Thursday, November 09, 2006

Lisa Randall on Discover

Today at 2pm the live-chat at discover.com with Prof. Lisa Randall from Harvard took place. It was a very balanced text interview in which Randall answered questions from the audience that were picked by the moderator (Amos Kenigsberg). He did a good job, it was interesting, and a lot of stuff was covered. As always, Lisa Randall was good in communicating her work.

Here are some of the more interesting questions (I wasn't able to copy and paste from the Java applet, so you'll have to endure screenshots). The person with name 'famous' is the moderator.

The first question was from someone called 'Xman':



In the answer I think Lisa Randall was referring to this paper on which I have commented in my post Why do we live in 3+1 dimensions.

As one would expect, there was a question about the LHC, and also one about the so-called string theory backlash:



She was also asked by 'qd_survivor' to comment on science blogging:



Anybody an idea what that guy was surviving? Here is my question, which I think she misunderstood:



(I don't know where these brackets come from, seems to be a software bug).
The question was whether string/loops/spin networks/other funny things out there can eventually turn out to be part of the same fundamental theory. And if mathematical consistency alone makes the fundamental theory unique. I agree on her answer, but it wasn't the answer to my question.

Besides this, I learned that the German version of her book 'Warped Passages' was just published. It seems to be one of the rare cases in which the German title 'Verborgene Universen' (hidden universes) makes sense, and imo is actually better than the English one.

A funny side remark: some time ago I piped a German research proposal through babelfish, which attempted as well to translate my reference list from German to English. The name Randall got 'translated' into Edge of the Universe. (Which is correct: the German word Rand means edge, and All is an expression for Universe.)

Update: Thanks to Georg, I found out that 'qd_survivor' has a post 'Two answers from Lisa Randall' on the blog A Quantum Diaries Survivor, where you also find a complete transcript of the chat.


TAGS: ,

17 comments:

Georg said...

"qd" is probably for "Quantum Diaries". Tommaso Dorigo's blog is called "A Quantum Diaries Survivor".

BTW, do you know if a transcript of the chat will be available online at some point?

Bee said...

Hi Georg,

yes, they said there would be a transcript on their website at some point. If I see it, I will add a link. Thanks also for the info about Tommaso! I should look at his blog. Best,

B.

Arun said...

Bee,

Trying to understand your question.

E.g., as far as I know, Newton's theory of gravitation is mathematically consistent, so mathematical consistency by itself is not enough. We of course know Newton's isn't a theory of everything because we are aware of other phenomena. So the question probably means does agreement with all currently known phenomena + mathematical consistency suffice?

I kind of doubt that we know enough. Then, just like for a certain range of gravitational phenomena, both Newton and Einstein are sufficient, there will be more than one mathematically consistent theory that will fit the observables. Only when we know more will we find that one theory is suitable and other is not. I don't think we know our grain of sand well enough to see the world in it.

I also think that science developed in the first place, by a ruthless focus on the questions for which experimental answers were within grasp. I mean, in some sense, Boyle's law is not very interesting, but Boyle did that rather than thinking uselessly about theories of everything (I'm sure he speculated about all that, but that wasn't his scientific output). And ultimately Boyle in just his one law will have done more science than many string theorists.

I suspect that Einstein onwards, the disregard of the very real frontiers of science in the quest of a theory of everything is going to be an occupational hazard for physicists.

Bee said...

Hi Arun,

I admit that mathematical consistency is very vague in the above context. I should try to come up with a clearer meaning of 'theory'. Regarding Newton: what is the mass that enters there and can it 'consistently' be explained? Best,

B.

Leucipo said...

Indeed at some time Newton does a remark about "two friends" (Barrow and Collins?) suggesting him to delay publication of his full theories until some more throughful mathematical checking. He uses this in the late days as a good excuse about the Newton-Leibnitz controversy. But some researchers have looked into the manuscripts to search for the more heavily edited pages. Proposition 2 of book 1, the law for Angular Momentum preservation, is one of them.

I do not need to remark what units does Planck's constant have, do I?

Bee said...

Hi Leucipio,

thanks. Regarding Planck's constant, I didn't want to go into the discussion how many constants a fundamental theory needs to have. You could also then ask the obvious question: why has the Planck mass the value it has. Not sure though how that would relate to consistency? Best,

B.

Arun said...

Can Newtonian mass be explained? - that is a physical, not a mathematical question, IMO.

Anonymous said...

Hi Arun,

I would agree that mathematical consistency is not a sufficient requirement for a physical theory. Particularly so, if there happen to be other observed phenomena that rule out the theory ... as would be the case for Newtonian gravity. So mathematical consistency and consistency with observed phenomena are necessary conditions. And again I would agree that a theory satisfying these conditions today maybe falsified tomorrow.

A third condition that can be imposed is one of usefulness: it is pointless to say that I have a theory xyz without connecting it to the physical universe.

To paraphrase Bee's question in these terms: does the set of theories describing everything and satisfying mathematical consistency have only one point t1 (which would then be the 'real' theory), or more than one point {t1,t2,...}, only one of which is experimentally (present and future) consistent?

While it would be nice to get the perspectives of smart people who work in physics, how much would Bee trust the answer?

Bee said...

wait wait wait, yes, no, yes, let me step back. My original question was: Do you think there is more than one mathematically consistent theory of everything?

Which imo includes the connection to observed reality (= everything). So far about my question to LR.

Now to what I wrote above, it's not clear to me whether we all mean the same with 'theory' (it's not even clear to me what I mean.) E.g. regarding Arun's question

Can Newtonian mass be explained? - that is a physical, not a mathematical question, IMO.

If you work with Newtonian gravity, you'll have to start with some axioms, but that's not it. You need to say what are the 'things' in your theory. Here you have a set of x(t), R^3 valued functions over R that you name 'particle trajectories' (Since you don't have any object in that theory that you could name 'particle' this doesn't actually mean anything.)

Do you require these functions to be smooth? Differentiable? Singularity free? To address the word 'fundamental', why R^3 and not R^9? Is d a parameter? If it is fixed, what fixes it? If we require some 'modern' theory to explain that, we can't just drop that question for Newtionian gravity.

Now to come back to my problem with the mass: what is the mass in Newtonian gravity. Is it a scalar? Is it the same for all x(t)? Is it a function? (\delta on x(t) ?). What is it?

To make my problem here more transparent let me ask the question differently: Newtonian gravity predicts that particle trajectories can meet. Since massive objects tend to attract each other this is the generic case. What happens then? Does the theory say anything about it? Esp: what happens to m in that case? Do you add it up? If so, why? Is Newtonian gravity sufficient to explain this? How can you possibly say anything about that without having a picture of interactions? Without knowing where the mass comes from? How may ways are there to come up with an interaction. A mechanism of mass generation? Consistently?

it is pointless to say that I have a theory xyz without connecting it to the physical universe.

The question was whether that's possible at all. Or whether we'll all end up in the same place if we follow different approaches very stubbornly.

To paraphrase Bee's question in these terms: does the set of theories describing everything and satisfying mathematical consistency have only one point t1 (which would then be the 'real' theory), or more than one point {t1,t2,...}, only one of which is experimentally (present and future) consistent?

Yeah... right, one could say so. Just that it sounds suspiciously landscapy if you put it this way.

While it would be nice to get the perspectives of smart people who work in physics, how much would Bee trust the answer?

Well. My question wasn't 'Is there ....', but 'Do you think there is.... How much I would or wouldn't trust in Lisa Randall's opinion is a completely different issue. The reason why I asked is that I'd find it a good idea if there was a stronger interaction between different approaches to the holy grail of TOE, because they'd actually turn out to be part of the same theory.

Best,

B.

Arun said...

Bee,

In Newtonian gravity, presumably point bodies simply pass through each other :). Mathematical consistency is not enough to figure out that that is not what happens in the real world. Mathematically, I will have to show you how to integrate the equation of motion through one of these collisions where the gravitational potential blows up. I'm reasonable certain it can all be worked out.

To come to the question, why R^3 rather than R^9 and a fundamental theory's "responsibility" to explain it - no, only in a theory which lives in something other than R^3 is it necessary to explain why we appear to live in a R^3.

Otherwise, there are an infinity of questions about why various mathematical objects do not apply to the physical world that no theory can ever answer. E.g., why doesn't the Newtonian world exhibit supersymmetry? Poincare invariance? etc., Why is the axiom of choice valid in this physical theory? etc. etc. Why isn't the monster group a global symmetry?... A theory only has an obligation to explain objects within the theory, and not those outside. Otherwise, we have had absolutely zero successful physical theories.

Anonymous said...


Which imo includes the connection to observed reality (= everything). So far about my question to LR.



Aha, so your question was "can there be two apparently inequivalent, completely valid descriptions of everything?"

I say apparently because because if they predict the same thing in every situation there does not seem to be a disagreement between them, if they do not then at least one of them can be falsified in principle.
So, the in-equivalence would be in formulation, where the equations would look different, but give the same results when applied to a physical problem. An analogy in well-known physics would be working in different gauges, without knowing about gauge transformations.




To come to the question, why R^3 rather than R^9 and a fundamental theory's "responsibility" to explain it - no, only in a theory which lives in something other than R^3 is it necessary to explain why we appear to live in a R^3.


I don't agree. While it may be OK to have a theory which takes this as an axiom; I don't think the responsibility changes with the dimensions it proposes. Yes, you could ask for experimental proof of *whether* it really lives in R9. But the question of why is really at the same level for both R3 and R9.

Anonymous said...

Which imo includes the connection to observed reality (= everything). So far about my question to LR.



Aha, so your question was "can there be two apparently inequivalent, completely valid descriptions of everything?"



Sorry I forgot to ask in the post: I seem to remember that Smolin was an active supporter of this kind of view ... that string theory could be probing Quantum gravity perturbatively while LQG would be doing so non-perturbatively. So, these 'different' theories are in fact complementary parts of the full theory.

Is my memory simply wrong, or is he still saying this?

Bee said...

Hi Arun,

I've already agreed that mathematical consistency isn't sufficient. E.g. it seems to me mathematicians don't really have a problem with singularities, whereas I would consider them as 'unphysical', and say a 'consistent' theory better not run into any.

Mathematically, I will have to show you how to integrate the equation of motion through one of these collisions where the gravitational potential blows up. I'm reasonable certain it can all be worked out.

I'm reasonably certain that you can come up with a recipe, but you'll run into a problem with deterministic evolution there. If the force acting on the particle diverges, the theory 'forgets' about the initial state. E.g. put a particle at rest at distance R to the attractor, and let it plunge into the (singular) potential. If it hits the singularity it doesn't matter at what R it was at rest. So, if you want it to pass through the singularity, how can you do that and not spoil determinism?

To come to the question, why R^3 rather than R^9 and a fundamental theory's "responsibility" to explain it - no, only in a theory which lives in something other than R^3 is it necessary to explain why we appear to live in a R^3.

Well, I am not sure what you mean with a theory being 'responsible' for something, so let me ask you what does a theory have to explain in order to be considered as 'fundamental'? In Newtonian gravity, the number of space-dimensions being 3 is another axiom of the theory.

I agree here with what Anonymous said above.

Best,

B.

Bee said...

Hi Anonymous,

Yes. If two theories make exactly the same predictions I would consider them as being equivalent. As I wrote above, I agree on what Lisa Randall said regarding this point. However, in this case it should be possible to show an equivalence between both approaches (or, to use a more in-fashion word: a duality). Best,

B.

Bee said...

Hi Anonymous,

I seem to remember that Smolin was an active supporter of this kind of view ... that string theory could be probing Quantum gravity perturbatively while LQG would be doing so non-perturbatively. So, these 'different' theories are in fact complementary parts of the full theory.

Is my memory simply wrong, or is he still saying this?


I think your memory is right. But as to his current opinion: please ask him and not me. Best,

B.

Bee said...

interestingly I just noticed that Lubos reproduced my 'funny side remark' -- obviously without linking here. well, in case you are wondering who copied from whom, ask yourself what reason he possible could have had to translate 'Randall' from German to English.

arivero said...

(hello, here A.Rivero pos(t)ing as the blogger identity Leucipo)

About Newton consistency, it has been a lot time ago but I think I got the remarks from an article of Mordecai Feingold or some other Newton scholar. I am sure it was an article, not a book. The argument was on one side, of course, about the nature of infinitesimals (or "atoms", as Cavalieri calls them using the latin word instead of the greek one). But then it was argued that other problem was about circularity of definitions when introducing time, ie about the fact that the parameter of a curve is a geometrical thing, no Time, but it must be interpreted as geometrised time. The preservation of angular momentum lets him to get out of the problem, because then he can map time intervals to geometric objects (chunks of area). It is a strange explanation if you think how Newton tends generaly to ignore Kepler work.

About mass I think that Newton climbs of the shoulders of Hooke, but still it is interesting to remark that it is a proportionality constant attached to a fluxed quantity, acceleration.

Now I think about, the area law hold even with a varying mass, does it? It should not hold with a tensorial quantity, but with a scalar the "effective force" is still a central force.