Chapter 18
Defining terms; the nature and method of definition; disputes about the meaning of words.
180 passages · 64 principal · Covers the 49 volumes of primary texts
Definition is among the first acts of reasoning and among its most contested. The idea covers the defining of terms; the question of what a definition is, whether a name imposed or a statement of what a thing is; the rules by which definitions are framed; the limits beyond which they fail; and the disputes that arise when words are used without them. It is a great idea for these books because nearly every author in them defines, refuses to define, or complains that others have not. The geometer, the naturalist, the judge, the lexicographer and the critic all meet at this point, and their encounters show how much depends on whether we agree about the meaning of our words.
The sharpest statement of the central distinction comes from Pascal. Geometrical definitions, he holds, are arbitrary impositions of names that abbreviate discourse, as when “even number” is fixed to stand for a number divisible into equal halves1. Such a definition designates the thing named. It does not declare its nature, and so it cannot be disputed2. If a definition keeps a word’s ordinary meaning, however, it becomes a proposition that needs proof3, and confusing these two kinds of definition is what perplexes controversy4. The nature of things is immutable, while names such as army, town or ten are capricious5. Euclid could exclude unity from number because we are masters of our terms, provided we give notice6.
Against this conventional view stands the ancient practice of Marcus Aurelius, who tells himself to define or describe each presented thing in its substance, name and components, so as to see it distinctly7. He would ask what a thing is in itself, what its material is and what its causal nature8, a habit his commentator Long singles out9. For the emperor, definition is a discipline of seeing the real. The Buddhist writings take a third position. Nagasena shows that “chariot,” and even his own name, are conventional designations for assembled parts, with no separate entity beyond them10. The Buddha, asked what makes a Brahman, answers by listing qualities11. Confucius, for his part, demands only that names be rectified so that words correspond to things12.
Burke reverses the usual order. He distrusts definition as a way of circumscribing nature before nature has been examined, and holds that it should be the result of inquiry rather than its precondition13. Accordingly he defines taste with as little theory as possible14. He cares little what his terms are called, so long as the sublime and the beautiful are granted to be different things in nature15.
On method, Hobbes and Pascal speak almost with one voice. Geometers begin by settling the significations of words, and precise truth-seeking requires the same16. An error in a definition multiplies as reckoning proceeds, so anyone who seeks knowledge must examine or make his own definitions17. Certainty rests on names settled in this way18, and geometry’s conclusions are indisputable precisely because it alone begins from definitions19,20. Hobbes even defines reason itself as the reckoning of the consequences of general names21. Pascal makes defining terms the first part of the art of persuading, which also requires evident axioms and the substitution of definitions for the things defined22. His rules are that obscure or ambiguous terms must be defined, that definitions must use only known words, and that the definition must be mentally substituted for the thing defined23,24. Practised in this way, substitution foils the equivocations of sophists25. His own demonstration about indivisibles proceeds in exactly this manner26.
Not all definition is geometrical, however. Mill’s father taught him to besiege the meaning of an abstract term by fitting it into a larger class and then dividing downward with distinctions until its limits were marked27. He also made the boy define “idea” and “theory”28. Hume defines complex ideas by enumerating their simple parts, and the simplest ideas by tracing them to impressions29. Augustine distinguished the “fair” from the “fit” by corporeal examples30 and divided the passions into species31. Franklin chose more names with fewer ideas attached to each, so that his virtues might be clearly defined32. Harvey distinguishes veins from arteries by function and the direction of flow, not by structure33, and Kelvin defines tides by their cause34.
Definition has limits, and the authors who press it hardest are the first to mark them. Pascal holds that primitive terms such as man, light and being should not be defined at all, and that attempts to define them, like Plato’s featherless biped, are circular or ridiculous35. Being, soul, time, space and truth cannot be explained without using the word itself36. The want of definition here comes from obviousness, not obscurity37: we cannot define such things without obscuring them, yet we speak of them with assurance38. Mathematicians therefore err when they demand definitions in matters of intuition39. Burke likewise counts pain and pleasure as simple ideas incapable of definition40, and he finds the rights of men to lie in a middle that is discernible but not definable41. Hume admits that the feeling of belief can only be described42. Voltaire reports that no one agrees on the definition of soul or matter43, and Burns calls mankind a science that defies definitions44.
Johnson puts the logical point most exactly. Nothing can be defined except by plainer words, just as nothing can be proved except from what is self-evident45. Circular definitions such as hind and stag are therefore unavoidable, and examples must supply what explanation cannot46. At the other extreme, Hobbes condemns definitions that cannot be understood at all. New words left unexplained are insignificant sounds47, and “spirit incorporeal” is a contradiction whose definition is unintelligible48.
A humanist suspicion of definition runs alongside these logical limits. Cicero scorns philosophers who push definitions of “good” and “wise” to an accuracy of no practical use49, though he gives a definition of friendship himself50. Montaigne sees no need to anatomize death and pleasure by definitions and etymologies when he already understands them51. Sidney finds that learned definitions of the virtues, which moral philosophers boast of52, give only a wordish description that never strikes the soul53. Browne prefers allegory to rigid metaphysical definition of mystery54. Wordsworth objects to a definition imprisoned by etymology55. Arnold refuses to define poetic excellence in the abstract lest he darken the question56, and Emerson, warned by the philosophers’ fate, lists the qualities of beauty instead of defining it57. Goethe’s Faust is mocked for having defined God and man with unblushing confidence58.
When definitions are lacking, disputes follow. Hume diagnoses long-running controversies as the work of disputants who attach different ideas to ambiguous terms, and he holds that exact definitions would end them59. A few intelligible definitions, he thinks, would close the quarrel over liberty and necessity60,61. He blames undefined terms such as “innate,” and Locke’s loose use of “idea,” for prolonging dispute62,63, and he contrasts the exactness of geometrical terms with the ambiguity of moral ones64. Yet his own two definitions of cause are both drawn from circumstances outside the cause, and neither captures the connexion itself65. He presses his opponents to define the necessity they claim to find in material causes66, arguing that any definition of cause without necessary connexion yields unintelligible or merely synonymous terms67. Finally he concludes that non-mathematical propositions such as no property, no injustice are merely definitions68. Carlyle similarly treats the question whether Scott was great as a dispute about words69.
Berkeley dramatizes the same danger. Philonous makes Hylas define a sceptic70 and then amend the definition, since doubting is suspense of judgment while denying is as positive as affirming71. He repeatedly accuses Hylas of changing the signification of “matter”72,73. Shifting a definition at will is contrary to the rules of reason, and a thing whose ideas are repugnant to one another is impossible74. Material substance, on this test, has a repugnant definition75. Calling water fire would pervert speech76, so naming must accord with common and philosophical usage77. Even so, the two speakers insist they are disputing about meaning and not merely quarrelling over expressions78, and they note that philosophers have never agreed on a definition of identity79. Hobbes had already observed that disputers keep the meaning of terms of art only as long as it serves them80.
The remedy, often, is to distinguish senses. Smith separates the two meanings of “value”81 and the two senses in which a revenue is understood82. He tells the reader to attend to the nature of the thing rather than the word when bounties and premiums are confused83,84. Burke fixes a technical meaning for “Delight” because common language, framed for business, confounds distinct affections85. Long fixes the sense of “Nature”86, and Ruskin holds that a few words well distinguished do more work than a thousand used equivocally87. Hobbes separates worthiness from worth and merit88,89, and Swift separates good manners from good breeding90. Even Aristophanes’ Aeschylus defends his use of “come” against “return”91.
In the particular sciences the question becomes most searching in Darwin. No definition of species or variety satisfies all naturalists92. Ranking often rests on individual opinion93,94, and to dispute about species before any definition is accepted is vainly beating the air95. Species, he concludes, is a term given arbitrarily for convenience to closely resembling individuals96. Fertility, the favoured criterion, is used circularly97,98, and only an indefinite amount of difference remains to separate species from varieties99,100. The Natural System, which might be read as a scheme of likeness or a plan of the Creator, he interprets as community of descent101. Trifling characters therefore define groups because they are correlated with many others102. Some groups, such as the crustaceans, can be defined only by chains of affinity103, and none could be sharply bounded even if every form were known104. Darwin offers a definition of high organisation as the specialisation of parts105 while admitting that higher and lower can hardly be defined clearly106. He declines to define instinct107 and uses “struggle for existence” in a large, metaphorical sense108. Systematists, he predicts, will at last be freed from seeking the essence of species109. On this ground Darwin rejoins Pascal’s nominal definition and leaves Marcus Aurelius’s search for essences behind.
The other natural sciences seek precision rather than abandon it. Kelvin finds that sailors and landsmen mean different things by “tide”110, defines it by its cause, and then tests the definition by applying it to the sun111, holding that definition is needed to avoid error112. Pasteur insists that alcoholic fermentation be defined strictly, lest confusion of language pass into confusion of ideas113. He warns against identifying a fermentation by one of its products114, defines a germ without deciding whether it is animal or vegetable115, and reasons from the definition of a ferment’s work116. Faraday makes the learner distinguish gas from vapour117 and defines “soft” in terms of attraction between particles118.
In morals, religion and politics, definitions carry the weight of doctrine. Socrates secures agreement that death is the separation of soul from body119. Dante examines Paul’s definition of faith120 and defines hope121. Pascal defines a miracle as an effect exceeding the natural power of the means employed122, and observes that the ultimate end gives things their names, so that “enemy” means whatever hinders that end123. Hobbes replaces the schools’ definition of will124 and defines injustice as failure to perform covenant125. Marshall reads “necessary” in the Constitution as needful or conducive rather than absolutely indispensable126,127. Freeman finds the meaning of “race” threatened if community of blood cannot be proved128, settles on language as a test129, and defines the Greek by speech and faith130. Lowell complains that few try to learn what democracy really is131. Herodotus defines Egypt by its inhabitants rather than by continental boundaries132.
Letters bring the problem home to language itself. Johnson had resolved to limit every idea by a strictly logical definition133, yet he found that shifting senses defeat a reciprocal interpretation of word and explanation134 and that shades of meaning merge imperceptibly135. Sainte-Beuve traces “classic” to its Roman origin136, rejects the Academy’s narrow definition137, and concedes that his own exceeds usual notions138. Wordsworth finds Coleridge’s definition of fancy too general139. Fielding defines the comic epic in prose by its parts140, while noting that Aristotle never defined the Ridiculous141.
Several questions remain open. These authors agree that definitions should be clear. They do not agree whether definitions should bind the inquirer at the outset, as Hobbes and Pascal want, or reward him at the end, as Burke and Darwin’s practice suggest. Nor do they settle whether disputes about words are trivial, as Hume sometimes implies, or are disputes about things in disguise, as Berkeley insists.
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: Matter 5b · Language 3b
See also: Language 3b
See also: Evolution 2d · Animal 1c
See also: Language 3b
See also: Language 9c
See also: Language 9b
See also: Language, Evolution, Reasoning, Mathematics, Matter, Logic