My guide to assessing agency

[Content note: mostly personal musing I wanted to get out of my system before finally turning the page on this general topic.  Perhaps anticlimactic, but I hope this succeeds in tying some things together.]

My last three posts here have focused on the grand metaphysical debate over free will.  Well no, not really.  For the most part, my primary concern hasn’t been in directly tackling the question of the existence of free will, but in treating the practical consequences of interacting with people on the assumption that a certain degree of agency lies behind their choice-making.  And that’s not mentioning other essays I’ve put up here (also about degrees of agency, or which agents get moral responsibility, etc.) which generally relate to this.

As far as I’m concerned, it’s perfectly justified for me to keep harping on this one general area.  To me, the question of how to evaluate choices and in what way we could do better (or worse) necessarily needs to take front and center stage if we seek to arrive at normative truths (i.e. answers about the morally optimal paths), and that is in large part what I’m interested in nowadays.  However, I feel that before I can put the topic to bed (for now; I make no promises of not bringing it up again!), one final aspect of it still deserves discussion.

You may have noticed in all the thousands of words of my last three blog posts, I managed to almost completely avoid the question of my own personal relationship to the contrasting approaches of assuming high agency (libertarian free will) or low agency (deterministic mechanisms behind choices), whether I usually prefer the high-agency (free-will-ist) goggles or the low-agency (determinist) goggles.  Well except of course that I’ve denounced the practice of always relying on either pair of the goggles, but I haven’t really gone beyond that into whether one side is preferable (either to myself personally or for everyone to aim for) or if there is some practical approach to follow for evaluating the level of agency of someone’s decision, etc.  That is what I want to write about today.  Although I’m afraid I’m not going to arrive at any really solid or satisfying answer in this essay, I feel that this sequence of posts would still be lacking something if I didn’t at least try to talk about it.

Let me start with my initial naïve solution, which goes back to the preferred framework I outlined here, where our evaluation of agency is for all means and purposes equivalent to our evaluation of how we should react to someone’s decision.  The “algorithm” is stated in very rough terms as follows.

The degree of agency behind a given decision is proportional to the likelihood of such a decision being altered by treating it as coming from a position of high agency.

Or in other words, someone contemplating doing X should be assumed to hold a high level of agency if and only if telling them “X is bad, mkay?” (perhaps along with supporting arguments for X being bad) is likely to change their decision.

Note that by proposing this “answer”, I’ve added exactly nothing to the discussion, beyond what I said in the above-linked post when I characterized free-ness of choices that way in the first place.  How do we know we’re reliable at making assessments of the dependent variable [likelihood of decision being altered by reacting as though it comes from a given amount of agency]?  What should we expect the overall distribution of results to look like?  Should we expect to find that choices are fairly high-agency in general, or fairly low-agency?  I have failed to provide answers to any of this.

The above attempted solution goes hand-in-hand with another general and even more useless assertion which is almost its corollary: there is a significant range of degrees of agency behind human actions, and it takes objective intellectual honesty to assess them properly.  In other words, be careful not to assume one extreme or the other.  This pretty much follows tautologically from my recent rants about the potential traps coming from high-agency goggles and low-agency goggles, or even just from my much more general spiel about keeping one’s mind open to “inconvenient” possibilities.  Take off your favorite pair of goggles once in a while (or make sure to keep switching them), open your mind’s floodgates to discomforting gadflies at least a crack, etc.  Same old stuff.

None of this answers the real question of which pair of goggles yields a map of the world that is actually closer to the territory, or whether I personally am more of a high-agency goggler or a low-agency goggler.  That is for me a tricky one, and I’m obviously still doing plenty of self-reflecting over it, but I’ll attempt a mildly stream-of-consciousness account of my personal approach below.  Spoiler alert: I still won’t come to any conclusion which gives a concrete answer going beyond the tautological one I stated just now, but I hope maybe I’ll arrive at some small insight along the way.

I reached the level of maturity where I began to form my own full-blown independent belief system around high school age, which I think is fairly typical.  Starting with that time, I can roughly break up my life into three major segments excluding the one I’m in now (I try to avoid extensively analyzing periods of my life until they’re well over): high school, college, and graduate school.  Although I wouldn’t characterize these stages precisely in terms of my views on free will or which pair of goggles I was wearing at the time, this partition provides a useful frame of reference for describing my personal journey.

I started out leaning very strongly in the direction of low-agency.  Determinism was simply the correct way to model the universe.  It followed directly from naturalism and the scientific method, which ran counter to irrational things like religion and was the mode of rationality itself as far as I was concerned.  And viewing people’s behavior in terms of their circumstances was not the most reasonable view, but the most moral one — it was the very definition of compassion.  The more fortunate among us should always extend a hand to the less fortunate; it was as simple as that.  If I ran across someone who was evidently facing some struggle or challenge, I was inclined to think it was my obligation to help them or at least to show them mercy for whatever they might do to me, provided it stemmed directly from the battles they had to fight.  I was empathetic (or maybe it’s better to say, sympathetic) to an extreme, always trying to see other people’s hostile positions from their point of view, always playing devil’s advocate when I saw someone else lionized or demonized.  At times it drove some other people (e.g. my parents) crazy.

That was me by the time I got into college.  I’m massively over-simplifying things here, of course, and when it came to some issues, because of my relative immaturity and ignorance I was far more judgmental of others and unaware of my own privilege than even a few years later.  But I think it would be safe to call me a low-agency goggler at that time.

I consider my college years a period of profound personal change for me that is unrivaled by any other part of my life, and one of them can definitely be expressed in terms of this low-agency/high-agency thing.  I entered university as an unabashed champion of the determinism-leaning attitude and the flavor of sympathy and compassion that went with it, but left the undergraduate phase of my university life with a dramatically (though not entirely) different perspective.

The change came about as a result of a series of social experiences I had during my first years living away from my family.  Without going into any details, I’ll just say that I found myself at close quarters with some people who not only steadfastly clung to low-agency goggles themselves, but were struggling to get by under circumstances more difficult than my own.  Their low-agency-gogglesism didn’t seem to be providing them a net advantage; rather, it was further entrenching them in their ruts.  Even worse, this ideology, after a few tempting modifications that I described in my anti-determinist-goggles essay, seemed to provide these people with excuses to attack those nearest them who appeared to be more fortunate.  This led to some destructive fighting between themselves and certainly plenty of nastiness that went my way, as I was clearly perceived as more fortunate than they were (I would claim that there was indeed a disparity but that their view of it was exaggerated and distorted by the low-agency lenses they gazed through).  Moreover, my own tendency to wear low-agency goggles wasn’t doing me any favors and ultimately wasn’t even doing them any favors.  I came to recognize that my instinct to act as selflessly as possible to those near me who were struggling may have helped them in the short term, but it opened me up to manipulation and sometimes full-on bullying, and eventually some of them dragged me down further than I was able to pull them up.  By the end of this period, when I was switching schools to pursue my graduate degree, I had realized that I seriously needed to rethink this part of my creed.

My general belief system mostly hasn’t changed any further.  Starting from that point and continuing through the present day, I have kept up a conscious effort to be very guarded against indulging those people who both appear to be wearing the low-agency goggles and are likely to see me as just “free” enough to use to their advantage.  I have been mostly successful at this, although most of my closest friends have certainly been the type whose worldview leans in the determinist direction.  However, my social experience in graduate school put me close to a few people of the opposite mould to this, both in the sense of being high-agency leaning and of having more favorable circumstances to work with.  At first I remember just how refreshing it was to actually feel envy for another person again, never mind how destructive I knew that emotion could be.  But eventually I was reminded firsthand how it felt to be oppressed from the other side — by the high-agency gogglers, the ones who felt threatened on multiple levels by my struggles and ineptitudes and told themselves it was for my own good as well as theirs that they chose to be harsh with me.  A lot of the fodder for my anti-free-will-goggles essay was a product of knowing several such individuals closely enough to be able to see (at least I believe) parts of their reasoning process.

I’m sure it goes without saying that there has been a very obvious correlation between the whole high-agency/low-agency split and the backgrounds and circumstances of different people.  In general, those who have seemed better better off than I am have treated me according to what they saw through high-agency goggles, and those who have seemed worse off through low-agency goggles.  I see this as another example of how each of our worldviews are influenced (even sometimes by self-modification) by the situations we’ve found ourselves in.  And probably I too have slipped into some of the failure modes of each side depending on how high on the ladder I sit relative to the people I’m dealing with.  But because of everything I’ve seen, I’ve made a very deliberate point of being as self-aware as possible about these failure modes, which is what in large part led to my writing extensively about them.

But I’ve drifted a bit off topic.  What does all of this mean for the question about levels of agency?  Okay, so I started out as a low-agency goggles-wearing adolescent, then took a sharp turn in the dawn of my adulthood and wound up trying to stick to a straight and narrow middle road or whatever.  My original intention was to say in this essay that at my core I’m obviously still on the low-agency side and the only difference is nowadays I go out of my way to qualify it.  In some ways this interpretation seems to check out: for instance, I am passionate about certain generally liberal positions (such as prioritizing deterrence and rehabilitation over harsh punishment) while pretty much my only nods to traditional conservatism seem to be careful qualifications to liberal positions (maybe the welfare system could be abused or lead to perpetuation of poverty).  At the same time, one of my most major concerns currently is with what appear to me as excesses in the low-agency mindset among younger generations of liberals, and I would say that this critical attitude has gone beyond “careful qualification” level.  So the landscape is starting to look a little confusing and my place in it is hard to characterize as being by default on one side or another.

At the same time, there is nothing qualified about my commitment to empathy, which I wrote about a year ago and continue to stand by, perhaps with fewer reservations than I hold for anything else I’ve written here.  And isn’t empathy, almost by definition, a way of understanding, by viewing conscious behavior in terms of its underlying mechanisms?  (It can practically be seen as the human side to my pro-science, anti-supernaturalism beliefs which I also feel very close to my core.)  It’s hard to shake off the intuition that to be empathetic towards other people is to cling steadfastly to a low-agency model.  This seems to contradict the “middle of the road” approach.

Except I can now resolve the apparent inconsistency, I think, having realized that the “empathy = low-agency” idea is subtly wrong.  And really, I should have reread my own essay on “A Principle of Empathy”, and I would have known that!

A major part of my whole thesis when I made that post a year ago was that empathy is not the same as sympathy or charity; it does not imply that we excuse someone’s behavior just because we understand what causes it.  Understanding means in particular knowing just how high someone’s agency is — maybe it turns out to be quite high, and we can determine that treating it as quite high will have the desirable effect on them; maybe applying empathy allows us to conclude in this case that someone is worth of praise or condemnation for what they do.  My “Principle of Empathy” is not low-agency-goggles-ism in disguise; if anything, it’s objective-rigor-in-determining-how-high-agency-our-model-should-be-ism in disguise!

In the last few days another framing of this occurred to me.  Someone commented on last weekend’s post by suggesting (among other things) that compatibilism, the philosophical position on the free will debate which I endorse, is really just hard determinism in essentials, the only real difference being an altered definition of “free” which allows for a conclusion that more people will find palatable.  And this is true.  Compatibilism doesn’t really lie at an equal distance between hard determinism and metaphysical libertarianism.  It agrees with hard determinism that the concept of “freedom” that the metaphysical libertarians are chasing after cannot exist.  But compatibilism, from my point of view, is essentially determinism plus a concept of freedom that does have the useful feature of allowing one to talk about different degrees of agency, moral responsibility, etc.  (Weird random thought that occurs to me while writing: this Determinism Plus can almost work as an analog to the Atheism Plus movement that was attempted a few years ago.  Hmm.)

It is easy to associate the empathetic worldview with the deterministic one, because both emphasize prediction of events by causation.  And this association is perfectly valid.  But let’s not overlook the fact that siding with determinism doesn’t necessitate giving up on a model that allows for agency.  If “determinist goggles” were really only about determinism, then I’d happily confess to having stubbornly clung to them pretty much all my life.  But my conception of that metaphor is meant to imply an aversion to recognizing genuine choice and responsibility, which is a very different matter.

In the end, even after all this rambling and meandering, I can’t propose anything beyond the “caution against being blind to side or the other” idea, except for a positive assertion that there is no concrete formula beyond “caution against being blinded by one side or the other”.  Actually, I think maybe this is a general truth about any conundrum for which there are multiple narratives that each have (at least partial) validity: the only meta-level a priori rule that can safely be applied is to avoid becoming enamored with one of them while being blind to the rest.

We are not gods.  We have no way of being able to peer into the inner workings of other sentient brains to see what is driving the decisions made in them, and there’s only so far we can go towards standing outside ourselves and understanding the inner workings of our own brains.  We just have to put our best effort towards being objective, honest, and open-minded in our quest to understand these things.

I can only think of one really concrete technique that can be used towards ascertaining the truth about someone’s agency, which is best employed between two people who trust each other to be operating in good faith — we’ll call them Alex and Beth.  In order to figure out how much control Beth has over something she does, they make the following agreement: Beth will, as objectively as possible, do a little soul-searching to get the most honest possible idea of how much agency she has.  And Alex, in return, will trust and accept whatever answer Beth returns with.  (I know I’m not the first to come up with such a scheme, because I clearly remember seeing it suggested on the rationalist internet somewhere, but I can’t remember to whose credit.  I had certainly already been considering something similar at the time, though.)

Apart from that, I can only lay out a few platitudes (i.e. pleasant-sounding but valid suggestions that follow pretty much tautologically from what has already been said).

We all feel compelled to do certain things, some of which appear from others’ points of view to be very free choices, and yet we each know we have something that at least feels like free will.  Our consciousness and self-awareness comes with impressions of control and responsibility.  In order to determine how to live ethically and which choice is the Right Thing To Do, we have to consider what kind and degree of control and responsibility lie at the base of actions.

But there is no completely reliable, cut-and-dry way to do this.  We are each stuck inside our own brains.  So the very best we can do is to approach such questions in the spirit of intellectual honesty and objectivity, even knowing that pure objectivity is impossible.  We have to examine not only other peoples’ minds but our own, remembering that even those decisions which seem entirely conscious and free are heavily influenced by outside factors, and that in many of even the most uncontrolled actions, at least a little power over them can be found beneath the surface.

We should especially keep in mind that when each of us is dealing with fellow humans who look like they’re managing worse than we are, we are more biased towards assuming high agency.  Conversely, when each of us is dealing with fellow humans who look like they have it easier than we do, we are more biased towards assuming low agency.

Above all, we are struggling to wade through this together.  Let’s try to be understanding of ourselves and towards each other.  And as long as it’s warranted by what we’ve uncovered in our quest for understanding, let’s be kind to one another, not forgetting that it is no less important to be kind to ourselves as well.

3 thoughts on “My guide to assessing agency

  1. ADifferentAnonymous

    Just binge-read all your agency/free-will posts. Thanks for writing about this, as it’s an essential topic in my own life as well. Some scattered thoughts:

    I’m frankly surprised you got pushback on saying low-agency=left, high-agency=right. Seems obviously true to me.

    I’m dating someone with a lot of executive dysfunction, and I’m basically all-in on seeing them as low-agency… but I get a lot of inner gadflies asking, “what if tough love would actually work?”

    The legal concept of “strict liability” is related to high-agency goggles–except that the former openly accepts that it will make false-positives as to who should be blamed. So while reading “Failure Modes of Determinist Goggles”, I built up a steelman that said “Always hold people accountable for their actions; sometimes people really can’t help their actions, but trying to make exceptions will reward manipulators more than it spares the innocent.”

    I usually work without deadline pressure and often have trouble focusing on work for long. Once, an unusual situation created acute deadline pressure. A low-agency view of myself would say that incentives to work wouldn’t influence my behavior, and so I would remain unable to sustain work. A high-agency view would say that I would make myself endure the unpleasantness of long-sustained work due to the strong incentives to do so. What actually happened is that I found myself wanting to focus on work and having simultaneously some of the most work-focused and hedonically positive workdays of my career. This kind of flies in the face of both pairs of goggles, but I suppose it could be seen as a high-agency view of my entire brain, subconscious included.

    My deepest worry on this matter is that humans have dedicated firmware to exploiting each others’ agency-detectors, and the prevalence of low-agency ideas actually reduces agency. This could be an explanation as to why modernity sucks even though it should be great.

    Liked by 1 person

    1. Thank you for reading and for your feedback! I’m glad to finally hear from another person who sees the similar connection between political alignment and assumed levels of agency — you may be the first one I’ve heard from.

      I know from experience that this issue of where we each fall on the assuming-high-agency vs. assuming-low-agency spectrum can cause difficulties in intimate relationships. I consider it largely to blame for the eventual doom of my one serious relationship — this was a long time ago but that personal situation certainly played a major role (though not the only role) in shaping me to eventually come up with the framework I’m using in these recent posts. But I have the optimistic feeling that when this kind of difficulty arises in a relationship, it can often be overcome as long as the people involved are able to view it in these or roughly similar terms and are willing to act in good faith.

      In general and in particularly with regard to your story about how deadline pressure affected you, I don’t know if you spend time around rationalist Tumblr, but you may be interested in the recent discourse about Bryan Caplan’s views on mental illness and deserving vs. undeserving poor (which themselves are a very interesting read). In particular, my main comments are here, here, and here, but a lot of other people posted with interesting points.

      I’d appreciate elaboration on your last paragraph, as I’m not sure I understand it properly. Maybe you mean that recent trends in the direction of low-agency-goggles-ism might result in actually reducing agency, in which case I certainly share your concern.


      1. ADifferentAnonymous

        Yup, found this sequence via the Caplan stuff 🙂

        (the following will eventually expand on my last paragraph)

        As far as I’ve seen, no one else is making much use of the consciousness vs. rest-of-brain. The controversial cases (clearly mental illness, but also I think deserving-poor stuff) revolve around cases where the factors limiting agency reside outside the individual’s consciousness but inside their head. Low-agency goggles correspond to considering the conscious mind the agent; high-agency goggles to considering the whole brain the agent. And of course the true answer lies between these two extremes in a complicated way; sometimes the non-conscious can respond rationally to incentives

        Now however much we might disagree over specific cases, pretty much all humans want to hold people accountable for stuff in their control but not for stuff outside their control. I’m no anthropologist, but I’d wager this is true in some form for more or less every human society back to the paleolithic and probably other primate societies too. But as long as this is the case, this can be exploited by pretending to be unable to control something you actually can.

        Now, Robin Hanson’s homo hypocritus theory states that the conscious/non-conscious divide exists largely to exploit exactly this kind of thing.

        Now say Og is a gatherer who’s supposed to bring food back to share equally with the band, but has the opportunity to pick out and eat all the sweet berries herself. Homo hypocritus theory would predict that the non-conscious brain would have some kind of ‘can I get away with this’ logic, and if the answer is ‘yes’, the conscious mind would experience a sincere desire to be fair to the band but also an irresistible compulsion to eat the berries, allowing her to later express to the band her sincere regret that she was unable to stop herself.

        (Okay, even I find that just-so story unconvincing, but it’s for hypothesis illustration, not evidence.)

        So I am worried about low-agency goggles reducing agency, but with the complication that trying to assess the correct level of agency might be an adversarial game in a way that people don’t realize. And I’m worried that the way the non-conscious pulls off this kind of exploitation tends to involve creating unpleasant conscious experiences.

        Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s