When is Common Knowledge Necessary?

Common knowledge is necessary whenever we want to act on knowledge, as a group.

That's it. Meditate on this for a bit.

Without common knowledge, we are back to the situation mentioned in the page about social knowledge. To make plans involving others, we must think through what each sub-group of people knows and skills they have, and what they know about others, and so on.

Common knowledge collapses all the complexity of social knowledge into a single layer: everything that is common knowledge to a group is fair game for plans. Everyone is expected to know it. We can always rely on common knowledge in our plans, without having considered whether a specific individual knows it or not.

The plans may be explicit and relatively complex, like organising a party or solving a problem at work. But they can also be more implicit and immediate. For instance, in a conversation, before using some advanced knowledge, we quickly think about whether it is common knowledge. If it is not, we either pay the cost of explaining it, or we just don't use it.

Thus, common knowledge is necessary not only for big things like big punctual events, but also just for fluidifying everyday conversations.

Improving Group Decisions: Public Schooling

Most decisions are made based on the status quo. This is so because deviating from the status quo costs resources, like time, attention and effort. By virtue of living in a physical world, we can not afford to pay that cost on everything.

While art likes to shit on the status quo, the status quo merely represents what is currently common knowledge. It can be extended, improved and built upon.
This is what I believe to be the cheapest way to improve a group's decision making ability: not to systematically fight what is common knowledge, but to instead extend and improve it.

Unfortunately, very few groups are purposefully trying to improve their common knowledge.

The most ambitious attempt I know of was public schooling, which de-facto ensured that everyone in a society built common knowledge. In conversations though, people do not really see it as the main benefit of public schooling. They are more focused on teaching kids skills for jobs, or make their own children as smart as possible.

From my point of view, in a world without jobs, public schooling still has its place. It is the most natural way to ensure that everyone in a society builds common knowledge: we ensure that everyone, at the beginning of their lives, is taught for years a specific corpus of knowledge.

To some extent, this is what makes us a society in the first place: we have a body of common knowledge that we all come to rely upon.

Group Norms and Shibboleths

Another example for where common knowledge is critical is group norms.

There are many norms that are beneficial to people when everyone follows them. However, no one wants to be a sucker and be the one to follow a prosocial norm in a group that does not follow it.

For instance:

• Preparing meetings: coming to a meeting where everyone is prepared is great and gains a lot of time. But coming to a meeting where we are the only one prepared is a mood killer.
• Talking over people: in a group where there is a norm of talking over people, it is stupid to not do it. It won't improve the quality of the discussion, and will just lead to the quiet person not being heard.
• Cleaning communal spaces: no one wants to always clean after others.

Thus, a common situation that can happen is one where everyone in a group would agree that a rule is a good one to follow, but no one is following it for fear of being a sucker.
This is a good illustration of the difference between shared knowledge and common knowledge. While there might be shared knowledge that the rule is good, until there is common knowledge around that, it will be costly for people to follow it.

This also helps understanding the need for shibboleths. A shibboleth is a word or phrase that is used to identify members of a group. It is a way to quickly ensure that everyone knows a specific piece of common knowledge. If I am from a group that has many prosocial norms, it is useful to have a quick way to check that someone else is from the same group. In which case, we just trade shibboleths, and can follow our norms without fear of being a sucker.

Breaking Taboos

There are many behaviours that we want to prevent on principles, without having to consider them on a case by case basis. This is the main rationale behind laws.

Taboos go beyond this. They are not just rules about what must or must not be done. They are more like trauma: merely thinking about breaking them or discussing them is enough to trigger a strong negative reaction.

Taboos can in theory be useful. Some ideas might be so seductive and damaging that it is worth crippling our collective intelligence to prevent them from being considered.
In practice, I don't think we are collectively smart enough to know when we should use them. They are all too often used for petty political reasons, as an intimidation tool.

Specific laws may or may not be worth it, we may want to pass them and commit to them for 20 years, or we may want to repel them. Both can happen. But banning the fact of discussing them is extremely costly.

This is why there is often a need to break taboos. And thinking about it in terms of common knowledge helps a lot. Scott Aaronson and Alexander both wrote about it, so I won't repeat the point here.

My main stance on this topic is that we do not have reliable ways to build common knowledge in groups. As a result, we can easily get fucked by taboos. We become unable to discuss topics of importance, thus sticking to the status quo, even when it is detrimental.

Getting fucked by taboos most commonly happens in politics. People are fucking stupid when it comes to discussions with any political stakes. Support for any social program is qualified of communism, while support for any punishment program is branded with fascism.
The peak extreme of this is how discussions of demographics, phenomena that are objectively true and measurable, are now coded as racist.

This is partly why I believe that mechanisms to build common knowledge are so important. Else, we are just condemned to repeat the same mistakes over and over again.