Back to LessWrong

Talk:Rationality

From Lesswrongwiki

Jump to: navigation, search

This is a brief analysis of the wiki page.

Claims made by the page:

(1) Rationality is the characteristic of thinking and acting optimally.

(2) An agent is rational if it wields its intelligence in such a way as to maximize the convergence between its beliefs and reality; and acts on these beliefs in such a manner as to maximize its chances of achieving whatever goals it has.

(3) For humans, this means mitigating (as much as possible) the influence of cognitive biases.

(4) Instrumental rationality is the art of choosing and implementing actions that steer the future toward outcomes ranked higher in one's preferences. Said preferences are not limited to 'selfish' preferences or unshared values; they include anything one cares about.

(5) Epistemic rationality is that part of rationality which involves achieving accurate beliefs about the world.

(6) It involves updating on receiving new evidence, mitigating cognitive biases, and examining why you believe what you believe.

(7) It can be seen as a form of instrumental rationality in which knowledge and truth are goals in themselves, whereas in other forms of instrumental rationality, knowledge and truth are only potential aids to achieving goals.

(8) Someone practising instrumental rationality might even find falsehood useful.

Support given and implied:

(1), (2), (4), (5), (6) could be taken as definitions, with (3), (7), (8) as their consequences. If this were done, it would need to be shown that (1), (2), (4), (5), (6) are good definitions. This would consist in the following. Firstly, a purpose for stating the definitions would be given. This purpose could be, for example, to capture our pre-theoretical use of "rationality", in order to utilise our intuitions about statements involving "rationality" as evidence for more complex statements. Alternatively, this purpose could be, for example, to provide a technical definition of "rationality" which will be used later on, in the same way a mathematical symbol is used as a place holder for a more complex mathematical object. Secondly, the definitions would be shown to be consistent, well-formed and clear, each to an appropriate level depending on the purpose. For instance, if the purpose was to provide a technical definition, the definitions would need to be fully consistent, precise and formed out of other precise terms in a way consistent with those terms. If the purpose was to capture our pre-theoretical usage, the definitions would need to be consistent insofar as we believe our pre-theoretical usage, in the realm of the intuitions we wish to utilise, to be, as precise as our pre-theoretical usage in the relevant realm is, and to be as close as possible to our pre-theoretical usage in the relevant realm. Once all this was done, it would be shown that the definitions are good. Finally, it would need to be shown that (3), (7), (8) are indeed consequences of the definitions.

The blog posts linked may contain support for the claims in the manner outlined above, or in another manner.

Suggestions

The way in which the blog posts linked are intended to support the claims could be made explicit on the page. If any of the blog posts intend to support the claims by taking (1), (2), (4), (5), (6), or some combination of those, as definitions, the purpose of those definitions could be made explicit on the page.

Merge with Rationalism?

I would like to rename and merge this page with Rationalism - comments? Deku-shrub (talk) 04:07, 15 April 2017 (AEST)