Governance is a word I use carefully. In the work I do — designing operating systems for complex organizations — governance refers to the set of structures, processes, and accountabilities that determine how decisions get made and how commitments are honored over time. It is not a synonym for bureaucracy or oversight. It is the architecture of reliable operation under conditions of uncertainty and change.
I have spent years applying governance thinking to organizations, to ventures, to systems that involve many people and many moving parts. I came to it more slowly with my own intellectual output — the writing, the positions, the frameworks I publish and teach. The delay is worth reflecting on, because the principles are the same. Publishing without governance produces the same failure modes in an individual intellectual body of work that ungoverned operations produce in an organization: inconsistency, drift, accumulated debt that eventually demands reckoning, and a relationship with accountability that is more performative than real.
What Intellectual Debt Actually Looks Like
In software, technical debt is the additional work created by choosing a faster or easier solution now over a better solution that takes longer. The debt is real — it eventually has to be serviced — but it is often not visible until you try to build on the shortcut.
Intellectual debt works similarly. It is the accumulated gap between the positions you have published and your current best thinking. Every time you publish a position faster than you have thought it through, you take on intellectual debt. Every time you avoid revisiting a published position because the revision would be inconvenient, you let that debt compound. Every time you implicitly maintain a position in a body of work that you have explicitly moved past in private, the debt grows.
The servicing cost is not theoretical. It shows up in several recognizable ways.
Contradictions emerge between pieces published at different times. A reader who has followed your work for two years notices that your current framework contradicts your earlier one in ways you have not acknowledged. They either lose confidence in your consistency or conclude that you do not actually track your own positions. Neither outcome is good.
You find yourself unable to cite your own prior work accurately. You know you wrote something about a topic, but you are not sure whether your current view is consistent with what you wrote or whether you have silently moved on. So you avoid citing it, which means the prior work is floating in public with no clear relationship to your current thinking.
Readers who made decisions based on your published positions feel misled when your position shifts without acknowledgment. This is particularly acute in advisory relationships, where clients or students may have acted on your framework and now discover you have moved past it.
The debt does not go away by itself. It has to be addressed explicitly, and addressing it after the fact is significantly more costly than preventing it through governance in the first place.
Intellectual Output Governance: Four Controls
The framework I use for managing my own intellectual output has four controls. They are not complicated. The difficulty is not in understanding them but in maintaining them under the operational pressure to produce quickly and the social pressure to appear consistent.
Position tracking is the first control. At any point, I should be able to describe my current position on any topic I have written about — not just in general terms, but with enough specificity to distinguish it from adjacent positions I might have held before or might hold in the future. This requires maintaining some kind of running record of what I have published and what I currently believe. Not a formal database — something more like a maintained map of my intellectual commitments.
The test for whether position tracking is working: if a reader asks me whether my current view is consistent with something I published eighteen months ago, I should be able to answer specifically. Not "I think so" or "probably." I should know.
Update discipline is the second control. When my position on something changes — because I encountered evidence that shifted it, or because I thought the problem through more carefully, or because I was wrong about something — I should publish the update. Not bury it in a footnote in a later piece, but address it directly: here is what I thought before, here is what changed, here is where I am now.
This is the control that most people skip. Updating publicly feels like admitting error. In a framework where thought leadership is about credibility signaling, admitting error seems like a threat to the signal. In a framework where thought leadership is about demonstrating how you think, updating publicly is the strongest possible demonstration. It shows that you are actually tracking your own thinking, that you are responsive to evidence, and that you hold yourself accountable to the positions you have taken.
Retraction protocol is the third control and the hardest to execute. Retraction is not revision — it is the explicit withdrawal of a position that turned out to be wrong in a way that matters. Not "I would say this differently now" but "I was wrong about this and I do not want the published version to stand as representing my current view."
Retraction requires a specific act: going back to the original piece, appending a clear note that the position has been retracted and why, and publishing the retraction where the original audience is likely to see it. Most people in the public intellectual space do not do this. It feels like unnecessary self-criticism. The argument against doing it is that the original piece has reached its audience and the damage, if any, is already done.
The argument for doing it is that a body of work is not just the content of individual pieces. It is the relationship you have established with readers over time. A reader who sees you retract something — clearly, specifically, with an account of why — learns something about your relationship with your own positions that they cannot learn from any other act. They learn that you mean it when you take a position, because they have now seen what you do when you turn out to have been wrong.
Quality threshold is the fourth control. It answers the question: what is the minimum standard a piece needs to meet before it is published? This is not primarily about prose quality or production value. It is about intellectual substance: has this been thought through carefully enough that I am willing to be accountable for it? Not perfectly — perfection is not a useful threshold for anything. But carefully enough that I can defend the reasoning, acknowledge the limits, and explain what would change my view.
The quality threshold is what prevents the publication rate from outrunning the thinking rate. If the threshold is too low, you are taking on intellectual debt with every post. If the threshold is so high that nothing meets it, you are not publishing at all. The right threshold is the one that keeps publication roughly synchronous with actual thinking.
How the Four Controls Work Together: A Worked Case
The controls are easy to state in isolation and easy to misunderstand in isolation, because their value is in how they constrain each other. A concrete case shows the interaction.
Suppose I published a framework two years ago arguing that governance friction should always scale with project size — bigger project, more rigor. Position tracking is the control that catches the problem first: when I sit down to write something new that touches governance scaling, the maintained record forces me to confront that I no longer believe the size-based version. What I now think is that friction should scale with irreversibility, not size — a small but irreversible decision warrants more rigor than a large reversible one.
Now the other three controls decide what I owe the reader. The quality threshold tells me the new position is thought through well enough to publish: I can defend the reasoning, name its limits, and say what would change my mind. Update discipline tells me I cannot simply publish the new view as if the old one never existed — I have to mark the change explicitly, because readers who adopted the size-based rule are still using it. And retraction protocol decides the strongest case: if the old framework was not merely refined but wrong in a way that would lead a reader to govern badly, I go back to the original piece and append a note saying so, where its original audience will see it.
Skip position tracking and none of the others can fire, because I never notice the divergence. Skip update discipline and the new piece silently contradicts the old one. Skip the quality threshold and I publish the revision before it is ready, taking on fresh debt to service old debt. The controls are not a menu. They are a sequence that only works whole.
The Failure Mode on the Other Side: Churn
The framework so far guards against under-revision — letting debt accumulate. There is a symmetric failure mode that the discipline must also guard against, and it is easy to miss because it looks like rigor: over-revision. Churn.
A body of work whose positions change constantly is not demonstrating responsiveness to evidence. It is demonstrating that the positions were never thought through to the quality threshold in the first place. If every quarter brings a new framework that quietly replaces the last, readers cannot rely on any of it, and the willingness to revise — which is supposed to make the maintained positions more credible — instead signals that nothing is maintained.
This is why the quality threshold is load-bearing for the whole system. It is not only a gate on what gets published; it is the brake on revision. A position that cleared a real threshold should change only when something real changed — new evidence, a flaw exposed, a context shift. If a position is revised more often than the underlying reality is shifting, the problem is not insufficient update discipline. It is that the threshold was too low and the original position was published before it was actually held. Governance of intellectual output has to hold both edges: revise when warranted, and do not mistake instability for honesty.
The Accountability Relationship
Governance of any kind is ultimately about accountability — the relationship between commitments made and commitments honored. Intellectual governance is about accountability to your own published positions.
This is a strange kind of accountability because there is no enforcement mechanism. Nobody compels you to acknowledge when you have changed your mind or to retract what turned out to be wrong. The accountability is self-imposed. Which means it depends entirely on whether you actually take the commitment seriously.
I have noticed that the practitioners and teachers I find most credible over time are people who have visibly revised their positions on things that mattered. Not constantly — churn is its own credibility problem. But in the cases where revision was actually warranted, they revised. Publicly. With specific acknowledgment of what changed and why.
This pattern of visible revision does something counterintuitive: it makes the positions they do maintain more credible, not less. If I know that someone revises publicly when warranted, I interpret their maintained positions as genuinely maintained — not just as positions they have not gotten around to revising. The willingness to revise is what makes the consistency meaningful.
The inverse is also true. If someone's public positions never change — if they can always be found to be consistent with whatever they said five years ago — one of two things is happening. Either they have been right about everything from the start, which is unlikely. Or they are not tracking the relationship between their published positions and their actual thinking, which is the failure mode I am describing.
Governance Under Operational Pressure
The real challenge of intellectual output governance is maintaining it when you are busy.
For practitioners — people whose primary work involves doing things, not writing about them — the intellectual production happens in the margins. You are running a venture, serving advisory clients, teaching students, building systems. Writing is something you fit around the operational work, not the other way around. Under these conditions, the governance controls are the first things that get dropped.
The position tracking does not get updated. The piece that should be revised gets left standing. The retraction that needs to be published gets deferred. You know it needs to happen, but the operational work always has a more urgent claim on your attention.
This is the governance failure I have experienced most directly. Not a dramatic collapse of intellectual integrity, but a slow accumulation of deferred maintenance. Positions I should have updated sitting in the queue. Revisions I have been meaning to publish. The gap between what I have written and what I currently think, widening gradually.
The response I have found useful is treating intellectual maintenance as a recurring obligation rather than a discretionary task. Specifically: building explicit review cycles into the working cadence rather than leaving revision as something that happens when time permits. Time does not permit, in my experience. The review cycle has to be structural.
For me this looks like a periodic review of recent published positions — not exhaustive, not as long as the operational cycle — specifically asking whether anything has shifted enough to warrant public acknowledgment. It does not need to be frequent. Quarterly is enough if the threshold question is taken seriously: has something shifted enough that I would be misleading readers who are relying on my published positions to stay silent about it?
It is worth being honest about what this discipline cannot do and where it is not worth the cost. It does not apply with equal weight to everything you publish. A short observation made in passing does not carry the same accountability as a framework you teach and that clients build on. Trying to govern every sentence you have ever written to the same standard is not rigor — it is a different way of producing nothing, because the maintenance burden becomes large enough to crowd out new work entirely. The governance effort should concentrate on the positions that carry weight: the frameworks people act on. For the rest, a lighter touch is not a lapse. It is correct proportionality.
What This Has to Do with Advisory Work
The context in which intellectual output governance matters most, for me, is advisory relationships. Clients engage advisory practitioners partly on the basis of their published thinking. They have read something, found it credible, formed a view of how the practitioner thinks — and hired them partly on the basis of that view.
If the published thinking diverges from the advisor's actual current thinking — because positions have drifted without acknowledgment, because intellectual debt has accumulated, because governance has lapsed — then clients are making decisions based on a representation that is no longer accurate. That is not a theoretical problem. It is a real misalignment between what the client believes they are getting and what they are actually getting.
The advisor who maintains rigorous governance of their intellectual output is the advisor whose published thinking actually reflects their current thinking. When they write something, it means something — because they have demonstrated, through the governance practices they maintain publicly, that they take their published positions seriously. When they change their mind, they say so. When they are wrong, they acknowledge it. Their public record is an accurate map of their actual intellectual commitments, not a curated highlight reel of their best-looking positions.
What You Can Do This Week
The full discipline takes time to build. One move does not. This week, pick the single framework you publish that other people most rely on — the one a client or a student would point to if asked what you think about your core subject. Then answer one question in writing: is my current best thinking still consistent with what that published piece says?
If the answer is yes, you have just done a unit of position tracking, and you are clear. If the answer is no, you have surfaced a live piece of intellectual debt, and you now have a small, bounded decision in front of you: refine it with an update note, or retract it if it was wrong in a way that matters. Either way, you have done the actual work of intellectual governance on the one position where it counts most — not by building a system, but by checking whether your most load-bearing published claim is still one you hold.
This is a harder standard to maintain than the alternative. But it is the standard that makes the advisory relationship actually work — not just the transaction of it, but the substance of it. Clients who can trust that their advisor's published thinking is accurate to their current thinking have a reliable access point to how someone who thinks well about their class of problems actually thinks about them right now. That is the value intellectual output governance produces. Not visibility. Not credibility signals. Reliable access to current thinking — the thing that makes the relationship worth sustaining.
Continue in this series
This piece is part of What Is Organizational Governance? A Systems Practitioner's Complete Guide, my systematic guide to organizational governance and operating systems. Related reading:
- Governance Debt: How Accumulated Shortcuts Create Organizational Failure
- The Real Cost of Moving Too Fast in Complex Systems
- Stocks and Flows: The Systems Thinking Concept Leaders Keep Missing
- The Five Signs Your Governance Is Theater, Not System
Working through this in your own organization? I help technical leaders design it directly — advisory engagements.
