I don't see how there can be such a thing as negative reputation capital. Wouldn't that mean B believes the opposite of what A says? If you anti-believed someone in a consistent manner, couldn't they exploit that?

Of course they could exploit that. But you're not "anti-believing" the information an untrusted party is passing. Rather, you are not acting on their information in *any* fashion. You're ignoring them. In the case of the common killfile, you may not even know that they're talking at you.

Also, you are speaking only of 1-to-1 reputation-relationships. But that is inefficient. The mere fact of having to evaluate each person's reputation yourself is significant work.

It is significant work. How much work depends on how valuable the transactions you are considering are to you. It's not as if the notion of reputation capital doesn't have precedent. When I make a purchase for my business, I do quite a bit of checking on the background of individuals and businesses I am purchasing from. The amount of reference work I do depends on how much I intend to spend. Furthermore, I am much more apt to do business with individuals who have been referred by other trusted parties (a transfer of raputation capital). Also my initial investment with a relative unknown is usually small. The repuatation capital of both parties increases relative to one another as a relationship is continued.

On the other hand, you could talk about the transmission of reputations. [..snip..] For instance, when a certain infamously-low-reputation (deservedly so) individual recently joined the cypherpunk lists, others who had endured him in the past tried to relay their impressions of him. It proved very difficult to convey, and they were somewhat attacked for their efforts and not entirely believed.

You seem to view the notion repuatation capital as absolute. It is relative to each user of it. The unnamed you are referring to did indeed come in with quite a bit of negative repuatation capital attached to his name. While I might regard the opinions of others on the list as being valuable, I chose to see for myself if the unnamed person warranted his bad repuation.

In other words, he *could not* spend down to 0, despite years of unflagging effort.

In other words, I decide if an when he has "spent down to zero" with me -- even if we never directly communicate.

Now, suppose J. Anonymous Gourmand spams all of Usenet, and millions of people who have never heard of J. Anonymous Gourmand before read a plausible but false account of the disgustingness of McDonald's food. Perhaps the same detailed study, just fake. (Not to intertangle this with other issues, let's further suppose that Ms. Gourmand sneaks in underneath spam-watcher's radar, and cleverly appears to be on topic in every group.)

Nothing about her reputation has changed, but surely when her claim is read by millions it will hurt McDonalds a non-trivial amount.

Why? That depends on the sophistication of millions of Usenet readers (heh). I think your extension of this example is not useful. It's impossible not to "intermingle it with other issues" such as the substance of the message itself, the ability to verify any factual information, etc. etc. ad nauseam. Reputation is but one factor in many and has a mostly negative effect if J. Anonymous Goumand is indeed anonymous. Change J. Anonymous Gourmand to say, C. Everett Koop, and you can envision a more tangible example of reputation capital in action. me -------------------------------------------------------------- Omegaman <mailto:omega@bigeasy.com> PGP Key fingerprint = 6D 31 C3 00 77 8C D1 C2 59 0A 01 E3 AF 81 94 63 Send a message with the text "get key" in the "Subject:" field to get a copy of my public key. --------------------------------------------------------------