The Game of Life: Discussing Determinism in Behavioral Science

Behavioral science, put simply, is the study of human behavior. The discipline scrutinises human movement scientifically, as a means of approximating the stimuli that determine patterns of individual and group activity (Banerjee, 1995). A core tenant of its scientific rigor therefore is its focus on causality, in that different behaviors are responses to an identifiable stimulus, and occur in a systematic fashion that lends itself to predictive analysis (Simkins, 1969).

This belief that human behavior can be explained in terms of cause and effect — like the atomic space in which it occurs — can be traced as far back as Ancient Greece, with philosophers such as Democritus and Plato (Osler, 2003). Many disciplines and paradigms have similarly emerged with the understanding that human behavior, with all its baffling complexities, can be explained in terms of deterministic associations and laws, the knowledge of which can improve public policy, personnel wellbeing, and corporate efficiency.

Determinism in behavioral psychology, for example, can take the form of environmental determinism, where the source influencing someone’s behavior is external to them — like when Bandura (1961), suggested that violent parents produce violent children (McLeod, 2013). Additionally, deterministic factors can be internal to the person, resulting from unconscious desires and motivations in the subconscious mind (as argued by Freud), or genetic and biological predispositions, such as specific genes that lead to high IQ levels (Chorney et al., 1998, cited in McLeod, 2013), or different personality types that foster specific behaviors (Alarcón, Foulks and Vakkur, 1998).

In addition, the notion of situational determinism in economics claims that the behavior of actors is influenced by the “logic of the situation” (Gustafsson, Knudsen & Mäki, 2003: p17), or that different internal preferences at different times lead actors to pursue a set course of action (Mäki, 2003). By observing these patterns, researchers can predict when people are most likely to behave in certain ways, and what changes in an environment can promote the actions most desired by interested groups.

Overall, whilst these disciplines are in no way uniform in their approach or unanimous in their findings, the general idea that stands out to people who may not know much about behavioral science is that our conscious actions result from environmental and internal stimuli — and are, as such, open to prediction. But what does this mean? Are these causalities set to play out regardless of our desire to act freely? Are our futures already planned? Behavioral science may not always be concerned with these questions, but they undoubtedly play on the minds of those not fully acquainted with the field.

Few would fret over the causal inferences of the natural sciences, such as the Laws of Thermodynamics, genetic predispositions to medical disorders, or the prediction of tomorrow’s weather. However, through implying that behavior is the determined outcome of a multitude of internal and external factors, audiences may interpret that they are not in control of their own decision making. Referring to the anonymous limerick, we may prefer the idea that we are free-willing autonomous agents with the ability to act as we wish, with the freedom that a bus (technically) has over its own movement, and may be hostile to the idea that we are instead destined to act in set ways because of our personal environments and internal cognition, like the constrained movement of a railed tram.

To address this dread, I invite readers to, in the same way I was as an undergraduate, become acquainted with “The Game of Life” — a mathematical automation game introduced by Cambridge mathematician John Conway.

The game of life consists of a two-dimensional rectangular grid of usually white cells, or pixels, each with the ability to turn on (turn black) or off (stay white). Picture it like a computer screen.

Fig 1. https://www.math.cornell.edu/~lipa/mec/lesson6.html

Off pixels will turn on if they are bordered by exactly three live pixels. Pixels will remain in their starting state if they have exactly two live neighbours, and live pixels will turn off if they either have fewer than two live neighbours and more than four in total (for more see Conover, 2009). The exact mathematical principles of the Game of Life are quite complicated, and may seem irrelevant to behavioral science. However, it is not the maths at play that is important, but the inferences we can gain about the relationship between pre-set laws and the kinds of patterns than can be observed from their enactment.

There is nothing else dictating the activity of the pixels on this grid. No pre-set pattern or graphics program is put in place. Only a small shape or a few pixels are turned on in a small location on the grid before the simulation is run. What follows is sensational. Whilst only adhering to a couple of very basic deterministic laws, the grid lights up in what looks like early 8-bit computer graphics, displaying detailed and explosive patterns that make you speculate that the complexity of the graphical action must all be pre-planned. But, as previously stated, none of the patterns are anticipated or deliberate. All that was pre-set were a few very basic principles determining under what conditions a pixel is to turn on or off. The collection of shapes, formation of entities, and the movement of what looks like the arrangement of cells in the body or avatars on a computer game all have their routes in very simple precepts.

What Conway’s Game of Life shows is that, whilst something can be purely deterministic, in that it can only act in a certain prescribed way, its behavior can still be unpredictable and unimaginably complex. The deterministic nature of the simulation is coupled with a huge sensitivity to a group of pixel’s initial condition, and the slightest difference in the starting state (where the pixels are turned on prior to the simulation, and what shape they form), radically effects subsequent states and patterns once the simulation is run (Ibid). A square shaped collection of live pixels, will move in an immeasurably different manner to a structure of a different shape in a different position on the grid. The kinds of patterns that are shown have a seemingly infinite number of possibilities in their movement, and how they interact with other strands of patterns.

Now, when applying this to behavioral science, this is not to say that human behavior is reducible to a handful of consistent laws and directives, like the pixels in The Game of Life. As stated earlier, Conway’s simulation provides an accessible heuristic tool to demonstrate that causality is more complicated than people may think. Though, in The Game of Life, we can say with certainty that any square on the grid will turn on under conditions A, stay as they were in condition B, and turn off in condition C, we cannot then predict with similar certainty what shapes and patterns we can expect to see from a specific shape occupying a certain starting point.

Similarly, just because behavioral science teaches us that we can expect individuals to behave in a certain way given a certain environment, it is not then to say that we can anticipate the extent of the ensuing behaviors and how they will change and adapt when met with others adhering to a similar deterministic logic.

For example, from approaching the criminal justice system with research about how changing criminal law does not deter individual criminal conduct (Robinson, 2004), or proposing to a business that a customer’s religious commitment determines the importance they place on personnel friendliness/assistance when evaluating their services (McDaniel and Burnett, 1990), the audience may infer that these associations are certain, whilst the sample demographic may resent being portrayed as so predictable. Behavioral outcomes that end up deviating from expectations may fuel resentment toward the usefulness of the science, and individuals may express discontent as being treated as mere formulaic axioms.

Using the heuristics of simulations like The Game of Life, can convey that attributing behaviors of interest to fundamental deterministic laws does not mean that we can fully account for or predict how those laws will play out in a wider context. Essentially, The Game of Life can help to teach the importance of vigilance and open mindedness about the predictability of asserted causalities. This way, audiences adapting to new insights from behavioral science will not be so surprised or feel “cheated” if the outcomes are different to what may have been approximated.

This is the essential take-away from Conway’s Game of Life for behavioral science audiences. Just because human behavior may have a quite simple deterministic origin, it is not to say that resulting behavior will be uniform, stable, and always predictable. The future therefore, is still largely unknown. We may provide hypothetical conjectures about the basic laws of human interaction, but there are still many such laws that we do not know. Conversely, we may not know the full extent of outcomes relating to those laws that we do know. This isn’t a critique of behavioral science research itself, as such limitations of the causal and predictive claims of any piece of research are always at the forefront of scientific assessment and enquiry. Instead, this heuristic device applies to the perceptions of the audiences who could benefit from a better understanding of the field.

In the Game of Life, as in life itself, we may formulate certain axioms about how things behave in a given environment, and from these axioms infer certain isolated outcomes (e.g., Condition A will make pixel X turn on). But, on the whole, the product still involves a high degree of context-specific randomness (e.g., from where did the game begin?). The dual challenge facing behavioral science is to formulate those axioms such that they reflect the true sequence of human action in a given environment, but also to acknowledge the limitation of those insights when extrapolating to other contexts. Ultimately, as a tool for formulating policy, behavioral science may be less concerned with whether human beings are more like buses or trams, and more concerned with making sure we get where we have set out to go.