lørdag 5. mai 2007

Moldbug: Two kinds of repeaters

This is an abbreviated post. The original can be found here.

A little while ago I noted that "religion" is a confusing tag. Whether or not an ideology includes an assertion of anthropomorphic paranormal entities matters a lot to the principle known as "separation of church and state". Unfortunately, ideologies can drop their wizards and still have largely the same beliefs about the real world. This makes the principle suboptimal for defending against things such as natural law, which may be considered an example of Catholicism without the wizard. In short, we have a security hole on our hands.

The general approach when you find a security hole is to (a) fix it, and (b) figure out what-all has crawled through the hole. This is going to require more than one blog post, but we might as well start on (a).

If we have a rule for the separation of Mithraism and state, the state can freely be taken over by Baalbots, which we don't want either. So this rule is overly specific. I argue that separation of church and state is still overly specific, because a church can drop its wizards and become a nonprofit foundation in order to take over the state while leaving church policy in place.

So I suggested the terms "kernel" for a set of assertions, such as a religion, and "repeater" for an institution propagating them, such as a church.

Your personal beliefs constitute your kernel. In theory everyone could construct their own individual kernel, but in practice people are social (and lazy), so they will share kernels. Furthermore, shared kernels will cluster in groups. This lets us identify patterns and speak of "prototype" kernels, such as Methodism, which is maintained among Methodists by making Methodist assertions at one another and checking for divergence from the prototype. Continuing with the analogy to computer programming, let's call such an assertion a "packet". The receiver may choose to accept or reject the packet.

A repeater is an institution which sends packets. The point of going to some repeaters, such as a Christian church, is to accept their packets. (If you frequently reject the packets from the church you go to, you are likely to switch churches soon.) We can call a person who does this a "client". Clients have trusting relationships to some repeaters. For example, a lot of people are clients to the repeater known as Wikipedia.

Finally, we need to come up with some way of defining "good" and "bad" assertions, or packets, so that we can fix this security hole and reactivate our firewall.

Let's divide assertions into metaphysical, factual, and ethical. We can disregard the metaphysical packets for reasons explained above. Bad factual assertions are those which contain misperceptions of reality. I think a packet denying the Holocaust is a bad packet, because I think the Holocaust is well documented. That leaves ethical ones. An internally inconsistent set of ethical assertions is bad. For example, the American South around 1850 asserted both slavery and human equality.

Such packets shouldn't be so hard to detect, and yet they persist. Take the repeaters known as Daily Kos and Free Republic. Their clients disagree a great deal, but my guess is that if you polled said clients for their ethical views, they would broadly agree with one another and with the American Christian tradition. So one or both of the repeaters is transmitting either misperceptions of reality or internal inconsistencies of ethics, which means that one or both of the kernels is bad, and we should separate it from the state.

So toxic packets are flying all around us. Why?

This post is getting long and I'll continue later, but my line of attack is to divide repeaters into two groups: the "disinterested" repeaters which transmit to clients whatever the clients request, and the "concerned" repeaters whose transmission pattern derives from some other source.

Which is better? And why? Hm...

Ingen kommentarer:

Legg inn en kommentar