Human beings everywhere and always have observed: the clouds that portend fair or foul weather; the heft and texture of the animal fur that will spin into the finest wool; the faint markings that distinguish a succulent from a poisonous berry; the path of the sun through the ecliptic that heralds the seasons; the slightly raised eyebrow that betrays a courtier’s intrigue; the effects of an herb on stomach pains; the discrepancies between words or hands that signal a variant text. Long before and long after scribal techniques such as writing, institutions such as libraries, and methods such as statistical correlation were invented, observers discerned subtle correlations between present signs and future events in both the natural and the human worlds. Because observed phenomena were often highly variable and sometimes unfolded on superhuman timescales, such correlations were the work of generations, codified, remembered, and transmitted in the form of proverbs: “Red in the morning / Sailors take warning.” Observational proverbs both antedate and postdate written versions, whether in the exhortations to sow and reap in conjunction with the movements of heavenly bodies in the Hesiodic Works and Days (ca. 700 BCE) or in the recommendations of the latest editions of the Farmers’ Almanac. Before the invention of reading and writing, sagacious observers were reading the world.
Observational knowledge is first and foremost correlational and predictive, rather than causal and explanatory. The Roman philosopher and statesman Marcus Tullius Cicero (106–43 BCE) described the correlations based on observation as “natural divination,” which he deemed more reliable than divination based on dreams, auguries, sacrifices, and prodigies. But this distinction between reliable and spurious correlations among observations has always been blurred, whether in the case of medical observations of the symptoms of a patient that point to a diagnosis or financial observations of the stock market that indicate boom or bust. Observation and divination seem to have been intertwined since ancient times, and the correlations that survive from that period exhibit the same difficulty in sorting out reliable from unreliable conjecture that bedevils modern analyses of *big data. In ancient Mesopotamia scribes correlated bad harvests with the likelihood of war and other misfortunes; in ancient Rome and in many other cultures portents as diverse as the flights of birds and the entrails of sacrificial animals were read as heralding future events. Some of the most ambitious, systematic, and long-lived observational projects ever mounted, such as the six-hundred-year continuous tradition of the astrometeorological diaries kept by Babylonian scribes from the sixth century BCE to 60 CE, mingle the most stable (astronomical) and unstable (meteorological) phenomena. The art of observation consisted not only in the careful, persistent monitoring of things and events but also in the choice of those phenomena likely to yield strong, durable regularities.
Observation has always been an art, dependent on sagacity, attention, and the trained senses, and no one has ever doubted that it is a useful, indeed essential art. But its status as a scientific practice has been more controversial, rising and falling as the definition of science itself changed. In the Latin Middle Ages, scientia was understood by university-trained Aristotelian scholars as certain knowledge of universal causes, in the best instance knowledge that could be shown to be necessary through demonstration. According to this scheme, scientific knowledge was explanatory but rarely predictive. Although the Aristotelian corpus of books about nature is rich in observations of all kinds, from the color of various animals’ eyes to the variety of smells and tastes, Aristotle himself considered such particulars to be history (historia, which embraces natural history as well the study of the past), a branch of knowledge preliminary and inferior to philosophy and even poetry, which deal in universals. Because Aristotle’s empirical remarks aim at philosophical universals, they are almost always generic in character, about species rather than specifics. The other great treasure trove of empirical particulars bequeathed by Greco-Roman antiquity to later learned traditions in the Middle East, Mediterranean, and later in northern Europe, Pliny’s Natural History (70s CE), did contain occasional items about individual cases (for example, a dog so loyal that it fished up its master’s corpse from a river) but for the most part also described genera rather than specifics. Only in the ancient medical literature, notably in the Hippocratic corpus, were individual cases recounted in detail, a tradition that was to be enormously consequential for scientific observation when Hippocratic medicine enjoyed a revival in sixteenth-century Europe.
It is notable that these empirical remarks in ancient Greek and Latin and medieval authors are seldom called “observations”: Aristotle used the Greek word teresis rarely, and then mostly in connection with animal behavior; the Latin word observatio, in keeping with its associations with the natural divination practiced by farmers, shepherds, sailors, and other outdoor laborers, was also not generally considered a learned activity, much less a philosophical category. The two most important exceptions were astronomy and medicine. In astronomy, observations of heavenly bodies stretching in some cases back to Babylonian as well as to Greco-Roman sources were at least partially transmitted through the works of the Alexandrian astronomer, astrologer, and geographer Claudius Ptolemy (ca. 100–ca. 170 CE) into learned traditions in, among others, Arabic, Persian, and Latin. However, the difficulty and expense of making new observations, much less of improving the sighting instruments (e.g., the sextant and quadrant) needed to improve their accuracy, meant that sequences of continuous observations were rare until the sixteenth century in Europe. Only a wealthy nobleman like the Dane Tycho Brahe (1546–1601), whose passion for astronomy was unusual enough within his aristocratic circles to have to be hidden from his family and tutors, could afford to devise new instruments like his great mural quadrant and to maintain a permanent staff of assistants (briefly including Johannes Kepler [1571–1630]) to conduct the long-term series of observations that eventually made Kepler’s remarkable innovations in the Astronomia nova (New astronomy) (1609) possible. Galileo’s (1564–1642) spectacular telescopic observations, reported in his Sidereus nuncius (Starry messenger) (1610), took the learned world by storm and also glamorized the once humble word observation in both academic and courtly milieus.
In medicine, the sect of the Empirics and the Skeptical philosophical school of Sextus Empiricus (ca. 160–ca. 210 CE) used observation as a term of art, and the word gained currency in learned circles in the mid-sixteenth century when Sextus’s works were partially recovered and translated into Latin. It is also notable that the Latin words experimentum and observatio were rarely found together in medieval texts, and neither referred primarily to learned practices: an experimentum was a test of the sort that an artisan might make in the workshop of a new recipe for dyeing cloth or tempering steel; an observatio took note of something or event, usually outdoors tilling fields, navigating by the stars, or watching flocks. Although university-trained scholars certainly observed their own world constantly, they did not theorize the word and practice as a path to scientia: observational knowledge was too contingent, variable, local, noncausal, and uncertain to qualify.
In the sixteenth and seventeenth centuries, this situation changed radically as voyages of exploration, commerce, and colonization to the Far East and Far West, technological inventions such as the printing press, religious *reformation and counter-reformation, and theoretical upheavals in astronomy, medicine, geography, and *philology deluged Europeans with novelties of all kinds: new flora and fauna, new peoples and languages, new drugs, new religions, new inventions, new stars, new continents, new cosmologies—as well as the rediscovery of old texts from antiquity read in new ways. Amid this ferment, the meaning of “science” began to shift, if only because knowledge once considered unshakably certain—for example, the geography of Ptolemy and Strabo or the cosmology of Aristotle—was revealed to be false on the basis of the observational discoveries of Columbus, Galileo, and others. The stock of observation as a learned practice rose accordingly.
Systematic observation structured by instructions, questionnaires, tables, and *genres emerged in several contexts in the sixteenth century. Northern European humanists wrote treatises on the art of travel, the ars apodemica, for their students about to embark on grand tours of Italy and other parts of southern Europe, full of minute instructions on what to observe and how to act: converse with everyone, but don’t divulge your religion, and take no books except your journal of observations; note the chief commodities of each region and their prices, the manners of its people, the court beauties and other favorites of the prince, the climate, the value of local coinage, and “infinite other particularities.” Such instructions aimed to improve the culture and character of the young gentleman and keep him too busy to frequent taverns and whorehouses in the sinful south. More economic and political in character were the questionnaires prepared at the order of Philip II of Spain in the 1570s to collect information about his possessions in the New World, with an eye toward governing colonies and exploiting resources. A century later, the Royal Society of London issued similar questionnaires to travelers in order to standardize natural history reporting.
Starting in medicine and astronomy in sixteenth-century Europe, and spreading to jurisprudence, natural history, philology, travelogues, and eventually natural philosophy by the mid-seventeenth century, the word observations started to feature prominently in the titles of both Latin and *vernacular books. Although the new genre was as diverse as its subject matter, its exempla exhibited certain common features: the observations were often collections, gleaned either from the author’s own experience or from other publications; they were organized as short, often numbered, entries, with little or no attempt at systematization; even in collections, firsthand observations (by autopsia) were granted privileged status; novelty was emphasized; and conjecture concerning the import of the observations was notably absent or even explicitly withheld. A seamless lineage connects these books of observations, increasingly designated as such in their titles, with the articles that appeared in the earliest scientific journals of the latter half of the seventeenth century. For example, the aptly named Miscellanea curiosa (Miscellaneous curiosities), sporadically published by the Academia Naturae Curiosorum (Academy of those curious about nature, later the Leopoldina, established in Schweinfurt in 1652, Germany’s oldest academy of sciences), was composed entirely of mostly medical observationes, numbered and titled as such, sent in by doctors from all over the Holy Roman Empire and beyond. “Observations” also featured prominently in the Philosophical Transactions published by Henry Oldenburg, secretary of the early Royal Society of London (established 1660) and the Histoire et mémoires (History and memoirs) of the Paris Académie Royale des Sciences (established 1666).
Along with the words experiment and history, observation emerged as a key practice in the self-styled new science of the late seventeenth century, and one increasingly theorized in relationship to the other two terms. Although Francis Bacon’s (1561–1626) call for a reformed natural philosophy based on a vastly expanded natural history influenced many of the proponents of the “new experimental philosophy” in the latter half of the seventeenth century in the Royal Society of London and elsewhere, his own use of the words historia, experimentum, and observatio had been quite fluid. By the late seventeenth century, however, these had become terms of art, designating refined forms of empiricism that were learned (as opposed to popular), distinct from each other, and grounded in *epistemological reflection. Whereas historia increasingly referred to compendious collections of information, both first- and secondhand, on a given topic (e.g., the winds), on the model of the natural histories sketched in Bacon’s own uncompleted Sylva Sylvarum (Forest of forests) (1627), experimentum and observatio referred to more particular inquiries. Experiment referred to an “artificial” (the word was Bacon’s) intervention in nature in order to produce a certain effect and thereby tease out its causes; observation referred to the close monitoring of nature undisturbed. Although the two words had rarely been conjoined in earlier centuries and still more rarely associated with scientia, by 1700 they were an inseparable and complementary pair, both emblematic of a new kind of science that produced probable rather than certain knowledge, aspired to prediction as well as explanation, and was pursued collectively through systematic inquiry into particulars.
The rise of observation as a scientific practice was propelled by a battery of techniques, some as old as note taking and intense attentiveness, others as new as making tables (well known in astronomy and other mathematical sciences since antiquity but new to disciplines like chemistry and natural history) and image making. Already in the sixteenth century, botanists, most of them physicians familiar with the new genre of medical observations, had lavishly illustrated their publications with the then-new image technology of woodcuts. By the late seventeenth century, the word observation could refer to an act of attentive perception, the verbal description of such an act, or an image (usually an engraving) accompanying the description. Works of observation, especially in anatomy, geography, and natural history, but also in booming new fields in physics such as the investigation of electricity and magnetism, were lavishly illustrated, and eighteenth-century savants such as Linnaeus vied for the services of the best artists. Observation also still carried resonances to a much older sense of the word as the faithful execution of certain obligations, especially carrying out religious rites. During the *Enlightenment, when the prestige of scientific observation was at its zenith, observation was more than a scientific practice; for virtuoso observers like the Swiss anatomist and naturalist Albrecht von Haller (1708–77) or the Dutch natural philosopher Pieter van Musschenbroek (1696–1761), it was an arduous way of life, demanding great dedication and often the sacrifice of money, social life, and health. Writing to von Haller in 1757, the Genevan naturalist Charles Bonnet (1720–93) vaunted observation as “the universal spirit of the arts and sciences”; observational prowess was the yardstick by which savants took each other’s measure; it was even possible to become a “genius of observation.”
Although observation never ceased to be an essential and steadily evolving practice in the arts and sciences, sharpened into a precision instrument for collating evidence, detecting patterns, and mounting arguments in the hands of great naturalists such as Georges Cuvier (1769–1832), Alexander von Humboldt (1769–1859), and Charles Darwin (1809–82), its star waned in nineteenth- and twentieth-century science—or at least philosophy of science. Prominent nineteenth-century men of science distinguished observation sharply from experiment, describing the one as passive and requiring minimal skill and the other as active and demanding the utmost ingenuity. In his Introduction à l’étude de médecine expérimentale (Introduction to the study of experimental medicine, 1865), the French physiologist Claude Bernard emphasized that “the mind of the experimenter must be active, that is to say he must interrogate nature and pose questions in every sense, following the various hypotheses that suggest themselves to him,” whereas the observer embodied “the passive senses that obeyed the intellect in order to realize an experiment designed with a preconceived idea in view.” The British astronomer John Herschel, in his Preliminary Discourse on the Study of Natural Philosophy (1830), demoted observation to an amateur activity, to be discharged by an army of volunteers who would diligently “observe regularly and methodically some particular class of *facts” and fill out standardized forms consisting of “distinct and pertinent questions, admitting of short and definite answers.” Twentieth-century philosophers of science took the devaluation of observation in science a step further by positing a “neutral observation language” that barely differed from mere looking—a form of empiricism too rudimentary to be suspected of being “theory laden.” In the late nineteenth and early twentieth centuries, new tools of mathematical statistical inference were developed that eventually led to the partial mechanization of the discernment of reliable correlations—for example, through significance tests, which had been the backbone of the long tradition of observation for millennia.
One sign of the waning scientific prestige of observation has been the massive enlistment of amateurs, most recently under the banner of “Citizen Science” and via websites like Zooniverse, in observation-intensive disciplines such as entomology and astronomy. Enormous trawls of data from satellites, CCTV cameras, and space-based telescopes have created an urgent need for volunteers to register new insect species, monitor video footage, and classify galaxies, and professional scientists are increasingly dependent on such assistance. To date, pattern recognition algorithms cannot rival the acuity and sagacity of the trained human eye of a conscientious observer: once again, sensory, epistemic, and moral qualities blend to make a virtuoso observer.
Whatever its fortunes as a scientific practice, the cult of observation flourishes among travelers, ethnographers, physicians, connoisseurs, detectives, and of course spies. Whether the task at hand is detecting a new disease such as AIDS or SARS on the basis of a scatter of symptoms in individual patients or the forensic examination of clues at the scene of a crime or judging whether a putative Vermeer painting is a forgery, the lynx-eyed observer detects what others overlook. As Sherlock Holmes famously reproached Watson: “You see, but you do not observe.”
Lorraine Daston
See also algorithms; cases; data; horoscopes; journals; knowledge; learning; libraries and catalogs; notebooks; printed visuals; teaching; travel