Metis matters

A detour into James C Scott via Dan Davies

I’m bad at resisting very interesting arguments, so a few quick reflections prompted by:

• Reading The Unaccountability Machine by Dan Davies

• The death of James C Scott

• A couple of posts from Dan on Scott (one, two)*

The main thing to say is: Dan’s book is very good. Read it to understand the polycrisis/why the world has lost its mind/if you’ve been meaning for ages to get into cybernetics and not gotten round to it.

Also now is your chance to read Seeing Like a State if you haven’t already.

What intrigued me last week was Dan’s take on Scott’s concept of metis, which I think I disagree with, although it’s possibly a difference of semantics/emphasis.

Before I jump into the difference, a quick recap on Scott’s distinction between techne and metis, which is something like the following:

• Techne is technical, scientific, and formal knowledge. It describes methods, procedures, and rules that can be systematically learned and applied. It tends to be codified, explicit, formal, standardised, top-down, abstracted, generalised, and universally applicable.

• Metis is practical, experiential, tacit knowledge. It’s the accumulated wisdom or knowhow that comes from practice or lived experience. It entails adaptability and the ability to respond intuitively in context. It tends to be local, situational, informal, decentralized.

Dan’s view, which I take to be a little provocative, is that:

in an important sense, there’s actually no such thing as metis.

The argument, as I follow it, is that:

• Techne is our best effort to codify the most important bits of metis

• We have no choice but to do this because we need ways to attenuate variety / compress information that would otherwise overwhelm our systems of government. (NB: Dan’s book is a study of cybernetics, hence nice phrases like ‘attenuate variety’.)

• The safeguards we have against this going very wrong are that: (a) bureaucrats live in the real world, so they have a decent sense of when the rules are veering away from daily reality. And (b) we can construct what Dan calls “red handle signals”. Ways for people to scream when the rules of techne are leading to unbearable outcomes, so that we can escalate this feedback to higher levels of the system.

• When feedback is escalated up the system in this way, the way we respond is essentially techne. i.e. we modify the rules / procedures of the bureaucracy, or we reallocate resources, to make outcomes more bearable. So it’s techne all the way through.

The result of all this — when the red handle signals work well — is stability. So it’s not all that bad an outcome, even if things sometimes get unbearable for a while. i.e. we don’t need to follow Scott into anarchism.

Plus, remember where we started: we have no choice but to compress information/attenuate variety, so this is the best we can do. It’s a non-answer to say ‘we should metis our way to a better outcome’.

The reason I’m labouring this point is that I think the metis / techne distiction is surprisingy important — politically, for social stability, etc.

And I feel just a bit icky about the conclusion that ‘there is no such thing as metis’, or the idea that this could be a stable outcome.

I think fundamentally I’m still hooked on Scott’s account of how profoundly techne differs from metis, and by his really quite brilliant and lengthy exposition of techne’s blindspots.

This isn’t to say, or isn’t just to say, that we lose something when we compress/abstract metis into techne (for example, when we count things in a bureaucracy, or when we draft abstract laws). As Dan explains so well in his book, this loss of information/compression isn’t just inevitable, it’s one of the valuable jobs of bureaucracy. It’s what lets us govern.

The worry is more that techne loses, irreplacably, certain types of information that are vital to responding to certain types of red handle signal. (NB: I’ll come back to that word ‘information’ in a moment, because it might be part of the problem.) And so maybe without metis we’re incapable of responding to some types of scream.

I’m thinking, for example, of screams that say things like ‘I’m experiencing an unberabable loss of dignity’, or ‘I am suffering from an existential loneliness borne of a lack of deep human connection’.

These are languages that techne can’t speak, so a bureaucracy that speaks only techne will be incapable of responding to these types of scream.

Actually, it’s worse than this, because the bureaucracy will respond, but it will do so by translating the scream in a way that doesn’t just lose information but that distorts the scream’s meaning, so that when the answer comes back it won’t mollify the screamer, it will actually antagonise them. i.e. the system will respond in a way that has manifestly failed to hear the person as a human. And this causes the last thing we want: instability in the system.

Maybe we can channel Scott for a moment and test this way of thinking with an over-extended analogy from nature.

Let’s say our systems of government are harming a forest because of slice-and-dice techne, so that a scream is coming back from forest mushrooms who are shouting in a cute little fungal chorus: ‘your monoculture forestry system is severing our mysterious subterranean mycorrhizal networks!’

The worry isn’t that techne will fail to ‘solve the problem’ the mushrooms are screaming about, or not quite anyway.

Techne may well come up with an ingenious solution — maybe some chemical fertiliser will be developed that feeds fungal networks, compensating for the now sparse and homogeneous leaf litter (I am speaking at the edge of my knowledge of mycorrhizal networks).

But maybe the fertiliser doesn’t quite hit the spot, i.e. maybe the mushrooms get bigger, but their nutrient balance is all off, and so deepdown they’re still unhappy/unwell. And maybe the fertiliser causes other imbalances in the ecosystem, harming the moss that was growing symbiotically with the fungus (now I’m just making things up).

And so maybe what we realise, eventually, is that the problem is the method: you can’t foster a healthy ecosystem by isolating individual components/applying abstract technical solutions to parts of the system.

Scott gives some nice examples of this kind of thing in Seeing Like a State. e.g. we replace a mixed-species old woodland with straight lines of Douglas Firs and then we deal with the collapse of biodiversity by fitting plastic bird boxes and hedgehog houses, etc, which is something we actually do.

I’m falling back into Scott’s original argument, but the point is just that there are are deep epistemological differences between techne and metis, related to what it means to know something and how we relate to things in the world. And so when the customs of metis are transcribed into techne these differences are lost in translation in a way that it not just information-lossy but dangerously distorting.

I’m sure, by the way, that there are ways to express all of this in better cybernetic language. e.g. maybe this speaks to Dan’s point in the book that stable systems don’t just need information, they need information they can act on in ways that are stabilising. So maybe what I’m describing is a failure of the red hand signal to capture sufficiently rich information (which would speak to live debates about richer forms of public feedback and engagement, e.g. deliberative mechanisms, collective intelligence).

Or maybe — back to that word information — what I’m fumbling for is a deeper critique of any theory of information. Maybe what makes metis distinctive is precisely that it goes beyond information. Maybe so much so that this could serve as a kind of definition. (Is knowhow information? Is the intuition that arises from years of practice a function of information?)

I appreciate that this sounds academic/dancing-on-pin-heady and I’m not wholly sure if these are just differences borne of semantics.

But the reason I think it’s worth sitting with these distinctions / definitions for minute is that there’s a fair case to say that a one line history of the world since 2008 would be:

technocrats respond to people’s human screams in overly-technocratic ways that only serve to further antagonise them

Which is similar in some ways to the argument Dan makes in his book about why the world has lost its mind, but is, I think, a different way of framing the problem, or at least a different emphasis, that takes us to different solutions.

Still, I’m putting this out there lightly. (Not least because I’m writing this on a break from writing The Book in the margins of life, which I’ve only allowed myself on the condition that I hurry up and get back to writing The Book quickly.)

For now, though, my instinct is: there is such a thing as metis and it matters.

Where does this leave us?

Interestingly, to me anyway, my instinct that metis matters a lot doesn’t lead me full circle back to Scott. i.e. I don’t buy that the way to make things more bearable for more people more often is anarchism.

At root I’m an optimist in thinking we can get better and better at developing schemes to improve the human condition. But where I am on board with Scott is in thinking that these schemes cannot be dreamed up in the high modernist fashion of Le Corbusian blueprints/Leninist state planning/large-scale technocratic capitalism.**

I guess the way I’d put it these days — and the thought that I find to be increasingly a source of optimism — is something like: it is possible to build systems of government the enable metis at scale.

And, no, I don’t think I buy the obvious rejoinder ‘errrr, yes, metis at scale is what we call techne’.

I think a lot of what’s happening in debates at the frontier of governance, both within the UK and globally, is essentially work to learn (re-learn?) how to foster the conditions for metis at scale.

Or actually, the way I tend to see this work these days is as a project to broaden our repertoire of government into one that’s more integrated, i.e. an approach that integrates, or incorporates, both techne and metis.

I appreciate this sounds all a bit California woo woo, and that I’ve at best loosely assembled the argument (if only I didn’t keep getting distracted…). But I guess to finish, here is a grab bag of some of the materials I’d use to more fully construct the argument:

• There are now thousands of people spending whole careers at the frontier of governance / public service around the world engaged in what is, I think, a metis-heavy approach to governance or problem-solving, e.g. people running local services that are centred in agency, or services that start by strengthening individual human relationships, or people applying techniques for evaluating and improving public policy in situ. Arguably this is part of my day job too, i.e. practising a way of solving complex problems that is a lot about repeated learning by doing.

• This work is now maturing or ‘going mainstream’ or spreading, while staying consistent with the basic idea of working in situ. e.g. people are maturing ideas like relational state capacity, and people are applying evaluation techniques like Theory Based Evaluation, adaptive management, etc., that are all about improving outcomes in context through repetitive cycles of doing and learning (as opposed to the evaluation/scaling/replication paradigm of RCTs).

• This work has some deep philosophical foundations that feel consistent with the concept of metis-based governance, e.g. I think of Elizabeth Anderson’s pluralistic theories of value, which speaks to the damage caused by abstracting, and also her work on the philosophical significance of personal human relationships and the particular (and hard to measure) qualities these relationships have. There also strong roots stretching back into the work of Jane Jacobs.

• I’ve spent quite a lot of time recently with people who are experts in indigenous forms of governance and knowledge, especially during a visiting fellowship in Australia and Aotearoa New Zealand and a Governance Futures workshop in Hawaii.*** I was embarrassed at how eye-opening I found this and it left me with a deeper, more sophisticated sense of the options for sustainable, humane, situated approaches to governance.

• It feels increasingly clear that care is the domain of life that most exposes the inadequacy of techne, and even of a system of techne with well functioning red hand signals. I wonder if care might be the itch that technocratic capitalism proves incapable of scratching, and the societal challenge that most urgently demands a more integrated approach to governance. e.g. see Madeleine Bunting’s Labours of Love and Hilary Cottam’s classic Radical Help as ways into this.

I’ll just leave that all out there as material to mull on.

As I wrote above, I’m still not 100% sure this isn’t semantics; some people would say I’m describing precisely what Dan was describing in his blogs linked above: a process through which the higher-level parts of our systems of government hear people’s screams and respond by updating techne, in this case by integrating more humane methods.

I think, though, my view is that this interpretation loses something. And that the work of improving our systems of government is more likely to succeed — i.e. it’s more likely to achieve a more stable system/society — if it’s conceived as enabling metis at scale.

Anyway, that’s my permitted time limit for deviating from what I should be writing. Interested as always in thoughts, critiques, etc.

Foonotes

* Yes I know this naming convention is inconsistent but it feels too weird either to call Dan ‘Davies’ or Scott ‘James’, so I’m running with it.

** I’ve always thought Scott is responsible for one of the most throwaway lines of all time when he writes in Seeing Like a State: “Large scale capitalism is just as much an agency of homogeneisation, uniformity, grids, and heroic simplification as the state is”, before quickly getting back to his book-length critique of the state as an agent of homogeneisation. Surely yes, that point about capitalism is very right and deserves more unpacking.

*** Let me emphasise that these are pretty much the only two jollies I’ve ever had, neither funded by my day job, and yes they were fun but also genuinely productive I promise.