4 ms·

A couple years ago I heard an interview with someone that has stuck with me. He said that human “reasoning” didn’t evolve to reason in the logical sense, but to

A couple years ago I heard an interview with someone that has stuck with me. He said that human “reasoning” didn’t evolve to reason in the logical sense, but to _provide reasons_ likely to be accepted by other humans, allowing better survival by manipulating other humans. This matches my perception of most people’s reasoning. What’s funny is that AI is now being trained by a human accepting or rejecting its answers, probably not on the basis of the rigor of the answer since the temp worker hired to do it is probably not a logician, mathematician, or scientist. I suspect most people’s reasoning is closer to an LLM’s than we would be comfortable admitting.

AnimalMuppet 2y ago
If you believe Darwin, then human reasoning evolved to give a right enough answer fast enough for the individual to have higher survival odds in the most common situations that it ran into. Even persuading other humans is an extra, not the real "purpose".
- llamaimperative 2y ago
  If you believe Darwin, there's no distinction between primary, "extra", or "real" purposes. All of these forces are in play all the time.
kelseyfrog 2y ago
I'm of this camp. I believe we rationalize our decisions rather than decide rationally [1]. The two are so intertwined in our minds that we confuse even ourselves. Outside of specific rationality frameworks, we can't help but do it. The smarter someone is, the more likely they'll be able to better rationalize and more clearly articulate their rationalizations, so intelligence isn't even an escape hatch. 1. The vast vast majority of the time
randcraw 2y ago
Yeah, that kind of reasoning (abduction) tends to be included with deduction and induction as the third form of reasoning. But IMO, it's really not. It's essentially confabulation, or legalistic argument -- where you're not attempting to discover an outcome (entailment) but you're trying to convince others that a desired outcome is plausible because it's possible to get there by cherry-picking a line of reasons. If abduction has an Achilles Heel, it's that the sequence of reasons proposed is only ONE OF MANY possible chains of reasons that could have led to the current outcome. That tells you nothing about one possible chain's correctness and little about its likelihood to be correct (since you need to understand the degrees of freedom in each scenario to do that). That makes abduction more useful as a means to express personal bias and manipulate the beliefs of others than an attempt to fair-mindedly understand how something came to be. It's hardly surprising that LLMs employ abduction rather than deduction or induction. Abduction is really just storytelling, where you recount a similar tale that resembles the current scenario (and ends where you want it to). It doesn't require any ability to generalize similar events into precise logical rules based on instances that share the same dependencies, mechanism of action, and outcome. Only by creating such rules using induction can deductive reasoning then take place using them. But LLMs associate, they don't equate. To date, I see no path to that changing.

AnimalMuppet 2y ago

If you believe Darwin, then human reasoning evolved to give a right enough answer fast enough for the individual to have higher survival odds in the most common situations that it ran into. Even persuading other humans is an extra, not the real "purpose".

llamaimperative 2y ago
If you believe Darwin, there's no distinction between primary, "extra", or "real" purposes. All of these forces are in play all the time.

kelseyfrog 2y ago

I'm of this camp. I believe we rationalize our decisions rather than decide rationally [1]. The two are so intertwined in our minds that we confuse even ourselves. Outside of specific rationality frameworks, we can't help but do it. The smarter someone is, the more likely they'll be able to better rationalize and more clearly articulate their rationalizations, so intelligence isn't even an escape hatch. 1. The vast vast majority of the time

randcraw 2y ago

Yeah, that kind of reasoning (abduction) tends to be included with deduction and induction as the third form of reasoning. But IMO, it's really not. It's essentially confabulation, or legalistic argument -- where you're not attempting to discover an outcome (entailment) but you're trying to convince others that a desired outcome is plausible because it's possible to get there by cherry-picking a line of reasons. If abduction has an Achilles Heel, it's that the sequence of reasons proposed is only ONE OF MANY possible chains of reasons that could have led to the current outcome. That tells you nothing about one possible chain's correctness and little about its likelihood to be correct (since you need to understand the degrees of freedom in each scenario to do that). That makes abduction more useful as a means to express personal bias and manipulate the beliefs of others than an attempt to fair-mindedly understand how something came to be. It's hardly surprising that LLMs employ abduction rather than deduction or induction. Abduction is really just storytelling, where you recount a similar tale that resembles the current scenario (and ends where you want it to). It doesn't require any ability to generalize similar events into precise logical rules based on instances that share the same dependencies, mechanism of action, and outcome. Only by creating such rules using induction can deductive reasoning then take place using them. But LLMs associate, they don't equate. To date, I see no path to that changing.