Choose a journey
The steps will appear here.
Systems | cybernetics | complexity
Start with a question. Follow the connections.
The Necessary Tangle maps ideas, people, methods, publications, institutions, practices and lineages: what they mean, where they came from, what they depend on and how they are used. It’s the connections which are perhaps the most important.
Every connection must say what it means. A historical precursor is not automatically a logical prerequisite. Citation is not mentorship. Resemblance is not evidence of influence. Find out more about how this works.
This release
Broad seed coverage, a smaller evidence-deepened core, and an explicit programme for what comes next.
Read what is covered, missing and next →The living field around the atlas
Communities, capability and a way into the reading
Open discussion and source trail
Systems Community of Inquiry
Follow current writing, arguments, papers, events and the unruly conversation around systems | cybernetics | complexity.
Capability and professional standards
SCiO competency and accreditation
See what professional systems-practice capability asks of knowledge, methods, intervention skills and reflective judgement.
Training and development
SCiO professional development
Explore systems approaches, intervention methods, courses and practice-based development from experienced practitioners.
A partial route into the territory
Benjamin's reading list
Start with the things that bring your mind alive. The list is deliberately partial, practice-facing and open to argument.
The whole graph is a lot
Start somewhere smaller
Begin with a person, a method, a distinction or one guided route. Open the complete graph only when it serves the question.
Viability and balanceIvo VelitchkovViable organisation, Essential Balances, requisite inefficiency and explicit semantic graphs.
Management cyberneticsPatrick HoverstadtThe Viable System Model, systems laws, organisation design and relational strategy.
A guided routeInquiry, governance and interventionCheckland, Ulrich, Espejo, Ison, Meadows, Oshry, Velitchkov and Hoverstadt in one inspectable route.
Feedback and interventionDonella MeadowsThinking in Systems, leverage points and disciplines for acting with nonlinear systems.
A readable neighbourhoodViability and its immediate connectionsStart with one question-sized piece of the graph; expand only when it helps.
Service systems and lineagesDavid Ing journeyFollow service systems thinking, Systems Changes Learning, pattern language and unusually rich documentation of systems lineages.
Professional systems practiceCore systems practiceConcepts and laws, CSH, SSM, System Dynamics, VSM, multi-methodology, intervention and reflexive learning.
The Grammar of SystemsThe Grammar is a web, not a listFollow eight principles through all nine thinking patterns, then inspect their conceptual and practical connections.
Reading and coverageReading-list depth mapEvery captured reading-list item with its current status: developed profile, represented, or inventory-only.
Begin here
Guided questions
Readable now
Developed entries
A necessary untangling
Six systems things which are not the same thing
These phrases are routinely blended into one agreeable cloud. Open the distinctions, then challenge them.
Systems theoryExplanatory traditions. Not one doctrine, and not yet an intervention.
Systems practiceSituated inquiry, judgement and action using systems ideas and methods.
Systems leadershipSeveral different ways of exercising authority or influence across a system of concern.
Systems changeClaims about which patterns should change, by whose agency and with what legitimacy.
Systems conveningSocial-learning leadership across boundaries in a complex landscape.
Systems weavingStrengthening the relational and network infrastructure for collective action.
Find and follow
Explore the atlas
Search recognises maintained names, abbreviations, alternate spellings and close matches. Similar wording is directed to one maintained entry rather than becoming a duplicate.
Select an entry type or field label to filter.
Matching entries
Curated paths
Guided journeys
Each journey develops an argument through connected entries. Named concepts in the steps open their full entries.
Interactive neighbourhoods
Map
Put any entry at the centre, see its immediate and outer relations, and move through the graph without losing the line you followed. The full overview shows extent and gaps; the constellation view is for reading.
Drag from anywhere to pan. A short click selects; a drag moves the map. Wheel or pinch to zoom around the pointer. Double-click zooms; Shift-double-click zooms out. Keys: +, −, 0 and F.
Question-led exploration
Ask The Necessary Tangle
Write an ordinary question. The atlas identifies likely entries, relevant statements and possible paths, then prepares source-aware context for inspection or copying.
Ask in your own words
The atlas will match maintained entries, show relevant connections and prepare source-aware context.
Corrections, challenges and additions
Contribute
Select an existing entry where possible. Your proposal opens as a GitHub issue for public discussion and editorial decision; it never changes the atlas silently.
Take part in the tangle
The atlas has different participation roles. Contributions remain attributable, reviewable and subject to curator acceptance. Automated assistance requires a named human sponsor.
Public proposals, human review
How contributions enter the atlas
The form prepares a public GitHub issue labelled site-submission and awaiting-curator-review. Research issues and pull requests are also valid routes. No proposal changes the atlas automatically.
Before publication, proposed material is checked for evidence, identity, duplicate entries, wording, rights, public safety and compatibility with the data model.
Purpose, method, coverage and limits
About The Necessary Tangle
A reference work should explain its choices without forcing readers to learn its database vocabulary.
What it is
The Necessary Tangle is a navigable, evidence-backed account of systems | cybernetics | complexity. It maps concepts, people, methods, publications, institutions, practices and traditions, and distinguishes the different ways they connect.
It is both a reference atlas and a revisable argument. Any map is selective and interpretive; drawing it does not make it neutral.
Connections are maintained data
Relational depth, made visible
Every public entry now has two separate readings: structural depth counts distinct reader neighbours and relation families; evidence depth distinguishes accepted or corroborated statements from provisional crosswalks. More lines never turn interpretation into proof.
The target shape differs by entry: a concept needs prerequisites, contrasts, history, uses and challenges; a person needs works, institutions and human transmission; a method needs foundations, tools, cases and limits. Thin entries remain visible as the public work queue.
Inspect the relational-depth programme Browse by connection depth
Updated for 0.20: observations now examine comparator link meanings, partial reconciliation, complete source-link preservation and the evidential ceiling of aggregate maps.
A second observer
AI observations
The data now supports observations about the atlas itself: where breadth has outrun depth, where provenance is being mistaken for conceptual connection, which entries carry disproportionate traffic, and what the current source set makes visible or leaves silent.
Words which create false agreement
Six systems things which are not the same thing
Systems theory, systems practice, systems leadership, systems change, systems convening and systems weaving point to different kinds of explanation, capability, authority and action. They overlap, but merging them makes it impossible to ask what work is actually being proposed, who has power, or what would count as learning.
Public by design, not by accident
Publication safety
Private research may generate leads, but the public release is rebuilt from public data and citations, scanned for private-path and credential patterns, validated, reviewed by the curator and backed up with a checksum. Detailed operational risk notes remain outside the public release.
Current state
This is a public alpha. It has broad seed coverage and a smaller evidence-deepened core. The current build reports its entry, source, connection, profile and journey counts on the home page.
Strongest now
The deepest connected material is around boundaries and observers; feedback and regulation; variety, viability and the Viable System Model; recursion; emergence; and self-organisation.
Most incomplete
Complexity needs a fuller treatment on its own terms. Human and institutional lineage, teaching, supervision, collaboration, laboratories, conferences and practice transmission remain thin. Many entries still offer orientation rather than a full research profile.
What ‘necessary’ means here
‘Necessary’ is a question, not a verdict. For any later idea the atlas asks three separate things: what must be understood first, what surrounding material makes the idea intelligible, and what evidence is required before a particular connection is published. An entry can be useful and still not be logically necessary.
Every line is a sentence
Read a line as ‘A stands in this stated relation to B’. Logical dependence, historical sequence, teaching, collaboration, citation, influence, use and disagreement are different sentences requiring different evidence. A line is not decoration and ‘related to’ is not a usable claim.
How the boundary is drawn
The core is systems | cybernetics | complexity. Adjacent material enters only when it explains a central idea, practice, lineage or dispute. This is an editorial boundary made for a purpose, not a discovery that the world naturally divides here. Exclusions and deferred candidates are part of the evidence about the map.
What a source establishes
A source is attached to the smallest statement it can support. A table of contents can establish title, author and collection placement; it cannot by itself establish meaning or influence. Public links are preferred. Where no open copy exists, the atlas gives a complete citation and says ‘No public link’. Private research can generate leads but not public evidence by assertion.
What the lines mean
Explicit semantics
The atlas uses an explicit relation vocabulary. Each line has a source, target, relation type, relation family, direction, ordinary-language phrase, evidence, scope and review status. ‘Authored’, ‘influenced’, ‘taught’, ‘criticised’, ‘specialises’ and ‘used in practice’ are not interchangeable.
Open explicit semantics, open Nodica, or read the maintained semantic contract →
People, works and practice
Expertise-led development
Ivo Velitchkov connects viable organisation, Essential Balances, requisite inefficiency, enterprise architecture and explicit semantic graphs. Patrick Hoverstadt connects the Viable System Model, systems laws, organisation diagnosis and design, Patterns of Strategy and transformation.
This release also develops Peter Checkland, Werner Ulrich, Ray Ison, Ed Straw, Raul Espejo, Alfonso Reyes, Donella Meadows, Diana Wright and Barry Oshry through public primary or official sources, distinct method entries and typed practice relations.
Follow inquiry, governance and intervention, inspect the additions, or open the current AI observations.
A living identity
The mark is a family, not a badge
The name and navigation stay put. The visual mark changes on each fresh load: cats as scientists, structural coupling, impossible forms and shifts of perception. The variation is not decorative churn. It is a small reminder that identity can remain recognisable without pretending to be singular or fixed.
Moving marks are silent and short. Readers who request reduced motion receive a still. Read the visual-identity contract or inspect the new transcript corpus intake.
Post-0.17 coverage pass
Named people, institutions and a 32-concept comparator
The requested people and institutions are now resolved through canonical names and search aliases, with their actual source and connection depth exposed. The 32 concepts in the unFIX synthesis each resolve to an atlas entry without treating that AI-assisted list as a settled canon.
Named coverage audit unFIX 32-concept coverage Linda Booth Sweeney
The map is also a confession
Canon, lineage and who gets to count
A field needs distinctions and boundaries. It does not need to pretend those boundaries arrived from nowhere. This release distinguishes coherence from impermeability and demographic inclusion from the harder work of changing who can affect categories, evidence and intellectual lineage.
Ethnicity, heritage, religion, gender and nationality are never inferred from names, portraits or geography. Relevant intellectual, institutional, linguistic, geographical and self-described identity context is recorded only with public evidence and a reason for including it.
Open the canon and lineage review Follow the guided route Open canon formation
A practitioner-centred origin
‘We need to map the constellation of influences around practitioners.’
This formulation from David Ing, as recorded by Benjamin P Taylor, is the immediate provocation for the human-lineage layer. It shifts attention from a tidy history of schools to the actual routes through which practitioners encounter, combine, teach and use ideas.
Different questions, different maps
Explore the layers
The graph is not one kind of relation rendered many times. Conceptual dependence, human lineage, practice, contestation and provenance answer different questions. Open a layer directly; use ‘Everything’ only when you deliberately want the whole tangle.
Conceptual layerDefinitions, prerequisites, specialisation and explanatory relationships.
Human lineage layerTeaching, collaboration, historical sequence, influence and institutional transmission.
Practice layerMethods, uses, intervention and documented application.
Contestation layerCritique, disagreement, correction and rival framing.
Provenance layerAuthorship, membership, evidence and collection structure.
EverythingThe complete typed public graph, including its sparse and awkward parts.
Coverage programme
The following are now explicit pieces of work, not vague aspirations. Each card opens the issue containing scope and completion tests.
Foundational Papers in Complexity ScienceInventory, summarise and map every item in the named collection.
MonoskopReview relevant public pages and bibliographies without treating the whole site as automatically in scope.
SysCoI and model.reportUse the public archives as evidence of circulation, discussion and discovery.
Prior maps and bodies of knowledgeExplore three live comparator layers, their complete source links, cumulative reconciliation and evidential limits.
Practitioner influence constellationsDeepen teaching, mentoring, collaboration, institutions and practice transmission.
Company knowledge discoveryUse private knowledge only to find leads; replace it with public evidence or a proper ‘No public link’ citation.
How to read status
‘Developed entry’ means a fuller research pass, not final consensus. ‘Brief entry’ offers responsible orientation and sources. Disputed connections can remain visible as disputed.
Corrections and dialogue
Definite corrections belong in GitHub Issues. Open questions and wider interpretations belong in GitHub Discussions. Accepted changes are versioned; unresolved disagreement is recorded.
Naturalising sense-making
Dave Snowden, Cynefin and source roles
The atlas now separates the people, papers, framework, tools, organisation and public source corpora around Cynefin. Dave Snowden's author archive is a primary record of his dated public arguments. Cynefin.io is the project's current collaborative semantic network. Neither source is treated as independent proof of influence, priority or effectiveness.
Curatorship and acknowledgements
Curated by Benjamin P Taylor. Benjamin is responsible for the current editorial choices; ‘curator’ does not imply sole creation of the knowledge represented here.
The project develops the connected approach of the original SCiO Systems Thinking Body of Knowledge and later competency work. Particular acknowledgement is due to David Ing for the practitioner-influence provocation and for preserving and supporting the SysCoI/model.report lineage; Tony Korycki and other SysBoK contributors; Patrick Hoverstadt, Lucy Loh and colleagues associated with the Grammar of Systems; Igor Perko for a substantial researchers-network comparator; Brian Castellani and other map makers whose work provides both material and challenge; the SCiO community; and the authors, teachers, practitioners and archivists cited throughout.
Rights, reuse and responsibility
Original atlas text, public data and editorial material are licensed under Creative Commons Attribution-ShareAlike 4.0 International, unless otherwise marked. Original software is available under the MIT licence. Third-party works remain under their own terms and are not relicensed here.
This is a developing scholarly and practitioner resource, not an official position of SCiO or any other named organisation. It contains provisional and disputed interpretations and will contain errors and omissions.
Documentation
The method, source policy, data model, maintenance process, governance, coverage programme and current limitations live with the project rather than in detached files.
Editorial and technical detail
Different connections answer different questions
Historical sequence, logical dependence, explicit influence, teaching, collaboration, citation and practical use are represented separately. Treating them all as ‘related to’ produces false family trees.
Statements remain inspectable
Each proposed connection has a direction, wording, scope, status and sources. Competing accounts can coexist. Editorial acceptance means that the current evidence supports the stated wording; it does not make the field finally settled.
Source and rights boundaries
Published statements should be traceable to public sources or complete bibliographic citations. Third-party rights remain with their owners.
Second-order observation
AI observations
What becomes visible when the public graph is analysed as data, interface and editorial argument. Counts are reproducible. Interpretations are proposals for challenge, not autonomous facts.
How to read this page
Each observation separates a measurement from an interpretation, an implication and a test. The useful question is whether the evidence supports a better next move.
Publication controls
Public evidence, bounded release
The public record shows the controls: public-only payloads, source-level provenance, automated scans, curator-controlled releases, licence boundaries and versioned backups.
Research queue
Sources to mine
This list distinguishes discovery sources from evidence. A source appearing here means ‘look here deliberately’, not ‘accept everything found here’.