Genes, Culture, and Chaos:
the Nature of Free Will
Free will. Everyone knows what it means, but no one can explain what it is. Scientists are as puzzled by it as anyone - perhaps more so, because they want to know not only what it is, but where it comes from, and how it fits in with what we know of human biology generally. Free will implies freedom from any sort of determinism, whether genetic, cultural, or divine, in exercising choice. But does such freedom really exist? Almost every school of human philosophy, including every major religion, has struggled with this question. Particularly in societies that emphasize the individual over the collective, the freedom of individuals to make choices, and the accompanying personal responsibility for the choices they make, is one of our most cherished concepts. Every legal system assumes that individuals are free to choose how they act in a given situation; individual responsibility has no meaning in the absence of unimpeded choice.
Free will is a particular problem for those studying human behavior. Behaviorists want to know why humans behave so differently in the same situation. Variability is certainly common in humans. We are different physically; tall or short; thin or fat; light or dark. Everyone agrees that those characteristics are affected by genes. We know from classical genetics and molecular biology that individual genes themselves come in different forms - called alleles - and that depending on the particular combinations of alleles an individual possesses - one's genotype - he or she may display any combination of the standard range of human physical attributes. Of course, even physical characteristics such as size and shape can also be influenced by the environment - diet and disease, for example.
But what about the tremendous variability in the way we behave? There is as much variation in the way humans react to their surroundings and to each other as there is in the way they look. Psychologists often assort human personalities along five major behavioral axes (see Figure). Most people can be described reasonably well by deciding where they lie on each axis. Humans also show considerable variability in traits such as mental ability, sexual preference, or religious feeling. How do we account for differences of this sort?
Explanations of the variability in human behavior tend to be defined by one of two major viewpoints. Biological determinists believe that our every reaction to what happens around us can be predicted from what is written into our genomes at the moment of our conception. Biological determinism posits that once we have identified all of the genes affecting a given behavior, and understand how all of the various alleles of those genes distributed among human beings affect that behavior, then we will be able to predict how individuals with any particular combination of those alleles - any particular genotype - will behave. Hardcore biological determinism provided the seeds for sowing eugenics earlier in this century.
"Cultural" or "environmental" determinists take a different approach. They regard a newborn human as essentially an empty slate. What an individual becomes as an adult is thought to be entirely predictable in terms of the cumulative environmental/cultural experiences to which he or she has been exposed. The more ardent environmental determinists believe that, given an infant at birth, they could - at least in theory - shape it into anything at all, simply by controlling the content and sequence of its environmental experiences, without reference to its genotype. It is from this school of thought, more than any other, that the notion of the malleability of individual human beings through social interventions arises. It's potential consequences have been explored in works such as "A Clockwork Orange" and "Brave new World."
Neither of these position holds out much hope for free will. In fact, they argue strongly against its existence. If our every behavior can be predicted from what is written into our genomes before we are born - what does that say about our freedom to choose, or about individual responsibility for the choices we make? On the other hand, the view that we come into the world as some sort of blank template, and that what we become, as cognizant adults, is simply the totality of our previous experiences, renders our behavior no less predictable - and our freedom to choose and act no less constrained - than if we are simply the sum of our genes.
Most scientists actually believe that human behavior is best explained by a combination of biological and environmental determinism, with genes and previous experience contributing roughly equally to the variability we observe in the way humans behave. Genes affect behavior because different combinations of heritable gene variants, encoding nervous system components, function differently in different individuals. The environment affects behavior because, as we can readily demonstrate, experience can have longterm effects on the way nerve cells interact with each other. But combining two different forms of determinism as an explanation of behavior does not help with the question of free will.
Free will is ultimately an expression, or at least a component, of human behavior, and behaviorists agree that behavior is rooted in the nervous system. Where in this dense complex of interacting nerve cells is free will to be found? And why is free will thought to be a uniquely human attribute? Animals make choices all the time. Every time there is more than one possible way to react, an animal is forced to make a choice. Do such choices represent free will? If not, why not? What is it we believe is unique about the human decision making process?
Human beings are clearly different from animals in that humans have developed something called mind, and from mind has come culture. A great deal of what we do - a great deal of our behavior - originates in the mind, with relatively little immediate input from the physical environment. The environment, when we respond to it, may itself be a product of mind - in many cases our behavioral responses are directed to the intellectual and cultural constructs with which we surround ourselves. Some have argued that culture now has a life of its own, even that it evolves independently of our genes, and controls human behavior in ways that simply do not apply to animals.
That may well be true, but as far as we have been able to tell, intellectual and cultural information from our environment is absorbed and processed by individuals through exactly the same neurological pathways as information about the physical environment. Exactly the same neural pathways that process the meaning of songs and novels, political arguments and religious beliefs, also process our responses to predators, food, the weather and potential mates. It thus seems highly likely that variability in neurological pathways will affect our response to cultural information in the same way that it would affect any other incoming information.
So let us assume, along with the majority of behavioral scientists, that human behavior, whether in response to our physical environment or our cultural/intellectual environment, is based in our nervous systems, and is controlled by some combination of our individual genotypes and our previous experiences within those environments. Let us further accept that the neurological mechanisms involved in making behavioral decisions in the context of either environment are the same. What then is the neurological basis of free will? Insight into this question may well come from a seemingly unlikely source: mathematics, and the science of prediction.
Predictability is the essence of science. A phenomenon is observed; various hypotheses are put forward to explain it; those hypotheses are then tested by their ability to predict what will happen under a given set of circumstances. Competing hypotheses are gradually winnowed down until a unifying theory, with maximal predictive power, emerges. But always it is the ability to predict that determines how long a theory lasts; when a theory is unable to predict an outcome, the theory is modified or, in some cases, abandoned altogether.
The physical laws derived to explain the universe have tremendous predictive powers. They allow us to predict tides and eclipses, the motions of planets and galaxies, or the rate of decay of a radioactive substance. But in a surprisingly large number of cases, even when we understand in great detail all of the physical forces acting in a given situation, we are unable to predict what will happen: the flight path of a balloon propelled by escaping air, for example, or tomorrow's weather. One characteristic shared by most of these seemingly unpredictable systems is that they can be greatly perturbed by slight changes in initial conditions. For example, imagine a small boulder balanced at the crest of a steep hill. We give it a nudge, and it begins tumbling down the hillside. Try to predict its pathway through its descent. Every twig it encounters on the way down, every collision with a bush or rock, sets up a new sequence of future interactions as the boulder descends. Whether, and at what angle, our boulder rolls over a small pebble in the first inch of its pathway can have highly unpredictable consequences for the remainder of its journey.
For a long time scientists assumed that the inability to predict behavior in such situations was just a computing problem: given enough detailed information about all of the conditions on the hill, and about the conditions of the initial push that started the boulder on its course, and a big enough computer, they could predict where the boulder would end up. They couldn't. Aside from the fact that the number of possible combinations of conditions in even the first meter of drop on a hillside approaches the infinite, we know from quantum mechanics that the initial conditions of any event - like nudging a boulder from its resting place - cannot itself be known with absolute precision; measuring position or momentum at a fine enough level alters either position or momentum.
Unpredictable behavior also shows up in simpler situations. Imagine a ball that is caused to bounce up and down between two metal plates, with the top plate imparting an occasional slap to the ball to keep it bouncing. Under some conditions of ball size, distance between the plates, and energy input, the ball bounces in a manner that is entirely predictable from all of the known parameters of the system. But under other conditions, often only slightly different, but equally well-described, the bouncing pattern of the ball becomes completely unpredictable. The same thing happens with a dripping faucet. Under some conditions of faucet diameter and water pressure, the rate of drop formation and fall is predictable from the properties of the system. Under slightly different conditions, the drop rate simply cannot be predicted. This inability to predict behavior in physically well-described situations gave rise to an entirely new way of thinking about physical phenomena called chaos theory.
One of the major insights of chaos theory is that oftetimes, out of apparently impenetrable randomness, a new kind of order can emerge. If the apparently random behaviors of the bouncing balls and drip patterns are followed long enough and plotted mathematically, they result in what are known as fractal patterns (Figure). This unexpected finding of coherent patterns in apparently random systems events has allowed scientists and engineers to harness chaotic behavior in highly productive ways, such as compressing images for computers and stabilizing lasers. Equally importantly, it has also made it possible to detect the presence of true chaos as an internal generator of randomness, as opposed to randomness resulting from unpredicted outside interference; the latter will never generate the beautiful internal fractal patterns of true chaos.
The brain certainly offers plenty of opportunities for the generation of nonlinear, chaotic behavior. Individual nerve cells (neurons) are themselves incredibly complex (Figure). Information is brought to each cell through numerous dendrites, each of which is extensively branched. Each branch is in turn connected to a different nerve cell; a single nerve cell may be able to receive information from a thousand or more other neurons. The electrical potential used to forward information through the cell's single (but again extensively branched) axon is controlled by hundreds of ion channels, which regulate the flow of sodium, potasium and calcium ions in and out of the cell. Cells within a given pathway or tract may interact with each other in feedback loops, as well as with cells in targeted regions of the brain. Nerve cell tracts thus provide a good breeding ground for the kind of nonlinear behavior associated with chaos. The slightest variation of initial conditions in one neuron - the particular selection of dendrites activated at any given moment, the number and state of active ion channels - could produce alterations that rapidly evolve into effects on subsequent signal distribution that are as unpredictable as the pathway of our hypothetical boulder down a mountainside.
And in fact there is increasing evidence that chaos may be a common property of the human nervous system. "Brain waves," the electrical properties of nerve cells measured in an electroencephalogram (EEG), are accurate indicators of the state of arousal of the brain, but their precise relation to underlying nerve cell activity is not entirely clear. One problem is that EEG recordings are made over large areas of the brain involving many millions of cells. These cells are often grouped into coherent nerve tracts that operate in parallel, and to a large extent independently of one another, so correlation of the resulting highly complex patterns with specific brain activities is extraordinarily difficult. It has been difficult if not impossible to use EEG patterns in any predictive sense with respect to nerve cell function. Nevertheless, when EEG traces are analyzed using software that detects fractal-generating chaotic behavior, evidence for chaotic activity abounds at many different levels of brain activity (Figure).
There is another level at which chaos is manifest in the brain. The branching patterns of axons and dendrites share a feature common to all fractal patterns: self similarity. At low magnifictions, one sees a particular branching pattern in the main trunks of dendrites, for example. At a slightly higher magnification, it can be seen that the spacing and pattern of smaller branches mimics the branching of the trunks. At higher magnification still, the branches themselves are seen to have branches, with again the same pattern. This same process can also be seen in the branchings of blood vessels, of kidney tubules, and even in the complex foldings in the lining of the intestines. It appears that a common fractal-generating algorithm is at work during the creation of many of the body's branched structures.
Why would we want to have programs embedded in crucial organs such as the brain that generate unpredictability? One advantage is diversity in the ability to respond to the environment. Diversity of response is present in our species as a whole by the presence of multiple gene forms (alleles) within the species. Some alleles of a given gene will work best under one set of conditions; other conditions may favor a different allele. Human beings on the whole are thus equipped to face a wide range of conditions. Within an individual, systems such as the immune system can even scramble certain genes internally to create the incredible diversity needed to deal with the wide range of pathogens that invade our bodies and cause disease. But that is a rare exception; inherited genes are in general not altered during an individual's lifetime.
Chaos does not alter genes, but it creates new possibilities in the pathways specified by genes. Nonlinear amplification of small fluctuations in initial conditions can result in the creation of random and novel structures, and even behaviors, that are not embedded in DNA per se, and are not related to previous experience. These new features will be subject to selection in the same way as gene-encoded properties, because they affect the ability of the organism to survive and reproduce. Unlike gene-encoded behavior, however, these characteristics are not automatically passed from one generation to the next; like changes imposed by the environment, they are limited to the lifetime of the individual in which they arose. But the chaos algorithm - the ability to generate diversity through chaotic behavior - is passed along; it is inherent in the very way many biological pathways are organized in the first place. The importance of this internal chaotic behavior is suggested by the fact that chaotic patterns decrease, rather than increase, during sickness and with increasing age in human beings.
We are just at the beginning of our attempts to understand the role of chaotic processes in human behavior, but already a number of possibilities suggest themselves. Creativity has always puzzled us. We sometimes make strange mental leaps, arriving at conclusions for which nothing in our previous experience seemingly could have prepared us, but which allow us to see things in an entirely new way. And what about decision making? How many times have we thought back on a particular situation and wondered why we acted as we did? Chaos provides us with a rational basis for expecting that behavior may be too complex to be predicted, that behaviors can emerge that seemingly have no rational basis, either in genotype or in terms of previous experience. The nervous system is at one level deterministic, but we cannot always predict what it will do.
And that brings us back to free will. We speak of free will as though it is something we control. But is that necessarily true? Our decisions to act in a certain way in a given situation arise entirely within us. But the very definition of chaos suggests that it would operate outside of human consciousness and memory. So if chaos is a factor in generating human behavior, then it may be that what we are calling free will is simply a way of accounting for a certain level of indeterminancy in our behavior, of trying to fit it into a pattern that we can understand - and think we can control.
What does this view of free will say about individual responsibility? Our genes and our past experiences, as best we understand their role in human behavior, are highly deterministic. If that's all there were, there would be little or no free will; we could plead for our every action that we are the helpless victim of one or the other form of determinancy. But the indeterminancy of chaos that frees us both from our genes and our past history, also forces us to accept responsibility for how we act. Chaos allows us to experience things not scripted in either genes or experience, but we have the power to learn. We can see and understand fully how our behavior affects our own lives, and the lives of those around us. That is the nature of free will; the ability to choose among possibilities dictated by neither genes nor experience. And in this definition of our humanity, chaos may be the most important factor of all.