Sociocracy meets formal semantics

This article is not really meant for the public. But at the slight chance that maybe 5 people out there might find this interesting, I am publishing it.

The way my mind works is that things need to make sense. If they don’t make sense, it haunts me. I have always felt a bit like that when it came to the difference between operational decisions and policy decisions. Operational decisions are decisions that can be made by any circle member, within policy, as part of their ordinary work. Policy decisions are more generic guidelines on how we do things.

I have taught this difference in sociocracy trainings. But it never really rang true to myself. I admit, it was more on an intuitive level. There are some operational decisions that just smell more like policy. And some just feel like one could argue either way. For instance, selecting people into roles: if we select a facilitator, we do this by consent. Sure, this is a generic decision with impact on how we govern ourselves. Makes sense. If the facilitator is absent for a meeting, if there is no policy around it, we often use a pragmatic approach: the leader appoints someone to facilitate that one meeting. Some groups also make this a consent decision. And let’s imagine the facilitator is absent and we already know that member will be absent for two meetings in a row due to business travels. Now what? Do we make a selection by consent, or an executive decision (for instance by the circle leader)?

This seems like semantics but the difference is huge when you consider that the method of decision-making changes. Policy decisions require consent. Operational decisions don’t. It seems like it would be helpful to know the difference between operations and policy. And to me, there was just too much gray zone there.

So I started using a tool I know well from a previous life, the life of a linguist, on the whole thing. What we do as linguists is to drill down until we find the most basic building blocks and see whether we can understand how the interaction between those building blocks follows patterns.

Operations are situations in the life of the organization

Typically, we start an organization because we want to do something. We set ourselves an aim (something specific like “producing and selling vegan chocolate”), and then we do things that we think serve the aim. I will call those things we do situations from situation semantics. (Another tool would have been event semantics which is very similar.) Situations typically contain individuals and the thing they do, for instance Mark stirring soy milk. Situations are nested and make up our world. There are small events like the stirring milk situation, but also bigger situations like making vegan chocolate, or the production circle having a meeting. Or the webmaster designing the website. (A small situation here would be the webmaster selecting a background color in a dropdown box which is just one small situation out of the design process.)

The aim of the circle serves as a container that holds all the activities of a circle. Let’s assume for simplicity now (we will refine that later) that it is pretty obvious what situation is part of the set of situations that serve the aim of the organization. All — factual or possible — situations that are happening in support of the aim for the aim. For instance, any chocolate making situation is in support of making and selling vegan chocolate.

If we have all the situations of an organization in our box and all of them serve the aim, things are easy. That’s the situation of 4 friends in a very young start-up, everyone just doing things that support the business. But soon they organization will want to differentiate. We split our one big circle into circles that have as their area of responsibility (=domain) a subdomain of the organization. For instance, we might have production and sales. We will now have to define the aims and domains of each circle, making sure there are no gaps between the domains and no overlap. (If there are gaps or overlaps, tension between the circles is waiting just down the line.) Every circle also comes with an overhead, the process level. That’s where we make decisions on how we hold meetings etc.

Within our aim and domain, we are free to act. The sales circle just does what is needed in that domain. Remember that in sociocracy (just like holacracy), circles are semi-autonomus, i.e. they have authority over their domain. For simplicity, I will assume that all the activity in the circles is supposed to serve the aim. (The system might take some noise in addition but let’s ignore that for now. Noise would be someone goofing or making a side comment in a meeting.) I will also assume that we know that immoral or illegal action is prevented by other factors — it is not part of the set of allowed situations. If we accept that, it is safe to say that within the circle’s domain, all actions are allowed that serve the aim. We now know what operations are. Operations are all the situations (or events) that are within the domain of the circle. Some of those situations will be actual situations (the situation of Mike stirring the milk at 5.45am in the kitchen by the sink), but there are also possible situations, things you would have been allowed to do. If you’re familiar with possible world semantics, that’s basically what it is. All the potential situations where Mike is stirring milk, or even all potential situations where someone else is stirring milk, are part of production circle. Mike could have stirred soy milk at 6am but he did not. Still, this is a situation that is a situation that is within the circle’s aim because it would have been allowed. Sidenote: in order to understand policy, we have to have a sense of situations that would or would not be allowed — we use possible situations to imagine a situation that we want to prevent. That is why possible situations and not only actual situations are part of the picture here. If this sounds complex to you, just remember that we all have a pretty clear sense that a milk stirring event would belong into production circle, so your mind seems to have no issue with possible situations.

What is policy?

Thanks for hanging in there! Now it gets interesting, and we’ll also pick up speed.

If we conceptualize that all operations are concrete or potential situations in a circle, then what is policy? On a tangible level, policy is what reduces or limits the amount of possible situations within a circle. For instance, let’s imagine a non-vegan chocolate factory now turns vegan. Before, we had all kinds of milk stirring events. If we turn vegan, all the situations of cow milk stirring events are now banned by policy, and all the vegan milk stirring events are still happening. The good thing about situation semantics is that it allows you to cut and slice anywhere that is conceivable. Everything that your mind perceives as a situation (small or big), is a situation. (Since linguistics aims to capture how the human mind processes language, that makes perfect sense.) For instance, if we decide to only produce chocolate in heart shape, then we simply rule out all potential events where the chocolate has bunny shape or whatever else. Situation semantics operates on a concept of ‘good enough for now’; if the heart is more or less a heart, we’ll count it until we have a reason not to. The other good thing about using situation semantics and elimination of situations is that we do not have to define what it is the circle can potentially do. We do not have to codify that choosing the background color is entailed by designing the website within sales circle because (1) everything that serves the aim of selling chocolate is allowed within the circle, and (2) situation semantics delivers us the common sense aspect. We have an idea, a working hypothesis, of what a webmaster does and how it relates to sales, and that’s good enough for a start.

Only make policy if you have to

I like to think that policy only needs to be made if it is necessary. Ideally, we would just read each others mind, have the same needs and the same style of doing things, and we’d flow together like a school of fish. Of course, that’s not the reality, but it is a wonderful place to start from because it is starting from trust that the other members of the organization have some shared reality with me.

Policy is what we make when we realize we need clarity. When something is not working. Typically, that is when we realize that our actions interfer with someone else’s needs, or when what we do turns out not to be serving the circle’s aim, despite good intentions. For instance, the webmaster might forget that a website design relying on the the difference between green and red is not smart design considering color blind users. Or when we realize that using sexist commercials make our potential customers more angry than eager to buy vegan chocolate hearts. In the first example, the color blind example, a simply way of giving feedback might suffice. I, personally, always err on the lazy side when it comes to making policy. I avoid making policy if the problem can be fixed by giving feedback to the designer. Too many organizations make policy because they are afraid to give honest feedback. “Hey, webmaster. Green and red, what about color blind people?” If things like that keep happening, what we might do is make policy is that the circle has to approve the website before it goes up. I would not want to make policy around banning all designs that rely on the difference between red and green, that’s not sociocracy, thats bureaucracy. Accordingly for the sexist commercials. Feedback might suffice but policy might help to prevent similar situations. Let’s assume we make policy that commercials have to be non-discrimatory.

Semantically, what we have done is reducing all possible situations of making commercials to only the non-discrimatory ones. The freedom in sales circle is now limited by policy. Strategy works in a similar way, when we decide that we only want to produce commercials that caters to women between 30 and 35 with children that are vegan. Policy is self-limitation by choice of the circle because it either improves how we work together or it improves how we reach our aim. Some situations are just not contributing to the aim in the way we thought. So we rule them out.

Back to semantics. Let’s say one of our employees keeps doing cartwheels in the hallway. It would never occur to us to make policy around gymnastics at the workplace, we have given the feedback, but they keep flipping down the hall, so here we go, now we are forced to make policy.

Using the universal quantifier, we can now say, for all gymnastics in the hallway situations, they are against policy.

∀s (s is a gymnastics situation & in the hallway) → not allowed.

In this case, gymnastics moves in your own office are allowed, just the hallway is banned. We can make this very complex by adding more and more conditions, for instance only banning gymnastics in the hallway between 9 and 5. This is quantification over potential situations, situations we expect might happen in the future.

∀s (s is a gymnastics situation & in the hallway & between 9 and 5) → not allowed.

Again, all policy decisions are to be made by consent of the circle that has this decision in its domain. Sales circle makes policy about their domain, production circle makes decisions about their domain. (Are you wondering right now who would have gymnastics in the hallway in their domain? Yes, me too.) The point is that if we make decisions that apply to all situations of a certain kind, this is a decision that gives direction to the circle or the organization, i.e. the circle decides together. Whenever there is quantification over situations, it is policy.

Back to operations

Operations are all of those situations, remember? Mike deciding to stir the soy milk after his lunch break, as he knows it is not limited by policy and it serves the aim. Let’s say there has been consent earlier on not using cow milk. The policy has been established by consent. Operations means to carry out situations that remain in the set of allowed/possible potential situations. (I think by carrying out policy and doing something, we just actualize a potential situation which means to anchor it in the actual world at a specific time.)

Operations and policy on process level

Besides the content of the circle, any circle also has a process level in their domain which guides how they self-govern. The facilitator faciliating agenda item A is an operation in that domain, or the secretary typing up the notes. What’s policy in that domain? For instance the policy that minutes have to be sent out within 48h after the end of the meeting. The quantification idea applies again:

∀s (s is a note-taking situation & s is ≤ 48h after the end of the meeting)

Or, stated in the negative (elimination of undesired situations from the set of possible situations in the life of the circle):

∀s (s is a note-taking situation & s is > 48h after the end of the meeting ) → not allowed

Or just having a facilitator is a policy decision because we’re making a statement that all meeting situations have to be facilitated meeting situations. Please note that the way I formalized the relationship between operations and policy, not having one is declared undesireable, but, when we look at it from this angle, it seems to me that in ordinary organizations, we make a wild mix between positive and negative statements, sometimes declaring things desireable — strategy — , sometimes forbidding things — policy. Obviously, declaring that all situations must be having-a-facilitator-situations just means to exclude the ones where that is not true.

∀s in the circle (s is a meeting situation & s has a facilitator) → not allowed.

This operator now eliminates potential situations where there is no facilitator from the set of potential situations.

Set theory

How big is the set?

The difference between policy and operations is that operations are situations. And policy is quantification over situations. That’s it!

Why is the difference between policy and operations sometimes so obvious and sometimes to muddy? The difference lies in the size of the set.

Let’s say an outreach circle of a young volunteer organization plans to approach a major donor. They have never worked with major donors and they don’t really know how to navigate it. What can’t we do? For instance, we might decide to make policy that when someone has already approached the potential donor, no one else from the circle should approach them again. Therefore, we can make a policy decision using the universal quantifier although we know that the number of relevant situations is just a set of one. We simply look at it from a more generic angle.

s (s is an approaching major donors event) → no one else but s approach the donor

Why would a circle choose to do that? Why do we need policy/quantification when we assume that there will only be exactly one actual situation that will be an instantiation of this kind of situation? Isn’t that using a chain saw to cut butter? Yes, it is. Probably, we would not make policy here. We might still do it because although in the circle’s domain, this kind of situation might be unchartered territory in which case it makes sense to get clarity as a group, hearing more voices about it and making sure everyone is on board. If approaching major donors was a regular part of their routine, they might just treat it as an operations (and there would most likely be some policy formed around it).

At the same time, the guidelines of approaching new donors is also a good example of how policy and strategy can blend into each other (that’s why strategy is also referred to as strategic policy). Again, in the framework I am presenting here, policy rules out potential activities. The policy says that if a potential donor has been approached by someone else, we do not go an try a second time. Let us talk about strategy, and then let us come back to this example.

What is strategy?

What is strategy? Strategies are this-over-that decisions. For instance, as a social enterprise, we might decide to focus on online trainings instead of in-person trainings. Whatever the intention might be (reducing our carbon footprint, or spending less time away from family), we say online trainings go before in-person trainings. Another strategy would be to do whatever pays best. The difference between policy and strategy is that policy rules out possible situations while strategy weighs them. We can express this using a degree variable.

λd (online trainings are d-desirable) > λd (in-person trainings are d-desirable)

More correctly, we are not comparing trainings but we are comparing situations that are training situations

(λs (s is a training situation))

Which, or how many situations are we comparing? We are making a statement about all situations, saying that it is generally true that online trainings are more desirable than in-person trainings.

s [λd (online training situations s are d-desirable) > λd (in-person training situations s are d-desirable)]

A strategy like “by the end of the year we need to have 10,000 followers on twitter” is a statement where we say that any world in which we have more than 10,000 followers by the end of the year is more desirable than worlds where we don’t.

Why are strategy and policy so similar?

If we assume (which might be falling short) that strategy is a this-over-that decision and we define which situations are the most desireable then in an extreme case this can mean to rule out all the other options. If the difference in the degree of desirability is big enough, it might get close to “in 99% of the cases to A over B” which basically means ruling out B. (Plus, what we are not touching on at all here is what follows from a violation of policy.)

I am sure there is more to say about strategy but the point for here is that strategic policy decisions are quantifications (they have a universal ∀-operator in front of them) which means we are making a generic statement about a certain set of situations. Therefore, we need consent to make those decisions. All our operational decisions should be consistent with our governance policies and guided by our strategic policies.

What are (s)elections?

I want to extend this idea of using situation theory for governance models to selection processes. Selection processes in sociocracy are most typically used to elect people into roles, for instance a circle member into the role of the facilitator. (It can also be used for any kind of selection, however. For instance, one can run a selection process to assign 5 people into different rooms or to select a city to go on vaction to.) To keep it simple, let’s say we talk about elections of people, and we only select individuals from within the circle. So we already know what the domain (in the mathematic meaning) is in a circle with circle members A, B, C and D.

Domain(R_facilitator) = {A, B, C, D}

What roles to we want to select them into? Let’s say, one of them has to take care of the office fridge, one of them has to take care of the coffee machine, and one of them takes care of trash in the shared office. Each of those roles stands for a set of situations (all the possible situations that would be within the policy set for that role). The circle would have to have a common understanding of what it means to be the fridge czar. Let’s say it includes all the situations of throwing out food that’s moldy and all the situations of restocking the fridge with milk.

s (s is a situation of throwing out moldy food ∨ s is a situation of restocking milk) → fridge czar is in charge of s

This gives us a set of both throwing out situations and restocking situations. Fridge czar refers to whoever gets selected into that role. There might be different ways to conceptualize this but I’ll go an easy way of just forming a subset of those fridge-related situations. If we select Mike into that role, then we are reducing the number of fridge-czar situations to the ones where it is Mike who throws out food and restocks milk.

s∈D: (s is a situation of a throwing out moldy food ∨ s is a situation of restocking milk) ^ a is agent of s

What we do in a selection process is that we set the variable by setting the value for the agent a. Since this is within quantification, we are dealing with policy; hence the decision of who will fill that role has to be made by consent.

What was this good for?

Yeah, to be honest I don’t know. I think the biggest advantage of this article is to explain what policy is and what operations are which then clarifies what decisions require consent and which don’t. It also clarifies the relationship between aim (purpose) and policy — policy restricting further the set of possible situations described by the aim.




Sociocracy, Non-Violent Communication, Linguistics

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

USB3.0 Portable Monitor For Linux

Using Generics For Mapping API Response Swift

Create Lifecycle Rules to Manage Your S3 Files

Not checking your code can cost millions of dollars

Quick remote FPGA lab setup without VPN, port-forwarding, firewall configurations and other…

A Day in the Life of our former Engineering Interns, also know as our new Insider Engineers!

How To Implement Form Validation In Rails Application

Sylius — Product import

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Ted Rau

Ted Rau

Sociocracy, Non-Violent Communication, Linguistics

More from Medium

The Much Misinterpreted Graph

The Return of Pristine Seas and Life-Filled Oceans

Local Journalism, a Link to Good Governance

War and economic sanctions may set global sustainable development back for years to come