It should be noted that certain publications written by creators of Star Trek's various series are sometimes referenced in later episodes of those series. Somewhat paradoxically, the novels themselves remain apocryphal in such cases, while whatever information was mentioned on-screen becomes canon.
The only case where that really happened was with Jeri Taylor's novel Mosaic, which told Janeway's life story; Taylor later incorporated elements from that backstory into "Coda." And maybe it also happened with the books The Ferengi Rules of Acquisition and Legends of the Ferengi by Ira Steven Behr & Robert Hewitt Wolfe -- I think a couple of Rules were introduced in those books before showing up onscreen, but I could be wrong.
And it's not really paradoxical. Anything written by members of the production staff, whether it's the writers' bible or a technical manual or an encyclopedia, is something that can potentially be a source of ideas for the show itself, but it isn't meant to restrict the creativity of the show's own makers. The show is the core work, and everything else exists only to support it. So there has to be the option either to use elements from outside sources or to ignore/contradict them, depending on what works better for the show.
It only seems like a paradox when people misinterpret "canon" to mean "real" or "true." Any long-running canon contradicts and rewrites itself all the time (Are they Vulcanians or Vulcans? Is it UESPA or Star Fleet or Starfleet? Does Data use contractions? Has Deanna ever kissed Riker with a beard?). The canon is simply the core work itself as distinct from secondary or derivative works.