Oddly Influenced

Welcome to Oddly Influenced, a podcast about how people have applied ideas from *outside* software *to* software. Episode 38: The trajectory of a collaborative circle.

The collaborative circles described in Farrell’s 2001 book of the same name share, in rough terms, a similar life cycle or trajectory. Farrell names four distinct stages that a circle goes through from initial formation to the creation of a shared vision and its supporting techniques. The nature of the circle changes sharply from stage to stage. I’m not fond of sharp divisions in general, and I don’t think his evidence really supports them here. I prefer to think in terms of a trajectory where focus and activity shift more gradually. That is, I think it’s enough to say that circles shift continuously from talk to work, from criticism to creation, and from safe consensus to risky disagreement to a new consensus. Along the way, trust increases but emotions and tensions do too. Typically, the tensions come to dominate and the circle disintegrates. However, that happens after what I called “going public” in the first episode, and that comes after what I’m going to talk about here. I’m imagining a team that has a creative breakthrough that leaves them at a new, stable, non-public status quo. That is, the tension and emotions peak and subside, everybody becomes comfy, and there’s no move toward going public.

After describing the typical trajectory, I’m going to dive into two topics that Farrell emphasizes: trust and friendship – but only with some disclaimers about how I’m the wrong person to talk about those topics.

≤ music ≥

But first, let’s talk about the difference between a team and one of Farrell’s circles. Both are groupings of people in what we might call the same profession. Farrell has painters, and writers, physicians treating mental diseases, and social reformers. We have software developers. For my purposes, I’m going to say testers, programmers, product owners, and the like are all part of the same profession. If you want to think of narrower professions – just programmers, say – that’s OK with me. I don’t think it changes much.

No matter who you include in the team, it is a collective, and it produces *group work*. Each individual contributes something to a whole that is usually not even associated with the team, at least not by those who consume the final work.

In contrast, a circle member typically produces *individual* work. “The Impressionists” don’t paint paintings; individual Impressionists do. It’s one person’s signature down there in the corner. Now, there are examples of pair work: for example, Joseph Conrad and Ford Madox Ford coauthored three novels, but that was unusual, and it was part of developing their shared vision for the future of the novel and didn’t continue after they had.

Circle members have more ego involvement with their work than software team members tend to. Rejection of the finished work, or even of an unfinished work, is perilously close to rejection of *them*. This results in what you might call “trust issues” that are different than those within a software team. More about that later.

The other difference is that I’m going to assume that a team’s creative work happens *on the job*. You expect to be paid to do it. This was not the case for the Impressionists, who were supported by their parents, odd jobs, or – sometimes – each other. Elizabeth Cady Stanton’s role as feminist firebrand was something she did when she could find time away from her main job: mother to six children and manager of a small business: that is, her household. Tolkien and C.S. Lewis weren’t paid to be writers; they were paid to be teachers at Oxford. And so on.

In effect, none of these people had bosses who could tell them “no”. I expect you do. As such, I’m going to have to assume you’re operating under the old semi-formal contract that the early “lightweight methods” teams made with their management: you’ll be given leeway to experiment so long as you keep delivering. That makes you better off than Elizabeth Cady Stanton or J.R.R. Tolkien, but worse off than the Impressionists. I’m going to find it useful – if perhaps artificial – to separate your “circle work” from your “normal work”.

≤ music ≥

In the beginning, the circle is a fairly-ordinary social grouping with a few specific characteristics: everyone’s doing the same kind of work (painting, writing, treating patients with nervous troubles), they like talking about work at both the abstract level (what kind of work is worth doing) and the detailed level (which of these two techniques is better), and they are vaguely dissatisfied with the status quo.

In the beginning, the circle’s emotional energy is directed outward, against that status quo. It seems to me that Step One is the harvesting of concrete stories about Bad Things and Bad People. Stories – events happening in time, with cause and effect – grab people’s attention. The best of the stories get told and retold, often to new people who enter the periphery of the circle.

As a consultant and speaker, I’ve told the story of Dawn’s “bright and dull cows” endless times. (See the show notes for a link to one of the earlier versions.) The moral got clearer in the retelling, and at least a couple of times, the retelling enriched or altered that moral. That’s what retelling stories does.

I imagine a lot of early progress in a collaborative circle involves extracting generalizations from stories. Where the stories originally came first, you can now state a proposition about a thing that’s bad, selecting from stories for illustration. It’s the generalizations that are used to move forward.

It seems to me those generalizations are often labeled by what I’ll call “trigger words” or “trigger ideas”. Those are words that in themselves provoke a strong emotional reaction. An interesting example of that I saw recently was a Mastodon thread in which people described how much they hate capital-A Agile. As I read through the thread, I kept recognizing trigger ideas, and eventually I wrote them down. I’ve lost the complete list, but some of them were “people treated as interchangeable (or ‘fungible’) cogs”, “micromanagement”, “pointless meetings”, “Taylorism”, and “empty ceremony”.

What’s grimly ironic is that all of those were *also* trigger words for the people who *created* capital-A Agile, including me. We used to use Fredrick Winslow Taylor and his “scientific management” with its “time and motion studies” as a whipping boy. And here he was again! A quarter of a century later. Said to be one of us.

It threw me into quite a funk, but then made me hope that maybe some of those people will form collaborative circles that will do something about it, and that this episode might be of some use to them.

The trigger words ground the collaborative circle: “whatever we’re about, we’re not about *that*.”

Note the “we”. An important early result of a collaborative circle is solidarity. Yes, good old “us vs. them”. I’ll discuss the distinct roles that bolster such solidarity next time.

Farrell notes that the circle sometimes has the character of a delinquent gang deliberately trying to shock “the straights”.

“Particularly in the early stages of the development of a circle, as members are developing a culture that encourages creativity, they often are more than simply innovative – they are deliberately provocative toward those in authority. During this period the work sometimes resembles acts of vandalism, in which the members desecrate symbols of authority in their field. The group members see themselves as rebels, and responding to one another’s dares, they take pride in devaluing the sacred icons of the previous generation. An example of work done during this stage of circle development is Marcel Duchamp’s replica of the Mona Lisa. The replica is nearly identical to Leonardo’s original, except […] Duchamp penciled in a moustache and goatee on Mona Lisa’s face.”

Farrell actually gets the story a bit wrong, in that Duchamp didn’t make a painting but just defaced a picture postcard for publication in a magazine. Still, a commentator wrote:

“In 1919 the cult of [the Mona Lisa] was practically a secular religion of the French bourgeoisie and an important part of their self image as patrons of the arts. […] Duchamp’s salacious comment and defacement was a major stroke of [freaking out the bourgeoisie].”

Or, as we called it in 1980’s science fiction fandom, “freaking the mundanes”.

≤ short music ≥

In the introduction to this episode, I said that circles move through a trajectory where things like trust increase. At this point, there’s not much need for trust: it’s not risky to criticize the same people and things everyone else is criticizing. And smallish deviations are easier. Here’s H.G. Wells describing a quarrel with Joseph Conrad:

“I remember a dispute we had one day as we lay on the Sandgate beach and looked out to sea. How, he demanded, would I describe how that boat out there, sat or rode or danced or quivered on the water? I said that nineteen cases out of twenty, I would just let the boat be there in the commonest phrases possible… But it was all against Conrad’s oversensitized receptivity that a boat could ever be just a boat. He wanted to see it with a definite vividness of his own. But I wanted to see it and to see it only in relation to something else – a story, a thesis… ultimately to link it up to… my world outlook.”

I think this can be interpreted as a disagreement about what in the tradition of the Victorian Novel they should reject. Wells was a moralist, as were the Victorians, rather than a stylist, and the others of the circle were, roughly, the opposite. Eventually, the disagreement could not be tolerated within a single circle, and Wells left.

So let’s say that the early part of the trajectory is about pumping up the emotional energy and solidarity of the circle, both of which will be needed for the work ahead.

Over time, the focus changes from challenging the status quo to building an alternative. Part of that is done by group discussions like the ones the circle has been having all along, but increasingly group discussions become discussions about sample works. As I’ll discuss in the section on trust, these works steadily become more ambitious and more contrary to the status quo. You can think of them as examples, or exemplars, or test cases that give the circle things to point at and ask, “Is *this* what we’re about?” Farrell doesn’t give enough detail for me to say for sure, but I imagine the discussion would range from the very abstract (“Is the goal of this work worthy?”) to the very specific (“this trick of brush stroke here – does it actually contribute to what the artist was trying to accomplish?”)

It seems that the abstractions – what Farrell calls the shared vision – settle down while there’s still plenty to do discovering and refining specific techniques. I’m also going to suggest that arriving at the shared vision is the point at which tension is highest and solidarity is most fragile (at least until the circle goes public, which typically is a prelude to disintegration). Codifying the shared vision is the point when people like Wells say, “Enough. I’m out.” Once the “fundamentals” have been established, I expect that the excitement of discovering and refining supporting techniques will reduce tension and smooth over any remaining bad feelings. Maybe that’s just because I like experimenting more than arguing.

In his book /Facilitator’s Guide to Participatory Decision-Making/, Sam Kaner describes the “groan zone” where progress seems to slow to a halt. That, Kaner says, has to be endured until the breakthrough comes. I imagine the development of the shared values is often something like a prolonged groan zone. You’ll need to be patient.

Farrell defines the “shared vision” as quote “a shared set of assumptions about their discipline, including what constitutes good work, how to work, what subjects are worth working on, and how to think about them.” That’s more vague than I like, so I’ll propose that the shared vision is mostly or largely three components that are specific alternatives to components of the status quo.

The first is alternative *goals*. For example, late-1990s “lightweight methods” – I shall no longer use the word “Agile” – changed a goal from “reduce the need to change the code because of new discoveries” to “make it cheap to change the code.” In the Context-Driven School of testing, a goal changed from “write enough tests” to “find enough bugs”, which I could generalize by saying the emphasis changed from “increase confidence in the product” to “provide useful information *about* the product – and about the project.

Another important component is what I think of as alternative habits. An example would be the lightweight methods motto “if it hurts to do it, do it more often”. Or: both the context-driven and lightweight methods have a habit of “turning to the code”. You see that in context-driven testing with its emphasis on manual exploratory testing: you’re not implementing a decision made in a test plan, you’re finding out more about the product by poking at it. You see it with a slogan of Kent Beck’s that I remember as “No design discussion should go on for more than 15 minutes without someone turning to the code to do an experiment.”

Then there are the opposite of trigger words. Whereas trigger words are words that *cause* a problem, anti-trigger words are words or ideas that are *evoked by* problems. Riffing on the old “if all you have is a hammer, everything looks like a nail”, these various words are what you *think with* when searching for a solution. In a 2008 talk on the lightweight style, I proposed certain such words that hadn’t been much publicized, even though they or the ideas behind them were common in insider conversation. Today I’d list them as trust, ease, reactive (as opposed to proactive), fast feedback, intentional naiveté, visibility, and joy.

There is perhaps not so much difference between such words and habits, but I think of it this way. A habit is an action, something verb-like. Each anti-trigger word is noun-like, something like a goal. Were you infected by my 2008 words, you’d tend to think that, whatever the problem, you stand a good chance of solving it if you push toward, say, ease or feedback. Or you know that increasing trust or visibility or joy will probably play a role in the solution.

That’s *my* idea of a shared vision, for what it’s worth. I should note that the shared vision isn’t static – after all, I was trying to add new anti-trigger words in my 2008 talk – but the rate of change will slow way down. I expect churn in techniques will go on considerably longer, though no doubt the rate of innovation will keep decreasing. At some point, we’d be justified in claiming a steady state. That might be enough for a circle or a team, depending on individual ambitions. For example, some of the Impressionist circle felt they, personally, had exhausted Impressionism as it applied to them, and moved on. But Sisley, for example, kept true to the Impressionist style, and I think that’s a perfectly fine way to live your life.

≤ music ≥

One thing that stuck out in Farrell’s exposition is how *skittish* people were about exposing new work to their fellow circle members. There’s a common sequence of developing trust:

1. People only share finished work, works they know they can defend.
2. A breakthrough comes when one circle member trusts another enough to share *unfinished* work. The case of Tolkien and C.S. Lewis is representative.

Tolkien grew up fascinated with myths and epic poems. He wrote poetry in that style. Only once had he shared any samples with someone else, a mentor who advised him to drop the whole idea. (Way to go, unnamed mentor!) But there was a day when he and C.S. Lewis stayed up until 2:30 a.m. talking about Nordic mythology. Quoting Farrell: “a few days after the late night conversation, Tolkien gave Lewis one of the unfinished poems to read”.

As was the custom of their time and class, Lewis sent the equivalent of an text message to Tolkien saying, “I can quite honestly say that it is ages since I have had an evening of such delight, and the personal experience of reading a friend’s work had very little to do with it. […] Detailed criticisms (including grumbles at individual lines) will follow”.

The way Lewis criticized those individual lines was brilliant. He pretended to be reading an English translation – by some imaginary scholar – of an original Old Norse poem, together with footnotes from other imaginary scholars. So, when he thought one of Tolkien’s lines fell with a thud, he had one of his made-up scholars claim the *original* Norse line had been mistranslated and then offer up a preferred translation. See how such scholarly playfulness softens the blow compared to “your line 38 sucks because…”?

Part 3. Frequently, this first experience translates into a ritual of sharing unfinished work with the whole circle, often in a formalized way. It did with Tolkien’s Inklings and the Fugitive poets, for example. The pattern writers’ workshops I mentioned in bonus episode 36 are an example of a ritual deliberately constructed to make sharing safer. Trust in the process reduces the need to trust each individual reviewer.

4. Such group sharing is still a bit skittish. People don’t put forth their craziest ideas. That happens when people pair off and work together at length. The nature of the pairing varies. Impressionists Renoir and Monet worked side-by-side on their individual paintings, commenting on each other’s as they went. The Fugitive Poets Davidson and Tate exchanged unfinished manuscripts by mail (Tate was off being treated for tuberculosis) and got criticism back, criticism that “gradually ratcheted one another into taking greater and greater risks – exposing vulnerable experimental work and responding with more and more honest, detailed, and helpful criticism.”

I’ll have more to say about the collaborative pairs next episode, but for now I want to talk about what’s required for people to, first, put unfinished work up for criticism and then, later, risk exposing even their most half-baked ideas. Farrell uses the blanket word “trust” for both. I don’t have a visceral understanding of the issue, because – unless I’m fooling myself – I don’t feel particularly nervous about being corrected or expressing half-baked ideas. Because of my particular screwed-up childhood, I don’t mind being wrong. What I hate is being *ignored*.

Until writing the scripts for these episodes, I was inclined to feel smug about our superiority over all these tentative sharers: Freud and Tolkien and Davidson. They, I thought, were a match for the bad old days of software – 1981 – when programmers would disappear for a month or more, then reveal their complete, finished, as-perfect-as-possible code, nervous and a bit resentful about bug reports and stylistic suggestions alike. Surely nowadays people are blasé about publishing (via Git or Slack or whatever) a smallish piece of unfinished work and getting reviews of it.

Or maybe not.

Maybe your team will also have to make the progression from low-trust stage 1 (revealing only the safest work) to high-trust stage 4 (unfazed by revealing any idea, no matter how tentative or outrageous). I can’t see that Farrell has any circle-specific advice to give (except, perhaps, to do it gradually), and nor do I, really.

However, I do have one thought. Suppose you think you have a high-trust, or high-safety environment. A team member is not afraid of criticism because they know it’s offered in an attempt to help that member succeed, not for the critical team member to “win” some status contest. However, let me suggest there are two different kinds of trust, because of two different kinds of fear, of two different kinds of criticism.

Tolkien knew C.S. Lewis shared his enthusiasm for Nordic mythology and epic poetry. So the fear would be that he wasn’t *good enough* at doing a good thing.

He’d have had a different kind of fear if he’d given the poem to one of his modernist academic colleagues, who thought mythology and epic poetry were silly things for an educated, grown man to attend to. He would have likely feared something like would have likely provoked something like Wolfgang Pauli’s (somewhat apocryphal) comment on a physics paper: “That is not only not right; it is not even wrong.” Or, another comment of Pauli’s: “What you said was so confused that one could not tell whether it was nonsense or not.”

I want to suggest such a comment is more wounding. It goes far beyond saying “you’re not good enough (yet)” to “you’re incapable of recognizing what good even is. You do not belong here.”

And so I want to suggest that you may think your team is full of trust and psychological safety, but it may only be of the first sort. And you’ll need the second sort to make real creative breakthroughs rather than small increments to the status quo.

Also note that a “not even wrong” idea will surely not pass muster according to the “good thing done poorly” criterion. That’s the reason I think “normal work” needs to be separated from “circle work”. If you and a pair are trying out a wild idea, the results will very likely not be suitable to be committed into the official code base. Half-baked team organization ideas might for a time need to be an addition to the accepted ideas, rather than a replacement. No getting away from daily standups while you work out how not to need them. And so on.

≤ music ≥

Finally, friendship.

Collaborative circles typically – but not always – start as a set of people who enjoy each others’ company, both individually and in groups. Farrell characterizes those as developing into friendships: “Friendship dynamics” is right there in his subtitle.

When translating that to teams, I get an uncomfortable echo of “We’re all a family here in BigCorp”. We’ve recently seen how that then translates into “… and we have to throw 10,000 of you children into the jaws of Moloch because we overhired during the pandemic”.

Fortunately for us skittish about mixing our work and personal selves, I don’t think “friendship” really deserves the prominence Farrell gives it. I admit I have no great talent for friendship – I need one true friend and I married her three and a half decades ago – but a lot of Farrell’s examples of friendship seem pretty shallow. There are many sentences like “The dangers of personal intimacy were avoided through abstract intellectualization” or “Tolkien had marital problems throughout the time he participated in the Inklings circle, but as far as we know, he never discussed them with Lewis.”

Within circles, people did do favors for each other: invite each other to dinner, or to stay a few days if they were passing through, or lend money. And the wealthier of the Impressionists supported the others: “When [Monet] was short of money and supplies, which was often, [he] stayed with Bazille.”

But with so many having so little intimacy about things *other* than work, I don’t think friendship is *necessary* for successful circles. Respect and cordiality and good behavior would suffice, what I might characterize as a business friendship rather than a real one.

≤ music ≥

So that’s the trajectory of a collaborative circle. If you’re ambitious and want a big change, I think it’s useful to know that perhaps-boring complaining is part of the process, or to attend specially to trust rather than assuming it, or to pay attention to the seemingly crucial role of pairs who work together and bring their somewhat-worked-out discoveries back to the larger circle.

Next episode, which I think better be the last one, will be about roles. A lot of us rail about how corporate structures want us to be interchangeable, fungible cogs in the machine. And yet we don’t pay attention to the different *informal* roles each of us play, roles Farrell gives names like “Lightning Rod”, “Peacemaker”, and “Conservative or Radical Boundary Marker”. I have this hope that, if you know Farrell’s names for roles, you’ll be more likely to recognize them and so more likely to make good use of them.

Thank you for listening.