Let’s say a human brain can have ten thoughts per second; or the rate of human awareness is ten perceptions per second. Fifty years of torture means just over one and a half billion tortured thoughts, or perceptions of torture.
Let’s say a human brain can distinguish twenty logarithmic degrees of discomfort, with the lowest being “no discomfort at all”, the second-lowest being a dust speck, and the highest being torture. In other words, a single moment of torture is 2^19 = 524288 times worse than a dust speck; and a dust speck is the smallest discomfort possible. Let’s call a unit of discomfort a “dol” (from Latin dolor).
In other words, the torture option means 1.5 billion moments × 2^19 dols; whereas the dust-specks option means 3^^^3 moments × 1 dol.
The assumptions going into this argument are the speed of human thought or perception, and the scale of human discomfort or pain. These are not accurately known today, but there must exist finite limits — humans do not think or perceive infinitely fast; and the worst unpleasantness we can experience is not infinitely bad. I have assumed a log scale for discomfort because we use log scales for other senses, e.g. brightness of light and volume of sound. However, all these assumptions can be empirically corrected based on facts about human neurology.
Torture is really, really bad. But it is not infinitely bad.
That said, there may be other factors in the moral calculation of which to prefer. For instance, the moral badness of causing a particular level of discomfort may not be linear in the amount of discomfort: causing three dols once may be worse than causing one dol three times. However, this seems difficult to justify. Discomfort is subjective, which is to say, it is measured by the beholder — and the beholder only has so much brain to measure it with.
I suspect that I would prefer the false memory of having been tortured for five minutes to the false memory of having been tortured for a year, assuming the memories are close replicas of what memories of the actual event would be like. I would relatedly prefer that someone else experience the former rather than the latter, even if I’m perfectly aware the memory is false. This suggests to me that whatever I’m doing to make my moral judgments that torture is bad, it’s not just summing the number of perception-moments… there are an equal number of perception-moments in those two cases, after all. (Specifically, none at all.)
That said, this line of thinking quickly runs aground on the “no knock-on effects” condition of the initial thought experiment.
I suspect that I would prefer the false memory of having been tortured for five minutes to the false memory of having been tortured for a year, assuming the memories are close replicas of what memories of the actual event would be like.
Actually, from what I read about related research in “Thinking, Fast and Slow”, it’s not clear that you would (or that the difference would be as large as you might expect, at least). It seems that memories of pain depend largely on the most intense moment of pain and on the final moment of pain, not necessarily on duration.
For example, in one experiment (I read the book a week ago and write from memory), subjects were asked to put their hand in a bowl of cold water (a painful experience) for two minutes, then they were asked to put their hands in cold water for two minutes, followed by the water being warmed gradually over another 5 minutes. (There were reasonable controls, obviously.) Then they were asked which experience to repeat. The majority chose experience two, even though intuitively it is strictly worse than experience one.
Of course, you’d have to find the actual related paper(s), check how high the correlation/ignoring-duration effect is, check if there’s significant inter-individual variation (whether maybe you’re an unusual person who cares about duration), but, regardless, there are significant reasons to doubt your intuitions in this scenario.
This suggests to me that whatever I’m doing to make my moral judgments that torture is bad, it’s not just summing the number of perception-moments… there are an equal number of perception-moments in those two cases, after all. (Specifically, none at all.)
True — we need a term for moments of discomfort caused by contemplation, not just ones caused by perception.
It seems to me, though, that your brain can only perceive a finite number of gradations of unpleasant contemplation, too. The memory of being tortured for five minutes, the memory of being tortured for a year, and the memory of having gotten a dust speck in your eye could occupy points on this scale of unpleasantness.
How bad is the torture option?
Let’s say a human brain can have ten thoughts per second; or the rate of human awareness is ten perceptions per second. Fifty years of torture means just over one and a half billion tortured thoughts, or perceptions of torture.
Let’s say a human brain can distinguish twenty logarithmic degrees of discomfort, with the lowest being “no discomfort at all”, the second-lowest being a dust speck, and the highest being torture. In other words, a single moment of torture is 2^19 = 524288 times worse than a dust speck; and a dust speck is the smallest discomfort possible. Let’s call a unit of discomfort a “dol” (from Latin dolor).
In other words, the torture option means 1.5 billion moments × 2^19 dols; whereas the dust-specks option means 3^^^3 moments × 1 dol.
The assumptions going into this argument are the speed of human thought or perception, and the scale of human discomfort or pain. These are not accurately known today, but there must exist finite limits — humans do not think or perceive infinitely fast; and the worst unpleasantness we can experience is not infinitely bad. I have assumed a log scale for discomfort because we use log scales for other senses, e.g. brightness of light and volume of sound. However, all these assumptions can be empirically corrected based on facts about human neurology.
Torture is really, really bad. But it is not infinitely bad.
That said, there may be other factors in the moral calculation of which to prefer. For instance, the moral badness of causing a particular level of discomfort may not be linear in the amount of discomfort: causing three dols once may be worse than causing one dol three times. However, this seems difficult to justify. Discomfort is subjective, which is to say, it is measured by the beholder — and the beholder only has so much brain to measure it with.
I suspect that I would prefer the false memory of having been tortured for five minutes to the false memory of having been tortured for a year, assuming the memories are close replicas of what memories of the actual event would be like. I would relatedly prefer that someone else experience the former rather than the latter, even if I’m perfectly aware the memory is false. This suggests to me that whatever I’m doing to make my moral judgments that torture is bad, it’s not just summing the number of perception-moments… there are an equal number of perception-moments in those two cases, after all. (Specifically, none at all.)
That said, this line of thinking quickly runs aground on the “no knock-on effects” condition of the initial thought experiment.
Actually, from what I read about related research in “Thinking, Fast and Slow”, it’s not clear that you would (or that the difference would be as large as you might expect, at least). It seems that memories of pain depend largely on the most intense moment of pain and on the final moment of pain, not necessarily on duration.
For example, in one experiment (I read the book a week ago and write from memory), subjects were asked to put their hand in a bowl of cold water (a painful experience) for two minutes, then they were asked to put their hands in cold water for two minutes, followed by the water being warmed gradually over another 5 minutes. (There were reasonable controls, obviously.) Then they were asked which experience to repeat. The majority chose experience two, even though intuitively it is strictly worse than experience one.
Of course, you’d have to find the actual related paper(s), check how high the correlation/ignoring-duration effect is, check if there’s significant inter-individual variation (whether maybe you’re an unusual person who cares about duration), but, regardless, there are significant reasons to doubt your intuitions in this scenario.
… huh.
I wonder if we might actually value experiences this way?
Daniel Kahneman suggests that we do. We remember thing imperfectly and optimize for the way we remember things. Wiki has a quick summary.
True — we need a term for moments of discomfort caused by contemplation, not just ones caused by perception.
It seems to me, though, that your brain can only perceive a finite number of gradations of unpleasant contemplation, too. The memory of being tortured for five minutes, the memory of being tortured for a year, and the memory of having gotten a dust speck in your eye could occupy points on this scale of unpleasantness.