§38. A New Organon

The value of experiments is Bacon’s tireless theme. He knows that Aristotle made gestures toward experimental inquiry, but he finds them desultory and dogmatic, content to remain on the surface, never chasing down hidden causes.

Bacon's empiricism is musical, its relation to the senses orchestral; they are instruments whose value depends on the art with which they are played. We need observations and experiments for natural knowledge, as we need viols and pianoforte for music, but we also need a control that is not itself just more sensation. Natural science requires experience duly ordered and re­flected, not clumsy and erratic, everything depending on the art with which experience is controlled. What Galen called qualified experience is the moral of Bacon's beautifully imagined analogy. While the empirics of the mechan­ical arts, with their trade secrets and inscrutable rules of thumb, are like the ant, collecting and using without understanding, and the professors, the nat­ural philosophers of the schools, like the spider, making webs from their own substance, the wisdom of the bee is to gather material from the flowers but digest and transform it by its own powers, making something useful of what is collected.

Bacon is not a naive inductivist, as if he proposed merely to collect all the positive instances we find of quality x and generalize. Mere enumera­tion is puerile, and he understands perfectly (as had Epicurus, Philodemus, Grosseteste, and Ockham) the value of the disconfirming instance. “In the establishment of any true axiom, the negative instance is the more forcible of the two.” He certainly had no faith in the senses, which tend to miss nature's subtler parts. To rely on the senses is to measure nature in human terms and is fallacious. “For it is a false assertion that the sense of man is the measure of all things. On the contrary, all perceptions, as well of the senses as of the mind, are according to the measure of the individual and not according to the measure of the universe.”¹¹¹

Bacon's natural philosophy is about qualities. Nature is a mosaic of qual­ities which experiments sift in search of some few basic ones from which the rest are generated. To pursue this inquiry rationally we have to start by collecting histories of the different qualities, heat and cold, heavy and light, wet and dry, and so on. From these we make tables of positive instances, neg­ative instances, and concomitant variations and comparative intensities, which record the interaction of qualities with other qualities. First we search for relatively low-level generalizations before a second phase of inquiry ascends to higher-level generalizations and their experimental test. To go directly from a few instances to the universal, as Aristotle expected, is “the mother of error and the curse of all sciences.”¹¹²

Bacon’s most well developed example featured the quality of heat. Having made his experiments and constructed his tables, he comes to the point of a “first vintage” generalization on the form of heat, using the word “form” in his new sense of necessary and sufficient cause:

The form or true definition of heat (heat, that is, in relation to the universe, not simply in relation to man) is in a few words as follows: Heat is a mo­tion, expansive, restrained, and acting in its strife upon smaller particles of bodies.

Bacon did not want natural philosophy to be pure theory, as if contemplative truth were good in itself. He expected science to be operative, and his dem­onstration satisfies that condition too:

Viewed with reference to operation, [heat] is exactly the same thing. For the direction is this: If in any natural body you can excite a dilating and expanding motion, and can so repress this motion and turn it back upon itself, that the dilation shall not proceed equably, but have its way in one part and be counteracted in another, you will undoubtedly generate heat.¹¹³

That is not a shabby induction. The idea that heat might be an invisible corpuscular motion was not a commonplace, and with due allowance it re­mains true by our lights. His argument is unexpectedly cogent in context, the example of heat being especially subversive of scholastic natural philos­ophy, for which a distinction of primary and secondary qualities was a core doctrine, with heat the central instance of a primary quality. The heat of one thing making another hot is also their model of causation. But if heat is a kind of motion and not primary at all, then the Aristotelian theory of quali­ties, of causation, and practically the whole natural philosophy falls apart.¹¹⁴

We should never rest with experience, like an empiric ant. The ways of reason have to penetrate and organize experience with problems and hypotheses, and invent experimental conditions where they can be tested. Sensations, perceptions, observations, and memories become instruments of science only when their experience is “duly ordered and digested, not bun­gling or erratic.” The experience from which we most efficiently learn is not untutored, naive experience, but rather experience artfully orchestrated, ex­perimental experience. Experimental controls on perceptual data eliminate the worst anthropomorphism. “For the subtlety of experiment is far greater than that of the sense itself, even when assisted by exquisite instruments....

Bacon is criticized for supposedly failing to appreciate the value of a cre­ative hypothesis. Claude Bernard, a profound experimentalist whose ideas I shall turn to (§58), felt the value strongly, and laid down as an “absolute principle” that experiments “must always be devised in view of a precon­ceived idea, no matter if the idea be not very clear nor very well defined” “We must give free rein to our imagination; the idea is the essence of all rea­soning and all invention. All progress depends on that.” Reading Comte, who takes this line, may have confirmed Bernard's own experience, and he repeats the canard, probably from De Maistre, that Bacon neglects this cru­cial element of experimental research. Bernard dismisses Bacon's methods as “vague and inapplicable precepts that we must hasten to forget if we wish to become true experimenters.” Even less hostile critics think Bacon leaves no place for the creative function of hypothesis. “If Newton had taken Bacon for his master, not he, but somebody else would have been Newton”; “Whether we look to Galileo... to Gilbert... or to Newton and Descartes, Leibniz or Huygens... we find that discovery was achieved by the opposite method of that advanced by Bacon.”¹¹⁶

This is much too easily said. A philosopher of later science observes that “the method of hypothesis is beyond question to be found in Bacon's work,” eloquently refuting De Maistre's sneering claim that no scientist ever made use of Bacon's method. “The instrument in itself is an excellent one and—in one or another of the particular forms that it takes, according to the kind of material dealt with—is fruitfully employed by scientific inquirers every day.” Tables of presence and exclusion are theoretically sufficient to isolate the nec­essary and sufficient cause (Bacon called it a “form,” usage Boyle follows) of the qualities.

Bacon's method is never undirected, his questions expressing hypoth­eses and suggesting experiments. “My course and method... is this—not to extract works from works or experiments from experiments (as an em­piric), but from works and experiments to extract causes and axioms, and again from those causes and axioms new works and experiments.” Nor were his natural histories a mere catalog of qualities, being carefully structured, often quantitative, compiled with the use of instruments and experimental techniques, adequately described for replication, and organized in a self­augmenting chain of experiments generating experiments. Under this dis­cipline experience becomes learned, literate, more consistently experienced experience, clarified and consummated by systematic experiment. Illiterate experience is barren, superstitious, and obsessive, Gilbert on the magnet being his example of experiments not informed, as they should be, by hypo­thesis, and resulting in a sheaf of disconnected facts (a mistaken criticism as we shall see). This is illiterate experience, undisciplined, the experiments of a sleepwalker. Literate, lucid, experimental experience is the right way to dis­cover nature’s alphabet and tease out her operations.¹¹⁸

Despite dogged modernity Bacon retained a scholastic expectation of certainty for the conclusions of natural philosophy. He did not suppose cer­tainty could be attained all at once, but that was the goal. The preface to his New Organon (1620) distinguished “probable conjecture” from “certain and demonstrable knowledge,” and speaks of “progressive stages of certainty,” idealizing a knowledge ultimately complete and immune to error. Shoring up the expectation of certainty is the assumption that his new logic is quasi­automatic, placing “all wits and understandings nearly on a level.” He seems not to consider that conclusions about unobserved cases induced from premises about the observed must remain fallible.¹¹⁹

Bacon shifts the ground of scientia, from Aristotelian demonstrations to experimental inductions.