The force field under the change model¶
Organisations hold their state, and resistance carries information about why they hold it and how. That claim runs through ChangeShop and through the Satir model. Both inherit it from Kurt Lewin, whose version is more specific than the phrase that survived him.
Lewin was a psychologist, trained in the Berlin Gestalt school, out of Germany in 1933, and at work in the United States from then until his death in 1947, at Cornell and then Iowa, and finally at the Research Center for Group Dynamics he founded at MIT.
Several of the sources have no web page in front of them, so those links download a PDF rather than opening a page.
Quasi-stationary equilibrium¶
One of Lewin’s last papers, Frontiers in Group Dynamics, appeared in the first issue of Human Relations in June 1947, four months after his death. It starts from a refusal: group life is never at a steady state. What looks like stability is a level held in place by driving forces pushing towards change and restraining forces holding it back, in rough balance. Lewin called these quasi-stationary equilibria, and the qualifier does the work. A process running at a constant level is not resting. It is being held there.
Two moves can follow. A level shifts when the driving forces increase, and it shifts when the restraining forces are reduced. Both land on the same line.
add driving force remove a restraint
│ │ │ │ │ │ │ │ │
▼ ▼ ▼ ▼ ▼ ▼ ▼ ▼ ▼
L2 ────────────────────── ──────────────────────
▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲
│ │ │ │ │ │ │ │ │
L1 - - - - - - - - - - - - - - - - - - - - - - -
Both work. They do not cost the same. Adding force raises tension across the whole field, and in Lewin’s account the tension comes out somewhere. The frustration experiments he ran with Roger Barker and Tamara Dembo in 1941 found it coming out as regression, children playing below the level they had already reached. The leadership climates his group ran later found aggression looking for a target. Taking a restraint away moves the level at lower tension. Lewin argued for the second, and said so plainly.
That reads onto a security programme without much translation. Exception processes, standing waivers and the workarounds nobody logs are restraining forces with reasons behind them. Mandating harder, escalating, adding a policy line and a sign-off: all of that increases the driving force, moves the level, and raises the tension, which comes back later as the quiet non-compliance the escalation was meant to prevent. Finding out why the waiver exists, and removing that reason, moves the same level without buying the tension.
B = ƒ(P, E)¶
Behaviour (B) is a function of the person (P) and the environment (E) together, Lewin wrote in 1936. E is not the room the behaviour happens in. It is the situation as that person reads it, which Lewin called the lifespace: the facts, people, goals and obstacles they are actually reckoning with at that moment. A control nobody has heard of sits outside it, and does no work.
Awareness training is an intervention on P. Making the safe path the easy path is an intervention on E. The second is usually cheaper and often more durable, and the equation says why: E is not the setting a behaviour happens in, it is half of what produces the behaviour. A policy harder to follow than the workaround has already changed E, in the wrong direction, whatever was covered in the training.
Learning a system by trying to move it¶
Lewin named action research in 1946, writing up work done for the Commission on Community Interrelations, the research arm he had organised for the American Jewish Congress to take on antisemitism and racial discrimination. What he described is cycles of planning, acting, and fact-finding about what the action actually did, run on a live problem rather than a reconstructed one. The knowledge and the intervention arrive together, because a system under a push reveals things about itself that a system at rest keeps quiet.
The same year, running a two-week workshop at New Britain in Connecticut for the state’s Inter-Racial Commission, his team found the part nobody had planned. Observers had been recording the behaviour of the groups for the research staff to go over in the evenings. When participants asked to sit in, and heard themselves described, the feedback turned out to do more work than the training had. The Basic Skills Training Group, later shortened to T-group, is that accident given a name. The National Training Laboratories at Bethel in Maine is the accident turned into an institution, founded in 1947, a few months after Lewin’s death. The participants who brought their own conflicts, the staff working on them and the observers describing what happened are the client, consultants and observers of ChangeShop under other names.
Action research and the T-group both rest on a wartime finding. With meat supply squeezed and the army rejecting recruits on grounds that looked nutritional, the National Research Council’s Committee on Food Habits wanted American households eating the organ meats nobody wanted: kidneys, brains and hearts. Lewin took the assignment because resistance to them ran deep, through physical disgust, social standing and superstition. Anything that shifted them had been tested against the hard case, not the easy one.
He recruited 85 women into six groups and ran nutritionists’ lectures against group discussion ending in a decision, same information in each, then checked what they had actually served at home a week later. Four of the 41 in the lecture groups had served one of the three. Twenty-three of the 44 in the group-decision groups had. Among the fourteen who had never used any of them, four tried one; none of the eleven with the same starting point in the lecture groups did. Persuading individuals to know better did poorly against a group deciding in the open, together, which is roughly the result any awareness programme rediscovers at its own expense.
Unfreezing, moving, freezing¶
The part of Lewin that travelled furthest is a three-step schema: unfreeze, then change, then refreeze. It reaches change-management courses as his model, usually without the rest of him attached, and it gets attributed to Frontiers in Group Dynamics, the 1947 paper. Section 6 of it is headed “Changing as Three Steps: Unfreezing, Moving, and Freezing of Group Standards”, and runs: A successful change includes therefore three aspects: unfreezing (if necessary) the present level L1, moving to the new level L2, and freezing group life on the new level.
Stephen Cummings, Todd Bridgman and Kenneth Brown trace in 2016 how that became foundational to change management, and argue that Lewin never developed such a model: it took shape after his death in the hands of textbook writers.
Bernard Burnes answered in 2020, arguing the three steps were a worked-out approach grounded in field theory rather than a later simplification.
Lewin’s own paper supports Burnes’s account. The three steps are not a stray sentence but a numbered section. Unfreezing is worked through by way of Gordon Allport on catharsis, the argument that a prejudice has to be discharged before it can be reduced, which Lewin turns into the emotional stir-up needed to break the shell of complacency. Allport wrote it for a 1945 issue Lewin edited himself, with Paul Grabbe. The catharsis argument was not something he reached for: he had commissioned it.
Freezing gets a mechanism: an organisational set-up equivalent to a stable circular causal process, an arrangement that feeds its own output back as its own input, so the new level holds without anyone pushing. Lewin returns to the sequence three more times, calls it “the three step procedure”, and reads the group-decision results against all three steps at once.
What did not survive him is the force field underneath. L1 and L2 are levels in one, and freezing is about group standards holding, not about a system coming to rest. Set beside quasi-stationary equilibrium, refreezing cannot mean stability, because nothing in the account is ever stable. It means a new balance of forces holds the new level without continuous effort.
That is a narrower claim than the textbook version makes, and it sits more comfortably alongside the Satir curve, which does not promise a resting state either.
Drawing the field¶
Lewin drew his force fields, using topology to lay out the lifespace as regions with boundaries and forces running across them. The diagrams are claims about a field, made by someone standing in a particular place, which puts them in the same category as an architecture diagram: useful because incomplete, misleading the moment they are read as measurement.
The notation goes further than the account behind it. Topology gives regions, boundaries and paths, none of which have a size. The forces drawn across them are arrows, which can be read as having one. Nothing in the account fixes a scale for a force, so a longer arrow is a judgement about the field rather than a quantity, and the balance in a quasi-stationary equilibrium is asserted rather than computed. The sketch of the two routes to L2 is built the same way: six arrows against three, and nothing anywhere saying what one arrow is worth. Bernard Burnes and Bill Cooke trace the decline of field theory after Lewin’s death partly to this: a formalism that satisfied neither the mathematicians nor the psychologists, and that buried the insight it was built to carry.
Lewin knew. The 1947 paper says the forces “have to be identified and will have to be measured quantitatively”, and calls measuring social forces a task “thus far accomplished only in a limited area of individual psychology”. The diagrams promised a measurement. It never arrived.
Leadership study and sample¶
Lewin, Lippitt and White found in 1939 that groups run democratically do better than autocratic or laissez-faire ones. Their experimentally created social climates were five-member after-school clubs of ten-year-old boys at the Child Welfare Research Station in Iowa, six clubs across two series, each passing through more than one climate under more than one adult leader.
Aggression was the thing being counted. Under autocracy the clubs went one of two ways. Some turned aggressive, and the aggression landed on scapegoats inside the group, not on the leader. Others turned flatly apathetic. The apathy turned out not to be calm: aggression jumped when the autocrat left the room, and again on the day a group moved to a freer climate. That is the tension argument in miniature, and the reason a programme held in place by pressure looks quiet right up until the pressure lifts.
The finding has travelled a long way from that sample and arrives in the leadership literature as a general claim about groups under pressure. It may well hold. The 1939 study is not what establishes it, and a security programme leaning on democratic decision-making has to earn the point on its own evidence. Six clubs of ten-year-olds in one American town is not a sample of the human population, and was never offered as one.
Inheritance¶
Not the three steps. A stalled programme is not sitting still, so the question worth asking is which restraining force could be taken away. A policy harder than the workaround has already lost, whatever the training covered. And a room where people work on their own problem while others describe what they see is Lewin’s method with the labels changed. The three steps came through intact. The field they were cut from did not.
Last updated: 9 August 2026