Learning From the Octopus

by Rafe Sagarin

  Signaling is essential to security, but it is not without vulnerabilities. Signals can overestimate or underestimate the true risk of a threat. They can be perfectly correct in the information they give but be so far removed from the source of the threat that they don’t seem particularly relevant to the receiver. Or they can be so repetitious that either the receiver becomes habituated to them and they foment a much smaller reaction or they annoy the receiver to the extent that the signal is shut down if possible.

  There seems to be a consistent reaction to repetitious, unchanging signals: checking out. Dan Blumstein, the marmot lover, and ecologist Elizabeth M. P. Madin investigated this phenomenon, what behavioral ecologists call “habituation,” by examining visits to the Department of Homeland Security’s Ready.gov website and calls to the DHS information line—which is supposed to inform civilians about what they should do to be prepared for homeland security emergencies—in relation to the threat level. They also looked at public opinion polling data to see how worried people reported feeling about terrorism. They found that seeking information about terrorism decreased over time and had no relationship with the threat level.25

  Egypt cleverly exploited the habituation effect of signals in its conflict with Israel. In the weeks preceding the 1973 Yom Kippur War, Egyptian forces conducted forty military exercises in plain view of Israeli forces along the border. Israeli sentries had come to expect the movement of armed soldiers and armor in the region. The first morning of the war started as many of the preceding days, with Egyptian military exercises, but this time the Egyptian forces kept right on moving across the border and initiated a “surprise” attack on Israel that was planned in plain view for weeks.26

  Egypt turned a pernicious problem of repetitive signaling on its head to create a security problem for its enemy. Most of our signaling technology, while designed to keep us safer, in essence provides the same repetitive habituation, and ultimately makes us more vulnerable.

  An ever-present, never-changing signal is about as effective at keeping a vulnerable population alert as a diorama of a grizzly bear is at scaring a museum visitor. But we certainly all jump when an infrequent signal, such as a fire alarm, suddenly goes off. Nonetheless, as I’ve already mentioned with regard to the tsunami alarms stoned to death in Banda Aceh and the millions of disabled smoke detectors in American homes, most alarm systems don’t provide us with the security we would like them to give us. Likewise, explosive gas alarms on the Deepwater Horizon rig were disabled months before the deadly explosion there. Collision alarms on the Washington, D.C., metro system were routinely ignored before a deadly crash on a red line train there; air traffic control alarms have been ignored, leading to crashing planes27; and Vietnam-era fighter pilots, who suffered high casualty and capture rates, shut down surface-to-air missile alarms because they just had too much information coming at them in the cockpit. It’s not like any of these alarms are warning about trivial matters. Nor is it the case that these alarms don’t save lives. One study from Ontario, Canada, showed that nearly 30 percent fewer deaths occurred in house fires where a fire alarm was present and working compared to those homes without working fire alarms, whereas U.S. reports place the percentage of fewer deaths in homes protected by working alarms at about 50 percent.28 Most of these alarms are disabled or ignored because they produce too many false alarms.

  So how can alarm systems work better? Re-sampling—in other words, bringing in a redundant element—can help a detector determine if the original stimulus was in fact a real threat. Mike Dziekan reports on a company that had an optical fire detector installed above the camera that made their photo IDs. One time when a new ID badge was being made, the flash from the camera triggered the alarm, causing the whole facility to be evacuated. If the alarm simply sampled the environment multiple times it would have caught the false alarm.29

  Marmots and other animals in groups are able to use a form of re-sampling to ensure that they are getting honest signals. They can listen to the alarm calls of several individual marmots to determine if a threat is real. If only Nervous Nelly is making alarm calls, there is a much smaller risk than if Nervous Nelly and Cool Hand Luke and other marmots are making a call. Likewise, as I mentioned in Chapter 5, honeybees check multiple sources of information, essentially counting votes of many individuals in the colony, before making a decision such as the need to move the hive because of imminent danger.30 These represent the same type of adaptive redundant features discussed in Chapter 5. Obviously, a key factor in all of these redundant information systems is speed—they won’t work if taking multiple samples for the presence of predators or the presence of smoke takes a long time.

  But even the most robust signaling strategies can be overcome. When an enemy learns your signaling pattern, it can use it against you. A U.S. soldier can salute, wear a uniform, and march in line to signal his loyalty, but it doesn’t mean he is loyal. This contradiction was tragically seen when U.S. Army Major Nidal Malik Hasan killed thirteen people and wounded thirty others on Fort Hood Army base in Texas on November 5, 2009. Hasan was shot four times by a soldier on the base, ending his shooting rampage. Although he barely survived, it would seem likely that an attack of this type at a military base was a suicide mission, indicating both the dishonesty of his signals of loyalty to the United States and the honesty of his signals of loyalty to radical Islamic ideals.

  Suicide attacks are the ultimate expression of an honest signal—in this case, of commitment to a cause. Several of Hasan’s truly honest signals were known through his idiosyncrasies—he once launched into a spirited defense of radical Islam and suicide bombing during what was supposed to have been a medical lecture, and he communicated with Anwar al-Awlaki, a radical Muslim lecturer31—but these signals were not interpreted by his superiors as overriding his more routine signals of loyalty to the U.S. Army.

  YOUR FACE GIVES YOU AWAY

  Criminals like Hasan may be able to use a number of superficial signals—their rank and uniform—to send dishonest signals, but their biology and their biological legacy send far more honest signals out, and we need to become better at detecting those signals. In particular, humans give away a lot of information in their facial expressions and body language. Even just the shape of your face may reveal your level of aggressiveness. A study of hockey players showed that the larger the width-to-height ratio of a player’s face, the more likely he was to have accrued penalty minutes for fighting and other aggressive transgressions.32 It’s believed that this wider facial structure is linked to higher testosterone levels during development and thus higher aggressiveness.

  Humans also have a very strong capacity to discern this signal: people representing several different cultures were able to accurately assess upper body strength (also a proxy for testosterone levels) of strangers even when they could only see their faces.33 But even the toughest hockey players are typically not engaged in terrorist activities (although a heavy hitter like Alexander Ovechkin certainly terrorizes his opponents on the ice), so while unchangeable features such as facial structure may be a subtle clue to the personality of our fellow humans, they don’t give us the fine-scale signal of nefarious intent that we would need to identify who is immediately dangerous.

  Using behavioral traits to screen people entering secure areas exhibits multiple features of natural security systems. Like other social organisms such as wasps, wolves, and primates, we have a highly evolved ability (including brain specialization) for facial and behavioral pattern recognition.34 Crows, who are also highly social, can remember for their entire lives the face of a human that has mistreated them when they were younger, and they may even share this learning among their social group.35 Behavioral recognition is a redundant defense effective against multiple types of criminals (e.g., drug smugglers or terrorists) even if their planned crime involves no concealed contraband. It can operate independently of racial profiling—Darwin, through observing many different cultures during his voyage on the Beagle, recog
nized that there were common expressions across human cultures and between different animal species as well.36 Finally, behavioral screening returns control of uncertainty to the population it is trying to protect because it can be conducted from hidden vantage points or video. As a head behavioral screener at Dulles Airport (one of 161 airports where behavioral screening has been deployed by the U.S. Transportation Security Administration 37) remarked, “The observation of human behavior is probably the hardest thing to defeat. You just don’t know what I am going to see.”38 The efficacy of layering discrete behavioral screening with other levels of verbal and nonverbal intent detection systems is currently being investigated.39

  Behavioral screening is also not a perfect system. Scientists and the Government Accountability Office bemoan the fact that the TSA’s behavioral screening was deployed with very little scientific evaluation.40 They would like to see more experimental testing, but I’m skeptical. As an ecologist studying large-scale ecological changes like climate warming and fisheries collapses, I am increasingly convinced that many problems can’t be ethically or effectively studied with a manipulative experiment. Just as we can’t design an experiment to replicate the incredibly complex suite of changes that climate warming will bring—we have to rely on detailed, real-world observations of how warming has already affected the planet and then project forward—it’s hard to imagine an experiment that can study the effectiveness of behavioral screening before it’s implemented in a real-world setting. Probably no actor is good enough to mimic the human physiological changes and outward expressions of those changes that occur when the mind and body are under extreme stress, as a real-world terrorist or drug smuggler would likely experience. To really study behavioral screening, you have to set it up and carefully monitor how well identification of certain traits leads to arrests of criminals.

  Other critics argue that behavioral screening just hasn’t worked. USA Today recently reported that sixteen passengers evaded behavioral screening and later turned out to be linked to terrorist plots.41 But this shouldn’t be surprising. Humans are very malleable. If these people, who were later linked to terrorist plots, were not actually involved in a plot at the time they went through behavioral screening, it’s likely that they didn’t express the outward signs of concealed behavior. Behavioral screening is only one part of a multi-layered screening system.42 If these sixteen passengers were truly a threat to air travel, they should have been identified by intelligence operations and placed on “no fly” lists before they ever got to the behavioral screening phase.

  Still other critics argue that behavioral screening works too zealously, pointing out that many travelers are nervous about flying and likely to exhibit tense behaviors. This argument is backed up by the fact that of the 152,000 or so people singled out for secondary screening at airports where behavioral screening was used between 2006 and 2009, only 1,100, or less than 1 percent, were subsequently arrested, with critics pointing out that this is a low positive rate for the number of passengers singled out. There is an undertone of indignation (and some pending lawsuits) on behalf of the violated civil liberties of these unfairly targeted people. But the reality is, all of us are unfairly targeted at airports; from the moment we walk in under video surveillance to the time we shed our shoes and subject our luggage and bodies to X-rays, we are presumed guilty until shown to be innocent. For the most part, you check your civil liberties at the curb with your luggage.

  A wider look at the numbers weakens the arguments of the critics. Consider that roughly 2 billion people traveled through these same airports during the study period, meaning that less than 0.01 percent of passengers were detained somewhat longer than the normal delays for screening that everyone receives. If we expand the view even more to include arrests of all U.S. air travelers (including airports that don’t use behavioral screening) at screening checkpoints using the best numbers I could get from TSA, we find that 855 people were arrested at security checkpoints in 2008–2009,43 a two-year period in which roughly 1.4 billion people travelled in U.S. airports.44 This general arrest rate—0.0000006 percent—makes the 1 percent arrest rate of people detained through the more targeted behavioral screening seem like a pretty effective process after all.

  Going from animal signaling to the unique facial and behavioral attributes of humans begins to close the loop. We started by looking at nonhuman natural systems that give us some insight into how humans might design and implement security systems. Now we are looking directly at what human nature can tell us about security failures and opportunities in human societies. In the next chapter I further blur the lines between nature and human society through an exploration of how human nature and the belief systems that we develop as humans growing up in different environments become the source of many of our security problems.

  chapter eight

  THE SACRED VALUES OF SALMON AND SUICIDE BOMBERS

  SALMON GO TO GREAT LENGTHS to kill themselves. After a short few years frolicking in the open ocean, they may travel thousands of kilometers to get back to the precise stretch of the same river in which they were born. On this journey they will have to slip past the birds, bears, sea lions, and humans that gather at river mouths to feast on them. They must swim exhaustively upstream for many miles, using most of their energy reserves to leap up waterfalls or swim ladders (artificial waterfalls constructed on the sides of artificial dams) until they reach their spawning grounds, where their last gasps are spent producing eggs or fertilizing them with sperm before collapsing in death, never to see the ocean again.

  From an evolutionary standpoint, it’s not hard to find sense in these suicide missions—the salmon are passing on and multiplying their genes in a habitat that has already been proven (by the adult salmon’s own experience) to produce strong and reproductively fit salmon. People tend to admire the determination of the salmon. At the very least, we generally don’t call the salmon “irrational” or “crazy” for their journey. We do, however, freely launch those pseudo-psychological assessments on human suicide bombers. Yet salmon and suicide bombers are not as different as their outward appearance would indicate. The most important difference between them is neither fins versus arms, nor gills versus lungs, but that the salmon (despite the dams choking up the rivers) still lives in the environment its ancestors evolved in for thousands of generations, while the suicide bomber does not. I’ll argue in this chapter that suicide bombing is just an extreme case at the far end of a spectrum of behaviors related to establishing and reinforcing self-identity that impart survival to organisms.

  Those behaviors tend to be the things that many people would call irrational beliefs, and they turn up all the time in security questions. Pakistanis and Indians fight a seemingly endless war with potential for mutual nuclear annihilation over a narrow strip of barely habitable territory. People stay in their homes despite clear warnings and even as the floodwaters rise into their attics. And we ignore over 100 years of collected scientific wisdom while we watch human-induced climate change alter our entire planet.

  The naturalist’s view on security doesn’t allow us to simply label something “irrational” and then dismiss it. Just as a biologist wants to get to the root of what makes a peacock grow such outlandish feathers or an immune system suddenly turn on its own host’s body, a natural-security approach tries to get inside these behaviors that compromise our security, tracing their roots back as deep in evolutionary time as possible and figuring out what they mean in today’s society.

  Evolutionary psychologists, who study the ancient roots of modern human behavior, argue that religious fervor didn’t develop in the modern world but in a world completely unlike the one we have briefly inhabited now.1 In this early world, humans lived in small close groups that struggled constantly to obtain enough resources to survive. Only rarely did they encounter small groups of other humans, and if their interaction wasn’t about trading resources, it was likely because one group was trying to take the other group’s resources by
force.

  Yet almost all political analysis of human behavior tries to explain it within the narrow confines of the immediate sociopolitical environment. Some public commentators try to get us to broaden our thinking. Journalists try to remind the most short-sighted among us that there were clear signs of terrorist activity against Western targets years before 9/11. Historians admonish us to open our eyes and look at the thousands of years of history in places like Iraq and Afghanistan.2 Political scientists urge us to look at individual security crises within their global context. I fully support these viewpoints, but I suggest that analyses digging back ten, a hundred, or even a thousand years must be nested within a perspective that goes back orders of magnitude deeper into human, and biological, evolution.

  If we convert our years as humans on Earth to words in a book, analyzing security only in the context of the past few thousand years of human history is like trying to understand all of War and Peace by reading only the last word. If we don’t understand the true nature of human behaviors we will only, at best, address the most superficial manifestations of deep-seated hatred, mistrust, and suspicion. At worst, by trying to eliminate behaviors (such as religious belief systems) that we fear but don’t properly understand, we will create a rich environment for strengthening and replicating them.