Warning: this comment is a ramble without a conclusion. Horses participating in tell culture? Cool. Preferences and consent are complicated.
This line of thinking seems to lead to some interesting places about the idea of consent.
I’m increasingly of the opinion that the whole notion of “consent” is socially constructed (that is, learned) — that it is desirable but cannot be assumed to be natural or inherent. People have to learn, not only to ask others’ consent, but to recognize when their consent is being asked: not only to ask “Do you want this?” but to know when someone wants them to have and express a preference.
Indeed, the idea of developing preferences of one’s own has to be learned. (Possibly the whole notion of having an identity, too.)
People raised in very controlling households seem to have trouble with this — with formulating and communicating preferences and seeking consent, rather than just ① going ahead and doing things that affect others and then seeing how those others react, or ② expecting others to do the reciprocal. They expect interactions to be, not necessarily forced, but certainly not negotiated. “Better to ask forgiveness than seek permission” is one thing as a maxim for decision-making in a bureaucratic office, but quite another thing in personal relationships!
This leads to communications problems between these folks and people who have been taught to exchange consent. For instance, “Would you like to do thus-and-so with me?” for one person can mean “I expect you to do thus-and-so with me and will be disappointed or angry if you don’t” whereas for another it can mean “I actually don’t know if thus-and-so would be worth doing for us; what do you think?”
Previously I thought that this difference was that (to put it overly strongly) people from controlling households had had their free will beaten out of them — that they had been abused or neglected in a way that made them alieve that people would not respect their preferences or dissent, and so did not bother to express any. But now I think the opposite: “just do stuff and see how others react” is the state of nature, whereas “formulate and express preferences and negotiate with others” is socially constructed.
And as a society, it seems we are demanding more and more of it. That sounds like a pretty good thing to me, especially for people whose preferences would otherwise be denied or disregarded. But it isn’t free or obvious; it’s a big structure of socially-constructed-stuff that people have to learn.
Computationally speaking, preferences aren’t free. Even if we model people as agents with utility functions (which I’m not sure we should!), having a utility function doesn’t mean having explicit knowledge of what your utility function is! In order to express preferences, an agent has to notice facts about itself, notice regularities about those facts, figure out what it might want another agent to do … and so on. All that requires brain power.
Teaching a horse to express preferences — that it can communicate something that will influence its handler’s actions, to get something done that it can’t do for itself — seems like a pretty big deal. Affirmatively communicating about a specific action is more “consent-like” than, say, merely expressing an emotional state of dissatisfaction or contentment.
I get the sense that people who live with animals generally do have a notion of what the animals like or dislike. But that isn’t the same as communicating preferences or consent.
On zoophilia/bestiality, I at one point thought something like: “A dog or horse can obviously express dissatisfaction with physical acts it doesn’t like — by pulling away, kicking, biting, etc. Some animals can clearly ‘propose’ sexual acts with humans, such as a dog humping a person’s leg. And we don’t expect people to seek animals’ consent to a hell of a lot of things that we do expect them to seek consent from humans — such as medical treatment or being put in a cage. So what’s the big deal?”
But a dog humping someone’s leg isn’t proposing a sexual act or consenting to one; it’s initiating one. If a human did the equivalent, to random people they didn’t have an existing relationship with, well, we wouldn’t want to put up with that sort of thing.
People (and, I suspect, horses) have different degrees of insight into their own preferences. It is perfectly possible to be wrong about your preferences: to believe that you would be happier if you ate a bag of candy, when in fact you would give yourself a stomachache and be less happy.
Imagine a woman sitting at the bar. The woman knows what she’s doing and knows that when she smiles in a certain way at a man there a 90% chance that the man will approach her, however only in 10% of the cases the man has an idea that the woman did something to make the woman approach.
If the woman initiates an interaction like that does she have informed consent? Is there some ethical imperative for her to inform the man that she initiated the interaction?
To frame the question in another way, if all you are doing is trigger the system 1 of the other person do let the person engage in certain actions, but you never ask a question to give system 2 the opportunity to reflect, do you have consent?
If it is indeed the case that everyone knows for certain what the signals mean, then they can be very specific communications of intent and consent: there is not actually any guessing going on! But if the point of using facial expressions and gestures rather than words is that the former are deniable, then it probably can’t be the case that everyone knows for certain: deniability relies on ambiguity.
If two people have slightly different interpretations of what the signals mean, then they can end up with extremely divergent interpretations of what happened in a particular exchange.
For that matter, if everyone in the bar grew up in the same town and went to the same schools, that’s a pretty different situation from if the bar is an assemblage of people from wildly different backgrounds who happen to have landed in the same location.
(I may be computing from stereotypes in saying this … but I expect that guess cultures prize uniformity, and fear diversity as a source of confusion; whereas tell cultures may consider uniformity boring, and prize diversity as a source of novelty.)
Sexually, it seems to me that if all you are doing is triggering the System 1 of the other person and neither person is waiting around for System 2 to engage and reflect, that may be very hot indeed — Erica Jong’s “zipless fuck” — but the failure modes are correspondingly huge.
If it is indeed the case that everyone knows for certain what the signals mean, then they can be very specific communications of intent and consent: there is not actually any guessing going on!
It’s possible to send signal A and the other person not understanding what the signal means and doing nothing.
But it’s also possible that they don’t understand the signal but the signal causes them to feel a certain emotion and that emotion lets them engage in an action without them having any idea of the casual chain.
The more I learn about how humans work the more I get those practical ethical dilemmas. Even worse, to really know what I’m doing I have to experiment and I’m curious ;)
A general training in do want, don’t want for ordinary things like blankets and types of food could go a long way to solving the problem.
Warning: this comment is a ramble without a conclusion. Horses participating in tell culture? Cool. Preferences and consent are complicated.
This line of thinking seems to lead to some interesting places about the idea of consent.
I’m increasingly of the opinion that the whole notion of “consent” is socially constructed (that is, learned) — that it is desirable but cannot be assumed to be natural or inherent. People have to learn, not only to ask others’ consent, but to recognize when their consent is being asked: not only to ask “Do you want this?” but to know when someone wants them to have and express a preference.
Indeed, the idea of developing preferences of one’s own has to be learned. (Possibly the whole notion of having an identity, too.)
People raised in very controlling households seem to have trouble with this — with formulating and communicating preferences and seeking consent, rather than just ① going ahead and doing things that affect others and then seeing how those others react, or ② expecting others to do the reciprocal. They expect interactions to be, not necessarily forced, but certainly not negotiated. “Better to ask forgiveness than seek permission” is one thing as a maxim for decision-making in a bureaucratic office, but quite another thing in personal relationships!
This leads to communications problems between these folks and people who have been taught to exchange consent. For instance, “Would you like to do thus-and-so with me?” for one person can mean “I expect you to do thus-and-so with me and will be disappointed or angry if you don’t” whereas for another it can mean “I actually don’t know if thus-and-so would be worth doing for us; what do you think?”
Previously I thought that this difference was that (to put it overly strongly) people from controlling households had had their free will beaten out of them — that they had been abused or neglected in a way that made them alieve that people would not respect their preferences or dissent, and so did not bother to express any. But now I think the opposite: “just do stuff and see how others react” is the state of nature, whereas “formulate and express preferences and negotiate with others” is socially constructed.
And as a society, it seems we are demanding more and more of it. That sounds like a pretty good thing to me, especially for people whose preferences would otherwise be denied or disregarded. But it isn’t free or obvious; it’s a big structure of socially-constructed-stuff that people have to learn.
Computationally speaking, preferences aren’t free. Even if we model people as agents with utility functions (which I’m not sure we should!), having a utility function doesn’t mean having explicit knowledge of what your utility function is! In order to express preferences, an agent has to notice facts about itself, notice regularities about those facts, figure out what it might want another agent to do … and so on. All that requires brain power.
Teaching a horse to express preferences — that it can communicate something that will influence its handler’s actions, to get something done that it can’t do for itself — seems like a pretty big deal. Affirmatively communicating about a specific action is more “consent-like” than, say, merely expressing an emotional state of dissatisfaction or contentment.
I get the sense that people who live with animals generally do have a notion of what the animals like or dislike. But that isn’t the same as communicating preferences or consent.
On zoophilia/bestiality, I at one point thought something like: “A dog or horse can obviously express dissatisfaction with physical acts it doesn’t like — by pulling away, kicking, biting, etc. Some animals can clearly ‘propose’ sexual acts with humans, such as a dog humping a person’s leg. And we don’t expect people to seek animals’ consent to a hell of a lot of things that we do expect them to seek consent from humans — such as medical treatment or being put in a cage. So what’s the big deal?”
But a dog humping someone’s leg isn’t proposing a sexual act or consenting to one; it’s initiating one. If a human did the equivalent, to random people they didn’t have an existing relationship with, well, we wouldn’t want to put up with that sort of thing.
People (and, I suspect, horses) have different degrees of insight into their own preferences. It is perfectly possible to be wrong about your preferences: to believe that you would be happier if you ate a bag of candy, when in fact you would give yourself a stomachache and be less happy.
Consent is really tricky.
Imagine a woman sitting at the bar. The woman knows what she’s doing and knows that when she smiles in a certain way at a man there a 90% chance that the man will approach her, however only in 10% of the cases the man has an idea that the woman did something to make the woman approach.
If the woman initiates an interaction like that does she have informed consent? Is there some ethical imperative for her to inform the man that she initiated the interaction?
To frame the question in another way, if all you are doing is trigger the system 1 of the other person do let the person engage in certain actions, but you never ask a question to give system 2 the opportunity to reflect, do you have consent?
Guess cultures are really tricky!
If it is indeed the case that everyone knows for certain what the signals mean, then they can be very specific communications of intent and consent: there is not actually any guessing going on! But if the point of using facial expressions and gestures rather than words is that the former are deniable, then it probably can’t be the case that everyone knows for certain: deniability relies on ambiguity.
If two people have slightly different interpretations of what the signals mean, then they can end up with extremely divergent interpretations of what happened in a particular exchange.
For that matter, if everyone in the bar grew up in the same town and went to the same schools, that’s a pretty different situation from if the bar is an assemblage of people from wildly different backgrounds who happen to have landed in the same location.
(I may be computing from stereotypes in saying this … but I expect that guess cultures prize uniformity, and fear diversity as a source of confusion; whereas tell cultures may consider uniformity boring, and prize diversity as a source of novelty.)
Sexually, it seems to me that if all you are doing is triggering the System 1 of the other person and neither person is waiting around for System 2 to engage and reflect, that may be very hot indeed — Erica Jong’s “zipless fuck” — but the failure modes are correspondingly huge.
It’s possible to send signal A and the other person not understanding what the signal means and doing nothing.
But it’s also possible that they don’t understand the signal but the signal causes them to feel a certain emotion and that emotion lets them engage in an action without them having any idea of the casual chain.
The more I learn about how humans work the more I get those practical ethical dilemmas. Even worse, to really know what I’m doing I have to experiment and I’m curious ;)
That seems like a huge leap in terms of capability, though, to add the free parameter of “condition to be started/stopped” somehow.