Monday 29 September 2014

"Black Holes don't exist:" giving context to sensational science news

A friend of mine recently pointed me to this article about how "Black Holes don't exist." The article concerns two recent papers

http://arxiv.org/abs/1406.1525
http://arxiv.org/abs/arXiv:1409.1837

Since it came out the following edit has been added to the article:

"Due to some confusion, we feel it is important to clarify. The notable word in the title is “mathematically.” In science, there are conflicting predictions that come from different theories, assumptions, and equations–different equations result in different outcomes and different proofs. In short, one set of assumptions leads down one path and give us new (potentially important) things to consider. But there are many paths. It seems that many people were not sure how to situate or read these findings. Hopefully, this clarifies things. We’d like to apologize to anyone who took this out of context or who was confused by the implications. In the coming days and weeks, more physicists will weigh in with their findings. Things will update as they develop. Science on."

And this edit is actually exactly why I wanted to discuss the article. What does a claim like this mean, and how can a non-expert interpret it? The following is basically word-for-word what I wrote in reply to my friend who sent me the article.

I skipped reading the article at first [I have since read it, and you don't learn all that much], as lots of articles like it miss the point with physics. People love to say "Einstein was wrong!" or basically "[New sensational thing about physics hopefully with the word "Quantum" somewhere!]" while not appreciating what is really going on. And it annoys me. A lot. Anyway, onwards with constructive things...

I'm no expert, so I'll be arguing from authority, basically. First things first, Mersini-Houghton, the author, appears a totally respectable physicist who has published highly cited work on a variety of topics in the past, and works at a respectable institution. I'm in no way trying to slander her or her work. It's completely within the confines of work on these topics (black holes, the information paradox), as far as I can tell. Her recent work on these topics isn't highly cited (so far: and that is only a few months, but physics moves fast these days), but citations aren't a perfect gauge of a work's relevance. The important point to take away is: the author is certainly no two-bit crank posting on vixra (I hope I don't get bombarded by any more cranks than usual, but just look at the kind of stuff that goes on there! If you've never lost a few hours on vixra, I highly recommend it.)

What I want to discuss here is how the paper was reported, and more importantly how non-physicists should read reporting on physics in general. It's all about context, and what we mean by "exist" I think...

The paper referred to in the article was published in a respectable journal, so it went through peer review and someone with more knowledge in the field than me thought it was correct. However, let's put it in context. There has been a whole lot of interest in issues related to black holes recently, thanks to a famous paper on "firewalls." (http://en.wikipedia.org/wiki/Firewall_(physics)) There seems to be something we don't understand about black holes in that they lead to a paradox that requires giving up one of three major pillars of physics (three according to wikipedia, I thought it was just two: unitarity and locality, but there ya go). The name "firewall" comes from the easy way out: there is a wall in the way that prevents anyone seeing the paradox. I'm not sure how seriously the firewall itself is taken. I think of it as a last resort to sweep things under the rug. But like I said, I'm no expert.

This debate over what to do with firewalls has led to a huge number of papers proposing different resolutions to the problem. To give that a number, the original firewalls paper has gotten 249 citations in the 2 years since it was published. Experimental papers, and confirmed theory papers, get more than that. But for a "pure thought" paper, that is astounding.

(Although, the original AdS/CFT paper, a paper on pure thought depending on your take on AdS/CFT applications, just became INSPIRE's most cited paper. It took over the model of leptons, which is manifestly about the real world, but there ya go, that's physics!)

Many people have said we may need to do away with black holes in one way or another. For example, Hawking argued that there is no paradox if black holes aren't "eternal." Another solution is that black holes are really fuzzy quantum or turbulent things, and that makes the calculation that led to the paradox incorrect.

Okay, enough context. Mersini-Houghton's idea is one of these many that say black holes never form. The idea is that when a star collapses on the way to forming a black hole the "Hawking radiation" kicks in causing a pressure that stops collapse before a true black hole forms. If it never forms, there is a never a paradox. In her second paper out this month she works numerically with collaborator to show this happens in realistic situations (for example breaking the assumption of spherical symmetry).

So, is Mersini-Houghton correct? Her first paper came out in June and hasn't really been picked up in the community. The excitement over firewalls has died down, so maybe it's that. But if she was correct, it would be enough to set people off. It hasn't, so I judge there must be something about it that isn't compelling. Maybe she made some simplifying assumptions that people believe would make her argument wrong in a realistic situation? I'd like to ask an expert. The new numerical work makes me think she is correct, within the parameters she's set herself at least. It's whether those parameters are right.

But do I think it means anything? I'm inclined to say "no" for the following reasons:

1) It appears that we do see black holes out there in space. I'm not familiar with the observations, but as I understand it, it is an established fact. Maybe Mersini-Houghton can get around it, and her system still forms these things, but they aren't "formally" black holes because they miss the little paradoxy bit. Then, for all practical purposes (i.e. in astrophysics), it is a case of walking and quacking like a duck, but without the particular nuanced existential consequences. Practically, then, "black holes" still exist. This has to be true of all the other firewall solutions. The black holes are still there, because we see them, but some little bit in the middle is subtly different.

[edit: the Event Horizon Telescope hopes to directly image black hole event horizons in the coming years. Things like "Saggitarius A*" are pretty good candidates for them. Any other comments about direct evidence for black holes are very welcome!]

Now you have to go out and find an observation that can confirm that. People are trying, but it ain't easy (see this interesting proposal to tell proposed firewall solutions apart observationally using gravitational lensing!). An example discussing Mersini-Houghton's work is that the "bounce" of the star instead of forming a black hole that she predicts could be the source of some high-energy cosmic-ray type things (fast radio bursts in this case). We see stuff like that, and by looking at them in detail maybe you can see what they came from. But this is messy astrophysics, and the number of explanations for these things is often very wide.
I'd like to see whether (i) Mersini-Houghton and co can still make real things that look enough like black holes that they are consistent with what we have seen already, and (ii) if they can make any novel observational prediction to test their theory. If the answer to either of these questions is "no" then they are dead in the water.

2) On a purely philosophical level, Mersini-Houghton's solution doesn't solve the firewall paradox in my opinion. It may solve it "in real life" if there are no collapsing stars that form tricky black holes. But it doesn't solve the problem for theory. In theoretical physics you can still set up a "thought experiment" and if that makes your theory inconsistent then you are in trouble. The whole firewall thing began with a thought experiment. Black holes are solutions to Einstein's theory, as long as they are you can always imagine one just sitting there. Conjure it out of nowhere. It doesn't have to be made by a collapsing star (because in the thought experiment you also conjured the star from nowhere too).
Mersini-Houghton's solution doesn't alter Einstein's theory or quantum mechanics, and so you can still do that thought experiment, and the firewall problem is still there.

So, Mersini-Houghton hasn't solved that problem in my opinion. And indeed, black holes still "exist" in the theory, so for many theorists they are still just as relevant as thought experiment probes of whether we really understand the universe. And many theorists are platonists anyway, so black holes, even of the thought experiment kind, do "exist."

3) Outside of thought experiment, as long as black holes are still this "formal" solution of Einstein's equations (and they are in Mersini-Houghton's theory as far as I can tell) then they do still *really* exist. This is thanks to quantum mechanics, where things are allowed to "pop in and out of the vacuum" (http://en.wikipedia.org/wiki/Virtual_particle). Black holes do the same thing in quantum gravity, at least in string theory they do as far I know (I've heard people argue that they needn't in other theories, but I don't find these arguments compelling: in quantum theory you need a very good reason for things not to happen, or else they do). So, black holes still exist at the quantum level even in her theory.

Does that have any relevance to real life? Well, virtual particles do. We have very strong evidence that virtual particles are important. They have observable effects on particle collisions. They are the reason the electron has the charge and magnetic properties it does, so probably phones and microchips and such wouldn't work without them. (Does anyone have a better example of the reality of virtual particles? I mean, an accessible one, not just "electroweak precision observables, duh")

What about virtual black holes? (i) I don't know if they suffer form the firewall paradox, because of the popping quantum business, so maybe they aren't a problem. (ii) They are much harder to see. They are intrinsically quantum gravity things, and that is, as far as we know, not relevant to anything we have a hope of measuring on earth. But there is hope, and this is exactly why I study cosmology: the early universe is a lab for very high energy things, and we can hope for signals of quantum gravity in the sky (in fact this is a lot of what Mersini-Houghton did in her earlier career, too).

So, in summary:

* There are lots of theories like this. Maybe this one's right, maybe it isn't.

* There are at least "black hole-like" things out there in space that we have seen.

* Philosophically, black holes-proper still exist in this theory thanks to quantum mechanics.

* We need to come up with observational and experimental tests and consequences of all this. 

Thursday 11 September 2014

The "Yes! And?" of science.

I personally believe that the academic "brand" of Impostor Syndrome (IS) is particularly tricky to deal with because underlying it is a certain type of arrogance. It took me quite a lot of time with a coach (thanks to the wonderful SupporTED program I participate in through the TED Fellowship) to realise that I really was arrogant in my Impostor Syndrome: anyone could say what they like about me being talented, but I was holding onto the belief that I the only person qualified to make judgements about myself. So with a slight of hand, I can disregard your positive statement. Easy Peasy. My coach had to bring out the Logical Data Big Guns to deal with me, but she did so, wonderfully. She showed me this internally arrogant attitude was seriously flawed. My data analysis software, my ability to process external feedback, was broken. I realised that I was rejecting data points based on my faulty Bayesian prior, and then refusing to quote the prior when making statistical inference. I know! I know! Bayes would be rolling in his grave! I was shocked, and chose to rename the problem Self-Data Malfunction.

So, if you know that this part of yourself is faulty and you want to repair it, what can you do? Well this is all happening subconsciously to some degree, so it isn't a case of just hearing and accepting the opinions of others. If it were that easy, I would have done it already!

When talking about the issue with people, I often heard a phrase that I realise was intended to be helpful, but to me expressed exactly the wrong idea: "fake it ‘til you make it."

The idea is that even if you don't feel worthy to be in your job, position of authority or degree program, you just "fake it" and act like you are worthy until some time later you realise, hey - you are in fact the woman who deserves to be there! And there are lots of strategies online and in books to help you build up the skills to "fake it".

But this just hit right to the core of my Self-Data Malfunction. If I was "faking it" at all, surely there must be some truth in my "you don't belong here" Bayesian prior? So then maybe my self-data analysis software was right after all! Cue the spiral of non-productive thinking.

And then I remembered a wonderful thing I've learned from doing improvisational comedy (which, by the way, I highly recommend - it's like emotional version of walking in traffic: all the excitement, none of the physical harm). The improve technique is the principle of "Yes! And?"

Here is how it goes.

Say you’re doing an improv scene with someone on stage and they suggest something, like they are your long-lost sister, or the floor you are walking on just happens to be made of fire, rather than rejecting it outright for being crazy (as these improvised suggestions often are), you imagine and accept the universe they've just created. You say "Yes!" to the idea. And? Then you run with it!

The "And?" part means that you build on it and immerse yourself in it. That often involves justifying the suggestion they just made - making it work within the context of the scene and your established characters. And then, ‘hey presto!’, you're doing improv.

When I was thinking about the Self-Data Malfunction, I realised that rather than faking it 'til I make it, I can "Yes! And?" my life in academia. It is incredible what that subtle change in emphasis did for my outlook on academic life.

So, what happens when you find yourself on the shortlist for a job you didn't think you could possibly get? You say yes! And? Go give a great talk/interview! You now live in the reality where you are a viable and attractive candidate for the position. Yes! And?

What about when you think you aren't good enough at writing this code, doing this derivation, finishing this paper? You remember that yes, you already got here, and you have skills that will enable you to tackle the task. And... then you go and smash it!

What happens when you are invited to submit that review paper or chapter and you feel like they may have asked the wrong person by mistake? You remember that yes, you have interesting things to contribute. And you now live in the universe where people want to hear/read them.

And what happens when someone like me wants to write about impostor syndrome, but there have already been great posts by incredibly smart, talented and accomplished men and women (for example John John’s post, Amy Cuddy's post on body language and how it can change your life, Ed Bertschinger's post on his own struggle) on the subject? What if I don’t yet have a faculty post, and the authority that comes with that to be able to write about impostor syndrome without fear of the affect it may have on people’s perception of me?

I remember that yes, I think I have something new to add to the mix, and then I remember that as a graduate student and postdoc I would love to hear from someone who wasn’t so accomplished or high up the academic ladder to tell me about things they’ve learned and are dealing with. And so I write this here blog post!

Does it mean you will always then succeed at things? Definitely not. I imagine your rate of success may be exactly the same as before. But your rate of trying new things, and putting yourself out there and taking risks will definitely improve, and with more opportunities come more chances to do an awesome job and succeed. And as we all know, it's all about statistics really.

It isn't easy to do all the time. The "No, but" voices are much more skilled and generally shriek banshee-like in my head, but this feels to me like a much more holistic way of enabling me to live and grow into my career and my life. The change is slow, but what I find happens is that I start to really enjoy new challenges and scary things, not because I’m trying to prove myself, but because I enjoy taking that journey to the “and” part of myself and find that it isn’t so crazy a world in the first place.

So…. Yes! And?