Inside the Rhizome: Four Social Groupings, and How Each Should Count

Pedagogy,Pedagogy, Andragogy, Heutagogy, Rhizome maps Dave Cormier's rhizomatic learning onto Personal Learning Maturity's top level as "Level 5, extended socially" — self-adaptation that's no longer a solitary act, negotiated instead with a network of other people in a field where the right answer keeps moving. That page leaves the network itself as a single undifferentiated mass. It isn't one. Someone learning the same topic as you, someone actively researching it, someone with a stake in the outcome but no research role in it, and someone who has never heard of it at all are all, in principle, part of Cormier's "community" — and they don't carry the same weight, offer the same kind of help, or stay in the picture for the same length of time.

Four groupings are worth separating out: co-learners (others learning the same topic, at a similar or nearby stage), researchers (those actively producing new knowledge about it), interested stakeholders (those with a direct stake in the topic who aren't producing new knowledge about it — an employer, a professional body, a project supervisor, a downstream user), and the largely unaware (everyone else, who has no particular engagement with the topic at all). Each differs along three dimensions: how much its input should count in a trust calculation, what kind of support or critique it's actually capable of offering, and how long it sticks around. Cormier's own claim is only that "the community is the curriculum" — he doesn't break that community into differently-weighted parts. The four-way split below, and pairing each part against a specific trust, support, and turnover literature, is this page's own extension of his argument, not a synthesis any of the sources cited below made themselves.

Worth stating up front: these are positions in a network, not fixed identities. The same person can occupy more than one of them depending on the topic — a lecturer is a researcher on their own specialism and largely unaware of a colleague's; a final-year student is a co-learner to their cohort and, on the one narrow question their dissertation happens to answer, briefly closer to a researcher than most people supervising them. Rhizomatic learning's whole point is that no node holds a permanently privileged position — that applies to which grouping a given person is in just as much as it applies to which person holds the answer.

Co-Learners: The Nearest Ring

AA co-learner is anyone else working on roughly the same material, whether that's a course cohort, a study group, or a stranger asking the same question on a forum a week before you did. Their expertise is close to yours by definition, which is exactly what makes them useful in a specific way a genuine expert usually isn't: a co-learner slightly ahead of you is operating inside what Vygotsky called the zone of proximal development — the gap between what you can do alone and what you can do with the right support close by, rather than a fully worked-out answer from someone too far ahead to remember what the confusion actually felt like¹. Scaffolding, the ZPD, and Using GenAI Well covers the same mechanism applied to a literature or an AI assistant; a co-learner is the original, human version of it.

Trust here is mostly self-calibrating: you can usually check a co-learner's reasoning directly, because it's pitched at a level you're equipped to follow, and they have little reason to misrepresent it — nobody fakes confidence about a topic they're visibly still learning. What they offer isn't authority, it's company and pace: a study group, a shared confusion turned into a shared question, sometimes a structured arrangement (a formal study group with roles and a schedule) and sometimes entirely free-form (a thread that happens to be active this week). Their availability is usually highest of all four groupings while a course or project is actually running, and drops sharply once it ends — a cohort is a social structure with a fixed lifespan, renewed every year with different people occupying the same role, per the site's three-tier account of knowledge half-life's own point that the surface layer of a fast-moving field turns over constantly while the underlying role persists.

Researchers: Where the Field's Current Answer Actually Lives

ResearchersResearchers are the people actually producing new knowledge about the topic, not just applying or teaching an existing version of it. Patrick Wilson's concept of cognitive authority — the standing a source has earned to be believed on a given topic, distinct from formal social authority or simple confidence — is built almost exactly for this grouping: a researcher's competence on their own specialism is the clearest case of legitimately deferring to someone else's judgement rather than verifying everything yourself from scratch². That deference needs a second check the cognitive-authority idea doesn't supply on its own: Sperber and colleagues' account of epistemic vigilance separates a source's competence from its benevolence — whether it's actually trying to give you an accurate answer rather than a self-serving one³. A researcher scores high on competence almost by definition; that says nothing about whether a specific claim of theirs is a settled result or a stake being planted in an ongoing argument they have a career interest in winning.

The practical move is to apply the half-life series' foundational/applied/ephemeral split to the person, not just the fact: the same researcher's word on a settled, decades-old foundation deserves far more trust than their word on this month's still-contested finding, even though it's the identical person speaking with the identical confidence. Their support style is usually the most structured of any grouping when it's mediated by peer review, and the most free-form when it isn't — a conference hallway conversation, a blog post, a reply on social media. And the grouping itself turns over at the field's own pace: in a slow field, "the current researchers" is a stable list for decades; in a fast one, whoever counts as the current expert on the newest sub-topic can change within a year, which is the social-network mirror of the ephemeral tier's own turnover, not a separate phenomenon from it.

Interested Stakeholders: Expertise With a Stake in the Answer

AnAn interested stakeholder has real, often substantial, domain knowledge and a direct interest in a particular answer, without being in the business of producing new knowledge about it — an employer describing what "industry actually wants," a professional accreditation body, a project supervisor, a vendor describing their own product. This is where the benevolence half of Sperber and colleagues' epistemic vigilance does the most work of any grouping: a stakeholder's competence can be entirely genuine while their account is still shaped by what answer benefits them, and the two need to be weighed separately rather than treating a confident, well-informed answer as an impartial one by default.

Their critique tends to be the most structured of all four groupings — a performance review, an accreditation panel, a supervisor's formal feedback on a project brief — precisely because institutions, unlike individuals, mostly interact through designed processes rather than ad hoc conversation. That same institutional character gives stakeholders an availability pattern the other three groupings don't share: they're usually the most durably persistent presence of the four, because an institution's own update cycle is generally far slower than the pace of the field it's overseeing. The half-life page makes this point about a syllabus fixed years before it's taught; the same lag applies to the accrediting body, the employer's stated requirements, and the professional standard sitting behind them, which is exactly why a stakeholder's account of "what's currently needed" is itself a claim with its own half-life, not a fixed ground truth to check other groupings against.

The Largely Unaware: No Direct Expertise, and a Genuine Use Anyway

TheThe fourth grouping is everyone with no particular engagement with the topic — not a deficient version of the other three, but a distinct position with its own value. On Sperber and colleagues' benevolence axis, this grouping is close to ideal: someone with no stake in the topic has no incentive to shade an answer toward a preferred outcome, because they don't have one. What they lack in competence they partly make up for by supplying something the other three groupings structurally can't: a reaction uncontaminated by the field's own assumptions — the same mechanism, at a small scale, behind Mark Granovetter's finding that a distant, infrequent contact is more likely than a close one to bring genuinely new information, precisely because a close contact tends to already know what you know⁴. Learning by Teaching, Learning by Building already makes the related point that explaining something to someone unfamiliar with it exposes gaps a knowledgeable audience lets pass unquestioned — the largely-unaware grouping is the live, external version of that same check, not a rehearsal for it.

Their "critique" is almost never a designed relationship the way a co-learner's study group or a stakeholder's review process is — it's incidental, in the form of a confused question, a blank look, or silence where engagement was expected. That confusion is a genuine signal about the explanation's clarity, not about the correctness of the underlying claim, and the two shouldn't be conflated. A separate tradition arrived at the same value from a completely different direction and made it deliberate rather than incidental: Synectics, William Gordon and George Prince's formalised creative-problem-solving method, builds a role for exactly this into its working group on purpose, on the basis that someone with no stake in the problem and no prior attempt at it brings a genuinely different kind of contribution than another expert would⁵ — Synectics training material describes this deliberately-included role as the "Naive Resource." Temporally, this grouping is the most stable of the four by a wide margin, for a specific reason: because it isn't tracking the field at all, it doesn't decay at the field's pace the way the researcher and co-learner groupings do. The most epistemically distant grouping is, ironically, also the most reliably available one.

How a Leaderless Network Still Produces Shared Terms

Cormier'sCormier's "the community is the curriculum" claims that a field's network, not a syllabus, decides what's worth knowing — but it doesn't say how a network with no central authority converges on anything stable enough to call a curriculum at all. Etienne Wenger's account of a community of practice gives the mechanism: meaning gets negotiated through two intertwined processes, participation (the lived, ongoing business of doing the thing together) and reification — giving form to that shared experience by producing objects, terms, and concepts that congeal it into something treated as a fixed "thing," even though nobody formally ratified it⁶. A field's jargon, its shorthand, the name everyone now uses for a pattern that used to need a paragraph to describe — all reification, produced by participation, with no single author.

Thomas Vander Wal's folksonomy — a bottom-up categorical structure built from many people's independent tagging, with no central taxonomist deciding the categories in advance — is the same process observed in tagging systems specifically: enough independent, uncoordinated acts of labelling something start to function as if they were a designed classification scheme⁷. Combining the two ideas directly is the move this site's own author made in an earlier piece of work, Folksonomological Reification, co-written with Karsten Lundqvist, Edwin Porter-Daniels, Richard Hussey, and Shirley Williams — applying Wenger's reification specifically to folksonomic tagging on social and learning tools, and adapting activity theory's subject-tool-object-community-rules-division-of-labour model to fit it⁸. That piece was never put through formal peer review — it was released instead for open, crowd-based review through a personal learning network, which is worth weighing the way this page's own trust calculus asks anywhere else: real domain grounding, without the specific check a formal editorial gate would have added.

Put together, this is the mechanism behind the four groupings above jointly producing a working vocabulary without anyone holding the pen: a co-learner's shorthand, a researcher's careful term, a stakeholder's institutional phrase, and an outsider's naive rephrasing all feed the same reification process, and whichever version gets repeated enough by enough of the network is the one that sticks — not because it was voted on or centrally defined, but because reification, per Wenger, is what participation does over time regardless of who's doing it. Reading Wenger's and Parslow et al.'s ideas together this way, as the mechanism specifically behind how the four groupings above jointly negotiate a shared vocabulary, is this page's own connection — neither source makes that link themselves.

Four Groups, Side by Side

Building This Deliberately

TheThe practical use of this isn't a taxonomy for its own sake — it's a check against the much simpler thing people actually do by default, which is weight input by how confidently it's delivered rather than by which of the four positions it's coming from.

• Seek out a co-learner slightly ahead of you, not only people at your exact level. The ZPD argument specifically needs someone close enough to remember the confusion but far enough ahead to have resolved it — a study partner at exactly your own stage can validate that you're both stuck, but can't supply the next step.
• Separate a researcher's settled claims from their live ones, even within the same conversation. The same person's word on an established foundation and on this month's still-contested finding warrant different amounts of trust, regardless of how similarly confident both are delivered.
• Before accepting a stakeholder's critique at face value, ask what answer would serve their interest. This isn't a reason to dismiss it — a stakeholder's competence is usually real — it's a reason to weight it against what they'd want to hear regardless of the facts.
• Use an unaware reader on purpose, not by accident. Before finalising something meant to be broadly understood, put it in front of someone with no background in the topic and treat their confusion as a finding about the explanation, not about them.

Related Topics

• Pedagogy, Andragogy, Heutagogy, Rhizome — the rhizomatic-learning row this page unpacks into four distinct social groupings.
• Personal Learning Maturity — the Level 5 maturity this page's "extended socially" territory builds on.
• Scaffolding, the ZPD, and Using GenAI Well — the zone of proximal development mechanism behind the co-learners section, applied there to literature and AI assistance rather than people.
• The Half-Life of Knowledge — the foundational/applied/ephemeral split this page applies to people and institutions, not just facts.
• Learning by Teaching, Learning by Building — the related mechanism behind why an unaware audience exposes gaps a knowledgeable one lets pass.

References