And even then, Scridb is still holding on to the document, meaning it is still stored in their database, which posts yet another legal problem because they store copyrighted documents without explicit authorization by the document's owner.
Copyright constrains your right to copy, not to store or to receive. Only through distribution can you infringe copyright.
For physical items, that's true, but to receive data, you make copies of it. Does your file upload process buffer an upload in RAM before writing it to disk? Two copies. That's why EULAs can even exist, because to install your software, you have to make a copy of it from the original medium to your hard disk, and to run it, you have to make a copy of it from your hard disk to RAM. scribd is almost definitely violating copyright by holding on to documents. Especially if they have backups, caches, and multiple servers, which are even more copies.
Incidental copying like that has never been tested in court, and neither have EULAs for that matter (I'd bet my ass they're invalid). Everything gets buffered somewhere, even in analog processes, even in physical processes. Stop making infringing copies of my work in your brain!
Even if it was deemed within the purview of copyright, it's squarely within the realm of fair use -- they aren't violating copyright any more than a corporate mailroom that scans and duplicates incoming correspondence. They took it in in good faith, they aren't publishing it outside their own organization, and they aren't making copies to get more 'use' out of it, so it doesn't matter.
"Secondly, there are a number of publishing businesses - sleeping with the enemy"
"The absence of "editorial intervention and approval" in combination with a highly flawed sign-up process"
http://www.writersweekly.com/this_weeks_article/005717_11182...