In a recent Google Webmaster Help video, Matt Cutts, discusses the use of the canonical tag for duplicate content. He references specifically short-term duplicate content. The example Cutts uses is a news site that would have content surrounding a story that perhaps is breaking news that is being updated constantly as more details emerge surrounding the event. In this instance, there is a legitimate reason for content to be duplicated across several different url’s on the same site.
Cutts suggests using the rel=canonical tag to avoid PageRank being shared across all multiple url’s containing this content. This would consolidate PageRank to one main page so when searchers looking for information surrounding the event would find a preferred page with links to the corresponding stories.