The evolutionary roots of paranoia

The survival of the suspicious

There is no such thing as paranoia. Your worst fears can come true at any moment. — Hunter S. Thompson

Paranoia—the unfounded suspicion that others are deliberately trying to harm you—pervades human experience. Most dramatically, paranoid (or “persecutory”) delusions are the most common symptom of psychosis, that profound "loss of contact with reality" characteristic of various psychiatric and neurological disorders. Consider, for example, John Nash, the mathematical genius and Nobel laureate who believed he was the target of an elaborate conspiracy orchestrated by the Pope and CIA—a classic example of paranoia's role in schizophrenia (depicted in the film A Beautiful Mind).

However, paranoid delusions extend far beyond schizophrenia, appearing in bipolar disorder, dementia, delusional disorder, Parkinson's disease, brain injuries, epilepsy, and substance abuse. In fact, as Nichola Raihani and Vaughan Bell observe,

“In terms of the causes and contexts in which it appears, paranoia is perhaps most remarkable for being associated with such a wide range of difficulties, impairments and stresses.”

Moreover, paranoia is not confined to full-blown persecutory delusions characteristic of mental illness or dysfunction. Such beliefs exist at the extreme end of a paranoia continuum distributed throughout the general population. As Daniel Freeman writes, “Many people have a few paranoid thoughts, and a few people have many.”

We all know paranoid people and most of us experience paranoia occasionally ourselves.

Why?

Why are propensities towards paranoia so widespread? And why is paranoia associated with such a wide range of psychological disorders, disturbances, and difficulties?

More fundamentally, why is paranoia such a common failure mode of human cognition?

The short answer, I think, is this: Other people constitute the most dangerous and complex threat humans confront, and the cognitive processes and strategies we have evolved to manage this threat are unavoidably fragile.

Here is the long answer.

Hell is other people

Humans are intensely social (in fact, ultra-social) animals. Most of our most important goals are social (e.g., relationships, community, approval, status, etc.), and even our ability to achieve non-social outcomes (e.g., food, shelter, warmth, etc.) depends and has always depended on extensive social support and cooperation.

This intense sociality has a nice side. However, because we are products of evolution by natural selection, it also has a dark side. Social competition and conflicts of interest are baked into human social life.

Put simply, other people and the relationships, community, and status they provide are not just the fundamental source of meaning and purpose in human existence. They are also the greatest threat we must navigate.

Social threats

People can often promote their interests by eliminating, hurting, controlling, manipulating, and deceiving others. This is why warfare, raids, slavery, hierarchy, domination, bullying, conspiracies, and exploitation are ubiquitous throughout human history, including so-called “pre-history”.

Admittedly, the profound dangers of human social life are often not legible to many upper-middle-class professionals living in affluent liberal democracies today, where there is a lucrative market for the idea that, deep down, humans are cuddly, prosocial, and unthreatening.

(The Dutch historian Rutger Bregman wrote a bestseller in which he literally refers to our species as “Homo Puppy”).

Nevertheless, this perception is an artefact of recent centuries of institutional development in Western countries (a strong state, rule of law, courts, police, democracy, etc.) designed to clamp down on the most violent forms of social predation and exploitation, not an accurate understanding of human nature or the conditions in which our species evolved.

Moreover, even in this historically and culturally unusual context, the main thing people must worry about is still other people. Even if most people in, say, the Netherlands are not worrying about being murdered or enslaved, they are worrying that others might cheat on them, manipulate them, gossip about them, destroy their reputation, and so on.

Social threats—those threats that originate in the hostile, exploitative, or manipulative intentions of others—are and always have been the most dangerous feature of the environment humans must navigate.

Moreover, they are a unique threat. Unlike, say, predators, diseases, or physical accidents, social threats originate in rational agents with the ability to engage in sophisticated forms of reasoning, planning, social coordination, deception, and concealment.

The evolution of social vigilance

Humans have evolved cognitive processes and strategies designed to manage social threats.

This “social vigilance” is complex. For example, it includes “epistemic vigilance”, cognitive processes designed to manage the risks of deception and misinformation posed by our species’ unique reliance on communication. Such processes cause us to reject what people tell us unless we deem the message plausible, have positive reasons to trust the messenger, or are provided with persuasive arguments.

However, epistemic vigilance is just one facet of social vigilance.

Most fundamentally, social vigilance is concerned with detecting, anticipating, and responding to social threats. Although there is still much we do not know about how this works psychologically, it seems to involve four aspects:

1. Error management

2. Calibration

3. Reasoning and Rumination

4. Response

Error Management

When detecting threats, cognitive systems can make two distinct errors: false positives (mistakenly sounding an alarm) and false negatives (failing to detect a threat). Since mistakes are inevitable, detection systems must weigh the relative risks of such errors and adjust their sensitivity accordingly.

For example, suppose false alarms are much less costly than false negatives. In that case, systems will be biased towards producing false alarms to guard against the more dangerous mistake of failing to detect real threats. This is known as the “smoke detector principle”.

The smoke detector principle has shaped human psychology in powerful ways. For example, it plausibly explains why people over-attribute agency to events (because failing to detect the role of agency can be catastrophic), why men over-perceive sexual interest from women (to avoid the risk of failing to detect genuine interest), and why superstitious tendencies are so widespread (because detecting illusory cause-effect relationships in random patterns is less costly than overlooking real ones).

The smoke detector principle is also likely at play in social vigilance. To be clear, excessive vigilance for social threats can be costly. For example, it can undermine the cooperative relationships we depend on. However, this mild paranoia is typically less costly than the risks of being exploited and manipulated.

Calibration

Adaptive threat detection should be calibrated to the degree of risk an individual confronts. In the case of social vigilance, this depends on two overarching factors.

First, how dangerous is the social environment? There are significant differences between navigating a civil war in the Democratic Republic of the Congo and navigating a university campus in the UK, and adaptive social vigilance must track this difference and adjust its sensitivity accordingly.

Secondly, how vulnerable are you to other people’s exploitation? Once again, there is significant variation here. A highly competent individual with extensive material resources and social support is far less vulnerable to social predation and manipulation than someone who lacks these cognitive, material, and social assets, and adaptive social vigilance must reflect that difference.

There is considerable evidence that individuals track these variable features of themselves and their environment and adjust their social vigilance accordingly.

For example, minority groups typically exhibit proportionally higher rates of paranoia than majority ethnic groups within a population, plausibly because minorities are more vulnerable to marginalisation and exploitation. In support of this explanation, this tendency is reduced when minorities live in communities with a high prevalence of co-ethnics (the “ethnic density effect”), which likely increases their feelings of social support and safety.

Moreover, as with most cognitive processes, this calibration is likely especially influenced during so-called “sensitive periods” such as childhood and adolescence, when brains extract broad and deep-rooted expectations about the world that are difficult to unlearn. This might explain why things like childhood abuse and bullying are so strongly correlated with paranoia.

Reasoning and Rumination

In thinking about social vigilance, it is tempting to understand the detection of social threats as a quasi-perceptual task involving scanning the immediate environment. However, this is misleading.

First, social threats involve invisible intentions which can only ever be inferred. Unlike, say, predators, sharp objects, or heights, you cannot perceive people’s hostile or exploitative intentions.

Moreover, this invisibility is greatly exacerbated by the fact that manipulative or hostile people often try to conceal or misrepresent their intentions. “Detecting” social threats, then, must involve complex and uncertain inferences, including consideration of the possibility that agents are deliberately trying to evade detection.

Second, detecting actually-occurring exploitation or manipulation (“Oh, I am a slave”) is much less valuable than anticipating social threats so that one can avoid the relevant risks. For this reason, social vigilance must involve inherently forward-looking cognitive processes.

People must be motivated and able to consider, explore, and meditate on the possibility of complex plans involving hypothetical and potentially elaborate scenarios. In other words, social threat “detection” necessarily recruits the most sophisticated capacities for reasoning, imagination, and mental time travel humans possess.

Responding to social threats

Finally, there is no point in tracking social threats without responding to them appropriately. Social vigilance must, therefore, guide strategic responses once suspicions of social exploitation and manipulation arise.

This is immensely challenging for multiple reasons. First, the adaptiveness of different responses depends on the nature of the suspected threat, which can vary enormously and which individuals can rarely be certain about. Second, the appropriate response to a given threat is also highly uncertain and depends on a range of factors that are difficult to estimate or predict. Finally, any responses must take into consideration how the threat will respond to the responses.

Should you respond with anger or submissiveness? Should you let it be known that you have detected the threat or keep it secret? Should you try to escape and isolate yourself or recruit support from others?

There is no simple answer to these or countless other questions, even when the social threat is known with certainty, which it rarely is.

The roots of paranoia

These reflections on the complexity and challenges of social vigilance can illuminate why paranoia is such a common failure mode of human cognition, both in the non-clinical (“neurotypical”) population and among those experiencing psychological disorders, impairments, or distress.

The roots of non-clinical paranoia

There are multiple reasons why paranoia emerges so readily in human cognition.

First, and most obviously, because social vigilance is likely designed to err on the side of caution, our psychology tends to generate false positives rather than risk false negatives. This inherent bias toward overdetection creates a natural tendency toward paranoid thinking.

Second, as we have seen, social threat detection involves unique complexity and radical uncertainty. Unlike physical threats, social threats involve invisible intentions that must be inferred from ambiguous evidence. This creates an intrinsic difficulty in calibrating our responses appropriately. It is easy to misinterpret innocent actions as threatening.

Relatedly, social threats involve game-theoretic aspects that make them uniquely challenging to reason about. Because people and their plans can anticipate our responses and adjust accordingly, we must often "expect the unexpected”. And because social threats originate in sophisticated rational agents, the range of possible threat-related scenarios is effectively infinite. This can create cycles of strategic thinking that easily spiral into paranoid speculation about concealed plots and elaborate hidden agendas.

Moreover, once paranoid suspicions are activated, they are inherently difficult to disconfirm. What evidence could disprove the suspicion that you are the victim of a carefully concealed conspiracy?

Finally, responses to anticipated social threats can easily exacerbate the situation. For example, responding with hostility to a merely imagined social threat can make the target of that hostility more hostile—and, hence, more threatening. More commonly, trying to seek safety from perceived exploitation and social dangers—for example, by avoiding others and seeking to be alone—makes it difficult to learn that one’s suspicions are unfounded. As Daniel Freeman puts it, safety-seeking

“defences—eg, avoiding activities, shutting curtains, avoiding eye contact—prevent the receipt and processing of disconfirmatory evidence. In other words, what the defences end up protecting isn’t the individual but rather the false ideas.”

For all these reasons, social vigilance is inherently fragile.

Of course, most of us avoid the worst forms of this. Because of the adaptive importance of successful social vigilance, we have evolved to be remarkably skilled at this challenging task. In addition, we often design social environments and institutions to reduce the profits of exploitation and predation, effectively "outsourcing" much of our vigilance for social threats to collective mechanisms. For this reason, most of us manage to avoid living perpetually paranoid lives.

But not all of us.

The roots of clinical paranoia

As noted at the beginning, extreme and often delusional forms of paranoia are common across a surprisingly wide variety of psychiatric disorders and disturbances.

Misunderstanding clinical paranoia

One popular way of understanding this phenomenon is to posit malfunctions within a special-purpose cognitive module underpinning social vigilance. For example, in an excellent book on delusions, Ian and Joel Gold argue that social vigilance is underpinned by a modular “suspicion system”, a highly compartmentalised, special-purpose system designed to detect social threats. When this system becomes disordered, they argue, the result is delusions, including—most commonly—paranoid delusions.

This explanation faces two significant problems.

First, social vigilance does not seem like the kind of task that could be underpinned by a special-purpose, compartmentalised mental module. For reasons already identified, it is a global, expansive task that must recruit almost all cognitive processes and access virtually all the information at an individual’s disposal.

For example, there is no natural class of “input” for social threat detection. In principle, anything could qualify as evidence of social threat depending on broader context, history, and circumstance: an unexpected look from someone you know well, the discovery your furniture has moved position, finding a stranger’s cigarette butt in your garden when you get back from a trip, and so on.

Second, clinical paranoia emerges from a remarkably diverse array of conditions and disorders, from the global disturbances found in schizophrenia to various forms of brain injury and drug abuse. It seems implausible that all these diverse conditions somehow narrowly affect the same highly specialised cognitive module.

Understanding clinical paranoia

A more plausible explanation involves interactions between several factors.

First, precisely because social vigilance involves complex, global, and fragile forms of cognitive processing, it is uniquely vulnerable to disruption. Many psychological disorders involve profound disturbances to how individuals experience and think about the world, producing general confusion and global alterations to how they process information. Even if such disturbances do not selectively target social vigilance, they are likely to have far more dramatic consequences on social vigilance than on the performance of simpler cognitive tasks.

Second, paranoia feeds on feelings of vulnerability. When we feel unsafe, we instinctively become more suspicious. One effect shared by a diverse range of psychiatric conditions and disturbances is to increase an individual’s feeling that they are vulnerable to exploitation or predation. If you are disabled, confused, or distressed, it is easier for others to take advantage of you.

As Raihani and Bell speculate, increased paranoia in response to impairment might, therefore, reflect the operation of a "cognitive failsafe" because cognitive impairment increases vulnerability to exploitation.

(Note that this reaction makes more sense if you imagine the far harsher worlds our ancestors navigated before the emergence of professional medicine, healthcare, welfare, and so on).

Third, psychological disorders and disturbances can easily interact with epistemic vigilance. As Hugo Mercier observes in Not Born Yesterday, people’s openness to listening to what others tell them (e.g., “You’re being paranoid!”) is precarious at the best of times. When people are distracted, confused, or distressed, it tends to evaporate, a response which usually protects people from deception and manipulation but can become highly maladaptive in delusions. When combined with heightened feelings of social vigilance brought on by feelings of vulnerability, this can make it difficult to talk people out of their unfounded suspicions.

Finally, when people are confronted with strange events they cannot understand or explain (a common experience in psychotic disorders like schizophrenia), it is natural to consider the possibility such events have their roots in the intentional behaviour of other agents. (This natural tendency also likely plays a role in religion). And if one feels vulnerable and fearful, these explanations will likely be biased toward paranoia.

For all the reasons already documented, it should be clear why such suspicious ideas can then quickly become unmoored from reality when individuals are experiencing disorders or distress.

Not only does the complexity of social vigilance make it easy for this process to become unreliable, but once individuals seriously consider paranoid explanations of strange events, their minds can naturally gravitate to far-fetched scenarios, and they will find it challenging to disconfirm their suspicions. Moreover, they will likely respond in ways (e.g., employing safety-seeking defences) that exacerbate their paranoia.

Conclusion

In summary, paranoia is such a common characteristic and failure mode of human cognition because it involves cognitive processes that are inherently complex and fragile. Social vigilance operates under profound uncertainty, targets invisible intentions that can only be inferred, and guides responses to threats that can anticipate and adjust to those very responses.

That this process frequently generates false positives, occasionally spirals into paranoid delusions, and is especially vulnerable to psychological disturbance is unsurprising. In some ways, the more surprising thing is how successfully most people navigate the dangers of social life.

[Note: this post involves ideas clarified, elaborated, and justified in much greater depth in my forthcoming book on delusions co-authored with Kengo Miyazono and Sam Wilkinson, tentatively titled ‘The social roots of delusions’].