but I find this explanation improbable, bordering on spurious

As the author points out, finding Greek σιδηρος (sideros) 'iron' and Latin sidus, -eris, 'star, constellation' cognates goes against how Greek and Latin developed phonologically. Latin sidus has been compared to the Germanic words for silver and slag, but those comparisons are not without problems, too. In the end the usually PIE root suggested (by Buck, Walde, Pokorny, et al.) is *sueid- 'to glow'. (Also, the -r- in the Latin word is probably a result of rhotacization (-VsV- => -VrV-).) The Greek word has no appealing etymology, and is thought to be a loan from some unknown language. A terminus post quem for the borrowing would be the (hypothesized) time of sV- => hV- in Greek historical phonology.



Ceci n'est pas un seing.