Cardano Constitution v2.4

A note on the record. This rationale is entered after the fact. At the time this amendment was voted, this DRep had stepped back from active voting to concentrate on more foundational constitutional work, and the analysis below — though largely complete — was never turned into a submitted rationale. I record it now not to re-litigate a settled outcome, but because the framework I committed to in my original rationale obliges me to evaluate trajectories over time, and enough time has now passed that the trajectory this amendment set is legible in ways it was not on the day of the vote. Several of the concerns registered here as risks have since become observable. Consistent with that framework, this is the rationale I would have submitted.

Contents§

• 1.0 Introduction
• 1.1 Summary
• 1.2 Description of Governance Action
• 2.0 Discussion
• 2.1 Method
• 2.2 The bundling problem
• 2.3 The changes, one at a time
• 2.4 The change beneath the changes
• 2.5 The trajectory in one view
• 3.0 Conclusion

1.0 Introduction§

1.1 Summary§

We are voting NO on this governance action to amend the ratified Cardano Constitution to version 2.4.

I want to be as clear about the character of this NO as I was, in my original rationale, about the character of that YES. The YES on the delegate-endorsed Constitution was a YES with reservations — an endorsement of starting conditions and of a monitorable trajectory, not of finality. This NO is its mirror. It is not a rejection of clarity, of rigor, or of enforceability, each of which this amendment genuinely advances. It is a judgment that the amendment, taken as the single indivisible object the vote actually presents, moves the constitution along a trajectory that draws down capacities I identified as load-bearing, without specifying the institutions that would replace them — and that it does so in a bundle that makes the drawdown difficult to see, difficult to attribute, and difficult to correct.

Three findings carry the vote. First, the amendment bundles heterogeneous kinds of change into one take-it-or-leave-it action, which is a procedural fail-state under my framework independent of the merits of any single change. Second, several of the substantive changes are individually defensible but, in combination, activate precisely the self-defeating dynamics my Hippocratic floor exists to detect — compression of deliberation into fewer, higher-stakes moments, and the concentration of interpretive authority by accretion rather than declaration. Third, and least visible on the day but most consequential in retrospect, the amendment performs a quiet, coordinated re-grounding of the constitution’s account of money and rights onto ownership — a move that is operationally nil and ontologically decisive, and that runs directly against the rights trajectory the Constitution promised to walk.

That third finding is the reason this rationale reads differently from a routine NO, and the reason it took me some time to know what to do with the analysis behind it. Arriving at this vote is what showed me that a YES/NO vote was the wrong instrument for the constitutional problem that mattered most. I will return to that at the close.

1.2 Description of Governance Action§

This governance action is of the type “Update to Constitution,” voted on by DReps and the Constitutional Committee. It replaces the ratified constitutional text with a revised version, v2.4, and — as required by the amendment process enshrined in Article VIII — references the prior constitution’s hash within the same action.

The amendment consolidates, into one action, at least five distinct categories of change: the removal of Budget Info Actions as constitutionally meaningful instruments, with budget requirements folded into Treasury Withdrawal actions; a requirement that Treasury Withdrawals be fully self-contained; stricter standards for referencing external documentation, via immutable links and document hashes; the removal of non-binding, aspirational, or “encouragement” language in favor of enforceable obligations; and a set of clarifications and scope adjustments to the Constitutional Committee’s role, alongside definitional cleanups. Several of these are framed in the surrounding discourse as clarifications rather than substantive changes. Under my framework, that framing is never accepted at face value, because clarity can shift power without adding a single new verb.

I take each in turn below. But the first thing to evaluate is not any of the five. It is the decision to present them as one.

2.0 Discussion§

2.1 Method§

I carry forward, without alteration, the evaluative framework established in my rationale for the delegate-endorsed Constitution, because the whole point of building that framework was to be able to use it again. The reader who wants the full derivation will find it there. In brief: I decline to evaluate from a purely internal (participant) or purely external (observer) perspective, and instead move between them using usefulness as the bridge; I favor the analysis of institutional and process trajectories over before-and-after snapshots, while minding the gap between constitutional text and observed practice; I evaluate against four mid-range criteria — sociological legitimacy, the channeling of political conflict into institutions, the limiting of agency costs, and the creation of public goods; and I hold all of this above a Hippocratic floor of minimum viable governance, the requirement that a design not contain elements likely to conduce unintentionally to the breakdown of the very governance it aims to enable.

Applying this to an amendment rather than to a whole constitution asks something slightly different of the framework, and it is worth making the amendment-level method explicit, because it is a method I would like to see the ecosystem adopt. For any proposed change I ask, in order: what actually changed — in text, in role powers, in process constraints, in interpretive authority, in the addition or removal of institutional pathways? Does the change introduce a new path to governance deadlock, interpretive monopoly, emergency normalization, or conflict externalization — the Hippocratic question, asked first, because a yes here is a presumptive failure regardless of intent? How does it move each of the four mid-range criteria? Where does it land on the nine evaluative dimensions, and especially on temperance, deliberative capacity, adaptability, and durability? And finally, is this a tightening for clarity and enforceability, or a narrowing that collapses pluralism and checks — a question of trajectory, not of motive.

To that method I add one discipline that the original rationale implied but did not name, and that the analysis of this amendment forced me to name. A change must be read at more than one level. A change can be operationally nil — moving no procedure, altering no threshold, redistributing no power — and still be significant at a level the operational lens cannot register. I learned this the hard way on one of the changes below, having first dismissed it as symbolic, and the correction is instructive enough that I have kept the error visible rather than quietly fixing it.

A note on register, which I mean to hold throughout. What follows is diagnostic, not accusatory. The findings are fitness gaps and trajectory readings, not charges of bad faith. The amendment was authored under real budget-season pressure; the enforceability gains are real; several of the individual choices were, at the time, temperate and defensible. The concern here is system-dynamic, not moral. Nothing in the analysis requires bad actors. The risks emerge from interaction effects, from the cumulative removal of buffers, and from increased reliance on interpretation under pressure — which is exactly why a Hippocratic lens, tuned to self-defeating design rather than to intent, is the right instrument for reading them.

2.2 The bundling problem§

Before any substantive clause, there is a change in strategy to evaluate: the decision to present grammar, structure, operational rules, normative removals, and role-power adjustments as a single consolidated action. This is not a process nit. Under my framework it is a procedural fail-state, and it is a fail-state independent of whether one ultimately supports or opposes the substance.

The reasons are structural. Bundling defeats trajectory analysis, which is the method I committed to: when many changes ship as one, causal attribution collapses, clean counterfactuals become impossible, and multiple trajectories fold into a single inseparable outcome that no good-faith observer can evaluate incrementally or learn from. Bundling weakens legitimacy by forcing false coalitions: voters must accept changes they oppose to secure changes they support, or sacrifice beneficial changes to block harmful ones, and the residue is a durable sense of having been procedurally steamrolled — most sharply among the minority, whose disaffection is precisely what erodes sociological legitimacy over time. Bundling forecloses the rollback-like behavior that atomic amendments would otherwise permit, so that correcting a single problematic change requires re-opening the whole constitution, which discourages correction altogether. And bundling raises Hippocratic risk directly, because harmful interaction effects become visible only after adoption, yet cannot be isolated once surfaced — the packaging conceals structural risk until remediation is harder.

There is a rollback-friendly remedy available without any protocol change, and I record it because I would rather propose a norm than only diagnose its absence. The ecosystem could commit to a changelog broken into atomic deltas, each tagged by kind — clarifying, organizational, operational, power-or-authority — with a standing convention that operational and power changes ship alone, or in a very small cluster, while grammar and definitions ship separately; and it could maintain a revert-template that restores prior wording for a single subsystem when downstream effects prove harmful. That is nothing more than my emphasis on monitorability and trajectory evaluation, made procedural.

I flag bundling first, and as a fail-state in its own right, because I believe it is right to do so before touching substance. Everything below inherits the cost of it: there is no clean counterfactual for any single change, no isolated metric to attribute, and no way for dissent to stay specific rather than generalizing to “the amendment.” That is a recursive harm — the bundle degrades the very deliberative and corrective capacities whose erosion is the substance of my objection to what the bundle contains.

2.3 The changes, one at a time§

2.3.1 The removal of Budget Info Actions as constitutionally meaningful. The budget pathway was, before this amendment, a two-stage affair: a non-binding but constitutionally recognized Info Action used to socialize proposals, structure deliberation, and surface objections cheaply and early; followed by the binding Treasury Withdrawal. The amendment strips the Info Action of constitutional standing and requires the withdrawal to carry everything itself. The constitution now recognizes a single authoritative budget gate where it once recognized two.

I do not defend Budget Info Actions as they stood. They were confusing, inconsistently used, and poorly socialized; the Constitutional Committee’s votes on them were widely misread, and net-change limits were misinterpreted by influential voices. Clarification was going to be necessary. My objection is that excision is the most extreme form of clarification available, and that it resolves confusion by shrinking the system rather than by helping the system learn. Under my framework, confusion is diagnostic data — a signal of insufficient institutional literacy, not of institutional invalidity — and the appropriate response to a misread instrument is to improve its affordances, not to remove the surface that generated the signal. Clarification could have tightened the language around Info Actions, explicitly bounded CC authority over them, separated budget signaling from budget compliance, or sunset them against a specified replacement. Instead the system now says, in effect, figure it out socially — which relocates a constitutionally enshrined deliberative stage into extra-constitutional space and hopes it survives there.

This is the amendment on which I most sharpened my own position during analysis, and the sharpening matters. I accept tradeoffs in principle. What I reject is a tradeoff that removes capacity without a compensating mechanism, and I treat efficiency claims as hypotheses rather than justifications. On that distinction, this is not a tradeoff we happen to dislike; it is a regression with an unproven hope of compensation. In my original rationale I named the deliberative capacity around the budget — the primary public good the constitution aspires to deliver — as under-specified and in need of protection, and I identified Info Actions as one of the few formal deliberative levers the constitution actually contained. Removing that lever raises, rather than lowers, the bar for whatever is meant to replace it, and no replacement is named.

2.3.2 Treasury Withdrawals must be fully self-contained. This change is often read as a formatting requirement. It is not. It changes the constitutional posture of a treasury decision, converting it from the endpoint of a distributed, multi-stage, multi-artifact process into a singular constitutional object that must contain its own rationale, scope, references, compliance evidence, and constraints, to be assessed by everyone at once. The review burden that was formerly spread across time and artifacts is compressed into a single submission, a single window, and a single judgment.

Paired with the removal of the deliberative pre-stage, this compression is where the Hippocratic floor comes under real pressure. Decision compression tends to increase false negatives — reject, it is too risky to assess whole — and to push disagreement into adversarial and reputational registers, because an actor who loses here does not revise and resubmit; they exit or escalate. That is the textbook shape of conflict externalization, and it is the opposite of the conflict-channeling function I credited the original design with. Compression also concentrates interpretive power, because “constitutionally complete” is itself an interpretive judgment, and it falls disproportionately on the Committee — which touches my judicial-supremacy concern directly. And it creates a standing pressure toward emergency normalization: when stakes are high and timelines tight, actors begin to argue necessity, exceptions come to feel reasonable, and emergency logic becomes habitual rather than exceptional. That is how emergency powers inflate — not by text, but by practice.

Underneath all of this is a category error worth stating plainly. This change treats a budget decision as though it were a parameter change — a legally sufficient artifact rather than a socially negotiated public good. Budgets are contested, iterative, value-laden, and political. Making them behave like parameter changes is a classic governance mistake, and it is one my framework is specifically built to catch, because it mistakes formal correctness for governance health.

2.3.3 Immutability of proposal references. Here the framework points the other way, at least in isolation. Requiring governance actions to reference off-chain documents via immutable links and content hashes shifts context from persuasive narrative to cryptographically anchored evidence, and that closes a genuine breakdown vector: people voting on one thing, opponents claiming it was another, authors later clarifying into a third, and the Committee forced to adjudicate a moving target. Anchored documents reduce post-hoc reinterpretation, procedural suspicion, and the bait-and-switch accusations that corrode legitimacy. On its own, this change contributes positively to legitimacy, to conflict channeling, and to the limiting of agency costs, and it likely passes the Hippocratic floor.

The concern is interaction. In a regime that has already removed its soft deliberative stage and now demands constitutionally complete artifacts, immutability can turn ordinary iteration into formal failure — a frozen but incomplete context becomes grounds for rejection, participation costs rise, tooling advantages accrue to institutional actors, and authorship concentrates. Immutability shifts, at that point, from accountability to procedural gating. So the watch-question is not “is anchored referencing good hygiene?” — it plainly is — but “does immutability, in this regime, function as accountability, or as a gate that moves power toward reviewers and insiders?” The honest answer requires monitoring: the share of actions failing on reference-format rather than substance, the median time-to-submit, the diversity of submitters, the resubmission rate, and whether “insufficient documentation” comes to dominate the Committee’s stated reasons. This change is a conditional positive, and the condition — low-friction tooling and an explicit, non-penalized iteration norm — is exactly what the surrounding amendment fails to supply.

2.3.4 The removal of non-binding and aspirational language. The stated principle is that only rules which can be adjudicated belong in a constitution. This has a legitimate Hippocratic benefit: aspirational text can be weaponized as enforceable law, and its removal lowers the interpretive discretion available for exactly the “spirit of the constitution” overreach my judicial-supremacy concern tracks. On its own, the change is narrowly defensible.

But non-binding clauses do constitutional work that is easy to underestimate. They create shared direction without coercion, they legitimate practices before those practices are enforceable, and they let institutions learn before being locked in. They are, in my framework, part of the system’s error-tolerance and learning layer — developmental scaffolding, not sloppy law. A constitution does not only rule; it also orients, and it establishes a field of meaning rather than merely a rulebook. Removing the soft layer forces disagreement into binary compliance questions, increases adversarial framing, and loads the adjudicator further — conflict concentration, again, rather than channeling. And it does this at the moment in the sequence when the last remaining constitutional space for norms to develop before enforcement is being closed. On its own the change passes the floor narrowly; in combination with the three preceding it, it fails, because it removes the learning surface just as the system’s stakes and rigidity are rising. This also touches the final question I reserve for constitutions I judge good — whether they inspire — because a text that speaks only in enforceable obligations speaks less in the community’s own voice, and citizens experience governance as compliance rather than as participation.

2.3.5 Constitutional Committee clarifications. These are the changes where “looks harmless” most easily becomes “structurally decisive,” and they are the last category that can shift the balance of the system rather than merely harden it. They are almost always framed as clarifications, not expansions — the Committee reviews constitutionality and not merits, does not substitute for democratic judgment, does not initiate actions. On paper, each such clarification reads as a constraint.

The framework asks not whether a line exists but where it is drawn, and in this context the line moves in a direction the text does not admit. When the constitution narrows the Committee’s explicit remit while simultaneously hardening what counts as constitutional compliance — via the removal of the deliberative pre-stage, the self-contained-withdrawal requirement, the immutability standard, and the deletion of soft guidance — the Committee ends up exercising more power in practice, even while disclaiming intent, because compliance judgments have become decisive and compliance is now more complete, more rigid, and more all-or-nothing. A “we only review constitutionality” clarification, in this regime, increases effective authority. This is functional expansion via boundary hardening, and it is precisely how judicial review drifts into judicial supremacy in my account: not through overt grabs, but through cumulative responsibility concentration. Transparency and reporting improvements, welcome as they are, do not resolve this — added without a structured response window, a mandated iteration pathway, or an appeal mechanism, transparency becomes finality theater: clear, legible, and unchallengeable. Legible authority without contestability accelerates legitimacy decay. My position on the Committee is therefore not “the Committee has too much power.” It is that the system is removing buffers faster than it is adding safeguards.

2.4 The change beneath the changes§

There is a sixth change, and I nearly missed it. Among the definitional cleanups, the amendment revises the defined term “ada holder” to “ada owner” throughout. In my working analysis I flagged this, assessed it as operationally nil — moving no procedure, altering no threshold, redistributing no power — and concluded from that operational nil that it was insignificant. I am correcting the conclusion, not the finding beneath it. The operational reading was right and remains right. What was wrong was inferring insignificance from operational nil, and that inference held only because I applied a single lens to a change whose significance lives at a level that lens cannot register.

Read ontologically rather than operationally, the term change is not a lone datapoint. It is the surface of three coordinated moves introduced at v2.4, all leaning the same way — toward defining participants and their rights in terms of ownership of the token rather than the agency the token serves. The Preamble’s foundational promise, which in the ratified text pledged to guard the rights of those who utilize the system, acquired the phrase “and the rights of ada owners” — the promise itself gained an ownership predicate. Tenet 10, which in the ratified text asserted nothing about what ada fundamentally is, acquired a definition of it as a “store of value” — the monetary tenet acquired a reification ontology. And the defined term shifted from holder to owner throughout. Taken singly, each reads as housekeeping. Taken together, they are a consistent semiotic drift: a quiet, coordinated re-grounding of the constitution’s account of money and rights onto ownership.

That re-grounding deepens reification — the token as a possessable thing that contains value — at exactly the moment monetary integrity should be pushing the other way, toward the token as an index of productive trust. It runs against the rights trajectory. And it is, I have come to think, the most rights-relevant thing v2.4 did — done in three places at once, and made nearly invisible by the same bundling that obscures everything else. I record the error I made about it deliberately, because it is the clearest illustration I have of why a change must be read at more than one level, and of how a governance instrument that only ever asks the operational question will be structurally blind to the constitutional move that matters most. The fuller treatment, with the ratified and source texts in hand, lives outside this rationale; here it is enough to say that the operational lens returned “nil” and was correct, and that stopping there would have let the single most consequential move in the amendment pass as a typo.

2.5 The trajectory in one view§

Set the six findings down together and a shape appears that no individual finding carries on its own. Across the amendment, the system now emphasizes legally sufficient artifacts, enforceable completeness, binary compliance judgments, and centralized review boundaries; and it has reduced soft coordination, learning surfaces, staged contestation, and constitutionally protected experimentation. The vectors of drift are consistent: reduced deliberative capacity, increased decision compression, increased procedural rigidity, increased interpretive concentration, and reduced constitutional learning surfaces. The resulting system is more enforceable and less developmental. That is not, in itself, illegitimate. It is a change in the kind of system the constitution is becoming — a shift from developmental governance toward compliance-centric adjudication — and it is a change made without the compensating institutions that would keep the newly hardened enforcement layer from becoming brittle.

I want to hold open, rather than resolve, the strongest case against my own reading. There is a genuine defender’s position that Tenets-as-criteria was the correct temperate choice, that premature enumeration risks ossification and a scramble of ad-hoc claims on public goods — a risk my own rationale named — and that the enforceability gains here are real. Nothing in my analysis refutes that position. What my analysis does is relocate it. Even granting every enforceability gain, it is the developmental account of the constitution that this amendment weakens — and the developmental account is the one that justified so much of the original YES. Whether the regression proves harmful in practice remains, as I have always held, an empirical question, answerable only by disciplined observation of participation rates, rejection patterns, authorship concentration, the Committee’s reasoning trends, and off-chain conflict intensity.

But there is a deeper finding, and it is the one that moved this from a close call to a NO. The ratified Constitution did not enshrine rights; it placed the Tenets in the rights slot as an explicit placeholder — an early embarkation for a process of discovering, defining, and enshrining the rights the Preamble promised to guard. That placeholder was accepted on credit. The rights were a constitutional debt, deferred on the understanding that the system would develop toward paying it. Paying that debt requires exactly the faculties this amendment spends: deliberative capacity to surface and test candidate rights, learning surfaces to let norms emerge before they are formalized, contestability and soft language as the scaffolds of that emergence, and a developmental rather than a compliance-centric governance philosophy. The drift removes each of these. So the debt is not merely left unpaid; the faculty that could pay it is atrophying. The system was meant to develop from Tenets to rights, and v2.4 reduced its capacity to develop at all — and then, in the ownership-framing drift, took an active step in the wrong direction on the very monetary question the rights derivation must later get right. That is why I cannot record this as a YES with reservations. The reservations are not at the periphery of the change. They are aimed at its center.

3.0 Conclusion§

We are voting NO. The vote rests on three things: that bundling heterogeneous changes into one indivisible action is a procedural fail-state under my framework, degrading evaluability, forcing false coalitions, foreclosing correction, and masking interaction effects; that several individually defensible changes, in combination, breach the Hippocratic floor by compressing deliberation into fewer high-stakes moments and concentrating interpretive authority by accretion, without any compensating institution named; and that beneath the procedural changes, the amendment quietly re-grounds the constitution’s account of money and rights onto ownership, in a direction that runs against the rights trajectory the Constitution exists to walk. Individually, much of this is defensible. Collectively, it is a regression in the system’s capacity to deliberate, to learn, and to develop — the capacities on which every future right depends.

This NO does not argue for reversal by default. It argues, under the same framework that justified the original YES, for balance restoration: rebuilding deliberative capacity, protected iteration pathways, and contestability to match the newly hardened enforcement layer, and — most concretely — for a standing commitment that future amendments be submitted as discrete, independently evaluable changes wherever possible. A constitutional system that means to remain a living document must preserve its capacity not only to decide, but to learn.

I will close where the analysis behind this vote eventually took me, because it is the real reason this rationale exists. Working through this amendment is what taught me that a YES/NO vote is not the right instrument for the constitutional problem that matters most here. The problem that matters most is not procedural; it is that the rights the Preamble promised to guard were deferred as a debt, and that the amendment, in re-grounding money onto ownership and store of value, made that debt both harder to pay and easier to mistake for already paid. No ballot can reach that. It can be reached only by going underneath the vote — by asking what Cardano actually is, such that its participants have rights at all, and then deriving those rights from that nature rather than decreeing them or voting them up or down. That question is what I left the ballot to pursue. This rationale is the last thing I owed the ballot before I did.

I offer it now not as a reversal of a settled outcome, but as a record — of a trajectory read early, of a method I would like the ecosystem to adopt, and of the moment the work turned from voting on the constitution to deriving what the constitution was always reaching toward.

DRep ID: [re-registered DRep ID — to be filled] DRep Profile: [profile link — to be filled]

Stay in touch. X: https://x.com/styg50