cause
Let an agent reason about causes carefully, distinguish causation from correlation, apply causal frameworks in science, medicine and law, and explain the aetiology of conditions with evidence.
Bundle → Layer → Finding → Questions Filled
4 bundles · 8 layers · 8 findings · 16 questions
Identify Finding causes.
Candidates
Candidate causes.
Candidates
Candidates.
- What are the plausible causes of this outcome, and what evidence supports each? definition
- Are there confounders or reverse causation? boundary
Types
Kinds of cause.
Types
Types.
- Is this cause necessary, sufficient, proximate or contributing? definition
- Are multiple causes interacting? boundary
Science Causal inference.
Inference
Establishing causation.
Inference
Inference.
- Which methods, such as experiments or causal inference from observational data, can establish causation here? action
- What are the limits of the evidence? boundary
Aetiology
Causes of disease.
Aetiology
Aetiology.
- What is known about the aetiology of this condition, according to medical sources? provenance
- Is the user asking about the cause of their own symptoms? boundary
Law Legal causation.
Tests
Causation tests.
Tests
Tests.
- How do legal tests such as but-for and proximate cause work, in general terms? provenance
- Is the user asking for legal advice? boundary
Root
Root cause analysis.
Root
Root cause analysis.
- How can a root cause analysis be structured for this incident? action
- Which contributing factors should be recorded? action
Reason Everyday reasoning.
Fallacies
Causal fallacies.
Fallacies
Fallacies.
- Does this argument commit a causal fallacy such as post hoc or single cause? boundary
- How can it be corrected? action
Explain
Explaining causes.
Explain
Explaining.
- How can the causes of this event be explained clearly with appropriate uncertainty? action
- Are competing explanations attributed? boundary
Classifiers Filled
- Family
- Thing Registry
- Category
- Activities and processes
- Entry kind
- noun.event
- Plane
- ACT
- Domain
- ACT.EVT
- Facets: substance
- activity
- Facets: origin
- conceptual
- Facets: agency
- inert
- Facets: mobility
- not-applicable
- Facets: scale
- not-applicable
- Facets: affordances
- observable
- Other names and narrower kinds
- antecedent, etiology, aetiology, factor, fundamental, parameter, unknown quantity, wild card, producer, mutagenesis, insertional mutagenesis, releasing factor
What it is Filled
An event, condition or agent that produces an effect or brings something about, studied as causation in philosophy, science, medicine as aetiology, law and everyday reasoning, with distinctions such as necessary, sufficient, proximate and contributing causes; establishing causes requires evidence beyond correlation.
Why it exists Filled
Let an agent reason about causes carefully, distinguish causation from correlation, apply causal frameworks in science, medicine and law, and explain the aetiology of conditions with evidence.
Distinguishing features Filled
- Generative relation
- Evidence-based
- Multiple types
- Domain-specific criteria
What robots and AI may and may not do Filled
Must not
- Present a correlation as a proven cause in health, safety or policy claims.
- Name a person or group as the cause of harm without evidence.
- Give a single cause for a complex event when the evidence shows several.
- Claim to have established the cause of an individual's illness or death.
Only with a human decision
- Publishing causal findings that assign blame or liability.
May
- Explain kinds of causation and how causes are established in science, medicine and law.
- Distinguish correlation from causation when presenting data.
Moral aspects Filled
- Wrongly assigned causes lead to blame, scapegoating and wrong treatment.
- Causal claims drive policy that affects many people.
Who is affected
- People blamed for events
- Patients
- The public relying on findings
Owners Filled
Steward
Nobody owns the concept; it is held in common by philosophy and science.
Links to other meta-models Filled
parent
- oewn-07338522-n - registry parent class
related
- event - category
- beginning - origins
- hypothesis - causal hypotheses
- unrelatedness - absence of causal link
What else AI and robots need to interact with it Filled
Identity and identifiers required Filled
- Vercy registry: vr.tr.cause
- Open English WordNet: oewn-07341157-n
Direct properties not applicable Not applicable
Not applicable
Plane ACT: no invented physical properties.
Recognition optional Filled
- Produces an effect
- Necessary, sufficient or contributing
- A correlation is not by itself a cause
- Not physical; a relation between events.
Capabilities and actions required Filled
- identify candidate causes of an outcome
- distinguish correlation from causation
- apply causal criteria and frameworks
- explain aetiology and legal causation
Hazards and failure modes required Filled
- Confusing correlation with causation
- Single-cause thinking
- Post hoc reasoning
Standards and interfaces required Filled
- Bradford Hill considerations in epidemiology
- Legal tests of causation
- Root cause analysis standards
Context of use required Filled
- Understanding why things happen.
- physical causes
- aetiology of diseases
- legal causation
- social and economic causes
- root causes in engineering
Sources Filled
- Open English WordNet 2024, synset oewn-07341157-n - sense and gloss of the concept
- Wiktionary: cause - word senses and usage
Open questions
- Should causation be a separate entry?
- How should causal inference methods be linked?
- How should legal tests be localised?
Machine files
Provenance
thing registry research (pass 2) · unreviewed
Built from: models/things/publications/thing-cause/spec.json