1. Gosling Emacs was not the first Emacs in C either, I'm pretty sure Montgomery (later CCA) Emacs came earlier, plus there were some so-called "ersatz Emacses" like ELLE (Elle Looks Like Emacs) and JOVE (Jonathan's Own Version of Emacs), not sure of their chronology. It certainly wasn't an audacious idea of Gosling.
2. GNU Emacs did for a while have a mix of files with Gosling's copyright (those were the ones that contained code that Gosling wrote) and files without it (those were the ones that had been written or rewritten by Stallman). Eventually all the Gosling code was replaced and so Gosling's copyright notices also went away. One could claim this is a "ship of Theseus" but that's a hell of a lot different than saying Stallman stripped away Gosling's copyright headers while re-using Gosling's code. The headers stayed in until the code was gone.
3. Gosling Emacs (before it became a Unipress product, and before GNU) was in fact distributed/redistributed by a great number of people in those days. Gosling was completely cool with that as far as I can remember. Later when he sold it to Unipress he announced something like "the program is too good to waste on the public domain". I don't remember if the gratis version stayed in circulation after the Unipress version came out, but it might be possible to research this.
Source: was there.