tldr: CLHS was programmatically generated from public domain standard, and it's the result of that work ("html code") that's copyrighted. CLHS and the standard are otherwise near identical and anyone can take the standard and produce their own html reference. CLHS is the oldest such effort, the copyright is permissive enough, that nobody has bothered to.
so the history behind CLHS is explained by Kent Pitman, the editor of the spec and the person who made Hyperspec https://nhplace.com/kent/Papers/cl-untold-story.html (starting at 6.2)
The standard committee produced several drafts of the spec, the copyright of the drafts was never transferred to ANSI, and the committee's explicit mandate was that the spec will be available to all interested parties to do with as they please. It's not clear who owns the copyright on the spec, but it's intentionally obscured so that nobody can claim it. I guess the most likely claimant is Kent Pitman, but he himself says that he was doing work for hire for an entity that doesn't exist.
Three draft proposed American National Standards (dpANS) were subsequently produced. dpANS 1 was a working draft and received extensive review and revision by X3J13. dpANS 2 was then created to resolve all remaining technical issues; X3J13 made no intentional changes in technical content after dpANS 2. dpANS 3 improved wording in a few places and extensively reworked the Credits section, not a normative part of the standard. Further formatting changes and boilerplate additions were executed by ANSI in their process of producing the standard, but these have no technical implications.
dpANS3 was shared freely to all interested parties. and while only published ANSI Standard is considered to be a definitive reference (and not hyperspec!), according to Franz's blurb above the only changes between dpans3 and the standard are cosmetic.
Later Kent Pitman while still working at Harlequin (which became Lispworks) programmatically converted dpANS to html, and Harlequin/Lispworks own copyright on that effort. Because they distributed it freely and were the first ones to produce it, it was also the most widely used standard equivalent. Franz has their own version of the conversion, https://franz.com/support/documentation/current/ansicl/ansicl.htm which is (c) franz inc. They explain the details here https://franz.com/search/search-ansi-about.lhtml which is also where I got the blurb above. Compare for example a random page of Hyperspec http://www.lispworks.com/documentation/HyperSpec/Body/03_bd.htm and corresponding Franz's section https://franz.com/support/documentation/current/ansicl/subsecti/literalo.htm
So Hyperspec, dpANS, and ANSI spec are all identical: same chapters, same form descriptions, same examples, etc. etc. Hyperspec is "copyrighted" Lispworks, dpANS is public domain, and ANSI spec is "copyrighted" ANSI.
You can grab dpANS from here https://github.com/xach/dpans. It's written in TeX using custom macro definitions. There is pdf of its output https://franz.com/support/documentation/cl-ansi-standard-draft-w-sidebar.pdf (NOT copyrighted franz). There is a custom parser for it https://github.com/robert-strandh/dpANS-parser, a converter to Texinfo https://github.com/rebcabin/dpans2texi and I believe there was an effort to do a "modernized" pdf by hacking up dpANS macros a little bit, but I can't seem to find it.