A floundering natural philosophical experiment

Twitter, under Elon Musk, promised a different type of social media with different philosophical assumptions. This is proving difficult in practice.

Towards the end of last year, I argued that Elon Musk taking over Twitter would provide us with something of a natural philosophical experiment. His stated philosophical assumptions (as a 'free speech absolutist') differ markedly from those in charge of other similar platforms. Therefore, seeing how the different platforms evolve could give us insight into the real world consequences of different philosophical positions. There has been enough time to see some results and, as I anticipated, the experiment has been compromised/confounded in various ways. What we are seeing is much less of a clean natural experiment than we might have hoped.

Things are always complicated

The first complication is that, while Musk has strongly professed a commitment to freedom of speech, he has been much less open to it in practice. One of the clearest examples of this is what has happened with the independent journalists he brought in to look through the records of the previous management to publish the 'Twitter Files'. Since December, when this started, Musk has managed to fall out with these journalists for saying things he doesn't agree with. This weak practical commitment to freedom of speech, and the principles underlying it, are undermining any clean philosophical experiment.

Incidentally, the Twitter Files have revealed how widespread the view I described in the post Misinformed about Misinformation is. There is a large group of people in influential positions who believe that certainty about truth has been achieved and anything that contradicts this deserves to be suppressed.

A second complication is that many media and other commentators fundamentally disagree with the direction Musk has stated he is taking Twitter in, and so it is very difficult to get anything like a neutral perspective on what is going on. In many cases, commentators struggle to even understand how someone could think differently and so tend to reject any changes almost on principle. A recent video excerpt from an interview of Musk by a BBC journalist captures the dynamic well. It has been widely shared, but do watch if you are interested. In it, the journalist challenges Musk on the basis that hate speech incidents have gone up since Musk took over. However, Musk challenges him to provide an example, and from then on the interview becomes disjointed. There is clearly a gulf of worldview and comprehension.

While these two complications are not that surprising, I have been noticing a couple of deeper, more philosophical complications.

The medium is the message….

In the 1960s, cultural theorist Marshall McLuhan became famous for the phrase "the medium is the message". Part of his core insight is that the medium you use, whether that be television, handwriting a letter, social media, an instruction manual, a novel or a poem, communicates something separately from anything actually written or said. It can also can significantly shape what can or can't be said persuasively. For example, using the sort of powerpoint deck beloved by consultants to communicate your care and feelings for someone would undermine your message. Or writing an airline safety manual in sonnets would be impressive but miss the point.

This raises the question of how much the particular form of social media that Twitter has adopted - with short messages, focused on reactions to others, inbuilt tools to spread tweets virally (e.g. retweets) and instant feedback via voting from other users - shapes what can and can't be said with any effect on the platform. The structure of the medium may in turn shape moderation policies (or at least how they work) and the epistemic usefulness of Twitter discussions.

Jon Askonas, for example, has described Twitter as a type of computer game called a multiplayer online battle arena (MOBA), which creates rules of engagement and incentivises different behaviours, regardless of what anyone says. In other words, the nature of the platform itself creates antagonistic, tribal behaviour and the influence of different moderation policies is likely to be minimal.

If this is correct, then attempts to moderate or run the platform differently will only have an effect if they lead to significant changes to the design of, structure and nature of interaction on Twitter. In other words, Elon Musk might have to turn Twitter into something different to make it the ‘online town square’ that he described as his ambition.

Notably, there are signs that he is transforming the medium of Twitter, deliberately or not. Tweets are no longer limited to 280 characters, he has foregrounded Community Notes and has recently released the source code for Twitter’s Recommendation Algorithm. These may, or may not, be sufficient to change the platform enough to change the nature of the medium and therefore communicate different messages (aka McLuhan). But this raises the question that if Musk significantly changes the medium, would this lead to the collapse of Twitter as a popular platform?

Societal versus individual priorities

A second complication is that there is significant distance between system or societal ambitions for discussion and individual motivations. Arguments for the freedom of speech, and JS Mill is a classic example, assume that a primary goal of conversation and debate is to reveal, or get closer to, what is true. While this seems natural to philosophers, it is clearly not the primary motivation for most people much of the time. We discuss and debate to entertain, to win others over, to prove others wrong and ourselves right, to make jokes or friends, to gain social status, and so on. All of these natural human tendencies may depend on the idea that some things are true and others aren't, but truth is, at best, a secondary consideration.

The challenge is that at societal and group levels, we often want or need truth to emerge from conversations and debates.¹ However, people's motivations for engaging in debate are often not focused on truth. An important question then is the extent to which dominant media, and the norms around them, make the discovery of truth more or less likely - even if the individual motivations of those taking part in debates aren't to find truth.

Adversarial court systems are a good example of an environment that is set up to make the discovery of truth more likely, even though the motivations of key people are elsewhere. In these court systems, the prosecution and the defence are expected to try and win their case by whatever means are allowed within the rules. But the rules, including cross-examination, presentation of evidence, and the role of the judge, are designed to make it more likely that the truth emerges by virtue of each sides' attempts to win.

Looking at Twitter in this way leads me to be pessimistic about how useful it will ever be for uncovering truth. The types of interactions encouraged - following others, likes, retweets - prioritise and reward a wide range of human interactions over truth. People are more likely to like or retweet something that is funny, cute, cutting, or confirms what they think, rather than something that is true. So it seems unlikely that the medium of Twitter is well designed to make, at the group or societal level, the discovery of truth more likely. This works against the success of any natural philosophical experiments we might want to see, as well as a number of Musk's plans for the platform. Unless Twitter becomes something different, we are unlikely to see anything much different on the platform.

Another natural experiment?

You may have heard that Substack has released a new platform that works similarly to Twitter - called Notes. It is more explicitly committed to epistemic humility and freedom of speech and so may be, over time, a better candidate for the natural philosophical experiment. It may also reveal whether the concerns raised here, that the interface in the medium itself shapes whether it can succeed as a place for respectful conversations.

Currently, the design of Notes is very similar to Twitter and so it may end up recreating the same behaviours. Alternatively, there are some ideas floating round about how to change the interface and therefore change the medium. To promote my own contribution:

In my post last year I identified four hypotheses to later test. I will come back to them at a later point, and consider them both with respect to Twitter and Substack Notes.

1. To be more accurate, it is in the interests of societies over the long run for truth to emerge. Otherwise reality will hit us in the face at some point and we'll get in trouble. But sometimes societies prefer not to accept truth.