brian_knutson's picture
Professor of Psychology and Neuroscience; Stanford University

I worry about worry. Specifically, is our worry directed at the right targets?

The adaptive value of worry is that it helps us to avoid death without having to experience it first. But worry can only save us when targeted towards actual threats, and only by eventually eliciting action (in the form of avoidance). Although worry seems to be caused by external factors, it is not. The neural worry engine is always on, searching for its next target (like Freud's free-floating anxiety).

The ancestral environment probably "tuned up" our worry engines, since individual differences in levels of worry show significant heritability (around 50%), and are normally distributed (i.e., most people experience moderate rather than minimal or excessive worry). This bell-shaped distribution implies that over generations, those who worried too little died (or were eaten), while those who worried too much failed to live (or reproduce). Thus, our forebears' menu of environmental threats likely selected an "optimal" level of worry.

Less appreciated is the notion that the ancestral environment selected not only the level, but also the most compelling targets of worry. Consider common targets of phobias, such as public judgment, snakes, spiders, and heights. Unless you live in the jungles of New Guinea, these are probably not the existential threats that deserve the energy of your worry engine. In fact, leading causes of death in the United States typically involve more "boring" fates like heart disease, cancer, stroke, and accidents; encouraged by more proximal causes including smoking, alcohol consumption, poor diet, lack of exercise, cars, and firearms. This "worry gap" between imagined and actual threats implies that our worry is often misdirected towards the wrong targets.

My metaworry (or worry about worry) is that actual threats are changing much more rapidly than they have in the ancestral past, which could widen the worry gap. Humans have created much of this environment with our mechanisms, computers, and algorithms that induce rapid, "disruptive," and even global change. Both financial and environmental examples easily spring to mind—witness crashes of global finance bubbles and the palpable rise in global temperature over the past decade. Not only are these changes rapid with respect to an evolutionary timeframe, but they plausibly result from human causes. I worry that our worry engines will not retune their direction to focus on these rapidly changing threats fast enough to take preventative action.

One could try to close the widening worry gap with actionable data. Personal change is one option. Drawing from targets described above, if we could calculate the relative risk of death by snakes, spiders, cars, and guns, we could theoretically compare the discrepancy between the ranks of what we actually worry about (e.g., spiders) and what statistically causes more deaths (e.g., cars). Then we could try to swap our lower ranked worry with a higher ranked worry. One could even sketch the outlines of software that facilitates more optimal reallocation of worry (I'm looking at you, mobile application developers). As long as we're going to worry about something, it might as well actually threaten our existence.

Another option for closing the worry gap involves policy change. Governments undoubtedly collect valuable data on the relative potency of various threats, so as to direct their limited resources towards reducing the most pressing dangers. But information alone is not enough—it must motivate behavioral change. Thus, laws (with enforceable sanctions) are sometimes necessary to transform this information into action. Unfortunately, in the case of global threats, governments must coordinate their laws. This is not impossible. It has happened in the past, when governments came together to ban chlorofluorocarbons in order to stop atmospheric ozone depletion. It may happen in the future with respect to carbon sequestration and global climate change. But it requires a massively coordinated and continued effort.

For both personal action and public policy, I am advocating metaworry rather than hyperworry. Hyperworry feeds on itself. Escalating worrying about worrying could fuel a positive feedback loop, ending in a fearful freeze. Since the brain has limited energy, we should probably view worry as a resource to be conserved and efficiently allocated. Beyond increasing the level of worry, metaworry here implies redirecting worry based on incoming information. Turning our ancient worry engines in new directions may be difficult, but not impossible.

Consider laws requiring the use of seatbelts while driving. What I find miraculous is not that these laws have become mandatory in most American states, nor even that they reduce fatalities as predicted based on studies of crash test dummies and cadavers, but rather that the laws now grab me at a visceral level. Twenty-five years ago, I would never have thought twice about driving without a seatbelt. Now, when I am unable to buckle my seatbelt, I feel uneasy, anxious, and tense. In a word, I worry—and seek to close the gap.