On Human Nature(s) 5: Applying the Kludges in Case Analysis
Marc Green
This final section places the kludges discussed previously in context so that they can reveal a cohesive story. After all, we humans like stories, and they are perhaps the best way to learn new facts and lessons. As Kletz (2001) says, “We learn more from stories, true or fictional, than from statements of principle and exhortations to follow them.”1
I first note that another page, Perceptual Factors in Police Shootings of Unarmed Suspects has already applied many of the kludges to a real-world mistake-of-fact shooting. The principle issue in this event was visual perception in nighttime low visibility conditions. The police officer mistook a man’s wallet for a gun and fired. The situation described below occurred in daylight and there were no visibility issues. The emphasis is more on schema and on non-visibility sensory factors, especially looming.
The Facts
The incident to be analyzed was described at the beginning of this series. Not all the kludges discussed are relevant to this or any other single event. However, I’ll add a few more additional concepts and also discuss the topic of perception-reaction time. To refresh memories, here are the major events in the shooting in question:
- A police officer was patrolling in his one-man cruiser on a cold winter’s day. He received a call saying that a stolen car had been seen at a local high school. The call also said that the car had been associated with several ATM robberies;
- The officer drove to the school and saw the stolen car in the parking lot. He positioned his cruiser across the only exit/entry to the parking lot. There were several cars parked in nearby spaces, and the lot was bordered in many places by snow banks created when snowplows had cleared snow;
- The officer exited his cruiser and walked part way to the car with his weapon in low-ready position. He stopped just to the right (vehicle left) of the driver’s side. According to the accident reconstruction report, the car’s front wheels were aligned 15 degrees leftward toward him;
- The driver revved the car engine three or four times. One witness testified that the driver looked at the officer “mockingly.” The officer told him to put his hands out through the window, which he did briefly before pulling them back in. The officer then repeated the command and the driver put his hands out the windows for three to four seconds. The officer moved to a new position about 12.7 feet from the car, again just outside the front left fender. At that point the driver pulled his hands back into the vehicle and out of sight;
- The driver hit the accelerator. The car’s rear tires initially slid rightward on ice causing it to yaw as it move forward leftward on a curved path toward the officer at a somewhat broadside angle;
- The officer fired twice in rapid succession, the second shot being fatal. The angle of the second bullet trajectory shows that the officer was near the driver’s side window. As the car moved past the officer, it passed within an inch or two of his leg or even possibly brushed him before stopping up against a snow bank;
- Subsequent accident reconstruction concluded that from the time that the driver initiated acceleration to the first shot was about 2.65 seconds. In the first 0.5 second, the car moved only six inches. In the first 0.75 second the car moved a little more than a foot toward him, and in the first full second the movement was about four feet; and
- The officer was put on trial for murder.
To start the analysis, I want to back up a bit and talk about the officer’s experience. This is always critical because humans are time travelers, their behavior is governed by the present situation but also by the past (experience) and the future (goals). The past is important because it reveals what his reasonable expectations might have been. He was a veteran officer. He had extensive experience in fire arms training, although he had never fired at a person in the line of duty. He had also previously made many traffic stops of cars that were reported stolen. All of these cases played out in one of two scenarios: 1) the driver simply surrendered, or 2) more often, the driver exited the car and took off so that the stop ended in a foot race.
This explains why he moved to a position relatively close to the car. He doubtless expected the event to play out according to one of the familiar scripts. The driver would surrender or more likely take off running. He approached the car because he did not want to give the driver much of a head start in case he bolted.
When the car started accelerating toward him, several psychological effects occurred:
First, suddenly his strong expectations were violated. Something that had always been true in the past was no longer true. He had never received any training on this situation, so he had no schema to fall back on. The first time is always different. A moment of shock and confusion likely ensued. As explained earlier, it is often the case that the actor feels the emotion prior to identifying exactly what is wrong. System 0 is immediately on alert.
Second, the car was looming toward him, eliciting the anxiety that such motions produce. As explained, looming is more threatening when the object is close and large. The broadside yawing movement made the car especially large. It was also very near. The initial 12.7-foot distance put it in or near his defensive peripersonal space (DPPS), but as the car moved forward, it undoubtedly entered the DPPS. The loud engine sounds probably also increased the stress. Finally, humans interpret loud engines sounds as meaning higher speed.
What to do at that moment? When an imminent threat arises, the general options are, to use the common phrase, “freeze, flee, or fight.” This alliterates nicely but is a bit misleading. It would be better said as “freeze, flee, or confront.” Freezing was likely to result in being struck by the car, so was it not an option. Fleeing was also not an option for reasons discussed below. Instead, he held his ground and confronted the threat. Just as a person trying to ward off collision with a suddenly stopped vehicle ahead triggers a well-practiced motor schema and slams on the brakes, the officer did what he had done so often on the practice range. He raised his weapon and fired.
To put this in terms of the previously discussed kludges, in analytical System 2 mode, behavior is mainly "top down" driven by events occurring in the brain’s cognitive processing. In System 0/1 mode, the "bottom-up" sensory information is more directly connected to response. It is a short-cut that saves time by bypassing System 2 rumination. It seems instinctive. The speed of System 0/1 mode is advantageous in dangerous situations where time is short and response must be made very quickly. Its downside is that there is little opportunity to analyze the situation and consider response alternatives. That's part of what is meant by perceptual narrowing.
Why Didn’t He Move Aside?
One of the prosecution's major points was that the officer could have chosen to flee. He could theoretically have simply avoided the approaching car by stepping aside to the left or right. As explained, humans automatically move left to avoid and threat coming from the right and to the right when the threat comes from the left. A problem arises when the threat is coming straight toward the actor. There is no obvious, automatic direction to safety. Turning around and running in the opposite direction would take time and assume that he could outrun the car.
The car’s yaw as it generally moved rightward toward him made matters worse. The car essentially grew in both the left and right directions. That is, the front bumper moved leftward while the rear moved rightward. This was magnified by the increase in retinal image size (i.e., looming). In sum, moving in either direction would seem to put him in the car’s path. Moving in either direction would have seemed to be moving toward the threat, which people do not like to do.
In an ideal world, he would have fallen back to System 2 and perhaps formed an escape plan. But time was short and deciding on an escape plan would require assessing several variables such as the route the car was taking, the direction it was safe to move and so forth. According to the engineering report, the realm of possible paths was very large, so it would be difficult in an analytical mode to be certain about where to escape or that any escape was possible. The car was looming and yawing as it approached him. This both magnified the rate of looming and the unpredictability of its path.
Another important factor was that he did not have a mental map of the area. People who are familiar with an environment develop mental maps, meaning they know the spatial relationships of objects in the environment without the need to look at them. For example, you know the layout of your house. If it were suddenly filled with smoke and fire, you would know where the best exits were to get out without much thought. If you were in an unfamiliar place, you'd have to spend time looking around and figuring out where to go. The officer had been to the school before and knew the parking lot’s general layout. However, the position of objects such as parked cars and snow banks varied. When the need to flee arose, he did not know whether there were any safe routes or where they would be. This would have required analysis consuming time that he did not have.
Perhaps most importantly, perceptual narrowing focuses attention tightly on to the threat. He was then not likely noticing his surroundings and possible flight paths. It has been found that motorists, for example, frequently have collisions that they could have theoretically avoided if they had swerved at the last moment. Instead, they focus so intently on the vehicle ahead that they only slam on brakes as a response. They simply seldom consider the steering option. It is difficult to convey how the threat of collision with an imminent object causes an extremely narrow focus.
Perceptual narrowing drives people toward simple, immediate responses. The officer’s simple, overlearned motor schema—firing his gun—could be performed without planning or much cognitive delay so it was much faster. Further, a rule of thumb is that it faster to move small body parts than larges one, e.g., arms are faster than legs.
Why Was He so Close to the Driver When He Fired The Second Shot?
In many police shootings, two facts often come in conflict. First, the officer fires multiple shots. Second, officers are only permitted to fire if they believe that their life is in danger. In this case, that would be the fear that the car would hit him. As events unfolded, the officer fired through the open driver’s side window, so he must have been very close to the front bumper as it passed by. However, the second shot occurred later, and the bumper may have already passed by signaling that danger was over.
The timing of threat and response inevitably leads to the topic of perception-response time (PRT) that seems to arise in almost every police shooting. Technically, the term refers to the interval between the moment that a person perceives the need to respond and the moment that a response is initiated (note: not actually performed). Because of the PRT lag, it is theoretically possible to make the decision to fire while the threat is present but for the effect of the response to occur after the threat has passed.
Because of PRT lag, I can say for sure that the car was farther from the officer when he perceived a signal to shoot than the time when the bullet left his gun. The exact distance depended on the car's speed, on his initial position, and on the PRT. The accident reconstruction report estimated speed (actually acceleration) and a distance of about 12.7 feet. This leaves PRT.
PRT is very difficult to specify with great accuracy and confidence. Consider that PRT is a time interval that begins with a clock starter and ends with a clock stopper. The clock stopper is usual easy to identify. It is the initiation of the response. Determining the clock starter is much more difficult. There are two obvious candidates. The first was the time when a physical event began. Here, it is when the car started moving, which was 2.6-2.75 seconds prior to response. The second, is when the person perceived the need to respond. This is the usual definition3, but it is also the more difficult to pin down. Recall that the car moved only about six inches in the first 0.5 second, about one foot in the first 0.75 of a second, and about four feet in the first full second. Only then would the motion perhaps become obvious, but at that slow speed the car may not have seemed highly threatening. At the estimated acceleration, the car would strike him in about 2.6-2.75 seconds. Lopping off about a second to perceive the motion, the officer had about 1.6-1.75 seconds or quite possibly less to respond.
The other important threat-related percept was the looming rate. While the car was moving toward him, its retinal image was growing. Figure 1 shows the approximate looming rate as a function of time. Looming is a highly nonlinear function of distance and time. As the figure shows, it grows slowly at first and then suddenly takes off at just 2.0 seconds from the car’s initial motion. (The graph stops 2.25 seconds because the number at 2.5 seconds is so great that is literally off the chart.) This suggests that severe perceptual threat would only be perceived somewhere between 1.0 and perhaps 2.0 seconds after car movement began. At this point, the car moved would four to eight feet and the officer would be roughly between 11 and 5 feet away from the front bumper. The window was about six feet behind the bumper so was 16 to 11 feet from that part of the car.
Figure 1. Looming rate as function of time after the car’s initial movement.
As to actual PRT, studies vary somewhat, but suggest that the time to raise and fire from a low ready position is about 0.4 to 0.75 second. Using these values, the officer was just in front of or just past the front bumper when he reacted. However, none of the studies that I’ve seen put the shooter in a real life-threatening position. The shooter has been able to practice the same response repeatedly. And, of course, there are no surprises. So what would a realistic PRT estimate be? It’s anyone’s guess because of the many differences between laboratory studies and action in the real-world. Trying to be very exact about PRT is, in my view, a waste of time. It simply isn’t possible for many reasons, some explained in the page Let's Get Real About Perception-Response Time and the book Green (2024).
The final issue is the second shot. Because the car was moving toward him, it is undoubtedly true that the officer was closer to the driver when he fired his second shot than when he fired the first shot. Was he still in danger at that point? The correct answer is that it doesn’t matter. The two shots were likely part of a single chained response that began with the first shot. It is only the conditions existing at the time of the initial shot that determined the officer’s behavior.
Summary
This case analysis has used a lot of numbers, depending heavily on the accident reconstruction report. The accuracy and robustness of such numbers is always somewhat uncertain, so it is wise to avoid leaning too heavily on them. It is perhaps better to look at the big picture of how the behavior fit into the normal flow of human nature. From this vantage point, events can be described this way. When the car started moving toward the officer:
- The officer entered the situation with a well-rehearsed script (System 1 was primed));
- He became shocked and emotionally aroused when the schema/script failed (System 1 failed);
- The emotion was magnified by the looming car and loud engine (System 0 in control);
- He perceptually narrowed and did not consider alternative plans (System 0 doesn't "think"); and
- He automatically responded strongly to the threat using a well-rehearsed motor schema (System 0 directly connects sensory input to action);
One final note. In terms of the previous discussions, the officer could not use System 2 due to time constraints. However, this does not mean that System 2 played no role in the event. In section 2, I noted that treating Systems 1 and 2 as absolutely independent is probably going too far. Before the car moved, for example, the officer knew two facts: 1) the car had been associated with other crimes, and 2) the driver had show erratic behavior and made "mocking" expressions. System 2 may have then judged the situation as potential trouble even before the car moved.
Conclusion
This five part series has had several goals. One is to highlight the importance of a scientific view of human behavior. Another is to explain why an understanding of human nature(s) is necessary to determine causation. I also have attempted to provide a general overview of human nature(s). This includes a decomposition of human natures into an array of kluges that evolution has pasted together so that we can overcome (to some degree) our inherent limitations. The discussion has largely focused an explanation of how the human natures express themselves as an array of these kludges that are designed to cover the spectrum of possible situations. The range extends from situations that require careful, accurate analysis, to those which are routine and can be automatically handled by preprogrammed behaviors, and finally to those that are one-of sudden emergencies leaving little time for more than a primitive defensive response. Last, I have attempt to use these kludges to explain a real-world event.
My parting piece of advice is: don’t pay too much attention to books like The Darwin Awards and to research like Kahneman's artificial experiments which similarly attempt to paint humans as stupid, irrational, and prone to poor decision-making. Certainly, people sometimes do stupid things (or things which appear stupid from the outside or in hindsight), but people are generally not stupid. Or irrational. Or prone to reasoning errors. Quite the opposite. We are amazingly adaptable and successful beings who are able to go far beyond our inherent limitations by means of a patchwork of clever kludges. We mistakenly take our success as normality rather than the extraordinary feat that it is and only notice the rare times when things go awry. As a result:
We make thousands of decisions every day in automatic mode without a mistake. Yet we don’t reflect and celebrate this wonderful mode of human decision-making at work rather, we put the blow torch on the one moment when it doesn’t work and something goes wrong. (Rob Long)
Sure, our human natures encounter situations where they sometimes fail, but the failure often lies in the situation and its incentives, not in the person who must act. And certainly, even the greatest source of our adaptability, the Law of Effect, can lead us down the trail to practical drift or some other rabbit hole when taken to extreme, but that does not detract from the critical role it plays in the normal successes of our everyday lives. Last, never forget that although we are not perfect, we could never be perfect because, like any machine ever that has ever existed, we have limitations built into our architectures. Anyway, we would find that perfection, even if possible, usually extorts a price that would be irrational to pay.
Endnotes
1The internet is full of websites that criticize the use of "anecdotes" in decision-making. While anecdotes without science should certainly be viewed skeptically, it is equally true that science without anecdotes should raise some questions, at the very least, about empirical validity.
2The yawing introduces some complications in computing values for car's movement. I have generally ignored the yaw to make the analysis simpler and easier to follow. It's one sure effect was to increase looming.
3For example, the National Cooperative Highway Research Program Guide says “PRT does not start until drivers can see and, to some degree, recognize the hazard." The need for the phrase "to some degree" shows how uncertain the clock starter, and hence then entire concept of PRT, really is.
.
