The self-erasing idea

Some ideas have a curious property: they destroy themselves.

The self-erasing idea

Some ideas have a curious property: they destroy themselves.

We're all well acquainted with memes – those little scraps of images and ideas that seem to replicate themselves endlessly, especially online.

There are other ideas that have the opposite property: spreading them destroys them. Or, they have inhibitory qualities that prevent them from being spread in the first place. I'll call these ideas anti-memes.

Passwords are a kind of anti-meme. They are only useful when they are not shared. The moment a password is compromised, it ceases to be a secret. Less philosophically, if a user becomes aware that the password is compromised they will change it – "destroying" the original information in the process.

Claude Shannon, pioneer of information theory, famously had what he called "the ultimate machine" on his desk. This box contained a single switch. Activating it would open a lid and cause a small robotic hand to come out and flip the switch back to the off position. The purpose of the machine is an anti-purpose: to turn itself off. You can watch it in action here:

Some information is like that: knowing it only precipitates its own erasure.

Dreams are another common anti-meme. They are slippery and forgettable. When you don't outright forget them, you inevitably unconsciously mutate the details to fill in gaps. Even when you do remember them clearly, they almost always make for bad stories. They are both inherently hard to remember and hard to share.

Some anti-memes rely almost entirely on mechanisms that limit their transmission. Taboos  are something of a special case of anti-meme. Spreading the idea doesn't destroy it – instead, it potentially destroys you within your social context. Yet, taboos have a certain attraction. In fact, they can easily lead to something similar to the Streisand effect, where an attempt at keeping something secret paradoxically draws more attention.

Anti-memes arguably have both an internal, intrinsic structure and a social dimension. Dreams are an example of information that is inherently "slippery," and forgettable due to the neurophysiology of sleep. Similarly, very long sequences of numbers prove difficult to remember and therefore to share. Other anti-memes, like taboos, have largely social mechanisms that prevent their transmission. Yet others, like passwords, have an ultimate-machine like logical structure that self-destruct – once a secret is known, it is no longer a secret, etc.

We've already seen that there are artificial and natural anti-memes. In other words, some anti-memes are a product of the limits of an information system – the human brain in the case of dreams. Others anti-memes are distinctly human inventions.

Nearly all of them seem to work by interfering with how we interpret information.

Humans are reflexively sense-making animals. Neuroscientist Michael Gazzaniga described what he called the "interpreter complex" – essentially a neurological subsystem that integrates information together into coherent narratives.

As part of his work, he studied patients who had undergone an epilepsy treatment that severed the corpus callosum, a thick bundle of nerves that connects the left and right hemispheres of the brain. In these individuals, the left and the right hemisphere can no longer communicate causing a number of perceptual abnormalities stemming from discordant information.

This short video demonstrates how a patient is unable to verbally identify anything in the left visual field – he is only able to draw what he sees:

In fact, this ability of the hemispheres to coordinate and process information also interferes with simple tasks. For example, another split-brain patient asked to solve a non-verbal puzzle finds his two hands fighting each other  beyond his conscious control:

One does not need radical brain surgery to fail to integrate sensory information. Simply being distracted or focused on extraneous detail is enough. Many people by now have experienced this effect, but if you haven't seen it before I'll simply show you this video without comment:

Experiments routinely show that incongruent perceptual elements are filtered out of consciousness entirely, especially when people are given specific attentional tasks.

The mere presence of contrasting information can hijack and disorient the interpretive machine. Take for example "Navon figures." These are patterns built with a "global" shape or symbol (say a triangle, or the numeral five) but compromised with a contrasting "local" symbol (a circle or the numeral 8). Here are a few examples:

Navon Figures: https://www.researchgate.net/publication/264698665_Executive_function_of_shifting_Factorial_structure_and_relations_to_personality_and_intelligence_domains

Numeral Navon figure

Studies demonstrate that looking at these types of figures at length will reduce performance in face recognition tasks, and even psychomotor tasks.

In other words, merely perceiving a pattern with contrasting global and local information causes a kind of "interference" that briefly makes you less capable and less coordinated.

Aside from purely perceptual techniques, language is also a powerful tool to shape perception. For instance, psychologists and hypnotists speak of "transderivational search." This takes the form of language that prompts a reflexive, if internal response, in the listener.

For example, imagine someone says: "You did it again, didn't you!"

The "it" in the sentence is unclear and could be anything. However, you likely subconsciously replaced "it" with something specific from your personal experience.

Hypnotists use these linguistic instincts to discern subtle patterns in the thinking of their subjects. For example, a classic hypnotic prompt uses a nonsensical speech fragment like "and those thoughts you had yesterday," to elicit a response that can be analyzed or otherwise used therapeutically.

The power of transderivational search questions is that merely hearing the question is to know the answer. The mind compiles language into a search instruction, and dutifully complies. These prompts can easily alter the information processing capabilities of the listener. For example, if I say to you "that happiest day, when you felt the best," most people will likely "anchor" to some positive experience and slightly alter their mood.

The Duck Rabbit Illusion

Transderivational search prompts are like a multidimensional version of the "duck rabbit" illusion. There are a massive number of potential interpretations, all of them based on an idiosyncratic frame of reference. It is somewhat obvious that an ambiguous stimulus offers many possible interpretations. What is more interesting is that in some circumstances, the mind cannot help but to reflexively offer an interpretation – and then to subsequently adhere to it.

The perception and integration of information – shaped by attention and governed by a pre-existing frame of reference – produces a narrative representation of the world. Each step along the way can be hacked in a way that compromises the perception. It isn't just that perception is unreliable; it has very specific mechanisms that are vulnerable to exploitation.

Anti-memetic information is structured so that one or more of these mechanisms are impacted. Ambiguity is a common factor, as we've seen, though it is usually not enough on its own to engender a self-erasing or self-limiting effect. Plenty of things in life are confusing, sometimes profoundly so. Often they are ignored. It is only a small subset of curiosities that become dangerous obsessions – psychological Navon figures that leave us knowing less than when we started.

One simple formula to create an anti-meme is to combine ambiguity, complexity and social taboo. Confusing information prompts an initial transderivational reaction (what could this be? what does it mean to me?) Taboo and fear of seeming stupid or crazy then further shapes discourse. Subsequent transmitters of the information will invariably distort it either to make the information more palatable (e.g. less taboo) or, freighted with a personal hypothesis. Often both occur simultaneously. Strangely enough, information is removed from an already ambiguous prompt.

The net result is that the inciting information becomes garbled far beyond the typical "telephone game" effect.

Occasionally, subcultures form that become preoccupied with this type of information. Readers of this blog know them well: conspiracy theories, UFOs, paranormal and occult ideas, etc. This instigates an entirely separate process of the creation of "pseudoknowledge" through what scholars Introne, Elzeini and Iandoli call "stitching" and "mutating."

In essence, a process of participatory story-telling or communal sense-making begins. People begin to pool hypotheses (likely highly personal ones) and try to fit them together. Invariably, some key individuals take on a role as "custodians of the story" – facilitators who decide what new ideas and facts will be added to the master narrative.

However, if the original information has anti-memetic properties, repeated cycles of storytelling only further corrupt information. Honest attempts to "make sense of a confusing picture" end up inadvertently contributing new biases and presumptions about the data. Before long, the story takes on a life of its own and rapidly becomes highly complex.

Typically, this complexity also simultaneously reinforces taboo. Psychologists have long recognized a "simplicity" bias, where people tend to prefer simpler or more comprehensible inputs to complex one.

Here is a simple test: try telling someone completely uninterested about a conspiracy theory. Do you feel embarrassed? Do you lose their attention? Do you find it difficult to know which parts of the "story" to emphasize, and which to minimize? Do you find yourself censoring certain details to avoid seeming crazy?

Strangely, I think it is possible that the "smoothing" tendency in these collective narratives about taboo topics can eventually bring about a powerful inversion. An idea with anti-meme elements can actually eventually become a supermeme.

Here is how. Repeated cycles of communal storytelling add to the narrative until it is something a very large number of people will relate to, if only peripherally. Repeated cycles of trying to "censor" the taboo elements eventually helps sanitize and simplify a core narrative.

Before long there is an equilibrium point: the central idea has become simple enough to spread, but ambiguous enough to admit new interpretations. The idea is odd enough to be fascinating; palatable and simple enough to spread.

Attacks on the idea only make it stronger. Since the core idea is ambiguous, it can be stretched to subsume and integrate new facts. It becomes a world view in of itself – an explanatory frame that can be used to interpret almost any information whatsoever.

"Stitching" and "mutating" within adherent communities become so second nature, that any new information is instantaneously smoothed and incorporated. Everything becomes proof.

Worse, becoming "initiated" into the super-meme requires accepting a small dose of taboo – a binding "secret" between individuals. Earlier converts will have to swallow a larger dose of taboo because the core idea has not been sufficiently simplified yet to become hyper-charismatic.

In the right measure, taboo can actually power belief formation. Scholars of religion like Rodney Stark have documented how "demanding" religions can actually paradoxically become more powerful over time. By asking adherents to sacrifice something, religious movements can essentially mitigate the "free-rider" problem. In other words, only the dedicated join – and the truly dedicated make the best evangelists, and provide the best benefits to other adherents. Once the movement becomes large enough, especially when it starts to incorporate celebrities and other high status figures, it becomes unstoppable.

Anticipation is usually a key theme of an anti-meme that has become a super-meme. Recall, the original thing that causes an anti-meme to form is ambiguity – an unanswered question that seems to have vast yet vague personal significance. The subculture initially forms to "figure out" the underlying ambiguity. As the cycles of community story telling take place, there is usually an underlying theme of "resolution" or in religious contexts, "revelation."

One suspects this is the psychological engine behind the "disclosure" movement in UFO communities. The subculture is animated by the promise that the ambiguity will be solved; the answer is just around the corner. As a relatively young not-quite-religion, its adherents have to tolerate a lot of taboo – therefore there is a widespread focus on "commitment" that will someday be rewarded. There is also a hunger to recruit, especially individuals perceived to have high social value such as celebrities, authority figures, etc.

The sad problem is that by the time organized groups emerge, the anti-meme has done its work and the original information has been hopelessly corrupted, if not entirely lost. Belief contaminates evidence; rumors and struggle for narrative control overwhelm investigation. The global pattern has become so strong and self-reinforcing it denies even the ability to see contrasting local patterns.

Each attempt to evaluate the underlying information causes a reflexive, binary divide. One is either a "skeptic" or a "believer." Both sides bemoan the binary, but fall prey to it all the same. Usually, because a central element of the anti-meme is an ontological question: is this strange thing real or not? Is there actually a pattern in the first place? Each piece of information is evaluated in light of this question; each attempt to think through the problem succumbs to a shifting mosaic of global and local patterns.

At the outset, I claimed that some anti-memes are natural (like dreams) and others are engineered (like secret passwords.) It stands to reason that super-memes like popular conspiracies or UFOs are a mixture of both, although the distribution remains difficult to assess – an inherent problem of dealing with weird, mutating ideas.

One thing however is clear. An anti-meme that has become a super-meme is potentially very dangerous. Over time, it will eventually become almost entirely consumed with pseudoknowledge, a negotiated social construct of meaning on something inherently inscrutable.

The inherent anti-meme properties will squeeze out any real information – either because it is too strange to be incorporated, or because it is hostile to the pseudoknowledge narrative. Attempts to "correct" it via debunking run the risk of only improving its fitness by removing "bad facts" and leading to new, smoother narratives.

In religious contexts, often "mystic" movements are an attempt to recover the original "mystery" or animating ambiguity. Inevitably these movements aren't very large, because they deal with a lot of complexity – something most simply are unwilling to do.

A good example of a religious anti-meme/supermeme is the angel. The Bible is among the most well known and read texts in human history. Angels are a key "mystery" of the Bible, yet very few people are aware of the actual description of angels and their anthropological context. This brief entertaining video describes some of the details and historical context:

The profoundly strange idea of non-human entities is extremely old, and it has been persistently "slippery" in the sense of the anti-meme. A spinning wheel of eyes, or a floating human-animal chimera, is "smoothed" over time into a charismatic baby or a beautiful woman.

Everyone knows the super-meme of the charismatic angel. Meanwhile, the anti-meme of the original angelic figure is nearly deleted from cultural memory despite appearing in one of the most well known and studied texts in history. Though not "deleted" entirely, it hides in plain sight.

Emphatically, this is not a matter of conspiracy. I don't think anyone has endeavored to "hide" the idea of angels in a purposeful way. Instead, the ambiguous but compelling idea behind them has been "smoothed," and mutated into something simple and pleasing. In the process, the original idea has arguably been entirely lost and replaced with something conceptually adjacent to it. In other words, mass behavior and cognitive psychology engenders a dynamic wherein we have "hidden" the original textual description of angels from ourselves because the information is so strange it is difficult to recall and therefore to tell stories about.

A good experiment might be this. Set a reminder for yourself to write down a description of the angelic beings you saw in the video a month or so from now. How much did you remember? Did you remember all of them, or did you forget or mutate the details of some? When I talked to my wife about this idea, I only remembered the wheeled beings (because they are the strangest and most disquieting to me personally) and completely forgot the description of the cherubim.


In our current era of conspiracy theories and collective madness, it is more important than ever to be conscious of pseudoknowledge and of both natural and engineered anti-memes.

At their most pernicious, they can become psychosocial Navon figures – ideas that make us stupid and less capable. As Frank Hebert sagely wrote: “Knowing where the trap is—that's the first step in evading it.”

I think the trap is largely in certainty – especially certainty in the absence of data.