Saturday, January 02, 2010

LHC alarmists: a judge would rule "stop LHC"

In the previous article, I discussed various beliefs in the end of the world (Y2K, AGW...) and Carl Brannen has brought our attention to the following new physics.soc-ph preprint:
Eric Johnson: The black hole case: the injunction against the end of the world (PDF)
The lawsuits by the LHC alarmists have so far been rejected on technical grounds such as wrong jurisdiction. Thank God.

The author of the paper, an assistant professor of law in Tennessee, analyzes what a judge would probably think and say if he had to evaluate the case on its merits. To save the time of most of you, after having written 90 pages of the preprint, he would probably stop the LHC. I just wanted you to know what you may expect from the lawyers if they get a chance to have their say. ;-/

There's a dead lawyer and a dead hare on the road. Do you know how can you distinguish them? You will find a braking track in front of the hare. :-)

Johnson reviews some positive and negative things that people in many diverse groups have said about particle physics. He discusses the dynamics that killed the SSC. And he summarizes some statements about the "dangers" from the LHC alarmists on one side and the active physicists on the other side. He seems capable to extract the relevant pieces of the texts - but he is manifestly incapable to determine which arguments are valid and which of them are not.

I think that in this respect, the typical judges would have the same strengths and limitations. It's not too good news but it's a kind of tautology because a lawyer is defined as someone who doesn't understand particle physics too well. ;-) You may see that his limitations are pretty severe. For example, simple pop science papers by Giddings and Mangano filled with trivial power laws that every physics PhD should be able to reconstruct (and that have deservedly received only 5 or 17 citations, respectively) are called "mammoth" papers discussing "unfathomably complex" issues. ;-)

He dedicated sections to extra dimensions, black hole evaporation, and so on. Given his being a lawyer, the writing is intelligent. But you can see that he is incoherently combining arguments with very different degrees of complexity, reliability, coherence, accuracy, experimental verification, and required sophistication of the physical theories. He can't distinguish them. He doesn't know which of them is superior. No lawyer would be able to determine such things.

That's why his sources range from John Ellis on one side to Luis Sancho on the other side. Now, Luis Sancho is a twat - see e.g. his unification theory - but it's obvious that every person who knows even less physics than Luis Sancho - and most lawyers may fall into this category - will be unable to see this not-so-subtle fact. So the ordinary people won't like that I am saying that Sancho is a twat - because they don't quite see that he is one. With their subjectively conceivable possibility that Sancho is not a twat, my (or Brian Cox's) statement that he is a twat is insulting.

Because the lawyers can't see even self-evident errors in argumentation, no argument can ever be settled in this environment. And Johnson tries to evaluate the risks.

It is not too surprising that his conclusion would be "stop the LHC". Even if you assume that the "strength" of the LHC alarmists' voice is one millionth of the voice of the particle physicists, the risks end up being unacceptable. The alarmists argue that there is an O(1) probability for the destruction of life by the LHC. Even if you reduce the estimate by a factor of one million, the expectation value for the number of casualties will be in thousands of people. An experiment that kills thousands of people will be deemed inhuman by most people.

He also calculates the average expected losses - the risk R - by multiplying the probability P of a disaster with the losses L. Of course, in the zeroth approximation, this is an indeterminate form, 0 times infinity. To regulate the expression, he puts P to be something like 10^{-12} while L is something like 4 septillions - the price of all the people who live or will live on Earth. ;-)

With these huge losses, the LHC is a bad idea once again. The risk calculated in this way would be in trillions of U.S. dollars while the "normal cost" is just in billions.

The quantification of the losses connected with the complete destruction of the Earth is a task that is affected by many assumptions and moral values but it is one where the expertise in particle physics doesn't help (except for giving an analyst a certain intelligence). Let me start with this factor, L. What is the price of destroying the civilization and/or life on Earth?

It's reasonable to expect that the bulk of it - or at least a substantial portion - is the price of all human lives: their assets and other animals are a small perturbation, relatively speaking. So we should essentially consider the somewhat unethical product, namely the number of the people multiplied by the price of a human life. How many people should we consider? Well, I think that we can only count the people who are actually living today - or when the LHC destroys the world.

Why shouldn't we count the people who will live in the future billions of years (before the Sun stops) but who are not alive yet into the losses L? Well, because the LHC won't directly kill them. Their price is already counted in the price of the currently living ancestors because the ancestors would "produce them" later.

To make this point more clear, let me say that you may evaluate the price of a life by looking at the integrated life insurance people are ready to pay throughout their life divided by the risks. They are getting insured against their death - which includes the non-existence of all the sons and daughters whose conception would occur after the early death - and all of their offspring; the ancestors are the only ones who can defend the interests of the future generations. Despite all the people indirectly killed in the billions of years, the total price of a life ends up being $50,000 or $5 million, according to different sources. It's very finite and it includes all the future generations.

So the future generations shouldn't really be separately counted. At any rate, I admit that the destruction of all the life is a special event, the additivity may break down, and some people could add the factor of (7 billion years over 70 years) i.e. 8 orders of magnitude to L. At any rate, you will get a large but finite number - between hundreds of trillions and octillions.

The physical part of the problem

See also: LHC alarmists, other TRF articles about the topic...

Of course, the real "expert" physics is hiding in P, the probability of an LHC-induced doomsday. Eric Johnson says it's 10^{-12} which is way too high. What is the best scientific estimate for P? Quite generally, the more scientifically correct method you use, the smaller P you will get. If you don't know anything about science, you could say that P=0.5.

However, it's important to realize that P is surely not "quite" zero and it cannot be. In fact, the Earth may cease to exist in 2010 even without the LHC. After all, we don't know whether our Universe is stable. It can decay to another vacuum: we just know that the lifetime shouldn't be much shorter than 13.7 billion years because that would make our survival unlikely and the hypothesis about a speedy decay would be awkward and disfavored by the evidence.

Now, I tend to think that the lifetime of our Universe is much longer than 13.7 billion years. And it may be "completely" stable which would lead you to use the de Sitter Poincaré recurrence time for the life expectancy of our Universe. But the point is that I can't be sure that the Universe won't decay in 2010.

If we knew that our Universe were naturally unstable, its estimated life expectancy would be of order 10 billion years and the probability that the whole Universe will decay in 2010 will be 10^{-10} - one over ten billion years. Even if we accept that there is a 1% probability that the laws of physics want to use metastable de Sitter spaces (e.g. because more than 1% of the top cosmologists will tell you that this is a likely fate of de Sitter universes like ours), the final "averaged" probability that the Universe will die in 2010 will be 10^{-12} - without any help from us, humans.

This observation addresses one important point of my criticism of their calculations. And it is a point that every lawyer should be able to understand even without a special knowledge of high-energy physics. To be brief: always put the numbers in the context. When they evaluate the risks of a global destruction, they must assume that such a destruction is "conceivable" to start with. But if it is conceivable, it is also conceivable that the world will die because of other reasons or no reasons. This "background" has to be considered and the "LHC signal" has to be compared with it and with other causes of a doomsday.

Johnson ends up with a "high average risk" of the LHC but he fails to calculate the risk of many other scenarios that could lead to the world destruction. Needless to say, they would give him comparably "substantial" results as the LHC doomsday calculation. The LHC would modify the background risks just a little bit, even if you assumed that Stephen Hawking, Steve Giddings, your humble correspondent, and other physicists are making a stupid error in their analysis of the Hawking radiation, extra dimensions, or black hole cross sections.

The paragraphs above assume that such catastrophes are "conceivable". As long as you use this (unlikely) assumption consistently, you will inevitably see that the LHC doesn't change much about the risks of the global destruction.

However, it's more sensible to use more realistic estimates of the lifetime of the Universe and stars. The Universe is probably much more long-lived than 10 billion years. The catastrophes are therefore always "local". The background risk is much smaller than what we previously considered.

If you perform a consistent calculation in this framework, you will simply end up with P which is much smaller than one septillionth, bringing the risk R down to less than one dollar. Recall that in U.S. English, one septillion is just 10^{24}. When you take just two phenomenological arguments - the absence of observed "collapsed stars" or "collapsed planet" and the survival of the Earth after billions of years of interactions with high-energy cosmic rays - you may essentially multiply the small odds calculated for the Earth and the number of similar "survivors" that we observe and obtain a number smaller than 10^{-24}.

What I would recommend you instead is to calculate what actually happens according to the best and most complete theories we have - with the uncertainties inserted as error margins for various parameters. You will see that the doomsday caused by the LHC is pretty much impossible - much smaller odds than 10^{-24}. The hypothetical black holes do decay - and the probability that they haven't decayed yet decreases exponentially, with a microscopic e-folding time. Even if they didn't decay because of a crazy reason, they wouldn't be able to acquire enough mass in time (before they leave the Earth).

Eric Johnson says that his hypothetical judge should stop the LHC because that would give the lawyers the power to stop many other risky technological enterprises that may become relevant in the future. Well, it's not surprising that a lawyer wants all lawyers to be more powerful.

I think it would be very unfortunate if lawyers and random plaintiffs became able to block the work of others whose importance vastly exceeds the importance of anything the plaintiffs have done in their life just by inventing some speculative "almost infinitely damaging threats". Such a new power of arbitrary complainants would dramatically reduce the people's freedom and the speed of the scientific and technological progress. And by the way, such a new power to block progress would vastly reduce my estimate for the price of this whole civilization.

Of course that I do think that it is important for the courts to be able to stop e.g. a particular kind of biotechnological research that threatens the health of a whole nation if not life on Earth. However, all the risks must always be quantified consistently and put into their proper context. I am completely convinced that even with an imperfect knowledge of the technical issues underlying a similar lawsuit, a consistent approach - and one that calculates the risks not only for the particular "threat" but also for many other, qualitatively similar threats, using the same method - will inevitably show that most of these hypothetical "new threats" are simply not genuine and the lawsuits are frivolous, randomly invented with the purpose to damage someone (or a particular experiment).

So I would emphasize that it's not enough to use a questionable method to determine that the "risk of the doomsday" caused by a particular activity is "substantial". You must check your methodology and, while avoiding biases, use it to calculate the "risk of the doomsday" caused by other activities and processes in the Universe, too. If you obtain a "substantial" probability for many other situations, it strongly suggests that your calculated threats are spurious.

What's really going on is the following: most people believe the myth that the life on Earth inevitably lasts forever. Everything is (or "should be") completely safe, right? So whenever they hear an argument that the activity XY represents a threat, XY is the first example where they have heard that there actually exists a risk. And they will hysterically oppose XY. In reality, life on Earth doesn't last forever. The Earth will become uninhabitable in billions of years or much earlier. And there exist all kinds of natural and man-made "background risks" besides XY that may make the life much shorter. Many of them are much bigger threats than XY. Before someone derives huge conclusions out of one of them, XY (such as the recommended cancellation of a $10 billion experiment), he should have some rough idea about all of them - otherwise his argument simply can't be balanced and rational.

While I am sure that lawyers will always have to rely on the testimonies of the other experts - because they don't really understand the stuff - and they will hear contradictory voices, it's important for the lawyers to demand various checks of the methodologies offered to them. One of the major checks is the calculation of the risk of other phenomena, using the plaintiffs' own methodology. The defendants should never forget to emphasize that methodologies and arguments must be tested, the lawyers should understand this argument, and they should actually be doing these things themselves. The lawyers should be able to understand that statements must be evaluated in the right context.

Sometime in the future, people may be doing things that will bring a significant risk for the life on Earth as a by-product. But we're simply not there yet. The only possible method to destroy the life on Earth that we have found so far is brute force: a sufficient number of nuclear bombs and H-bombs. The risk for life on Earth caused by anything else we have invented is negligible relatively to the "natural background".

And that's the memo.


  1. Hi Lubos,

    surely there must be a minimum mass required for a black hole to be stable.This should, I presume, lie outside the masses of the patrticles used at LHC.

  2. Dear lutzj,

    More precisely, a black hole (except for an extremely charged, "extremal" black hole) is never quite stable. It evaporates.

    But you probably mean a black hole where the evaporation and the growth from the accretion of matter are in balance.

    Such a limit exists and the black hole with this property has far greater energy = mass.c^2 than what can be produced at a collider, regardless of the (not-yet-falsified) scenario with additional dimensions that you choose.

    Such a "balanced" black hole is really "macroscopic", in a sense, but we will surely never produce a macroscopic black hole in colliders and a microscopic one has no chance to survive and grow for long enough time to become a macroscopic one.

    Best wishes