How a mind can become aware of a “nothing”
Explaining the apparent paradox in positively affirming a negative
By Yervant Kulbashian. You can support me on Patreon here.
From being we can never derive negation. The necessary condition for the possibility of saying “not” is that non-being should be constantly present within us and outside us, that nothingness should haunt being — J. P. Sartre, Being and Nothingness
Every sensory experience you have is of something. You can’t perceive a negative, an absence, or a lack, at least not through your senses. There are an infinite number of things that are not present at any given time; the presence of a cat is also the absence of an airplane, or computer, and so on. Visual stimuli from the world can perhaps indicate when you see the cat, but there are no sensory inputs that correlate to when a cat is not present. Nor would it make sense for your brain to generate a “no cat” signal — and a separate signal for every other absent thing — at every moment when they are not present.
The only way you can know that something is not there is if you first go looking for it, and are disappointed. When listening to a song and you notice a note was dropped, it was because the preceding notes made you expect to hear it. A negative is therefore created internally, as an expectation or request instantiated by circumstance, and is prior to confirming a presence. As Sartre pointed out negation (including in concepts like destruction, removal, incompleteness, etc.) is only created by the human mind. It exists inside you and arises because you momentarily wanted something. The English term “want”¹ even encapsulates both these senses — it is both a “desire” and an “absence”.
We can follow this same reasoning to also explain every positive act of recognition. Looking at the image above you might identify it in a number of ways based on context: it is an image, an elephant, an animal, a patch of grey or yellow, a collection of pixels, a savannah scene, wildlife, digital art, an AI-generated image, etc. There are an infinite number of “things” you can broadly become aware of or interpret the same stimuli as. Whichever of these you pay attention to — i.e. whichever you go looking for — is what you capture. Becoming aware in this positive sense is therefore the fulfilment of some searching need, the moment when you consciously recognize that something you were looking for has been found. And it only happens because you were somehow oriented towards finding it.
Even an act of affirmation — saying “that is true” — always involves an original query to which you expected a response. There are an infinite number of assertions you can decide are true at any given time — e.g. I am in a room, 12 x 12 = 144, blue is a colour, and so on. Nor does having a random thought in your mind automatically mean you consider it to be true — an additional confirming action is necessary. To select which of your millions of possible thoughts you affirm, there must first be a need, a pressing reason why this request was made, and why it was made now. You may not be aware of what that reason is — and we’ll show why later. Nevertheless you can only identify something as true if you first go looking to confirm or deny it. The process of recognizing a truth is in this sense similar to becoming aware of the presence of an object like a cat, or spotting a cup of water when you are thirsty. They are all triggered by an absence or need.
In short, a negative always precedes a positive. A negative is created internally because you are looking for something, and then gets filled in by whatever you are looking for.
However, this model is still missing something. It suggests that your mind can only become aware of positive things — e.g. once you have found the cat. How, then, can you positively know that something doesn’t exist. How could you ever assert “the cat isn’t there”?
One possibility is that being aware that you can’t find something — that it is absent — is also a type of resolution to a search. In this case you are trying to express a failure. Your mind might decide to frame the negative (failure) as a positive assertion whenever it wants to communicate it in some way, such as to assert its absence to other people, or just in your own mind. So you momentarily switch from searching for the entity, to proclaiming that negative outcome.
This behaviour seems odd, though. It implies there could be three conflicting resolutions to the same initial tension. For example, if you are searching for a cat the first resolution is that you find the cat, the second is to tell others that the cat is missing, and the third is to simply give up finding the cat — “this cat can’t be found”. How could the same tension be resolved by all three of these? The second case is perhaps easier to align with the first: telling other people you can’t find the cat might encourage them to assist you, as a sort of plea for help. But giving up on the search entirely seems irreconcilable with the first two.
The confusion goes away once we remember that every such search is based on a deeper underlying need. The tension that triggers the search for the cat is not necessarily a search for the cat itself. No human is born with a desire to search for cats. That need would have been rooted in another source, say, in boredom, in which case giving up may be a good option if the search takes too long. Or you may be searching for the cat because you are worried about its safety, in which case you might not give up, and your actions would become more frantic as you are unable to find it. We’ve seen how people whose children go missing often refuse to admit their child may be gone forever — for them that is not an acceptable answer to the underlying tension.
The relationship of a tension to its many contradictory resolutions can be tricky to wrap your mind around. To explain it more clearly let’s break down a concrete example.
Imagine you have built a software program which looks through a camera to control robots in an industrial setting. For safety, you want to ensure that no human enters the perimeter while the robots are active. So you add a motion detector connected to another piece of software that senses if someone has entered the space and triggers an alarm. Unfortunately, like all hardware, the motion sensor is noisy; it may signal movement when there is no one actually present, giving you false positives. So you will likely want the model you create to use the camera to look and confirm that someone has actually entered the space before the alarm is triggered.
Training this model will proceed through operant conditioning. Every time a person enters the forbidden space and the model fails to trigger an alarm, it receives a punishment, like a reprimand. Early on this happens every time, since the model has not yet learned the proper response. Such situations also always coincide with the motion detector firing. So the model learns that a motion detection signal usually predicts a punishing event.
Until now our model had no reason to even care about someone entering the space. Through punishment we are giving it a reason to care. Going forward the model we’ve designed will be “twitchy”, a sense of anxiety (or low-value state) arising whenever the motion detector goes off, even if there is nothing actually there. That anxiety merely represents a learned inner tension whose nature is that it should be removed — it demands that the agent do something to reliably protect it from the punishment it has learned to anticipate. In doing so we have set up the negative, the searching absence or lack. Each subsequent signal from the motion detector starts the search chain going.
The triggering of the motion detector is still entirely “unconscious”, or pre-reflective, in Sartre’s words. Awareness of the problem has not yet entered the picture; it has only gained a tension. To learn that something is wrong is a positive act, like all awareness. And that has yet to happen. Positive awareness comes only after this initial searching trigger is presented, as its resolution. The negative causes it to turn its attention and try to ascertain — i.e. positively confirm — what is there. It is like looking to see what just brushed your arm. “That’s what that was”, or “that’s what happened” are both such resolutions.
Unfortunately, the means by which its inner tension may be resolved is as yet unknown to the model. The situation is quite dire too: not only is it unaware of how to resolve the tension when it is triggered, it doesn’t even know what would count as a resolution. The motion detector signal by itself cannot teach it what the resolution should be. The source of truth for that resides somewhere outside it, in the trainer and their actions. The model is starting from a position of ignorance, and has so far only been given a new motivating tension. It must now learn to carve, out of the space of as yet undifferentiated inputs, some sense of how to avoid its punishment. It must be made conscious of what is happening.
While tense, and through random actions, the agent may or may not trigger an alarm. In addition, sometimes the punishment does not arise as was expected since there wasn’t actually anyone there. So a resolution to its problem is hiding somewhere in there which the model would do well to discover — “why didn’t I get punished that time?”. In such cases, the internal tension or anxiety would be unnecessary and can be suppressed. In yet other cases triggering the alarm itself causes punishment (i.e. if it triggers a false alarm). To learn what counts as a resolution, it must figure out how to distinguish between all these cases. The motion detector signal by itself does not give enough information to distinguish between them — as mentioned, it is unreliable. So it must discriminate based on additional inputs from the camera.
Given enough experiences with a pattern of punishments, or lack thereof, our agent can learn two tension-resolutions by which it suppresses its anxieties².
- In the first case, it learns that activating the alarm — conditioned on the fact that it detected visual movement — leads to a resolution. This is the equivalent of: “I sensed an intruder” a.k.a. asserting a positive.
- In the second case, a lack of visual movement leads to a resolution only if the alarm is not triggered. This is like confirming that no intruder is present a.k.a. asserting a negative³.
The initial anxiety caused by the motion detector can now be suppressed by two seemingly contradictory paths. But they are only contradictory if you forget that the goal was not to find or identify an intruder, but to avoid punishment. When considered from that higher-level perspective, they are perfectly aligned.
The important lesson of the above system is that in order to recognize both a negative (no intruder) and a positive (intruder), some initial tension must be active beforehand, as a sort of heightened alertness looking for a resolution; and both those responses are types of resolution. This is how we overcome the apparent contradiction in asserting a negative. Awareness of any form of negative — e.g. pain, destruction, absence, removal, even desire — is not awareness of a negative in itself, but a positive resolution to some deeper need, one which was not in itself looking for that entity.
There may be any number of other responses that the model in the example above can learn to resolve the parent tension, such as ejecting the intruder or calling the foreman, and these will all come under the umbrella of correctly addressing an intruder. The reason all of these can be encapsulated under one tension is because the tension itself was never “looking for” an intruder — it didn’t even know what that was. It simply learned to conceive of intruder as shaped by the feedback it received. The real motive and conditioning was outside the agent’s mind — it drove the behaviour of the trainer. And like a magnet shaping iron filings, the fear of punishment lead it to form a working, practical conception of intruder, including in all its positive and negative aspects⁴.
What we’ve described explains how a mind can seemingly create novel concepts without knowing what it is creating beforehand. Every difficult conception in human cognition can be explained by looking to the set of needs that the understanding has been used to address; needs for which the concept’s various facets became unexpected resolutions. Most importantly, positive awareness of anything, including one’s own self or mind, must arise from a negative need that prompted you to search for it. It would not be too much to say that the entirety of conscious awareness, both introspective and of the external world, arises from this same pattern of search and response.
¹ Like Sartre’s manqué.
² It will ignore any paths that lead to punishment as irrelevant; all it really cares about is avoiding the tension.
³ It is also possible that the motion detector does not fire when someone is present. In that case some other triggering source must be employed, even if it is an internal prompt like “perhaps there is a trespasser?” Without some motivating trigger no search can begin, and so no awareness can occur.
⁴ It has transferred the trainer’s structures of intruder to itself through reinforcement. Nor is a human trainer actually required: the world itself, devoid of other agents, can equally well train a mind through reward and punishment.
