I don't see anything in the official stance as to what to do if you don't know who the original author is though. For example, when I started here I inherited a bunch of legacy code and I don't know who the original author was or where the code came from. Now I have rewritten most, if not all of this code from the ground up over the years, but I do wonder if I had wanted to post it, how the proper way to handle it would have been?
I always try to say, at minimum who I believe the author was or may be even if I don't know where I got that code. Some people, like Tony Tanzillo, Vladimir Nesterovski, MP, ElpanovEvgeniy, Renni have so much code in so many different locations it's hard to pinpoint where you got the code. Often times that was a judgment call on my part based on the style of code I was looking at (some code you can just tell who authored it after you study enough of that person's code). ...I think there was only one time where I couldn't list at least the author so I noted that in my post and included a header saying that I assumed "temporary ownership" of that code until the actual author was noted.
When I was learning AutoLisp I tried to consume as much code as I could often neglecting where I got the code. Only later did I wise up and start documenting "who, what, where, why and when".