Scientific Anti-Realism: Constructive Empiricism

Scientific Realism vs. Scientific Anti-Realism

• Scientific realism: the view that the best explanation of the success of science is that our mature theories are (approximately) true and their central, theoretical terms refer to real, mind-independent entities.
• Scientific anti-realism: umbrella term for positions that reject truth (or approximate truth) as science’s primary aim and/or deny that we can justifiably believe in unobservable entities.
• The transcript focuses on one prominent anti-realist form: Constructive Empiricism (CE).

Constructive Empiricism (Bas van Fraassen, 1980-present)

• Name breakdown:
  • “Constructive”: emphasizes the active, intentional construction of models that scientists build.
  • “Empiricism”: emphasizes that knowledge claims should not outrun the observable evidence.
• Semantic agreement with realists:
  • Both CE and realism take the language of science “at face value.”
  • Theoretical terms (planet, electron, neutrino, etc.) are treated as purporting to refer to real entities.
  • No reinterpretation or instrumentalist “fiction” about language is needed.
• Epistemic disagreement with realists:
  • Realist: we should believe theories are (at least approximately) true.
  • CE: we should withhold belief in the truth of theories about the unobservable. We need belief only in their empirical adequacy.
• Definition:
  • A theory T is empirically adequate iff everything T says about observable entities, events, and regularities—past, present, and future—is true (i.e.
    it saves the phenomena).
  • Truth about unobservables is unnecessary.
  • Thus, the aim of science = empirical adequacy, not truth.

Observable vs. Unobservable

• Observable: objects & properties accessible to unaided human senses (or, in some stricter versions, those we could in principle see without theoretical assumptions).
  • Example: mineral hardness, melting point, visible color.
• Unobservable: atoms, electrons, protons, molecular structures, atomic numbers, quarks, DNA’s exact atomic geometry, etc.
  • Gold’s atomic number: 79 (defined via number of protons/electrons) → unseen by the naked eye; learned through theory-laden instrumentation.

Role of Scientific Models

• Scientists build idealized, abstract models to represent phenomena. CE stresses:
  • Models are tools for calculation, explanation, and prediction—not literal pictures of reality.
  • Adequacy criterion = match with observable data.
• Examples discussed:
  1. Crystal lattice models of minerals
  • Colored balls & sticks indicate atomic/molecular arrangement.
  • Useful for predicting cleavage planes, combination tendencies, etc.
  • Do not guarantee truth about actual atomic shapes or bonds.
  1. Stick-and-ball double helix model of DNA
  • Requires multiple idealizations:
    • Perfectly spherical atoms of uniform color.
    • Vacuum-like environment (ignores messy cellular context).
    • Picked-out right- or left-hand helix orientation.
  • Explanatory and pedagogical power despite idealized distortions.

Historical & Philosophical Context

• Ancient Greek astronomy (“save the phenomena”): Ptolemy & Simplicius accepted that human models can at best fit appearances; only the divine mind could know underlying truth.
  • Post-Galileo era removed theological basis; modern CE’s reasons are different: metaphysics & model idealization.
• Metaphysical prudence argument:
  • Believing in current unobservables is a high-risk strategy; past science is littered with abandoned entities:
  • \text{Ether}, \text{Phlogiston}, \text{Caloric}, epicycles, etc.
  • Electrons, protons, neutrinos, DNA, etc. might meet the same fate in a century.
  • Therefore we should not incur excess metaphysical commitments.
• Darwinian/Selectionist explanation of success:
  • Competing theories undergo a “struggle for survival.”
  • The ones we keep are those that fit observational evidence; survival ≠ truth.
  • This provides an alternative explanation to the realist’s “they work because they’re true.”

Idealization, Abstraction, and Model-Based Science (last ~30 yrs of philosophy of science)

• Growing literature documents that:
  • All theories deploy simplifications, ideal boundary conditions, deliberate omissions.
  • Even if a model is approximately true in a very narrow sense, it is literally false in many respects.
  • Thus, truth may be a misguided or excessively ambitious goal for everyday scientific practice.

Implications & Debates

• CE offers a middle way:
  • Accept the realist’s sophisticated mathematics, large-scale experimentation, and literal syntax of theory.
  • Reject the necessity of belief in the unseen.
• Challenges posed to realism:
  1. Pessimistic meta-induction: track record of discarded entities suggests caution.
  2. Underdetermination: multiple empirically adequate theories can exist; why privilege one as true?
• Realist responses (not in transcript but contextually linked):
  • No-miracle argument: best explanation of reliability/prediction is that theories are at least approximately true.
  • Explaining novel success often seems to require belief in underlying structure.
• CE counters: “empirical adequacy + selectionist success” is explanation enough; metaphysical humility preferred.

Key Takeaways & Study Reminders

• Memorize empirical adequacy definition.
• Distinguish clearly between semantic (language-meaning) vs. epistemic (belief-justification) aspects.
• Know why CE can accept everyday scientific practice yet remain anti-realist.
• Be ready to illustrate with examples (gold atomic number 79, crystal models, DNA helix) why observable/unobservable divide matters.
• Understand why idealization undermines straightforward truth claims.
• Recall meta-arguments: Darwinian survival of theories, metaphysical risk.
• Recognize CE’s lineage: from ancient “save the phenomena” to van Fraassen’s modern empiricist stance.