Posts

Coming To Grips With Free Will

/38 Comments/in /by

 

Index to posts in this series

Agency

Michael Tomasello didn’t write about the evolution of free will. His book is called The Evolution Of Agency. Even so, I think we should understand Tomasello’s model as a partial defense of free will.

The idea behind the book is that the psychological processes that characterize our species are the result of evolution and evolutionary pressures. That includes agency. Recall from this post that Tomasello gives this description of agency:

…[I]n the current case, we may say that agentive beings are distinguished from non-agentive beings … by a special type of behavioral organization. That behavioral organization is feedback control organization in which the individual directs its behavior toward goals—many or most of which are biologically evolved—controlling or even self-regulating the process through informed decision-making and behavioral self-monitoring. Species biology is supplemented by individual psychology. P. 2,

This description of what we mean by agency doesn’t explain how we set goals. But I think as a first approximation that we set goals “through informed decision-making and behavioral self-monitoring”, heavily influenced by our families and communities through what Tomasello calls socially normative agency. We examine as many aspects of our situation as we can think of and handle, we apply our decision-making tools, we decide. Among the constraints for decision-making we consider the incentives and constraints of our society.

Once our goals are set, we consider the ways we might reach them, and choose the one that seems most likely to enable us to reach the goal. We monitor our results, and make adjustment as we go along, including changing the method of reaching the goal, or the goal itself, if that seems better to us.

Determinists

Some scientists deny the existence of free will, including  Robert Sapolsky, a Stanford neurobiologist.  He explains why he thinks we have no free will in this LA Times interview about his recent book Determined.

Here’s an essay in The New Yorker by Nikhil Krishnan, a philosopher at Cambridge, discussing the book in the context of philosophy.  This article says that Sapolsky doesn’t define the term free will, but offers

a challenge. A man, Sapolsky invites us to imagine, “pulls the trigger of a gun.” That’s one description. Another is that “the muscles in his index finger contracted.” Why? “Because they were stimulated by a neuron,” which was in turn “stimulated by the neuron just upstream. . . . And so on.” Then he throws down the gauntlet: “Show me a neuron (or brain) whose generation of a behavior is independent of the sum of its biological past, and for the purposes of this book, you’ve demonstrated free will.”

First, how exactly would that kind of free will have evolved?

Second, that’s not how people think of free will. In normal usage free will is about the ability of the individual to make choices among alternatives, a view central to Tomasello’s model. I could shoot my gun, or I could not shoot my gun. Both are within the range of possible actions, and I can choose between them. Sapolsky thinks the fact that I don’t shoot is the result of every bit of experience in my past, and that I had no real choice. Someone else with a different past might not have any choice but to shoot. Tomasello, I think, would say that I can think rationally about whether or not to shoot the gun, examine the possible consequences, determine which action accomplishes my goals, and act on that reasoning

Examples

1. If someone had asked me 30 years ago what my favorite color is, I would not have had much of an answer. I might have said I don’t have one, or I might have said British Racing Green; or maybe blue, which is close to a non-answer. Today I would say jewel tones: ruby red, dark blue sapphire, intense emerald green. I can point to several reasons for this change. One is seeing the lapis lazuli blues of early Renaissance Sienese paintings of the Virgin Mary, and a ring we bought, gold with tiny sapphires.

2. Lake, a deeply conservative Trumpish Republican, attends a work-related dinner with their partner. Lake doesn’t know anything about the people at the table. The conversation turns to politics. Lake doesn’t want to impede their partner’s career, and keeps quiet.

3. Albert Einstein at the age of 16 imagines what he would see if he were riding side by side with a beam of light. A few years later he suddenly realizes the implications of the answer.

Analysis of examples

1. Favorite color doesn’t implicate goals. It seems to be about recognizing a thing that gives us pleasant feelings. The example asks if we can know whether a thing gives us pleasure, not whether we can choose what gives us pleasure.

It seems likely that we can train ourselves to take pleasure in things. I like opera, but that wasn’t always so. I learned to like opera by attending operas, listening to opera singers, and eventually singing opera chorus. How exactly does that relate to free will? Would Sapolsky say I had no choice in the matter?

2. This example seems fairly close to the foraging examples used by Tomasello, including the ones about our early modern human ancestors. Each person in the group has to play a role. Lake’s role is not to irritate the other people at the table and hurt Lake’s partner’s ability to bring home the bacon. Was that an exercise of free will by Lake?

3. I chose the Einstein example because I’ve always thought it was a singular insight into an otherwise intractable problem. The greatest works of art, music, literature and inquiry also show us a singular insight into our world, other people, and ourselves.

This example seems to combine elements of the first two. Why was Einstein thinking about this bizarre hypothetical at age 16? How much of the solution he eventually reached depended on the fact that other people were thinking about and working on that problem? Would Sapolsky agree that this is so far outside normal human behavior that it qualifies as free will? Is the concept of free will relevant to this example?

Conclusion

Of course, there isn’t an answer to this disagreement, so here’s what I think. Our bodies, including our cognitive processes and our psychological processes, co-evolved in a way that encouraged collaboration as a survival tactic. We learned to cooperate in gathering food, making simple tools and clothing, and protecting the group. It turns out that the cognitive and psychological processes we evolved are useful for other things, like making music, decorating plates and bowls, and inventing airplanes. They can be used for darker purposes. They can be used for highly abstract purposes, like set theory and surreal poetry.

We can also act rationally, just like Einstein thinking about the nature of light. We can force ourselves to examine as best we can the likely outcomes of our actions. We can use that skill to decide what we want and how best to get it. We can choose to act on the results of that rational thought or not. That’s enough free will for me.

Great Apes As Rational Agents

/15 Comments/in , /by

Index to posts in this series

In Chapter 5 of The Evolution Of Agency, Michael Tomasello discusses the nature of the agency displayed by the great apes. This group consists of five species, chimpanzees, bonoboes, orangutans, gorillas and humans. The first four of these are the subject of this chapter. African great apes seem to have emerged about 14 million years ago following millions of years of evolution of mammals. The changes were far-reaching.

There are three relevant threads in this chapter:

  1. evidence of the rationality of great apes
  2. evidence that they recognize that others of their species act intentionally, possibly including humans
  3. the evolutionary pressures that might have contributed to the development of these two mental capacities

Tomasello offers an explanation of this rationality as dependent on a second tier of executive control.

Rationality of Great Apes

Tomasello gives a number of examples of evidence from field observations and experiments that shows the great apes are capable of observing their environment and acting on it it rational ways. One is their understanding of tools. For example, they use sticks to fish for ants and termites to eat. If there is no stick nearby they will tear off a twig from a tree, and strip the leaves if there are too many. They will drop stones on above-ground termite nests to flush out the bugs. Here’s how Tomasello ascribes rationality to this practice:

… [W]hen faced with a novel physical problem, great apes can also take control of the causal process and make new tools that will work in the new context. In this case, they are first imagining an effect that is needed to solve the problem, and then going back to create a cause. For example, in the wild, chimpanzees routinely modify too-leafy branches by stripping leaves from them so that they will fit into termite holes. P. 72, cite omitted.

Here’s another example. The experimenter shows a piece of food to a chimpanzee and then puts it into one of two cups. The experimenter shakes the empty cup. The chimpanzee is allowed to pick a cup, and chooses the unshaken cup. This shows a reasoning chain: no noise means no food; therefore the food is in the other cup. Tomasello says that these are forms of logical organization that we should consider as rational.

Great apes understand some cause and effects created by others

Great apes understand cause and effects created by their own actions, as do other mammals. Unlike other mammals, they also understand indirect causes of results, as with the use of tools. They also recognize that other creatures can themselves cause effects through their actions. Tomasello cites a paper reporting

… that three human-raised chimpanzees selectively reproduced actions that a human demonstrator intended to perform over actions she performed only accidentally; the chimpanzees also performed actions that a human intended to perform but did not actually succeed in performing. P. 75.

A two-part experiment tests whether chimpanzees can “use self-experience to infer what another sees”. (Abstract here). Great apes will took at what another is looking at, which is referred to as gaze-following. The subjects are taught the visual properties of two screens, one opaque, one see-through. The first experiment tests gaze-following when the experimenter is using each screen. The subjects don’t seem to distinguish between the two types of mask in the gaze-following experiment.

The second experiment uses a competition model, where the chimpanzee and the experimenter are dealing with food in two boxes. One has an opaque lid and the other a screen or a transparent lid. The subject is taught the effects of the three lids. Then food is placed in both boxes. If the chimpanzee tries to get food when the experimenter can see it (transparent or screen lid), the experimenter takes it away. To get the food, the chimpanzee must know from its own experience whether the experimenter can see the food. In this setting the chimpanzees get the food significantly more often, leading the experimenters to “conclude that chimpanzees successfully used their self-experience to infer what the competitor sees.”

Tomasello also cites some evidence of social learning in great apes. He says they can learn by noticing the results of the actions of other great apes, and then doing the same thing or something similar.

Evolutionary pressures

Tomasello suggests two types of environmental pressures that might have led to the evolution of these skills. First, fruit is an important part of the diet of chimpanzeees. Fruit trees grow in small clumps, and don’t put out fruit at the same time. Chimpanzees tend to sleep in large groups at night, and split into small groups for foraging. The smaller groups somewhat reduce the competition for food.

Great apes do not usually share these finds. The dominant member of the small group takes all it wants. Even so, predicting the behavior of competitors makes it more likely that subordinate individuals will obtain sufficient food.

Another factor might be that the great apes depend on social learning to maintain their groups, and to understand whether a specific behavior is or is not tolerated. Great apes have longer juvenile periods than other mammals, and much of their time is spent in groups where they learn to align their behavior with that of others. This requires them to be able to attribute their own experience to others of their groups.

This attribution seems to extend to their own mental states. As an example, juvenile chimpanzees use a specific arm gesture to indicate a desire to play with another. The juveniles know that the other must be looking at them in order to see the signal just as they do, and they wait until the other is looking at them to make it.

The psychological processes of great apes

Recall that Tomasello proposes a modes in which small mammals have an executive tier that supervises and controls the operational tier of mental processes. He suggests that great apes have a second level of control which he calls the reflective tier. Its function is to “… to monitor, troubleshoot, and intervene in processes of executive decision-making and cognitive control….” P. 82. In effect, Tomasello says it gives the great apes access to their own mental processes. This fits with the evidence he cites. For example, it explains how great apes can attribute their mental processes to other.

It also explains the results of this experiment cited by Tomasello. The subjects were rewarded for setting a group of blocks on end. Then a block was added that wouldn’t stand on end because of an internal weight. The subjects frequently inspected that block carefully trying to figure out why it wouldn’t stand up. The subjects are trying to reach a goal but failing. Tomasello says in this case the reflective tier in intervening in the intentional action to try to figure out why what works for most blocks doesn’t work for this specific block.

Discussion

1. Of course great apes can’t explain why they make these choices, so perhaps we humans don’t immediately think of them as rational. But think of the number of decisions we make without using strict logical constructions. In many of these cases rationality is buried so deep in our brains that we don’t really need to use language to work out the solution. This is similar to what Kahneman describes in Thinking Fast And Slow, which Tomasello cites in an earlier chapter for a similar proposition.

2. So far we’ve looked at three categories of agency, goal-directed agency as in lizards, intentional agency as in squirrels and cats, and rational agency as in the great apes. Tomasello’s thesis is that the psychological processes of human agency evolved through these groups.

Next he takes up humans. And so will I.