External anchor text is an orthogonal issue to the problem of having to succinctly describe what you are looking for in very few distinctive terms. Think about it this way: Imagine if you replaced every document on the web with the “anchor enhanced” version of that document, i.e. you added a few paragraphs on to the end of every document with the text from all the inbound anchors.

Now, every document is has more text associated with it, correct? So it’s more descriptive. But now every other page on the web also has more text associated with it, because every other page on the web has now been enhanced. So at the end of the day there is a higher chance that the words you use to describe your query will be found on the relevant page. At the same time there is now a higher chance that those same words will be found on a lot more non-relevant pages, too.

And, as Jeremy noted, Mike does confound two limitations of conventional web search: it not only doesn’t leverage dynamic interaction between people a la social search, but also doesn’t offer the user-machine interaction of exploratory search.

But I concede that the word “social” is a bit poorly defined. In some sense, everything we do–or at least everything with regard to communicating information–is social. I just feel that’s not a particularly useful sense.

I’m not sure about the distinction here. When web search engines add external anchor text to the description of a page, they’re already widening the vocabulary and using the static social networking of links to improve results. And every time the engines index a Yelp or Consumerist review page, they get user created content. As usual, I think “it depends”.

Avi

And I wanted to put it in the record that social search isn’t the only way of adding context to your information seeking. Social information is one sort of context. The things that exploratory search systems do is another type of context.

But context is what we’re really all after, here.

There are other types of search that are not reference oriented, such as exploratory search. Social search is one such type. Exploratory search is another.

But is that what people mean when they refer to Twitter as a search engine, recommender system, etc.? In that case, it seems they should be arguing that offline communities are search engines that compete with Google–and I’ve never heard that argument.

Search is only one component of Information Retrieval. Too often, “search” people assume that all there is to IR is searching…

I am not putting down the value of search, but twitter is definitively a recommender system, and thus, an actual IR tool, of some sort.