“Everything is much more complicated. At every moment it is much more complicated. ‘They got married because they fell in love and wanted to share their life’…’he lied because he didn’t want to hurt’. What ridiculous stories! We are stratified creatures, creatures full of abysses, with a soul of inconstant quicksilver, with a mind whose color and shape change as in a kaleidoscope that is constantly shaken.”
from Night Train to Lisbon.
Transformative experience leads naturally to some form of virtue ethics. Because it is arrived at from the perspective of rational decision-making, it avoids some of the biggest challenges to a virtue ethics. Nor is it committed to any form of objective stance about virtues. On the other hand, the idea that transformative experience can be used to shape cognitive dispositions is, like virtue theory, vulnerable to the argument that stable ethical dispositions do not exist. This may seem surprising, but serious arguments – bolstered by psychological research – have been levied against the idea that reasonably stable ethical or personality dispositions exist (they may be like astrological signs) and whether they can be reasonably disentwined into separate, understandable categories. The two questions are related and they’re a serious challenge to the very idea of self-altering decisions. If stable (not unchanging but reasonably consistent) dispositions, excellent or otherwise, don’t exist, then there isn’t much point in trying to build them with experience. If they exist but our current categories don’t well describe them, we need – at minimum – a new way to catalog and think about cognitive dispositions before we can reason well about how to acquire them.
The idea that virtues might not exist is surprising. Though we are all out of practice in thinking about virtues seriously, we nevertheless dumbly accept that “he is honest” and “she is brave” are understandable sentences with meaningful content. But a lot of folk concepts turn out not to have much content. Perhaps saying “He is honest” is not much different than saying “She is an Aquarius” or “He is an INFP”.
This kind of critique is called “situationist” and it comes in two flavors. A weak situationist believes reasonably stable dispositions exist but that they are broadly contextual. On this view, dispositions apply within a business-as-usual view of life and culture. A weak situationist would describe the statement “He is honest” as carrying with it a broad contextual frame that linguistically might be of the “Within the normal range of temptations and inside and by the standards of his current micro-culture…he is honest”. A strong situationist thinks nearly all dispositive attitudes are vulnerable to change based on very slight alterations in the surrounding set of conditions. A strong situationist would likely deny that the statement “He is honest” has any real meaning at all since it would need so many qualifiers to be true as to render it useless.
Since ethical dispositions are not meaningfully different than most other types of dispositions, the strong situationist can’t just be a skeptic about ethical dispositions. The strong situationist is a skeptic about personality – the existence of reasonably stable dispositions in almost any form except skills (it’s impossible to be a skeptic about cognitive skills).
From a cognitive science perspective, it’s easy to see why the weak situationist critique is probably true. The adaptive learning model suggests that dispositions are both stable and changeable. Taking a normal person and immersing them in Nazi culture is likely to make them worse (and when it doesn’t, perhaps it will make them much better). If you must be a collaborator or a resistor, there’s not much room for muddling along as you are.
What’s more, the cognitive structures we have are often brought into conflict by real world situations. Having a stable disposition to honesty in a culture with limited temptations and high expectations of honesty is relatively easy. But when other cognitive structures (like risk avoidance or going along or getting wealthy) are brought into conflict, there’s no way (from the inside or the outside) to predict what the result will be. In other words, someone with a stable disposition to honesty would have cognitive structures that generally conform to whatever definition of honesty exists in the local culture. But the outputs of those structures might easily be overwritten by outputs of other structures when the context/environment is changed. What’s more, connection systems are utterly different than algorithmic systems. They don’t generate inexorable answers based on logical deductions. They generate weighted responses driven by complex initial inputs. We’d expect ANY connectionist system to be probabilistic and fuzzy about responses to complex stimuli.
On any plausible reading of human history and past lives, this kind of cognitive switching and fuzziness based on changes in culture or milieu is common. People think contextually, which is why having a good culture is important.
But if the weak situationist critique seems undeniable, the strong situationist critique looks implausible. Everything we know about cognitive science and connection systems suggest that they do provide stable structures that, once trained, adapt slowly to new learnings and stimuli. It’s hard to believe that a thinker whose primary tool is a large set of connection systems wouldn’t have some form of stable dispositive structures – even if all those dispositions are relative to specific cultural situations as the weak situationist believes.
What’s more, it’s hard to see other than semantic differences between dispositive tendencies to some virtue like behavior and cognitive structures implementing some kind of skill. Since skills undeniably exist, it’s hard to see why cognitive structures that are more free floating in application wouldn’t also exist.
So why take strong situationism seriously?
The impetus behind strong situationism is not common experience or cognitive science, it’s social psychology. There is a considerable body of evidence suggesting that minor changes in experimental conditions evoke substantially different dispositive responses in most people. For more variation, in fact, than we’d intuitively expect based on our general folk interpretations of character.
In Lack of Character, John Doris laid out the case for a strong view of situationist ethics. He describes a series of social psychology experiments suggesting that people’s ethical behavior is highly determined by specific and relatively gratuitous circumstances. His favorite example is the “Found Dime” experiment. Study participants were placed in a phone booth (you know it’s an old study!) in which there might or might not be a left-over dime. Then a person walked by the booth and dropped a sheaf of papers. Folks who found the dime were much more likely to help than ones who didn’t.
This seems to suggest that rather than having a fixed disposition to kindness or helpfulness, our ethical impulses are driven by tiny variations in our mood (a dime made us kind!).
The dime experiment is just one example in the situationist canon. Similar studies suggest highly situation dependent responses around honesty, courage, kindness, willingness to inflict pain, etc. Across almost every conceivable dispositive virtue, there’s a social experiment that makes it look massively contextual. In experimental testing, people’s dispositions seem to be at the whim of short-term moods, tiny variations in local conditions, and small changes in instructional emphasis.
Everyone who is alive is familiar with the way moods impact our responses to situations. But though life is constantly throwing minor perturbations at us (what work-day doesn’t involve dozens of minor ups and downs that are more important than finding a dime?), we nevertheless believe that our co-workers (and ourselves) have many fixed and easily describable dispositions.
Can the science be wrong?
In this case, the answer is both yes and kind-of. Social psychology studies are fascinating, but there’s real risk in taking them at face value. Doris’ book was written before the replication crisis that swept social psychology in 2015 and the years that followed. It turned out that a shocking number of important classic social psychology experiments could not be replicated.
That means that when experimenters tried to redo the experiment (often with much larger sample sizes), the original results weren’t observed. Many foundational studies that had become received wisdom turned out to be baseless.
It’s a reminder that while the institution of science works astonishingly well it is a human institution with the fallibility we associate with most human activities. In this case, it wasn’t that researchers were consciously misleading anyone. They didn’t falsify their results or rig their studies. They simply didn’t understand basic experimental statistics.
Many “classic” social science studies involve tiny numbers of participants. Many didn’t adequately shield the experimental apparatus. And many were simply the inevitable result of publication bias while running lots of experiments, a few of which yielded anomalous results.
So, there’s plenty of reason to doubt the authority and weight of most situationist studies, and it’s difficult to believe that strong situationist views will hold.
More recent studies also provide additional context on situationist findings. Because it does turn out that context is important and that people’s response vary quite a bit. However, it also turns out that while in any given situation a person’s responses may vary, over time, people tend to have fairly consistent response types. In other words, we see someone as kind or empathetic because in most cases they are – even if that kindness displays a fair amount of short-term variation. A really good free-throw shooter will still miss at least 10% of the time – and if the circumstances are distracting enough that percentage might be much higher. And even a poor free-throw shooter will make a couple shots out of every ten. That doesn’t mean we can’t tell the difference between Steph Curry and Shaquille O’Neal at the free-throw line.
The fundamental problem with strong situationism is that it wants to have it both ways. Doris puts an immense amount of stock in the idea that if a found dime radically changed behavior, character can’t be real. But here’s Doris trying to explain why people don’t seem completely random:
“It likewise courts misunderstanding to suppose that situationism is embarrassed by the considerable behavioral regularity that undoubtedly is observed; because the preponderance of people’s life circumstance may involve a relatively structured range of situations.”
Let’s court misunderstanding. Variable response doesn’t suggest that character isn’t real. It does suggest that it has more inherent variation than we might otherwise have imagined. It’s undeniable that our lives are constantly awash in small contextualizations. If character didn’t exist, we’d expect people’s behavior to be wildly inconsistent and unpredictable or that people have managed to place themselves in such deep ruts that they almost never find dimes, get rained on, see a good movie, get an annoying email, have a meeting run on too long, get greasy fries for lunch, hear a funny joke, or have a disappointing conversation with a partner.
Yet we see people exhibit behavioral regularity and the type of contextualizations Doris cites are far too small to be explained by behavioral ruts. That fairly stable cognitive structures exist in the brain is not debatable. We don’t have to wonder if the next piece Yo-Yo Ma plays will be awful because he recently lost a dime. And while even stable structures will sometimes yield cases where tiny perturbations in the conditions impact the outcome, that can’t be common or both stable performance and consistent learning would be nearly impossible.
That people are highly successful in predicting what others will do doesn’t refute a weaker situationism. Nearly all our predictions occur within stable contexts. The fact that we can consistently predict that a person will be honest in the completion of their duties doesn’t suggest that there aren’t significantly altered contexts in which they might not be honest at all. Many a person who is honest in a well-run culture will turn corrupt in a corruption culture. Many a person who is normal in a healthy society will turn into a beast when made a guard at a concentration camp. This kind of situationism is entirely plausible given what we know about adaptive learning and connection systems. Not only do people change, but the types of learning that are brought to bear in any situation are bound to be contextualized.
And yet…not only does weak situationism seem valid, but it’s fair to wonder whether and to what extent the virtues – as descriptive categories – capture anything real in cognitive structures. Part of what critics of virtue ethics like Doris are attacking is the tendency of described virtues to be simplified and compounded in ways that make them unsound as category descriptions.
This is firmer ground.
Cognitive structures are hierarchical arrays of complex connection systems. These systems are shaped by experience and are often combined in unexpected ways to deliver responses. One of the natural properties of connection systems is that it’s almost impossible to capture the way they function in words. Thousands or even hundreds of thousands of words are insufficient to capture the way a connection system analyzes images or plays a game. So how likely is it that a single word: courage, honesty or wisdom is going to capture the reality of what’s going on in the brain?
Not bloody likely.
It isn’t a matter of our folk culture being dim or unimaginative. It isn’t a matter of our novelists being poor at their jobs or English being an unexpressive language. Describing a person’s cognitive structure or how It will respond to the world using words is not a doable job.
It’s made even harder by the fact that virtues (and dispositions in general) are almost certainly never properties of a single system in the brain the way (the way there is for certain aspects of vision). There isn’t a courage system. There isn’t a wisdom system. There isn’t an honesty system. We can’t build up and train a connection system dedicated to a single disposition.
This is quite likely true for most dispositions. It’s why the military doesn’t teach courage – it teaches behavioral responses (drill, obedience, discipline) and emotional ties (the squad buddy system) that produce behaviors that evince courage.
It’s the same with wisdom or honesty. We don’t have a wisdom system that we can apply to any problem and that can be independently trained. Wisdom might be described as the ability to apply substantive principles from a substantial body of learned experience to new problems. What that means is that for a person to have wisdom, they must have a substantial body of real experience, they must have developed good skills at principle extraction from those experiences, and they must have developed the habit of applying those principles (bringing different connection systems in the brain) to bear on new problems. They probably also need temperamental dispositions that help them apply those skills instead of flying off the handle in anger or getting distracted by other problems.
Perhaps you think that’s a terrible description of wisdom, but no matter what your description is, it’s likely to be a whole set of natural and learned cognitive capabilities and not a single thing.
There’s a close corollary here to genetics. In folk parlance, there’s a tendency to think of the magical single gene that drives a behavior, a trait, or susceptibility to disease. Every once in a while, such a gene does exist; usually there is no single genetic marker. For height or intelligence or heart disease there are hundreds or thousands of genetic markers involved. Dispositions are almost certainly similar. There may be a few dispositions associated quite closely with a single connection structure in the brain. Dispositions that can be isolated and trained. But that’s going to be the exception not the common case. There is no reason to believe that any folk virtues (or, really, any of the dispositions we might care about) are single entities. That means that training them is always going to be difficult.
This, too, fits with weak situationism. Since the brain is a master of re-use, it’s likely that the multiple structures that drive what we blithely call courage or wisdom or honesty WILL get re-used across new problems and new situations. It would be natural to expect dispositions to transcend specific circumstances, so even weak situationism is probably too strong. But it would also be reasonable to expect that such re-use has limitations and might not always occur; the military lessons of discipline, obedience and drill might carry over into a work-place environment but not to a night on the town or a sojourn in Tahiti.
Aside from providing further support for a (fairly) weak situationism, this suggests that disassembling folk virtues into a closer mapping of cognitive structure would be beneficial. Or perhaps that we should expect experience to shape us in broad ways that are not easily categorized and are best summed up as “becoming more like him or her”. Disassembly makes it easier to see how virtues could be cultivated, which virtues are practical for us, and why we might sometimes fail (or succeed) in unexpected ways. However, even with a thoughtful disassembly, it’s likely that whatever descriptions we generate will be, at best, rough metaphors.
That’s just the reality of trying to map complex connection systems with simple words.