Rebuilding the Room

Four essays have made a single case from four directions. Organisations appoint on the certificate because it is legible, train by the fortnight because it produces a document, hand the result machines that reward obedience, and sometimes select against the very people who would have argued. Twice I have promised that the remedy costs less than the certification budget and considerably more trouble. This essay makes good on that promise, and most of it is given to practice rather than argument, because the argument has been made.

The remedy is not a course. An organisation that answers this series with a module in critical thinking or a certificate in psychological safety will have reproduced the error it set out to correct, and will have a document to prove it. Nor is it to hire better people, which is the [[10-19 Areas/11 RW Ink Column/11.01 Published/The Saviour-Industrial Complex|saviour’s answer]] to every institutional failure. The faculty of judgement, as Edmund Burke (1729–1797) insisted, is common to everyone; what differs is its exercise (Burke [1759] 2015). The task is to rebuild the conditions in which that exercise happens.

What the Room Requires

Burke’s phrase—proper and well-directed exercise—has two working parts, and most organisations supply neither. Exercise means cases: real problems, carrying real consequence, met in sufficient number for discernment and judgement to form. Direction means somebody who notices: somebody competent to tell you that you got it wrong, and why, soon enough for the lesson to attach to the case. Hubert Dreyfus (1929–2017) and his brother Stuart located the step from competent to proficient precisely there (Dreyfus et al. 1986). It is how surgeons, barristers and airline captains have always been made. What is new is how seldom organisations outside the protected professions bother.

The room has two enemies, and the second is the less obvious. The first is the manager who takes the case away. The second is the machine. Pierre Heudel and colleagues, surveying the evidence on clinical deskilling, observe that AI triage strips out the foundational, easy cases on which skill consolidation depends (2026, 4). Organisations automate the easy work first because it is cheapest to automate, and those are precisely the cases on which novices learned. How the assistance is designed matters as much as how much of it there is. Hamsa Bastani and colleagues found that students given an unrestricted model did worse once it was taken away, while those given a version that offered hints rather than answers suffered no such penalty (2025; 2025). Assistance that withholds the answer builds capability; assistance that supplies it does not.

The room also needs the people in it to speak, and the fourth essay showed that silence in organisations is rationed by self-interest: people speak readily when their own agenda is threatened and fall quiet where speaking offers them nothing. Three levers change the price. The first is consequence. APRA found that remuneration at the Commonwealth Bank had little sting for senior managers when poor outcomes materialised ((APRA) 2018, 3), and where nobody senior bears the cost of an ignored warning, nobody senior has a reason to hear one. The second is procedure. Dennis Tourish and Paul Robson suggest that managers scrutinise good news as rigorously as bad, and that meetings ask as a matter of routine what criticisms the decisions in hand have drawn (2006, 725). Its virtue is that it depends on nobody’s courage, which matters, because James Detert and Amy Edmondson found that people withhold what they know even under a good boss (2011, 484). The third is protection. James Westphal and Poonam Khanna found that the deterrent to challenging a chief executive outweighed directors’ own shareholdings and the pressure of institutional investors (2003, 393). You cannot pay people to bear a social cost the institution keeps imposing. You can only stop imposing it, and be seen to stop, by promoting the people who argued.