## Monday, September 25, 2006 ... //

### Do the laws of nature last forever?

Several people have informed me about an article in New Socialist

The author seems confused what is science, what is physics, what is a law of nature, how it looks like and how it may look like, and why we believe such laws. He is profoundly confused at many levels: basic levels as well as the technical ones.

Cosmological natural selection

Lee Smolin promotes his cosmological natural selection. Just during the last month, five independent people have mentioned this issue in discussions with me or in their own articles; the list included famous names like L.S. or A.V. All of them are convinced that it is trivial to falsify Smolin's hypothesis and it has, in fact, been done immediately when Smolin proposed it.

A decade ago, Smolin had conjectured that the laws of our universe are optimized for black hole production because every new black hole is a new baby whose properties are similar to the parent universe but it is not quite identical because there is also a cosmological mutation going on. The most prolific universes - those who create many black holes - are going to dominate the ensemble of the universes. Lee Smolin has written a whole book whose content is isomorphic to this paragraph.

It is easy to see that if you change some parameters in our universe, for example if you reduce the hierarchy between the electroweak scale and the Planck scale, many more black holes will be created. The theory is dead. Trivially dead. Period. Why does Smolin revive this nonsense all the time, without having any new arguments or mechanisms? Does a lie become the truth when it is repeated 100 times?

Smolin also misleadingly suggests that he is behind the word "landscape" in theoretical physics even though the only scientifically plausible meaning of the word in theoretical physics was explained by Leonard Susskind.

Why do we believe that the world is governed by laws

The main reason why we believe that the universe obeys laws is that the laws we have found are the simplest satisfactory explanation of the experiments and observations that we have done. Lee Smolin doesn't seem to mention this technical detail. Instead, he focuses on battles between philosophy, religion, and science that should place no role in science whatsoever.

Evolving laws vs. evolving knowledge

Lee Smolin clearly fails to distinguish the evolution of the laws themselves and the evolution of our knowledge about them. He claims that Einstein's discovery that the geometry of the real space is not flat is "another example" of evolving laws of nature. In reality, of course, Einstein's equations had the same form and applied to the motion of cellestial bodies long before Einstein realized that. The laws are not changing.

What does it mean for the laws to change?

What does a physicist mean by the basic laws of nature? He or she means the most fundamental possible mathematical reasons or rules that predict or imply a class of observed phenomena or all other phenomena. If the laws were evolving, they would not really be laws. If the question whether it is fine to kill other people had an evolving answer, there would be no law about the murder. Of course, at a longer time scale or length scale, this social law indeed doesn't exist. But it exists within an effective theory.

Physics is more fundamental than sociology and its laws are thus more lasting, too.

A physicist always picks the most fundamental law she can pick. If the laws were evolving significantly but if they were still accessible to science, the primary thing that a physicist would be interested in would be the laws that govern the evolution of the "simpler" laws. No doubt, we would call these rules "laws" again, even though Lee Smolin tends to call them "metalaws". The prefix "meta-" only means that the things are perhaps getting too complicated for Lee. But it doesn't mean that they're getting too complicated for everyone else.

In his recent book, he also incorrectly uses the word "meta-theory" for string/M-theory. In reality, the word "meta-theory" could have been used a few times at the beginning of the duality revolution but no one would use this terminology today. String/M-theory is simply a theory. It is impossible to divide it to pieces.

Even though most crackpots are unable to understand this simple fact, string/M-theory is the best textbook example what a theory in physics means. It is a logically coherent structure including a finite number of concepts and a finite number of equations and other mathematical rules that can be used to predict the outcome of many experiments. I say that string/M-theory is the best representative of the word "theory" we have in science because it is the most complete and the most logically coherent description of the widest possible set of phenomena that we have ever had, namely all of them.

Technically, by a theory, we mean a choice of the Hilbert space and/or basic degrees of freedom together with rules to determine the dynamics - the Hamiltonian, the action, or more general rules to calculate correlators or the S-matrix.

The word "theory" does not require the system of concepts to be already proven experimentally. We use the word "theory" for theories that are not yet proven or that are unlikely to be ever proven (little Higgs theory?) as well as for theories that have already been falsified (Glashow's old SU(2) theory) although the word "model" is often a substitute, especially for more concrete theories that have many conceivable "siblings".

In science, we couldn't use the word "theory" just for one of these options - a correct or a wrong theory - because the validity of any sufficiently interesting theory we discuss or investigate has yet to be determined and this fact would make the word "theory" unusable in most situations: another point that the "critics of science" completely misunderstand. Virtually neither of the "Not Even Wrong" crackpots understands that if we already knew for sure that a theory is correct, we wouldn't be developing it anymore - we would only be looking for its other consequences and we would be moving to a more profound theory.

The theory that most of the theoretical physicists, especially the cutting-edge theoretical physicists, work on at a given moment of time is necessarily an unproven theory, essentially by definition. If a theory is already proven, then it is not at the cutting edge.

Meaning of the evolution in the actual theories we have

In the state-of-the-art theories, we know exactly what it means for the theories to evolve in time. For quantum field theories, it means to change their relevant and marginal parameters into functions of the cosmological time: the masses of elementary particles and the renormalizable couplings may depend on time. We know that such a possible evolution is severely constrained by observations. Although there are a few controversies, it is fair to say that it seems that there exists no such evolution.

We shouldn't be thinking about creating new degrees of freedom because a transition from a theory with some degrees of freedom to a theory with other degrees of freedom would prove that at least one of them had to be incomplete at the mass scale of the new degrees of freedom, and thus a more complete theory was needed. Alternatively, such a transition would be completely discontinuous and it wouldn't allow us to use the knowledge about the previous regime to learn anything about the new regime.

In string theory, it is completely impossible to change the laws of nature as a function of time because string theory has no parameters to be adjusted. There is only one unique and eternal mathematical structure called string theory and it cannot be changed or contaminated. It can be twisted but the twisted version is not a realistic theory. :-)

Any evolution is governed by the laws of string theory. This is why string theory can be used to prove that spacetime topology can change, among other things. The laws of string theory don't break down and don't have to be - and cannot be - changed even when an extreme effect such as topology change occurs.

If the laws of a theory needed to change in order to describe a certain transition, we would have a proof that the theory is an incomplete description of reality.

Needless to say, this is not the case in string theory. String theory is a complete description of reality even though we don't understand its predictions in some extreme situations, especially those that have something to do with the ultratiny expanding universe. Can we really tunnel into another vacuum and which observables we should talk about in this setup? Although people assume that an answer similar to the answer of effective field theory with a "common sense" discontinuity is the right one, we are not guaranteed that we have fully understood all implications of string theory for this situation.

But at any rate, whatever the allowed observables, rules, questions, and transitions in this context are, they are a part of string theory and we are not allowed to mess up with these laws by inventing some meta-laws or laws saying how the previous laws should change.

Mathematical character of laws

In any theory that remotely resembles the theories that have been successfully used to describe reality for centuries, the most fundamental laws - even if some people would like to call them meta-laws - are given by systems of mathematical equations constraining certain mathematical structures and quantities. If this were not the case, we couldn't describe the universe quantitatively.

Lee obviously disagrees with the previous paragraph. In order to show how intensely he disagrees with the basic thesis of theoretical physics that the world is based on mathematical laws, he even quotes Roberto Unger, a Harvard philosopher, who has called the eternal mathematical laws thought to be relevant for physics "a poisoned gift of mathematics to physics". Wow.

Lee feels that physics was consistently getting rid of time and he suggests that it could be a good idea to give time the same role it played before Galileo. If the laws are evolving, we would need to return before Galileo. In other words, if his article reflects what Lee thinks, he wants to return us to the age of the Inquisition in this respect, too. Very nice.

This is the physicist who is spamming us with the nonsensical comments that we should be doing physics without a pre-existing spacetime; on the other hand, the very fact that we have permanent laws that are valid at all times - which has been the case at least for 300 years - is already too bad for him. Do you think it is a consistent approach to try to kill time where it seems almost necessary and restore its key role in contexts where it has been impossible for 300+ years because the eternal laws work so well?

In philosophy, one could spend several centuries by thinking about the evolving laws of nature. Needless to say, it would lead to the ballpark of the same realm as most other investigations in philosophy, namely the realm of nowhere. In physics, we have very different methods. According to these methods, the concept of the basic laws that evolve in time is an ill-defined concept because we can't really define any global "time" coordinate (because of general relativity and other deep insights) and because we can't define what it means for the laws to evolve unless we have other laws that don't evolve. It also seems to be a useless idea that doesn't help us to explain anything that we know about the universe, and therefore its weight in physics is tiny even though it can still generate several pages in New Socialist.

And that's the memo.