How to teach the local style

Hi everyone,

This is the weekly free issue of “Overthinking Everything” which goes out every Wednesday. There is also a paid issue every Sunday. The last one was Some notes from behind the scenes, in which I reflected on the direction of the newsletter¹.

As a consequence of those reflections, I’m going to be having a bit of a change of direction: The newsletter has got a bit too self help recently, and I need to change that. I like self help and I’m happy to do it on occasion, but it’s become too much the default, and I’d like to take a step back from that.

So for at least a few weeks I’ll be returning to my first love: Books. I’ll be basing each issue around something I find interesting from a book I like. It won’t be quite the random book prompt approach that I’ve advocated for before (A Guide to Starting a Daily Writing Practice), but will be a bit more targeted than that.

The Uncertainty Mindset

This is a book by Vaughn Tan, who also writes here on substack on a newsletter of the same name.

It’s a study about high end experimental cooking, the sort you do in restaurants who can afford to invest in R&D and do genuinely frontier pushing food research, and what we can learn from world that applies more broadly.

His idea of the “uncertainty mindset” is that it is a worldview required for genuine innovation, which accepts the future as in some sense fundamentally unknowable, but also that we have to figure out how to navigate it anyway, and he argues that high end food R&D is an excellent example of people embracing this mindset.

He uses “true uncertainty” in contrast to “risk”. With risk he means something along the lines of “One of these ten things will happen, and we have a reasonably good idea of how likely these are”. With “true uncertainty”, the number of things that might happen might be effectively infinite. R&D exists in a world of true uncertainty.

There’s a line that is almost not certainly due to Einstein as usually claimed, but which I still quite like: “If I knew what I was doing, it wouldn't be called research.” Part of Tan’s argument is that it goes deeper than this: With research, we don’t only not know what we’re doing, often we don’t even know what we’re trying to do.

Teaching open-ended style

One of the ideas I find interesting in The Uncertainty Mindset is its idea of open-ended style. An open-ended style is a style that you can innovate within: Things in this style feel familiar, but the style has enough room for expansion that things within it can feel both familiar and genuinely novel.

In a restaurant context (at least, the sort of restaurant Tan is writing about), you have to be able to produce dishes that both fit your restaurant’s brand - there has to be a sort of thing that people come to you for - and provide a constant stream of novel dishes within this brand. There is a distinctive thing you can experience as “A Fat Duck dish”, even as they constantly come out with new dishes.

In contrast, without that need for innovation, you get a closed-ended style. A closed-ended style can be codified - you can say exactly what is and isn’t in the style. With an open-ended style, you generally cannot:

Rick Billings, ThinkFoodGroup’s head of pastry R&D, told me about a restaurant he knew where “they tried to lock down the style by documenting the hell out of it. They locked it down all right, but it destroyed the spark. They’ve stopped making really new things there.” Understanding open-ended style in sufficient detail to create new things that are infused with it requires knowledge that cannot be articulated. Moreover, attempts to put this tacit knowledge into words seemed to destroy it. Knowledge of open-ended style is inherently tacit: knowledge that cannot be put into words for easy communication

“Knowledge that cannot be articulated” is not, I think, quite right, in that it turns out that the way to communicate it is, partly, to articulate it. Specifically, it’s to articulate it in the context of a particular attempt to create something in that style. The teams in question did rapid prototyping, with short iterations, and in which the whole team would provide feedback on the item according to their own sense of what the style should be.

This process of feedback is illustrated well in the construction of a sandwich:

HERNANDEZ: I like the baguette here but the crust on the urchin is too thick.

GARCIA: Why do you say the urchin crust is too thick?

HERNANDEZ: The urchin’s too creamy and by the time you bite through the bread and the crust it’s oozing everywhere.

LOZANO: Yes, that’s a problem, but it’s because the crust on the baguette is too tough. The crust on the urchin is too heavy also, but the more important thing is we should be using a different kind of bread.

GARCIA: The idea here is to have it be like our other courses, a two-bite with perfect textures and flavors. The crust on the urchin should be much more delicate, and you should be able to get the crust through the bread—

TURNER: —like the calamari fritti [a dish previously on the menu at another ThinkFoodGroup restaurant]?

GARCIA: Yes, but make sure it isn’t overcooked, it has to be delicate. What are you thinking for the bread?

TURNER: Maybe like a brioche?

HERNANDEZ: Or one of the steamed breads? Those have barely any crust at all.

GARCIA: Good idea. Try to make the difference in textures [between the fried urchin and the bread] really big. What did you all think of the flavors?

LOZANO: More acid, and more bright.

HERNANDEZ: I thought it needed more salt, too. Maybe salt plus lemon zest after frying—

GARCIA: —and at service finish with a bit of [lemon] juice so the urchin doesn’t get soggy. Done. Can you get the new bread ready by today?

By directly discussing the ways in which something is “not quite right”, people both improve the quality of the dish being constructed and also collectively refine their shared understanding of what “quite right” looks like.

This kind of process of collective iterative feedback is what creates a shared style. Somehow I’ve never really seen that working out in software development. I’m not entirely sure why - it may be that we just don’t really do this sort of open-ended style that well - but it feels like something we are missing rather than something that we don’t need.

A collective felt sense

One of the things this description of open-ended style put me in mind of was Focusing - the skill of trying to articulate the emotional felt sense of some nebulous preverbal thought. I’m particularly put in mind of Kaj Sotala’s article explaining it, and a passage he quotes from Mark Lippmann:

When you say something, and it doesn’t come out right, you try again. Where your mind goes before you try again, that’s felt meaning.

When someone says, can you explain that in different words? Your mind goes back to that, in other words, felt meaning.

When someone says, what do you mean by that? Your mind goes back to that, in other words, felt meaning.

When you’re writing something, and it’s hard, what you’re searching for is felt meaning. Where you’re searching is where felt meaning will be.

When you’re writing something, and it’s easy, you’re drawing upon felt meaning.

When you lose your train of thought, you’ve lost your sense of the felt meaning you were speaking from. When you remember what you were saying, the what is felt meaning.

When you have a particular style, there is a thing that it is like for something to be in that style, and you have this kind of preverbal feeling for it that you are trying to communicate.

The way to do that is to encounter things and articulate the direction to the boundary - what makes them that, what makes them not that. In this way, you construct a partially shared felt sense - nobody has exactly the same felt sense, but everybody has a felt sense that they can refer to as “the house style”, and by repeated encounters with each other in the course of working together, these are gradually brought into alignment.

Connections

When reading a book, it’s natural for me to want to connect it up to other books, so here are some interesting connections that come up for me.

The first is that this process of “constructing a shared felt sense” framing of open-ended style seems very like what Thomas F. Green describes as normation in his “Voices: The Educational Formation of Conscience”². The cooks in question are part of a very specific community of practice (their kitchen) and are learning a “voice of conscience” that teaches them a sense of taste conveying what that community is trying to achieve.

The second is that the process of feedback reminds me very much of the shop talk in Talking About Machines, where photo copier repair technicians would talk a great deal about their work with each other, as part of sharing knowledge and calibrating their shared sense of what a technician does.

This kind of collective back and forth as part of establishing the norms and preverbal shared understanding of a team seems like an underappreciated feature of work, and is something that I’ve often felt was missing as part of my software development work.

Postscript

I’ve been running coaching auctions for my newsletter issues recently. As part of the change of direction I’m going to temporarily suspend those, though I hope to resume them at some point. In the meantime, if you’d like a coaching session with me about any topic, you can book one-off sessions using Calendly.

Also, as many of you already now, there is now an Overthinking Everything discord server, which I’ve been enjoying having, and I encourage you to join us there by clicking this invitation link.

The image for this post is of Gordon Ramsay’s restaurant “Petrus”, and was taken by Nick Webb and obtained from Wikimedia commons.

1. And also threatened to start writing about Stargate. This probably won’t happen. Probably.

2. This book is hard to find, but I wrote a bit about it in Jiminy Cricket Must Die. My opinions on the subject have moved on somewhat since then, though I still think it’s basically sound