Discretion¶
Core Idea¶
Discretion is the structural arrangement in which a rule system deliberately leaves a gap open and delegates the choice within that gap to a situated agent's judgment, rather than fully determining the outcome by a fixed rule. The defining commitment is not vagueness but a designed delegation: the rule fixes the outer boundaries—the permissible range of action—while reserving the selection among permissible actions to the agent who confronts the particular case. [1] This trades the predictability of rigid rules for the adaptiveness of case-by-case judgment, and it answers a recurring structural problem that any rule-governed system eventually faces: the impossibility of enumerating in advance every contingency a rule must govern, as H.L.A. Hart (1961) analyzed in his account of the "open texture" of legal rules. [2] The classic statement comes from Kenneth Culp Davis (1969), who defined discretion as existing "whenever the effective limits on [an official's] power leave him free to make a choice among possible courses of action or inaction"—locating discretion precisely in the space the rule declines to close. [3]
The prime is a relation between a rule and an agent, not a property of either alone. A rule with no gap leaves no discretion; an agent with no governing rule exercises raw will, not discretion. Discretion lives in the structured middle: bounded freedom inside a frame that someone else set.
How would you explain it like I'm…
Letting the Helper Choose
Choice Inside Set Limits
Bounded Delegated Judgment
Structural Signature¶
Discretion encodes a structural pattern: bounding rule → reserved gap → situated agent → bounded choice → reviewability. It separates two regimes—what the rule fully determines and what it deliberately leaves open—and names the delegated authority to choose within the open region, along with the accountability that ought to attach to that delegation. [1] Ronald Dworkin's (1977) often-quoted image captures the bounding relation exactly: discretion "is like the hole in a doughnut" and "does not exist except as an area left open by a surrounding belt of restriction." [4]
Equivalent framings:
- Bounded latitude to choose among permissible actions
- A gap deliberately left open inside a rule system
- Delegated judgment exercised by a situated agent
- Choice not fully determined by a fixed rule
- The trade of rule-predictability for case responsiveness
- Reviewable authority within another's frame
- The doughnut hole defined by a surrounding belt of restriction
The structural insight is robust across substrates: a sentencing judge choosing within a statutory range, a manager acting within delegated spending authority, a clinician selecting among guideline-permitted treatments, a caseworker deciding which applicants merit closer review, and a human operator authorized to override an automated system all exhibit the same relation—an agent applying judgment in a region the governing rule intentionally did not fully specify. Dworkin further distinguished weak discretion (judgment required because a standard is imprecise, e.g., "take the most experienced sergeants") from strong discretion (the agent is simply not bound by any standard the authority has set on the matter), a distinction that sharpens what kind of gap is in play. [4]
What It Is Not¶
Discretion is not mere rule-vagueness or sloppy drafting. A vague rule produces uncertainty as a side effect; discretion is the deliberate reservation of a choice to an agent, a designed feature of the system rather than a defect. The whole point is that the rule-maker could have specified further and chose not to, judging the agent better positioned to decide the particular case. [1] Vagueness can create de facto discretion, but the prime names the structural arrangement of delegated choice, not the linguistic imprecision that sometimes accompanies it.
Nor is discretion the same as arbitrariness. Arbitrariness is choice unconstrained by reasons; discretion is bounded choice that is supposed to be exercised for reasons appropriate to the purpose of the delegation. The standard administrative-law formulation is that discretion must be exercised reasonably, within its purpose, and not capriciously—an official who acts arbitrarily has abused discretion, not exercised it. Davis's central reform argument was precisely that unstructured discretion drifts toward arbitrariness unless it is confined, structured, and checked. [3] The prime therefore carries an implicit accountability dimension: the gap is delegated, not abandoned.
Discretion is also not unlimited authority. It is latitude within bounds. An agent who exceeds the permissible range is not exercising more discretion; they are acting outside the grant entirely. The bounds are constitutive: remove them and you no longer have discretion but raw power.
Finally, discretion makes no claim that the delegated judgment will be correct or good. The prime describes a structural arrangement—choice reserved to a situated agent—not a guarantee about outcomes. Discretion can be exercised wisely or badly, fairly or with bias; the same gap that enables responsiveness to particulars also enables inconsistency and discrimination. Naming the structure does not endorse any particular exercise of it.
Broad Use¶
Law: Judges exercise sentencing discretion within statutory ranges; prosecutors exercise charging and plea discretion in deciding whom to charge and with what. The defining tension between rule-bound consistency and case-sensitive judgment is foundational to jurisprudence, with Dworkin and Hart staking out competing accounts of how much genuine discretion judges actually possess. [4]
Public administration: Michael Lipsky's (1980) "street-level bureaucrats"—police officers, social workers, teachers, clinicians at the front line—convert abstract policy into concrete action through on-the-spot judgment the rulebook cannot fully specify, so that their accumulated discretionary decisions effectively become the policy as citizens experience it. [5]
Management and organizations: Delegated authority gives managers latitude to act within defined limits without seeking approval for each decision; the design question is how much judgment to delegate, to whom, and with what reporting and review—a problem of distributing decision rights through the hierarchy.
Medicine: Clinical judgment fills the space that practice guidelines deliberately leave open for the individual patient. Guidelines bound the reasonable range; the physician selects within it based on the particulars no protocol can fully anticipate.
Monetary policy (non-obvious): The "rules versus discretion" debate in macroeconomics—crystallized by Kydland and Prescott (1977)—asks whether a central bank should bind itself to a fixed policy rule or retain the freedom to judge each situation. Their time-inconsistency result showed that retained discretion can be worse than commitment, because the ability to re-optimize each period undermines the credibility that makes policy effective. [6]
Artificial intelligence: Human-in-the-loop override reserves discretion for a person wherever automated rules are judged insufficient, and the contemporary design problem is exactly where to draw that boundary and how to keep the reserved human judgment meaningful rather than a rubber stamp.
Clarity¶
Naming discretion lets practitioners see a system's flexibility as a designed quantity—how much judgment is delegated, to whom, within what bounds, and subject to what review—rather than as mere rule-vagueness or administrative slack. [1] It sharpens the perennial tension between consistency (rules treat like cases alike) and responsiveness (discretion adapts to particulars rules cannot foresee). Once a system designer sees discretion as a dial rather than an accident, the question shifts from "is this rule clear enough?" to "where should this rule stop specifying, and who is best placed to decide the rest?"
This reframing also clarifies what kind of problem one is facing. A complaint that decisions are inconsistent may reflect too much discretion; a complaint that decisions are rigid and unjust in particular cases may reflect too little. The prime gives a common vocabulary for diagnosing both as movements along the same axis—the placement of the boundary between mechanized rule and delegated judgment—rather than as unrelated grievances.
Manages Complexity¶
Discretion bounds where judgment is allowed and thereby lets a rule system stay finite. Rather than attempting to enumerate every contingency in advance—an impossible task given the open-ended variety of particular cases—the system delegates the residual variety to local agents who confront the particulars directly. This is a form of complexity management by deferral: the rule absorbs the regularities and hands off the irregularities to a situated decision-maker who has information the rule-maker lacked. Herbert Simon's (1947) analysis of administrative behavior framed this as the rational necessity of decentralizing decisions to where the relevant facts and bounded-rational capacity actually reside. [7]
The complexity does not vanish; it is relocated to the point in the system best positioned to handle it. A statute cannot anticipate every defendant; a sentencing range plus judicial discretion handles the variety without an infinite rulebook. A protocol cannot anticipate every patient; a guideline plus clinical judgment does. The structural cost of this relocation is that the system's behavior becomes harder to predict and audit, which is precisely why the prime pairs delegation with reviewability.
Abstract Reasoning¶
Recognizing discretion supports reasoning about the rules-versus-discretion trade-off as a general design problem: more rule buys consistency, transparency, and resistance to bias and capture, at the cost of responsiveness; more discretion buys adaptiveness and local knowledge, at the cost of predictability and the risk of arbitrariness. [1] It also supports reasoning about accountability—delegated judgment requires oversight, reason-giving, or audit to prevent the gap from becoming a license—and about where to place the boundary between mechanized rule and human judgment in any given system.
The same abstraction enables counterfactual transfer across domains. If structured discretion with reviewability tames arbitrariness in administrative law, could the same structure (bounded override plus audit trail) tame it in automated decision systems? If time-inconsistency makes retained discretion costly for a central bank, does an analogous credibility problem afflict any agent who could renege on a commitment whenever re-optimizing seems locally attractive? These are not literal transfers, but the structural logic—how a gap left open interacts with predictability, credibility, and accountability—travels.
Knowledge Transfer¶
The legal insight that discretion must be structured and reviewable to avoid arbitrariness—Davis's program of confining, structuring, and checking discretionary power—transfers directly to other governance substrates. [3] In AI governance it reappears as bounded human override with audit trails and contestability requirements; in management it appears as delegation with defined limits, spending thresholds, and reporting lines; in medicine it appears as guideline-bounded clinical autonomy with peer review and morbidity-and-mortality conferences. Wherever a system chooses how much judgment to delegate, the accumulated legal-administrative experience about how to keep that delegation from sliding into arbitrariness becomes available.
The transfer runs in both directions. The economic lesson about time-inconsistency—that the freedom to re-decide can destroy the credibility that makes a policy work—imports back into law and management as an argument for binding rules and pre-commitment where credibility matters more than flexibility. The shared structure (a deliberately reserved gap and its consequences) is what makes the cross-domain lessons more than analogy.
Examples¶
Formal/abstract¶
Judicial sentencing within a statutory range: A criminal statute prescribes that an offense is punishable by two to ten years' imprisonment. The legislature has fixed the boundaries but deliberately left the choice within them to the sentencing judge, who weighs the particulars of the offense and offender that no statute could fully enumerate. This is discretion in its paradigm form: a bounding rule (the range), a reserved gap (any point within two to ten years), a situated agent (the judge with the case before her), bounded choice (she cannot impose one year or twenty), and reviewability (an appellate court can correct an abuse). The structural-sentencing-reform movement responded to perceived over-delegation: sentencing guidelines narrowed the gap to reduce inter-judge disparity, trading responsiveness for consistency. Mapped back: The example exhibits every element of the signature. The reform debate—guidelines versus judicial freedom—is literally a debate about where to place the rule/judgment boundary, the core abstraction the prime names. Tightening the range is not the elimination of discretion but the resizing of the gap, which is exactly how the prime predicts such reforms should be read.
Rules versus discretion in monetary policy: A central bank could bind itself to a fixed feedback rule (set the interest rate as a mechanical function of inflation and output) or retain discretion to judge each situation afresh. Kydland and Prescott showed that the discretionary regime is time-inconsistent: knowing the bank will re-optimize each period, private actors anticipate inflationary temptation, and the resulting equilibrium is worse than under credible commitment. Mapped back: Here the prime's accountability and predictability dimensions become formal. The "gap left open"—the bank's freedom to re-decide—is the very thing that destroys credibility, demonstrating that discretion is not a free good: the reserved choice interacts with how other agents anticipate the choice. This is the rule/judgment trade-off rendered as a precise economic result, and it shows the prime is substrate-portable beyond its legal home.
Applied/industry¶
Delegated authority in a corporate hierarchy: A company grants regional managers authority to approve expenditures up to $50,000 without head-office sign-off, escalating anything larger. The threshold is the bounding rule; the manager's latitude beneath it is the reserved gap; the manager is the situated agent who knows local conditions; the cap and the escalation rule are the bounds; and quarterly expense review is the reviewability that keeps the delegation honest. Designing this scheme is exactly the question "how much judgment to delegate, to whom, within what bounds." Mapped back: The corporate spending limit is the doughnut's belt of restriction; the manager's freedom beneath it is the hole. Lower the cap and the hole shrinks; raise it and the hole grows. The example shows the prime is not a legal artifact but a general design pattern for distributing decision rights in any rule-governed organization.
Human-in-the-loop override in automated decision systems: A content-moderation or loan-approval pipeline applies automated rules to the bulk of cases but reserves a class of borderline or high-stakes decisions for a human reviewer who may override the machine. The automated policy bounds the routine; the reserved override is the delegated gap; the reviewer is the situated agent; the override is constrained to defined categories; and logging plus audit provides reviewability. The contemporary failure mode is hollow discretion—an override nominally reserved for a human who, under time pressure and automation bias, simply ratifies the machine—precisely the "rubber stamp" that converts delegation into a fiction. Mapped back: This is Davis's structuring-and-checking problem reincarnated in software: the gap must be real and reviewable or it collapses into either automated arbitrariness or unaccountable human override. The prime's insistence that discretion is delegated-not-abandoned judgment explains exactly why a token human in the loop fails to deliver the accountability the structure promises.
Structural Tensions¶
T1: Discretion trades consistency for responsiveness, and the trade cannot be escaped. Every increment of delegated judgment that lets an agent adapt to a particular case is, by the same token, an increment of inconsistency across cases. A system cannot maximize both treating-like-cases-alike and responding-to-each-case-on-its-merits; the two goods are in structural tension because they make opposite demands on the same gap. Designers who tighten the rule to gain consistency lose case-sensitivity, and those who widen the gap to gain responsiveness invite disparity. The prime does not resolve this tension; it names it as the axis along which any such system must be positioned.
T2: The bounds that constitute discretion are themselves contestable. Discretion exists only inside a belt of restriction, but where that belt sits is often disputed. An agent may believe a choice falls within their delegated authority while an overseer believes it exceeds it; the "permissible range" is frequently clear only at its center and fuzzy at its edge. This means disputes about discretion are often not about how judgment was exercised but about whether the matter was within the agent's grant at all. The structural gap has a boundary, but the boundary is itself a site of conflict, and the same act can be characterized as legitimate discretion or as overreach depending on where one draws the belt.
T3: Delegating judgment relocates complexity but also relocates the opportunity for bias and capture. Pushing decisions to a situated agent puts them where the relevant facts are, but it also puts them where the agent's particular interests, prejudices, and vulnerabilities to influence operate. The same proximity to the particular case that makes discretion valuable makes it a channel for discrimination and favoritism that a uniform rule would have suppressed. Street-level discretion can humanize a bureaucracy or it can let an officer's bias determine who gets stopped; the structure that enables responsiveness is the identical structure that enables abuse, which is why oversight is constitutive rather than optional.
T4: Reviewability is necessary to discipline discretion yet partly defeats its purpose. Discretion is delegated because the situated agent has knowledge or judgment the rule-maker lacks; but to prevent arbitrariness the agent must give reasons that an overseer can evaluate. The more fully a decision must be justified in terms an outsider can review, the more it is pushed back toward articulable rule-like criteria—eroding the very latitude that motivated the delegation. Demanding complete justification collapses discretion into rule-application; demanding none invites arbitrariness. Every accountability mechanism therefore sits in tension with the autonomy it polices, and systems must choose how much of the gap to expose to review.
T5: Retained discretion can destroy the credibility that makes commitment valuable. The time-inconsistency tension is that the freedom to re-decide each period—the essence of retained discretion—can be anticipated by other agents and priced in, producing a worse equilibrium than a binding rule would. An agent who could renege whenever re-optimizing looks locally attractive cannot credibly promise not to, so rational counterparties act on the expected reneging. Here discretion is not merely costly in predictability; it actively undermines outcomes that depend on others trusting a commitment. The tension is that the option to exercise judgment later can be worth less than the ability to foreclose it now.
T6: The same act can be classified as exercising discretion or as abusing it, with no neutral vantage. Because discretion is bounded choice exercised for appropriate reasons, the line between a legitimate exercise and an abuse depends on a contested judgment about whether the reasons were appropriate to the delegation's purpose. A lenient sentence is sound mercy or unjustified disparity; a regulator's forbearance is sensible flexibility or unaccountable favoritism. Observers with different views of the delegation's purpose will classify the identical decision differently, and there is no view from nowhere that settles it. This makes "abuse of discretion" an inherently interpretive charge rather than a factual measurement, which is why its adjudication is itself entrusted to further discretionary judgment.
Structural–Framed Character¶
Discretion is a framed prime on the structural–framed spectrum: it names the arrangement in which a rule system deliberately leaves a gap open and delegates the choice within it to a situated agent's judgment, rather than fully determining the outcome by a fixed rule. The commitment is not vagueness but designed delegation — the rule fixes the outer boundaries while reserving the selection among permissible actions to the agent facing the case.
The concept comes straight from law and governance, and that origin saturates it: its vocabulary of rules, judgment, authority, and reviewability travels intact, it carries clear normative weight about how delegated latitude ought to be exercised, and it cannot be stated without presupposing agents, rules, and institutions of authority. Naming a situation discretion recasts it as a rule-and-judgment system rather than merely spotting a structure already there, as when a judge sets a sentence within a statutory range or an administrator decides a case the regulations leave open. On every diagnostic, it reads framed.
Substrate Independence¶
Discretion is a moderately substrate-independent prime — composite 3 / 5 on the substrate-independence scale. The structural commitment is clean — a deliberate gap left open inside a rule system, delegating bounded judgment to a situated agent — and it transfers across legal (judges), organizational (delegated authority), medical (clinical judgment), and computational (bounded AI override) settings. The catch is that all of these are governance and decision substrates populated by rule-following agents; there is no physical, biological, or formal-systems instance. The rule-versus-judgment tradeoff is real structure, but it stays inside the social-cognitive governance family.
- Composite substrate independence — 3 / 5
- Domain breadth — 3 / 5
- Structural abstraction — 4 / 5
- Transfer evidence — 3 / 5
Relationships to Other Primes¶
Parents (1) — more general patterns this builds on
-
Discretion presupposes Authority
Discretion presupposes authority because the rule system's deliberate leaving-open of a gap and assignment of the choice within it to a situated agent requires that the agent hold legitimate power to make binding decisions in that space. Without authority's recognized right to bind, the discretionary choice is mere opinion that others may ignore. Discretion is authority operating in the rule's silences: the rule fixes the boundary, and the authority delegated to the agent fills the interior with case-specific judgment that binds through the agent's recognized standing.
Children (1) — more specific cases that build on this
-
Equity is part of Discretion
Equity is a constituent piece of discretion. The discretionary gap that rule systems leave open is filled in part by equitable reasoning: when strict rule application would yield an unjust or inadequate result, the decision-maker examines specific facts and tailors the remedy to achieve fairness in the particular case. Equity is not a separate mechanism but a substantive content layer within discretion -- the principled-flexibility move Aristotle's epieikeia introduces that turns delegated judgment into justice-tracking rather than arbitrary preference.
Path to root: Discretion → Authority
Neighborhood in Abstraction Space¶
Discretion sits among the more crowded primes in the catalog (29th percentile for distinctiveness): several abstractions describe nearly the same structure, so a description that fits it will tend to fit its neighbors too — transporting it usually means disambiguating within this family rather than landing on it exactly.
Family — Rules, Enforcement & Property (11 primes)
Nearest neighbors
- No One Is Above the Rules — 0.83
- Role — 0.80
- Modal Reasoning — 0.80
- Authority Delegation Under Uncertainty — 0.80
- Conflict of Interest — 0.80
Computed from structural-signature embeddings · 2026-05-29
Not to Be Confused With¶
Discretion must be distinguished from Sovereignty, which is final, supreme authority over a domain—the power that recognizes no superior and is, in the limit, unreviewable. Sovereignty answers the question "whose word is final here?"; discretion answers "within whose rules, and within what bounds, may this agent choose?" The two are nearly opposite in their relation to a surrounding frame. Sovereignty is the frame: it sets the bounds for everyone else and is not itself bounded by a higher rule within the domain. Discretion is constitutively inside a frame someone else set—it is bounded, delegated, and typically reviewable latitude. A sovereign legislature has no discretion in the prime's sense over its own domain because there is no surrounding belt of restriction; it has plenary power. A judge applying that legislature's statute has discretion precisely because the statute bounds her. The relationship is hierarchical: sovereignty is the source from which delegated discretion flows downward, and the defining feature of discretion—the doughnut's belt of restriction—is exactly what sovereignty, at the apex, lacks. Conflating them obscures the central structural fact that discretion is granted authority while sovereignty is originating authority.
Discretion is also distinct from Precedent, which constrains present judgment by binding it to the pattern of prior decisions. The two operate on the same raw material—a present agent deciding a case—but in opposite directions. Precedent closes the gap: it imports the resolution of past cases as a constraint on the present one, treating like cases alike by force of prior decision and thereby reducing the agent's freedom to choose afresh. Discretion is the gap: it is the freedom that rules and prior decisions deliberately leave open for the agent's situated judgment. Where precedent binds, discretion is correspondingly diminished; where precedent runs out or is distinguished, discretion reappears. This makes them complementary measures of the same quantity from opposite sides: a domain densely covered by binding precedent affords little discretion, while a domain of first impression affords much. The relationship is dynamic, because today's discretionary decision can itself become tomorrow's binding precedent—the exercise of discretion can generate the very constraint that narrows future discretion. They are not the same prime: precedent is a source of constraint drawn from prior choices, whereas discretion is the reserved freedom that exists in the space such constraints have not occupied.
Finally, discretion is unrelated to Taboo, which is an absolute, often sacred prohibition—a categorical "never" that admits no weighing or case-by-case judgment. Taboo and discretion are structurally opposite in how they treat choice. Taboo forecloses choice entirely within its scope: the prohibited act is simply off the table, not a permissible option to be selected against. Discretion opens choice within bounds: it presupposes a range of permissible actions among which the agent may select. A taboo is not a tight bound on discretion; it is the removal of the option from the permissible set altogether, so that no exercise of judgment could legitimately reach it. Where a discretionary range says "choose any point between two and ten years," a taboo says "this is never to be done, by anyone, under any judgment." The two can coexist in a single system—a judge has discretion within a sentencing range but faces absolute prohibitions on, say, accepting bribes to set it—but they are different structural devices: discretion delegates bounded judgment, while taboo withdraws a region from judgment entirely. Treating taboo as merely a strong form of bounded discretion misses that taboo denies the very thing discretion grants, namely a live choice.
Solution Archetypes¶
No catalogued solution archetypes reference this prime yet.
Notes¶
Discretion is frequently conflated with "flexibility," but the two are distinct. Flexibility is a capacity for change or adjustment; discretion is a specific structural arrangement in which a bounded choice is reserved to an agent by a rule that could have specified further. A flexible system might achieve its flexibility through many mechanisms; discretion is the particular mechanism of delegated, bounded judgment inside a fixed frame.
The prime carries an implicit normative companion it does not itself supply: the expectation that delegated judgment will be exercised for reasons appropriate to the purpose of the delegation. Davis's reform program—confine, structure, check—presupposes that the bare structural gap is dangerous unless disciplined. Practitioners who treat discretion as simply "freedom for the agent" miss that the prime's stability depends on the accountability scaffolding that surrounds the gap. Strip away review and reason-giving, and discretion degrades into the arbitrariness it is supposed to be distinguished from.
Dworkin's weak/strong distinction is worth keeping in view because the two carry different stakes. Weak discretion (judgment required because a standard is imprecise) is compatible with the view that there is, in principle, a right answer the agent should reach; strong discretion (no governing standard binds the agent on the matter) raises the harder question of whether the agent is genuinely creating the resolution rather than discovering it. Much of the Hart–Dworkin debate turns on how much strong discretion judges really have, and the prime is neutral between these positions—it names the structural gap without settling whether the gap is filled by discovery or by choice.
The substrate range is real but bounded: discretion travels cleanly across legal, administrative, managerial, medical, and computational governance settings, but all of these are decision substrates populated by rule-following agents. There is no physical, biological, or purely formal-systems instance of discretion, which is why its substrate-independence is scored moderate rather than high. The rule-versus-judgment trade-off is genuinely structural, but it stays within the social-cognitive governance family.
References¶
[1] Galligan, D. J. (1986). Discretionary Powers: A Legal Study of Official Discretion. Oxford: Clarendon Press. Canonical legal-theoretic study of official discretion: develops discretion as a designed delegation of bounded choice within a structure of standards, the bounding-rule/reserved-gap/situated-agent/reviewability relation, its distinction from mere vagueness, and the general consistency-versus-responsiveness trade-off. ↩
[2] Hart, H. L. A. (1961). The Concept of Law. Oxford University Press. Analytical-jurisprudence treatment of legal systems as rules of recognition, change, and adjudication; develops adjudication as the rule-bound institutional practice through which secondary rules apply primary rules to particular cases—foundational for understanding procedural fairness as a constituent of legal-system legitimacy. ↩
[3] Davis, K. C. (1969). Discretionary Justice: A Preliminary Inquiry. Baton Rouge: Louisiana State University Press. Defines discretion as existing wherever effective limits on an official's power leave a free choice among courses of action; argues unstructured discretion drifts toward arbitrariness and advances the confine–structure–check reform program that transfers across governance settings. ↩
[4] Dworkin, R. (1977). Taking Rights Seriously. Cambridge, MA: Harvard University Press. Source of the doughnut-hole image (discretion exists only as an area left open by a surrounding belt of restriction), the weak-versus-strong discretion distinction, and the challenge to Hart over how much genuine discretion judges actually possess. ↩
[5] Lipsky, M. (1980). Street-Level Bureaucracy: Dilemmas of the Individual in Public Service. New York: Russell Sage Foundation. Argues that frontline public-service workers convert vague policy into concrete action through on-the-spot discretionary judgment, so that their accumulated decisions effectively become the policy as citizens experience it. ↩
[6] Kydland, F. E., & Prescott, E. C. (1977). Rules rather than discretion: The inconsistency of optimal plans. Journal of Political Economy, 85(3), 473–491. Time-consistency problem in macroeconomic policy: optimal plans formulated at distance are abandoned at the moment of execution; institutional analogue of individual preference reversal in long-horizon corporate commitments. ↩
[7] Simon, H. A. (1947). Administrative Behavior: A Study of Decision-Making Processes in Administrative Organization. Macmillan. Introduces the zone of acceptance (and the related zone of indifference, after Barnard): the pre-defined region within which a subordinate accepts orders or acts on delegated authority without case-by-case approval — the canonical organizational formulation of pre-positioned scope. ↩