Origins of Polarization with Yphtach Lelkes
Ceejay Hayes:
This is CounterPol. Today we're talking with Yphtach Lelkes, co-director of the Polarization Research Lab, on the relationship between identity and political ideology. Going into this episode, I had a suspicion that identity, more so than ideology, influences how voters vote in highly polarized democracies. In searching for a meaningful distinction between political pluralism and polarization, perhaps we can look to political group dynamics and communication strategies to better understand what exactly polarization is. The U.S. is the primary point of focus throughout this interview, which leaves me wondering if citizens of other democracies recognize familiar patterns in Yip's observations. Let's find out together.
Yphtach Lelkes:
My name is Yph Lelkes. I am an associate professor at the Annenberg School for Communication at the University of Pennsylvania. I'm currently a fellow at the Princeton Center for the Study of Democratic Politics. I identify as a political scientist. I've been working on topics related to partisanship, political ideology, and polarization. what I'm most known for is this idea of effective polarization, which is something I started working on as a PhD student with Shanto Iyengar, and then Sean Westwood, Gaurav Sood, and then a lot of other people started working on this stuff as well. And the history of this stuff is that when we were in grad school in the 2000s, there was a pretty widely debate in political science about were Americans, the mass public, ideologically polarized? So were people moving to the extremes on policy issues? And the debate was essentially, if you looked at the public opinion data, Democrats and Republicans and Americans as a whole don't really seem to be moving to the extremes. they don't hold more extreme attitudes than they did 20 years ago, 30 years ago. So by that, I mean, if you ask them their opinion on certain issues like abortion or gay rights or whatnot, especially economic issues, this is definitely true. People aren't necessarily becoming more extreme. And that's one definition polarization that we often call ideological divergence. At the same time, another kind of group of people were making a claim that that may or may not be true, but what has certainly been happening is ideological sorting, which is this notion that Democrats and Republicans now hold consistently liberal slash conservative positions. it's not necessarily extremity, it's what we would call constraint or homogeneity within groups. So if you tell me that you're a Democrat, I could pretty much guess where you're going to stand on a variety of issues. Forty years ago, there was lots of conservative Democrats and lots of liberal Republicans, and this has changed a lot. So this debate was kind of raging on. We came at this with this notion that One thing that is very important in American politics is partisan identity. That Americans don't necessarily have a great grasp of political issues. We're kind of famously uninformed about politics. Most people don't think about politics, and when they do, their policy attitudes don't necessarily make much sense. But partisanship is important. Meaning, like, if you tell me that you're a Democrat, there's a 90% chance that you're going to vote for the Democratic candidate. And once we start thinking about partisanship as kind of the central focus, it kind of shifts the focus of what we're interested in from policy positions to identities. And there is a rich literature in psychology about social identities and what these mean. And going back for some years now, we know that even in minimal groups, even when these identities are just randomly assigned, you develop warmth towards your own group. So what we said is let's move the conversation beyond policy-based polarization to what we called at the time effective polarization. Even if we do or do not disagree with one another a lot more, we may hate each other a lot more than we used to. And so we went into the data and we gathered what data we could find starting from the 60s. And you see that, yes, it is certainly the case that Democrats and Republicans just dislike each other a lot more than they used to. And it is related to policy police, but it's been occurring kind of independently.
Ceejay Hayes:
What comes first, identity or ideology?
Yphtach Lelkes:
Yeah, it's a great question. And I think the answer is both matter and both may come first, but different policies may come first and different policies may shift your identities. So let me start with your identities are very heritable. There's a very strong correlation between the political identity that you hold and the political identity that your parents hold. You're born into a group, and then when you're born into this group, you adopt the policies of that group. So in that case, I think the consensus is that partisanship generally comes first most of the time. Politics is dynamic though, and parties kind of shift over time, and there are moments in history when people feel so strongly about an issue that they will switch their identity. This has happened a few times in history. I think the most famous example is when the Democratic Party passed the Civil Rights Act in the 60s. You had a lot of white Democrats in the South who cared the most about racial issues. And when they saw that the Democratic Party was trying to be more inclusive, they switched their identities to the Republican Party, right? There's work by Vanya Washington showing that essentially what you have is in the 60s, Democrats who were a lot of them were poor and held economically liberal policy positions because of the race issue, they switched to the Republican Party. And then over time, they started adopting the policy positions of the Republican Party. One policy switched people's identities. And then once you hold an identity, you start trying to fit in with the group. And you listen to the elites, and the elites tell you if you're a Republican, you also believe in free trade, and you also believe in low taxes. So this relationship is dynamic. So the answer, I guess, which comes first is you're born into an identity, but it is malleable when the environment changes.
Ceejay Hayes:
You've done some research on the relationship between, like, close family ties and kinship and closed-mindedness and open-mindedness. Can you just explain a little bit about that research and how that remains relevant today, especially in Western democracies where you would think that there are actually more community-based ties and it's less family akin, but you still find quite a bit of closed-mindedness?
Yphtach Lelkes:
So this kind of relates to what comes first. policy or identity. The work that we did there, which is heavily based on work by like Ben Enka and Joe Henrich at Harvard, is that we have a value system that served a function for our ancestors. And the value system was kind of based on the environment that we lived in. So if we lived in an environment where we had to interact with lots of strangers, we would have more of an open worldview. We'd be more trusting of strangers. And if we lived in an environment where instead we were heavily reliant on each other, like a small farming community, a clannish community, and we didn't interact with outsiders much, and outsiders could even be potentially a threat. So what you see is that like areas where there's more malarial risk, where there's more risk of disease coming in from the outside, people are less trusting of outsiders. So that's the work of Joe Henrich and NNK. And then this carries on through socialization. And the effect is not particularly big, but it's still there. But if I know who your ancestors were 100 years ago or 200 years ago, there's still a correlation between how tight-knit that group was and how ideologically conservative you are, especially on cultural issues. Cultural issues are strongly correlated with authoritarianism, and they potentially pose a threat to the status quo. People who are more authoritarian, people who are more closed-minded, are more wary of a disruptor to traditions. Family kinship is kind of a proxy for how tight-knit the group was. So if you married within the family, if there's a lot of cousin marriage, endogamy, whatnot, it's indicative of you coming from a society where you don't really interact with strangers. What we find is that knowing who your ancestors were, 100, 200, 300 years ago, the kind of ethnic group they came from, the kind of family structure they came from, is correlated with their cultural attitudes. If we're worried about threats toward the group, if we're worried about threats to traditions, then we're probably going to go to the party that is So what you see in some of our data, not all, is that there's a consistent link between cultural attitudes and your ancestry. But among those who are politically engaged, they also become economically conservative. And in our mind, the story there is that they go to the party that protects them culturally. They see what the party's positions are on other stuff, and then they gather the package wholesale.
Alan Jagolinzer:
You mentioned identity, and then you mentioned policies. Is it possible those two can disconnect? What I've seen is some cases where identity seems to override policy, so it seems to be much more dominant, and in fact, a lot of the policy that's behind the identity is actually self-defeating. and if they actually cared about policy, they wouldn't align with that identity. Can you speak a little bit to how those two forces interplay? Because there's a lot of hype around identity, and I just see people picking identity when, in fact, it's very, very, very clear they're undermining their own interests.
Yphtach Lelkes:
So this is kind of like the what's the matter with Kansas hypothesis. I don't know if you remember that from 2004, but it was essentially why are poor white Republicans voting for the Republican Party. When that came out in the early 2000s, there was a lot of pushback, but it seems to be the case now. There's a couple answers to that. One is that on less salient policies, this is certainly the case that identity always trumps policy. And there's a study by Barber and Pope in APSR that basically Trump in 2016 was just flip-flopping all over the place. Sometimes he was liberal, sometimes he was conservative. And when you show Republicans Trump's position on these issues, they would flip-flop as well. You know, they just would follow the leader. And this notion of following the leader is pretty well established in political science. So there's a couple explanations. One is that there's scope conditions to that, where those elite cues don't work on all issues. They work on non-salient stuff. They work on stuff where people don't have strong attitudes initially. The other aspect, which is kind of related to Larry Bartels' work, like Homer gets a tax cut, is that people, Americans especially, really don't know much about politics. And if you ask someone on Medicare who's going to protect Medicare, and they're a Republican, they're going to say the Republican Party. And this gives politicians a lot of leeway. They can use rhetorical strategies to be ambiguous about what they're actually going to do and actually what they want to do. The other thing is you also see that parties are shifting over time to match their electorate. So it's not always just unidirectional. The parties are persuading voters to pick up issues, but on certain issues voters are less movable and you see parties beginning to move towards the voters on issues where they're okay with moving on. You see this with like the Republican Party's increasing increasingly becoming populist and anti-free trade. And that's probably because their base is increasingly white working class and would support those policies. Again, these things are dynamical. I think the effect generally starts with parties to people, but parties also change their platforms over time.
Ceejay Hayes:
I think there is this understanding, just at a very basic level, if we're looking at effective polarization, those who are aware of how polarized the U.S. is, or who are just feeling more polarized within the U.S., kind of locate that shift in 2016 with the rise of Donald Trump and his ascension into the Republican Party. But there's one of your papers where you sort of locate the start of the deepening affective polarization at the start of the Reagan era from 1980. Can you just give us a quick history of the rise of affective polarization since the 1980s, and if you feel comfortable, the origins of affective polarization in the U.S. specifically, where you decide that sort of originates from?
Yphtach Lelkes:
I think we're political scientists. had been arguing is that this stuff really started earlier. It started in the 60s, but then ramped up in the 80s. And the Liliana Mason argument is essentially that what happened in the 60s is that the parties sorted. And by sorting, I mean, both ideologically, but also demographically, right? So that the Democratic Party became more diverse and became more urban, and the Republican Party became rural and white. And eventually, especially in the 80s, it became much more religious. Lilliana Mason's argument is that the party essentially became this heuristic for your outgrowth. That if I knew what your party was, then I knew you were probably different from me or the same as me. If I was a Republican and I knew you were Democrat, I probably not only knew your policy positions, because there's this ideological story as well. I knew you were probably pro-life. I knew that you were maybe a minority, and I knew you were probably not an evangelical Christian. And so the process started in the 60s among elites, but then it took a while to trickle down, and I think it really came to a head In the eighties, when the moral majority came in and then the evangelical party said, we are definitely Republican in seventy six, we had an evangelical democratic president. There's a bunch of other stuff that's happening simultaneously. Everything feeds each other. You can't say it's this or that, but there are kind of these engines of polarization that are occurring that speed up this process. So in the 80s as well, you have the rise of 24-hour news. Education levels go up. And when education levels go up, people become more partisan. And then you start having this rise in constitutional hardball and partisan warriorship. People usually point to Newt Gingrich in that. So as I kind of said before, we in this lab, and I think in political science, very much believe that the fish rots from the head and that the problems start with elites. Citizens are looking to their groups, to their prototypical group members to figure out, how am I supposed to feel? What am I supposed to think? And when they see that political leaders in their party hate the other side, then they're going to hate the other side. If you look at the data, you see elite polarization rising and then jumping after the Civil Rights Act, and then jumping again in 1994. And there's a lag that citizens are trying to figure out what to think. And there's variation by how attentive they are. you have interest groups who are becoming more political as well around this time. Like the NRA was 40 years ago, basically an outdoors club. Now it's clearly a political interest group that makes the engine go faster. So you have all these kind of inputs, whether it's the media, whether it's the elites, that are kind of at the mezzo level that are kind of speeding all this up. You have the internet starting in 2000. There's some debate about the role of the internet. And then 2016, I think we're already kind of at a ceiling here. I'm not sure it's at the rise. I think we've peaked. And what you see actually in 2016, is a rise in what some people call negative partisanship, which is that what is driving kind of the difference in effective polarization in 2016 is not like loving your party. It's just everybody hated their own candidate. Democrats didn't like Hillary Clinton very much. Republicans didn't like Donald Trump very much, but they really hated the other party's candidate. So that's what you see in 2016. I'm not convinced that it's gone up that much since then, to be honest. In the past 10 years, the way we collect survey data has changed a ton. The people who are answering surveys are very different because it's all done online. And so I'm not sure how much we can compare trends pre-2008 to what we see now with survey data.
Ceejay Hayes:
You mentioned earlier the fish rotting from the head. And so elites taking on more extremist policy positions and then pulling the electorate towards them in this sort of way that like identity trumps ideology. Even if the electorate is not that extreme themselves, they align themselves with the party and thus co-sign a sort of extreme policy framework. So there is a sort of like ideological asymmetry. Why does that exist? Because you would imagine that in a democracy, if someone is too extreme, you know, like, actually, I don't really believe this, you know, I don't align myself with this. I'm not going to vote for that, even if they associate with a particular party, but that does not play out. What motivates someone to enable a policy position that they don't particularly believe simply because it aligns with that ideological identity?
Yphtach Lelkes:
There's a couple of questions in there. So elites, their main goal is to build a winning electoral coalition. So they can't go way off the rails or they alienate too many people in the middle. So there's kind of this incrementalism that occurs where Republicans and Democrats are trying to build a coalition and pull people from both sides. They want to not alienate people that are if you're a Republican to the right by becoming too centrist and you also don't want to alienate people from your left who are too much in the middle. And then, because of identities, people start adopting the policy positions. So there is a slow kind of persuasion that's occurring. And this is how the polarization may occur, is that there's this, again, dynamical relationship as the policymakers kind of move slightly to the extremes, they persuade the middle, to come with them, and they no longer alienate them. Now you can move even further to the extreme and bring in even more policymakers. So it's this interaction between party has an enormous effect on persuasion, but people try to hit this middle ground of not alienating too much, but the process evolves.
Alan Jagolinzer:
I have a question to push back a little on that or to probe that. You mentioned elites going off the rails and or some sort of kind of control mechanisms, but what I'm seeing is that violence and violent threats seem to overcome all that infrastructure. And the fear of violence and kind of the energy and excitement around the potential use of violence seems to allow for a minority of a belief structure to kind of own the rights to the pathway. And I wondered if you had any thoughts about that.
Yphtach Lelkes:
Yeah, I mean, I think that we have kind of a system where policymakers are most concerned with getting primaried from the more extreme side. I suppose that's another element that disrupts any control mechanism is that we know that if you're a Republican, you will vote for the Republican candidate, no matter what their policy positions. And there's various studies that show that it takes a lot of policy positions you disagree with before you're willing to defect. Policymakers are cognizant of that. They know that if they win the nomination, everyone's going to rally behind them. So they're concerned about winning the primary nomination. People who tend to vote in the primaries tend to be more extreme. And so they're kind of shifting to meet the primary voter.
Alan Jagolinzer:
Because I was going to say, at some point, the votes don't matter. And I think, particularly in the U.S. context, it seems as though it's moving that pathway where there's alternative mechanisms to override the votes is sort of what's, at least in the last election cycle, that was the attempt. And then in this cycle, There's all kinds of ways in which the traditional notion of I'm gonna try to win via election and win via persuasion have kind of succumbed a little bit to I'm gonna win by trying to marginalize people and not allow them to vote, and or I'm gonna win by threatening people who are trying to vote, and or I'm gonna try to win by taking over the tallying of the votes and discarding some votes and things like that.
Yphtach Lelkes:
I would push back a little bit in Andy Hall's work shows that extremists tend to lose general elections. Elections are very close most of these days. There are certainly safe seats, but the ones that matter are very close and you can't alienate the independent voter. So I would point to Andy Hall's work showing that moderates still win. You saw this clearly with Biden in 2020. I think that was a take-home message. In 2022, what you saw is that in the open seats, where it was close elections, the election deniers lost. I'm not being Pollyannish about this. I think that we have to safeguard, you know, I think there's clear attempts to kind of override this, to disenfranchise people, to make votes not count. And I think that's because people know their policy positions are not popular. So because their policy positions aren't popular, what they try to do is not change the policy positions, but they try to change the rules of the game.
Ceejay Hayes:
You know, I wonder about your point on extremist losing, because what you've just described is that like political elites, political power brokers will move themselves slightly to the extreme or far enough out extreme where it makes a distinct policy shift but it's not so extreme that it alienates voters. And then in doing that, it pulls voters to their positionality. And so a position that might be extreme, say in 2004, becomes less extreme in 2012. And so it's not that extremism doesn't win, it's that it has to play a more gradual game before it can win. Like, I wonder if politicians are seeing that extremism can win if they go slower with it.
Yphtach Lelkes:
Based on the theoretical models that I've read, that would be the correct strategy, that it's kind of a shifting of the latitude of acceptance. We're talking psychology, where if you suddenly show somebody something extreme, it will seem insane. But if whatever came before it, was a little bit closer, that policy seems less insane and more acceptable. So it is kind of the shifting of the Overton window. And I think if people want their policies passed, it occurs through persuasion, it occurs through moving people along. And something like on the Democratic side, suddenly saying we need like a Green New Deal, everybody needs to do this or that. may have seemed very extreme to people. But nowadays, as people are shifting to electric cars, as people are more cognizant of climate change, it seems less extreme and more feasible. So I mean, ultimately, politics is about bringing people along with you. And one way of doing that is not saying you have to come all the way over here. It's classic social psychology. They just move people along slowly until they're already there. You know, it's kind of like the frog boiling in water kind of analogy. If you throw a frog into boiling water, it'll jump out immediately. But if you turn up the heat slowly, they boil to death. And you kind of see this with like anti-democratic policies, where if you suddenly ban Muslims or you suddenly do something, people lose their minds. But if you slowly chip away at rights, people accept it.
Ceejay Hayes:
We've talked about the foundations that make democracy polarized. But you all have a very interesting position. You are the co-director of the Polarization Research Lab. So you perhaps have one of the clear insights into towards a path towards depolarization. Can you talk to us about what that would look like, getting people to be less hostile towards outgroups and more willing to build consensus and participate in a collaborative, not so competitive democracy?
Yphtach Lelkes:
I think the problem is with political elites. And I'm not sure how many of them actually believe that Trump actually won the vote. And I don't know how many of them actually believe the insane things that they say. But they feel like they need to say stuff like this, that they need to follow Trump right or wrong. because otherwise they'll lose their elections. And so what, in my mind, one of the most effective ways of solving this issue is somehow changing the incentive structure for elites. So there are big ideas which are probably not reasonable, but probably would be very effective. And that would be something like, we have a two-party majoritarian system which incentivizes zero-sum thinking. If you win, you win and the other side is out. I have a strong inkling that if we went to a system which was kind of more proportional, where it was more collaborative, like a proportional representation system, where parties need to work together to pass laws, that they can't just say We don't need to compromise. That would probably help a lot. Whether or not that would happen, I'm skeptical. But I think the most promising, the least likely strategies are institutional. Change the electoral system. Think about how to make the Constitution have minority groups like Wyoming have less power over the rest of the nation. Make things a little more equal. I think those kinds of things would probably go a long way. What we've been doing in the lab and what we are hopefully going to release soon is this dashboard of elite rhetoric. where we've been tracking what politicians say over time and we're using large language models ingesting their speech from all sorts of sources and we are building something where voters, maybe just as importantly donors, can use it to figure out which policymakers are talking about public policy and doing the job and which ones are just conflict entrepreneurs, which ones are just behaving in a way that is dividing us not working towards policy goals, whatever they may be. Our goal is not to move people to the middle. It's to create a system where there's kind of mutual toleration and there's forbearance, where you accept the rules of the game. So we are hoping that by changing elites and the incentives for elites to kind of divide us, we may see some changes.
Ceejay Hayes:
Before I interviewed you today, I was thinking it'd be so interesting if the PRL did a ranking of politicians and gave politicians a grade on how well they work in forwarding public policy. So not being like a polarizing figure, but they actually work across the aisle and across ideologies to engage their leaders.
Yphtach Lelkes:
That's exactly what we're doing. And we're going to release a suit where we rank people. And things are like what you would expect. Someone like Marjorie Taylor Greene is incredibly vitriolic and doesn't talk about policy very much. And we hope that people will use this when it comes out, partly because politicians want to get reelected. They're single-minded election seekers. And voters don't like conflict. Voters don't like negativity. They just are more scared of the other side. Voters support these people because in their mind, they see the other side as a danger to democracy. Republicans think Democrats are a danger to society. Democrats think Republicans are a danger to society. And we are hoping that whatever we release changes the incentives as much as we can.
Ceejay Hayes:
Because it would be non-partisan. It's not saying that this politician is good for a particular thing.
Alan Jagolinzer:
I'm a little skeptical, though. I think it will be spun as partisan. What I'm hearing from particular wings are if you collaborate, any sense of collaboration is some sort of surrender to the enemy kind of vibe, and it's being spun. So the rhetoric from those who are intentionally trying to charge all this up is one of us versus them, period. It is pure zero-sum, to the point where in some cases, some people are celebrating the notion of civil war in a literal sense of zero-sum. We cannot at all abide with them, and that's why I find it interesting to get our head around it, because I do think that there is a whole swath of human beings who really just want things to just kind of be calm, and to some degree, they're being outmaneuvered, or outvoiced, represented, I don't know what the right term is, by the people who are the carnival barkers. So I actually look forward to what you're doing. I really think it could be very interesting to see.
Yphtach Lelkes:
I truly believe that most people are sick of vitriol, and we are kind of captured by the carnival barkers. And for various reasons, people have a very skewed perception of reality right now. They think that everyone in their party probably thinks this way. It's a lot of pluralistic ignorance that's going on. So I guess shining a light on this, showing people that there is a broad democratic group of people in America. That's actually the majority of America.
Alan Jagolinzer:
wouldn't go a long way. It would be interesting to correlate your list with the retirements and the number of people who are now being put up against the early primaries. How many of them are being primaried? It'd also be interesting to see if and when you're called in to testify in front of the House of Representatives.
Yphtach Lelkes:
Yeah, I can imagine the politicians don't really like being studied. So I'm a little worried about cease and desist letters.
Ceejay Hayes:
We're studying polarization from the top down. You know, I think there's a level of consciousness here, that thing like, actually, this is looking a bit insane, how divided the populations are. And I'm not sure how much of that is filtering down into the electorate population. I think they're just experiencing it. And so it'd be interesting to see if you name what is happening and you point to people and be like, these are the people that are dividing apart from just the political affiliation, non-politically affiliated. It'd be interesting to see if that becomes a motivating factor to shift electoral favor.
Yphtach Lelkes:
Jana Krupnikov, she wrote a book called The Other Divide, which is basically most people hate this stuff. Most people don't like politics. And the people who are the loudest on social media are not representative of it. But we get this biased prism, as Chris Bell calls it, of reality right now.
Ceejay Hayes:
Before I let you go, I'll just ask one more question. What is giving you, some hope, a scenario, a circumstance, a thing that's happened that points to a potential light at the end of the tunnel in terms of this hyper-polarized democracy that we live in.
Yphtach Lelkes:
In our survey data, you ask people, do you think it's the end of democracy? Are you resigned? And we find very little resignation. The people feel more optimistic About democracy, then I think the media tells us, I think the other thing is that, I mean, it's funny because I never really thought of myself as an optimist, but I do think that there's ups and downs in justice. There's ups and downs in violence and whatnot, but I think things. 20 years ago for lots of different groups. There's fits and starts, there's pushback, and there's certainly threats to democracy. We have to be on the lookout, but like standard of living is so much higher than it once was. The number of people in dire poverty around the world is much, much lower than it once was. Again, I don't want to be Pollyannish about it, and I think there'll be ups and downs, but I'm hoping we're moving in the right direction.
Ceejay Hayes:
I find it intriguing that this conversation, as many conversations had on this pod, landed on political elites and their influence on democratic infrastructures. What I found surprising, however, is the fluidity of the dynamics between political parties and voters. Influence isn't unidirectional. It seems that parties will adjust themselves to voters as readily as voters adjust themselves to parties. If this is the case, then I have more questions on the responsibility of the voter in polarized democracies. If political opportunists successfully operationalize polarizing narratives to win elections, how culpable are voters for democracy's deterioration? Thanks again to Yphtach Lelkes for sharing his knowledge with us. You can find more of his work at polarizationresearchlab.org. The podcast is produced and hosted by me, Ceejay Hayes. Alan Jagolinzer is my co-host today. This podcast is edited by Jac Boothe of Neon Siren Studios. If you enjoyed CounterPol, please rate, review, and share with friends. Thanks for listening.