METHOD

How we research and classify.

Ends and Means starts with a familiar question, gives a plain answer, and keeps the evidence and uncertainty close enough to inspect. It separates what people wanted, what institutions promised, what happened in a particular setting, why it may have happened, and how anyone judges the result.

PUBLIC COMMITMENTS

The commitments behind each explanation

Keep claims close to their evidence

Substantial factual claims remain individually addressable and connect to the particular source consulted, an exact passage or table, and the source’s role in the claim.

Preserve differences and limits

Ideas, traditions, institutions, movements, labels, and bounded cases can overlap without becoming interchangeable. Place, period, actors, disagreement, and missing evidence limit each conclusion.

Use tools without surrendering judgment

AI may assist discovery, synthesis, drafting, and consistency checks. It is never a source or authority. People remain responsible for source fitness, framing, interpretation, wording, and publication.

Read the technical protocols

These repository documents give researchers and reviewers the detailed rules behind the public commitments.

Read the editorial principles

WHAT THE SITE DOES

One subject can involve several different things.

A name such as democracy, socialism, or monarchy may refer to an ideology or idea, a family of traditions, a proposed institution, a movement’s own description, or a label applied by someone else. A subject guide brings those meanings together without declaring them identical. It also keeps fictional depictions separate from evidence about the world.

The site records disagreement instead of choosing a classification by word matching or party name. When serious sources classify the same subject differently, each view should name who made it, what they meant, and what evidence and period they considered. No score or automated rule settles a contested identity.

How can I check an explanation?

AskBegin with the subject or question you recognize.
ReadGet the explanation before the supporting claims and sources.
Open a claimReveal supporting claims and citations near the explanation.
Check the citationSee the contributor, year, exact location, and link to the source.
Go to the passageUse the page, article, chapter, or section named in the citation.
Keep the limitsRead qualifications and open questions with the claim.

The main explanation stays readable while details remain available on demand. A citation is not enough by itself: the named passage must support the particular claim, and a source used for background must not be presented as direct proof.

What kinds of judgment stay separate?

Evidence about what happened

Descriptions of rules, actions, participation, outputs, and outcomes need sources. The site distinguishes a stated aim or formal design from reports of how people acted under those rules. Missing or conflicting evidence remains visible.

Explanations of why it happened

A sequence is not automatically a cause. Causal accounts remain attributed and should face serious alternatives. A counterfactual makes the comparison explicit by asking what would likely have happened if a named condition or institution had differed.

Judgments about whether it was good

An assessment depends on values and priorities as well as evidence. The site states the standard being used, the assumptions behind it, and the evidence it would require. It does not turn several standards into a hidden overall score.

FROM A SOURCE TO AN EXPLANATION

What does the evidence trail contain?

The site keeps the material being studied, the copy actually consulted, and the claim drawn from it as separate records. That makes it possible to correct one part without quietly changing the others.

Works and sources

A work is the underlying book, article, law, speech, dataset, image, or other creation. A source is the particular edition, archive record, translation, scan, or webpage consulted here. A citation points to that source and always names a precise page, table, section, timestamp, or other locator.

Claims and citations

Each substantial proposition is recorded as a focused claim rather than hidden inside a long paragraph. Claims can state an observation or definition; attribute a value or purpose to a named actor; describe a proposed design; record a classification or causal hypothesis; or make an editorial interpretation. This keeps factual, attributed, and analytical language visibly distinct.

A citation can support, challenge, qualify, or provide context for that exact claim. Context helps a reader understand the claim but does not prove it, and a background reading link has no evidentiary role.

Relationships and qualifications

Connections among ideas, institutions, cases, and claims have a direction and a stated scope. Their evidence can support the connection, challenge it, or limit where it applies. A missing connection is not treated as proof that no relationship exists.

Names and external identities

Wikipedia links can help a reader recognize a subject, while identifiers such as Wikidata records can help distinguish one identity from another. These links are orientation aids, not evidence. They cannot support a claim, import an outside classification, or erase a disagreement about what a term means.

BOUNDED CASES

A case shows a setting, not a perfect example.

Each case bounds the place and period and identifies the people and institutions its evidence addresses. Where possible, it separates written rules from rules in use, interactions, and outcomes. A country is not treated as one timeless system, and an institution found in a case does not make the whole case an embodiment of an ideology.

Within a case, an episode narrows attention to a particular phase or institutional arrangement. Its evidence is sorted into five categories: conditions, formal rules, rules in use, interactions, and outcomes. Empty categories remain empty rather than being filled by inference, and dates describe the episode being examined rather than the lifetime of every institution involved.

Qualifications and limits attach to the claims they constrain rather than forming another evidence category. This keeps a caveat beside the proposition it changes and avoids treating a general warning as evidence about the episode.

Case selection and sources create limits. Records may preserve official voices better than ordinary experience, cover one region unevenly, or describe a translated term through colonial categories. Those gaps narrow what the case can establish and what can be transferred elsewhere.

OPEN QUESTIONS

What happens when the evidence is not enough?

A specific unresolved question stays attached to the claim or explanation that exposed it. It may ask for a serious counterargument, contrary evidence, a missing comparison, or a causal counterfactual. The question names what evidence could address it instead of using “more research is needed” as a general disclaimer.

An open question can be partly answered, resolved with new claims, or withdrawn with an explanation. Until then, it remains a visible limit on the conclusion. Uncertainty does not assign a numerical confidence score; it narrows the language, period, place, or population to what the available evidence can sustain.

UNCERTAINTY AND CORRECTION

Human judgment remains accountable.

AI may assist discovery, synthesis, drafting, and consistency checks, but it is not evidence or authority. Researchers and reviewers decide which questions matter, how a claim is worded, which sources are relevant, and whether a classification is fair. Checks can catch missing links, unsupported publication states, and malformed citations; they cannot make those choices neutral.

Uncertainty should change the wording and scope of a conclusion. Fairness does not give every account equal weight when their relevance or evidentiary quality differs. Consequential objections, conflicting evidence, causal alternatives, and unanswered questions remain visible as open research questions. Better evidence can revise a claim, a classification, or the method itself.

The live site includes only material that has been checked against its sources, fitted to the current definitions, and accepted for publication. Earlier notes and unused research may inform later work, but readers do not encounter them as published claims, citations, or pages.

To suggest a correction, name the page or claim, explain what should change, and link the evidence and exact passage when possible. A person reviews the proposal and its sources before publication. Accepted material changes remain inspectable in the public repository history. The editorial governance policy explains who decides, how corrections and reconsideration work, and which safeguards apply to sensitive subjects.

JUDGMENT

Which assumptions shape an assessment?

Accountability

Can affected people inspect, contest, and replace delegated decision-makers?

Assumption: Public and quasi-public power should be answerable to affected people.
Evidence needed: Formal appointment rules and evidence of representation, transparency, contestation, and practice.
Limit: Formal appointment rules do not establish accountability in practice.

Affected-community accountability

Can affected community members inspect decisions, demand accounts, contest officeholders, and replace delegates?

Assumption: People subject to governing decisions should have effective means to contest and replace decision-makers.
Evidence needed: Formal selection and removal rules, located accounts of practice, participation evidence, and counterevidence about military or informal authority.
Limit: Rotation and declared accountability do not alone establish equal participation or civilian control in practice.

Distribution

Who receives income, wealth, control, benefits, and adjustment costs?

Assumption: Concentrated wealth and involuntary burdens require justification.
Evidence needed: Separate measures of wages, wealth, ownership, control, employment, and adjustment costs.
Limit: Wage equality does not establish wealth equality.

Planning correctability

Can affected institutions identify shortages or excess claims, revise allocations in time, and document who bears the adjustment?

Assumption: Institutions exercising binding allocation power should make consequential errors detectable and corrigible.
Evidence needed: Formal revision rules and timing. Records of changed requirements, allotments, shortages, and resulting production adjustments.
Limit: Fast revision does not establish democratic accountability, distributive fairness, or efficient outcomes.

FURTHER DETAILS

Read the commitments and technical definitions.

The project vision describes the site’s purpose. The editorial principles explain fairness, judgment, source use, and revision. The editorial governance policy explains accountability, corrections, reconsideration, and safeguards, while the contribution guide gives further detail for offering evidence. Researchers can also inspect the technical definitions used to keep records distinct.