Partisan Olive Branches: The Case of Marjorie Taylor Greene
Why praising your opponents can be rational
I want 2026 to be the year of partisan olive branches. Simply put, I want it to be the year in which people publicly recognize when their political opponents get something right. By ‘get it right’, I mean either getting it right according to your own political views or according to overlapping shared norms (these might be few in number).
My argument is that this practice can have enormously positive ripple effects for stable political cooperation across deep disagreement. And if your interest is in the collective common good—however thinly you conceive of it—then sometimes you ought to give credit where credit is due, even when doing so is personally or politically costly.
To make things concrete, I will use the example of Marjorie Taylor Greene throughout.
Marjorie Taylor Green, Guilt, and Olive Branches
If you’ve read the New York Times interview with Greene, you’ll notice that she has engaged in a series of fairly costly partisan admissions of guilt, that is, public admissions of her errors and excesses as a partisan. Multiple times this year, she’s broken with Trump and the Republican party line in ways that I think deserve recognition. She declared the war in Gaza a ‘genocide’, criticized Trump for enacting tariffs that hurt her constituent’s businesses and made it her mission to get the government to release the Epstein files — much to the chagrin of Trump and others.
Besides this, Greene has been the object of threats by Trump and his allies. Rather than backing down or responding with similar hostility, she responded in a notably unusual way. When A CNN host pushed her to clarify whether she was only speaking out because Trump's divisive rhetoric had been directed at her personally, she replied:
“Dana, I think that’s fair criticism.” And I would like to say, humbly, I’m sorry for taking part in the toxic politics.”
Let me be clear about what this is not. Commending Greene for particular actions is not an endorsement of her broader political agenda, her past behavior, or her worldview. There is a great deal to disagree with her about. The narrower claim is that some actions deserve public recognition in virtue of what they signal, not who performs them.
If people have the moral courage to be publicly honest about their own excesses, then we too should have the courage to commend that behavior publicly. Even—and especially—when it comes from people we strongly oppose. That is what it means to extend a partisan olive branch.
Why Olive Branches Matter: A Signaling Argument
But why should we extend partisan olive branches at all?
The obvious answer is that doing so is a credible way of signaling to our opponents that we have cooperative intentions—and people tend to prefer cooperating with those who plausibly signal that they are willing to cooperate in return.
To defend this point, I want to appeal to signaling theory: the study of how agents convey information about their intentions or dispositions through observable, often costly behavior under conditions of uncertainty.
Here’s how I see it. Many people on the right believe the left are militantly bent on destroying the country as they see it (and vice versa). But, there are also people who think some form of cross-party cooperation is in the collective interest. For simplicity, suppose there are two broad types of political actors: hostile agents, who view politics as zero-sum, and cooperative agents, who believe there is value in coordination across disagreement. How can someone tell who is who?
Because only you have access to your own mind, others must infer what kind of political agent you are based on what you do. So, even if you were a cooperative agent, that feature of your psychology is basically hidden to them. What the opposing partisan needs from you is a signal, i.e., an observable feature of your behavior that credibly conveys information about your intentions and beliefs.
That point about credibly conveying information is crucial. In signaling theory, what typically makes a signal credible is whether it is costly for the person to perform. In other words, when it would be harder for bad-faith or hostile agents to perform than genuine cooperators. In today’s political climate, publicly praising an opponent is incredibly costly because it risks in-group backlash and accusations of betrayal.
When we extend an olive branch then, we are not merely being polite. We are paying a visible cost in order to signal that we have cooperative or pro-social intentions.
Further Considerations: Dilution, Hostile Interpretation, and Reputations
At this point, you might doubt whether partisan olive branches work or whether they work in the way I’m saying. Here are a few points worth considering. For a start, the costliness of a signal obviously depends on who is watching. That’s why there’s a big difference between me talking to my brother about Greene versus writing about it on Substack.
Second, signals can become dilute. If everyone starts extending partisan olive branches, the cost of doing so may diminish to the point that the signal loses credibility. If every Substack writer is constantly praising their opponents, then praise ceases to be informative.
This is a fair concern. But notice that if we got to a stage where everyone was doing this, then people would just have to signal in even more costlier ways their willingness to cooperate with others. Paradoxically, it’s a feature of my view and not a bug that extensions of olive branches will be initially costly but tend toward being less costly in the long run. Perhaps politicians and activists will be forced to go beyond praise and toward proposing policies that garner cross-partisan consensus and are in the interest of every citizen.
A third point is more challenging: not everyone will interpret your signal as you may have intended.
This has happened with Greene to some extent. My reading of her recent behavior is that she has publicly recognized that partisan zealotry distracted her from what mattered most: her constituents. Others disagree. Alexandria Ocasio-Cortez, for instance, has suggested that Greene’s actions are driven by self-interest rather than genuine contrition or cooperation.
Does that mean extending an olive branch to Greene is misguided?
Not necessarily. When observers disagree about the meaning of a signal, the appropriate solution is often more information. In signaling contexts, consistency over time is what transforms ambiguous gestures into reputations that are difficult to dismiss. As my favorite music theorist Adam Neely likes to say: repetition legitimizes.
To return to Greene, I think AOC is wrong because Greene has not done this once, but several times and in increasingly more vocal ways. And, at least for now, publicly disagreeing with Trump seems to be costly. Greene's pattern of behavior is precisely what makes the signal credible, and why I think dismissing it outright is a mistake. Correspondingly, that's why it's not a mistake to extend an olive branch toward this consistently practiced behavior.
Coda: Polarization versus Olive Branches
My hope is consistent extensions of partisan olive branches can surprise political opponents and gradually make cross-party cooperation seem possible again. If the practice succeeds and becomes widespread, it may even incentivize cooperators to go further — toward joint policymaking or bipartisan legislation that genuinely serves citizens.
That said, polarization makes this practice much harder. Partisans on opposing sides are deeply affectively polarized: they attribute negative traits almost exclusively to those on the other side. When the signaling environment is hostile, even well-intentioned signals can be ignored or punished.
But the partisan dam has to break somehow. Never conceding ground, always assuming bad faith, and refusing to build bridges is political suicide.
Green has said that she felt very comfortable speaking to the progressive hosts of The View. Maybe the hosts were motivated by ratings. But they could have generated attention by attacking her instead, as others have done. They extended an olive branch, and she took it.
Let's make 2026 the year of peaceable - yet strategic - political cooperation.



I agree with the political point of this post, but want to discuss some of the ideas about signaling, and in particular the idea that signals have to be costly to be credible.
That's either not always true, or if it is, it's only under a broader notion of "cost" than what we usually intend.
The classic case where it is true is where there are two types, wealthy and poor, and everybody would prefer others think they're one type rather than the other (wealthy rather than poor). In that setting, to send a credible signal that you're wealthy rather than poor, you need a signal whose cost is prohibitively expensive for the poor, but not for the wealthy. Saying "I'm wealthy" won't do, because poor people can easily say that. But wearing a Rolex, or driving a Ferrari, will do, because poor people can't afford to send those signals.
But just move a little bit to a setting where not everybody *wants* to be understood as the same type. E.g., we're handing out free ice cream, and the choices are chocolate or vanilla. You're asked which one you want. You say "chocolate". That is a credible signal that you prefer chocolate. Is it a costly signal? In the pre-theoretical, intuitive sense, obviously not. It costs no effort or resources to mouth the word "chocolate". Why does it still work? Because the signal *would* be costly if you were the other type-it would lead to your getting your less preferred flavor. So in the relevant sense for signaling theory, it's a costly signal, in that it's an easier signal for someone who likes chocolate to send than it is for someone who likes vanilla. But in the intuitive sense, it's not a costly signal.
I think lots of boring, regular communication is like this. The signals are not costly, but they're still credible, because not everybody wants to be seen as the same type. If you ask someone where they're from, and they say "New Jersey", that's credible, because if they weren't from New Jersey, they probably wouldn't want you to think they were. It's only in the special case where everybody wants to be seen as the same type that you need signals to be costly in the traditional sense.
What's the relevance to politics/your post? I don't think there's any obvious tendency for olive branches to become less credible/costly over time, so long as you have a bunch of people who want to be seen as partisan warriors/don't want to be seen as cooperators. The idea that you'd need a signal "costlier" than olive branches only holds if even non-cooperators want people to think they're cooperators. And (sadly!) I don't think that's where we are. Though I admit it would be nice if we got there.
Great piece, Kyle. As you note, saying a good word about MTG doesn't amount to forgetting/excusing past vitriolic behavior.
I'm not sure I necessarily believe her statement about regret (internal states are unobservable), but her opinion change about the president seems genuine and works as an occasional reminder that some people supported Trump because policy promises he made (https://journals.sagepub.com/doi/10.1177/1532673X241295688).
A fashionable and comforting story is that people fell into a personality cult. Stories like MTG ending her loyal soldier posture suggest that people pay attention to substance too.