Chapter 45
Hypotheses and suppositions; conjecture in science and reasoning; testing assumptions.
263 passages · 81 principal · Covers the 49 volumes of primary texts
Hypothesis is the act of supposing: laying down something not yet known to be true so that reasoning can proceed from it, and so that its consequences can be compared with what is found. The idea covers the deliberate postulate of the geometer, the tentative guess of the naturalist, the conditional reasoning of the moralist about origins no witness saw, and the emendations of the textual critic. It belongs among the great ideas of these books because nearly every inquiry they record moves by supposition. When Plato’s Socrates says he assumes the strongest principle he can find and accepts as true whatever agrees with it1, he names a procedure that Kelvin, many centuries later, calls simply unavoidable2. Between these two lie questions that are still open. What entitles a supposition to belief? How is it tried? When should it be abandoned?
The first question is whether reasoning can do without hypotheses at all. The authors who answer most firmly are the ones who use them most skilfully. Harvey, though he prides himself on demonstration rather than the conjecture of others3, assumes quantities of blood merely as a ground for reasoning, and notes that his conclusion holds whatever figures are supposed4. His circular motion was first conceived as what must be the case if the blood is to return from arteries to veins5. Descartes explains why he calls certain propositions hypotheses even while claiming to deduce them from first truths. Causes and effects, he says, lend each other support, and this is no vicious circle6. Locke recommends the Copernican motion of the planets to the learner as the simplest hypothesis and the likeliest true7. Yet he advises reading the philosophical systems to learn their hypotheses, not in the expectation of satisfying knowledge8. Lyell allows that the assumed course of nature may be supplied hypothetically, but only if it is kept consistent with probability and with analogy to what is observed9. Carlyle draws a moral line here: what is unproven should stay on the hypothetical side of the barrier until it has truly been acquired10.
A second use of hypothesis arises where facts cannot decide. Rousseau, inquiring into the origin of inequality, sets the facts aside. He calls his reasonings hypothetical and conditional, meant to illustrate the nature of things, much as naturalists frame systems of the world’s formation11. He declines to trace man’s bodily origin through comparative anatomy, judging that only vague conjectures are possible there, and so supposes man always formed as now12. He admits that his whole account rests on conjecture, defends it as the most probable and the only available means of finding the truth13, and adds that such hypotheses cannot be overthrown, since they cannot be given the certainty of facts14. Kant’s case differs in kind. For him freedom is a necessary hypothesis of reason in any being conscious of a will15. Practical reason must presuppose it, and it is consistent with physical necessity16. Rousseau’s supposition is necessary because evidence is lacking. Kant’s is necessary because action is impossible without it. Neither is meant to be tested as Harvey’s was, and this marks one boundary of the idea.
Most authors who conjecture take care to say so. Pliny offers several explanations of a fountain’s rhythm, from air currents to a hidden reservoir, and leaves his friend to judge17. Newton presents his suspicion about the forces of particles explicitly as a conjecture18. Franklin found that by offering his observations as conjectures rather than dogmas he was under no obligation to defend them19. Darwin’s journal is full of guesses labelled as such. He suggests that phosphorescence may come from decaying organic particles20 and that circling vultures signal prey to one another21, and he concedes that his speculations on rates of elevation are vague22. In the Origin he proposes a duck-like ancestor only as not incredible23, and he declines to speak with certainty about the origin of an instinct24. Helmholtz hedges his account of ice granules25, and Geikie calls his figures tentative and open to correction26. Milton places a different kind of modesty in Raphael’s mouth. Astronomers’ devices to “save appearances” are mocked27, and a string of “what if” suppositions about sun, earth and other worlds is left undecided28. Milton’s caution is reverent where the naturalists’ is methodical.
The more consequential question is how a supposition is put to trial. Copernicus treats the earth’s motion as a postulate to be tried, on the principle that a false hypothesis would not produce verified results29. He faults earlier astronomers for hypotheses that fitted the phenomena without corresponding to them30. Descartes adds that where several deductions are possible, crucial experiments yielding different results must decide among them31. Voltaire describes Newton conjecturing that gravity reaches the moon and then checking the supposition against measurement32. Faraday stages the method before his audience: steam, hydrogen or some other gas, each tested in turn33. Helmholtz devises an experiment to test Thomson’s explanation of the plasticity of ice34. He confirms a theoretical prediction quantitatively35 and recalls how Davy’s experiment with ice disposed of the latent-heat account of friction36. Even Dante’s Paradise contains such a trial, when the poet’s conjecture about dense and rare bodies is examined and found false37.
Prediction gives the trial its sharpest form. Darwin states plainly what would destroy his theory38, and he observes that contrary results on species and varieties would have been fatal39. He reasons from selection to how bees must build, then tests the reasoning with wax and vermilion. The bees stop excavating where planes intersect and leave flat walls40,41, results that also correct Huber42. He predicted slight revolving movements in non-climbing plants, which Fritz Müller later found43. The coral-reef theory is tested the same way. Fringing reefs should mark stationary or rising land44, dead and submerged reefs should exist45, and the coloured map should sort reef types by the movements of the land46. Darwin warns that his inference about volcanoes was hazardous until their absence from subsiding areas confirmed it47. Pasteur lays a wager that fruit cells will act as yeast in carbonic acid gas, and tests it48. He holds that theoretical ideas, however improbable at first, are confirmed when new facts bear them out49. He even predicts a patient’s recovery from his cultures50.
Not every theory can be tested crucially, and a further criterion appears: the range of facts a theory explains. Burke defends a theory grounded in experiment as good for what it explains, even if it does not account for every poetic image51. Darwin makes this criterion the backbone of his case. A theory should be accepted if it explains large bodies of facts that are otherwise inexplicable52. He likens natural selection to the undulatory theory of light and to gravitation53. He contrasts it with a creation view that explains nothing54,55. Huxley, quoted in the volume, calls descent an unproven hypothesis, yet the only one physiology supports56. The criterion carries an obligation to face objections, and Darwin does so candidly. He admits difficulties that might be fatal57, judges most of them only apparent58, and concedes that the theory stands or falls with the imperfection of the geological record59. He argues that negative evidence from that record is nearly worthless60. Here a tension appears. Explaining gaps by supposition can resemble the ad hoc patching that other authors condemn, and Darwin seems aware of it when he asks for allowances for ignorance only within limits61.
Rival hypotheses are therefore weighed openly. Darwin rejects in turn the theories of atolls that rest on protective instinct, on submarine craters, and on Chamisso’s growth at the outer edge, because each ignores the problem of the foundation62. Subsidence alone solves it63, and it accounts besides for the form, grouping and alignment of atolls64. Lyell traces the prejudices that held geology back. Earlier geologists conjectured what might have been instead of studying present causes65. Violent catastrophes were invoked without restraint from analogy66, and the Woodwardian hypothesis crowded countless rare events into a few months67. His imagined subterranean philosopher shows how theories mirror the narrow range of phenomena their authors observe68. Darwin follows him in refusing sudden revolutions of climate to explain the frozen Siberian animals69 and the debacle proposed by catastrophist geologists70. He also judges Forbes’s land bridges insufficiently evidenced71. Kelvin turns the same weapon on the geologists themselves. Tidal evidence shows the earth as rigid as steel, not a thin shell over molten lava72,73, and he denies any right to assume interior fluidity74.
The sharpest polemics are directed at gratuitous hypotheses. Voltaire charges that Descartes left geometry to frame hypotheses, making philosophy an ingenious romance75, and that the vortex theory heaps hypothesis on hypothesis without establishing that its matter exists76. Pasteur condemns convenient theories that receive a new hypothesis whenever facts contradict them, as with Liebig and Fremy77. He calls Liebig’s contact-action obscure and gratuitous78 and Traube’s ferment imaginary79. He discards spontaneous generation and an independent virus until only the hypothesis that the vibrio itself is responsible remains80. Newcomb finds Struve’s extinction of light unsupported by facts81. Berkeley carries the critique into metaphysics. Hylas’s brain-traces are absurd because the brain is itself only an idea82. Hylas finally concedes that material substance was a groundless hypothesis83, while Philonous claims to frame no hypothesis at all, only plain belief in what the senses show84. Whether this disclaimer itself rests on a supposition is a question the dialogue raises without settling.
Good inquirers also correct their own suppositions. Darwin supposed the air at sea was moist until measurement proved the reverse85. He abandoned his belief that a plain had been deposited by the sea once pottery and relics pointed to an earthquake wave86. He gave up a simultaneous cooling of the whole world for Croll’s explanation87, and he revised a view after a reviewer’s argument88. In Helmholtz’s history of physics the caloric theory fitted many processes89 until heat generated by friction from no known source undermined it90. The kinetic theory then came forward as a probable account91. Kelvin, by contrast, holds the ether a reality of science, inferred from the vibrations of light92,93, and finds support for the speculative hypothesis of an elastic ether in wax that yields like pitch and yet is elastic94. In the medical writings, Jenner offers the origin of cow-pox from the horse as presumptive evidence and not proof95. He leaves his conjectures for experiment to decide96,97 and hopes they will direct others to closer inquiry98. Lister conjectures that decomposition in wounds can be prevented by killing the floating particles rather than shutting out the air99.
Outside natural science, hypothesis takes on other functions. Hume offers custom as the only hypothesis that explains inference from repeated instances100, and proposes to test his account of experimental reasoning on animals101. He grants that the religious hypothesis may explain visible phenomena, but he holds that it licenses no added facts102. Any further attribute ascribed to the Author of nature he calls mere conjecture103,104. Browne, on the other hand, values guardian angels as a hypothesis that resolves doubts ordinary philosophy cannot105. Lessing defends transmigration as an old hypothesis that is not laughable merely because it is old106. Mill’s father would treat the Manichaean theory as a mere hypothesis107. Emerson reports metaphysicians calling the astronomers’ impulse an unreasonable postulate that begs the question108.
History and criticism have their own discipline of conjecture. Herodotus offers his own conjecture that the sun lowers the Nile in winter109, and he judges the unseen by the evident when he compares the Nile with the Ister110. Johnson observes that editors indulge conjecture beyond warrant111. Conjectural criticism, he says, has no axiomatic system and yields at best one probable reading among many112, so it is harmless only if kept in the margin113. Macaulay tests the rival explanations of The Prince against Machiavelli’s other works and refutes them114. Emerson finds the theories of Stonehenge built on coincidences of names115. Freeman holds that the origins of the Aryans remain unprovable conjecture116, while he entertains alternative suppositions about them117. Renan distrusts hypotheses that are more ingenious than solid118. Bagehot admits his reading of a woman’s motives is mere hypothesis119. Washington warns that changes made on hypothesis invite perpetual change120. Darwin notes that travellers fill the gaps in their knowledge with superficial hypotheses121.
Across these texts there is wide agreement that supposition is indispensable and that it must be marked as supposition. There is also agreement that deduced consequences, crucial experiments and explanatory range are its proper trials. The disagreements lie elsewhere. Some suppositions can never meet a test: Rousseau’s origins, Kant’s freedom, Hume’s deity. Where these belong in knowledge remains unsettled. So does the line between legitimate auxiliary supposition and the ad hoc patching that Pasteur and Voltaire denounce, which explanatory power alone cannot draw.
Introductory essay written by Claude Opus 5.5 from the outline and the notes on every passage below; quotations are checked against this edition.
See also: Experience 9e
See also: Geology 4b
See also: Geology 2b · Physics 6b
See also: Evolution 3a · Cause 9a
See also: Change 5a · Geology 6c · Science 3d · Cause 8b
See also: Physics 5b · Matter 1c
See also: Life and Death 1d · Medicine 9c · Cause 9b · Experience 9c
See also: Cause 2e · Nature 4d · God 1c
See also: Cause, Geology, Evolution, Science, Physics, Experience