Conversation in Cybernetics / Module One of Seven
Module One · Pask & the Founding Theory
Cybernetics usually points outward — to rainforests, flocks, and feedback loops in the living world. This module turns the lens around. Its case is a conversation, and the strange thing you will notice is that to study it, you must have one. The tools you meet here — the entailment mesh, the teachback — are the same tools every other module in this college already uses. They were Gordon Pask's inventions. You have been inside his theory all along.
In a London laboratory, a learner sits at a console. It is not a screen of multiple-choice questions. It is a teaching machine called CASTE — Course Assembly System and Tutorial Environment — and on it is displayed a map: a web of small concepts about probability, with lines showing which idea rests on which.1
The learner does not receive the subject in a fixed order. She chooses a concept to explore, and the machine asks her to do something specific: explain it back — to demonstrate the relationship she claims to understand by rebuilding it. If she can reproduce the idea, and show why it depends on the ideas beneath it, the machine accepts that she knows it and opens the concepts it entails. If she cannot, the conversation goes back a step.
Knowing, on this machine, is never assumed from a correct answer. It is shown by reconstruction. The lesson is a genuine back-and-forth: the learner proposes, the machine probes, and the two reach agreement about what has actually been understood. Pask called the whole exchange a conversation, and he meant the word literally.
Watch what is missing. There is no lecture poured into an empty vessel. There is no test that rewards recognising the right option. There is only a structured exchange in which understanding is continuously made visible, agreed, and built upon. The machine is crude by today's standards; the idea behind it is not. That idea is the subject of this module.
Pask built a formal theory beneath that lesson. Five ideas do the work, and — in the proper Paskian manner — each one entails the next. Read them as a chain, not a list.
Conversation is the unit of learning. Not the individual mind. Learning happens between participants, in the exchange of proposing and probing until meaning is shared. A single brain in isolation is the wrong place to look; the right place is the loop between two of them.2
A P-individual is whoever is doing the conversing. Pask distinguished the biological body — the M-individual, the “mechanical” processor — from the P-individual, the psychological participant that actually holds a point of view. One body can host several P-individuals; two bodies can, in a good conversation, briefly become one. The participant, not the person, is the unit.3
An entailment mesh is the shape of the knowable. A subject is not a line of facts but a web of concepts, each one entailing others. A → B means understanding A is needed for B; A ⇔ B means the two require each other; A | B means they are alternative routes. The map on the CASTE screen was an entailment mesh.4
Teachback proves understanding by reproduction. You have not understood a concept until you can rebuild it — teach it back, derive it, show why it holds. Recognition is cheap; reconstruction is the test. This is why every module in this college ends by asking you to reproduce, not to recognise.5
Serialists and holists travel the mesh differently. Some learners advance step by careful step along one thread (serialist); others leap to the whole picture and fill in the parts afterward (holist). A real conversation lets either route succeed. Pask's evidence: mismatch the strategy to the learner and understanding collapses, even when the content is identical.6
Every device here is a feedback loop. The learner acts, the conversation returns a signal about whether understanding holds, and the next action is shaped by that return — the same circular causality you met in the rainforest and the murmuration, now running between two minds.7 And because the conversation can turn to examine itself — the participants discussing how the learning is going — it is a cybernetics of cybernetics: second-order. The observer is inside the system being observed. This reflexive turn, which Heinz von Foerster named, is exactly what you are doing by reading this page about conversation as a conversation.8
To know a thing is to be able to rebuild it. To know that you know it is to rebuild it in front of someone who can tell you whether you have.
Here is this module's own entailment mesh — the subject matter drawn as the web it really is. The shape is the message: nothing here is a list to memorise; everything entails something else.
The mesh of Module One. The green arc returning from teachback to conversation is the feedback loop — what is proved by rebuilding re-enters and reshapes the exchange.
→ entails · ⇔ mutually entails · | alternatives (dashed)
You have met the five concepts. Pask's claim is that you do not yet know them — you will know them when you can reproduce them for someone else. Choose the route that fits how you have been reading. Both arrive at the same place.
→, ⇔, or |. You have made an entailment mesh.This module is reflexive by design: to study conversation you had to have one — with the page, and now, if you took the teachback seriously, with another person. That is not a trick of presentation. It is the whole claim. Pask held that knowing is not a substance stored in a head but a process kept alive between participants.9
So, this week: notice one exchange — a lesson, a meeting, a disagreement — and ask whether it was a conversation in Pask's sense or merely a transmission. Did anyone have to rebuild an idea to show they held it? Did the two of you reach agreement about what was understood, or only about who was right? The difference is the difference between learning and being told. Carry that question into Module Two, where a centre at Brunel University spent twenty-five years turning it into a practice you can run on yourself.