My attempt to Correct Myself
I was rushing, earlier, to try to answer Diogenes' objections, to meet a deadline, and I'm afraid that I started making mistakes -- note my need for two corrections, quickly posted. When returning from my apointment I started to think this out again...am I beginning to see a glimmer here?
I know how frustrated and dissatisfied Grw8ght and Diogenes are with me. I therefore start with the operative assumption, "PianoTeacher has made a BIG SLOPPY mistake that *only he* can't see. He is incapable of working thru the examples he has contrived to answer it, and obviously is missing the point of others because of this gargantuan error. Given that, let us re-state a fully-reduced actual impossibly paranormal challenge, and step-by-step, relate it to the parameters of the Wellfed claim."
'OK,' you say, 'But at least have the decency to *entertain* people, not bore them to death.'
Well...I'll try.
Paranormal Challenge No. 1:
We meet a an allegedly delusional character who is named Unfed. He is convinced that his next door neighbor is sending "evil thoughtwaves" his way. He is so certain of that, that he thinks that he can prove it; hell, he is *so* sure, he insists that he will be able to show *ten times out of ten tries* to detect them that he can really, truly perceive those evil waves.
His neighbor happens to be a busy salesman who only rushes home, from time to time, to pick up orders and gulp a cup of coffee. He whisks in-and-out, in-and-out, all day and all night long.
Unfed has two acquaintances: wise, patient James Randi; and the paper boy, who could care less about the guy's delusions, one way or another.
James Randi does not believe that it is rational for Unfed to make these assertions. He says, "Say: I bet you *can't* pick up "evil thoughtwaves", and I'll put up ten bucks that says so!"
The bet is accepted by Unfed, and James quietly, without telling anybody, arranges to find out whenever the next door neighbor is HOME, and when he is AWAY. James intends to keep that info to himself.
He asks Unfed to write down the exact times, up to ten occurrences, over a period of exactly three days, that he has the conviction that those "evil thoughtwaves" are impinging on him.
Meanwhile, James watches the neighbor's house all day and night, for a period of three days, noting down the time the neighbor is home, and when he has been away from the house. He covers up his sheet of paper while writing and does not let anyone see the list thereon; when finished, James puts it into a mayonnaise jar on Funk & Wagnall's porch.
At the end of the three day test period, the paper boy collects Unfed's list, and James' list, and makes a tally of the time periods. If Unfed is right and CAN perceive *evil thoughtwaves*, he'll claim they were emitted during times that James has recorded that the salesman neighbor was *actually* home. Any one false positive out of ten will prove that Unfed cannot assert with perfect accuracy that he CAN indisputably tell if he can ALWAYS perceive the "evil thoughtwaves".
When the tally is made, there are SOME positives, but not ten. There are enough false positives to make the result come out purely random; ergo, Unfed is imagining things, and can't win James' ten bucks.
-------
Adjusted Paranormal Challenge No. 2:
It seems to me that we have restated something very close to the test concept that Diogenes has summarized: it is reduced nearly to essentials. The only difference I can discern is that James' list has a time range, not merely ten fixed numbers representing ten individual times, because the neighbor was in the house for varying periods of time, even though he was perpetually in a hurry.
We can correct for this by turning the test into one that is a little different but has a closer fit, by stipulating that "Unfed believes that the neighbor only emits evil thoughtwaves the *exact moment* that the neighbor steps into his house; the rest of the time he is too busy thinking about something else." So James Randi need only record the EXACT moment that the neighbor steps into his house. Under the circumstances, there is no "range" for providing windows of time; the tally of times given by Unfed, and James, must agree PRECISELY, ten out of ten instances.
I believe may be an exact fit.
-------
I shall try to take PianoTeacher (me!) by the hand, and walk him thru the examples above, and the Wellfed Claim, step by step.
1. Wellfed's assertions that he can hear differences -- irrespective of GSIC, but as a hi fi listener -- are normal human cognitions and judgments. There is nothing delusional about his hearing differences under some conditions; his claim only begins to run parallel with the two Paranormal Claims above when he inserts the insistence that the differences HAVE to be caused by the GSIC device.
2. Similarly, Unfed can perceive evil words if his neighbor shouts profane insults at him within earshot; but the "perception of evil thoughtwaves" is indeed, in the considered view of rational Randi, an irrational delusion UNLESS by some strange quirk, Unfed's data match Randi's and prove Unfed's claim. THAT would be weeeeeeeiiiirrrrd, and surely paranormal!
3. There is an existing bias, of James, that Unfed is "irrational". It is strong and reality-based, and only PROOF can overcome it and show evidence of the paranormal, for no one thinks that a person can detect evil thoughtwaves, and nobody can measure them; certainly no one *else* claims to perceive them, in James' neighborhood.
4. Unfed only becomes "not irrational" when he proves his point and wins the ten bucks. No other situation allows for him to be "not irrational"; at least no way that rational processes can infer or demonstrate.
5. Wellfed, on the other hand, may be rational when he hears differences that have real causes, and differences that other people can hear, too. We don't know if he is imagining the differences, and don't care if GSIC is causing them; we only want to line up Wellfed's claimed hits, with the instances of which CD was in the player. The resulting calculation for positives in ten out of ten tries MUST, per Wellfed's boast, be exactly right every time; no false positives will be allowed.
6. We are not interested whether or whether not what Wellfed claims to be able to perceive are "audio nuances" or just self-delusions. We don't have to quantify them at all.
7. Ergo, we are unconcerned with cognition, etc.
8. As a matter of fact, our biases are that the GSIC can have no effect so there cannot be, by a known mechanism, a way for Wellfed to line up all true positives, and no false ones, except by a virtually impossible chance occurrence.
Now, I *think* this is correct.
-------
Probably the only thing to worry about here is Wellfed's boast, and its degree of practicality. He may actually hear nuances; our biases say "NO -- he isn't" but that's irrelevant. Proof of a line-up (and inferred coincidence with GSIC treatment) comes out of the data -- the true positives. A false positive immediately invalidates the boast that "ten out of ten tries" can be achieved; it in effect terminates the test and proves that Wellfed can't win the Challenge. Sorry: no million bucks.
Now, assuming -- and going against the biases of rational people; indeed the proofs that PianoTeacher himself has shown -- that Wellfed, though able to detect concrete and actual nuances and preferences between CDs, and that GSIC *does work*, Wellfed only wins the Challenge if he moves the bar so low that he CANNOT lose. This means that in the eventuality of (a) faint nuances from GSIC that vary all over the place, from undetectable to obvious; (b) Wellfed, though a good listener, makes the usual mistakes of judgment and evaluation errors because he can't reliably focus; and (c) his protocol is too loose in some other parameters, he must set the bar VERY low indeed in order to win: he must have it as low, or in fact lower, than the most minimal mistake of all that he will make.
With the bar set too high, Wellfed can't win even IF the effect is real and he can detect it. He simply makes too many mistakes to get ten out of ten true positives. He might also have set the bar higher than that for any listener who was more reliable and sensitive than he is, and who makes fewer mistakes.
At some point, the bar is set so high that: (a) no one can win; (b) a good test subject never could avoid false positives; and (c) it does not matter if the effect is real, and also if it is "Belethlike" (i. e., obvious to every listener, per my earlier discussion). I can see a problem with one exception of (c). It would be possible to set the bar too high to detect even Belethlike differences -- by choosing SO MANY hits as the required number of tries -- that Wellfed actually falls asleep or becomes completely disoriented and starts blowing it, while normal people listening to the original ten out of ten tries, always can detect Belethlike differences. Eventually there is a point where even nominally Belethlike differences become impossible to detect: when the test subject loses consciousness, or if total loss of the necessary mental coherence is caused by something.
------
I may have actually worked through "the paradox of Diogenes" here, right, and have located my mistake. Or, have I?
But, in looking over my flowchart, I see that the narrative above seems to fit into it, so the linkages I was seeing last night, which caused me to conclude that "the Wellfed protocol as stands is impossible under any circumstances", is still valid. For indeed the bar is too high, considering my EXTRA parameter added that "ten out of ten tries" is impractical and unachievable, given that a fallible human listener, with a loose listening protocol, would start to lose confidence in reliably detecting real but small nuances.
Unfortunately, my wife has just informed me "that I must go", so I only have time for a quick read to check spelling. We'll take a chance and post this and see how embarrassed PianoTeacher may become by his latest attempt to think this through!
Yours,
PianoTeacher, himself.