rmiller wrote:
At 03:58 PM 6/6/2005, you wrote:
rmiller wrote:

At 03:01 PM 6/6/2005, Pete Carlton wrote:

(snip)

The point is, there are enough stories published in any year that it would be a trivial matter to find a few superficial resemblances between any event and a story that came before it.

Let's look a little closer at the story in terms of gestalts.

On one side we have published author Robert Heinlein writing a story in 1939 about a guy named Silard who works with a uranium bomb, a beryllium target and a fellow named "lenz." We'll leave Korzybski out of this one (I suspect Heinlein borrowed the name from A. Korzybski, a sematicist of some renown back in the 1930s.) To me the interesting nodes involve the words "Silard" "lenz" "beryllium," "uranium" and "bomb." So let's agree that here is a story that includes a gestalt of the words "Silard, lenz, beryllium, uranium and bomb."

But you can't use that particular "gestalt" when talking about the probability that a coincidence like this would occur, because you never would have predicted that precise gestalt in advance even if you were specifically looking for stories that anticipated aspects of the Manhatten project.

Where on earth did *that* gestalt rule come from??? ;-)

It comes from the idea that if you are going to claim a "hit" for ESP or whatever that has a significance level of 1 in 10^9, then it should be true that *if* ESP didn't exist and the hit was a pure coincidence, then no more than 1 in 10^9 versions of you in different possible universes should be claiming to have seen a hit with that level of significance--that seems to be inherent in the meaning of "significance level", no? If 1 out of 15 parallel versions of you would end up claiming to see a hit with a significance level of 1 in 10^9, *even if there was no ESP and these hits were pure coincidence*, then obviously something is wrong with the reasoning that these various versions of you are using. Do you disagree?


I wrote:

It would make more sense to look at the probability of a story that includes *any* combination of words that somehow anticipate aspects of the Manhatten project. Let's say there were about 10^10 possible such gestalts we could come up with, and if you scanned trillions of parallel universes you'd see the proportion of universes where a story echoed at least one such gestalt was fairly high--1 in 15, say.



This means that in 1 in 15 universes, there will be a person like you who notices this anticipation and, if he uses your method of only estimating the probability of that *particular* gestalt, will say "there's only a 1 in 10^9 probability that something like this could have happened by chance!" Obviously something is wrong with any logic that leads you to see a 1 in 10^9 probability coincidence happening in 1 in 15 possible universes, and in this hypothetical example it's clear the problem is that these parallel coincidence-spotters are using too narrow a notion of "something like this", one which is too much biased by hindsight knowledge of what actually happened in their universe, rather than something they plausibly might have specifically thought to look for before they actually knew about the existence of such a story.



you responded:

Sounds like you're invoking rules of causation here--post hoc rather than ad hoc, hindsight bias, etc. Certainly I am not suggesting Heinlein's story caused Szilard to be hired (interesting thought, though!) And unless I want to invoke Cramer's transactional approach, I would not really want to think that the Manhattan Project caused Heinlein to write his story. That would require reverse causation, and we know that doesn't happen. This is very simple: we have instances in which Heinlein includes key words (definable as being essential to the story---without them, different story) that form a gestalt of. . .well, key words. These words are equivalent to those describing the Manhattan Project and not many other things. To show that there are not many other things these key word gestalts describe, one can wait a year and use Google Print to call up all the books and stories associated with these key words. Then we will have a probability to work with. Since the gestalts are separated by four years (or thereabouts) then we shouldn't have to invoke causation.


You are misunderstanding the meaning of "hindsight bias", it's not about the prediction causing the event being predicted or vice versa, it's about *you* making a retroactive calculation of the "probability" of a particular successful prediction which is illegitimate because you incorporate knowledge of what already happened into your choice of the "target" whose probability of predicting you want to calculate. This is illustrated by an anecdote involving Richard Feynman which is quoted at http://www.numeraire.com/value_wizard/probable.htm :




What came to Feynman by "common sense" were often brilliant twists that perfectly captured the essence of his point. Once, during a public lecture, he was trying to explain why one must not verify an idea using the same data that suggested the idea in the first place. Seeming to wander off the subject, Feynman began talking about license plates. "You know, the most amazing thing happened to me tonight. I was coming here, on the way to the lecture, and I came in through the parking lot. And you won't believe what happened. I saw a car with the license plate ARW 357. Can you imagine? Of all the millions of license plates in the state, what was the chance that I would see that particular one tonight? Amazing!" A point that even many scientists fail to grasp was made clear through Feynman's remarkable "common sense."



Of course in this example Feynman did not anticipate in advance what licence plate he'd see, but the kind of "hindsight bias" you are engaging in can be shown with another example. Suppose you pick 100 random words out of a dictionary, and then notice that the list contains the words "sun", "also", and "rises"...as it so happens, that particular 3-word "gestalt" is also part of the title of a famous book, "the sun also rises" by Hemingway. Is this evidence that Hemingway was able to anticipate the results of your word-selection through ESP? Would it be fair to test for ESP by calculating the probability that someone would title a book with the exact 3-word gestalt "sun, also, rises"? No, because this would be tailoring the choice of gestalt to Hemingway's book in order to make it seem more unlikely, in fact there are 970,200 possible 3-word gestalts you could pick out of a list of 100 possible words, so the probability that a book published earlier would contain *any* of these gestalts is a lot higher than the probability it would contain the precise gestalt "sun, also, rises". Selecting a precise target gestalt on the basis of the fact that you already know there's a book/story containing that gestalt is an example of hindsight bias--in the Heinlein example, you wouldn't have chosen the precise gestalt of Szilard/lens/beryllium/uranium/bomb from a long list of words associated with the Manhattan Project if you didn't already know about Heinlein's story.

Jesse


Reply via email to