated, objectively defined goals for research he turned to the Chair of the Physics Department, and said
“Ron, what are the goals of Physics?”
The physicist answered
“To understand the nature of the universe.”
The VP asked the Chair if he wanted to be more specific. The Chair said “No.”
The Chair was taking the classic reductionist view of science. Physics deals with the fundamental nature of physical forces and objects. Chemistry deals with the ways in which the basic elements uncovered by physicists can bond to create the materials we deal with. Biology then deals with how these materials form living things, geology deals with non-living things. Psychology sits on top of biology (with special concern for a particular species; us). Continuing on the reductionist path, Sociology and Economicsshould be derived from Psychology. And on we go. According to this view the purpose of a theory in Psychology is to connect behavior to its biological roots. Viewed this way, the modern emphasis on neuroscience is both understandable and highly desirable.
The reductionist approach assumes that “the truth” lays somewhere, and that science progresses by closer and closer approximations to that truth. This view is rather like the 19th century explorer’s quest for the source of the Nile. The Nile had to arise somewhere. Starting in Roman times, successive explorations developed new approximations, until finally it was realized that the Nile was produced by the confluence of two rivers, one arising from a huge spring and swamp in Southern Sudan and the other arising from Lake Victoria. Similarly, we know that the brain controls behavior. Sufficient research should reveal the mechanisms of its action.
While reductionism is one of the goals of theory building, there is another. Theories provide models to guide action. When a theory is used this way it has to be stated at the same level of generality as the actions it dictates. When engineers and highway architects lay out plans for a road they think of vehicles as solid objects, capable of certain degrees of acceleration, deceleration and turning…all concepts rooted in Newtonian physics. The civil engineer and the architect do not want, and could not use, a quantum-level description of a car. The variables they control; lane widths, highway curvature and so forth, are defined at the ordinary level of everyday perception. Their model of the object they want to control, the vehicle, has to be at that level. Newtonian physics works just fine.
Egon Brunswik, a Hungarian psychologist who emigrated to the University of Califonia, Berkeley, in the 1930s, provided us with an argument for thinking about theory this way in Psychology. Brunswik argued that because animals in general (and people in particular) are embedded in an environment their behavior has to be understood in terms that connect with the environment. Consider learning, i.e. changes in behavior over time as a result of exposure to an environment. From a reductionist viewpoint the challenge for the neurosciences is to explain how the behavioral changes are related to changes in the state of the brain. Brunswik would (probably) have argued that the challenge for psychology is to model how regularities in the environment map onto regular changes in behavior. More generally, in order to construct an environment for controlling behavior…be it a school or a prison or an advertising campaign…models of human thought should be stated in terms of psychological constructs, such as working memory capacity, speed of decision making, and probability of recall of information from long term memory, without regard to how these constructs are realized by brain structures.
Education: Teachers deal with students at the behavioral level, not the neurophysiological level. Teachers can use information about the relative benefits of spaced vs. massed instruction, different types of rehearsal, patterns of language use that indicate dyslexia, behavioral ways to aid students in maintaining attention (which may be as simple as seating the affected student in the front of the room), the cost and benefits of repeated testing, and the identification of gifted students (including but not limited to the use of intelligence tests). It is true that language is substantially a left hemisphere function, that attention is controlled by circuits in the prefrontal cortex and the cingulate gyrus, and that the development of retrievable memories depends upon the functioning of the hippocampus. All these findings are central to understanding the neural basis of behavior, but they are only of passing interest to educators.
Personnel selection: Personnel selection represents an attempt to predict what behaviors a person is likely to emit in a target situation. This can vary from passing university courses (predictive device: The SAT) to showing up on time for a job (test of conscientiousness). Looked at abstractly, selection tests present a person with a problem or situation and ask the examinee to make a voluntary response saying what he or she would do. The predictive power of the test comes from correlations between overt behaviors in the test and target situation, not from correlations between brain states that may underlie behavior in either test or target situation. Brunswikian symmetry applies.
Decision Making: Classical economics assumes that humans will be rational decision makers, who maximize their expected utility. See, for instance, Von Neumann and Morgenstern’s Theory of games and economic behavior, a work that is central to economics. Psychologists doubt that people are so rational, and as a matter of fact Nobel prizes have been awarded to two people (Herbert Simon and Daniel Kahneman) who pointed out that such rationality just is not there. Simon argued out that people have limited information processing capacity, and thus often select a satisfactory but easily identified alternative rather than searching for the best alternative available to them. (Think about buying a new car. At a certain point you just get tired of looking.) Kahneman and his colleagues extended Simon’s observations by showing that we have all sorts of heuristics for decision making that work pretty well, most of the time, but that can lead us astray.
In the last ten years or so research on the neurolscientific underpinnings of decision making has shown that (a) decision making, like most problem solving, depends largely on the working of the prefrontal cortex, in combination with the cingulate cortex but (b) when the decision makng involves emotional content the activities of the prefrontal and (especially) cingulate cortex are also influenced by signals from “emotion centers” of the brain, and in particular a structure called the amygdala.
Now, let us put these things together and, because it is an election year, apply the combination of facts to politics.
One of the strongest “irrational” influences on decision making is framing. For instance, Kahneman and his colleagues showed that people are more likely to take an action that is presented as a way of avoiding a loss than if it is presented as a way of acquiring a profit, even though the anticipated economic value of the action is the same in either case. George Lakoff, a psycholinguist at the University of Berkeley, has pointed out that politicians avidly exploit the framing effect. One of his telling examples is the phrase “tax relief,” which he points out implies that tax is an affliction for which relief is appropriate. An alternative would be to present taxes as dues that we pay for the privilege of living in a prosperous society. Viewed that way, taxes are not quite such an affliction.
Presumably some frames work and others do not. (Intuitively, I think “tax relief” does and “citizenship dues” does not.) Understanding how frames are used to sway opinions in electioneering and advertising is an important topic in social psychology. Studies of the use of frames do not require a neuroscientific analysis of the brain mechanisms involved in decision making. It is sufficient to know that emotional content influences decision making. Knowing where in the brain the emotional content arose, although certainly an important step toward the reductionist goals of science, is not a requirement for the study of propaganda!
Brunswikian symmetry again!