Hacker News Reader

Comment by khawkins | Hacker News Reader

Occam's razor is a heuristic used in the investigative process, not a principle. While it's fair to point out that a theory makes everything more complicated and thus might make it less likely to be sound, it's not fair to push back on the first bit of research into a theory because they haven't had the time to fully develop it.

Indeed, a scientific revolution generally starts by treating as false an assumption previously held as true. Not to say that this is a revolution, but if you push back too hard it won't be, whether it's true or not.

View on HN

History : Dun Scotus (really) probably invented Ockham's razor - William was his student and used it to avoid the thing that all medieval philosophers needed to avoid, namely being tied up and put on a bonfire.

Why did Dun invent it (probably)… well here's the thing, Christianity has a God with three faces, son, father and holy ghost thing. Why? The answer is - don't multiply entities beyond necessity, so God has the number of faces necessary to do the job, no more, no less.

I'm wittering on because this is where that heuristic came from, literally it's angels on pins stuff. So don't invest in it, I'd bet a bit that if we got Dun and Bill together with a few pints of mead they'd laugh themselves silly to here that 21st Century physics pins any weight onto their measure.

The Greeks wouldn't have, the Chinese didn't, why do we?

I mean, we still call it Occam's razor, but I don't think that it's really even the same principle by this point, it just shares the same name. In science today, we don't really care about the number of entities. We don't reject the idea that the stars are suns like our own just because that would imply that the number of atoms in the universe is drastically larger than the number of atoms in the solar system. Instead of hypotheses with the least number of entities, we favour hypotheses with the smallest Kolmogorov complexity. [1] As a slogan for the modern version, I like the phrasing by John Von-Neumann: "With four parameters I can fit an elephant, and with five I can make him wiggle his trunk." [2]

[1] https://en.wikipedia.org/wiki/Kolmogorov_complexity [2] https://www.johndcook.com/blog/2011/06/21/how-to-fit-an-elep...

Nitpick: The historical problem with stars being suns wasn't (just) the universe being larger than the solar system, but that the telescopic observations available at the time seemed to imply that every other visible star would have to be much larger than the Sun, in fact larger than the orbit of Saturn. This was because early astronomers didn't understand optical diffraction and thought the Airy disks visible around stars were the stars themselves, making their angular radius in Earth's sky seem vastly larger than the reality. [1] Both characterization of the Airy disk and observation of stellar parallax didn't occur until the 19th century, by which point religious objections wouldn't have had the same status as in Galileo's day anyway (for example, Darwin's work was published only a few decades later).

Source: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.612...

You ought to read a little bit into the history of astronomy, particularly planetary motion. If the ancient Greeks had applied Occam's razor then we would've abandoned geocentrism long before Copernicus.

We should've listened to (and built on) Aristarchus of Samos instead of Ptolemy's deferent/epicycle contraption. All of this because everyone wanted to hang on to Plato's assertion that the heavens obeyed the mathematical beauty of circular forms.

Although it's attributed to him, William of Ockham did not come up with Occam's Razor.

My metaphysics teacher showed me a line in the Summa Contra Gentiles where St. Thomas Aquinas enunciated the same idea behind the Razor almost a hundred years before Ockham did. Sorry, I don't have the reference handy. Furthermore, the Wikipedia article on Ockham's Razor traces the basic idea all the way back to Aristotle's Posterior Analytics.

Occam's razor has a practical interpretation in Bayesian statistics, in that a more complex hypothesis will tend to have exponentially lower prior probability, because in comes from a more heavily parametrized component of the hypothesis space.

That doesn't mean much without a concrete probability model, though, and there probably isn't one in this case.

Also that point seemed odd. Isn't unification of two previously separate phenomenon (dark matter and dark energy) simpler in Occam terms? It's using one thing to explain two, even if the actual mathematics are more complex to state. Also there's other troubling cosmological issues that this would appear to address, such as Hubble's constant having slight variation outside of measurement tolerance, which slightly differing fluid densities would explain (vs the current assumption of unknown sources of error).

Sabine's point, if it stands up, is that the author used a different equation of motion in their simulation than they write up in the paper. If true that's sloppy and bad, but it's not a point against the actual theory underlying the simulation, no?

Whether it is simpler depends on the details of the unification. Two comparatively simple theories are not made simpler with a very complex unification.

The electroweak force is most certainly not easier to describe than electromagnetism and the weak force separately. But we still talk about electroweak interaction as being a simpler theory than two separate fundamental forces, because it reduces the number of things that are needed to describe the universe, even with added computational difficulty.

One could also think of it in terms of the the amount of "magic" required to make the simplification work.

As far as I can tell Occam terms seem to differ depending on the convenience of whoever happens to be wielding Occam's razor.

It's not totally a heuristic, though it's intractable to use it the real way. Which theory has a shorter program that generates it?

As I read that paragraph, I thought to myself that the author should read Sabine Hossenfelder's recent book Lost in Math: How Beauty Leads Physics Astray. In that book, Sabine argues quite convincingly that our aesthetic judgements of new theories can mislead us. She says of simplicity,

> This dream still drives research today. But we do not know whether more fundamental theories necessarily have to be simpler. The assumption that a more fundamental theory should also be simpler—at least perceptually simpler—is a hope and not something that we actually have reason to expect.

> As I read that paragraph, I thought to myself that the author should read Sabine Hossenfelder's recent book Lost in Math: How Beauty Leads Physics Astray.

The author is Sabine Hossenfelder :)

I realized when I reached the bottom of the post. It is odd to me to see her saying that the new theory is disqualified for not being as simple as alternatives after reading her book.

It's not disqualified for not being simple, it's disqualified for not making testable predictions. Given two theories, neither of which make testable predictions, the simpler one (and the one that has been studied more thoroughly) is the more reliable.

Otherwise, why not just say that the missing mass is unicorns? And the energy driving the accelerated expansion of the universe is the energy of their love? There's no shortage of radical explanations for dark mass/energy. The scientific community takes them seriously to the extent that they are compatible with past observations and make testable hypotheses that distinguish them from other theories.

The author of the negative mass theory said that it will be testable via the cubic kilometer ice array in Antarctica when it comes online (though I'm not sure how). He mentions it in his chain of 17 tweets that summarize the theory .

I thought he was planning to use the Square Kilometre Array radio telescope which is still in the works, and not to be confused with the Ice Cube neutrino detector which has been operational for several years.

https://cdr.skatelescope.org/

https://icecube.wisc.edu/science/icecube/detector

Yep, you are right - I was confusing the Square Kilometer Array for the Ice Cube neutrino detector. Heres a direct link to his tweet: https://twitter.com/Astro_Jamie/status/1070304980224172033

Then we'll find out when it comes online! If the new theory is compatible with past observations, and correctly predicts something that GR gets wrong, that will be a huge finding! But the press has gone wild prematurely, as they often do. As others have mentioned, that seems to be the real issue to which Dr. Hossenfelder objects.

>It's not disqualified for not being simple, it's disqualified for not making testable predictions

That's not uncommon. In their early development, most theories don't make testable predictions, especially when they attempt to expand the status quo, and not just add some incremental detail...

> odd to me to see her saying that the new theory is disqualified for not being as simple

No. Her main argument is, the way I read her article, is:

"Farnes in his paper instead wants negative gravitational masses to mutually repel each other. But general relativity won’t let you do this. "

The way I understand it, the author of the paper fails to be compatible with the theory that was confirmed time and again during the last 100 years. She just avoided to formulate that so bluntly.

In which case she probably has read the book.

Reply on news.ycombinator.com