The Scientific Image

Citation

van Fraassen, B. C. (1980). The Scientific Image. Oxford University Press (Clarendon Press).

Chapter Summary

The book is a monograph of about 235 pages in seven chapters, the main text running to roughly 215 pages. It states and defends constructive empiricism, the view that science aims at empirically adequate theories and that accepting a theory involves belief only that it is empirically adequate. Having stated the view and answered the case for realism, it develops accounts of theories, scientific method, explanation and probability that fit it.

Chapter 1: Introduction

The introduction traces the opposition between empiricism and realism about science from the seventeenth century, and explains why the logical positivists' empiricism lost ground: their theories of meaning and language proved inadequate. A viable empiricism, it says, needs conceptions of theory structure, truth and empirical adequacy that do not depend on how a theory is worded. It introduces the name "constructive empiricism," treating scientific activity as the construction of models adequate to the observable phenomena rather than the discovery of truth about unobservable parts of nature, with acceptance carrying a pragmatic dimension and only modest belief.

Chapter 2: Arguments Concerning Scientific Realism

The chapter sets two views against each other. Scientific realism holds that science aims at a literally true story of the world and that accepting a theory involves believing it true. Constructive empiricism understands theories just as literally but takes the aim to be empirical adequacy. The chapter then criticizes the main arguments for realism, including the theory/observation dichotomy, the observable/unobservable contrast, inference to the best explanation and the principle of the common cause, engaging with Smart, Sellars, Reichenbach, Putnam and Dummett. It concludes that these arguments count against logical positivism but do not favour realism over constructive empiricism.

Chapter 3: To Save the Phenomena

This chapter asks what the empirical content of a theory is. If a theory is identified with a linguistic formulation, no non-trivial line between empirical and non-empirical content can be drawn; restricting the vocabulary to an "observational" one does not describe the observable world. Identifying a theory instead with its family of models gives a new approach, which the chapter illustrates with Newtonian physics, absolute versus relative motion, nineteenth-century ether theory and quantum mechanics. It also notes a hermeneutic circle in interpretation.

Chapter 4: Empiricism and Scientific Methodology

Scientific theories do more than answer empirical questions, so the chapter takes four challenges to the empiricist: whether rejecting realism leads to self-defeating scepticism; whether methodology and experimental design are intelligible without realism; whether the unity of science and the joint use of different theories make sense; and what status to give virtues such as simplicity, coherence and explanatory power. The answers rest on the pragmatics of inquiry, including a doctrine of experiment as a continuation of theorizing by other means.

Chapter 5: The Pragmatics of Explanation

Explanatory power is a theoretical virtue that does not reduce to empirical adequacy, so it poses a challenge to any empiricist account. After a critical survey of attempts to define scientific explanation, the chapter offers a pragmatic account on which explanations are answers to why-questions, with relevance fixed by the context, interests and presuppositions of the questioner. It argues that, on this understanding, evaluations of explanation cannot support scientific realism, and it claims to resolve a range of standard puzzles about explanation.

Chapter 6: Probability: The New Modality of Science

The chapter treats probability as a modality that entered science with quantum mechanics and irreducible indeterminism, and asks how an empiricist can interpret it. It criticizes frequentist and propensity interpretations of physical probability and describes the models of probability theory as constructible from families of sequences constrained by limiting frequencies, within a general view of modality based on the pragmatics of scientific language.

Chapter 7: Gentle Polemics

The final chapter, about a dozen pages, is a parody of Aquinas's Five Ways to prove the existence of God, rewritten as arguments in the style of those given for scientific realism, followed by a parody of the realists' defence.

Key Concepts

Chapters 1–4

• Constructive empiricism: the view that science aims at empirically adequate theories and that acceptance involves belief only in their empirical adequacy.
• Scientific realism: the view that science aims at a literally true story of the world and that acceptance involves belief in truth.
• Empirical adequacy: a theory's being correct about what is observable.
• Models and empirical content: a theory identified through the structures it offers for modelling phenomena, not through a linguistic formulation.
• Acceptance: attitude to a theory that includes belief about observables and a pragmatic dimension.

Chapters 5–7

• Why-question account of explanation: explanations are answers to contextually determined why-questions.
• Theoretical virtues: simplicity, coherence and explanatory power, which are not reducible to empirical strength and are given a place through the pragmatics of inquiry.
• Probability as modality: probability as a basic parameter of nature, interpreted without frequentist or propensity accounts.

Critical Analysis

Strengths

• Precise statements of the rival aims. Chapter 2 formulates realism and constructive empiricism as claims about the aim of science and about what acceptance involves. The Stanford Encyclopedia of Philosophy takes the view's definition from this chapter and credits the book with showing that anti-realism could survive the end of logical positivism.
• A full programme rather than one thesis. The later chapters supply accounts of theory structure, methodology, explanation and probability, so the position is tested on explanation and probability, where, as Chapter 5 itself says, an empiricist faces a challenge.
• Worked examples for the models approach. Chapter 3 grounds a technical idea in Newtonian mechanics, absolute and relative motion, ether theory and quantum mechanics.
• A self-contained account of explanation. Chapter 5 can be read as a free-standing contribution to the theory of explanation, and it is the longest chapter in the book.

Limitations

• The observable/unobservable distinction is the most contested step. The Stanford Encyclopedia records objections, such as Paul Churchland's, that the line is "very feebly principled", and further discussion of observability for microscopes. Readers should study these objections and van Fraassen's replies.
• Technical in places. Chapters 3 and 6 use the vocabulary of model families, sequences and limiting frequencies, and the book engages directly with authors such as Putnam and Dummett, so it is hard going without a prior introduction to the realism debate.
• Dated in parts. Van Fraassen revisited the view in later work (for example the 1985 reply to critics, the 2001 "Constructive Empiricism Now" and the 2002 book The Empirical Stance, all cited in the Stanford Encyclopedia), and this book does not contain those developments.

Where It Fits in the Course

This is a primary source for one of the two positions at the centre of the realism debate, so it comes after an overview. Samir Okasha's short introduction treats whether scientific theories describe reality at an accessible level, and Peter Godfrey-Smith's book follows the debate from logical empiricism through to scientific realism. Alex Rosenberg and Lee McIntyre's textbook covers explanation, laws, confirmation and theory change, which gives context for Chapters 4 and 5, and Alan Chalmers's book on observation, experiment and induction is useful background for Chapter 2.

Thomas Kuhn's The Structure of Scientific Revolutions is the natural companion. Kuhn describes how sciences change; van Fraassen asks what to believe about the theories that result. Reading both shows how a historical account of theory change and an empiricist account of acceptance can be set against each other.

How to Use This Book

The book suits students who have taken a course in philosophy of science and know the realism debate and basic ideas about explanation and confirmation.

A practical order: read the Introduction, then Chapter 2 closely, as it states both positions and the case against realism. Next, read Chapter 5 on explanation, which can stand alone, and Chapter 4 for methodology. Chapters 3 and 6 are the most technical and can be read selectively; Chapter 3 matters most if a course discusses the semantic view of theories, and Chapter 6 if it covers probability. Chapter 7 is short and can be read at any point after Chapter 2.

For exercises, write out the realist and constructive empiricist aims side by side and identify one scientific claim on which they would take different attitudes. Then take Churchland's objection to the observable/unobservable line, as reported in the Stanford Encyclopedia, and draft van Fraassen's reply from Chapter 2.