When you want to get connected to a person you haven't met, you will probably find an acquaintance who knows the target person. For instance, you may send your business plan directly to VCs, but you should NOT expect these VCs will get back to you if you are not introduced by those VCs trust.

So, the first step is to look for your acquaintances who directly know the target. If you find an acquaintance who agrees to introduce you to the target, basically you’re done. However, the introduction doesn't always work. In such cases, either the acquaintance doesn't fully trust you or the target doesn't trust the acquaintance (or both) happens. (Please remember this is a simplified case: in real life, you change your attitude depending upon the context.)

Let me make cases a bit clearer. Here, "-->" (arrow) means a one-way trusted link and "<==>" a two-way trusted link.

(a) success
YOU <==> ACQUAINTACE <==> TARGET
(b) failure(1)
YOU --> ACQUAINTACE <==> TARGET
(b) failure(2)
YOU <==> ACQUAINTACE --> TARGET

Remember the purpose is to get the path from the target to you to establish a bi-directional path b/w the target and you, because you already set the path from you to the target using a trackback.

So the next step is to find ways to make these procedures automated.

Suppose the hyperlink on A to B means a one-way social trusted connection from A to B. If you have a link to the acquaintance and vice-versa and the acquaintance has a link to the target and vice-versa, the case (A) show above is satisfied and the trackback can upgrade to the "trusted" trackback.

If you don't find any acquaintance who knows directly the target person, you will start looking for a path via two people. In this case, a discounted credibility is given to the trusted trackback (Mathematical algorithm or appropriate discount rate is to be figured out.)

To make this computation more realistic, a repository database should exist which stores linking information shown above. I think such distributed repositories are maintained by small communities and such repositories are interoperable, like DNS or CA.

Furthermore, it seems more effective that each "hub" person, who has more links to "spokes" than other and therefore leads a small group, has such a repository because these "hub" people tend to like creating and maintaining social networks, even it's not paid. (It's better if such operations for maintenance are built in the weblog tool.)

This is something like small blogrolling or LinkedIn servers each of which has a bit different settings but interoperable.

---
The next step is to apply this to the individual entry (topic). As Joi and others are trying, the approach using Technorati will probably work. But need to consider it for a while.