It’s reasonable to update as a result of the analysis of fiction (including fanfiction) for two reasons, neither of which are directly related to the events of the story in the same way that events in real life are related to updating. The first is: does this prompt me to think in a way I did not before? If so, it is not evidence, but it allows you to better way the evidence by providing you with more possibilities. The second is: why was this written? Even a truthless piece of propaganda can be interesting evidence in that it is entangled with human actions and motivations.
If considering a new hypothesis fundamentally changes the way you think about priors, and the arguments you used to justify ratios between hypotheses no longer hold, then, yes, you will have to look at the evidence again.
I feel a little odd about calling that process ‘updating’, since I think it’s a little more involved than taking into account a single new piece of evidence.
The first is: does this prompt me to think in a way I did not before? If so, it is not evidence, but it allows you to better way the evidence by providing you with more possibilities.
I think that this would only be true if it prompts you to think in a new and random way. Fiction which prompts you to think in a new but non-random way (that is, all fiction) could very well make it worse. It could very well be that the author selectively prompts you to think only in cases where you got it right without doing the thinking. If so, then this will reduce your chance of getting it right.
For a concrete example, consider a piece of homeopathic fiction which “prompts you to think” about how homeopathy could work. It provides a plausible-sounding explanation, which some people haven’t heard of before. That plausible-sounding explanation either is rejected, in which case it has no effect on updating, or accepted, making the reader update in the direction of homeopathy. Since the fiction is written by a homeopath, it wouldn’t contain an equally plausible sounding (and perhaps closer to reality) explanation of what’s wrong with homeopathy, so it only leads people to update in the wrong direction.
Furthermore, homeopathy is probably more important to homeopaths than it is to non-homeopaths. So not only does reading homeopathic fiction lead you to update in the wrong direction, reading a random selection of fiction does too—the homeopath fiction writers put in stuff that selectively makes you think in the wrong direction, and the non-homeopaths, who don’t think homeopathy is important, don’t write about it at all and don’t make you update in the right direction.
The scenario is not evidence at all, fictional or not. The reasoning involved might count as evidence depending on your definition, but giving a concrete example is not additional evidence, it only makes things easier to understand. Calling this fictional evidence is like saying that an example mentioning parties A, B, and C is “fictional evidence” on the grounds that A, B, and C don’t really exist.
The scenario is not evidence at all, fictional or not. The reasoning involved might count as evidence depending on your definition, but giving a concrete example is not additional evidence, it only makes things easier to understand. Calling this fictional evidence is like saying that an example mentioning parties A, B, and C is “fictional evidence” on the grounds that A, B, and C don’t really exist.
Interesting point. The sort of new ways of thinking I had imagined were more along the lines of “consider more possible scenarios”—for example, if you had never before considered the idea of a false flag operation (whether in war or in “civil” social interaction), reading a story involving a false flag operation might prompt you to reinterpret certain evidence in light of the fact that it is possible (a fact not derived directly from the story, but from your own thought process inspired by the story). While it is certainly possible to update in the wrong direction, the thought process I had in mind was thus:
I have possible explanations A, B, and C for this observed phenomenon Alpha.
I read a story in which event D occurs, possibly entangled with Alpha, a similar phenomenon to Alpha.
I consider the plausibility of an event of the type D occurring, taking in not only fictional evidence but also real-world experience and knowledge, and come to the conclusion that while D takes certain liberties with the laws of (psychology/physics/logic), the event D is entirely plausible, and may be entangled with a phenomenon such as Alpha*.
I now have possible explanations A, B, C, and D for the observed phenomenon Alpha.
It is important to note that fiction has no such use for a hypothetical perfect reasoner, who begins with priors assigned to each and every physically possible event. Further, it would be of no use to anyone incapable of making that second-to-last step correctly; if they simply import D* as a possible explanation for Alpha, or arrive at some hypothetical event D which is not, in fact, reasonable to assume possible or plausible, then they have in fact been hindered by fictional “evidence”.
What about fanfictional evidence?
More seriously, shouldn’t it be “don’t update on fictional evidence as if it were true”?
Certainly it’s reasonable for a story to make us reconsider our beliefs.
It’s reasonable to update as a result of the analysis of fiction (including fanfiction) for two reasons, neither of which are directly related to the events of the story in the same way that events in real life are related to updating. The first is: does this prompt me to think in a way I did not before? If so, it is not evidence, but it allows you to better way the evidence by providing you with more possibilities. The second is: why was this written? Even a truthless piece of propaganda can be interesting evidence in that it is entangled with human actions and motivations.
If considering a new hypothesis fundamentally changes the way you think about priors, and the arguments you used to justify ratios between hypotheses no longer hold, then, yes, you will have to look at the evidence again.
I feel a little odd about calling that process ‘updating’, since I think it’s a little more involved than taking into account a single new piece of evidence.
I think that this would only be true if it prompts you to think in a new and random way. Fiction which prompts you to think in a new but non-random way (that is, all fiction) could very well make it worse. It could very well be that the author selectively prompts you to think only in cases where you got it right without doing the thinking. If so, then this will reduce your chance of getting it right.
For a concrete example, consider a piece of homeopathic fiction which “prompts you to think” about how homeopathy could work. It provides a plausible-sounding explanation, which some people haven’t heard of before. That plausible-sounding explanation either is rejected, in which case it has no effect on updating, or accepted, making the reader update in the direction of homeopathy. Since the fiction is written by a homeopath, it wouldn’t contain an equally plausible sounding (and perhaps closer to reality) explanation of what’s wrong with homeopathy, so it only leads people to update in the wrong direction.
Furthermore, homeopathy is probably more important to homeopaths than it is to non-homeopaths. So not only does reading homeopathic fiction lead you to update in the wrong direction, reading a random selection of fiction does too—the homeopath fiction writers put in stuff that selectively makes you think in the wrong direction, and the non-homeopaths, who don’t think homeopathy is important, don’t write about it at all and don’t make you update in the right direction.
does anyone else find it ironic that we’re using fictional evidence (a story about homeopathic writers that don’t exist) to debate fictional evidence?
The scenario is not evidence at all, fictional or not. The reasoning involved might count as evidence depending on your definition, but giving a concrete example is not additional evidence, it only makes things easier to understand. Calling this fictional evidence is like saying that an example mentioning parties A, B, and C is “fictional evidence” on the grounds that A, B, and C don’t really exist.
The scenario is not evidence at all, fictional or not. The reasoning involved might count as evidence depending on your definition, but giving a concrete example is not additional evidence, it only makes things easier to understand. Calling this fictional evidence is like saying that an example mentioning parties A, B, and C is “fictional evidence” on the grounds that A, B, and C don’t really exist.
Interesting point. The sort of new ways of thinking I had imagined were more along the lines of “consider more possible scenarios”—for example, if you had never before considered the idea of a false flag operation (whether in war or in “civil” social interaction), reading a story involving a false flag operation might prompt you to reinterpret certain evidence in light of the fact that it is possible (a fact not derived directly from the story, but from your own thought process inspired by the story). While it is certainly possible to update in the wrong direction, the thought process I had in mind was thus:
I have possible explanations A, B, and C for this observed phenomenon Alpha.
I read a story in which event D occurs, possibly entangled with Alpha, a similar phenomenon to Alpha.
I consider the plausibility of an event of the type D occurring, taking in not only fictional evidence but also real-world experience and knowledge, and come to the conclusion that while D takes certain liberties with the laws of (psychology/physics/logic), the event D is entirely plausible, and may be entangled with a phenomenon such as Alpha*.
I now have possible explanations A, B, C, and D for the observed phenomenon Alpha.
It is important to note that fiction has no such use for a hypothetical perfect reasoner, who begins with priors assigned to each and every physically possible event. Further, it would be of no use to anyone incapable of making that second-to-last step correctly; if they simply import D* as a possible explanation for Alpha, or arrive at some hypothetical event D which is not, in fact, reasonable to assume possible or plausible, then they have in fact been hindered by fictional “evidence”.