To me status is how seriously a person is taken by others. It is about influence (not zero-sum) rather than dominance (zero-sum). Status is an ad hominem shortcut that avoids Pascal’s Mugging by giving low weight to the arguments put forth by low-status people. Often status is highly domain-dependent: for example, I would take Eliezer seriously on rationality, on fiction writing or on AI research, but not on MWI advocacy. In an extreme case you discussed before, a generally low-status person turns into a high-status one when credibly threatening someone with a bomb.
If you want a formal definition, I suppose you have to select
a group of people,
a member of the group (or maybe a non-member),
a set of domains this group cares about,
and find a distribution of the prior given to the statements of this member in a given domain. If the distribution is localized enough, it makes sense to talk about that person’s status in this domain for that group. If you find that for some person the distributions are similar across groups and/or domains, then you can talk about that person’s status in a more general sense. In an extreme case, there are people whose opinion is taken as gospel on the wide variety of issues by diverse groups of people, e.g. Einstein on anything science-related. More examples: Reagan’s quotes on anything policy-related among US conservatives, Krugman on politics/economics among US liberals, Kahneman on human thinking among LWers.
This definition allows one to actually calculate a person’s status by polling group members for priors, implicitly or explicitly.
To me status is how seriously a person is taken by others. It is about influence (not zero-sum) rather than dominance (zero-sum). Status is an ad hominem shortcut that avoids Pascal’s Mugging by giving low weight to the arguments put forth by low-status people. Often status is highly domain-dependent: for example, I would take Eliezer seriously on rationality, on fiction writing or on AI research, but not on MWI advocacy. In an extreme case you discussed before, a generally low-status person turns into a high-status one when credibly threatening someone with a bomb.
If you want a formal definition, I suppose you have to select
a group of people,
a member of the group (or maybe a non-member),
a set of domains this group cares about,
and find a distribution of the prior given to the statements of this member in a given domain. If the distribution is localized enough, it makes sense to talk about that person’s status in this domain for that group. If you find that for some person the distributions are similar across groups and/or domains, then you can talk about that person’s status in a more general sense. In an extreme case, there are people whose opinion is taken as gospel on the wide variety of issues by diverse groups of people, e.g. Einstein on anything science-related. More examples: Reagan’s quotes on anything policy-related among US conservatives, Krugman on politics/economics among US liberals, Kahneman on human thinking among LWers.
This definition allows one to actually calculate a person’s status by polling group members for priors, implicitly or explicitly.