The problem I've had with PHP ever since it was started is not that it's a bad language, but that it started out as a bad language in exactly the same way that Perl 4 was a bad language (unsurprisingly, as PHP developed from a Perl script) and then proceeded to fix those problems, one by one, in exactly the same way Perl had already done it, several years before - and yet PHP became very popular while most people trying to use Perl for web programming are still asking basic questions about CGI scripts written in Perl 4 like it's 1994. One of the best proofs of Worse Is Better that I ever saw.

Wait - "broken" doesn't mean "broken beyond repair", let alone that somebody else would be able to provide a better fix. At first sight this is merely a problem of Google neglecting to use a specific bit of information in their filtering, namely the information who copied content from who. It won't be possible to do this automatically with 100% accuracy, but I guess if you take all content from two webpages as prepared for indexing (that is, basically the plaintext conversion), compare them for similarity, and when found to be similar, downgrade the ranking of the one to appear most recently (which is not so very easy to determine) then you can mostly fix this problem. The basic problem I see is that even with smart optimization techniques (e.g. limit comparisons to pages with similar sets of keywords) the similarity testing required probably won't scale. However there are ways around that too (e.g. don't calculate it at indexing time but spread the calculation over queries at querying time, to make ranking progressively smarter, but I have no idea if Google's infrastructure allows stuff like this).