A renewed Ratio Club
Draft proposal, November 2026. This is a draft for discussion: nothing in it is settled.
Proposal in brief
We propose to renew the Ratio Club as a network of small, face-to-face dining clubs in the UK, the USA and anywhere else in the world. Each club has no more than 21 members, meets over dinner, and gives an evening to one or two talks on unfinished work followed by open argument. Membership is by invitation, and rank is left at the door.
The aim is to recover what made the 1949 club productive: people from different disciplines working on live problems, talking as peers. The renewal deliberately changes what the original got wrong for today, above all its all-male, narrowly drawn membership.
The original Ratio Club, 1949–1955
John Bates, a neurologist at the National Hospital for Nervous Diseases, Queen Square, founded the club in September 1949. It followed the publication of Wiener's Cybernetics in 1948. Members met in a basement room at Queen Square. Bates had planned a meal followed by discussion, but in practice a talk usually came first, then a buffet dinner with beer, then extended discussion [1]. The name grew from Donald MacKay's suggestion of Machina Ratiocinatrix, and after the first meeting Albert Uttley proposed the Latin ratio, meaning reason or calculation [1]. The club met regularly until 1955 and held a final reunion in November 1958 [1].
The club had twenty-one members in all [1], and professors were excluded so that rank would not shut down argument. Members included W. Ross Ashby, Grey Walter, Alan Turing, Horace Barlow, Donald MacKay, Albert Uttley, Philip Woodward, Thomas Gold, William Rushton, John Pringle, Eliot Slater and William Hick. Many members brought wartime radar and signal-processing experience to the study of the brain.
Ashby travelled from Barnwood House, Gloucester, and was among its most active members. The club gave him a critical peer audience for the ideas that became Design for a Brain (1952). His journal records his impressions of the meetings from the first onward [1, 5]; Bates's papers at the Wellcome Collection hold the meeting notices and correspondence [4]. The standard secondary account is Husbands and Holland, "The Ratio Club: A Hub of British Cybernetics", in The Mechanical Mind in History (MIT Press, 2008).
Why now
The conditions of 1949 have returned. Then, new computing machines and Wiener's book raised the question of whether machines could think, adapt and regulate. Today, large language models and other AI systems raise the same question with far wider public consequences.
The present debate is dominated by engineering, commercial and policy voices. The cybernetic questions about regulation, variety, adaptation, viability and the place of the observer are largely absent. A renewed club would bring those questions back into face-to-face conversation, across disciplines, without the formality of a conference or the hierarchy of a university.
The club must take on live problems. It is not a heritage society, though the history gives it a standard to meet.
Principles carried forward
The renewal keeps the reasons behind the original rules, not the rules themselves.
Twenty-one members
- Why it worked: everyone speaks; everyone is known.
- Renewed form: no more than 21 members per club.
No professors
- Why it worked: rank did not shut down argument.
- Renewed form: come as yourself, not your title. Senior people are welcome if they leave status at the door.
Talk, then dinner and discussion [1]
- Why it worked: the talk gave substance; the meal and beer loosened the argument.
- Renewed form: one or two short talks, then dinner, with discussion running through and after the meal.
A shared provocation (Wiener, 1948)
- Why it worked: a common question across disciplines.
- Renewed form: machines and minds today: AI, ecology, viability, regulation.
Members built things
- Why it worked: homeostats, tortoises and radar kept talk honest.
- Renewed form: bring something that works, or that fails interestingly.
Regular attendance
- Why it worked: continuity built a shared language.
- Renewed form: members commit to attending most meetings.
Form of a meeting
A meeting runs for one evening, perhaps every two or three months. Talks present problems, not finished results. A good talk ends with an open question the room can work on. Disagreement is expected and need not be resolved on the night.
1. Talks
Drinks on arrival, then one or two talks of 15 to 20 minutes each, on work in progress. A demonstration, model or prototype is encouraged.
2. Dinner
Dinner together, unhurried.
3. Discussion
Open discussion through and after dinner, which should take most of the evening. This follows the original club's practice [1].
4. The record
The keeper of the record notes what was said and what remains unresolved.
Membership and invitation
Membership is by invitation, because invitation protects the quality of the conversation. To stop the club closing in on itself, each member may bring one guest a year, and guests may be invited to join. There are no formal qualifications. Members are invited for qualities.
Active work on a real problem
In research, practice, art, clinical work, engineering, design or community work.
Willingness to be challenged
To present half-formed ideas and to have them challenged.
A different home discipline
Different from most others in the room.
Generosity and curiosity
In discussion.
Commitment to attend
A new member is proposed by one member and seconded by another.
A founding group chosen with care
A club's first members shape everyone it invites later.
Inclusion by design
The original club was entirely male and drawn largely from Cambridge, London and wartime science. Copied literally, the design would reproduce that. The renewal addresses it deliberately, not by hoping it sorts itself out. The tension between a small, chosen circle and an open one cannot be fully resolved; the guest rule and a varied founding group are the main safeguards.
Gender balance and age range
From the founding group of each club onward.
Practitioners beyond the academy
Artists, designers, clinicians, engineers, organisers.
Cost
Keep dinners affordable, and consider a sliding contribution or sponsorship so price does not decide membership.
Time and care
A long weekday evening excludes carers. Vary timing, and consider an occasional lunch or weekend meeting.
Venue and drink
Accessible venues; alcohol available but never central.
Geography
Clubs in many countries, and within each country well beyond the obvious centres (in the UK, beyond London and Oxbridge).
Many clubs, one charter
The clubs share a one-page charter and otherwise govern themselves. Growth comes from new clubs, never from larger ones: when a club wants more than 21 members, it seeds another.
Shared across all clubs
- The charter's principles, as set out in this proposal.
- The name, with a local identifier (for example, Ratio Club Bristol or Ratio Club Boston).
- An open record of meetings.
- Cross-visiting: a member of one club may attend another as a guest.
- A yearly gathering by country or region, and an occasional international gathering.
Decided locally
- Membership.
- Meeting frequency and venue.
- Language.
- Topics and format details.
Coordination
A small coordinating group, perhaps one person from each club, keeps the charter and arranges the gatherings. It has no authority over local clubs.
The record
Each club appoints a keeper of the record. We know about the original club largely because Ashby kept a journal; the renewal should not depend on luck.
The record notes who attended, what was presented, the main lines of argument and the questions left open. It is written up within a week and shared among the clubs. With members' consent, records could be archived alongside the Ashby materials at ashby.info, linking the renewed club to its predecessor's primary sources.
Recording should not inhibit frank talk. Each club agrees what is attributed by name and what is reported without attribution.
Appendix: Ashby's questions for the Ratio Club, 1950
Ashby circulated a first list of discussion topics on 18 February 1950 and a refined list of 28 questions dated 15 May 1950. The questions below summarise the refined list, grouped by theme, with Ashby's numbering in brackets. His exact wording is reproduced in Husbands and Holland [1], from the John Westcott papers. A renewed club could draw up its own founding list in the same spirit: questions open enough to bring different disciplines to the table, and some deliberately outrageous.
Brains as machines
- Machines defined only statistically, and how far that knowledge applies to the brain [1].
- Whether noise plays a part in brain function [2].
- Whether abnormalities of brains and machines can be described in common terms [3].
- The brain's indifference to the exact location of some processes, and physical parallels [4].
Memory, prediction and learning
- What present-day machine memories imply for the brain [5].
- How far machines have imitated the conditioned reflex, and what they still miss [15].
- How to design a machine that, like the brain, works out its own formulae for prediction [16].
Information theory
- Sense organs, optical illusions and delays in transmission seen through information theory [6–8].
- Information theory applied to insect communities [21] and to genetic control across generations [26].
Shared definitions and the observer
- Definitions of feedback, stability and servomechanism that apply to all systems [9].
- What the physiologist and the physicist share, each observing a system only partly accessible, and only by interfering with it [10–11].
- Whether mind is a physical unobservable, and what follows [12].
Mechanism
- The thermodynamics of open systems applied to the brain [13].
- How far present-day machines can imitate life [14].
- Which brain processes are analogue and which digital [17].
- How a many-part machine acts quickly or slowly without losing coordination [18].
- General methods of making systems stable, and their uses in physiology [20].
- Discontinuous servomechanisms and the brain [22].
Society and economics
- Steady states in economic systems [19].
- Reorganising the Civil Service on cybernetic lines [23].
- Economic vicious circles, and reorganising the economic system cybernetically [24–25].
The speculative edge
- Whether members could agree a conclusion on extra-sensory perception [27].
- A machine whose time is a complex variable, and any bearing on obscure properties of the brain [28].
What became of them
The first list ended with a topic dropped from the refined one: the effect of alcohol on control and communication, with practical work. Several questions became meetings. Pringle spoke on noise (1950), Grey Walter on adaptive behaviour (1951), and Bates and Good on extra-sensory perception (1951, 1953). Barlow recalled long arguments over analogue and digital processes.
References
[1] Husbands, P. and Holland, O. (2008) "The Ratio Club: A Hub of British Cybernetics", in P. Husbands, O. Holland and M. Wheeler (eds), The Mechanical Mind in History, MIT Press, pp. 91–148. A copy is on ashby.info.
[2] Holland, O. and Husbands, P. (2011) "The origins of British cybernetics: the Ratio Club", Kybernetes. doi:10.1108/03684921111117951
[3] Husbands, P. and Holland, O. (2012) "Warren McCulloch and the British cyberneticians", Interdisciplinary Science Reviews 37(3), pp. 237–253.
[4] Wellcome Collection, Bates, John A.V., and the Ratio Club, GC/179.
[5] Ashby, W.R., Journal, W. Ross Ashby digital archive, ashby.info.
Interested?
If you would like to hear when the first club forms, register your interest on the home page.