Philosophy of Natural Science

Summary

Philosophy of Natural Science is a short introductory textbook by the German-American philosopher Carl G. Hempel, published in 1966 by Prentice-Hall in its Foundations of Philosophy series. It is still under copyright, so the library holds no text; this record describes the book in its own words and quotes only two short sentences. In about a hundred pages and eight chapters Hempel sets out how hypotheses are formed and tested, what it means for a law to explain an event, and how theories and theoretical concepts relate to observation, using concrete episodes from the history of science rather than formal logic. The first chapter distinguishes the empirical sciences, which must be tested against experience, from the non-empirical disciplines of logic and pure mathematics. The second chapter, the most widely reprinted, tells the story of Ignaz Semmelweis at the Vienna General Hospital in the 1840s. Hempel follows Semmelweis as he considers and rejects explanations for the high death rate from childbed fever in the First Maternity Division: epidemic influences, overcrowding, rough examination by students, the fear caused by a priest's bell, and the women's position in delivery. The death of his colleague Kolletschka from a scalpel wound during an autopsy suggested that students coming from the dissecting room carried 'cadaveric matter', and washing in chlorinated lime brought the death rate down sharply. From the story Hempel draws the moral that hypotheses are not mechanically derived from collected facts but are guesses invented to account for them, to be judged by testing their consequences. The following chapters analyse the logic of testing. A failed prediction can refute a hypothesis, but a successful one cannot prove it; tests always rely on auxiliary assumptions; crucial experiments and ad hoc rescues are examined, with examples from Torricelli and Pascal on the weight of air and from the phlogiston theory. Hempel discusses what makes evidence strong: quantity, variety, precise new predictions, theoretical support and simplicity. Chapter 5 presents his best-known idea, the covering-law model, in which an event is explained by deducing it from general laws and particular conditions, with a statistical variant for probabilistic laws. The final chapters treat theories that posit unobservable entities, operational definition of scientific concepts, and whether biology can be reduced to physics and chemistry. Widely used in college courses for decades, the book was translated into many languages and remains one of the clearest short statements of the logical empiricist view of science.

Plot or Argument

• Ch. 1: Scope and aim of natural science

  Hempel separates the empirical sciences, whose statements must be checked against experience, from logic and pure mathematics, and announces the book's concern with inquiry, explanation and theory.

• Ch. 2: Invention and test, the Semmelweis case

  Semmelweis's investigation of childbed fever illustrates the elimination of rival hypotheses, the role of a chance event in suggesting a new one, and the decisive test by handwashing. Hempel argues against the 'narrow inductivist' picture of science as collecting facts first and generalizing afterwards.

• Ch. 3: The logic and force of a test

  Deductive implications of a hypothesis can refute it but cannot conclusively verify it; tests depend on auxiliary hypotheses; crucial experiments and ad hoc hypotheses are examined with episodes such as Torricelli's barometer and Périer's ascent of the Puy-de-Dôme.

• Ch. 4: Criteria of confirmation and acceptability

  The support a hypothesis receives depends on the quantity and variety of evidence, on successful new predictions, on its fit with other theories, and on simplicity, a notion Hempel finds important but hard to define.

• Ch. 5: Laws and explanation

  Hempel distinguishes laws from accidental generalizations and presents the deductive-nomological model of explanation and its probabilistic counterpart, the inductive-statistical model.

• Ch. 6: Theories and theoretical explanation

  Theories introduce unobservable entities and processes and are linked to observation by bridge principles; their power lies in unifying many laws.

• Chs. 7–8: Concept formation and reduction

  Hempel examines operational definitions and their limits, and asks whether biology can be reduced to physics and chemistry, weighing mechanistic and vitalist positions.

Lessons

• Facts do not arrange themselves

  Hempel argues that no one could make progress by first collecting all the facts and only then looking for patterns, because without a hypothesis one does not know which facts are relevant. Semmelweis needed a guess about what killed the women before he knew what to observe. Imagination begins inquiry; testing disciplines it.

  • Why does Hempel think the 'narrow inductivist' account of science is unworkable?
  • Which of Semmelweis's hypotheses came from chance events, and which from his own reasoning?
  • Where does imagination belong in science, and where must it give way to testing?

• A test can refute more than it can prove

  If a hypothesis predicts something that does not happen, it is in trouble; if the prediction succeeds, the hypothesis is supported but not proved, because some other hypothesis might have predicted the same thing. Hempel teaches students to see the asymmetry and to ask how much support a successful test really gives.

  • When Semmelweis's priest experiment failed, what exactly did it show?
  • Why could the success of handwashing not by itself prove his theory of cadaveric matter?

• To explain is to show why it had to happen

  On Hempel's covering-law model, explaining an event means showing that it follows from general laws together with particular conditions. The model makes explanation and prediction two sides of the same reasoning. Later philosophers questioned whether every explanation fits this form, and Hempel himself refined it.

  • Using Hempel's model, how would you explain why a puddle froze overnight?
  • Can you think of a good explanation that does not seem to cite any law?

• The cost of an idea before its time

  Semmelweis reasoned well and saved many lives, yet his conclusion was resisted for years. Hempel's chapter shows the method that led to the right answer; the history around it shows that method alone does not guarantee acceptance.

  • What evidence did Semmelweis have that his colleagues might have found convincing?
  • Why might a sound conclusion be rejected by competent people?

Context

Hempel belonged to the logical empiricist movement that grew up around the Vienna Circle and the Berlin Group in the 1920s and 1930s and moved largely to the United States as its members left Central Europe. Its aim was to clarify the logic of scientific reasoning: what a law is, how evidence confirms a hypothesis, and what an explanation accomplishes. Hempel's papers of the 1940s, especially 'Studies in the Logic of Explanation' with Paul Oppenheim (1948), were central to that programme.

Composition

The book was written at Princeton as an undergraduate introduction for Prentice-Hall's Foundations of Philosophy series, edited by Elizabeth and Monroe Beardsley, which commissioned short volumes from leading philosophers. Hempel condensed his technical work into plain prose and chose historical case studies in place of formal apparatus.

Reception

The book became a standard text in American and European courses and has been translated widely. It appeared four years after Thomas Kuhn's Structure of Scientific Revolutions, and readers have often set the two side by side as contrasting pictures of science. Historians of medicine have pointed out that Hempel's account of Semmelweis is brief and leaves aside much of the surrounding story, while philosophers have used the same case to compare inductive, hypothetico-deductive and explanatory approaches to inference.

Comparisons

• pasteur-and-lister-germ-theory-papers

  Semmelweis reasoned about 'cadaveric matter' without knowing of germs; Pasteur's cultures from puerperal fever and Lister's antiseptic surgery supplied the microbial explanation two and three decades later.

• popper-the-logic-of-scientific-discovery

  Popper also stresses that hypotheses are bold conjectures and that refutation is logically decisive where confirmation is not, though Hempel gives positive evidence more weight than Popper does.

• kuhn-the-structure-of-scientific-revolutions

  Kuhn's account of paradigms and revolutions offers a historical and communal picture of science that readers often contrast with Hempel's logical analysis of testing and explanation.

• mill-system-of-logic-book-iii

  Mill's methods of agreement and difference describe the eliminative reasoning that Hempel's Semmelweis applies to the two maternity divisions.

• peirce-illustrations-of-the-logic-of-science

  Peirce's account of hypothesis, or abduction, anticipates Hempel's insistence that hypotheses are invented rather than read off the data.

• bernard-introduction-to-experimental-medicine

  Bernard's account of the experimental idea, tested by its consequences, is a nineteenth-century physiologist's version of the method Hempel analyses.

Quotes

For Students

A reading companion for classical Christian high school students. See the student guide.

Reading Guide

Watch for

• This record has no full text because the book is under copyright; read a library or classroom copy alongside the guide.
• The list of rejected explanations in Chapter 2, and how each is ruled out: by existing facts, or by a deliberate test.
• The role of a chance event, Kolletschka's death, in suggesting the hypothesis that proved right.
• Hempel's argument against the idea that science begins by gathering all the facts without a guiding question.
• The two kinds of explanation in Chapter 5, deductive and statistical, and the examples Hempel uses for each.

Key terms and names

Hypothesis

A proposed statement offered to account for some facts, which must then be tested by its consequences.

Test implication

An observable consequence that should follow if a hypothesis is true, such as a falling death rate after handwashing.

Auxiliary hypothesis

An additional assumption, about instruments or conditions, needed to derive a test implication; a failed test may cast doubt on it rather than on the main hypothesis.

Ad hoc hypothesis

An assumption added only to rescue a theory from contrary evidence, without independent support.

Covering-law model

Hempel's account of explanation, in which the event to be explained follows from general laws together with statements of particular conditions.

Deductive-nomological explanation

The form of covering-law explanation in which the conclusion follows with certainty from universal laws.

Inductive-statistical explanation

The form of covering-law explanation in which statistical laws make the event highly probable rather than certain.

Childbed (puerperal) fever

A severe infection of women after giving birth, the disease at the centre of Hempel's case study.

Seminar Questions

1. Comprehension What were the death rates in the two maternity divisions, and why did the difference between them matter to Semmelweis?
2. Comprehension List the explanations Semmelweis rejected before reaching his solution, and say how each was ruled out.
3. Comprehension According to Hempel's covering-law model, what two kinds of statements must an explanation contain?
4. Interpretation Hempel says the priest experiment failed to reduce mortality. Did that test refute the psychological hypothesis completely, or could someone have rescued it? How?
5. Interpretation Why does Hempel think hypotheses must be invented rather than derived from data? Do you find his argument convincing?
6. Interpretation If explanation and prediction have the same logical form, as Hempel claims, why do we sometimes feel we understand something we could not have predicted?
7. Big Question Semmelweis's conclusion was correct and saved lives, yet many physicians resisted it. What do humility and charity require of professionals when evidence suggests they themselves are causing harm?

Faith Connections

• 1 Thessalonians 5:21 (Scripture)

  'Prove all things; hold fast that which is good.' Hempel's method of testing each hypothesis by its consequences is a disciplined form of this counsel applied to the natural world.

• Proverbs 18:17 (Scripture)

  'He that is first in his own cause seemeth just; but his neighbour cometh and searcheth him.' Semmelweis's rejection of plausible explanations one after another shows why first impressions need examination.

• Luke 10:33-34 (Scripture)

  The Samaritan binds the wounds of a stranger and pays for his care. Semmelweis's persistence on behalf of poor women in a public ward, as Hempel recounts it, is an example of learning put to the service of neighbours.

• Francis Bacon, Novum Organum I.95 (Great Conversation)

  Bacon compares the empiric to the ant that only collects and the dogmatist to the spider that spins from itself, and praises the bee that gathers and transforms. Hempel's mix of imaginative hypothesis and strict test is a modern version of the bee.

Logic and Rhetoric

Central argument: Science cannot proceed by collecting facts without a guiding question, since without a hypothesis there is no way to decide which facts are relevant; hypotheses are therefore invented to account for the facts, and their acceptability depends on deriving test implications and checking them against experience, as Semmelweis did; and a scientific explanation shows that the event to be explained follows, deductively or with high probability, from general laws and particular conditions.

Rhetorical device: extended example (exemplum). Hempel opens his account of method not with definitions but with the Semmelweis story, returning to it repeatedly in later chapters so that abstract points about testing are anchored in a single memorable case.

Debate: Resolved: Every genuine scientific explanation shows that the event explained had to happen, or was highly probable, given the laws of nature.

Writing Prompts

Essay: Using the Semmelweis chapter, explain Hempel's account of how scientific hypotheses are invented and tested. Then argue whether Semmelweis's handwashing result proved his hypothesis or only supported it.

Comparison with Germ Theory Papers (Pasteur and Lister): Semmelweis, Lister and Pasteur all attacked the same enemy, infection after childbirth or surgery. Compare the kind of evidence each relied on, and explain why Pasteur's cultures could establish what Semmelweis's statistics could only suggest.

Medical Lens

The book is philosophy of science, but its most famous chapter is a case study in clinical epidemiology: Semmelweis's comparison of mortality from puerperal fever in two maternity divisions of the Vienna General Hospital (1844–1848), his elimination of rival explanations, the inference from Kolletschka's death after an autopsy wound, and the trial of handwashing with chlorinated lime. Hempel reports the yearly death rates and the later extension of the hypothesis to putrid matter from living patients after a case of uterine cancer. The chapter is often taught to medical students as an introduction to hypothesis testing and confounding, and it stands just before the germ-theory papers of Pasteur and Lister in the history of infection control. Hempel's discussion of reduction in the last chapter also touches the old debate between mechanism and vitalism in physiology.

• Ch. 2

  Semmelweis rejects epidemic influences, overcrowding, rough examination, the priest's bell and delivery position as causes of the excess mortality in the First Division.

  Childbirth Maternal Fever Infection
• Ch. 2

  Kolletschka's fatal scalpel wound during an autopsy suggests that students coming from dissections carry cadaveric matter; handwashing in chlorinated lime follows.

  Anatomy Dissection Physicians Doctors Public Health Sanitation
• Ch. 8

  Mechanism and vitalism in biology and the question whether life processes can be reduced to physics and chemistry.

  Physicians Doctors

Creative Prompts

• Classroom

  Turn Semmelweis's investigation into a table: list each hypothesis, the prediction it makes, the test or observation, and the result. Then decide which tests were strongest.

• Essay

  Write the case for and against the covering-law model using an everyday explanation, such as why a car would not start, and argue whether Hempel's model captures what a good explanation does.

Further Reading

• Aspects of Scientific Explanation, Carl G. Hempel

  Hempel's collected technical essays (1965), including the papers on confirmation and explanation.

• The Etiology, Concept, and Prophylaxis of Childbed Fever, Ignaz Semmelweis

  Semmelweis's own 1861 account of the investigation Hempel retells.

• Inference to the Best Explanation, Peter Lipton

  Reworks the Semmelweis case to compare models of inference.

• The Doctors' Plague, Sherwin B. Nuland