Hacker News Reader

Comment by kmm | Hacker News Reader

I'm not an expert in GR, but the linked paper seems nonsensical[0]. It postulates a highly degenerate decomposition of the metric tensor, for which they postulate a contrived action which then seems to correspond to the Einstein-Hilbert, through mathematically unsound manipulations (how can you raise indices with the metric being nowhere even close to invertible?).

Besides, how can you talk about unifying general relativity and electromagnetism without mentioning Kaluza-Klein theory[1]? And what about one of the most beautiful principles in physics, gauge invariance[2]?

I don't want to be rude, but I'm very curious as to how this got through peer review.

0: https://iopscience.iop.org/article/10.1088/1742-6596/1956/1/... 1: https://en.wikipedia.org/wiki/Kaluza%E2%80%93Klein_theory 2: https://en.wikipedia.org/wiki/Gauge_theory

View on HN

> Besides, how can you talk about unifying general relativity and electromagnetism without mentioning Kaluza-Klein theory[1]?

No offense but the very first paragraph of the article's introduction mentions Kaluza's work:

> The earliest attempts can be reasonably traced back to the German physicist Gustav Mie (1868-1957) and the Finnish physicist Gunnar Nordström (1881-1923). Fruitful efforts came, for example, from David Hilbert (1862-1943), Hermann Weyl (1885-1955), Theodor Kaluza (1885-1954), Arthur Eddington (1882-1944) and of course also from Albert Einstein (1879-1955). It is less well-known that, for example, Erwin Schrödinger (1887-1961) had such inclinations as well, see [1]. For a thorough historical review, see [2].

Kaluza's work is mentioned in the first paragraph of the paper. How can you write such an inflammatory critique without having actually read the paper?

Kaluza's name is mentioned, but not Klein's, and there's no mention of their theory whatsoever.

Kaluza-Klein theory is the archetype of expressing electromagnetism purely through curvature, and I'd expect any paper doing the same to refer to it, as well as explain how the work in the paper differs from or expands on it.

The fact that the metric proposed in the paper corresponds to the term added to the 4D spacetime part of the Kaluza-Klein metric is already suspicious. It makes me think they're either repeating Kaluza and Klein's work, or aren't properly citing it when they should have.

I have never been in academia and this is the first time it clicked to me that citing a paper is not just an act of mentioning it as a networking artefact, and should be critical to the substance. Thanks for the explanation.

> I'd expect any paper doing the same to refer to it, as well as explain how the work in the paper differs from or expands on it.

But they do:

> The strength of the present approach is simplicity, there is no need for higher dimensions, torsion tensors, asymmetric metrics or the like.

(emphasis mine)

> The fact that the metric proposed in the paper corresponds to the term added to the 4D spacetime part of the Kaluza-Klein metric is already suspicious.

Why? There aren't many ways to write down a symmetric 2-tensor starting with a vector field.

> It makes me think they're either repeating Kaluza and Klein's work, or aren't properly citing it when they should have.

They're doing neither. Sure, they might have mentioned Kaluza's work more explicitly (AFAIK Klein's contribution was the suggestion to compatify the 5th dimension) but I suppose 1) they assumed that readers are familiar with it and 2) they wanted to mention several approaches to marry GR and ED and Kaluza's approach was just one of them.

Are you sure it was peer reviewed at all? It looks like a conference proceedings.

According to this https://iopscience.iop.org/article/10.1088/1742-6596/1943/1/... they're all peer-reviewed. An average number of reviews per paper of 1 seem a bit curious though.

Assuming that every paper is reviewed at least once, Markov’s inequality implies that every paper is reviewed exactly once!

I think this is the declaration for the volume containing the conference in question?

https://iopscience.iop.org/article/10.1088/1742-6596/1956/1/...

Higher acceptance rate, but more reviewers per paper.

Yeah, I think so too. While the authors might attempt to counter that they are only describing EM in vacuum, the fact is that GR is nonlinear. You can’t just add an EM sector by a rank-1 metric component and expect linearity to apply.

At some level, this feels like a vacuous result. If you start with a rank-2 massless field theory, and constrain it to act on rank-1 fields, is it surprising that you get a rank-1 massless field theory? Is this not just an elaborate form of completing the square?

Yeah first thought that popped into my head was Kaluza-Klein theory, and I'm far from an expert enough to dissect the paper but it has that smell of junk science.

Reply on news.ycombinator.com