Catabasis

‘Are you a communist because you read Karl Marx, Tommy? No. You read Karl Marx because you know you are a communist’

I really dislike it when people moralise or intellectually belittle their political opponents. I don’t mean when it is done in the sense of a casual jest in the company of friends, I mean it in the sense of political partisans genuinely feeling that their opponents only hold the opinions they do because they are either uninformed, unaware of their best interests, or just plain malicious. Naturally I’m going to focus on the left, because I am left-wing and almost everyone I know my own age is left-wing, and using a post about not dunking on one’s political opponents to dunk on one’s political opponents is tasteless, even by my standards. It must be mentioned though, that naturally this occurs just as much across the aisle. One of the most irksome examples of this is when fellow pro-choice advocates advance the view that those in the pro-life camp are motivated by a desire to control women’s bodies. They’re not. Almost all pro-lifers genuinely believe that abortion is the killing of a human person, with all the moral repugnance that that act entails, and thus it is completely reasonable for them to be so strenuously opposed to it. If I believed that abortion amounted to the murder of hundreds of thousands of people a year in the UK, I would be pretty morally defective were I not somewhat miffed. Of course, I disagree that foetuses in the early stages of pregnancy are human persons, and thus stems the disagreement with pro-lifers. This is similar to my hatred of non-vegans who get angry at vegans for being so pushy, as if they do it out of a frivolous desire to upset those who choose to eat meat. Again, almost all vegans believe that killing mammals such as cows and sheep is broadly analogous to killing humans, with all the moral repugnance that that act entails, and thus it is completely reasonable for them to be so strenuously opposed to it. Of course, I disagree that mammals are sufficiently similar to humans to qualify as moral agents, and thus stems the disagreement with vegans.

Of course, just because people hold laudable motivations for advancing a particular viewpoint, that doesn’t necessarily exculpate them for the consequences that those viewpoints entail. Pol Pot may have genuinely believed his actions in Cambodia would usher in a quasi-utopian society to the ultimate benefit of every future generation of Cambodians until, presumably, the heat-death of the universe. That doesn’t excuse him for killing a full ¼ of the Cambodian population and burying them in shallow mass graves during the Khmer Rouge’s reign over the country (One particularly gruesome and hopefully apocryphal detail of the Khmer Rouge’s reign of terror that has always stuck with me was that Cambodian soldiers would audibly laugh when bashing open the heads of new-borns against trees in the killing fields, because not doing so might have indicated sympathy for their victims and marked them out as potential dissidents to their superiors). With notable exceptions, I generally oppose the assumption that people who I’m arguing with are malicious. I think the left tends to take the view that most people they argue with are either systematically fooled into holding political views that are against their own interests (which the enlightened left-winger themselves has somehow avoided, of course), such as in the case of working-class people who vote for right-wing politicians, or in the case of clearly well-educated people who happen to skew right, are just rapacious apologists for a system which benefits them. The most egregious example of this I’ve ever come across is a white leftist who told me that liberals could only ever be concerned about the plight of Uyghur Muslims in China because of the economic potential such people as a potential market, rather than actual concern over the internment of religious minorities in concentration camps. It’s one thing to hold political disagreements with liberals, and another entirely to insist that liberals disagree with you because they fail to see other human beings as anything but potential sources of economic enrichment.

II.

Whenever the idea of whether ones’ political opponents should be treated as well-informed and well-intentioned comes up in a debate, I usually give a short, easily digestible answer, which I call ‘The Useful Lie’ argument for political civility. Fundamentally, it boils down to the idea that the assumption of ignorance or malice in one’s political opponents leads to political isolation and polarisation. Everyone by now has heard of implicit association tests, which were revolutionary because they reveal that people who claim to hold no antipathy towards ethnic minorities have implicit negative associations about them. It turns out, however, that if one conducts an implicit association test on the views that Democrats and Republicans in the US have about one another, the negative biases were held towards members of the opposite party are 50% stronger than those seen when shown a member of another race. We know that this deep political polarisation leads to the breakdown of the civil norms that allow a democratic nation to function.

Who doesn’t love democracy? Almost everyone across the political spectrum thinks democracy (in some form or another) is a valuable ideal towards which governments should orient themselves, from Burkean Tories (Yes, my personal Overton Window only stretches as right-wing as Edmund Burke) to Lockean liberals to the Marxist-Leninists of the CPB (The communist ideal society is ‘stateless, classless and democratic’, though the contours of Marx’s thought mean that it would be a democracy without any major political quandaries to sort out. This is a fascinating topic and forms the basis a lot of my disagreement with Marx, but that is a topic for another time). I get a lot of flack from fellow liberals in that I don’t see democracy as an intrinsic good, in the way that St John (Rawls, not Apostle) did. I tend much more towards instrumental justification, à la James Madison’s contributions to the Federalist Papers. The main thrust to this stems from a certain pessimism about the prospect of good government (‘If men were angels, no government would be necessary. If angels were to govern men, neither external nor internal controls on government would be necessary’[1]), insofar as I feel that democracy isn’t *really* valuable because it allows a nation to vote in good governments, but that it allows us to vote out bad ones. No matter how benevolent one believes a prospective dictator to be, short of truly divine omnibenevolence, one has no guarantee that this good behaviour will carry on once the keys to the presidential mansion have been handed over and the secret handshake that unlocks the nuclear codes revealed (This is also broadly analogous to my views on free speech. I mostly don’t buy idealist spiel about the value of differing opinions and the free market of ideas, because most views held by most people are not valuable or even true, though I am open to arguments from epistemic humility. My love for the freedom of expression comes mainly from not trusting anyone to be responsible when given the power to restrict it). One of the most resounding signs of the widespread acceptance of the ‘democracy good’ line of reasoning is the pains through which hideously authoritarian regimes go to pretend they are democratic in nature. I believe that this is a pretty serviceable justification for the necessity of democracy, and so if preserving democracy requires us to pretend that our political opponents aren’t stupid at best and evil at worse, at least we know we’re suspending our own beliefs for good reason. The reason I use this ‘Useful Lie’ argument is because most people I know implicitly already agree that democracy is a desirable end, and it also doesn’t challenge the core beliefs that one is either better-informed or more moral than one’s opponents. After all, almost everyone believes that they are more moral and more intelligent than the average person. Either the semantic and statistical connotations of the word ‘average’ must have flexed considerably without me noticing, or some people are a little too generous in their self-appraisals.

The only problem is, this isn’t actually the reason for my insistence on what may be called ‘political civility’. I use it because it is easy, and I am an undying proponent of expending the minimal amount of effort required to achieve a goal (the goal here not being convincing people of my viewpoint, but simply coming off as insufferably smug). The real reason requires some explaining, so here goes.

III.

Did you vote/would you have voted for Brexit? Now, what is your opinion on abortion? Should the UK spend more on its military? Do you think that differences in average personality traits between men and women are innate or socialised? The chances are, if you’re for remaining in the EU, you are also a pro-choicer, and you are also likely to believe that the UK should cut down on military spending. You’re also highly likely to believe in a vast variety of other predictable political viewpoints, from the need to respect the separation of church and state, the need to promote immigration and the necessity of gun-control laws. If you’re for leaving, you also probably hold a very predictable set of opinions, often those directly antithetical to those held by the Remainer. Does this strike you as odd? If it doesn’t, then you haven’t thought about it hard enough. The questions seem logically distinct. What relation does the UK’s membership with the EU have with the ontological status of foetuses? What relation does the ontological status of foetuses have with the very complex geopolitical calculation required to judge appropriate levels of defence spending? What do the vicissitudes of international relations have to do with the interdisciplinary sociological, biological and psychological considerations in the ‘nature vs nurture’ debate? Despite this appearance of independence between issues, you’ll almost never find someone who attends Remain marches one day, pickets an abortion clinic the next, attends a CND march on a third day and then speaks tops it all off with writing a dissertation on innate differences in brain structure between the sexes. What gives?

The temptation here is, of course, to simply say ‘My side is more well informed, and so while these questions seem to be logically distinct, there is one answer to all of them which will be picked by the informed’. This idea is actually not a completely untenable answer, but for a big flaw. If (for argument’s sake) the Right is the political grouping whose intellectual faculties are correct, then that explains why people on the Right trend to a certain set of answers to political issues, but it doesn’t account for why the Left seem to systematically choose the wrong answer. We would instead expect those who aren’t on the Right to have ‘randomly distributed and disparate beliefs about most of these topics’[2], and this would will be true if it turns out the Left were the grouping who were more intelligent. To make this clearer, let us talk about another field of knowledge, biology. There is obviously a group of people who are systematically well-informed about biology, namely professional biologists. People who are not professional biologists will have varying levels of correct belief about biology, but even the most uninitiated don’t systematically hold views that are directly contrary to those of the educated biologists. A biologist will tell you that the average human has five fingers, but even someone who has never received any education in biology would tell you the same thing. You won’t find anyone in the world that has systematically arrived at the opposite biological conclusions to professional biologists, yet this is what we seem to find in politics.

Of course, a response that I hear often is that political inclination is not to do with information and intelligence, but to do with morality. They don’t think their side is better-informed than their counterparts across the aisle, but that they are just better people than them. Well, maybe, but that doesn’t just posit the existence of one objectively correct answer to a wide range of moral quandaries, but also that one side of the political spectrum has a faculty by which to perceive such moral truths which the other lacks. Only problem is, reaching objective moral conclusions is proving to be rather tricky, take it from a philosophy student. If you could, for example, prove beyond all doubt that one side of the abortion debate is correct, then you have an endowed professorship at an Ivy League University and a Kyoto Prize waiting for you. Even if someone like that does exist, I’m damn well sure they wouldn’t bother reading my blog.

IV.

Early last year, Netflix put out a feature-length documentary, The Great Hack, which focuses on the Cambridge Analytica scandal. For those of you unfamiliar with the whole fracas, Cambridge Analytica’s advertising targeting model allegedly utilised the data of millions of Facebook users, originally harvested by academic Aleksandr Kogan, without the consent of said users. The documentary mostly follows David Carroll, a professor at The New School, in his attempt to take legal action against Facebook in a UK court for the breach of data privacy laws. As such, the documentary focuses on questions of business ethics and, well, data privacy law. While I don’t deny that these things are important, the possibility that Cambridge Analytica and Facebook criminally misused the data of millions of people is quite frankly the least interesting part of the story of CA, in my humble opinion. See, it is how and why the data was used that is the real corker, but this is a topic that the documentary spends barely ten minutes on.

I’m going to try to summarise this as succinctly as possible, but that will be quite the test. The fundamental insight which underpinned CA’s voter targeting model, and distinguished it from traditional methods of voter classification, was the recently investigated correlation between the Five Factor model of personality and voting inclinations. You see, the Five Factor personality model (also known as the Big Five model or OCEAN model) measures how people respond to a set of questions, and then use these answers to rank them on five metrics; Openness, Conscientiousness, Extroversion, Agreeableness and Neuroticism. I know a lot of people are sceptical of personality tests in general, thanks in no part to intensely dubious and unrigorous examples such as Myers-Briggs and Enneagram. The Five Factor model, while not without its critics, is worlds apart in terms of acceptance by serious psychologists and is by far the dominant model of personality used by actual personality scientists. Here is a good article explaining as much. In 2017, I came across a 2007 paper published in Political Psychology showing that certain results on a Five Factor test are strongly predictive of voting behaviour in the German electorate – namely, ‘Openness makes citizens more inclined to support parties endorsing social liberalism whereas low scores on Conscientiousness increase the likelihood of liking and voting for parties subscribing to economic or social liberalism as do high levels on Agreeableness. Finally, high levels of Neuroticism appear to render individuals more inclined to support parties that offer shelter against material or cultural challenges’. These headline results, that Openness, Neuroticism and Agreeableness are positively correlated with social liberalism and that Conscientiousness, and to some degree Extroversion, are negatively correlated with the same, are the most replicated results in research on the links between personality and political leanings. The phenomenon is by no means trivial, but the exact correlation between personality traits and political leanings is somewhat debated. I’ve seen some studies with an r as low as .24 and as high as .61, generally settling around .45. I would love to delve into the nitty gritty of this, but that would bloat this post beyond all reason. Possibly the most succinct statement on the state of play on this fascinating correlation is a quote from Anya Samek, Associate Professor of Economics at the University of Southern California – ‘Personality has a strong, statistically significant association with voting intentions, even when controlling for voter characteristics and past voting behaviour… These findings are in line with a nascent literature documenting association of personality with conservative or liberal views’[3]. I don’t want to overstate the case made here, I am not arguing that all politics is reducible to personality, which is why I suggest everyone go read the book and take in all the nuance (I realise I’ve only provided a link to a singular paper so far but this isn’t to pull wool over your eyes – I’ll provide a full complement of links to literature on the links between personality and politics at the bottom of this post). Political views aren’t set in stone from birth (though the received wisdom that people become more right-wing with age generally massively overestimates the degree), and I really, really, really cannot emphasise enough how much I do not want to be taken as a determinist in this issue. There really does just seem to be a lot of evidence that personality has a large role in determining political views, and t was only a matter of time, that the predictive power of such a model was put to use in real-world politics, and Cambridge Analytica were one of the companies to do it first.

Addendum: Many people have asked me what my own Five Factor results are, and what that would predict about my politics – I’ve taken a few different tests, and the average results are: 83rd percentile for Openness, 9th percentile for Conscientiousness, 83rd percentile for Extroversion, 58th percentile for Agreeableness and 12th percentile for Neuroticism. I am very high in openness and incredibly low in conscientiousness, which are positively correlated with liberalism, but I am also very low in neuroticism/high in emotional stability (Some tests use emotional stability instead of neuroticism – they measure the same thing, so high emotional stability is the same as low neuroticism), which is positively correlated with conservatism. I am also high in extroversion, which is positively correlated with conservatism in some studies, but other studies find it not to be correlated with either conservatism or liberalism. Make of all this what you will. I’m incredibly unfortunate to be low in conscientiousness, as it is, with the exception of a certain other metric that I shan’t mention, the strongest predictor of lifetime earnings and material success we have. On the plus side, I’m extremely lucky to be both high in Extraversion and very low in Neuroticism, which combine to produce high emotional well-being and life satisfaction. This is all to say, I’m going to be very jolly in the unemployment line sometime soon.

There’s a scene early on in The Social Network where Mark Zuckerberg is dragged in front of a Harvard disciplinary board to face retribution for various online misdeeds. In typical socially clueless, iconoclastic fashion, our intrepid and only nominally contrite protagonist demands recognition from the assembled Harvard notables for his invaluable contribution to identifying flaws in Harvard’s network security. When the head of IT responds to this by pointing out that the sophistication of Harvard’s network is what allowed them to track down the origin of the hack in under three hours, Mark brushes him off effortlessly – ‘That would be impressive, except if you’d known what you were looking for, you would have seen it written on my dorm room window’. This exchange came to mind recently, when I was doing the research for this article. I had become aware of the Schoen & Schumann paper in 2017, and so felt extraordinarily clever when I picked up on the psychological innovation of the CA model from the sliver of attention it was given in The Great Hack. This would be impressive, except if I’d known what I was looking for, I could have watched Alexander Nix, CA’s undeniably Bond villain-esque CEO, blurt it all out in an eleven-minute presentation at the 2016 Concordia Annual Summit. CA did not have direct access to a database of the Five Factor personality results of millions of Americans, because most Americans have never filled out a Five Factor personality test. The real genius of their model was a second level of correlation they could utilise. You see, Aleksandr Kogan’s research had paid a relatively small body of Facebook users (~300,000) to fill in a Five Factor personality test, and then correlated the results of such tests with the types of pages that each respondent had ‘liked’ on Facebook. This is what allowed Cambridge Analytica the power to predict the political leanings of millions of American Facebook users during the 2016 election cycle. They didn’t have the Five Factor results of these people, but they did have access to what pages they had liked, and so could use that data to predict what these people’s Five Factor results were, and then use that prediction, in combination with demographic and socio-economic factors, to further predict political leanings. The fact that the CA model incorporated ‘traditional’ methods of predicting voter behaviour, such as geography and socio-economics, alongside psychometrics is important, as it showed that at least some degree political partisanship was additionally predicted by psychometrics over the isolated use traditional divisors. This big data approach allowed Cambridge Analytica to sort people into three broad groups – Those who were already highly likely to vote for their candidate, those highly unlikely to vote for their candidate, and swing voters. Spending campaign money targeting these first two groups was pointless or inefficient, respectively, and so the CA model allowed their clients (the political campaign that had hired them) to spend money only on targeting the swing voters. Cambridge Analytica also claimed to be able to ‘microtarget’ messaging to different sets of swing voters, depending on what psychological research showed would best sway the inclinations of people with their Five Factor profile towards supporting a particular candidate.

Now, it’s not as if the efficacy of Cambridge Analytica’s model is unquestioned, and there have been a few pieces online casting some doubt over them. I don’t really feel that the points of contention affect what I’m trying to show with the whole example of CA (and do keep in mind, I’m only really using CA as an example to couch the underlying academic work done in psychology on this issue) but I know some people will take one look at my argument, see that there is some contention about the exact efficacy of CA’s model and will thus dismiss everything I’ve said. In order to avoid this, I shall have to talk about such criticisms.

The first point centres around what I shall call the ‘second-order correlation’; the use of Facebook likes to predict Five Factor traits, with many claiming that this method of building up a potential voter’s Five Factor profile is highly ineffective. This argument was actually advanced by Alexander Nix and Aleksandr Kogan themselves during the Parliamentary investigations into potential wrongdoing by CA, and was backed by Eitan Hersh, professor of political science at Tufts University, who in his written testimony to the U.S. Senate Judiciary Committee claimed that ‘building a psychological profile by connecting Facebook “likes” to survey respondents who took a personality test would lead to inaccurate predictions. Facebook “likes” might be correlated with traits like openness and neuroticism, but the correlation is likely to be weak.’ I would say this particular criticism of CA’s model is the most prominent. There is some pushback against the veracity of such criticism, with Christopher Wylie (one of the CA ‘whistleblowers’) arguing in his own written testimony to the same committee that Hersh was attempting to contradict ‘copious amounts of peer-reviewed literature in top scientific journals, including the Proceedings of the National Academy of Science, Psychological Science, and Journal of Personality and Individual Differences’. The fact that Alexander Nix and Aleksandr Kogan (The difference in spelling is infuriating, I can’t just call them the ‘two Alexanders’) have advanced this interpretation of the data cannot be taken at face-value, as it is now in their interest that CA’s activities be seen as politically inconsequential, lessening the chance that they will face ramifications for their questionably legal use of data. Nix even claims that Kogan’s dataset ‘proved to be useless’ and was deleted by CA back in 2015, before they undertook any serious political consultancy work in the US. This version of events is flatly contradicted by Business Insider, who report that when Facebook learnt of the use of the Kogan dataset by CA in 2015, they requested that CA delete the data, a request which was not honoured. Regardless, whether Wylie or Hersh is correct, it doesn’t matter for my overarching point. They are disputing the ‘second-order correlation’, questioning whether Facebook likes can be correlated with Five Factor traits, whereas my interest lies with the ‘first correlation’, the link between Five Factor traits and voting behaviour. This link has not been called into question, as far as I can tell.

The second big contention is the effectiveness of what Nix refers to as ‘microtargeting’. In short, CA claimed that they could sell policies more effectively by showing different ads to people with different personality profiles. This claim was wholly unfounded, with Vox’s coverage of this issue reporting that ‘There’s nearly no evidence these ads could change your voting preferences or behaviour’. This seems to be a pretty universal view, and I can’t find very many voices at all arguing that CA were able to back up Nix’s grandiose claims of being able use microtargeting to more effectively convince people of policy positions. Nonetheless, this doesn’t at all undermine my point. This is very conveniently summarised by the article put out by Nature on this issue – ‘being able to make simple guesses about their political partisanship would be “a big deal — even if [Cambridge Analytica’s] psychometrics were bunk”’. The fact that CA’s model could predict voter partisanship but could not influence their views actually reinforces my point about political inclination not being formed by political propaganda or education.

VI.

I hope the point I am making here has become very clear – politics to some significant degree is downstream of personality (You can now make sense of the title of this post). Knowing my readership, I’ll be accused of making a correlation-causation error if I don’t explain why personality causes political inclination and not the other way round, and so I’ll briefly address that. Firstly, Five Factor traits are stable in people from an early age up into adulthood, and young children are hardly political savants with radical views on planning permission law and the responsibilities of the nation state in the international sphere. Secondly, there is a large body of research which shows that Five Factor traits have a considerably degree of genetic heritability, with a recent metastudy arriving at an average heritability of .49. Generally, there is good evidence that insofar as personality plays a role in political views, it is politics that supervenes on personality.

All this aside, the fact that politics seems to be so closely linked to personality leads me back to the original point. I don’t like it when people assume that they hold a particular partisan line because they are so well-informed or so morally well-endowed, because this is wrong. To a significant degree, you hold the particular partisan line you do because your personality inclines you towards said line, or because of other intrinsic characteristics like age or ethnicity. This is why, of course, my friend didn’t need to actually read Thomas Sowell to know he disagreed with him – Sowell is a conservative, and despite whatever wonders of political persuasion he had at his disposal, they almost certainly wouldn’t have shaken my friend’s left-wing instincts. Of course, there are significant differences between levels of political education across any society, but to a very large extent political education is undertaken along the partisan lines – we seek information that confirms our pre-existing views, and this is exacerbated by increasing polarisation and ideological segmentation with echo-chambers. Here, I have to introduce an epistemic line of argument – there is a significant personality component to the formation of political opinions, and also a large demographic effect, and I don’t think it is controversial to say that one’s largely uncritically formed Five Factors personality or circumstances of birth are not a good way to make accurate judgements about incredibly complex moral and political issues, at least in any models of political epistemology I’m familiar with. These findings certainly conflict with the assumptions that voters are rational and form their preferences through examination and consideration of policy. Another big implication of our political views being so bound up with our personality and sense of self is that we interpret facts that are incongruous with our worldviews to be attacks on us, and hence we become incredibly resistant to conflicting information. Almost everyone who has strong beliefs on politics believes that they are right, and we can’t all be right. This is, at the end of it, an exhortation for epistemic humility – at least try be open to the possibility that you are not right, and in light of that, act in a way which is at least somewhat conducive of viewpoints that might be unfamiliar to your own. Even if you are the one person who has somehow divined the objectively correct view of all social and political issues, do keep in mind that other people have arrived at incorrect conclusions through no fault of their own, and this might serve to help us not hate one another quite so much – not because hating one another isn’t conducive to the whole democracy thing that I’m quite fond of, but because I really quite like people getting along, and I think it monstrously unfair to hold people to be evil or stupid on account of political views formed independently of morality or intelligence.

Links to resources, in no particular order:

https://www.annualreviews.org/doi/pdf/10.1146/annurev-polisci-051010-111659

http://lup.lub.lu.se/luur/download?func=downloadFile&recordOId=8992021&fileOId=8992024

https://onlinelibrary.wiley.com/doi/epdf/10.1111/j.1467-9221.2007.00582.x