I was just thinking today about how this LHC/RHIC deal would make a good section for the neverending thesis. My thesis argues that not only nonhuman, but nonbiological nature as well has an objective, intrinsic value that ought to be taken into account by humans. Thus, the value of the Earth itself--not counting humans--should be taken into account within a risk analysis of projects that could potentially cause global, existential catastrophes. So, yes, this is of academic interest to myself, and thanks to all for helping me out with your many comments.

But yeah, I am truly concerned that the LHC could destroy the Earth. I remember you earlier asked me what would make me change my mind about lighting off the LHC. I dodged your question, as you pointed out, by suggesting that the LHC be replaced by a 14 TeV single-beam collider. But now, after much more consideration of the issues, I am prepared to directly answer you, and that is I would consider the LHC to be reasonably safe if the p_catastrophe could be shown to be << 10^-18. And that's a nonconservative p_catastrophe since it only extends 100,000 years into the future, and only includes human values. And I just don't see that.

Let's go through the equation again:

1. Probability that we exist in a dangerous dimension: it seems that it's the 5th dimensional mBH's that are the really dangerous ones, and since there's basically 20 options and little a priori reason at this point to choose one over the others, I'll go with 5%.

2. Probability that 5-D mBH's produced at the LHC would get captured by the Earth's gravity: G&M said that the typical LHC produced mBH would not get trapped by Earth's gravity, they did say there was a finite probability that some would. So, I'm going to go with 95%.

3. Probability that Hawking radiation will fail and mBH's will be stable: I'll go with the median of the poll of physicists, 2%

4. Probability that the mBH will grow fast enough to be potentially problematic: this is complicated. I've already limited my analysis to just 5-D mBH's. And we've got to take into account the astrophysical evidence of numerous white dwarfs. In order for there to be a danger, it would have to be the case that CR produced mBH's pass through white dwarfs as well as Earth. Moreover, G&M argue that 5-D mBH's would be the most easily stopped mBH's. However, according to G&M's analysis of mBH stopping lengths in white dwarfs, their equation 5.34 (p. 35) that is supposed to constrain the sensitivity of their stopping length estimates contains (D-5) in the denominator. Thus, apparently, as far as I can tell, it is the 5-D stopping length estimates that are the most uncertain. (People please feel free to point out the flaws in my reasoning.) So what to choose as a probability? The Precautionary Principle says to be conservative. So, how about 1%. It's less than the probability that we are in 5 dimensions and less than the probability that Hawking radiation will fail.

Thus we have:


Very unlikely in other words. Yet the expected value of the number of deaths is 10^-5 x 6 x 10⁹ = 60,000. This is coincidently about what the worst case (most conservative) expected value in number of deaths that BJSW that Kent listed in the abstract of his paper above for BJSW's worst case p_catastrophe for RHIC.

Yes, I apparently pulled these numbers out of thin air. But what else can I do? Arguably, the 5% figure I used for the chance we live in 5 dimensions is much too high in light of the white dwarf data. That is, the probability that we live in 5 dimensions just is the probability that 5-D aren't captured by white dwarfs. But as it is, I'm multiplying the two probabilities, thus my subjective probability that we live in 5 dimensions is actually 0.05%.

Also, if CR-produced 5-D mBH's can in fact pass through white dwarfs harmlessly, that will make it much harder for a low velocity LHC-produced 5-D mBH to be captured by the Earth. On the other hand, if we live in 5 dimensions, the LHC will be a virtual black hole factory that will produce millions of mBH's over the course of it's useful life time. So I can change the 0.95 to 0.001 if you want, but that still leaves an expected value of number of deaths of 60. If the LHC were a nuclear reactor and fell under standard radiation risk minimization standards, that would be totally unacceptable--and that's not counting the value of future generations at all.

I would appreciate some help from the CERN scientists in determining the p_catastrophe for the LHC. However, the physicists are keeping their cards close to their chest this time around. In marked contrast to the earlier RHIC controversy, none of the CERN scientists have offered a quantitative estimate of p_catastrophe. At least after weeks of scouring the internet, I have not been able to find a quantitative estimate of p_catastrophe for the LHC from CERN. We're just told to trust them that p_catastrophe is really, really low.

That's what bugs me the most about this whole thing. How undemocratic this process has been. Physicists with no special expertise in risk analysis and who have a vested interest in seeing the LHC go through are deciding on an ad hoc basis what is an acceptable risk for the rest of us. I don't think that's fair. At the very least, they should lay their cards on the table and tell us what their best quantitative, worst case estimate for p_catastrophe for the LHC is, as well as a detailed breakdown of how they arrived at that estimate. But it doesn't look like that's going to happen. . . .