Humans were destined to fight online

The epistemology we built the digital world with cannot survive the reality of human knowledge and discussion

The 'post-truth' turn in political and social debate has been heavily discussed over the past decade or so, and most analyses point a finger at the internet and social media. However, this does not mean that there is anything inherently post-truth about the technologies themselves.

On my reading, these technologies are actually built on and instantiate traditionally robust positivist assumptions about truth. The problem is that these are flawed and unstable assumptions that degenerate when used by humans at vast scale. It is the realities of human perspectives, knowledge-dynamics and psychologies, when supercharged by modern digital technologies, that have produced a post-truth world.

This distinction may seem like a pedantic detail, but I always think an accurate diagnosis is the starting point for solving a problem. At the very least, it strongly suggests common approaches to resolving polarisation are destined to fail, because they are tackling the wrong problems.

The internet is perfectly positivist

I recently explained a little about Information Theory, the philosophical framework that lead to modern computing, and therefore the online world. On this way of thinking, all information - treated as series of discrete messages - helps us to identify knowledge. Each message stands alone and, by itself, increases the probability of certain states being true in the world. While attractive, this way of thinking runs into challenges when we are dealing with human-created messages that are intended to entertain, deceive or ponder, rather than inform.

Nevertheless, it has provided the mental framework by which information is coded and presented online. Everything online, whether it is a tweet, a Reddit post, a long form article or a short video, has the form of a discrete, disembodied, stand alone message. As such, any information online appears like it is a universal statement that is always and anywhere applicable. While this follows from the assumptions of Information Theory, it also fits neatly with logical positivism.

This approach to philosophy grew most strongly out of discoveries about logic and mathematics in the nineteenth century. One of the core assumptions is that any properly formed statement in a language will be either true or false, simply based on the facts of reality. It is a view of language exemplified by what is known as the T-Schema and historically taken to be the basic definition of the concept of truth. On this schema, a sentence P is true if, and only if, P. For example, the statement 'snow is white' is true if, and only if, snow actually is white.

The default form of presentation of any information online looks and feels as though it is a statement in this positivist sense. It looks like a general statement that is just true or false by itself. In other words, digital communication is structured in a positivist style - i.e. everything is universally true or false. The problem is that normal human communication doesn't work like this.

Human communication occurs deep inside a context

I have written previously about limitations with this positivist approach, notably with logical paradoxes. My recent articles have added a further challenge for this way of thinking. If all human knowledge is about some particular aspect of the world, then there are particular facts that matter, rather than a general notion of reality, in terms of deciding if something is true. A statement will often not be true completely in general, but only relative to a context.

My previous article noted that this holds in areas of quantum physics, but it is highly familiar in day-to-day life. We regularly misunderstand something someone has said or written because we came in part way through a conversation and didn't have the right context. Humans routinely communicate in short hand, with in-group references, in a particular time, place and context, as this enables efficient and precise communication with the people around us.

To pick one of many examples, we can consider a universal sounding statement like "Democracy is the most effective form of government". It looks like a universal statement, but our decision as to whether it is true or false will depend on what form of democracy we have in mind (e.g. Ancient Athens, modern liberal democracy, direct democracies where everything goes to a vote, etc); what we consider to be most important in government and even what time period we are in (for example, modern inventions like postal logistics and secret ballots change what we think is possible).

Instead, the nuances of what relevant words mean, in combination with what the sentence is truly about, often mean that whether a statement is true or not can genuinely depend. Unfortunately, all this nuance of ordinary context largely disappears once something is online.

The internet collides with the real world

The clash between how information online presents - as context-free, universal claims - with how human communication really works continues to reverberate throughout our societies.

It triggers and drives some of the bitterness and polarisation that pervades most online debate today. When statements appear to be context-free, people invariably read their own, assumed context into them. We often experience this when we join a conversation half way through. It also happens online, where it means that people are reading their own assumptions in to a vast amount of discussion and therefore regularly misunderstand what is being said. However, as the information online appears to be a universal statement we think we should understand, it is very common to get horrified or offended by what we think is being said and things easily escalate. It takes particular and rare self-discipline and cognitive empathy to seek to understand what the other is really trying to say given what their context is.

More generally, we have a clash between the way truth is presented and how it works. Online information appears to be a series of universal, context-free statements and we often take what we read in that vein. However, context, place and what the statement is about is critically important to what anyone actually says or writes - none of which is natively embodied in most digital communication. Once we couple this with the way anyone can publish online and gain huge reach, some kind of post-truth and polarised world has inevitably followed.

A different approach to supporting truth online

If this analysis is correct, it provides another reason why existing approaches to reinforcing truth online, focused on combating misinformation and fact-checking, will only have a limited benefit. The core problem is the clash between how information presents itself online and how we humans build and interact with knowledge. Improving the quality and reliability of information online doesn't address this disconnect. Instead we need to think differently about the way information is presented and communicated online and also tackle the human side of how we react. I do not have any good solutions, so can only offer some directional suggestions.

One ideal for the presentation of information online would be that it better reflects the context, purpose and assumptions the information was developed and presented within. While this occurs more naturally within, for example, longer form writing like I produce here, it is unnatural to much of what is produced online at the moment. Nevertheless, there are some, perhaps inadvertent, moves in this direction.

For example, X/Twitter has expanded its Community Notes function that allows users to add information to posts. Notably, given my argument here, the headline used to introduce a note is that “Readers added context they thought people might want to know”. It is at least framed to address the gap I have identified. This is one idiosyncratic option and there are likely many others. People need to try and see if anything helps reintroduce a sense of time, place and context to information online.

On the human side of how we react, the best option I can think of is to strongly encourage a praxis of charitable reading. Before coming to a conclusion or responding vigorously, stop to think: what are they trying to say and where are they coming from? That gives us the best chance to really understand what they are saying and talking about. Then we will be be in a much better position to decide whether we think it is true or not.

This type of behavioural norm would fit neatly alongside a shift to decorum, rather than content, moderation online - but neither naturally fits with how people are currently acting. Obviously, this is one of those solutions that is most attractive to those people who don't need it, but encouraging a norm of charitable reading in online communities may start to make a difference, at least at the margins.