Lost in Translation

I've written before about the problem of science journalism from the scientist's perspective. Here's a report on a physics colloquium at USC on the same topic, but from the journalist's perspective.

One interesting point is that the news media require an article to be "news" and science just isn't about making news in general. Almost any study builds upon earlier work, and new results are incremental. But a newspaper article needs a flashy headline and lede, so every study is a breakthrough. Worse, the publication of some study today doesn't mean that anything in particular happened on 27 March 2006. The research was completed months ago, and possibly took years to do. But the newspaper article has to take the perspective that something happened, which distorts the whole thing. Until the editors can get their heads around this distinction, science journalism will continue to be insipid.

Astronomy illuminates the glory of God

And here's a view of what we should be doing experiments on and why. (Also, you get lots of nifty pictures.)

Experiments on Prayer

There have been a number of published experiments testing the proposition that prayer aids sick people. Early ones claimed to see an effect, but later ones claimed no such effect leading to arguments and polemic. It seems to me, however, that from a Jewish perspective, such experiments are doomed to failure for their premise is faulty.

These experiments generally work as follows. The experimenter takes a sample of ill people and divides them into two groups. One has people praying for them, the other does not. The experimenter then looks at the survival rate of the two groups. Should it be the same, it is claimed that there is no effect. Should the group being prayed for do better, it is claimed that prayer is efficacious. I suppose that should the control group do better, it would be claimed that prayer is dangerous and should be stopped immediately. Whatever the outcome, the conclusion is founded on the assumption that the two groups are equivalent, and that any statistical difference in their outcome is a result of being prayed for.

Leave aside the twiddles and flourishes and the details of the experimental protocol. Never mind whether it matters that Jews are praying for Christians, or vice versa, or whether athiests or agnostics should be included either as the ill or the prayers, or any other such details. The experiment is fundamentally flawed from the moment the ill people are divided into groups. In the interests of not biasing his results, the experimenter uses some "random" method to divide the patients into groups. Perhaps he pulls names from a hat, or balls from a lottery machine. Or, more likely, a computer is involved. Whatever the case, the intent is that the two groups be completely equivalent so that the result can be attributed, or not, to the effects of the prayers.

But, remember, the very premise of the experiment is that there possibly exists something, call it "God" for convenience, who can be influenced by "prayer" to change the outcome for the sick person. What is the nature of this "God"? I don't know what the experimenters were thinking, but the G-d that Jews know of is omnipotent, omniscient, and ever present; directly involved in the day to day running of the world. So, when the experimenter "randomly" divided up his patients, was it not G-d who actually determined which group they would go into? And was it not already known before Him how long each would live, and what the effects of the prayers would be? So, then, was it not the case that the results of these experiments would be exactly, precisely, what G-d wanted them to be, whether positive, negative, or ambiguous? Perhaps there ought to be a theologian on the granting panel next time someone proposes to run experiments on G-d.

Customs of the day

15 Adar

On the day after Purim it is the custom to wear leather shoes buffed to a high gloss. If possible, one should travel to a railway station or airport to have the polishing done by a craftsman, who should be paid with an open hand. One who uses a machine, or does it himself, also fulfills the custom. One should also listen to shiny shoe music. In view of these customs, the day is known as Shoe Shine Purim.

The Number of Stars in the Universe

A reader asks

What can you tell me about chazal's take on the number of stars and what modern science says today and are the numbers the same and whether you feel that this is a proof of torah shbaal peh?

The relevant text is Gamara Brachos 32b:

אמר לה הקב"ה בתי י"ב מזלות בראתי ברקיע ועל כל מזל ומזל בראתי לו שלשים חיל ועל כל חיל וחיל בראתי לו שלשים לגיון ועל כל לגיון ולגיון בראתי לו שלשים רהטון ועל כל רהטון ורהטון בראתי לו שלשים קרטון ועל כל קרטון וקרטון בראתי לו שלשים גסטרא ועל כל גסטרא וגסטרא תליתי בו שלש מאות וששים וחמשה אלפי רבוא כוכבים כנגד ימות החמה

where Reish Lakish quotes Hashem as saying he created 12 constellations, for each constellation 30 chiel, for each chiel, 30 ligyon, for each ligyon, 30 rehaton, for each rehaton, 30 karton, for each karton, 30 gistera, and for each gistera, 365 thousand myriad stars. Multiply it out and you've got 12 x 30⁵ x 365 x 1,000 x 10,000 or about 10¹⁸ stars. An impressively large number, indeed.

So what does "science" have to say about this question? In context, we must be asking how many stars are there in the observable universe. This, at least, is finite, since the further away we look the further back in time we look and the universe has a finite age according to current understanding. Stars are found in galaxies, so, to estimate the number of stars in the universe we need to worry about three factors: The typical number of stars in a typical galaxy, the number of such galaxies in a typical volume of space (the galaxy density), and the total volume containing galaxies.

The most recent such estimate for the number of visible stars in the universe is that of Prof. Simon Driver, presented at the 2003 General Assembly of the International Astronomical Union. He comes up with 7 x 10²² stars. And this is probably a lower limit since he's only counting stars within the reach of his telescope.

Well, he doesn't really count the stars. What he's done is count the number of galaxies within a (relatively large) strip of sky and estimated how many stars are in those galaxies from their brightnesses. Then he's multiplied that number by the total number of such strips needed to cover the whole sky. Or at least that's what I assume he's done based on the article. Nothing has been published regarding the calculation.

My correspondent quotes some old mail-Jewish discussions on the subject.

The number itself is amazing (10^18). This is (cosmically speaking) pretty darn close to the current estimate of 10^22.

gemara (Brachot 32b) _does_ count the number of stars in the sky, or at least calculates the number, and comes up with 12*(30^5)*364*(10^7) = 1.0512 * 10^18. On a logarithmic scale, this is surprisingly close to the best modern astronomical estimate, which is about 10^21.

How accurate is Prof. Driver's number? Without knowing how he does it, I couldn't say. Given the sort of assumptions he's likely making I would be surprised if he's off by more than a factor of 100 or so in either direction, though. Which is important, because his number is 70,000 times larger than that of R"L. Those who would like to claim that the two numbers are equivalent are guilt of innumeracy. Just because two numbers are large, doesn't mean that they are the same, and a factor of 70,000 is far too large to brush under the table. Let me illustrate what this factor means. Pick a stretch 6 km long on your favourite highway. The Moon is 70,000 times further away. Can you actually claim that bit of highway is as far as the Moon?

Some attempts to deal with the numbers:

What I cannot figure out is which assumptions the astronomers are making that could be tweaked in order to make the numbers match better. The simplest would be to lower the number of stars in an average galaxy to 10^4. But that seems awfully small. Any thoughts?

By the way the wording of the passage is curious Reish Lakish doesn't have the vocabulary to state such a big number so he talks in terms of what we would call galaxies and galactic clusters.

Avg galaxy ("gastara") = about 4x10^9 stars
Avg local cluster ("karton") = 30 galaxies
Avg supercluster ("rahaton") = 30 clusters

It goes on to say that superclusters are grouped into clusters of about 30 (megasuperclusters?) and that these are in turn grouped into an even bigger pattern of about 30 (hypermegasuperclusters?) of which the universe has a total of (about) 365.

Moreover from my amateurish research it seems that one of the prevailing theories of cosmic structure is that it is fractal and is not the calculation of Berachot 32a an example of fractal structure (4billionx30x30x30x30x360)?

Is this a "proof" of Torah sh'bal-peh? I wouldn't say so, and I think it a mistake to argue so. From one perspective, R"L would need another three layers of factors of 30 to be in the same ballpark, but his way of calculating the number of stars has nothing to do with the way they are really distributed in the universe. All those groups of 30s correspond to nothing observed, and resorting to "fractal geometry" and "hypermegasuperclusters" doesn't make it any better.

From another perspective, who is to say that the Prof. Diver's number is correct? Let any of his assumptions be wrong, and that number could increase, or decrease, possibly significantly. I think it presumptuous to assume that as of today science has "the correct answer" which can be used to validate Torah.

Besides, the whole argument misses, I think, the point of that gemara. The context is that Knesset Israel is complaining to HaKodsh Baruch Hu (i.e. G-d) that she has been forgotten by him. R"L has Hashem answer that he's created an incredibly unimaginable number of stars, every last one of which are for the sake of the Jewish people; how can she say he's forsaken her? In that context, is the number really intended to be precise? One can read it, instead, as a deliberate hyperbole, with no intention that the actual number be taken seriously. What is much more interesting, perhaps, is that R"L is willing to suggest that there are vastly more stars than those that are visible to the naked eye. If you want to hang something on this gemara, I think that that is a somewhat more fruitful direction than innumerate and presumptuous comparisons of very large numbers of dubious precision.