More discussion regarding Daryl J. Bem on precognition

Following are the emails I sent to Daryl J. Bem, and his response to the first of these, regarding his article Feeling the Future: Experimental Evidence for Anomalous Retroactive Influences on Cognition and Affect (Journal of Personality and Social Psychology, DOI: 10.1037/a0021524). Actually, I realized afterward that my explanation for his "precognition" experiment would also apply if it was a human presenting the screens instead of a computer. It seems likely that in the same way a computer primes its response to user input, as explained in my emails, a human would prime her/his response when presenting a question, in order to react more rapidly to the anticipated answer. If that's indeed the case, we should see other examples of "precognition" if the person asked is capable of perceiving the priming by the person asking the question.

--------------------------------------- emails
From: Bertrand du Castel
Sent: Monday, January 31, 2011 9:50 AM
To: Daryl J. Bem
Subject: RE: Your experiment vs. electronic forensics

I had read your article thoroughly and I have indeed appreciated the double random algorithm. This is because you've been so thorough that I undertook to write you (I am very familiar with randomization which is a key element of computer security).

I need to explain to you that contrarily to what you think, there is something telling happening inside the computer before the participant indicates her/his response. Let me explain to you why:

(1) When the user clicks on the answer, the computer goes into a series of instructions that then triggers the appearance of the screen of choice. It is perfectly true that this happens after the user has clicked, and therefore, this sequence of instructions has nothing to do with the appearance of precognition.

(2) However, what's important here is what happens before the user is presented with the choice. In order for the proper screen to be later shown by instructions (1) after the user clicked, the computer has to prime its program. The way it does it is by associating, before presenting the choice to the user, what is called in programming the "click event" to the proper screen, so that when the user clicks, the right set of instructions is triggered. The very fact that this priming occurs is source of a set pattern of radiations that could be detected by the user.

If you show me the program code, I'll be able to pinpoint you exactly to where in the code the priming is done before presenting the choice to the user. But you can do that by yourself: you just have to look in the program at what's called the "click event handler". What you'll see is that the event handler is initially set to be ready to go to the right screen according to the answer, and only after that is the choice presented to the user. The way it subsequently works is that when the user clicks on the answer, a "click event" occurs, and the preset "click event handler" triggers the right set of instructions. But again, for this to work, the "click event handler" had to be set properly before the choice is presented to the user.

Please let me assure you that I have no pre-conception here; I just want to inform you of a possible source of bias.

Thanks. Bertrand.

---

From: Daryl J. Bem
Sent: Monday, January 31, 2011 9:06 AM
To: Bertrand du Castel
Subject: Re: Your experiment vs. electronic forensics

Thank you for your informative comments. Your concerns would be particularly relevant if my experiments were testing real-time clairvoyance; they are less relevant because the design is set up to test precognition. Accordingly, nothing happens inside the computer until after the participant has already indicated his/her response. Moreover, as I discuss on pp. 11-15 of my article (attached) , I used both an algorithmic random number generator (RNG) and a true hardware-based RNG to rule out a number of possible alternative interpretations, such as real-time clairvoyance.

---

From: Bertrand du Castel

Sent: Monday, January 31, 2011 8:31 AM
To: Daryl J. Bem
Subject: Your experiment vs. electronic forensics

Executive summary: Computer programming such as that used for priming screen display creates radiations that may be detected by a subject.

Dear professor Bem,

In examining possible biases for your screen experiment, you may have forgotten one which is familiar to security programmers such as smart card operating system programmers. As a reminder, smart cards are security token subject to attacks because they contain money, or means to access power, for example, military or other places. One common attack thieves may use on smart cards is sensing the patterns of activation of the electronic circuits via equipment such as thermal sensors or electron beam probes. The most obvious example of defense actually used in smart cards, the first one taught to programmers, is that when writing in a program a "if" statement, the two sides of the statement should be of equal length, otherwise it is easy to detect which side of the statement has been taken by measuring timing of electronic activation. Equal length is obtained by padding with bogus statements the side which would otherwise be smaller.

Typically, a non security-aware programmer doesn't know this, and would program your experiment without regard to patterns of electronic activation. The computer takes a different path when priming for one kind of screen or the other, because it associates the click instructions to one part of the computer circuitry or the other, which affects the internal pattern of the electronics in subtle ways, for example via caching algorithms of the operating system of the computer. It is entirely conceivable that a body primed for perception of erotic pictures via documented extra sensitization of the insula (cf. How do you feel — now? The anterior insula and human awareness (A. D. (Bud) Craig, Nature Reviews Neuroscience 10, 59-70, January 2009)) would be capable of detecting the differential radiations, and therefore make accurate predictions; since the signal is very weak, it is also very understandable that the deviation from the norm would be small, as it is in the experiment.

The way to avoid this bias is to have the computer itself connected remotely enough from the screen, while taking extra precautions in the insulation of the cable joining the computer and the screen. Your write-up doesn't say whether that was the case.

Regards. Bertrand du Castel.