Latticed Intentions
From endirectional to intentional. Updated: 2008.02.19
Animal Minds
A blog about animal behavior and cognition.
Recent Comments
 | AlummaAmax on SPACKLED 2/20/07 Kacelnik… |
 | DamionKutaeff on Canine Science Forum, Budapest… |
 | bibomedia on About |
 | jlm19 on SPACKLED 2/20/07 Ramsey et al.… |
 | merluccius on SPACKLED 2/20/07 Ramsey et al.… |
So, right off the bat (sports equipment, not furry animal), I’m unsure what to make of Gary’s opening assertion that “All adaptations, at least all those that are really adaptive, are selected to advance their holders toward some biologically valued end state.” Perhaps I’m just distracted by wondering what an adaptation that isn’t really adaptive would be. But ignoring that worry, I’m just not sure that adaptations are selected to do what is stated. Organisms have the adaptations they do because their ancestors had traits that were selected. The traits were selected not to advance their holders to the “end state” (after all, the trait holders already were in the state on which selection operated, so no advancement to that state required). Nor were the traits selected to advance the descendants (future holders of the trait) to the same end state — natural selection has no such foresight.
We can, however, ask whether the current holder (bearer) of a trait gains a similar comparative fitness advantage over its competitors who lack the trait as the ancestral trait bearers did over their competitors. If so, we might decide to say that the trait is serving “its adaptive ends” whenever those ends are historically the ones that led to selection for the trait. But it would still be inaccurate to say that the trait was selected to do this. The trait was selected because of something it did (for the ancestor), not because of something it is going to do (for the descendant).
Trees, trees, but where’s the forest? I think the basic distinction Gary wants to draw between agents who may become aware of some of the (ancestrally defined) functions of their adaptive traits, and agents who have no such access can still be defended. And we can even agree to use the term “endirected” to cover both the aware and unaware ways of conforming to biological heritage (which is not manifest destiny). But I absolutely want to resist the idea that, “actions must be endirectional before being aware of their outcomes allows them to be intentional.” If you cut off your nose to spite your face, it will most likely be an intentional action which is not endirectional: selection may even have worked against such actions in your ancestors, as it will in you. Maybe it’s true that organisms have to engage in some endirectional actions (man gotta eat!) before they can have any intentional actions, but we’ve been given no reason to think it’s true that each intentional action has to be endirectional itself.
-CA
Has been revised (see above) in response to the comments of the mysterious merluccius.
I should add, my goal is to frame intention within an organization constraint that will enable me to build a robot that can act with intention.