The first full paragraph on page 4 of Aaron’s review starts with

It is true [..] that there really is no such thing as ‘string theory’. It is, rather,
a collection of partial theories and calculational techniques bound by physical intuition and conjecture.

I was surprised when I read this first. A “collection of partial theories bound by intuition and conjecture” does not sound like anything which would allow tests like described in the next sentence:

[…] this skein is remarkably robust. Calculations
that have the possibility of destroying this structure invariably turn out to reinforce it.

Of course I immediately agree that the study of string theory involves a lot of partial theories, the links of which are often conjectural.

But I think it is important to point out that there is, apart from - and above - intuition and conjecture a binding force, which distinguishes the study of “string theory” from the study of a subjective menu of partial ideas that happen not to contradict each other.

Compare with some other field of physics. Say classical mechanics. There is a unifying essence, namely Newton’s laws, or Hamilton’s equations, if you like. People want to study this.

But only in the most simplest of situations can one actually practically study this essence itself. Most interesting phenomena in the world are - or at least have been for a considerable time of history - not directly accessibly using the fundamental equations. Often, they require inventing “new theories”, like thermodynamics, theory of friction, theory of fluid dynamics.

Today most of these “new theories” can be derived pretty well from first principles. But for various periods of history, scientists believed in the correctness of Newton’s laws but had to accompany them nevertheless by a “collection of partial theories and calculational techniques bound by physical intuition and conjecture”. Bound - that is - among each other, but, most importantly, bound by conjecture and intuition to the essence that they are thought to be derivable from in principle, namely some microscopic dynamics following Newton’s laws, or maybe following the laws of quantum mechanics, if that’s necessary.

If I remember correctly, there are aspects of something as ordinary as friction (certain special sorts of friction for certain materials) which even today remain puzzling, in that it is not clear precisely how they follow from a microscopic description.

The same case could be made for particle physics. While it is believed that the essence of particle physics is the study of the standard model, it is a fact of life that for describing most interesting phenomena, notably bound states of nucleons, one needs to resort to “theories ” (models) that are bound to the standard model only by intuition and conjecture. We expect them to be derivable from the essential concept itself, but we do not know how to do it.

I am emphasiziing this point because on the blogosphere one could frequently here various layman ask rethorically for “the equations” of string theory, implying precisely that the whole thing is made up from subjective choices without really being a “theory” which you could hand somebody with the words “this is what I mean by string theory, study this”. Like, for instance, you could do with classical gravity by handing somebody the Einstein-Hilbert functional.

For this reason it might be worthwhile to state clearly in a popular account like the above review is, what precisely it is in essence that people are studying when they are studying string theory, notwithstanding that this essence may incarnate itself in various different guises.

And - please correct me if you think I am wrong - the essence is this:

The study of string theory is the study of quantum theories that have a perturbative expansion which is given by a loop expansion in terms of Feynman-2-graphs computed using 2-dimensional gravity.

Whatever else you do, if it is not bound, maybe just by conjecture or intuition, to this essential principle, you will hardly be doing string theory.

For instance one can go ahead and study linearly-extended objects governed by an action principle involving a BF-theory ($\to $). Working out the dynamics, one finds that it does indeed describe propagation of string-like objects.

In a world where “string theory” would be just anything bound by intuition and conjecture, one could happily go ahead and declare that this particular theory is one of the many that constitute “string theory”.

But that’s *not* what happens. The dynamics of the strings in these BF-theories is not governed by Feynman-2-graphs with amplitudes governed by worldsheet gravity. So it’s something else, not “string theory”.

Of course it could happen, and I would be positively intrigued if it were true, that these BF-theory strings are related to one of the satellite theories that are indeed, slightly by conjecture and intuition, related to essential string theory proper: namely to the theory of D-branes and D-strings in particular.

With some “string theory” papers on DBI-actions and things like that, describing D-brane dynamics, one could get the impression that the theory investigated in these papers is just yet another random theory, that we happen to like to add into our “string theory” portfolio.

But of course this is not a matter of our free choice. The DBI-action, while it can be studied just by itself, is, by more or less robust arguments, believed to be *derivable* from the assumption that there is a theory whose perturbative expansion is given by a string theory S-matrix.

The same applies to all other topics studied in string theory. For instance, people are very interested in extensions of classical gravity which involve fields known as RR-fields. (Of course Aaron and Jacques know all this, much better than I do. I am just stating it for the record and for the general reader)

This is curious, as precisely for those fields it is *not* known how to relate such theories to a worldsheet description. But it is strongly believed that, while not known at the moment, it is possible in principle. Hence the “theory of RR-backgrounds” is related to the main body of string theory, partly by conjecture and intuition.

If, however, there were no indication at all that RR-backgrounds arise in quantum theories that have perturbatiuve expansion involving a string theory S-matrix, then we would not consider RR-backgrounds part of string theory. Would we?

Similar comments apply to Matrix Theory, AdS/CFT and so on.

I apologize for having probably emphasized the obvious. It just struck me as an important point in the discussion of the raison d’être of string theory.

## Re: Not Even Wrong

Very nice review, fair and balanced.

One quip, the SuSy calculations on the CC while they do cut the problem in half are rather troubling, b/c in principle there is absolutely nothing wrong with the calculation. They seem robust! Whereas for the vanilla SM such a calculation is rather unnatural and one could speculate about all sorts of ways out of it. 120 orders of magnitude or 60 orders of magnitude is still (at least to my mind) god returning ‘error, man made mistake’

I think that is what Dr Woit was reffering too -shrug-