What's valuable. A 1914 book that states, in plain language, several mechanisms most modern management writing only gestures at — and states them better. Her passage on how the choice of measure decides the behaviour is sharper than most things written on the subject since. It is also the cheapest way to see where the modern vocabulary of standards, records, incentives and training actually came from, and how little of it is new.
What you can take from it. Five things, all of which transfer out of a 1914 workshop without modification:
How it scores.
✓ HOLDS UP a real strength ▲ CARE works, with a catch ✕ FAILS doesn't stand up i CONTEXT neither good nor bad
| ▲ CARE | Is any of it proven? | The methods are untested, not disproved — nobody has ever run a trial on them. The claims about what it does to people are worse than untested: the last chapter says that if it doesn't produce a fuller life it wasn't really this system, which puts them beyond testing altogether. |
| i CONTEXT | Has anyone actually used it and reported back? | Not that I can find. No result either way, so there is nothing to score here yet. |
| ✕ FAILS | How many workplaces tried it, and how many did it work in? | Nobody counted. Not one number in 333 pages. That matters because it means the book's fame can't stand in for proof — and this book is very famous. |
| ✓ HOLDS UP | Can you tell the useful part from the sales pitch? | Yes, and they come apart cleanly. The methods are usable. The pitch — that the system is psychologically correct and therefore the final word on management — is where the untruths sit, and where the attitudes of 1914 are built in. |
| ▲ CARE | What's the best argument against it? | That taking the thinking away from the person doing the job is a way of controlling them, not a way of getting more done. Braverman made that case, and her own sentences hand him the evidence. |
| ✓ HOLDS UP | Could you actually do it? | The individual methods, easily — most are a single decision or a checklist. The whole system, with its planning department and its eight specialist foremen, no. Nobody is suggesting you should. |
| ▲ CARE | If you stopped, how hard is it to start again? | Depends which half, and she says so herself. The thinking tools you can pick up any time. The standards are unforgiving, because habits decay — and her own rule is that you can't judge a method until using it has become a habit. |
| i CONTEXT | How much would actually change if you used it? | Three different sizes, depending which part you take. The full system is a replacement — it changes how an entire workplace runs, all at once. The individual methods are small and steady — a bit better each week, no upheaval. And the record ladder changes nothing at all: it is a thermometer, not a treatment. It tells you where you stand so you can decide what to do next. |
| 01 | MANAGEMENT IS TAUGHT, NOT INNATE | It stopped being a knack the moment it was written down. |
| 02 | SEPARATE THE OUTPUT FIRST | Nothing individual — task, teaching, reward — exists until you can see one person's work. |
| 03 | THE PLAN CANNOT BEAT ITS RECORD | Four grades of record, each buying you a grade of forecast. No shortcut. |
| 04 | THE UNIT CHOOSES THE BEHAVIOUR | Pick the measure whose gaming is the result you wanted anyway. |
| 05 | NOBODY TEACHES IF TEACHING DOESN'T PAY | The teacher was never unkind. He was unpaid, and training his own replacement. |
"It has been assumed that management was a 'gift,' a 'knack,'--that the man who could manage was born, and could not be made."
That sentence is what the book exists to destroy, and everything in it serves that one argument.
In 1914 management was not a subject. It was something you either had or hadn't, picked up by standing next to a man who had it. Gilbreth's claim was that it had just become teachable — and that the missing half of the new science was the mind of the person doing the work.
I read it because the chapter titles looked useful. Having read the whole thing, the titles turned out to be the least useful part of it.
"Science, or even system, must be written."
The old management was oral. That is her diagnosis, and she calls the oral-versus-written split "the essence of the difference" between the two systems.
It is a sharper claim than it looks. If the method lives only in a foreman's mouth, then it cannot be inspected, cannot be compared with a better one, cannot be improved on purpose, and dies with him. Every apparatus in this book — the instruction card, the record, the standard, the programme — is a device for getting management out of speech and onto paper.
She takes it right down to the organisation chart. Of the eight functional foremen, the four in the planning room "communicate with the workers mostly in writing"; the four out on the job "usually communicate with the men orally." The split runs through the structure, not just the paperwork.
A knack cannot be audited, argued with, or handed on. A written method can be all three. That is the whole of the difference.
"The recognition of individuality is the subject of this chapter."
This is the cheapest move in the book and everything else depends on it.
Separate one person's output from the group's, and a chain opens up: you can record them individually, so you can set them an individual task, so you can teach them individually, so you can reward them individually. Under gang work not one link of that exists. Her own staging says as much — separated output is the first thing that appears when an organisation starts to leave the old way behind, long before anybody owns a stopwatch.
Her summary of what was wrong before puts it first of three: "men are handled mostly as gangs, and personality is sunk."
Two details worth having. The instruction card is signed by whoever wrote it, and it tells the foreman in large type that if the work cannot be done as written, he reports to the man who signed it. Accountability points upward: when the method fails, the named party is the planner.
And her argument for why records matter to the worker is not about data at all. It is about permanence — the stone-cutter who marks a block he knows will be walled up out of sight, the singers of her day rushing to make gramophone records because a live performance "dies with the day." She thinks people want evidence that their effort happened.
You cannot teach, task, or fairly pay a group. You can only ever do those things to a person.
"Record unconscious. Programme cannot be definite. Method is indefinite."
The single most useful table in the book. Four grades of record, each buying exactly one grade of plan:
That is a diagnostic anyone can run this afternoon. Ask what grade your record is at, and it tells you what quality of forecast you are entitled to. Most organisations are asking for grade four answers off grade two records, and calling the gap a planning problem.
What an unconscious record costs, in her words: the man "has no idea of his capacity"; every good method he stumbles on "is practically lost to the world, and perhaps also to him"; and — the line that stings — "all bad methods which he has fallen into will be fallen into again and again, as there are no warning signs to keep him out of them."
Then the sentence I would put on the wall:
"Improvements are not made between the programme and the following record — they find their place between the record and the following programme."
Read that twice. Under the old way, the job is planned badly, then rescued by improvisation while it runs, and the record comes out looking nothing like the plan. Under hers, the job matches its plan, and every improvement is moved to the moment of writing the next plan. An organisation that fixes everything mid-job never improves, because the learning never reaches a plan. That is the improvement cycle, stated in 1914 and usually credited to people born thirty years later.
This is where her argument about standards belongs, and it is the best argument for a written procedure I have read — because it never mentions consistency. The instruction card is "a most accurate memory in inanimate form, that neither blurs nor distorts with age"; it is "a form of transferable memory"; and, brutally, "experience, from an earning standpoint, does not exist when the mind does not retain a memory of the method." A written procedure is not a rule imposed on somebody. It is somebody's best answer, kept.
Two more that come with it. Standard records give "not a greater rigidity, but a greater elasticity," because standard elements recombine into many predictable plans while bespoke wholes recombine into none. And her test for whether a record should exist at all: "No record should be made which does not, directly or indirectly, actually reduce costs or in some way increase efficiency." Quality of records before quantity of records — from the woman usually blamed for inventing reporting bloat.
"Always select that unit of output that will, of itself, cause a reduction in costs."
This is Goodhart's law, thirty-one years before Goodhart was born, and unlike most modern statements of it she does not stop at the warning. She gives the design rule.
Her worked example is a concrete gang, and it is worth following:
Then the belt and braces: use several units to check each other, "for one alone may be misleading, or put an incentive on the workers to give an undesirable result."
Don't ask what you want to know. Ask what people will do to make the number move, and choose the number where the answer is the job.
The rest of her measurement chapter earns its keep too. Four views of a person's capacity — what they think it is, what their colleagues think it is, what their managers think it is, and what measurement shows it to be — with the observation that the manager is usually the most wrong of the three. And before measurement exists at all, "the only standard is what someone thinks about things, and the pity of it is that even this condition does not remain staple." Opinion is not merely unreliable. It drifts.
Her rule against secret observation lives here, and her first reason for it is not fairness but data: someone who does not know they are being watched cannot help you find the wasteful step, so the study "will probably be worthless." The fairness reason comes second and it is the one that lasts — measured in secret and found out, "he will lose confidence in the squareness of his employer," and will never again be sure he isn't being watched.
She also has the procedural fix for the observer's own bias, which is the part I did not expect from 1914: do not chart the data until every observation is in, because watching a curve build "develops a distinct theory in the mind of the investigator and is apt to bend the curve." That is pre-registration, in a management book.
"It is not to the pecuniary advantage of any particular person to give this teaching."
The best thing in the book, and it is an economic argument, not a moral one.
Why did nobody in a 1914 workshop teach anybody anything? Not because they were unkind. Because:
And the person they would have to ask for help "might discharge him for not already knowing."
Her conclusion is the sentence that makes this transferable: "The worker is not to be blamed for this attitude. The conditions under which he worked made it almost inevitable."
So the fix is structural, and there is not a word of exhortation in it. The interdependent bonus ties a teacher's pay to their learner's output. Nobody is promoted until they have trained their replacement. Teaching is made to pay, and it starts happening.
If knowledge isn't moving through an organisation, don't ask who is being unhelpful. Ask who is paid for it, and who is made worse off by it.
Three hard constraints come with it, all still usable:
The instruction card gives the what and the how. The System gives the why. Two documents, two jobs, and she is insistent about the second — "unless the man knows why he is doing the thing, his judgment will never reenforce his work. He may conform to the method absolutely, but his work will not enlist his zeal."
No method counts as standard unless the teacher can demonstrate it, on request, at any time. A written instruction nobody can perform on demand is not a standard. That also rules out a whole category of modern training material: a generated video can carry the teaching layer, but it can never pass this test, because there are no hands behind it.
Motions first, then speed, then quality — in that order, which inverts how every trade had been taught. Teach quality first and the learner reaches it by whatever motions they improvise, and those motions set hard; every later attempt to speed up is then fought by "an involuntary impulse of the muscles to follow the old methods." She admits the cost honestly — the motions-first learner is slower to good quality at the start — and argues the eventual combination repays it. Right habits are cheaper than corrected ones.
The five ideas above are the book at its best. This is the working behind the scoring at the top, and the reasons not to take any of it on trust.
01 · Taught, not innate. Writing something down does not make it true, and she treats the two as the same move. A written standard derived from one observation of one man on one morning is exactly as arbitrary as the foreman's opinion it replaced — it is just harder to argue with, because it is now in a file with a signature on it. She also never asks who is permitted to read the reasoning back. The worker gets the instruction card; the reasoning stays in the planning department.
02 · Separate the output. Visible individually is also comparable individually, and disposable individually. The group was a shield as well as a fog, and she never weighs what is lost when it goes. Her own chapter IX supplies the evidence against her: where the reward went to one winner, the men "got together secretly, and selected the winners for a week or more ahead" — sharing the prize rather than competing for it. They preferred the group. She reports it as a defect of the reward scheme rather than as a preference worth understanding.
03 · The record ladder. It measures the form of a record, never its honesty. A standardised record taken from someone who knew they were being timed, on a good day, with good material, is grade four by her scheme and still wrong — and it now carries the authority of a standard. She is aware of observer bias and writes well about it, then treats the resulting record as settled. Predictable is not the same as true.
04 · The unit chooses the behaviour. She names the trap and then walks into it. The whole system is paid on a bonus for beating a time — a single unit, set by the people who benefit from it, with quality policed separately by an inspector precisely because the unit itself does not protect quality. Her own chapter III concedes it when it quotes Schloss: a foreman bonused on quantity will trade away quality, so you must hire someone else to watch. That is not a well-chosen unit. That is a badly chosen unit with a guard bolted on.
05 · Making teaching pay. The fix works, and it works partly by doing exactly what the teacher feared. Their trade secret becomes the company's standard; they are paid for surrendering it, and afterwards they are more replaceable than before. She calls this conservation of knowledge. They would have called it the thing they spent thirty years avoiding, and they were not wrong about the mechanism — only outvoted on whether it was worth it.
Braverman's charge is aimed at exactly the mechanism this book spends its length defending. Scientific management, he argues, was never really a technique for producing more; it was a technique for producing control. Its defining move — what Gilbreth calls separating the planning from the performing — is the separation of conception from execution: management takes the knowledge, the judgement and the planning, and leaves the worker the execution.
On his account that is not a side effect of the system. It is the system. Knowledge built up by the people doing the work is removed to the office, and what comes back is an instruction card. Gilbreth's "conservation of memory" and Braverman's "monopolisation of knowledge" describe the same transfer and disagree only about who ends up better off.
She saw the objection coming and answered it three ways. The typist is no less human than the scribe. A standard is a memory aid, and objecting to it is like wanting workers to be animals with no memory. And a standard is a floor to invent from, so the worker can improve "from that point upward, instead of wasting, perhaps, a lifetime in striving to get up to it."
And then she wrote this, in the same chapter:
"It is out of the question to permit the deviations resulting from individual initiative. Individual initiative is quite as objectionable in obtaining the best results... as service would be on a railroad if each locomotive engineer were his own train despatcher."
Standards "must be followed absolutely by the worker," and initiative has "a distinct place... but that place is not outside of the planning department." The strongest evidence for Braverman's case is in her own sentences, on the same pages as her defence against it. Any account of this book that reproduces the defence without that passage beside it is misrepresenting it.
Where Braverman is himself vulnerable, and it should be said: Richard Edwards has argued that he takes Taylorism's own account at face value — treating managers' descriptions of how completely the system was installed as a description of what actually happened on the floor, and so missing both the real labour process and the resistance to it. The deskilling thesis also assumes control is an end in itself, when autonomy and reskilling can serve profit perfectly well.
Two things this note cannot claim, and will not pretend to.
Braverman was not read directly — the full text could not be reached, so his position here rests on three academic secondary sources that agree with one another. Whether he engages Lillian Gilbreth by name, rather than Taylorism as a movement, is unverified. Read the section as the strongest published attack on the book's central claim, not as a personal reply to its author.
No substantive published critique of this book as a book was found. It is treated in management history as a landmark and in feminist scholarship as an achievement, and appears to have been very little argued with. Nobody serious seems to have taken its psychological claims and tested them — which is a finding about the book's reception, and a poor one.
The chapter-by-chapter psychology is real and properly cited — James, Sully, Stratton, Calkins, Ebbinghaus, Montessori. The claims about what happens to people are not. Better health, better colour, better appearance, contentment, industrial peace, the removal of the fear of losing your job: the evidence offered is that "all scientific managers testify," plus a magazine interview at a showcase firm concluding "they are all contented." Not one measurement, in a book whose entire thesis is that everything should be measured.
And the final chapter closes the door on testing altogether: "if Scientific Management is what it claims to be, it leads to the development of a fuller life... If it does not do this it is not truly Scientific Management." Any installation that fails is thereby defined as not the real thing. That is what puts the human claims beyond weak and into unfalsifiable.
Two racist passages, in different chapters. In the standardisation chapter, the example chosen to illustrate that workers cannot find good methods for themselves is "the West Indian negro who carried the wheelbarrow on his head," called "impossible and absurd." In the teaching chapter, a footnote about why a man copies a teacher risen from his own rank tells an anecdote about "the old darkey" and uses a racial slur twice. One line would be an artefact of its year. Two, in separate chapters, is a pattern in the book.
Psychological records on every child, from birth, proposed three separate times. In chapter II minds are to be "studied, and the results recorded from earliest childhood." In chapter IV a child should be "accustomed to being measured, physically and psychologically, from his first years, just as he is now accustomed to being weighed" — and the stated benefit is that they will not object to being measured once they reach work. By chapter VIII it includes heredity: "the collection of vocational guidance data will begin with a child at birth, and a record of his inheritance will be kept." The recurrence is the finding. This is not an aside; it is a settled part of the programme.
Also on the page: a long passage on the "weaker brother," arguing that low output should be measured to establish whether the cause is weakness, unsuitability or laziness — though the same passage turns the instrument on the strong, and asks whether they are making the weak over-exert.
Read in full from Project Gutenberg eBook #16256, chapters I to X in sequence. Every quotation above is from that text and can be checked against it.
Braverman's and Edwards's positions come from three academic secondary sources rather than from the books themselves — see the flag above. The author's biography and the copyright position are recorded separately with their own sources and access dates.