Against Scientific Management

John Senior, Rational Grazing, and Poetry

In a talk on the Odyssey, John Senior remarked that it is entirely unnecessary for man to study science. However, it is of the utmost importance that he study poetry/literature. He went on to bemoan that engineering students would dread their “humanities” electives, and pick lousy ones like linguistics or economics - lousy not because these are bad in themselves, but because there is a neglect of the very fabric of reality: poetry.

Guilty as charged.

I took economics. Granted - a “history of economic thought” class that has been influential in getting me to where I am today. Also, other available classes like “bible as literature” (you may as well read the bible like a cookbook if you’re going to treat it like mere ‘literature’) weren’t exactly the Great Books courses that Senior advocated.

Nonetheless, as I wrestle with the likes of Senior, it has occurred to me that something we must unlock is a “poetic mode of thinking”. It is not the only mode of thinking that is useful, but it is actually the bedrock.

I have noticed that much scientific thought is becoming very divorced from reality. Clean, sterile, and behind a computer - rather than in the field, in the trenches. The poetic mode is engaged with the things, it uses metaphor, not principles and abstractions.

This past week I was reading Grass Productivity, when Voisin remarked:

Obviously a technical man would have said: “In view of the too short rest period I have given the grass it will not provide its maximum daily production, since I am not at the maximum of the productivity curve as shown in Figs. 3 and 4: the total output of grass will have diminished, as Zurn’s work has shown (Table 2)." The Breton peasant did not need to see our curves: he, and his ancestors, were intuitively acquainted with them. They knew that by cutting their grass after an insufficient period of rest their annual production would be diminished."

-Grass Productivity by Andre Voisin

This reinforces Senior’s point.

In my reading of the agricultural literature it actually seems that in many ways we have lost the plot by chasing the line. And that’s a great way of putting it - losing the plot. It’s not that the facts were wrong. They just aren’t the story.

Goodhart’s Law

xkcd #2899

If we lose touch with the facts - if we constantly bring things back to the ‘end results’, we fall prey to Goodhart’s law:

When a measure becomes a target, it ceases to be a good measure.

Put another way, whenever you make something a goal, everyone starts to game the system to get that metric. Or theologically speaking, we worship the creature (finite) rather than the creator (infinite).

Table that for a moment. As engineers, we want to design/facilitate good systems. Good systems are resilient.

The discussion of resilience has largely been around the idea that we should have ‘fault tolerance’. Or, even, that systems should get better when they are stressed - antifragile. That has some degree of nobility in it but it misses the mark. We need people - with their prudence - in the systems. Not just monitoring the output, but actually having their hands in the systems, because otherwise we fall prey to Goodhart’s law. If we just trust the metric, rather than being involved in the whole process, we lose something.

Men in the System

Let’s follow that to its conclusion: the ideal state of systems is not that people are less involved in them, but that they are more involved with them. But this can’t be just a silly make-work sort of involvement. This has to be a meaningful involvement - what some people point to when they say an “ownership mentality”. The man in the system shouldn’t have a bullshit job.

When machines go wrong, in America, we have PEOPLE that come up to the mark. - Richard Nixon

I’ll probably write about this further, but it’s the difference between a machine operator, who knows how to use a particular machine, but did not set it up, does not really know how to fix it, and is really only there as a stop-gap until “full automation” can occur - and a machinist, who can not only operate, but set up, diagnose, correct, and most importantly - redirect a piece of equipment. This is a matter both of the character of the person, and the freedom he is afforded to act in this virtuous way as a machinist.

But what is the machinist doing, really? If you get to know one, you’ll quickly realize that machinists are actually much less scientific and coldly “rational” than machine-operators. Their knowledge is not theoretical - it is incarnate, and they often think of things not from abstract principles, but from extrapolating actual experiences - from likening what they are facing to something they have in the past. What’s more - each machinist will solve the problem in a different way - even if they may ‘converge’ at a similar result, they will have differences in their art - they will have character.

Against Bugman Science and Towards Poetic Holism

So, how do we get past Goodhart’s law?

In a way, it’s a shaking-off of metrics. That isn’t to say to dive headlong into a cesspool of relativism. It means more that we need to awaken our senses - so that we know the good when we see it.

Nowhere is this more apparent than our food and medical systems. The powers that be would have you believe that overconsumption of simple carbohydrates has nothing to do with type-2 diabetes, which is something you can only say if you’re so head-high in scientific papers that you can’t see reality. Perhaps it’s not exactly the simple carbs. But it clearly is in that direction and only a fool needs exhaustive scientific proof to move away from that direction.

I posit this (it would be ironic to prove it): life is so complex and multivariate that the scientific method cannot produce holistic solutions.

How do we get there, though? How do we navigate the peculiarities of life?

The Limit of Science - Follow the ‘Direction of Goodness’ only for a while

An engineer at Tesla gave me this philosophy on how to apply laptime simulation to the development of Formula SAE cars. But you can do this with anything. It’s stuck with me. Don’t use calculations (scientific management) of what you’re doing to figure out the endpoint. Use them like a compass - use them to find the “direction of goodness” - the right set of tradeoffs to make right here, right now.

But that’s all it’s for. It’s not for setting the long-term goal. That’s for poetry, philosophy, and your innate sense of things that will be developed as you follow the path.

Eventually, as you walk it, you will become aware of new variables and phenomena. Accounting for these will lead you to make a new compass, with a new magnet.

Again this means we need men in the system - and not factory workers, not button-pushers, not even overseers - but philosophers with dirty hands.

Embrace Repair

We implicitly design systems that don’t need to be rebuilt or repaired. But… what if that’s a wrongheaded idea? Perhaps systems that break and are fragile are actually good?

I have noticed this in developing (complex, automated) systems: build a low-quality version first, or one that’s semi-manual. This way you will be forced to intervene in it, forced to really grapple with what exactly is going on. You’ll learn things by being forced to rip the thing apart.

What if instead of treating any particular system as “finished” - and only in need of a button-pushing operator - we left systems “in development”. We should look at this state not in “need” of further development, but with opportunity for further development.

The example which I have right in front of me is our farm. We adjust what we do every year - it is continuous improvement. But it is different than what auto manufacturers do when they come out with new models - we’re always refining the same land - not ripping it all up and starting again.

The Car Exemplar

What if cars looked more like this? The idea of modularity (swappable components, ship-of-Theseus style) has been harped on with very little success… perhaps because it is simply not radical enough.

Well, the trouble is, that cars (at least, playing their role in suburban transit) require optimization: cheap costs and high efficiency. Maybe it is possible to make a more utilitarian vehicle that meets those criteria. But perhaps it’s just asking too much. Cars demand temporary, high-maintenance pathways to travel on, using wearable parts: roads and tires. They carry an extra 1000 pounds per person. We could go on about the implicit constraints of road transit - but they are really demanding. To make such a system work with a modicum of efficiency requires optimization - scientific management.

Can we envision sorts of transit which meet our needs while doing it in such a way that doesn’t require scientific management? We could just scale back our needs - and then, the Amish cart and buggy seems to fare decently. Bicycles are often more convenient (and repairable!) than cars in dense areas - and e-bikes extend their utility. I don’t know what the replacement(s) for cars would be - that’ll require some creativity.

But it’s not electric cars as we’re coming to see them. EVs are fundamentally the same thing as ICE cars, it’s just that they weigh a little more, and outsource fuel combustion to a centralized power plant rather than an on-board engine.

I think the folks at Edison Motors are going in the right direction of goodness, as they design vehicles which use readily available components rather than custom ones. Plus, the diesel-electric thing is more sensible, and automobile technology makes sense for logging or hauling big stuff over long, unimproved distances.

Lives for People, not Jobs for Computers

Good systems don’t require scientific management.

The management they do require should not need hypotheses, extensive testing, p-values, peer-reviewed studies (cue shrieks of rage from the scientific-industrial complex) and the rest of the appurtenances belonging to the church of scientism.

No, the management should be artful. Intuitive. Embodied.

I’ll be frank: this is infuriating. It’s also very alien to our world where “you can be anything” as opposed to following in the traditional craft of your father. As I read through the books on farming, so much of it comes back to touchy-feely stuff. I can pretty well wrap my head around schematics and drawings of parts that I have never seen, I can analyze them, etc… but no amount of literature is sufficient to describe a plant or animal adequately.

Here’s the rub. Just doing that, not having the touching and feeling… turns you into a Chinese room. You can manipulate symbols, but not meaning.

Poetry does quite the inverse: it manipulates meanings of the symbols. It conveys meaning - it can convey truth (or falsehood) without being syntactically correct. This is an incredibly important skill to have. I have seen it over and over and over again in mechanical design where someone gets to a right solution - not for the wrong reasons - but for reasons that they could not express in a “technically correct” fashion.

This is the mystery of incarnation, and we are to embrace it.

But we the faithful marveling at the mystery cry out with faith:

Rejoice, vessel of God's wisdom;
Rejoice, storehouse of God's providence.
Rejoice, revealer of philosophers as fools;
Rejoice, exposer of the technologists as irrational.
Rejoice, for the fierce debaters are made foolish;
Rejoice, for the creators of the myths have wilted.
Rejoice, breaker of the webs of the Athenians' logic;
Rejoice, filler of the nets of the fishermen.
Rejoice, drawer of many from the abyss of ignorance;
Rejoice, enlightener of many with knowledge.
Rejoice, ship for those wishing salvation;
Rejoice, harbor for life's navigators.
Rejoice, O Bride unwedded.