More on New Scientist

I felt sort of bad saying all that stuff about Wired when the guy who wrote the piece I did like showed up to say “Thanks for the link.” But hey, I’m not going to stop criticizing bad science reporting, nor can I imagine shutting myself up about the practices which I think cause bad science journalism. (Nor do I have the vanity to think that by myself, I’ll make any difference.) I’d feel considerably more uncomfortable if Greg Egan didn’t go and provide a whole new plateful of reasons to be upset with pop science.

Egan has been masochistically plowing through New Scientist ever since the EmDrive incident, when he had found himself “gobsmacked by the level of scientific illiteracy” the magazine had put on display. Now, commenting at The n-Category Café, he gives two additional recent “absurdities.”

• an article on Boltzmann brains (full article for subscribers only) which insists that the correct way to estimate cosmological parameters is to take into account the fact that if there were a large number of such conscious beings in the far future, it would â€œviolate the Copernican principleâ€ by making non-Boltzmann brains like us atypical;
• a letter from a reader which suggests that maybe quantum entanglement is just like two twins leaving home with different-coloured wallets: if you see one twinâ€™s wallet, you know the colour of the other wallet. The reader asks â€œAm I missing something?â€ â€¦ and the magazineâ€™s editor responds â€œNo, itâ€™s exactly rightâ€. So much for QM violating Bellâ€™s Theorem.

A “Boltzmann brain” is a random fluctuation in a set of particles which happens to create a conscious brain (out of neurons or transistors or anything else). Somewhat like the case of Last Thursdayism, or a really advanced replicant, this brain has hallucinatory memories of its education as a scientist, the Hubble Deep Field images, the WMAP data and so forth. The fluctuation lasts just long enough for the brain to say, “Hey, I’m conscious!” before sinking back into chaos. A small fluctuation, just large enough to create such a brain, seems much more likely than a fluctuation big enough to create a whole Universe.

This is all fun to speculate about, but it’s a long way from the claim advanced in the New Scientist piece spotted by Greg Egan. The concepts involved are unsettlingly far removed from direct testability: for example, per Carroll and Chen, it matters whether the Big Bang was the beginning of everything or merely of our local patch of a much larger universe. I’d say that it would be foolish to “insist” upon anything right now, particularly when the claim involves something as vague as the “Copernican Principle” (which is the assertion that our current situation is most likely not “exceptional” in any way).

[UPDATE: For the point I should have made but didn’t, see the comments below.]

The second item is, to me, much more “gobsmacking” than the first. It’s one thing to oversimplify an idea on the edge of cosmology, or to give too much weight to one possibility among many; it’s much worse to mangle a central concept in a well-established area of physics. The situation described by the letter-writer and endorsed by the editor is, most emphatically, not quantum entanglement. It completely misses the point of entanglement and describes instead something entirely unremarkable. A person writing a letter to a magazine doesn’t have to know this; after all, they’re writing in the hope of receiving information. For the editor to get this wrong is downright reprehensible.

Here’s a quick primer on the whole “entanglement” thing:

Alice and Bob are playing a cooperative game. They each receive a stream of random bits: for example, each of them can flip a coin and get either heads (0) or tails (1). Neither can see the other’s coin, and the coins are fair. At each turn, Alice and Bob each read a bit and decide whether or not to raise their hand. A raised hand is an “output” of 1, and a hand kept low is an output of 0. The goal of the game is to have exactly one player raise a hand, and only when both input bits are 1. They don’t have to win every turn, but we’d like them to win as many turns as possible. What strategy can they choose ahead of time in order to maximize their chance of success?

We can state this more succinctly with a little notation, breaking out the Boolean operations AND and XOR. Let $$I_A$$ and $$I_B$$ be the input bits for Alice and Bob respectively, and call their output bits $$O_A$$ and $$O_B.$$ We want a strategy by which

$$O_A \hbox{ XOR } O_B = I_A \hbox{ AND } I_B.$$

What’s a way in which Alice and Bob can win 75% of the time? Yes: never raise a hand! If both players always output 0, then they will only lose in the 25% of turns when both coins happen to come up 1.

It turns out that this is actually the best possible strategy. As long as the players can’t exchange information after they’ve seen their coins (in which case the game would be trivial), their best bet is to output 0 every time.

Now, suppose that Alice and Bob are in widely separated locales, and each of them gets a visit from one of the twins described in New Scientist. Robbie visits Alice, and shows her the wallet he is carrying (it’s either red or green); Fred visits Bob and shows him the wallet he is carrying (which is the color opposite to Robbie’s). Does this information help Alice and Bob in any way?

No. If Alice sees a green wallet, she knows that Bob sees a red one, and vice versa. This doesn’t help them choose what to do in their Boolean coin-flipping game.

However, suppose that Robbie and Fred are each carrying one half of an entangled pair of quantum particles. Then there exists a strategy by which Alice can look at her coin, choose how to measure her particle and pick an output bit based on that measurement, such that if Bob follows the same strategy, they can win more than 75% of the time.

The details of that strategy involve some trigonometry and a little bit of elementary quantum mechanics; I won’t go into depth here, but these lecture notes have the story. The point is that looking at colored wallets doesn’t let Alice and Bob win more than 75% of the time, while taking measurements on entangled photons does. The quantum situation is fundamentally different from the one our letter-writer described, and the editor should have known it.

UPDATE (30 September 2007): Dave Bacon quotes the letter asking about entanglement and the editor’s response, and then says, “At which point my brain just exploded.” In the comments, Greg Egan gives some elaborations on Bell’s Inequality. For more on the Alice-and-Bob strategy problem, see quant-ph/0404076 and quant-ph/0608146.

13 thoughts on “More on New Scientist”

1. Just to clarify my problem with the Boltzmann brain article, my real objection concerns the “Copernican principle”, and can be stated independently of the concept of BBs.

Suppose a cosmologist asks you to choose between theory A, from which it plausibly follows that in the entire history of the universe there will be X oxygen-breathing conscious beings living on terrestrial-sized rocky planets (never mind how we can derive these consequences, just take it as given), and theory B, from which again it follows that there will be X oxygen-breathers on rocky worlds, and 1000X methane-breathing conscious beings living in the atmospheres of gas giants.

The “Copernican argument” — as put forward in the Boltzmann brain article — says “We are unlikely to be atypical”, and requires that we rate theory B as much less likely than theory A, on the grounds that we have observed ourselves and found ourselves to be oxygen breathers, which of course are not the typical conscious observers in theory B.

I think that’s nonsense; if O is the event “at least one conscious observer notices that he/she/it is an oxygen breather”, then P(O|A)=P(O|B)=1, i.e. both theories imply with certainty that someone will notice themself to be an oxygen breather. The observation doesn’t rely on the Hand of God descending and picking an oxygen-breather from the set of all observers with probability 1 for theory A, but probability 0.001 for theory B.

2. Graham Lawton says:

Blake – I can’t resist issuing you a challenge. Since you have identified the causes of bad science journalism, why don’t you show us hapless science journalists how it should be done? Here’s the challenge: write a feature-length popular science article on topic of your choosing and post it here. I’ll critique it for you. How does that sound?
From Graham Lawton, features editor, New Scientist.

3. Thanks for the clarification; I agree with your take on the “Copernican argument.” I wrote this post in a hurry between working on two other things, so there’s probably an egregious error lurking somewhere within it.

4. Graham Lawton:

Your comment got caught in the Akismet spam filter; sorry about that.

To be more precise, I wouldn’t say that I’ve identified the causes of bad science journalism, although both TR Gregory and I have tried to establish a taxonomy of the different ways it can go wrong. Also, at least in this case, my primary concern was not that an article suffered from slovenly reporting, but that the guy behind the desk should have said, “No, that’s not quantum entanglement; quantum entanglement is more complicated.” Photons can violate Bell’s Inequality; guys carrying wallets obey Bell’s Inequality. I tried to explain the difference with my Alice-and-Bob example (a XOR proof system, as discussed by Scott Aaronson). As I mentioned above, I wrote this post in a hurry between other things, so I’m sure my explanation could be improved.

How does that old joke go? “A critic is like a eunuch in a harem: he can see what is done, and he knows why it is done, but he can’t do it himself.” I do a great deal of criticism (here, here and here are some of my favorites), but I’ve also tried to explain things which haven’t been explained before, at least to a popular or semi-popular audience. My ongoing series on quantum mechanics (1, 2, 3 — with pandas! — 4 and 5) is an example. When I’m less harried for time, I’ll be taking it into supersymmetry.

Of course, that series involves actual math, so even though it’s much more conversational than a textbook, it’s a far cry from anything you could buy on a newsrack.

2,400 words, you say? Well, I’ve got a journal article which I need to revise for my research group today, and other stuff which needs to get done over the weekend. . . how about I put a story up on this website Tuesday morning?

5. Graham, I can’t resist issuing you a challenge: can you put some measures in place to rescue New Scientist‘s plummeting reputation? Hire a few consultants in a range of subject areas who know their field, and run every article past the appropriate person. I know there are several “consultants” listed on your masthead, but either they’re incompetent or they are not being consulted. Any competent physicist would have binned the EmDrive article in 5 minutes. I realise that your writers generally seek comments from a range of experts in the specific topic at hand, but in the case of the EmDrive there seems to have been a mixture of miscommunication, time pressure and selective quoting, because neither of the two people quoted with apparently positive opinions actually intended to endorse Shawyer’s ideas.

As I’ve said elsewhere, I think New Scientist does an excellent job 70% of the time, and most of the other 30% isn’t that bad. But you really need a mechanism in place to catch the spectacular misfires, and it shouldn’t be that hard. You don’t need Nobel prize winners, but London must be full of smart underpaid junior academics who could spare 30-60 minutes every week or two to assist with some quality control on the scientific content.

Snappy headlines, accessible metaphors, and engaging writing are all important, and most of the time New Scientist has these things in abundance. But when the science is flat out wrong, it makes the magazine a laughing stock in the scientific community, and a megaphone for misinformation for the lay public.

6. Graham Lawton says:

Yes. Thanks for accepting my challenge, that’s very cool of you.

To clarify my motivations, we science journalists get a lot of grief from bloggers. Some of it is valid but a lot isn’t. The criticisms often seem to be driven by a belief that if only science journalists operated by the norms of science rather than the norms of journalism, the scales would fall from the eyes of the great unwashed and science would take its rightful place as the undisputed font of all knowedge and wisdom.

Another common theme is that what we do is easy, and yet we still manage to do a crap job because we’re thick, ignorant, lazy and dishonest.

Hence my challenge. You’re quick to criticise, but let’s see if you can write something that gets the science so absolutely spot on that even Greg Egan can’t find fault (not easy) AND which can be read, understood and enjoyed by a general readership. This is the tightrope we walk every day.

On the entanglement thing, I don’t know what happened. The editor’s comment was cut back so heavily from the original that it ended up being wrong. I’m sure we’ll run a correction.

7. Graham,

For some reason, Akismet insists upon flagging your comments as spam. I don’t have anything to do with that, and I’m marking them “not spam” when I notice them so that hopefully the software will catch up. My guess is that you’re using an IP address which had previously been used to post spam comments on other blogs. Hopefully, this problem will clear up soon.

I accidentally missed this one, but I managed to drag it out of a cache.

Hi Greg,
Thanks for the challenge. Interesting, but as I don’t agree that we have a plummeting reputation I will have to decline.

You are, naturally, free to imagine you have whatever reputation you’d like. However, among physicists, linguists and computer scientists — to mention only the first few folks I’ve heard talk about it — New Scientist‘s reputation is not spectacularly high. Maybe these people aren’t your intended audience; OK, but engendering ill will among them hardly seems like good business practice to me. The telephone call which begins, “Hello, I’m researching a story for New Scientist” should not be an instant turn-off!

The criticisms often seem to be driven by a belief that if only science journalists operated by the norms of science rather than the norms of journalism, the scales would fall from the eyes of the great unwashed and science would take its rightful place as the undisputed font of all knowedge and wisdom.

I don’t think scientists (or at least those who think about the issue for any length of time) expect journalists to follow exactly the same practices that scientists do. Finding stuff out about the world is one job, and explaining it is another. When the explanations go awry, however, the people who find stuff out get rather nettled. So, I’m not sure what you mean by “the norms of science”, unless a foundation of integrity and respect for the facts is unique to science, which I simply don’t believe.

By analogy, suppose I were an expert in political matters. (It’s hard not to have an interest in politics nowadays, and I daresay I know a little more than many people, but plenty of people know much more than I.) Let’s say I’ve had extensive training at university, followed by years of hands-on experience inside the Beltway, learning the ins and outs of the business. Then I see, not once but multiple times, stories in glossy news magazines which get the political positions of major candidates wrong, perpetuate myths about the legal system, and give erroneous descriptions of how the Electoral College works. Let’s say that about a third of the articles I happen to notice have such goofs. Why shouldn’t I point these mistakes out to people and speculate why they occur?

To my eye, reporting on savings-and-loan scandals or foreign policy walks the same tightrope as that on astronomy or molecular biology.

Likewise for the “undisputed font” accusation. Just about every scientist I’ve read stuff from or met in person acknowledges that there are areas of human endeavor which are simply orthogonal to scientific research. The ones who don’t say this, say nothing at all. Science and technology give you an iPod, and art fills it with music: what could be less confrontational than that? Few researchers of my acquaintance are such monsters of vanity that they wish their judgments on Mozart and Picasso to be the gold standard of the realm. Furthermore, most everybody accepts the distinction between “is” questions and “ought” questions. There’s plenty of human wisdom involved in artistic creativity and morality, wisdom which (to briefly summarize a very broad topic) is informed by science far more than it is under the thumb of science. It’s when our answers to the “is” questions get misrepresented our outright mangled that we tend to come over all persnickety.

I am quick to criticize — guilty as charged! I also try to explain something each time I make a criticism, and in such a fashion that it is comprehensible to as many people as possible. Otherwise, my criticism would have no substance.

If I believed otherwise, your challenge would make me feel rather like Roger Ebert being asked to direct a movie.

I’ve enjoyed this little back-and-forth, and I hope to have something good to read available in a few days.

8. John Baez says something I meant to point out but was too absentminded to include in my post.

The funniest thing about this Boltzmann brain baloney is that after convincing themselves that randomly self-forming assemblages of particles having experiences just like ours are not only possible but predominant in an infinite universe, people still seem perfectly sure that they aren’t examples of this phenomenon — and then go ahead to draw conclusions based on both these ‘facts’. For example, the conclusion that the universe can’t really be infinite.

9. Graham Lawton says:

I’ve enjoyed it too.
The politics analogy is a good one. I’m pretty sure that all professionals despair about media coverage of their specialist subject. As just one example, I heard a historian on BBC radio the other day bellyaching about the BBC’s coverage of history. He was worried that people were getting the wrong idea about how history was done, and were not being taught historical facts in the right way… sound familiar?
(The irony is that, here’s me, bellyaching about the way that bloggers portray science coverage in the media.)
On boltzmann brains, you’re entitled to think there’s no merit in the science, but shooting the messenger won’t get you anywhere. All we did was faithfully report what cosmologists are thinking and doing. If you don’t like it, have a go at the people who fund the research, not New Scientist.
I’ll shut up now.

10. Graham, on the Boltzmann brain article, no one’s shooting the messenger. I don’t doubt that there are cosmologists whose views the article faithfully describes, but what I expect is for your readers to be told that there are also many cosmologists who consider this line of reasoning about as sensible as arguing that the Ancient Greeks’ estimates of the size of the Earth should have been discounted due to the weird corollary that the planet would be so large as to make Greeks an atypical minority.

I haven’t polled cosmologists on this issue, but surely the contrary view deserved a paragraph or two at the very least. New Scientist would (thankfully) never publish HIV or climate-change sceptics without putting their views in context. Reporting on more abstract issues might not carry the same moral burden, but the basic principle is the same: readers deserve to get a sense of the true range of informed opinion.

11. The inaugural issue of the Yale Climate Media Forum on Climate Change and The Media includes a fascinating piece by Philip Meyer, journalism professor at UNC Chapel Hill. I quote a few paragraphs.

In the age of the Internet, mere transmission no longer adds value to information. The way to add value to the surplus of data is to process it to help the reader select it and make sense of it. That requires interpretation, and interpretation requires objectivity in the scientific sense. I call this objectivity of method as opposed to the he-said/she-said objectivity of result. In other words, journalists should act more like scientists: collect information, look for patterns, construct a theory, and then provide an objective test of the theory. Objectivity in this sense means asking a question of the data in a way that will protect you from being fooled by the answer.

Journalism, like science, is tentative in its conclusions. It should be as transparent as science, leaving a paper trail of data that other investigators can retrace and arrive at the same or better conclusions.

The reporters who bought the White House line on weapons of mass destruction in Iraq were not being objective investigators. They were just parroting their sources, fearful of alienating them. That’s stenography, not reporting. Correspondents in the Knight Ridder Washington Bureau got the story right because, not having a Washington outlet, they were not in a symbiotic relationship with the official sources and had to use shoe leather to seek out the working stiffs in government and the military. Objective inquiry, not advocacy, made their effort successful.

(Words are nice; read the whole thing.)

12. In the latest issue of New Scientist they’ve published two letters from readers setting things straight on entanglement, about as well as could be done within the space limitations. The one thing that’s missing here is any clue as to how the local hidden variables / “coloured wallet” model is falsifiable, which AFAIK can only be explained by invoking more than one (joint) observable.