When you are building a site which has a search result page, (similar to https://webmasters.stackexchange.com/unanswered for example) which has excerpts from each article and a link to the article, how do you instruct the search engine to direct users to the actual article not the results page?
If we had the content replicated on another page we could use rel='canonical' which would essentially do this at a page level, but what happens when you have excerpts from multiple pages in the same physical page?
Does this even matter at all? SE doesn't seem to do anything special here (that I could see). I just really don't want users to ever be directed to a results page when they could get the article content instead.
Originally, aggregate pages were designed to be landing pages that would perform well in search engines. It was a way of attracting users to your site. Today, while this still can work, it is less of a factor. Some people de-tune these pages so that they do not outperform the original content. This makes sense if the aggregate page is for humans instead of machines. However, if you think that this page should perform well in search then the page would have to be tuned carefully. eBay does this to perfection.
You do not have to do anything "canonical" for this. You just have to decide if you want the page to perform well in search or not. From there, it is about tuning your pages.