Technology as the Crystallization of Rationality

Rationality with a hard R

note: this is a repost of an essay I posted October 2022, but decided to rewrite with the help of CJ The X. The heart of it remains the same, but I’ve added some new and better perspectives to it, including indulging myself in some Descartes slander. enjoy!

In order to understand the ethical issues of the tech industry, we have to look at the mechanisms that precede it. That is, the «rationality» of Rationality itself.

It should not be difficult to argue against something that «makes sense» if it feels wrong, but it is, and that’s the core issue I have with technology and the Western episteme at large.

After burrowing deep into tech criticism and philosophy, I have noticed that this form of rationale–namely, that objective «logical truths» overrule human intuition at every turn–is indistinguishable from the very mechanics that technologies consist of. 

There are two things I wish to outline here: the allure of this particular form of logical coherence, and the dangerous, invisible leap made from logical coherence to truth and morality. Fixing this is crucial in our quest to make humble, wholesome, beautiful technologies.

seductive clarity

As software developers, we like to believe that we are fully in charge. The programmer decides what the application’s objective is and exactly how it handles this. But how many compromises does she have to make along the way in order for the program to run? There isn’t much space left for nuanced, dynamic, subjective human values when the rules of coherence are literally computationally non-negotiable. The machine has strict instructions to follow and will not deviate under any circumstances, like any good foot soldier. 

In an attempt to make up for this compromise, we naively assume that a vast amount of data will make up for any shortcomings in our algorithms. Start off with a simple idea, then with enough data collected over time, it will add up to the nuance and complexity that was lost upfront.

It's not a lack of data that is the issue here. You can have trillions of data points about a customer, but even if these are parsed in a way that «makes sense», you are still not getting a complete picture of any of those people. In fact, you're not getting any individual person, only the mode they slip into whenever they need to buy stuff.

This is illustrative of the problems baked into the field. Sense is defined narrowly, but used broadly, achievable only through yes-no computations, and the human element is treated as an obstacle to overcome.

C. Thi Nguyen observes in his paper «The Seductions of Clarity»

¹

that we favor clarity and coherence above uncertainty because it's easier to grasp. We are cognitively limited beings. We cannot know everything all the time. And so we need to have some sort of rule of thumb that helps us come to rest, ease our anxious overanalysis, and rely on our most useful tools to navigate reality. 

Unfortunately, this convenience comes with a price. We lose nuance and complexity along the way, which is what makes our experiences so rich and full of depth in the first place. This seductive clarity makes us lose sight of this being lost.

When this tool of «sense-making» is used against us, it can hijack the affinity we have for clarity, taking us farther away from our actual goals and values. 

Offloading the responsibility of truth, value, and morality to various technologies implicitly permits the «sense» of Silicon Valley to dictate the values of the world. We have become emotionally invested in the idea of their neutrality, exactly because they give us the ease and convenience we sorely need in our otherwise hectic lives.

But this is exactly what makes them so dangerous. 

creating sense

Rationality is a method by which we organize knowledge. It is the ability to parse through information and make sense of things, and I mean this quite literally: it is the act of making sense. 

Rationalizing can be understood as the process of weaving ideas together to create and recognize useful patterns. When the connections between these ideas are harmonious and consonant, the ideas create «sense». This «sense» of clarity motivates us to consider the rationalization «true» and therefore worthy to be relied upon as a tool to achieve consistent and useful results. 

This sense of clarity relies on the premise a rationalization is based on. Premises act like the laws of physics. They shape everything they come in contact with. For instance, let’s say you have a simple idea:

The kitchen light is on. 

The premises of this statement inherently deny this idea:

The kitchen light is off.

This is the first thing you learn in any Logic 101 class: what logically follows from a premise? If you’re sitting on my left, then I’m sitting on your right.  All cows are female

²

. All cars are boys. Stuff like that. 

The logical process itself is crucial. Being rational is not only about reaching a desired end state, but also about how you arrived at this end state. Like solving any math equation, we are interested in getting to the answer under these specific logical constraints. Your teacher doesn’t care if you had a divine intervention in the middle of your 6th grade algebra test. You still gotta show how you got the correct answer, Kevin. 

Computation is the outsourcing of this logical process. This mechanism that we use to organize our thoughts is the very same one whirring inside all machines. The code is a set of instructions outlining premises, and what would logically follow from them. 

Rational coherence in computing (this «making sense») is the difference between a program that runs smoothly and one that crashes. When coding, the exact second your code stops making sense, you will get an error. When you create a contradiction or a statement that contradicts the premises at hand, the code won’t run. 

There is also something called semantic errors, where the code runs but isn’t doing what you expected it to. On the surface this might seem like an example of computational sensemaking not cohering with our ends. Still, these semantic errors are still premised on the cohesion of universal logical sequences, unlike intuitive human truths, which already have a diversified set of premises and other considerations they operate on. More on this later.

What becomes lost amidst all this is the fact that motivation for this entire endeavor has to be anchored. Kevin isn’t solving equations because equations should be solved. A reason for rationality cannot be found within its own scope. If we try to find rational reasons to be rational, it will lead to a slew of contradictions

³

. Like a black hole, it will collapse in on itself, swallowing itself whole.

Our desire to be rational then, must be rooted in something else. It cannot survive on its own. 

We’ve uncritically lauded coherence so far above any other considerations that we lose sight of this crucial fact. This however isn’t anything new, nor is it a mistake only made by techies. This can be traced back, as many such things can, to a silly French man. 

Descartes the war criminal
Early century philosopher René Descartes saw the fickle nature of what we so often take for granted as «objective truth». Most truth claims have valid counterarguments. Our senses can play tricks on us, as any false memory or optical illusion demonstrates easily. What objective truths can we really be sure of? 

This quickly developed into him doubting his very existence. 

He generated a thought experiment to elucidate this: What if all of this: reality, existence, my very free will, isn’t real? What if a powerful evil demon is tricking me to believe that all of this is real? After much pondering of the orb, in answer and salve to this existential dread, he came up with the following:

The very capacity for thought, the ability to come up with such a hypothetical scenario at all, reveals that if anything, thought itself provides proof of existence. This is referred to as the Cogito: I think therefore I am.

While we can doubt everything else, this one primary truth cannot be doubted because it affirms itself. Descartes believed we can use self-affirming truths as a foundation to build all our other truths upon.

These self-affirming truths are allegedly irrefutable. They snap shut like a circle; they can be tossed into any context while keeping their truth intact no matter what. 2+2 equals 4 even if the sky is falling while raining cats and dogs.

Though if we step out of these circles, we quickly realize their limitations. It is something so painfully simple and obvious, but because of its simplicity, deeply elusive and thus difficult to realize: Just because something makes sense, doesn’t make it true. 

isolated logic

«Sense» is not interchangeable with truth, value, or morality. 

Bigotry is a great example of this. A common misconception of any -ism is that it's nonsensical, but bigoted notions are perfectly coherent most of the time. Fascist and racist ideologies follow the laws of logic just fine. The premises get along great. And yet they lead to death, violence, and suffering. 

The problem isn’t that they don’t make sense. Racism isn’t wrong because of rational reasons.

The horrible feeling deep in your belly when encountering immorality, the repulsion caused by something trying to appeal to you rationally which you know to be wrong, reveals that something simply making sense is not enough.

As it stands, we have many modes of consideration available that are equally valid to rationality. We are able to think of things intuitively, associatively, imaginatively. Yet, this cartesian form of rationality has completely undermined them. 

Rationalizations only operate within their own internal logic and the premises they consist of. They intentionally silo themselves away from other considerations that might be of equal value and relevance: relationships with other existing things, the context it resides in, and their attached history, the intentions and purpose behind the question, and so on. This simplified context is part of their appeal, but the fact that cartesian rationalizations can only get along with their own premises is also where their limitations lie. The isolated coherence can’t withstand the fluctuating complexity of the world in which it resides.

Nice that your little thought experiment proves that a demon isn’t controlling you. But what compelled you to believe that in the first place? Does it make a difference if there was a demon controlling your reality? Wouldn’t this demon-controlled world functionally be your reality now, not really changing anything significant about your life at all? What was the point of this?

Yes, your code runs smoothly and without errors, but how well does it fit the world? Does it resonate with its environment? What need is it serving, and is this holistically addressed? How many resources were used to bring it about? What is it obfuscating in the process? 

Yes, 2+2 equals four, but who cares about that if the sky is falling?

Questions like this might be received by rational actors as «What About-isms», deliberately complicating perfectly coherent and reasonable objective thoughts. This frustrating overcomplication is famously looked down upon in ethical and political discourse, where nuance is often received negatively. Try to answer a question with “Well, it depends”  and you are seen as avoidant and deliberately difficult. Just answer the question! 

[The western episteme] maintains a self-perpetuating and self-referencing closed belief system that often prevents them from seeing or noticing the process of recursion. Tiffany Lethabo King, Humans Involved

In Western philosophy, once a notion is established, it is excruciatingly difficult to refute it, exactly because it's so insular, «self-affirming.»

As this self-affirming recursion is crystallized in technologies that we use daily, it becomes increasingly difficult to question their initial premises. Negative fallout is deemed as necessary collateral to irrefutable rationality. After all, the math checks out.

Buying into the irrefutability of rationality as «objective» makes us susceptible to harmful rational notions. These do the most harm when their mechanics are invisible to us, more successfully seducing us to their evil ways.

This is deliberate, and the entire appeal of objective rationalism from the capitalist perspective. 

Here is a justified, valueless way to ignore everything that isn't in your direct interest, as irrational.

towards a symbiotic sensemaking 

No code of ethics and no effort is justifiable in the face of the cruel mathematics which command our attention.  -Albert Camus, the myth of Sisyphus

Rationalization isn’t the only way we handle knowledge. By diversifying our methods, we access new perspectives, add more depth to our knowledge, and increase the tools available to reach for dynamic ideals like goodness, beauty, and morality. We can’t keep favoring a form of rationality that undermines our other methods of consideration for not acting as it would.

I believe if we were to naturally follow what is valuable, what is good, and what is moral, it will lead us away from some of these harmful cartesian self affirming circles.

This does not mean abandoning sensemaking altogether, but rather having all our modes of thought collaborate and rely on each other instead of undermining each other. I imagine the coherence and sense they would make together would be much more beautiful and compelling.

This might not make much sense, but don't worry, that's this selfish parasitic sense trying to wry away from the sun. 

Sometimes we need to stop making sense in order to really make sense. 

1. Here is a link to Nguyen’s paper. If you are not fond of reading long philosophical papers there is a video lecture available outlining these ideas.

2. because cow is the name for female cattle, all cows have to be female in order to not make a contradiction.

3. In why be rational, Niko Kolodny outlines the normative nature of sensemaking. Our spawn point is always subjective, and we don’t always follow the logical consequences of our ideas; only if we are already invested in them. Kolodny says:

When we say ‘you ought to’ in the sense of ‘you have reason to,’ we usually seem to be saying something about the relation between your situation and your attitudes. When we say ‘you ought to’ in the sense of ‘it would be irrational of you not to,’ we seem to be saying something about the relation between your attitudes, viewed in abstraction from the reasons for them. We are saying something, for example, about whether your beliefs are logically consistent, or whether your intentions for ends cohere with your intentions for means—things that are true, if they are, quite independently of whether there is reason for you to have any of those beliefs or intentions.