The idea that “shame works”—that stigmatizing behaviors and shaming the people who do them are necessary and honorable tools of public policy—is a recurring theme in both conservative and more communitarian/paternalistic liberal rhetoric. It’s often based on personal experience, or home truths from one’s mom, and because people do sometimes say that shame worked for them I had a hard time articulating why I rejected this rhetoric so completely.

But I recently finished reading Middlemarch for the first time. Shame motivates several of the characters, and it shapes their lives in sharply distinct ways: At least one person really does clean up his act in part due to being shamed, whereas another person becomes much worse than he needed to be and yet another reacts to shame by becoming defiant and a bit superior. Looking at these divergent reactions might illuminate what’s going on when shame “works”—and why it’s not, in my view, a valid tool of social control. Spoilers below, for those who prefer to go into their 19th-century novels with as little background as possible!

The three characters involved here are: Fred Vincy, a fun-loving, basically decent guy with more money than sense; Nicholas Bulstrode, a fervent Methodist and general God-botherer who is always Bloomberging around in other people’s lives (for their own good!); and Will Ladislaw, a romantic type with suspiciously un-English forebears.

Vincy gets himself badly in debt and isn’t able to repay the family of the woman he loves, Mary Garth. He’s foolish and thoughtless, and his actions damage a family who could ill afford the money they lent him. He confesses his wrongdoing to all of the adults in the family more or less sequentially, feeling pretty awful about himself the whole time, and they try to be forgiving and gentle with him but they also don’t forget what happened. Mary’s mother, especially, gives Fred a pretty strict talking-to which would sound horrible coming from almost any other character; it boils down to, “You need to know that everyone thinks very little of you because of your past behavior,” which is never a becoming thing to say. But because Mrs. Garth lacks a general sense of personal superiority (there are some very funny bits about her obstinate belief that she must be subordinate to her husband) this speech is simply a harsh mistake, not her finest moment, rather than serving as an implicit indictment of her own character. And Fred pulls himself together. He works very hard, he gives up his gentlemanly airs and his frivolous habits, and he earns the trust of the girl he loves. It’s clear that this change was wrought in him by a series of painful conversations with Mary, her mom, and a local clergyman, which forced him to see himself through others’ eyes–a humiliating perspective shift, and precisely the thing which advocates of shaming intend to provoke.

But then there’s Bulstrode. Bulstrode is hiding a truly awful set of secrets, including complicity in theft and depriving a woman of her rightful inheritance. When a dissolute companion from his past turns up and threatens to expose him, Bulstrode turns desperate, and his actions become progressively worse as he seeks to cover up his past crimes. The creeping horror of an Agatha Christie villain begins to hang over him as he plods downward into evil.

Bulstrode reminded me a lot of me. Specifically, he reminds me a lot of what I was like in the year or two before I quit drinking. During that time, when I knew that I had a serious drinking problem but hadn’t yet quit, shame completely corroded my moral sense. It isolated me. I felt like there was nobody I could trust or talk to. I had no hope of change and no sense that there was any way out. I was able to imagine taking actions to hide what was going on, but stopping was completely unimaginable.

And then there’s Will Ladislaw. Lots of people in Middlemarch try to shame him for his dubious background, and he basically just gets angry with them. He totally rejects their judgment of him (and he’s right to), and decides to make his way in the world while paying them as little attention as possible.

Two things separate these experiences of shame: authority and hope. When we talk about shame “working,” we usually fail to recognize the importance of these two categories.

Ladislaw frankly rejects the authority of the people who would judge him. He doesn’t slink off in defeat, because he doesn’t think he’s wrong. When some guy at the Brookings Institute says we should be shaming teenage mothers, I wonder why he thinks they’d listen to a poster more than they’d listen to the actual authorities in their world, like peers and family. We live in a society of chaotic, competing authorities. If one authority rejects you, unless you already love that authority it’s really easy to just reject it right back.

And even if you accept the authority of the people who shame you, two obvious responses to shame are Bulstrode’s responses: despair; and concealment, even at the cost of another human life. It should be obvious how that solution relates to shame over an unintended pregnancy: Shame is a huge motivating force for abortion.

Bulstrode really is trapped and hopeless. He feels that he has no pathway back to membership in the community. His identity has been irrevocably destroyed by shame; and so he despairs, and all of the changes shame provokes in him are bad ones. Fred Vincy, by contrast, believes that there’s a path open to him by which, once chastened, he can be restored to the community. It’s precisely this forward path for those who mess up which virtually all “shaming as public policy” programs ignore.

My sense is that when people say that “shame works,” especially when they’re speaking from personal experience, they are thinking of situations like Fred’s. They were confronted by an authority figure whom they loved and admired, and they could easily imagine themselves making better choices and relieving their shame. That experience provokes positive change. Is that kind of confrontation the best way to provoke change? I’m skeptical–I take a quasi-Tolkein line, where many people deserve shame but that doesn’t mean that you deserve to give it to them–but it can clearly be a part of loving or trusting relationships. (Rod Dreher’s post which I linked above gives several good examples of this, and is much more nuanced than the Brookings piece.) I don’t think this kind of “positive” shaming can be done by public policy.

And when people say shame doesn’t work, they’re speaking from experiences more like Ladislaw’s—or like mine.

So they’re thinking of how often shame attaches itself to a morally-neutral fact, like Ladislaw’s parentage. Shame is closely allied with disgust, and we attach it to things like poverty at least as often as to actual wrongdoing. Few people feel guilty for smelling bad or getting their period in gym class; plenty of people feel ashamed for those things. This makes shame inherently a more suspect tool. It’s also inherently more tied to outside opinion than guilt is. This is part of its usefulness–like I said above, shame forces you to see yourself through other people’s eyes, which can be a powerful corrective–but shame does involve a kind of outsourced or socialized conscience.

Or they’re thinking of how devastating shame is without hope and forgiveness. In Recovery Options: The Complete Guide, Dr. Joseph Volpicelli and Maia Szalavitz review the evidence which finds that shame-based treatments for addiction make addicts worse. People drink and use more when they feel worthless and hopeless. The more extreme shame-based “therapy” has an absolutely awful record of creating cruelty in well-meaning staffers and worsening or even creating addiction in patients. By contrast, if you offer a renewed, forgiven identity, people will often choose it even if you are actively working not to stigmatize their existing, messed-up identity.

George Eliot’s shaming characters are often presented sympathetically, and two of the three shamed characters really do deserve what they get. Even under these nearly ideal circumstances, she shows the cruelties and failures of shame as a form of social control.