No, not like that

There’s a certain froofiness that comes with the phrase “we’re all connected”.

It just seems like one of those wispy, granola sayings that is completely detached from science.

Maybe it isn’t?

Connections

Let’s do a simple illustration on why everything very well may be connected - as long as we’re on the same page about what “connected” means.

Direct

We’re going to set up four nodes in a network. Each node pair has a potential edge, or direct connection, between them.

Simple Network with Known Connectivity

In the above, we see that Node A is directly connected to Node C. But Node A is not directly connected to Node B.

Simple enough - we can say that Node A is not directly connected to Node B. Therefore, the statement “all the nodes are connected” would be false if we assumed connected refers to direct connection only.

But…

Indirect

In the same network above, we can start admitting indirect connections.

If we admit indirect connections, then A connects to B via C. This may seem like it’s not useful, except that in any dynamical system, the influence of A on C will spread to B, because B is influenced by C and C is influenced by A.

So, in a certain (but different) sense, A is indeed connected to B.

Why it matters

Science is all about studying real-world systems, typically by breaking them apart 1.

One of the core ideas of experimental science is isolation - breaking connections between variables inside your scope and variables outside your scope.

Ask yourself: if we designed a rigorous experiment to isolate variables here, would you expect that rigorous experiment to yield data showing the effect of A on B is statistically significant?

Summary

Yes, we are all connected in the sense that our actions can indirectly affect everyone/everything else. No, we are not all connected in the sense that we are directly connected.

Rigorous Science is obsessed with figuring out if two variables directly affect each other, it ignores the much more important question of how two variables can influence each other.

This has big implications for medical science.

  1. This does not have to be the case - we can do better science by keeping the system intact and characterizing its “transfer function” - or how outputs depend on inputs - without breaking it apart. This gives a more in situ understanding, while also being more useful for our goals of understanding the full system. ↩︎