“Information fatigue” has been called a distinctively modern phenomenon, but reference works—a culture’s most visible attempts to manage *information overload—are found written in cuneiform on clay tablets and in Egyptian hieroglyphs on *papyri. The putatively modern phenomenon is nearly as old as writing itself.
Reference as we know it, however, can be no older than writing; it is necessarily a written form. Oral cultures may have to reckon with their own forms of information overload, but their solutions differ from those of literate cultures. Those without access to writing can manage only as much information as fits in memory but writing makes it possible to store the overflow in books. Beyond reference works’ origins in writing, though, it is notoriously difficult to make generalizations about them, and there is no widely accepted definition of the form. They are so diverse in their contents, organization, and readership that it is probably wrong to refer to them collectively as a *genre. Any area of endeavor in which the amount of important information is too much for one memory—whether as elevated as the vocabulary of patristic Greek or as demotic as the market prices of baseball cards—will eventually be treated by at least one reference work. Reference books define words (whether of the entire vocabulary of a language or some specialized subset of it), identify library holdings, name stars and galaxies, locate half-remembered quotations, predict the phases of the moon, warn about the counterindications of medicines, compile the results of mathematical functions, and provide elaborately indexed tables of folkloric motifs and postage stamps and Jewish surnames and French watermarks.
No one has ever enumerated all the varieties of reference works—dictionaries (monolingual and polyglot, general and subject specific), *encyclopedias (general and specialized), bibliographies, atlases (terrestrial and celestial), classifications of maladies (physical and mental), almanacs, *concordances, chronologies, legal compendia, compendia of sports statistics, biological *taxonomies, trigonometrical tables, artists’ catalogues raisonnés, collectors price guides, and so on—let alone provided a comprehensive bibliography of the individual titles. Such a list would surely run into the millions and would include works in most of the world’s written languages, living and dead. There is, moreover, no agreement on the outer periphery of the category: are cookbooks, real estate listings, railway timetables, and newspaper horoscopes properly reference books? It would be difficult to make a principled case for excluding them, though in practice we reserve the term for works of a certain magnitude and gravitas.
Every reference genre has its own history or, more accurately, histories, since similar forms are often developed independently in distant cultures. Dictionaries tend to follow a pattern in most places where they appear: the earliest ones are either bilingual, subject specific, or devoted entirely to “hard words”; only later do general monolingual dictionaries that strive to cover the entire vocabulary appear, and historical scholarly *lexicons are a product of nineteenth-century scholarship. Legal *codes typically mark the high-water marks of empires; imperial conquest and exploration spur the production of new bilingual dictionaries and *gazetteers. Perhaps the most complex history is that of the encyclopedia, a term that was first applied to a reference genre in the seventeenth century, though countless ancient works—Greek, Latin, Sanskrit, Chinese, Arabic—can be called “encyclopedic,” leaving open the question of what precisely constitutes the genre. Some, like Pliny’s Naturalis historia (ca. 77–79 CE), inhabit the boundary between encyclopedia and science textbook; others, like the Chinese leishu, combine encyclopedic scope with the form of copious anthologies sometimes running to hundreds of millions of words.
The boundaries between the various reference genres are often unclear, and different languages demarcate the field in different ways. Speakers of English use Greek lexicon, for instance, as a synonym for dictionary, often with the connotation of a dictionary of the classical languages; in German, Lexikon is often used where Anglophones would use encyclopedia. Peter Mark Roget repurposed thesaurus—also used for dictionaries, medical texts, and textual corpora—to serve for what had been called a synonymy. Speakers in English sometimes limit dictionary to alphabetically organized works covering the vocabulary of one or more languages, but in many compounds—biographical dictionary, historical dictionary—it overlaps with encyclopedia, and dictionaries of proverbs and dictionaries of quotations are other kinds of compendia altogether.
Not content, however, but organization is paramount: a reference book is a work structured to facilitate rapid consultation. This means its parts—“entries”—are discrete, comprehensible individually, and organized to make finding them as efficient as possible. Because reference books are meant to be consulted in parts, the works as a whole often seem unreadable. Most books have readers, but reference books have users. Their creators, too, are often thought of not as writers but as compilers. While the names of a few lexicographers and encyclopedists are familiar and a few of their works express something of the character of their creators, the twenty-first-century norm is that reference books are prepared by teams of scholars—dozens, hundreds, even thousands—working in near anonymity.
It is the responsibility of those teams not only to write the entries but to put them in a useful order. Some information has an obvious ordering principle—tables of logarithms and chronologies, for example—but for information not intrinsically ordered, some structure must be imposed to make rapid access possible. History is full of examples of struggles to find “natural” modes of organizing complexity. Ancient encyclopedias and glossaries are often organized topically in an effort to keep like with like. In 1852 Roget followed a number of ancient models, including Philo of Byblos’s Greek synonymy (early second century CE?) and the Sanskrit Amarakosha (Immortal Dictionary) (ca. fourth century CE?), when he worked to classify nearly the entire English vocabulary: the top-level classes of his taxonomy are “abstract relations,” “space,” “matter,” “intellect,” “volition,” and “affections,” with dozens of subheads covering more than a thousand categories. Such systems, though, however perspicuous they seem to their creators, have proven difficult for users, and so many compilers give in to an arbitrary order—since the fourteenth century, typically the one provided by the alphabet. Though it offers ease of use, alphabetical order has the disadvantage of scattering related terms throughout the work (cat, feline, tiger) while grouping unrelated terms through accidents of orthography (casuistry, cat, catachresis), and so the quest for rational organizations persists.
We usually speak of “reference books,” but reference has been embodied in virtually every kind of information technology, and reference books are more liable than most written forms to be shaped, enabled, and sometimes threatened by developments in the technologies of the word. The earliest surviving *glossary, the bilingual Urra=hubullu from the second millennium BCE, translates between Sumerian and Akkadian in a series of clay tablets; the oldest surviving Chinese dictionary, the Erya, was probably written on bamboo strips sometime around the third century BCE. Many thousands of reference works date from Greco-Roman antiquity, including geographical surveys (Claudius Ptolemy’s Geography, composed ca. 150 CE), legal codes (Tribonian’s Corpus iuris civilis [Corpus of civil law] dates from the early sixth century CE), and proto-encyclopedias (Isidore of Spain’s Etymologiae [Etymologies] appeared in 636 CE). While the earliest of these may have first circulated on papyrus scrolls, none such survive. Instead the *codex form of a *parchment manuscript, bound at the spine, spread across the ancient world starting in the second century, along with Christianity, and encouraged the kind of random access that was difficult with scrolls. Vast works like the Domesday Book, more than sixteen hundred pages, compiled in 1086, and biblical concordances, a genre that dates to the early thirteenth century, became practical with the advent of easily turned pages. Paper appeared first in China and spread to the Islamic world, where it was the medium of choice for Avicenna’s medical compilation the Qanun (Canon) (1025) and Muḥammad ibn Mūsā al-Khwārizmī’s adaptation and expansion of Ptolemy, the Kitab surat al-Ard (Book on the appearance of the earth, early ninth century). The arrival of paper in Europe (the eleventh century in Italy and the fourteenth century north of the Alps) was a crucial prerequisite to the success of printing in the fifteenth century. Printing created new norms, including pagination and indexes that could serve every copy in an edition. In the nineteenth century the steam press and cheap wood-pulp paper helped to place dictionaries and encyclopedias in every middle-class household. Stereotype and electrotype changed the economics of reprints and revised editions.
We are only beginning to understand some of the consequences of the electronic revolution. Printed reference works are economically inefficient: almost by definition, they are useful to their readers only in parts. The user of a twenty-volume encyclopedia might, over the course of a lifetime, consult only a few hundred of the work’s ten thousand pages. For centuries, therefore, there has been an imperative to keep reference works as concise as possible to minimize materials costs. The *digital dispensation has changed those calculations, though, and freed from print, we have no need to “save space.” The materials cost of a ten-thousand-entry encyclopedia is now effectively the same as for a one-hundred-thousand-entry encyclopedia.
The change in the economics of the materials, however, has had a spillover effect on the economics of compilation. Reference works are often huge projects that stretch out over decades. The New English Dictionary, intended to be completed in four volumes and ten years, grew over the course of seventy-five years into the thirteen-volume Oxford English Dictionary, at which point it was time to begin a new edition—and as soon as that one was finished, the project began again, occupying hundreds of professionals over the course of a century and a half. Even without substantial materials costs, therefore, someone has to pay for the intellectual labor that goes into reference publishing. But the distribution of old reference works online and the creation of new ones, without any cost to the user, has depressed the value of that labor and inspired a mantra for demoralized twenty-first-century “content providers”: “You can’t compete with free.” We are now participating in a culture-wide experiment to determine whether reliable reference works can be assembled by unpaid and uncredentialed volunteers without central oversight. The number of specialized reference works continues to grow, their costs generally borne by institutional libraries, but more general works—dictionaries, encyclopedias, atlases—struggle to produce enough revenue to sustain their creation.
Still, constant creation is a necessity, for new reference works will always be needed to replace the old ones. Reference books are uncommonly subject to obsolescence, and only a very few older examples are as useful now as when they were created. Reference works are valuable for the information collected in them, but *facts themselves, our understanding of facts, and our interest in various kinds of facts are all constantly changing, making old reference works ever less useful. The most many reference works can hope for is to go through multiple editions before becoming obsolete. Novelty sells reference books: dictionaries tout their thousands of new words since the last edition, and encyclopedias boast of up-to-the-minute currency.
Obsolete works, however, have their own attractions for historians of information, for they give remarkable insights into the mentalities of their creators. If we want to know facts about the world there is no reason to turn to Isidore of Seville or the Encyclopédie of the French *Enlightenment. If, however, we want insight into how the church fathers or les lumières saw the world, few sources are more valuable than the reference books they created, both for what they consciously chose to tell us and for what they unintentionally reveal about the way they saw the world. Their selection tells us what information they considered most vital, and their method of organization gives us insight into how their creators mentally organized their worlds. Though we may read them for very different reasons than those of their original users, even the most obsolete reference books demand attention.
Jack Lynch
See also bibliography; excerpting/commonplacing; indexing; letterpress; maps; scrolls and rolls