3 ms·

Rationality is about internal consistency of beliefs and values. "Rationalism is inherently flawed in that it lacks a value system" strikes me as a category er

Rationality is about internal consistency of beliefs and values. "Rationalism is inherently flawed in that it lacks a value system" strikes me as a category error. Giving you a value system isn't its job; instead, it provides a toolbox for identifying and resolving contradictions in your preexisting beliefs and values. (Since we all work with imperfect information about the world, internal consistency is far from a guarantee of accuracy; but it improves the odds.)

sudosysgen 6y ago
There are two issues here. First of all, not all value systems are rational and optimizable. The most relevant ones to humans simply aren't. Building a rational system to maximize these value systems is an impossible task, it's a category error. Like trying to run regression with a non-differentiable cost function. Value systems are an "ought". No rational, logical system can ever change them. Resolving contradictions in your values is a categorical error, because there is no such thing as having more than one value means inherent contradiction. As a result, a purely rationalist perspective is ineffective. You can only achieve a certain level of rationality. Which is already exactly what people have been doing for millenia, trying to find the most logical way to follow their values, and trying to phrase their inherent, base values into a set of explicit values that represent them as well as possible. It's not anything different from what secular philosophy has done for over 4000 years now if you construe it that way. Finally, it's never possible to know that your system is without any contradiction, because that would be equivalent to solving the halting problem. So, do you have a notion that contradiction can be estimated? Measured? How do you come to the conclusion that your belief system, with that contradiction removed, is any less contradictory than before, knowing full well that you might have introduced a more complex, as of yet unknown contradiction, or ten thousand of them? And so it seems that having soft contradictions is not actually necessarily a bad thing, and that jumping to resolve them should only be done if they actually are terminal, because balancing contradictions in your cognitive process is actually very doable and a good tool. It's not just about information either, even with perfect knowledge it's impossible to be sure of the absence of contradiction. In other words, it's never possible to prove a the consistency of a model in that model itself. You can try to prove it using the framework of another model, but how do you know that that model is consistent? You can't. This is why I personally try to ground the internal consistency of my though process in the material world - a contradiction is only problematic if resolving it can materially improve the situation or understanding I derive from it. But then again, Philosophical materialism vs idealism is a whole other can of worms, so il leave it to that very small facet of it.
- chrchang523 6y ago
  It is not necessary to resolve all contradictions in one's beliefs and values to gain some benefit from rational introspection. Even professional physicists haven't arrived at a consensus on how general relativity and quantum mechanics fit together; that doesn't make the internal consistency of those two models worthless.
  - sudosysgen 6y ago
    Yes, and that is a completely useless remark that is the basis of Western philosophy since 4000 years. This isn't what "lesswrong rationalism" is about, that is about frantically trying to eliminate as many contradictions as possible even if it means being less effective at actually optimizing the values, whether it is by eliminating flexibility, increasing the likelihood of error, or abstracting away the contradictions to such a level that they can't be consciously dealt with anymore. As you've missed in my first post, the issue is that personal philosophies are fundamentally different from physics, in that in physics there is a correct model, that can be reached by eliminating contradictions and constantly iterating. As far as models to maximize personal success, however that is, there is no such thing as a correct model, there will always be contradictions, and hyperfocusing on the most rational possible model to optimize for an irrational goal will almost always yield worse results. Rationalism doesn't just mean "we can derive some benefit by carefully trying to use rationality to arrive to better models and decision", the same way that scientism doesn't just mean "we can derive some benefit by using science", rationalism is the idea that *all beliefs and decisions should be based purely on rationality". Unless your goal is purely quantifiable and free of contradictions, this is actively harmful. I never once suggested that rationality isn't useful or that resolving some contradictions is useless, I actively stated the reverse, that they are both useful insofar that they actually make the situation better for you, which is absolutely not what rationalism is about. Rationalism presupposes the supremacy of rationality, which itself cannot be rationally proved.
    - chrchang523 6y ago
      Any toolbox can be misused. Anything can be made into an idol. The fact remains that the "lesswrong rationalism" community was the first to popularize several rather sophisticated reasoning concepts. They have provided common reference points that have made it easier for me to communicate my thinking to others and vice versa, and I am far from the only writer/reader who has benefited in this way. As a consequence, I'm generally sympathetic to this group, and personally identify as a rationalist. I agree that the heuristic "internal consistency is always worth actively working on" can go very wrong. It looks like you're defining rationalism as structuring of one's life around that heuristic; I'm guessing you must have had bad experiences with people who actually do that.

sudosysgen 6y ago

There are two issues here. First of all, not all value systems are rational and optimizable. The most relevant ones to humans simply aren't. Building a rational system to maximize these value systems is an impossible task, it's a category error. Like trying to run regression with a non-differentiable cost function. Value systems are an "ought". No rational, logical system can ever change them. Resolving contradictions in your values is a categorical error, because there is no such thing as having more than one value means inherent contradiction. As a result, a purely rationalist perspective is ineffective. You can only achieve a certain level of rationality. Which is already exactly what people have been doing for millenia, trying to find the most logical way to follow their values, and trying to phrase their inherent, base values into a set of explicit values that represent them as well as possible. It's not anything different from what secular philosophy has done for over 4000 years now if you construe it that way. Finally, it's never possible to know that your system is without any contradiction, because that would be equivalent to solving the halting problem. So, do you have a notion that contradiction can be estimated? Measured? How do you come to the conclusion that your belief system, with that contradiction removed, is any less contradictory than before, knowing full well that you might have introduced a more complex, as of yet unknown contradiction, or ten thousand of them? And so it seems that having soft contradictions is not actually necessarily a bad thing, and that jumping to resolve them should only be done if they actually are terminal, because balancing contradictions in your cognitive process is actually very doable and a good tool. It's not just about information either, even with perfect knowledge it's impossible to be sure of the absence of contradiction. In other words, it's never possible to prove a the consistency of a model in that model itself. You can try to prove it using the framework of another model, but how do you know that that model is consistent? You can't. This is why I personally try to ground the internal consistency of my though process in the material world - a contradiction is only problematic if resolving it can materially improve the situation or understanding I derive from it. But then again, Philosophical materialism vs idealism is a whole other can of worms, so il leave it to that very small facet of it.

chrchang523 6y ago
It is not necessary to resolve all contradictions in one's beliefs and values to gain some benefit from rational introspection. Even professional physicists haven't arrived at a consensus on how general relativity and quantum mechanics fit together; that doesn't make the internal consistency of those two models worthless.
- sudosysgen 6y ago
  Yes, and that is a completely useless remark that is the basis of Western philosophy since 4000 years. This isn't what "lesswrong rationalism" is about, that is about frantically trying to eliminate as many contradictions as possible even if it means being less effective at actually optimizing the values, whether it is by eliminating flexibility, increasing the likelihood of error, or abstracting away the contradictions to such a level that they can't be consciously dealt with anymore. As you've missed in my first post, the issue is that personal philosophies are fundamentally different from physics, in that in physics there is a correct model, that can be reached by eliminating contradictions and constantly iterating. As far as models to maximize personal success, however that is, there is no such thing as a correct model, there will always be contradictions, and hyperfocusing on the most rational possible model to optimize for an irrational goal will almost always yield worse results. Rationalism doesn't just mean "we can derive some benefit by carefully trying to use rationality to arrive to better models and decision", the same way that scientism doesn't just mean "we can derive some benefit by using science", rationalism is the idea that *all beliefs and decisions should be based purely on rationality". Unless your goal is purely quantifiable and free of contradictions, this is actively harmful. I never once suggested that rationality isn't useful or that resolving some contradictions is useless, I actively stated the reverse, that they are both useful insofar that they actually make the situation better for you, which is absolutely not what rationalism is about. Rationalism presupposes the supremacy of rationality, which itself cannot be rationally proved.
  - chrchang523 6y ago
    Any toolbox can be misused. Anything can be made into an idol. The fact remains that the "lesswrong rationalism" community was the first to popularize several rather sophisticated reasoning concepts. They have provided common reference points that have made it easier for me to communicate my thinking to others and vice versa, and I am far from the only writer/reader who has benefited in this way. As a consequence, I'm generally sympathetic to this group, and personally identify as a rationalist. I agree that the heuristic "internal consistency is always worth actively working on" can go very wrong. It looks like you're defining rationalism as structuring of one's life around that heuristic; I'm guessing you must have had bad experiences with people who actually do that.