@intpdreamer replied to this post:
To not even go as far as ask for a definition of *root* insecurity, how would you define insecurity? Is it anything that causes fear? Anything that causes self-doubt? (In which case - does the degree of self doubt have to be inappropriate to be deemed an “insecurity”?)
I actually have been intending to write a definition post for "insecurity", because after understanding my own insecurities (and to a lesser extent others' insecurities) more and more, I recently realized no definition or explanation I have ever been exposed to actually explained it right, in a way that properly focused on the cognition mechanisms of it, the usefulness of it, or the empirically-learned by the brain logical-ness of it.
Insecurity is what happens when a mind has learned to predict that it will be hurt in some manner, even if only by the chronic absence of positive experience.
Insecurity is what happens when within the futures the mind is predicting, the good-enough outcomes have too little total probability mass/density.
Lucky you everyone, you just triggered a very important rant!
No, this needs big words, that's how important this is:
Step 1 - Prediction Functions
Remember that the human brain is in large part a prediction machine - that is one of its main functions and purposes, and it is subconsciously predicting all the time.
At any given moment, our brain is running a bunch of "prediction functions" with all of our current raw experiences and mental state as inputs (sensory data, internally maintained world model, memories, emotions, and so on), and those prediction functions spit out what experiences will come next, which update the world model and are themselves experiences and thus get fed into more prediction functions, over and over recursively, until the brain runs out of relevant and habituated prediction functions.
These prediction functions are instantaneous by human standards: when new information comes in, they happen faster than we even consciously notice that new information.
Technical detail below; feel free to scroll to step 2.
Prediction functions habituate thusly:
Neurons are regularly growing new connections, this works even if the growth is purely random but there might be any number of evolved heuristic optimizations to make it grow faster.
Neurons are regularly discarding connections whose firings are not reinforced or maybe even "deinforced", and reinforcing those that are reinforced.
How are they reinforced or deinforced? With brain chemicals various parts if the brain squirt around of course, but in response to what? Well:
The brain is always pattern matching different cognition together. A couple obvious sources to compare the result of any given ripple of fired neurons connections would be our raw sensory organ data and our slow conscious thinking. Two more will be described shortly. There may be others.
The pattern-matching wetware for each of these comparisons is always comparing its inputs and squirting out chemicals signaling good or bad pattern-matches, which get circulated around. This concept is enough to work, but you can devise optimizations, and evolution may have already implemented a bunch of them.
The prediction functions are pattern-matched with each other. This helps new prediction functions get developed and reinforced faster, even if they cannot share the same wiring with the older ones, because of too much logical difference or because they don't fully cover the same cases.
Stronger reinforcements due to trauma or other intense experiences are possible, causing neurons to retain and keep reinforcing connections which would normally get pruned out. This is probably done by certain brain parts being responsible for releasing the right type or amount or combination of brain chemicals or otherwise signaling in response to specific severe-enough signals, and then either neurons directly responding to that by treating that as a vastly stronger reinforcement for those connections that matched up with that event the right way, or probably a more advanced system where traumatic memories are stored redundantly or differently in brain parts which themselves are pattern-matching wetware for reinforcing prediction functions.
Note that this means that at any moment there may be any number of "ephemeral" or "nascent" prediction functions "implemented" by the brain, many of which are nonsensical or wrong, and they will be kept or culled as they empirically prove themselves accordingly, but also that prediction functions can get kept even if they were only correct in our earlier specific circumstances, and that older prediction functions might be contributing to the reinforcement or deinforcement of new prediction functions.
So the brain optimistically generates new connections, lets them fire as they will, and the ones that pattern match raw sensory data or conscious slow cognition or maybe each other or other sources get rewarded and retained.
Step 2 - Prediction Pyramids
The result is one or more final predictions logically resting on what I initially called "prediction pyramids".
One pyramid for each final prediction, where the ground at the base is all the inputs, and the root point is a prediction.
Each layer of the pyramid represents all the prediction functions which had the opportunity to execute at the same time - if a prediction depends on the result of a previous prediction, then the pyramid is "higher".
The pyramids can overlap, of course - some prediction functions might cause more than one new prediction function to activate, and so on.
When the predictions are mutually exclusive, that's just our mind seeing multiple possibilities, with how likely each one feels being determined by how strongly and thoroughly those prediction pyramids have been reinforced for similar situations before.
Before we even finish consciously processing what's just come in through our senses right this moment, the brain has already run through many prediction pyramids, at least to some significant height.
Technical detail below; feel free to scroll to step 3.
If we want to get more formal, we can think of them as "prediction trees", using tree in the mathematical or graph theory sense: each node is one habituated prediction function, the output of each node is a prediction, so the root is the final prediction.
Or if you want you can include the raw experiences in the tree and then those are the leaf nodes and the prediction functions are the non-leaf nodes.
If you want to complicate the picture further by representing
how the raw experiences may get fed into some or all of the prediction functions and not just the initial ones, and how prediction functions can in turn update the world model and thus new raw experiences, or
the overlap of multiple prediction trees using the same prediction functions,
then it turns into a directed acyclic graph (DAG).
We can keep going, because in practice it's kinda like a partially cyclic graph which just has some nodes that do not participate in cycles. A directed partially-acyclic graph, or DPAG, if you will.
Keep going with this long enough and eventually you're just representing an actual modern neural net implementation, or something like one.
For practical introspection, just the idea of prediction trees or pyramids is most of the value, and usually good enough.
There is also a complication I have been skipping: maybe some prediction functions share neurons. So besides data dependencies there is also a possible bottleneck in that case, and reinforcement of prediction functions actually doubles as increasing or decreasing the priority - maybe a neuron has some way of being conditioned to respond more readily to one signal versus another.
Or maybe this is simply prevented at a higher level - even if we imagine just very simple neurons that can only decide to fire on all output synapses or not on any, in responce to a strong enough input on one or more of its inout synapses, with only one neurotransmitter available, even then overlapping prediction functions would simply either misfire too often and get disenforced, or else prove accurate enough and maybe only cause occasional subtle errors.
And remember the idea about prediction functions being pattern-matched against each other?
Notice how these ideas combine: if you have two overlapping prediction functions, and that is the bottleneck in critical situations, and you need to shave off those tiny fractions of a second, then simply keep practicing at that edge or your performance envelope, and gradually a new copy or almost-copy of that original prediction function will grow and get reinforced which doesn't overlap with one that has to fire at the same time.
Basically, the height of your prediction pyramid/tree/DAG/DPAG is more formally determined not just by data dependencies but also refractory periods and any activation at the same time on overlapped neurons.
Step 3 - Prediction Substitution
Of course the brain also heuristically "cheats" to optimize this to be faster and take less energy, which is why this is able to stay fast no matter how tall the prediction pyramid gets.
If a second prediction is reinforced enough for immediately following an earlier one (that is to say if prediction one fires on some input, and prediction two then fires right after because of the result of prediction one), then the brain will simply eventually grow a prediction function that produces the second prediction directly for the initial input.
If two different prediction functions or pyramids are too consistent with their results, without in some situations producing enough different but still valuable/reinforced results, one will eventually be removed.
In other words, the brain will always try to flatten and simplify the prediction pyramids, eliminate all that nuanced reasoning in between. (If we want to resist this, we have to make a point to think and maybe even regularly put ourselves into situations in ways that keep discerning the difference.)
This is why introspection past a certain point, and extracting the logic from inside our intuitions, is part reverse-engineering and part historical inference - because the brain will readily optimize out the original reasoning/justification/"prediction pyramid" behind a given reaction!
The brain will substitute simpler or shorter prediction pyramids for complex or tall ones every chance it gets.
Step 4 - Emotional Reaction
Some parts of the brain are constantly streaming in this prediction data (or some reduced form of it as a heuristic optimization).
They react by activating some physiological process or releases some chemicals (neurotransmitters, hormones, whatever) into the brain accordingly, which we in turn consciously perceive as the often "immediate" emotional reaction.
Remember, all of the above happens basically pre-consciously, basically instantaneously by human conscious thinking speeds.
But why a reaction to a prediction? We can have raw reactions to raw experiences like pain and pleasure, that's obvious. But a prediction is this abstract idea thing, isn't it?
No, because what is the mind predicting? What is it a prediction of? What is the "language" or "format" of a raw prediction in a mind? Raw experiences.
Up at the start of the rant I said that the prediction functions produce "what experiences will come next", and this is a key detail.
At the level of these prediction functions, every single prediction the brain makes represents a raw experience, some of which include pain or pleasure or are otherwise experientially positive or negative, and our brain is ultimately reacting to that.
This, this is what I keep talking about when I mention "raw experience prediction analysis". The reverse-engineering of the pyramids of prediction functions in our mind, many of which have been substituted out for simpler heuristics over the years, to finally figure out why we do what we do, react how we react, feel how we feel, think how we think, and want what we want.
The realization that everything, everything about what motivates our minds, and maybe even all possible minds, can be understood as reactions to one or more interlinked chains of predictions of raw experience - that, that is the "raw experience prediction epiphany".
We apply that enough, while being ready to look at ourselves as unflatteringly as needed to do so, and eventually, the view we get of our mind is profoundly explanatory, comprehensive, predictive, and empowering as a result.
There are some things about how minds work that this doesn't cover, but everything it doesn't cover is far easier to satisfactorily explain with much simpler ideas. And the more time passes the more I find things I initially thought I understood well enough without this, only to find that this enriches it.
So anyway, "insecurity"...
... should make more sense now.
Our mind is constantly predicting raw experiences, based on what it has empirically learned from past experiences.
Our mind reacts negatively whenever it predicts it's experiences will be sufficiently preference-pleasure negative.
An insecurity is whenever our mind has learned to predict that its experiences will be below that threshold. So insecurity can be anywhere from intensely immediate to lightly gnawing about the indefinite future; anywhere from extremely specific to inscrutably general. The essential part is the below that threshold, a certain negative reaction, a compulsion to prevent that outcome, starts to kick it. It's a gradient, usually to some degree proportional to how severely below the threshold it is, but also importantly, how much our brain believes the outcome could be better if only we knew could do something about it.
But the actual follow-up beyond that point, including what emotional reactions are felt, is itself dependent on the prediction functions and other habituated cognition. Some examples different people might experience include:
Desperate or needy or manipulative behavior to try to get a wanted/needed thing.
Rage and hostility and violence because it helped them prevent hurting or to get their way before.
Eager sexual arousal because that helped stop hurt or got them safety and approval before.
Some other behavior which modifies the situation to temporarily soothe or mask a want or need with other experiences.
Going non-verbal or freezing up because their brain literally cannot think of any action which would be predictive of improving the situation.
Physiological activation of the body as in fight-or-flight just-in-case, without any compelled action.
Generalized anxiety, a weaker form of the previous point.
Except an insecure mind is prevented, by the same life strokes that cause and sustain the insecurity, from developing particularly reliable ways of avoiding those negative outcomes - if they had those, they wouldn't be insecure.
So often these reactions are not actually constructive to the goal, because the person simply doesn't know how to do that, or were constructive but only under particular circumstances (often less healthy ones, like abusive relationships, etc).
But just as often, the reactions are very constructive, sometimes even perfectly executed, skilled preventions or mitigations of the predicted negative outcomes, and otherwise totally justifiable independent of the insecurity.
The essence of insecurity, and it's literally in the name but I spent years missing it, is the lack of feeling secure, assured, confident, certain (about our raw experience being good enough in the future).
In saying that an insecurity is whenever our mind predicts that its experiences will be below some threshold, and feels negatively compelled to change that, I deliberately left unspecified whether or not it is objectively sound for the brain to make that prediction, or whether or not the fundamental compulsion to avoid the outcome is adaptive, or whether or not that threshold is in some way at the right level.
Those are all important questions, and touch on the essential point that insecurity isn't just this defect that some people have, but rather an essential ingredient of cognition for most people to some degree. But they are otherwise irrelevant to describing the essence of what an insecurity is.
