Canonicalisation
This refers to the removal of duplicated pages from a website and basically means that each URL on a website should bring up a completely unique webpage.
When search engines index a page of a website, through the use of their bots, they look at the importance of each page and also at the links contained within each one. If two pages have different URL's but the same content, the bots will split the authority of these pages resulting in the loss of "link juice". In broader terms, it means that if a page with a rank PR rank of 6 is duplicated, then its rank will be split with the other page(s). This means that its resulting PR will be reduced depending on how many times the content is duplicated.
Duplication may be done by mistake, however one of the most common reasons is that a URL may have a www version and a non www version. Although they may look like different pages, they are seen by the search engine bots to have the same content. Another mistake that people may make is creating a new page with the same content, but leaving the old page there as it still has links leading to it.