Axioms & Theorems
NDT proposes an account of how normative judgments form, what makes some of them moral, and why they can conflict. Explore the starting claims, then follow what they support.
These axioms include NDT’s original claims about moral cognition. Their consequences follow conditionally; the claims themselves remain open to argument and evidence.
Select an axiom to read its full claim and find the consequences that use it. The three questions organize the explanation; they are not three processing stages.
Withholding means this argument can no longer establish its conclusion. It does not establish the opposite conclusion.
Formal Specification 1.6.5 · Scope and Commitments 1.18 · Proposition Registry 2.11 · Empirical Audit 2.4
Normative Detachment Theory
Normative Detachment Theory (NDT) describes how people form judgments about what should or should not happen. It traces how a person’s understanding of a situation, and what matters within that understanding, support a judgment. The theory explains differences between moral and nonmoral judgments through the processes that produce them.
Mina is preparing breakfast for her young daughter. The milk smells sour, and Mina believes it could make her daughter ill. There is fresh milk in the refrigerator. She reaches a familiar judgment: “I should discard the milk.”
Concern for her daughter could guide Mina’s judgment. Or a food-handling procedure alone could supply her reason. The parent, situation, and sensible decision stay the same—but would Mina be arriving at the judgment in the same way?
Begin with concern, then compare a judgment based on procedure alone.
NDT’s larger wager is that familiar differences in judgment share an underlying cognitive organization. Explore the argument: What makes a judgment moral?
NDT examines how a person’s understanding supports a judgment—and what makes some judgments moral. The breakfast comparison asks what actually helps Mina form her judgment, not just whether she reaches a sensible answer.
We can agree about what should be done while reaching that conclusion for different reasons. We can also understand the same situation differently and disagree about what follows.
In another imagined family, Elena reaches two conflicting judgments about her daughter Mina’s journey. Follow Elena and Mina’s two judgment paths in Theory in Action.
A parent notices spoiled milk while preparing a child’s breakfast. “I should discard the milk” could follow from concern for the child, or from a food-handling procedure alone. The sensible answer can stay the same while the reasons differ.
These ideas depend on one another, but they are not three timed stages. They identify the materials and relationships NDT uses to explain one particular judgment.
NDT begins with the interpretation active when the judgment forms. An observer’s extra knowledge is not automatically part of that interpretation. NDT distinguishes considerations that helped form the judgment from those that were merely available.
A normative judgment requires a difference in what matters that is active in forming the judgment. That difference may be explicit, attributed, carried by how a representation functions, or derived from those sources.
Constitutive evaluativity is the narrower claim that a representation can carry evaluation partly through the role that makes it the representation it is. It is not another name for the broader requirement that normative judgment needs evaluation.
A normative judgment is moral, in NDT’s sense, when a represented moral relationship actually helps organize the evaluation used to form it. Moral cognition is therefore one kind of normative cognition.
Travel into the relationship diagram while a root-shaped view keeps the whole structure in sight.
To follow a relationship’s role in forming a particular judgment, choose Theory in Action in the module navigation. The Demonstrations button also gathers judgment paths, the relationship tour, and the full removal tests.
What follows from NDT’s starting claims? Explore ten starting claims in three groups, with the premises and limits of each consequence made explicit.
What must a moral relationship contain, and when does it help form a judgment? Inspect the three ingredients, or start with the contrasting reasons for discarding milk.
How do particular ideas explain guidance, concern, or contribution? Compare claims within their intellectual lineages, then inspect the complete source records.
Guided tours
The specification supplies exact definitions and the arguments linking them. Its sections separate judgment formation, moral classification, challenged support, and conflict between judgments. Open the section guide when you want to trace a claim into the formal account.
Read further
Begin with the Article, then choose the companion that fits your question. The formal authorities and supporting audits remain available for readers who need exact definitions, claim status, or project history.
Download the reader documents and boundary paper · PDF, Word, and illustrated sources
The breakfast comparison · relationship structure
Mina still sees her daughter as someone whose welfare matters, someone personally connected to her, and someone she can help. NDT calls the three kinds of represented information builders: they jointly constitute the relationship. Whether the relationship helps form “I should discard the milk” is a further question.
The map shows NDT’s proposed structure, not measured mental events. Select a node for its definition; the breakfast connection appears in the explanation panel.
Structure of a Moral Relationship
This module explores the structure of a moral relationship in NDT and how that relationship can contribute to forming a moral judgment.
Caro is a physician deciding how to respond to Eli, the patient assigned to her care. Caro can contact Eli and considers, “I should attend to my patient.” Eli’s welfare matters to Caro in its own right. Caro understands Eli as her patient and herself as able to respond. In the imagined case, those considerations help form her judgment. The map explains what the relationship contains and the further work it must do to give the judgment a moral basis.
Builders are the three required components of the relationship: value in itself, perceived agency, and positioning of the object relative to the subject.
Benders affect what follows from a relationship, such as the priority of a response. The controlled examples preserve the exact relationship core and its actual contribution. The formal term “strict bender” marks those preservation requirements; it does not mean that every urgent consideration works that way.
A gate marks a required condition, not a timed mental step. The represented features must satisfy the stated criteria.
Colors identify explanatory roles, not good and bad judgments. The same roles recur in the relationship tour and ingredient comparisons. Names and connections carry the distinctions even without color; an active highlight marks the current selection or demonstration step.
Select a node for its definition, significance, removal test, and formal status.
A hypothetical illustration of NDT, not experimental evidence. The steps explain different roles, not a measured sequence of mental events.
The example stipulates a moral relationship: positive inherent value with negative inherent value absent, an operative physician–patient role bridge, relevant perceived agency, and the required subject and object bindings contributing together in the judgment’s actual support. Positive value is an example, not a universal requirement. A professional title alone does not establish the role bridge.
The penalty-only comparison stipulates that no moral relationship, including a self-directed relationship, contributes. Self-interest and professional rules do not determine a judgment’s type merely by being named.
Urgency qualifies as strict bending only when the exact relationship core is preserved, the relationship actually contributes in both matched cases, and the separate downstream influence meets the required contribution tests. Strict bending does not, by definition, preserve the final judgment or complete moral classification.
Swipe the diagram horizontally to inspect every node; select a node for readable detail below.
Can a model distinguish a relationship that helps form a judgment from a relationship that is merely represented, while preserving the difference between the relationship and influences such as urgency? The model need not use NDT’s labels or divide the work in the same way. The Atlas illustrates NDT’s proposed distinctions; the illustration does not establish that a rival cannot represent them.
Nuances of Normative Judgment
Unratified v52 candidateThis module explores distinctions NDT draws among normative judgments by varying the representations and reasons that help form them.
People distinguish concern for another person from simply following a procedure—even when both lead to a sensible decision. A theory of normative cognition must explain more than the outward answer. NDT locates the difference in how a particular judgment is formed.
Begin with one ordinary situation
Mina is preparing breakfast for her young daughter. Before pouring the milk over her daughter’s cereal, Mina notices a sour smell. She believes the milk has spoiled and could make her daughter ill. There is fresh milk in the refrigerator, so replacing the carton would be easy. Mina reaches a familiar judgment: “I should discard the milk.”
Start with an equally familiar reason. Her daughter’s welfare matters to Mina for her daughter’s own sake; Mina understands their personal connection and knows she can protect her from harm. Those considerations help Mina judge that she should not serve the spoiled milk.
Or start with a different reason: a food-handling procedure. Imagine the same parent, the same spoiled milk, and the same decision, but suppose that the procedure alone supplies the reason for the judgment. Would the outward answer tell us what made the difference?
The arrows show how the represented relationship helps form Mina’s judgment.
NDT classifies the judgment as moral because the represented relationship helps form it—not simply because discarding spoiled milk is sensible.
NDT classifies the judgment as nonmoral under the stated condition: no represented moral relationship helps form the procedure-only judgment.
Mina has not stopped caring about her daughter, and the relationship has not been removed. The comparison changes whether that relationship contributes to the judgment. Calling the procedure-only judgment nonmoral does not condemn the action or identify a defect in Mina’s parenting. A procedure may handle the situation perfectly well.
Consider checking a carton’s date while putting groceries away. Someone can know that a family member’s health matters and nevertheless discard the carton solely because a storage rule says to do so. The breakfast comparison asks us to imagine that separation even with the daughter waiting for breakfast: the represented relationship remains available, but the procedure alone supplies the reason for the particular judgment.
Quick or habitual action does not establish that separation. Concern can operate without being consciously rehearsed, and a rule can express concern. The example specifies the reasons; it does not infer them from Mina’s speed, conduct, or later explanation.
No. Suppose concern and an independently sufficient food-handling rule both actually help Mina reach the judgment. Removing concern can leave the same answer supported by the rule. The remaining answer does not show that concern originally did nothing. Conversely, an available reason that played no part is not made a contributor merely because it could have supported the answer.
The two routes are specified possibilities, not observations of Mina’s hidden mental life. They make NDT’s classification rule inspectable: the same outward answer can be supported by different reasons. Deciding how an actual person’s judgment formed is a separate problem; outward behavior alone does not settle it.
The comparison preserves Mina as both the person judging and the represented person able to respond, with her daughter as the person affected. In the concern case, the relationship and those exact roles jointly shape the health-related reason and judgment. In the procedure-only case, no represented moral relationship contributes; retaining the displayed parent–child relationship is not enough to classify the judgment as moral.
Consult the Formal Specification, sections 4b.1–4b.2 for contribution and moral classification, and sections 4c.1–4c.2 for the parent-care and procedure-only cases. The diagram explains dependencies; it is neither a timed mental sequence nor an execution of the full formal model.
Change one represented feature while keeping the stated comparison conditions fixed. Compare the supported distinction before inspecting its formal record. The authored examples illustrate definitions; the examples are not new experimental evidence.
A committee asks Rae whether a former mentor should take a public role. Rae’s view of the mentor contains both positive and negative regard: the mentor appears good and bad in their own right within the same represented relationship, not merely useful or costly to Rae. Keep Rae and the mentor fixed while removing one evaluation or both. Neither profile, by itself, decides the appointment.
Both evaluations remain represented. They do not cancel into neutrality.
Hypothetical, definition-based schematic—not a measured result or an execution of the full model. Branches mean present or absent, not measured intensity.
Removing one component changes the profile. Removing both removes a required part of the tested moral relationship. Negative value alone does not mean evil, blame, punishment, or avoidance. A value profile alone does not determine whether the whole judgment is moral, nonmoral, or prudential; qualification and actual contribution require additional conditions.
Every result names what becomes unavailable, what is held fixed, and whether moral type or only a subtype or downstream effect changes.
1 Choose a test
2 Inspect the difference
The tests expose consequences of NDT’s current decomposition; they do not establish that every rival must use the same vocabulary. This model version’s positioning-route families—identity-mediated positioning, intimacy, and direct encounter—form an exact exhaustive disjunction and may be co-active. Relational Positioning is valence-neutral, and shared, contrastive, and complementary-role identity are forms of the identity-mediated family. By contrast, urgency, vulnerability, power, and sentience are nonexhaustive candidate bender sources whose roles must be demonstrated, not assumed.
NDT and other theories
A familiar decision gives different approaches a common point of comparison. Begin with the reasons for discarding spoiled milk, then inspect what a particular account explains—and what remains to be explained.
A parent notices sour-smelling milk while preparing a child’s breakfast. The parent believes it could make the child ill, has fresh milk available, and judges, “I should discard the milk.” Concern could help produce the judgment. Or imagine that a food-handling procedure supplies the entire reason, with no moral relationship contributing. The parent can still care about the child.
The specified contrast does not reveal anyone’s hidden reasoning. It gives the comparisons a common question: what would an account need to explain about the judgment beyond its outward answer?
Instead of treating a whole theory as a single match or disagreement, the review compares particular explanatory claims and mechanisms. Each encounter names the question, locates the source within its intellectual lineage, and explains the precise relation to NDT. A useful mechanism can be shared even when accounts disagree about what makes a judgment moral.
The selected encounters offer an entrance; the full catalog retains the exact comparisons and sources. NDT can represent reasoning associated with different moral philosophies without endorsing their verdicts or every claim made by their advocates.
Historical analysis tool. This earlier model grouped four explanatory tasks as “axioms.” Its rule check verifies its own encoded rules; it does not establish an independent axiom basis. Use Axioms & Theorems in the navigation above for the revised premise-to-consequence arguments.
Axioms and theorems
Earlier four-heading modelThe earlier rule-check results can be inspected below. The revised argument views distinguish a starting claim, an additional premise, a constructed case, and a consequence.
What these checks do: choose a set of results, then test which axioms are enough to derive every result in that set inside the small rule model used here. This does not prove that the axioms are empirically true or that the four headings generate every definition in the Formal Specification.
The trace follows declared dependency links from each separately tracked part of an axiom. It shows where that axiom participates; that reach is not itself a proof. NDT’s allowance for several local viewpoints follows from its descriptive scope and Polyphony definitions, not from another axiom.
The current structural audit finds 313 formal endpoint premises of several kinds across a 593-node / 1,909-edge graph. It verifies dependency closure and finite witnesses; it does not establish that this is the only natural or model-theoretically independent axiom basis. Read the independent basis audit ↗
Next step in two seconds. Select Pause to stay here.
Explore the full architecture: drag to pan, pinch or use zoom, select a node for details, and use arrow keys to navigate. Historical marks do not date current arrows. Worked cases illustrate the partial executable companion.