Is the claim empirically answerable?
What evidence could matter?
Scientific questions connect claims to possible observation, measurement or intervention.
Side 95
Philosophy
of Science
A study of what scientific knowledge is, how theories explain, what evidence can justify, and what science is entitled to claim about unobservable reality. It examines the logic and limits behind scientific practice rather than replacing that practice.
question→model→evidence→explanation→revision
06philosophy lenses
05theory problems
05science limits
95Side
Science is not defined by one universal procedure.
Different sciences use experiments, observation, modeling, historical inference and simulation, so demarcation must look beyond a simplistic “scientific method.”
Can the claim be tested against reality?
Risk of being wrong.
A theory gains content when evidence can count against it.
Which method fits the phenomenon?
Experiment, survey, telescope, simulation?
Methodological pluralism reflects different evidential situations.
Can others inspect and challenge the work?
Public criticism.
Science depends on institutionalized scrutiny as well as individual reasoning.
Can commitments change with evidence?
Correctability.
Fallibility is compatible with knowledge when error-correction is built into practice.
Scientific explanation does more than summarize regularity.
Explanations can appeal to laws, mechanisms, causes, unification, models or statistical structure.
Derive event from general principles.
Classical accounts emphasize subsumption under laws and initial conditions.
Show what made the outcome happen.
Causal explanation identifies dependence or production.
Open the black box.
Entities, activities and organization explain how a process unfolds.
Explain many phenomena with fewer patterns.
Explanatory power can come from showing apparently separate facts as instances of one structure.
Explain distributions and probabilities.
Some sciences explain why an outcome was likely rather than inevitable.
Explain by role in a system.
Biology and social science often ask what contribution a structure makes to a larger organization.
Do successful theories describe what the world is really like?
Scientific realism and anti-realism disagree about how literally to interpret claims about unobservable entities and structures.
| Position | Core commitment | Pressure point |
|---|---|---|
| Scientific realism | Mature successful theories are approximately true | Past successful theories later proved false |
| Entity realism | Some unobservables are real because we can manipulate them | Manipulation may still rely on theory-laden interpretation |
| Structural realism | Science captures real structure better than intrinsic nature | What counts as preserved structure? |
| Constructive empiricism | Science need only be empirically adequate | Why accept observable/unobservable distinction as fundamental? |
| Instrumentalism | Theories are tools for prediction | May understate explanatory and ontological commitments |
Scientists often reason with models that are knowingly false in detail.
The philosophical problem is how idealization and simplification can still yield genuine understanding.
Idealization
Assume away complications.
Frictionless planes and perfectly rational agents can isolate mechanisms despite being unreal.
Abstraction
Ignore features irrelevant to the question.
Useful models preserve selected structure while discarding detail.
Representation
Model stands in for a target system.
Similarity alone cannot explain all scientific representation.
Simulation
Explore model behavior computationally.
Simulation creates evidence about models and, indirectly, about the world they represent.
Robustness
Result survives model variation.
Agreement across different assumptions can increase confidence that an effect is not an artifact.
Evidence can constrain theory without selecting a unique theory automatically.
Theory choice also involves simplicity, scope, coherence, predictive success and explanatory power.
More than one theory can fit available evidence.
Data rarely interprets itself without auxiliary assumptions.
Tests depend on background assumptions.
A failed prediction need not identify which assumption failed.
Prefer fewer unnecessary commitments.
Simplicity is useful but difficult to define objectively.
Explain more with one framework.
Broader reach can be valuable if it does not merely add flexibility.
Predict before fitting.
Successful risky prediction can carry more evidential weight than accommodation.
Fit with well-supported knowledge.
Consistency with other successful theories matters, but revolutions can revise that background.
Science progresses without becoming infallible.
Progress can mean better prediction, deeper explanation, wider unification, improved measurement, or elimination of error—none of which guarantees final truth.
Anomaly
Evidence exposes limits of an existing framework.
Revision
Models, auxiliary assumptions or core theories change.
Comparison
Competing theories are judged across evidence and explanatory virtues.
Retention
New theories often preserve older results as limiting cases or approximations.
Humility
Scientific success supports confidence without erasing uncertainty or domain limits.
The Logic of Scientific DiscoveryKarl Popper · falsifiability and testing
The Structure of Scientific RevolutionsThomas Kuhn · paradigms and change
The Scientific ImageBas van Fraassen · constructive empiricism
Representing and InterveningIan Hacking · realism, experiment and practice