There's a really great chapter in Underland about mycorrhizal networks. Clever wordplay on 'wood wide web', because this discovery was contemporaneous with the emergence of that amazing digital tool. I like the hope in the riff; it's a postcard from the past with a faded image of the desirable future.

Anyway, the story goes like this: there were a group of forestry botanists who wanted profitable trees to grow, so they cut down the unprofitable mature trees of a different species to make space for profitable saplings. All the saplings died or were stunted as a result.

There's information missing in this system, about this system. Hubris lives here. Research by Suzanne Simard, a UBC professor of Forest Ecology provided the missing piece:

Simard's field studies revealed that trees are linked to neighboring trees by an underground network of fungi that resembles the neural networks in the brain. In one study, Simard watched as a Douglas fir that had been injured by insects appeared to send chemical warning signals to a ponderosa pine growing nearby. The pine tree then produced defense enzymes to protect against the insect.

This unknown network, in the case of the dead and stunted saplings, is used by young trees to communicate need for nutrients, and by old trees to respond to that need with their excess energy. This transport exists across species – the trees in the above example were paper birch and douglas fir. The fact of this breaks so many of our ideas about what a forest is, and more importantly what a species is.

There are no individuals in the forest, these facts tell us. Why do we insist on forcing reality to conform to our categories, even under the guise of science? Robert MacFarlane's book makes the point that our prejudice means we have to learn by scientific means what is indigenous common knowledge – the analogue to the English word 'world' in many languages is 'forest'. In fact, much of the ideas about connection to planet and membership in complex biological systems is simply absent from modern languages – how can it be that we don't have a word for our planet as one living thing? Because politics, I think.

It's a cliched example, but this is the function of the Ministry of Truth in Orwell's seminal work: to limit discourse only to those ideas deemed acceptable by the powerful. Sure it's a stretch. But here's a place where, lacking even the right question to ask, we injure what the entire discipline of Economics considers its root: that we act rationally in service of self. These forestry botanists killed their profit with their inability to consider the forest as a system – they were literally missing the forest for the trees.

The implications of this network under the forest blow my mind. This is a network like we know to build. It's ad-hoc, addressable, operating-system agnostic, high-throughput, and decentralized. The idea of intra-species communication and resource sharing is simply so alien to our individual-centric mode of thought. We approach objective study with this destructive prejudice; where else are we unknowingly injuring ourselves with our commitment to misunderstanding?

In conversation with a botanist, MacFarlane makes the argument that it's the work of artists, poets, writers, historians, and linguists to make language for these parts of the world that don't yet fit what our words currently do. I wholeheartedly agree that language is a living thing that must evolve, and that we must consider language carefully. I suggest that we should also understand what power is served when language is allowed to be stunted.

There's a parallel here to political thought. Throughout the 1970s-1980s there was an effort rooted in Cold War politics to mark philosophers and their individual works as 'totalistic'. 'Here be dragons', says the sign we erected near Hegel, Marx, Lenin, Goldstein and their contemporaries. Any thinker in the 'liberalist' Western tradition who elected to cite works in the 'totalitarian' category was in for trouble, so the ideas contained in these works did not gain the purchase they deserved.

This has had hilarious effects. International Relations, prior to 2001, did not have interest in the activities of groups within states, largely because that analysis had been deemed 'Marxist' due to its reliance on that author's 'Critical Theory' methodology. The rise of International Political Economy as a discipline is largely in response to the toolkit that was made suddenly inadequate on Sept 11, 2001 – not an attack by a state, so what do we do? How do we understand and respond?

Our choices, these mushrooms teach us, shape our language and thereby shape how we begin to understand the world.

This is the core danger of the current AI paradigm. Knowledge is lost in the shift from the language of doing to the language of requesting. When I ask an AI to write code that solves a complex but annoying task – form validation, for instance – I atrophy my own skill at handling that task. So when AI is unavailable, these tasks become increasingly difficult for my lack of practice. Request work from AI long enough, and tasks become impossible without it. In terms of language, the syntax and grammar become stunted.

This argument is not new. It was made when we moved from writing machine code to compiled languages. Then again when linkers were introduced. Then again with frameworks that abstract away strange parts of underlying languages. It's almost an anti-progress argument, and I almost hate making it.

But, consider ownership. As we dull our skills by asking AI for code instead of writing it ourselves, what happens when we ask for something the AI has been instructed by powers greater than us not to provide?

Power, sometimes, resides with the priests instead of the oracle. While we have agency we should be careful where we rest it, lest we kill our saplings for our structured ignorance.