npbuilds Values Excavator — The Dowser
Find the hidden values flowing beneath the surface of any decision. Every proposal, policy, argument, and choice embeds value judgments — assumptions about what matters, whose interests count, and what trade-offs are acceptable. Most of these are invisible to the person making them.
The values-excavator is a moral analog of the assumption-excavator in the logic subdomain. Where that skill finds hidden premises, this one finds hidden values. They complement each other: assumptions are epistemic; values are moral.
Input
From the ethics director or directly:
- The proposal, decision, policy, or argument to excavate
- Context: whose values? (the user's, an organization's, a policy's, an argument's)
- Depth: surface (identify the top 3-5 implicit values) or deep (comprehensive value mapping)
Process
Step 1 — Identify Explicit Values
Before excavating what's hidden, map what's visible. What values are stated?
- Mission statements, stated goals, explicit priorities
- "We believe in...", "Our priority is...", "What matters most is..."
- Criteria explicitly used for the decision
These are the above-waterline values.
Step 2 — Excavate Implicit Values
Dig for values hidden in the structure of the decision:
Values in Metrics
What you measure reveals what you value.
- What metrics are being used or proposed? (Revenue, engagement, satisfaction, safety, equity)
- What metrics are absent? (Often more revealing than what's present)
- How are trade-offs between metrics resolved?
Example: Optimizing for "user engagement" implicitly values time-on-platform over user wellbeing, advertiser revenue over information quality, and behavioral manipulation over autonomy — even if none of this is stated.
Values in Defaults
The default option reveals the preferred outcome.
- What happens if no one takes action? Who benefits from inertia?
- What's treated as the "normal" baseline against which alternatives are judged?
- Who bears the burden of proof — those proposing change or those defending the status quo?
Example: "Innocent until proven guilty" values individual liberty over collective security. "Precautionary principle" values safety over innovation. Neither is objectively correct — they're value choices.
Values in Scope
Who and what are included reveals whose interests count.
- Whose interests are considered in the analysis?
- Whose interests are excluded or invisible?
- What time horizon is assumed? (Short-term profit vs. long-term sustainability)
- What geographic scope? (Local community vs. global impact)
Example: A cost-benefit analysis that counts only shareholder value implicitly values shareholders over employees, customers, and communities.
Values in Framing
How the question is posed reveals what counts as a solution.
- Is this framed as an efficiency problem? (Values productivity)
- Is this framed as a rights problem? (Values autonomy)
- Is this framed as a safety problem? (Values security)
- Is this framed as a fairness problem? (Values equity)
Example: "How do we reduce healthcare costs?" (efficiency frame) vs. "How do we ensure everyone has access to healthcare?" (rights frame) — same domain, completely different value landscapes.
Values in Omission
What's not discussed reveals what's taken for granted.
- What obvious considerations are missing from the analysis?
- What stakeholders are not mentioned?
- What potential harms are not acknowledged?
- What alternative approaches are not considered?
Step 3 — Map Value Tensions
Identify where implicit values conflict with each other or with stated values:
| Tension Type | Pattern | Example |
|---|---|---|
| Stated vs. revealed | Organization says X but does Y | "We value work-life balance" + mandatory 60-hour weeks |
| Value vs. value | Two legitimate goods in competition | Privacy vs. security; innovation vs. safety; efficiency vs. equity |
| Short-term vs. long-term | Immediate benefits vs. future costs | Quarterly earnings vs. long-term sustainability |
| Individual vs. collective | One person's good vs. the group's | One employee's advancement vs. team cohesion |
| Means vs. ends | The process conflicts with the goal | Achieving equity through inequitable methods |
Step 4 — Trace Value Origins
Where do these values come from? Understanding the source helps the user evaluate them:
| Source | Description | Stability |
|---|---|---|
| Institutional | Embedded in organizational structure, incentives, or culture | Highly stable; persists even when individuals change |
| Professional | Derived from professional norms and training | Stable within a field; varies across fields |
| Cultural | Absorbed from broader cultural context | Often invisible to those within the culture |
| Personal | Individual beliefs and experiences | Variable; may conflict with institutional or professional values |
| Ideological | Part of a coherent political or philosophical worldview | Stable but often unexamined |
Step 5 — Assess Value Coherence
Are the excavated values coherent — can they all be satisfied simultaneously, or are trade-offs inevitable?
- Compatible: Values can all be served (rare for non-trivial decisions)
- Tension: Values pull in different directions but compromise is possible
- Contradiction: Values are mutually exclusive in this context — choosing one means sacrificing another
- Hierarchy needed: Values are compatible in general but this specific situation forces a ranking
Output
VALUES EXCAVATION
─────────────────
Subject: [the proposal/decision/policy/argument analyzed]
Whose values: [individual/organization/policy/argument]
Stated Values:
- [explicitly stated value] — Source: [where stated]
...
Implicit Values Excavated:
1. [Value] — Found in: [metrics/defaults/scope/framing/omission]
Evidence: [specific feature of the decision that reveals this value]
Visibility: [hidden / semi-visible / visible but unchallenged]
2. [Value] — Found in: [category]
...
Value Tensions:
T1: [Value A] vs. [Value B] — Type: [stated-vs-revealed / value-vs-value / etc.]
In this decision: [how the tension manifests]
Resolution requires: [what the user would need to decide]
T2: ...
Value Coherence: [compatible / tension / contradiction / hierarchy needed]
Most Critical Finding:
[The single most important implicit value or tension the user should be aware of]
Recommended Next Steps:
- [e.g., "Run dilemma-analyzer on tension T1 to see how ethical frameworks weigh in"]
- [e.g., "Route to stakeholder-mapper to check whose values are missing"]
Error Handling
No hidden values found: Rare for any non-trivial decision. If the proposal is genuinely transparent about its values, report this as a positive finding. More likely, dig deeper — values in omission are the hardest to see.
User is emotionally attached to the values excavated: Handle with care. "Your decision reflects a value of X" is descriptive, not accusatory. Values are not flaws — they're choices, and making them explicit gives the user more control, not less.
Values are culturally specific: Note cultural context without universalizing. "In this cultural context, X is valued" is better than "X is good/bad." The excavator surfaces values; it doesn't rank them.
Too many values to map: Focus on values that are load-bearing for the specific decision at hand. Group minor values into clusters ("productivity values," "equity values") and detail only the ones in active tension.