The Axes of Agency

A pragmatic discussion of the value of incentives and responsibility

Being an agent implies that you can be held responsible for your actions. It also means that you can, for the most part, freely choose to make decisions according to your particular desires and preferences. Of course, you will surely make many decisions that you regret, either because they didn’t lead to your expected outcome or because you later came to realize that the outcome you thought you wanted was actually bad for you. But, especially assuming you’re able to incorporate that information and respond to it by changing your behavior in the future, you’re unlikely to conclude that you ought to give up your agency in general in response to this regret.

I’ve written about sexual consent and the importance of recognizing women as full agents in this domain. And I’ve been somewhat dismissive of arguments made by

Louise Perry

,

Mary Harrington

,

Walt Bismarck

and others, that average sex differences across the Big 5 personality traits, are particularly relevant here. It’s not so much that I find it implausible that higher average agreeableness means more women than men have sex “out of politeness”, it’s just that I think that’s their problem, and only their problem, to deal with.

Louise Perry may be right that many young women would be better off if they followed a self-imposed rule to not sleep with men until they’ve demonstrated a sufficient level of commitment. But if they break their own rule that’s still on them. This doesn’t mean we can’t or shouldn’t judge any potential tactics used by the men they interact with as unethical, but we can also judge the woman for failing to live up to her own principles and values. And of course, using a self-imposed rule as a heuristic to improve your outcomes doesn’t imply any loss of agency.

On the other side of the coin, I’m sure that more men have sex with women they think of as “below their standard” out of desperation after a sufficiently long dry spell. If they get a girlfriend a few months later and regret sleeping with some woman while in their horny and deprived mental state… oh well! It’s not that I couldn’t respond with empathy if a friend was sharing their feelings of regret after either of these situations, but, largely for pragmatic purposes, I wouldn’t consider their agency to be meaningfully complicated by the context.

Agency, moral responsibility and justice

I don’t believe in libertarian free will, so in the deepest sense I don’t believe anyone is truly morally responsible for their actions. I take the Sam Harris view that violent psychopaths are supremely unlucky to have been born with or have had experiences which led them to have such a twisted mind. Because of this I favor criminal punishment that’s optimized to deter crime and keep known dangerous individuals away from everyone else while minimizing the harm done to the criminals themselves. To some degree we should also weigh the desire we have for retributive justice above and beyond deterrence and safety, but I think a strong desire for retributive justice is harder to maintain without a belief in libertarian free will.

When an individual is charged with committing some criminal act, there are mitigating circumstances which could fully exonerate them from paying any legal consequences for that behavior. If someone was found to have abused a dog, but they were being held at gunpoint by someone else during the entire event and were simply obeying orders in order to preserve their own life, we’re going to view them as far less culpable and deserving of punishment. In a less extreme example, knowing that a violent criminal was themselves the victim of extreme childhood abuse might soften our retributive urges when hearing about their case but it likely won’t meaningfully change our opinion that they need to be kept out of society for everyone else’s safety. 

Perhaps Martha’s stalking of her baby reindeer is wholly understandable if we take a sufficiently broad understanding of her life, childhood and neurochemistry. But as I’ve already suggested with the case of the violent criminal… understandable doesn’t mean unimpeachable. When anyone does anything it’s technically possible to tell a causal story about precisely why, it’s just that we generally don’t have access to knowledge of all the relevant facts or a desire to find those facts out. And whether that person is going to be held legally responsible for those actions depends less on whether we can construct a story explaining their behavior and more on what that behavior, in context, tells us about their propensity to behave that way again.

Our assigning of moral blame and appropriate punishment is largely pragmatic. We typically consider how bad the action which is being judged is (in terms of harm to the victim), whether the individual committed this action for direct or instrumental reasons (e.g. did they abuse the dog because they got pleasure from it or because they were being held at gunpoint and simply valued their own safety over the suffering of the dog), and whether there was some extenuating circumstance (insanity, past abuse) that makes the desire to commit this action more intelligible to us and evokes our feelings of empathy. If there’s reason to think the action was committed for instrumental reasons, depending on what those reasons were, we might radically change our view of both the individual’s moral culpability and what constitutes appropriate punishment. 

But not all types of instrumental reasons will be viewed the same way. The instrumental reason of “protecting his own life” in the dog abuse example is wholly absolving, but if he said he did it in return for a large financial reward, we wouldn’t see him as so innocent. Again, we might find his behavior much more understandable, knowing that he didn’t get pleasure out of the act, but we would still consider him deserving of blame and punishment. In the former case, where he was held at gunpoint, we conclude that the man “didn’t really have a choice” so he didn’t exercise meaningful agency when abusing the dog. Whereas in the case where he accepted the money to do so we consider him to have willfully participated in dog torture, albeit for reasons other than sadistic pleasure in response to the suffering of another animal.

As I mentioned above, it’s conceivable to imagine a world where we have much deeper insight into the preceding causes of individual actions. A high-level view of all the dominos that were knocked down to lead to each observable event. But it wouldn’t meaningfully change our application of criminal punishment, other than perhaps to reduce our natural retributive urges, which themselves evolved for pragmatic reasons. Understandable or not, a dangerous person is still a dangerous person and incentives matter because most people who are capable of violence are responsive to the threat of punishment.

Your personality is not a mitigating circumstance 

Of course, the same external causes won’t lead to the same behavior for every person. Their unique personality, an outcome of an individual’s genetics and life experiences (and maybe randomness too, but the degree to which things are randomly determined really doesn’t matter), will mean that the same external causes lead to differential individual outcomes. But we don’t generally consider someone's personality characteristics as reasons to excuse them from responsibility. 

If a man cheats on his wife, a defense that “he’s just a super horny guy!” will, in most cases, fall on wholly deaf ears. Maybe he’d get a slightly more sympathetic hearing if he’s able to really pathologize his behavior by getting a diagnosis as a “sex addict”, but in general he’s held fully responsible. Similarly, a man cannot excuse himself from punishment after assaulting someone by pointing to his long and consistent history of anger and violence. He can’t just say “hey I’m super easily prone to anger and I typically express that anger through violence, it’s hard to be me, give me a break.” In fact, this would make us likely to make his punishment even longer since this he’s making it clear that we can expect more behavior like this in the future. And on the other side of the personality spectrum, if some woman is convinced to engage in criminal behavior by a friend or colleague, she doesn’t get to defend herself by saying “you know, I’m just a real pushover, I don’t like making anyone upset so I can’t be held responsible because I’m a totally spineless individual.” 

The fact that we can make group level observations about differences in personality characteristics, interests or intelligence also doesn’t lead us to differentially treat people within those groups (with some exceptions like age or a very low level of general intelligence, and we don’t only protect them from consequences we also limit their freedoms). We have an absolute line which determines acceptable behavior and anyone who crosses that line can be punished for doing so. Men are not given lighter sentences for violent behavior as a result of the observation that men are more prone to violence. And women are not held to different standards for aiding and abetting crime as a result of the observation that women are more agreeable. 

Stepping away from criminal actions, incentives matter for all decision making. And the most basic requirement for an incentive structure to operate is that individuals must deal with the consequences of their actions, we must treat them as agents. In everyday life these aren’t legally constructed punishments but simply the natural outcomes of different choices. If you get into work late every day for 3 months you probably won’t get that promotion you were hoping for. And if your girlfriend guilts you into attending her expensive and tiresome bachelorette party weekend, you’re going to be stuck paying for the flight and smiling through the brunches with her squealing sorority sisters. You chose to go and your history of being easily guilted into such things is irrelevant.

Minors and agency

When it comes to children though, we don’t hold them wholly responsible for their actions. This can be seen in lowered sentences for equivalent crimes relative to adults and the fact that contracts with minors are generally considered voidable. The other side of this coin of course is that we also limit the freedom of minors to engage in a variety of voluntary actions and we often give their parents the power to consent on their behalf. 

This could mean requiring a parent to co-sign for something the minor wants, like a tattoo, or it could mean allowing the parent to force their child to do something which they don’t want to do, like being forcibly taken to a school for troubled youth or even just wrangling a screaming toddler into a car seat. And in some cases, it’s the state that makes the decision on what’s allowed, such as with statutory rape laws, where a minor generally cannot consent to sex with an adult regardless of how their parents see the situation (note that in some states it’s still technically illegal for same-aged teenagers to have consensual sex with one another).

So why don’t we consider children able to legally consent?

1. They’re still developing cognitively and lack important skills related to decision making, including impulse control.

2. They can’t understand certain decisions or their long-term ramifications sufficiently to make irreversible decisions so their ability to give informed consent is not clear.

3. They’re easy to manipulate and have insufficient experience to recognize when they’re being manipulated - they’re also often in relatively vulnerable positions in relation to adults and may be easy to coerce given a fear of or desire to please authority figures.

4. We want to protect them from likely harm that may result from their consensual behavior, and we think that generally, the cost of limiting their freedoms is worth the benefit of preventing them from doing something dangerous or which they’re likely to regret

5. In some sense we feel that the parents have ownership over their kids (partially for pragmatic reasons), although we’ve come to see children as more deserving of rights and respect as individuals over time. For example, we no longer allow physical abuse of children, something seen as the prerogative of the parents in previous eras.

Lines are always arbitrary and will always lead to imperfect classification

All of these points have merit, but the strength of each claim will depend on the particular minor and the details of any given situation. We can think of adults who are not considered mentally disabled in any way but who could conceivably have their ability to consent questioned for one of the reasons listed above. But we can’t have laws that are as vague and context dependent as the reality underlying the quality of decision making for different individuals.

People need to be able to predict with near certainty whether their actions would be considered legal or not. So even if an adult is dealing with what they see as a very cognitively mature minor, they know that contracts made with that minor will be voidable and that sexual activity with that minor is prohibited - they don’t need to consider whether their subjective assessment of the situation will be shared by others. And this also means that an easy to manipulate 18 year old is legally fair game, even if most third parties would consider the interaction morally unacceptable. Basically, we simplify much of this debate around the factors that influence ability to consent down to a single variable: how old are they? And if they’re “an adult” we consider them fully capable of consent while minors are often considered incapable. 

Of course, this arbitrary line drawing means there will be plenty of cases where we question the ability of someone over 18 to meaningfully consent and there will also be cases where a capable minor is prevented from exercising their own agency. But the ease of enforcement and predictability of whether an action will be deemed illegal are worth the downsides related to the inevitable categorization errors that come with the application of blunt boundaries.

Essentially, we decide that at 18 a person is more often than not the best person to make decisions that affect them personally. Before someone is 18, we assume that more often than not, either their parents or the state is in a better position than they are to make decisions that affect them. There are some cases where we make exceptions and allow a minor to consent to something within a class which would typically require parental consent, and I’ll discuss some of these cases within the medical context in an upcoming post. 

Minors are different from adults in that they won’t always be minors!

I think it’s important to draw particular attention to the first point, that minors are “still developing cognitively and lack important skills related to decision making, including impulse control” and note that their lack of impulse control is an extenuating factor largely because it’s expected to change. A minor is not just someone with poor impulse control, they’re someone who is still in the process of developing the ability to control their impulses.

We want them to face the optimal level of consequences so that they have the incentives to develop this control but not such strong consequences that they can’t recover from bearing them. We have good reason to expect that the future version of this person will have changed significantly and in some relatively predictable ways. While we undoubtedly continue to develop as adults, the pace of that development generally slows down and its direction becomes harder to predict. 

I’ll also draw attention to the fourth point, that “[w]e want to protect them from likely harm that may result from their consensual behavior and we think that generally, the cost of limiting their freedoms is worth the benefit of preventing them from doing something dangerous or which they’re likely to regret”. There are some adults for whom similar reasoning applies and who continue to operate with a legal guardian as a result of some demonstrated cognitive deficiency. But in the majority of cases, even when we see an adult who’s making many poor (but legal) decisions, we rarely conclude that we ought to limit their freedom. 

This is because we intuit that having your freedom limited feels really, really bad! And also because the heterogeneity of individual preferences, implicit time discounting rate etc. mean that it’s difficult to know how to best maximize someone else’s personal utility function. It’s something we do to kids because it’s temporary and it’s clearly in their long-term best interest to hand over the reins, at least when they are very young. You might draw the line somewhere different than where it currently is legally, but obviously we can see that toddlers getting to do exactly what toddlers want to do would generally be bad for the toddler!

And, pragmatically it’s problematic since we can’t freely interact with people who aren’t considered to be agents. If everyone wanted to be treated as a child who would that leave to take responsibility for them? And will those people reliably act in the best interests of those they’re taking responsibility for? More importantly, if we insulate individuals from the consequences of their behavior how will they be incentivized to change and improve? When we tell a young women who regrets sleeping with some pushy guy that it was not her fault at all, as she was helpless to resist due to her agreeability, how should we expect her to behave in a similar situation in the future? It’s important to treat women as agents because it’s key to creating the incentives that can lead them either to develop their assertiveness, or to learn to avoid situations where their lack of assertiveness poses a real problem.

In patriarchal societies women’s freedoms are often significantly curtailed. Most women living in gender-equal societies can easily see why we would not want to live in those societies and can tap into a sense of gratitude for the early generations of feminists who fought for the freedoms we now enjoy. Yes, of course, freedom comes with costs, which sometimes includes the burden of regret. But those costs pale in comparison to the cost of living as a perpetual child.