Think about a close friend. Maybe not your best friend, but someone pretty high up the list. If you ever needed a place to crash for a night or two, you’d happily call them up and they’d definitely say yes. If they were in trouble you’d go help them. Maybe you’ve been to their wedding, or they came to yours.

What would you do if they robbed someone? Let’s say it was nonviolent but clearly theft. They were walking through a parking lot, and saw a car window open with a wallet inside. They reached in, took the cash out, and then walked off. Unfortunately, a policeman saw the whole thing.

Now, what if they killed someone in a fit of rage? Or in a carefully planned murder, but with a good reason? What if they were unethical in their business dealings, or were mean to their dog?

How does this change if it’s part of a pattern? Or if you found out about it now, but it happened before you were friends? Would you have made friends with them if you knew?

In other words, would you be friends with a bad person?

The Isserow paper

Jessica Isserow discusses this question in her recent paper On having bad persons as friends1 , which I heard her discuss on the Journal Entries podcast. She starts from the moral intutition that one should not have bad people as friends, proposes and dismesses a few possible justifications, and then defends a different line of thinking that leads to the same conclusion.

This is an incredibly interesting and important question, but I disagreed with much of her thinking here. I’ll briefly go over her analysis, explain where I think she went wrong, and then put forth a few of my own thoughts.


Let’s refer to the two people here as Thelma – the subject – and Louise – the bad person 2.

Four limits to what we’re discussing:

  1. Thelma is not herself a bad person, though not necessarily a saint.
  2. Thelma does not suffer directly as a result of Louise’s badness.
  3. Thelma rarely sees Louise’s badness – it’s not salient in their friendship, though she knows about it.
  4. Louise is definitely bad, no doubt about it. Not a good person who sometimes makes mistakes.

First, a quick objection: can you even be true friends with a bad person? Aristotle says no, that friendship involves “mutual recognition of moral goodness.” Isserow cites contemporary disagreement here: “friendship can require us to act contrary to our moral obligations,” for example moving bodies.

This language is a preview of some disagreements I have. To me, friendship doesn’t conflict with morality: it creates new moral obligations. Your actions towards your friends carry moral valence within the context of the friendship that they would not have outside it. You have moral obligations to your friends. I don’t see the paper addressing this, at all.

The wrong arguments

The three arguments analyzed are as follows:

  1. Desert view: friendship is good, and bad people don’t deserve good things.
  2. Abetting view: you are abetting moral transgressions, and thereby guilty yourself.
  3. The risk view: it is imprudent to be friends with a bad person, since they may harm you.

Desert View

You are giving the friend something they don’t deserve. Isserow says this is wrong, because we often give our friends things they don’t deserve, such as sympathy and forgiveness. We don’t want to be moral book-keepers to our friend. My response: this ignores the changing context between “becoming someone’s friend” and “giving them good things within a friendship.” Deciding to become someone’s friend is completely different from deciding to stay their friend. We have obligations, moral and otherwise, to our friends that we do not have to strangers-who-might-be-friends.

Isserow is attacking the “weak form” of the desert view, but the strong one is much more interesting.

It’s also not obvious to me that “giving to the undeserving” is a vice. Indeed, I am more likely to count excessive generosity as a virtue.

Abetting View

Louise perceives her friendships as constituting an endorsement of her personhood. “I must be a decent person, look at all these great friends who willingly associate with me.” In the absence of friends, she might be forced into reflection about her character. So Thelma’s action contributes to the probability of Louise staying bad, or even getting worse, causing harm in expected value terms.

Isserow basically acknowledges this in in qualitative terms but denies that it holds to any “significant or morally interesting degree.” We are not, for the most part, responsible for our friends’ bad acts, just as we are not for their good acts. And we can attenuate the problem by being scolds: reproaching our friends.

I think Isserow here doesn’t fully reckon with the influence that people have over their friends. You tend over time to do the same thing as your friends, to act and think in similar ways. Moral responsibility comes with increasing influence. But this can act slowly. It seems to me that entering into a friendship with a bad person involves taking on a moral debt: in the near future, Thelma will be responsible for some harms because of her influence. But as her influence over Louise grows and exerts itself, and as Louise improves, Thelma repays that debt with interest.

Isserow goes on in a later section to say that there is relevant harm to other people through this friendship, but in an expressive way – we have a duty to hold particular attitudes towards some things, and friendship is an expressive act.

Risk View

Isserow here first lines up and knocks down two straw men. First, Louise may shoot Thelma tomorrow (or rob her, or cheat her, or whatever). But this is a practical risk, not a moral one. Second, Thelma’s reputation may suffer because she is friends with Louise. This too, is mostly a practical risk.

Then we get to moral risk, which is more interesting. The friendship may require Thelma to do something immoral. Isserow brushes this off by saying that all friendships involve moral risk, so this doesn’t “get to the heart” of why it’s wrong to befriend a bad person. I think the degree here matters quite a bit – surely exposure to a higher level of moral risk is bad.

The main argument

How do we become friends with someone? Not through intentional choice, but growing into them. But it is agentive and evaluative in some way. And all friends have shortcomings – we choose our friends in spite of those. So if Thelma is friends with Louise, it follows that Thelma has decided that Louise’s problems are “not that bad”. So then Thelma is “cultivating a particular fault of character”, and guilty of “an objectionable sort of moral complacency.”

This begs the question, in the formal sense. If it’s not bad in itself, then it doesn’t betray any defect in values. And if it is, then we need to explain why it’s bad, and we’re back to square one! “Murder is bad because it betrays a defect in values”. No, murder is bad for other reasons – from the outside, it may allow us to see the defect in values, but that’s not the reason.

I really feel that I must be missing something.

Let’s say that we need to wrap up a fish in a piece of newspaper. We have two newspapers to choose from: one that expresses moral views, and another that advocates all kind of immoral views. There is nothing moral about which newspaper we use. Our choice of one or the other doesn’t say anything, by itself, about which views we believe in or how important those views are to us. It is a question of the act itself: is there moral relevance to which newspaper our fish is wrapped in?

Similarly, our choice of friends only expresses our moral priorities if our friend’s morality is relevant to the act of friendship. I think that it is, surely, but the angle of “disordered moral priorities” doesn’t seem to add much.

My view

This is running long, but for now, what do I think?

  1. It is not inherently unethical to be friends with a bad person, but it can be depending on circumstance.
  2. We must be very careful in all cases to distinguish the acts of becoming their friend vs being their friend.
  3. One critical circumstance is whether they were a bad person when you became friends, and whether you knew that. Entering into a friendship with a bad person, with full knowledge of their badness, is the extreme case.
  4. Once you are friends with someone, you have an ethical obligation to remain their friends if they descend into badness, and to try and help them recover. This is part of friendship.
  5. It can even be virtuous to befriend bad people, because you may influence them to be less bad. In fact, a true friendship is itself virtuous, and so you may be introducing more virtue into their life at all.
  6. But you must be careful lest they drag you down instead of you lifting them up. Your morality will be influenced by your friends, and so you are taking a risk that you will adopt their badness. This is a moral concern – you have a moral obligation not to become a bad person.
  7. Another concern is social proof – you may “normalize” badness to other people, which would be bad – as in the Talmudic conecpt of “ma’arit ayin”.

I’ve read nothing on the philosophy of friendship before but here are some potentially interesting links to explore.

  1. Official link here

  2. This may prove distracting, as their friendship is pretty unlike what we’re describing here. But hey, it’s a good movie.