Assuming that agency is not all-or-nothing, one should be able to measure the degree/amount/strength of agency. Is this different from, say, intelligence as an “ability to reason, plan, solve problems”? Are there examples of intelligent non-agents or non-intelligent agents? Assuming the two are correlated but not identical, how does one separate them? Is there a way to orthogonalize the two?
CFAR’s notion of agency is roughly “the opposite of sphexishness,” a concept named after the behavior of a particular kind of wasp:
Some Sphex wasps drop a paralyzed insect near the opening of the nest. Before taking provisions into the nest, the Sphex first inspects the nest, leaving the prey outside. During the inspection, an experimenter can move the prey a few inches away from the opening. When the Sphex emerges from the nest ready to drag in the prey, it finds the prey missing. The Sphex quickly locates the moved prey, but now its behavioral “program” has been reset. After dragging the prey back to the opening of the nest, once again the Sphex is compelled to inspect the nest, so the prey is again dropped and left outside during another stereotypical inspection of the nest. This iteration can be repeated again and again, with the Sphex never seeming to notice what is going on, never able to escape from its programmed sequence of behaviors. Dennett’s argument quotes an account of Sphex behavior from Dean Wooldridge’s Machinery of the Brain (1963). Douglas Hofstadter and Daniel Dennett have used this mechanistic behavior as an example of how seemingly thoughtful behavior can actually be quite mindless, the opposite of free will (or, as Hofstadter described it, sphexishness).
So ants don’t have agency. The difference between intelligence and agency seems to me to vanish for sufficiently intelligent minds but is relevant to humans. Like ArisKatsaris I think that for humans, intelligence is the ability to solve problems but agency is the ability to prioritize which problems to solve. It seems to me to be much easier to test for intelligence than for agency; I thought for a little bit awhile ago about how to test my own agency (and in particular to see how it varies with time of day, hunger level, etc.) but didn’t come up with any good ideas.
One sign of sphexishness in humans is chasing after lost purposes.
How about “agency” as the extent by which people are moved to action by deliberate thought and by preferences they’re aware of—as opposed to by habit, instinct, social expectations or various unconscious drives.
That’s pretty much similar to Luke’s definition I guess.
Is this different from, say, intelligence as an “ability to reason, plan, solve problems”?
It’s different in that it also chooses which problems to seek to solve, in accordance to one’s own self-aware preferences .
Are there examples of intelligent non-agents or non-intelligent agents?
Lots of intelligent non-agents—a pocket calculator for example.
Trying to get a handle on the concept of agency. EY tends to mean something extreme, like “heroic responsibility”, where all the non-heroic rest of us are NPCs. Luke’s description is slightly less ambitious: an ‘agent’ is something that makes choices so as to maximize the fulfillment of explicit desires, given explicit beliefs. Wikipedia defines is as a “capacity to act”, which is not overly useful (do ants have agency?). The LW wiki defines it as the ability to take actions which one’s beliefs indicate would lead to the accomplishment of one’s goals. This is also rather vague.
Assuming that agency is not all-or-nothing, one should be able to measure the degree/amount/strength of agency. Is this different from, say, intelligence as an “ability to reason, plan, solve problems”? Are there examples of intelligent non-agents or non-intelligent agents? Assuming the two are correlated but not identical, how does one separate them? Is there a way to orthogonalize the two?
CFAR’s notion of agency is roughly “the opposite of sphexishness,” a concept named after the behavior of a particular kind of wasp:
So ants don’t have agency. The difference between intelligence and agency seems to me to vanish for sufficiently intelligent minds but is relevant to humans. Like ArisKatsaris I think that for humans, intelligence is the ability to solve problems but agency is the ability to prioritize which problems to solve. It seems to me to be much easier to test for intelligence than for agency; I thought for a little bit awhile ago about how to test my own agency (and in particular to see how it varies with time of day, hunger level, etc.) but didn’t come up with any good ideas.
One sign of sphexishness in humans is chasing after lost purposes.
How about “agency” as the extent by which people are moved to action by deliberate thought and by preferences they’re aware of—as opposed to by habit, instinct, social expectations or various unconscious drives.
That’s pretty much similar to Luke’s definition I guess.
It’s different in that it also chooses which problems to seek to solve, in accordance to one’s own self-aware preferences .
Lots of intelligent non-agents—a pocket calculator for example.