Elon's Twitter: a philosophical experiment begins

Watching for the impacts of different epistemic attitudes

When Elon Musk’s planned purchase of Twitter became news earlier in the year, I published a couple of articles that looked at the conceptual arguments for free speech, which was Elon Musk’s stated reason for the takeover, and the role of different epistemic attitudes in how people approach the necessary task of moderating social interactions online.

At the time, I wrote that:

Elon Musk owning Twitter will lead to an interesting natural experiment between different social media platforms. He will run Twitter based on a different set of fundamental philosophical and epistemic attitudes to Facebook or Instagram.

Now that he has completed the takeover of Twitter and begun remodeling the company, we can start watching this natural philosophical experiment and see if there are any interesting results.

So what is the experiment?

Unlike much of the current commentary on Twitter, our interest is not primarily on profitability or regulatory dynamics. Instead the focus should be on the lived experience and style of communication that may, or may not, arise from different epistemic attitudes.

Broadly speaking, the major social media platforms like Facebook, Instagram and (previously) Twitter, have adopted an attitude of epistemic confidence. They assume that knowledge, at least on important topics, is readily known and therefore the moderation role is largely to remove false or misleading claims. Hence there is a very strong focus on fact-checking and misinformation.

As argued previously, the major justifications for freedom of speech rely on a different philosophical attitude: epistemic humility. Knowledge is hard and we often get things wrong, so therefore we should be listening to a wide range of views to test our thinking. Starting from this attitude, it makes more sense to focus on behaviour or decorum moderation - ensuring behaviour is civil and reasonably respectful - rather than moderating content.

It is currently not yet clear whether Elon’s Twitter will follow this logic through - and this natural experiment will fizzle if it doesn’t. However, he has repeated his commitment to freedom of speech as the rationale for acquiring Twitter:

The reason I acquired Twitter is because it is important to the future of civilization to have a common digital town square, where a wide range of beliefs can be debated in a healthy manner, without resorting to violence.

This suggests that the moderation of Twitter, which in his words is necessary to avoid “a free-for-all hellscape”, will be based to some extent on an attitude of epistemic humility. This will set up an interesting, albeit not decisive, natural social media experiment: what effects will different fundamental philosophical attitudes adopted by different platforms have on communities and individual experiences? Is one preferable to another from users’ perspectives?

What might we learn?

One core justification of freedom of speech, strongly associated with JS Mill, is that it helps us better find and test truth. We have also argued that epistemic humility helps us do the same, which is why the scientific method assumes epistemic humility. On the other hand, a common argument for strong content moderation is that it helps prevent social media descending primarily into a source of falsehoods and conspiracy theories that are completely disconnected from reality.

This natural experiment should therefore provide some insights as to which philosophical attitude and moderation approach encourages greater knowledge and more effectively furthers the discovery of truth.

This is obviously a deeply philosophical question that is not going to be readily measurable. Experiences are also likely to differ across users. However, there are some areas where the people involved have interests that more closely align with this goal and who will therefore be potentially useful barometers.

One good example is to see how effectively different social media platforms facilitate discoveries and collaboration around science, research and policy development. This more commonly involves people who are personally motivated to seek new knowledge and defend positions, so differences in how moderation approaches support or hinder these goals are likely to be more noticeable.

The differences between platforms may also provide some insights into the consequences of different philosophical attitudes on social dynamics. Will different moderation approaches encourage or discourage tribalism and social fractures? Or will there be no differences at all?

There is a strong school of thought that many of the troubling social dynamics are inherent to the nature of social media and are therefore unavoidable. For example, Jon Askonas argues that Twitter is really a type of computer game called a “multiplayer online battle arena” (or MOBA) as the core dynamic is that “competing networks of accounts (informally led by influencers and blue checks) do battle.” On this view, fractured tribalism is unavoidable on social media.

The two different moderation approaches we have mentioned can both plausibly make the dynamics more toxic. Letting anyone say what they want without removing clearly false information may encourage wilder claims and cause different communities to drift further apart. On the other hand, regular and obvious content moderation can encourage people to appeal to the authorities to settle disputes and that means they may be less likely to engage in good faith discussion.

It is not clear which of these, if either, is a stronger driver of human behaviour over large groups. If we genuinely get a divergence of approaches between Twitter and other platforms, this natural experiment will shed some light on this question and therefore the effects of different epistemic attitudes.

Will there be clear results?

Given most individual social science experiments return equivocal results, it is unlikely that this natural experiment will decisively answer any of our questions. There is a high likelihood that considerations about income generation, coding and engineering and costs will mean that Elon’s Twitter does not operate purely on an assumption of epistemic humility. Thus there won’t be as clean a difference between different platforms as we would ideally like for good experimental design.

Nevertheless, it is worth watching to see how different moderation approaches foster different online dynamics - or whether they make very little difference.

My hypotheses, or predictions

A well conducted scientific experiment sets out hypotheses that predict what we think will happen so they can be tested. Although this is not a well conducted experiment, it is still worth setting some out and revisiting them later. In no particular order, my hypotheses are:

1. Twitter will remain a MOBA, but only because the dynamics that drive this have always existed in human social behaviour. Politics, tribal warfare and schoolyard gangs are only the more visible forms of this behaviour. Moving online doesn’t change any of this.

2. A new moderation approach at Twitter will shift the social power dynamics. Content moderation tends to privilege those who already have a powerful voice or institutional clout. Decorum moderation is likely to mean a wider range of people gain social power on Twitter, but this power is likely to be less permanent.

3. If we think about the range of positions on any particular issue, my hypothesis is that an approach based on epistemic humility at Twitter will mean that the extremes isolate into their own bubbles more completely, but the those in the broad middle will moderate as they are exposed to a wider range of good faith views.

4. Assuming something like this holds, then Twitter will be more effective at uncovering truth than other platforms for those motivated to look for it and test their thinking. However, as has always been the case, this may be uncomfortable as the truth is often not what we want it to be.

We’ll have to wait and see whether the experiment confirms or disproves these hypotheses. In the meantime, reactions to whether they are plausible or what you think will happen are welcome.