Uncertainty is epistemologically required

Epistemologically, I stand firmly with the arch-empiricists and against the idealists Plato and Descartes. Epicurus and Locke start sane. Plato and Descartes, not so much.

Issue #1: What we have to work on is entirely inputs from our senses, and built-in systems in our brains. The existence of the external world is a conclusion. Logic itself is a deduced regularity from our senses.

How do we get patterns? We look at the world, group things into groups that seem to act the same...and then name the groups. Later, we try to identify characteristics that will let us learn stuff about the groups. The essential element of what we're doing is ignoring parts of the world in order to better understand other parts.

This immediately brings up the useful question: are we ignoring parts that matter? This is my favorite reason for uncertainty. 999/1000 cases, we don't have a good handle on what elements matter...and yet we're ignoring many of them.

Note: discussions about imaginary stuff where we define the rules of what matters
are exempted. Math, for instance. Discussions about super-simple topics where we know ALL of the relevant rules at a 0.001% accuracy or better also exempted. Physics at near-human scales, for instance. It's just all the other topics where we have problems, and should expect to be wrong.

Shortened version...what you pay attention to trumps your calculations. AKA he who asks the questions wins.

Issue #2: Follow math for long enough, and you reach: Necessary uncertainty. Chaotic and agent-based systems are the natural state of the world. Different initial conditions, even at millionths of a percent, make for real differences in end-states over a reasonable period of time. Many results require perfect precision, and/or infinite computation to get correct answers...and perfect precision is impossible.

Shorter: Answers are often incalculable.

Issue # 3: We normally operate under conditions of massively insufficient information. To flip a coin...and then calculate whether it's heads or tails only requires precise numbers on initial forces (thumb and hand-lift), air resistance, height from the ground, and coin/ground collision elasticity...right? Since we don't have all the information, we have to guess. The sane response academically is that there's a 50% chance of landing heads. Indeed...it's largely true that even ignoring issues #1 and #2, our answers to questions (again, outside boring contrived problems) always land in the realm of probability. Sane supposition is that the answer is always p<1, nearly always p<.9, and for contentious questions usually p<0.6.

Shorter: All statements have probabilistic truth-values

Issue #4: Monkeybrains. Follow observation long enough, and you reach: evolution. Our brains are clearly and obviously not built primarily to find truth. Rather, our brains are cobbled together towards a local maximum approximating evolutionary success, just like every other part of any living thing. Were we calculating machines, we might be able to handle the first three issues. But as is, we're evolved gene-driven robots. Brains are constructed first to compete with other same-gender same-tribe folks for reproductive success. Second, they're built to compete for scarce resources on the savanna with other tribes. And third, to go about the relatively easy (for our evolved apex-predator selves) business of not getting dead before having kids. Furthermore, brains are expensive, calorie-wise. A brain, much like muscles, should be no bigger than necessary in order to do it's job. And so it's more surprising that we can fake a general purpose logic-engine than that we have a bunch of heuristics that sub for logic most of the time.

Shorter: The Monkeybrain dominates logic in all of us.


On 4 counts, we'd better not think that we can predict much successfully. And so the core insight of rationality is that predicting the future is hard. And it's greatest prophets are Hume and Hayek.