On Human Nature(s) 1: Orientation Toward Behavior
Marc Green
Overview: When an event such as a road accident or police shooting occurs, how should it be explained, i.e., how should causation (and blame) be assigned? The premise here is that determining the cause of any event requires an understanding of the mechanisms involved. When the event has human factors, then the mechanism is human nature. By this I mean the sensory, cognitive, and emotional "kludges" that humans employ to successfully perform tasks that achieve goals and avoid harm despite our many limitations. This five-part series presents a conceptual schema of human nature which enables scienttific analysis of human decision-making and behavior in events with both good and dire outcomes. It is based on a keynote presentation given at the Force Science 2024 Fall Conference. Most concepts discussed here are also covered in more detail in Green (2024).
A police officer receives a radio call that a stolen car was seen driven to a local high school parking lot. He proceeds to the school where he sees the car and its driver. Once there, the officer positions his cruiser directly in front of the car so that it blocks the only driveway exit that would allow escape. He exits his cruiser and takes a position between his vehicle and the front-left of the stolen car. With his weapon in low ready position, he orders the driver to put his hands out through the open car window. The driver guns the engine a few times with his tires turned slight to the left but then complies, so the officer moves closer, finally standing in front and just wide of the driver’s side. Then the driver pulls his hands back in and, with rear tires squealing, starts accelerating toward the officer. The officer raises his weapon and fires twice. The second shot goes through the open driver’s side window fatally wounding him as the car brushes the officer on its way into a snowbank.
Introduction
When bad things such as police shootings or road or other accidents happen, there is often an inquiry. The usual goals are determining why the mishap occurred, attributing the cause, and often (usually) fixing the blame. The inquiry may then go on to make recommendations for steps that will prevent similar mishaps from reoccurring. Of course, the success of any preventative measures depends on the accuracy of the causal attribution.
In this article, I explain why it is first necessary to have good conceptual model of human nature before causal attribution is possible. Although the term “human nature” is plastic and many sources use it in reference to vaguely defined aspects of the human condition, I mean it in a very specific sense: the collection of mechanisms that humans as material beings employ to survive and to successfully perform tasks that achieve goals and avoid harm despite our limitations. Any good conceptual model identifies the system objects and their relationships but also their limitations and constraints. I will argue that keeping human constraints/limitations in mind is critical for understanding and evaluating human decision-making and behavior. The discussion elaborates some themes introduced on other pages such as Thinking Like A Human Factors Expert and Death by Uber 4. In sum, this set of pages can be characterized as an attempt to convey my conceptual model of human nature to the reader.
The examination of human nature is divided into five parts. This first section shows why an understanding of human nature is necessary to determine causation when a mishap occurs. It then contrasts two orientations toward human decision-making and behavior, the prescriptive and the scientific, and explains why the scientific is more useful. The second section explains the natural constraints on human behaviors and the need to employ kludges1 (pronounced "kloodges") that circumvent them. The discussion then turns to dual process theory, the idea that humans have two distinct natures commonly called System 1 and System 2. It provides a brief overview of each, explaining that “buyers are liars” so System 2 is seldom the true driver of decision-making. Instead, the vast majority of human behavior is the result of System 1 processing. Two simple visual search problems demonstrate the differences between the two natures. Next, the discussion of System 1 kludges is broken into two sections. The third section delves into three of its main and best-known predictive kludges, affordances, heuristics and biases, and schemata. The fourth section continues with a discussion less-widely-discussed of System 1 predictive kludges, Kopfkino and predictive coding. The section ends with what might be term System 0, emotion and its role in behavior, especially in defensive reaction. The final section applies some of the discussed kludges in reviewing a real-life event, the police shooting described above, before summing up the discussion of human nature(s).
Human Nature and Causation2
The main reason to study human nature is that it is otherwise impossible to determine causation. Cole (2018) nicely highlighted the close connection between mechanism and causation by saying:
If we are going to learn how to drive a car, we need to know at least some basics about how the engine runs. This is important because it helps us diagnose cause and effect while we drive. Car dragging to the left after a big boom? We probably blew out a tire. We don't just trust our senses (the boom, the dragging). We rely on our knowledge of how the mechanism works.
In sum, simply observing an event is insufficient to draw conclusions about causation. It is necessary to understand how the mechanism operates. If the event involves humans, then the mechanism is human nature. We humans instinctively apply implicit conceptual models of the human mechanism to understand the behavior of others. The only questions are: 1) what is the model's basis and 2) whether it is reasonably accurate. A faulty conceptual model of human nature will likely produce erroneous conclusions. It will further lead to the misattribution of causation and to the proposal of ineffective preventative measures.
Prescriptive (Folk Psychology) vs. Descriptive (Scientific) Views of Behavior
The precursor to developing a good conceptual model of human nature is to adopt a proper view of behavior. It is important because inquiries usually aim, implicitly or explicitly, to determine whether human behavior was normal or aberrant. This leads to the question of how such a determination should be made. For didactic purposes, I'll contrast the two most common, yet conflicting, orientations when thinking about human behavior. The "prescriptive view" of normality is perhaps the more natural. The view expects that humans should conform to some preconceived notion of normal behavior. Their actions are then evaluated by their degree of adherence to the prescribed norm. People whose behavior departs from the norm are at fault, blameworthy, or even irrational. When an actor fails to meet the norm, the mishap will likely be attributed to some internal "Folk Psychology" flaw, an inferred mental state such as inattention, carelessness, etc.—which is little more than circular reasoning. This is the way the legal system and some professionals such as engineers and accountants tend to think. It also passes as commonsense reasoning in the population at large, who are especially prone to the "fundamental attribution error" and counterfactual thinking.
Preconceived norms arise from a variety of sources. Sometimes they are based on codes and regulations, which may or may not have some basis in empirical evidence. Sometimes they are based on mathematical models of optimal behavior that define rational choice. (See section 3.) Often as not, they are based on whatever would be convenient for designers, businesses, and safety authorities to keep costs down and to divert legal responsibility for any mishaps that may occur. Prescriptive norms often magically appear after the fact in hindsight and with counterfactual thinking.
However, the prescriptive view is problematic because it often demands that humans behave in ways that simply aren't realistic since they are contrary to human nature, e.g., pay attention to everything 100% of the time, never become fatigued, follow every rule, etc. Norman (1988) summarized this problem by saying:
[t]echnologies are designed to be used by people, ordinary people, people who grow fatigued, whose attention wanders, whose mind is preoccupied. It does no good to legislate against such properties of human nature. It does no good to complain that if only workers would keep their minds focused on the task, they would not be getting injured. Everyone's mind wanders, everyone daydreams, gets fatigued.
The alternative, "scientific view," is descriptive rather than prescriptive. It assumes that human mental processes and behavior can be studied and analyzed scientifically just like any other natural phenomenon such as light or gravity. By this I mean that they follow a lawful cause and effect that can be discovered by empirical observation. Admittedly, the laws may not be a precise as in some other sciences like physics, but that is only true because psychology is more complex than physics3.
In the scientific viewpoint, prescribing what someone should have done is simply the wrong way to think about behavior. Concepts such as "should," "ought," "blame," etc. don't exist in science. Behavior is what it is. Moreover, the "bad apple" approach promoted by Folk Psychology explanations of merely saying that the person was stupid or careless or should have done this or that explains nothing. It is unlikely to produce a correct understanding of causation and to suggest effective safety improvements. The scientific view can say whether a person acted within the realm of normal human behavior, but the notion of normality is statistical in the sense that is defined by empirical data. It is not based on some preconceived Folk Psychology notion of what someone thinks the actor should have done.
All of this leads me to another important psychological principle best summarized by psychologist Fred Skinner's deep philosophical insight that (now say it after me):
The rat is always right
I have explained what he meant on another page, Thinking Like A human Factors Expert, but in short, it was a warning to his fellow psychologists. If you perform an experiment, and the rat does not do what you want him to do, it isn't because the rat was stupid, lazy, or inattentive. It was because you set up the experiment incorrectly. Rats can only be rats. Granted, humans are more flexible and adaptable than rats. While this very flexibility and adaptability is what sets us apart and underlies our success as a species, it often leads us into trouble. If you want people to behave differently, then change the situations and especially the incentives. (See below.)
There are two main points to take away from this discussion. First, as a scientist, I never say that the person should have done anything. Scientists are in the explain business, not the blame business. At most, a scientist can say is whether the person acted within the bounds the typical behavior suggested by the data. Unlike the prescriptive view, my notion of expected behavior is not based on some preconceived idea but rather on the study of how humans actually behave. Of course there are bad rats. If some rat acts outsides the bounds of expected rat behavior, that could be construed as being blameworthy rat behavior, but that's not my business as a scientist.
Second, our human nature is immutable. It is who we are and what we are. It is the most difficult part of any man-environment system to change. Proposed safety improvements that require a change of human nature or worse, run contrary to that nature, are bound to fail. While human nature is immutable, however, human behavior is not. Human behaviors can be changed. The best way to modify behavior is to change the environment, as explained by the safety hierarchy. To quote James Reason (the guru of human error): "we cannot change the human condition, but we can change the conditions under which humans work."
The greatest human ability is adaptability to the conditions. What is it that humans adapt to? The answer lies in perhaps the most important principle of human behavior, Thorndike's Law of Effect:
Behavior followed by favorable consequences is more likely to be repeated. Behavior followed by unfavorable consequences is less likely to be repeated.
The Law of Effect may seem so laughably obvious that it need not be stated. Yet the distribution of favorable consequences (reinforcements) and negative consequences (punishments), the incentive structure, is an important environmental factor that is often overlooked in many analyses. Time and again personal and social disasters (e.g., the 2008 financial meltdown) occur because the incentive structure reinforces destructive behavior. The designers of any new system or policy should give hard thought to the intentional, and especially the unintentional, incentives that are being created. To quote Charlie Munger's paraphrasing of the Law of Effect, "Show me the incentive, and I'll show you the outcome." I direct you to a series of YouTube videos called "Great Moments In Unintended Consequences" for epic examples of incentives gone awry.
Conclusion
The important takeaway from this discussion is that, whether we are aware of it or not, we all use a conceptual model of human nature to judge the behavior of other people4. Such a model is necessary to explain what happened and to make causal attributions. For most people, however, the model is likely to be faulty if it is based on prescriptive notions derived from hindsight bias, counterfactual thinking, folk psychology, etc. The rest of this discussion presents a conceptual model of human nature based on scientific research with an added dollop of experience from my years of investigating human mishaps.
Armed with proper, scientific orientation of human decision-making and behavior, the next step is to examine human nature(s) in detail. The reason for "(s)" is that humans have two more or less distinct natures, one old and the other new.
Next: System 1 vs. System 2
Endnotes
1"Kludge" is an old computer term for a messy workaround or makeshift solution to a problem. This is the best description of human survival and information processing methods. They do not form a nice neat package. They were not designed to make life easy for scientists to uncover and classify.
2I had considered including a section on causal reasoning in mishap investigation, but that has already been somewhat discussed in the pages Death By Uber 4 and Causation.
3As Murray Gell-Mann, the Nobel Prize winning physicist, said "Imagine how hard physics would be if particles could think."
4Actually we all have many different conceptual models of other people that we apply under different circumstances. The most obvious is that we apply different models to ourselves and to other people. For example, the fundamental attribution reflects the tendency to attribute causation to dispositional factors when explaining the behavior of other people but situational factors when explaining our own. We also apply different conceptual models to different classes of people, e.g., our tribe vs. their tribe.
.
