The internet might destroy democracy
Mysterious Twitter user “The Matthew Principle” (I am going to simply call him Matthew from now on) has an interesting tweet-essay in which he argues against Fukuyama’s central thesis that liberal democracy is the ultimate form of government. Unlike other critics of Fukuyama, he posits that the real threat to liberal democracy comes not from its external rivals, like Communist China or Putin's Russia. Rather, he thinks it’s going to result from internal issues, particularly the rise of social media, which empowers emotionally volatile and less-informed segments of the electorate, undermining the control traditionally held by the upper classes. In a way, Democracy is suffering from its own success.
He argues that what we usually conceive of as liberal democracy has not been an actual democracy, with the elites (upper middle and upper classes) holding much more power than usually recognised, via their control of the press and the informational ecosystem more broadly.
“The upper and upper-middle classes controlled the public discourse through their control of the media and thereby stabilised the system. In such an environment, it was difficult, bordering on impossible, for “grassroots” movements to obtain more than minor influence, except where they had the support of the elite establishment.”
(excerpt from Matthew’s tweet)
But, according to him, the rise of social media poses a fatal threat to this equilibrium; This epistemic control can no longer be sustained due to the current decentralisation of our intellectual ecosystem. Free speech in the current era is not the free speech of the 20th century. As a result, we are doomed to devolve into permanent populism in the near future, with Trump being just a harbinger of this new era. Since he thinks this type of populism is not going to lead to a stable political situation, he deduces from simple evolutionary principles that, in the long term, democracy will be replaced by a new political system. He bets on a more limited form of democracy, with extensive social media censorship and the right to vote being restricted to only some segments of the population.
The viewpoint articulated by Matthew, positing the internet as a potentially harmful force for democratic societies, resonates across a broader spectrum of mainstream commentators. One can sense a growing unease among societal elites, a kind of apprehension akin to feeling unstable ground beneath them. Take for example this tweet by Noah Smith, highlighting the same concern.
This apprehension seems to be driving a call for more stringent regulation of our information networks. I have previously written about the rise of misinformation studies, the academic field often used to justify the need for increased control.
Should elites control the information ecosystem?
I agree with a lot of what Matthew and others are concerned about. The core of my agreement with Matthew can be summarised as:
I am not a naive believer in the “marketplace of ideas”. As I said in my previous post, I think in the end it is healthy and normal that epistemic authorities emerge. The opinion of a Biologist on the safety of the COVID-19 vaccine should natter much more than the opinion of a random member of the public. It’s good that people take certain things these elite authority figures say for granted. So in that sense, it is true that elites have much more political power than indicated strictly by the number of votes allocated to them. And I think that is good. I agree that social media *is* a threat to democracy and to our current political stability by undermining the trust in intellectual authorities and decreasing the centralised power of our elites.
So given this, a natural question emerges: Isn’t it good then that elites are trying to put brakes on this chaos through the whole misinformation studies endeavour? If you are familiar with my writing, you know I believe this is a bad idea. I have spent the last 3 posts (see this and this and this) warning against the dangers of top down control of our informational ecosystem by bureaucrats and argued that the field of misinformation studies is in large part enabling this. The other thing you might know about me is that I like both democracy and political stability. So how do all these ideas fit with each other? Shouldn’t I support misinformation people based on utilitarian grounds? Yes, the intellectual foundations of their field are shaky, but who cares about it when our system of governance is at stake?
Well, my first point will be that even if they try, elites won’t be able to exert the control they so crave. That era is gone. We need to adapt to a new one. Misinformation warnings might serve the careers of a few commentators and academics, but they will NOT meaningfully put a stop to the spread of false narratives. If you do not believe me, I’d like to remind you that, despite all the screaming and shouting from elites Trump, a bona fide populist, is more likely than not to become the next President of the United States. So in a sense the only real chance for intellectual elites is to regain trust of the population, as I argued in a previous essay. If that happens, I think it would keep us on the path of a relatively liberal democracy.
But is this even possible in theory? Matthew’s essay suggests that the increased scrutiny posed by the internet is intrinsically incompatible with liberal democracy. I think he believes that elites, the flawed humans that they are, simply cannot withstand the increased scrutiny without emerging as much less respectable than previously thought. And that the general populace will naturally converge upon conspiratorial, populist narratives. I do not fully agree with him.
Firstly, he overstates his case by underestimating the veto power that the general populace has always had via voting. Yes, elites could manipulate narratives. While elites can indeed influence narratives, their power is not absolute – ultimately, votes do matter. This dynamic is akin to a river shaped by its banks. Elites may set the boundaries for discourse and ideas, much like riverbanks guide the river's flow, but they cannot stop the river. Increased scrutiny of elites might put an additional evolutionary pressure on elites, but it does not necessarily follow that it would lead to a complete breakdown of our governance system.
Secondly, there's an inherent human tendency to seek intellectual leadership and authority. In the era of the internet, despite the challenges it brings, it's conceivable for a principled elite to emerge and maintain influence.
A tale of two vaccination rates
To support my somewhat optimistic perspective that the internet does not necessarily lead to spiralling into stupid populism and conspiracy, let's consider some empirical evidence. Take, for instance, the anti-vaccination narrative, which has been a prominent example of populist anti-elite messaging in recent years. However, the response to this narrative, specifically in terms of vaccination rates, differs notably between the United States and the United Kingdom. As of May 2023, 70% of the U.S. population had received two doses of the COVID vaccine, compared to a considerably higher 88.8% in the UK (by August 2022). This 70% figure in the U.S. also conceals variations across different states. It's important to note that these populations, both in the U.S. and UK, primarily speak English and have similar access to social media platforms. This suggests that factors such as the degree of trust in elites and cultural influences play a substantial role in shaping public response, even when internet access is a constant.
For example, we know that Democrats tend to have a higher level of trust in intellectual elites, including academia, and this is reflected in the higher vaccination rates observed in Democrat-leaning states. This evidence indicates that while the internet provides a uniform platform for information and misinformation alike, the impact of such narratives is significantly moderated by the trust in elites and cultural context. As I argued before, the best thing intellectual elites can do to defend liberal democracy is act in such a way that increases the trust of the population.
To close this, I’d like to add another argument against centralised control of information: the fact that it is now used to control intra-elite discourse. In an ideal world, different elite factions would battle with each other: this would give the general public more choice between what kind of propaganda they want to believe. This might suggest that the most crucial form of free speech is actually within the elite circles themselves. Such dynamic inter-elite discourse ensures a check-and-balance system, preventing any single faction from dominating and potentially spiralling into extreme or unchallenged viewpoints. Like a cancer taking over one’s body, insanity will inevitably take over any ideology that is universally favoured by a monolithic elite.
I have noted before that the end goal of many misinformation experts seems to be weakening of elite free speech specifically (in order to ensure the victory of their preferred side). I believe so-called misinformation can be classified into 2 types: brute misinformation (e.g. the kind of Infowars drivels) and haut bourgeois propaganda (stuff you’d find in the New York Times). While brute misinformation trades in explicitly crazy ideas, haut bourgeois propaganda is usually more subtle and relies on presenting partial truths or spinning narratives in a way that is favourable to one side or the other. As I have noted before, the anxiety around brute misinformation and populist narratives is often used to justify the encroachment of freedom of speech/thought in an arbitrary way in more subtle situations (e.g. and thus enforcing haute bourgeois propaganda).
Good essay. It really made me think about the problem of expert credibility, social media and nuance. Have you ever heard of Solution Aversion? It's a neat little bit of research which shows that the only difference which smart people demonstrate when asked to evaluate empirical evidence which contradicts a previously held belief is that they are able to come up with more means of discrediting said piece of evidence. I think this a highly informative piece of research, because it not only shows how human distributed networks tend to polarise, but also holds the key to understanding another phenomenon- the tendency of people to commit to an evidence gathering exercise which makes them believe that a problem they've been studying is far more serious than it actually is.
A good example of this would be climate change. Anybody who has read IPCC reports and summaries directly for more than a decade will know that although climate change is a very serious long-term problem, there is no feasible scenario which makes climate change an existential threat or a civilisation destroyer other than Black Swan events. I don't want to get into too much detail, but it's relatively easy to research online the assumptions that went into RCP/SSP 8.5 and see they bear no relation to a business as usual scenario an in all probably the current worst case scenario for 2100 is the Rocky Road scenario. Anyway, I was more interested in climate change as a means of explaining how governmental and institutional forces can find themselves exaggerating a problem as a means of attempting the galvanise a shift in public policy, only to find themselves in danger of losing their institutional credibility. The aims may be noble but the results are predictably catastrophic.
Here's why. The ability to detect dissembling, managed narratives or even information massaging and utilising fear as a means of inducing behavioural change simply doesn't correlate with general intelligence. If anything, the highly cognitive and highly educated have a blind spot in this regard because class affinity means that they far more likely to believe that 'people like them' are going to be more likely to be operating from 'pure' motives. The same simply isn't true of most of the population in Western countries- although 50% of the American population might possess 'some college', it's highly likely that only around 20% fit the educational background necessary to confer an automatic belief in the credibility of experts. Whose still, those who do tend to occupy elite governmental and institutional roles are highly susceptible to groupthink, because they tend to drastically overestimate support for their policies and ideas within the population. A good example of this would be the progressive component of pre-Musk Twitter. Progressives accounted for 50% of all tweets, despite being only 8% of the population.
So how did the UK differ in its handling of vaccines and Covid restrictions? Well, although the legacy media supported an elite narrative, the British government adopted a 'warts and all' approach to furnishing information. Government websites quickly admitted that their had been vaccines deaths including mRNA deaths, but that they were exceptionally rare. Even the most fringe conspiracy sites stuck to a limit of 1,000 vaccine deaths from AstraZeneca and 1,000 vaccine deaths from mRNA- in a population where roughly 90% of the over 18 category received the initial double jab.
It's also worth noting that although vaccine uptakes were high in the initial round of vaccinations, booster shots had a much lower rate of uptake. This is because although the British government acquitted itself pretty well on access to unvarnished and unmassaged data, it also quickly got itself into the habit of producing drastically overegged worst case scenarios for Covid case projections trying to induce people to change behaviour. What made matters worse was that although the Tories never actually went ahead on introducing vaccine status based restrictions (over than on flights), they continuously raised the spectre- and large segments of the British public, coded along lines of class rather than political affiliation quickly developed a reluctance to take further vaccines.
My final point is nuance. For anyone who has studied the data it quickly becomes apparent that the only time that lockdowns actually worked was right at the very start of the pandemic. This isn't because lockdowns actually stopped the virus (sea barriers being the exception). It's because it limited the initial wave of infections to relatively young and healthy real economy workers dealing with the public. In effect, it provided a limited from of herd immunity by infecting individuals who would have otherwise acted as super spreaders. But how many people would be willing to concede both that locking down two weeks earlier in most countries would have saved lives, whilst also agreeing that after this initial wave lockdowns became relatively ineffective- or at least only effective insofar as they convinced particularly vulnerable elderly individuals to cut off all human contact, other than the occasional open air contact with a delivery driver or a family member?
This can teach us a lot about how to maintain institutional credibility. First, furnish a 'warts and all' approach to information management. Second, no matter how noble the aim, don't try to massage the information to produce a morally beneficial result. Trust is more important, and people can see right through experts trying to persuade. Third, don't use fear- as we've seen with recent populist rejections of green policies, persuasion is far more likely to obtain far better results than fear aimed at producing behavioural change.
And if you doubt this last point- exactly how many additional days of annual leave did you get last year for using socially responsible green methods of getting to and from work? if it was less than three- they really haven't even begun to try to persuade people...
I think it's worth distinguishing a few different issues:
1) Control by elites vs control by elites whose values systemically differ from those of non-elites.
2) The effectiveness of explicit attempts to label and limit 'misinformation' vs the effectiveness of a system where people's information is organically controlled by elites because they can find elite sources they trust to appropriately inform them.
I'd argue that the thing we've lost is the ability to assume that the full Overton window of the populace is sufficently represented in expert opinion that you can trust relatively mainstream sources that consult these elites to adequately represent the best arguments for all sides.
Without this the problem is that many non-experts realize they can't always trust experts to give counterarguments/context if it existed so even when those experts are telling the simple truth they can't rely on them and thus they can't draw rely on experts to recommend media/information.