`LIBRARIES AND CATALOGS`

A remarkable chapter of the Austrian writer Robert Musil’s The Man without Qualities (1930) bears the title “General Stumm invades the State Library and learns about the world of books, the librarians guarding it, and intellectual order.” Vexed by the lassitude of the government committee on which he served, the besabered Stumm ventured into the Imperial Library in Vienna. The general expected quick results from a state information repository. He instead encountered “book stands and library tables piled high with catalogs and bibliographies, the concentrate of all knowledge, don’t you know, and not one sensible book to read, only books about books.” With technical precision he reported that the place “positively reeked of brain-phosphorous.”

By Musil’s time, libraries had emerged as the central information-processing centers of the industrialized West. Modern technological developments—the mechanical cylindrical printing press, integrated rail and postal systems, the telegraph, electric power, telephones, typewriters—coupled with social changes in the professionalization of knowledge created a communications environment in which increased volumes of information moved with unprecedented rapidity. Libraries served as both warehouses and clearing centers for the mass of information contained in books, periodicals, and the myriad publications of industry, commerce, and government.

The modern library aimed not simply to stockpile learning but to connect knowledge sites and manage information flows. Its creators recognized that the production of knowledge was a dynamic, ongoing process. Professional training, “scientific” methods, and “modern” procedures were confidently introduced into the space of the library. Stumm himself had been admitted into the library’s catalog room—“the holy of holies”—only after asking “about some sort of railway time-table that would make it possible to get cross-connections between ideas going in every direction.” The librarian, a sometime professor of library science, mistook Stumm for one of his own and steered the general toward a “bibliography of bibliographies.”

The largest of the new libraries—the British Museum Library (1857), the Bibliothèque nationale in Paris (1868), the Library of Congress (1897), the New York Public Library (1911)—were sprawling book machines that processed thousands of new books and periodicals each year thanks to specialized staff and devices such as card catalogs, conveyor belts, book lifts, and pneumatic tubes delivering call slips.

Libraries were everywhere. The popular lending libraries that dotted the urban landscape furnished day-to-day familiarity with library practices and paraphernalia. Government, business, and industry were served by specialized libraries—sometimes called “information bureaus”—that provided access to technical reports, scientific documentation, and institutional records. Meanwhile a large research library was regarded as the intellectual center of the modern university. Though differing in size and scale, libraries were the ubiquitous, everyday information-processing sites of the predigital age.

`LIBRARIES, SPACE, CATALOGS`

From the perspective of the history of information, libraries furnish an important example of sites where the recording, ordering, and sharing of large quantities of disparate data was an organized social activity. Library technologies such as catalogs and indexes were sometimes adapted to the organization of knowledge in other social spheres such as government and commerce. And library data management practices contributed to the early *digital landscape in meaningful ways.

Even before industrialization libraries functioned as sites of interconnectivity that managed information flows and provided an interface between people, books, and knowledge. Libraries grew exponentially in size from the *Renaissance to the early twentieth century. Remarkably the core practices for accumulating and structuring bibliographic data—describing, classifying, abstracting, indexing—remained much the same. Libraries combined assemblages of manual procedures with a variety of handwritten and printed media to structure information, order collections, and guide readers to books. Though catalogs changed in form and function over the centuries, they were always intimately connected to library space.

The physical space of the library was an essential element of library “information architecture.” Premodern libraries offered a deeply embodied experience of knowledge. Standard library furnishings of the Renaissance featured long, sloped lectern desks with books stored on shallow shelves either above or below the reading surface. Books were often chained in place, and readers were accustomed to moving around a library to read. By the seventeenth century books were more commonly kept on wall shelves and read at tables. Many libraries adopted a plan of jutting alcoves with accompanying desks in order to maximize shelf space. Books were normally arranged by subject: theology, law, philosophy, and so on. It was a rare library that possessed more than a few thousand volumes.

Catalogs were normally topographical inventories that itemized books according to their physical arrangement in the library. There was little preoccupation with elaborate description or complex classification. When the French librarian Gabriel Naudé observed in 1627 that the best order for a library was “that which is the most common,” he spoke the library wisdom of the ages. Classification schemes aimed to reflect shared intellectual categories. Memory, visualization, and embodied familiarity with library space played an important role in finding books.

Even as collections grew in size over the course of the seventeenth and eighteenth centuries, the cognitive contours of the library changed little. Library architecture enhanced the embodied encounter with knowledge. The Benedictine scholar Oliverius Legipontius observed in 1747 that “looking at a book collection should be like seeing the entire history of learning in a mirror.” Catalogs, in turn, sought to replicate and enhance a library’s visual order. A subject catalog reflecting shelf order was often the only working tool available to librarians and readers. As Naudé explained, subject catalogs allowed readers to see “in the twinkling of an eye” all books available in a given area. Library architecture responded with visual and spatial cues. Subject headings or pressmarks were inscribed above bookcases. Copies of catalog pages on individual sheets that were sometimes posted on library furnishings underscore the close relationship of catalogs and the spatial environment.

`Figure 1. A bird’s-eye view of the interior of Leiden University Library in 1610. Books are visibly arranged in the subject order of the library’s printed 1595 catalog. The lettering on top of the cases, which in reality would have been visible to no one, is an engraver’s conceit—subject classification was no doubt communicated differently. Engraving by Willem van Swanenberg. Image courtesy Leiden University Libraries (COLLBN Port 315 3 N22).`

The apotheosis of this phenomenon lies in the fantastical sculpted baroque libraries whose galleried, often curved library walls seemingly enfold the reader in bibliographic embrace. Books not only were stored openly; they were displayed. “Knowledge lies here” was the tantalizing promise, quite literally within grasp, even when guardrails and librarians intervened. There was no pretense that a library might contain all books. Rather, the library constituted a prism of universal knowledge that refracted but did not reproduce a larger order.

`PRINT, DATA, ORGANIZATION`

The *early modern print explosion meant that the informational tools used in libraries were called on to manage increasing volumes of data. Basic procedures for making catalogs remained the same for centuries. The first step was to arrange books in subject order. Most classification schemes were thematic and hierarchical. Books were grouped into broad subject areas—theology, law, history, literature, and philosophy were fairly standard—and subdivided into further subclasses that varied according to the size and nature of the collection. The next step was to draw up a shelf list to serve as a subject catalog. Empty shelf space was left to accommodate new acquisitions, and blank pages were left in catalogs. Alphabetized author tables or keyword indexes might be made afterward. This method was used, for example, to organize the Bibliothèque du Roi in Paris in the seventeenth century (twenty-three subject classes in fourteen volumes, cataloging some sixty thousand volumes; accompanied by a twenty-one-volume alphabetical index) and the library at the University of Göttingen in the eighteenth century (eighty-five subject classes in 147 volumes). While loose slips or cut sheets—the embryonic elements of the modern card catalog—were used as aids to sorting and arranging, a bound catalog in book form remained the habitual media support for storing bibliographic information.

Figure 2. The English architect Sir Christopher Wren designed an influential interpretation of alcove library architecture for the library of Trinity College, Cambridge (1695). Reading desks are situated in the midst of each alcove, with books an arm’s length away. At the end of each bay the Roman lettering used for shelf notation is prominently displayed. Wooden tablet-like leaves beneath open to reveal written lists of the contents of each bay. Photo: James Kirwan. Image courtesy Wren Library.

Each bank or case of shelves was assigned an identifier, very commonly a capital Roman letter. In some cases it served secondarily as a subject class *code. Individual shelves were assigned a distinct identifier such as a lowercase letter or Roman numeral, with individual volumes on each shelf assigned a number using Arabic notation. In other instances books were numbered sequentially within each class. In isolated cases superscript notation (a precursor of decimalization) was used to incorporate additions. What is now called the shelf notation or call number might serve at least four different functions: to provide a distinct alphanumeric identity for an individual volume; designate a book’s physical location; indicate a subject class; and furnish a universal notation for cross-referencing bibliographic data across different types of catalogs and finding aids.

To modern eyes the system is perplexing. The fixed character of both the physical arrangement of books and the bound catalog appears to condemn librarians to a purgatory of endless shelf rearrangement and recataloging. Yet new accessions were the exception rather than the rule in premodern libraries. Systematic shelf arrangement also offered frank recognition of the fragility of catalog systems and the enormous labor involved. No matter how approximate a physical subject arrangement, rooting around on library shelves could be the first as well as the last resort of any investigation. The alcove architecture adopted by many medium-sized libraries in the seventeenth century, and that endured as a model for more than two hundred years, actively fostered this sort of wandering, tactile encounter with books. It is telling of the entire system that some regarded catalogs as intellectual crutches, a poor substitute for the flexible, multifaceted “place” memory of the librarian.

The tools developed by Hernando Colón, son of the famous navigator Christopher Columbus, provide a snapshot of available premodern library information practices. Colón inherited his father’s wealth as well as his name and amassed a large library—some fifteen thousand volumes of printed books by the time of his death in 1539. Just as importantly, he devoted money and energy to its organization. An unusually large number of staff worked in the library and under Colón’s direction produced an ambitious array of catalogs, indexes, and other works. The primary information tool was a register of accessions where bibliographic data was recorded as books were purchased and assigned a number. The accession number signaled the physical location of a book and served as an identifier for cross-referencing in other catalogs. The register was supplemented by a suite of author, title, and subject catalogs and indexes. The most remarkable of Colón’s working tools is the recently discovered Libro de los Epítomes, or “book of summaries.” Assistants were employed to read Colón’s books and prepare abridgments. These were then reviewed, corrected, and copied into the Libro. Though the contours of the system changed slightly as the collection grew in size, its essential elements remained in place. A torturously complex system of numbers and symbols purportedly enabled staff to move from one informational device to another and to locate books on the shelves.

Colón’s system was complex and unique. It nonetheless shows how catalogs and indexes could be used to organize a large number of books and share knowledge-building information. Tellingly, many of Colón’s catalogs remain unfinished. Labor intensive, they were the work of years rather than months. Most early modern libraries had a sole, usually ill-paid, librarian who often performed other duties. Only larger institutions might have assistants, copyists, and the like.

Conrad Gessner provides a strikingly different example of how bibliographic data could be adapted and organized. While Colón sought to bring order to his library in Seville, Gessner sought to catalog the world of books writ large. When he coined the term universal library a few years after Colón’s death, it was not because Gessner had access to a fabulously stocked library but because he did not. The town library in his native Zurich possessed a meager one thousand volumes.

Hailed as a landmark of bibliography, Gessner’s Bibliotheca universalis (1545–48) offered readers a panorama of the world of Renaissance books. Published at the tail end of the first century of print, Gessner’s bibliography inventoried the available stock of learning and assembled what we would now consider a virtual library catalog. Of interest are the tools Gessner employed: earlier printed bibliographies; printers’ catalogs; handwritten library catalogs; a network of correspondents; the book collections of members of his circle in Zurich and nearby Basle. He also distilled data from title pages, *colophons, and tables of contents. And in the manner of Colón’s Epítomes he furnished readers with the fruits of his own reading. A reference tool intended for scholars, the Bibliotheca itemized some twenty-five thousand Hebrew, Greek, and Latin titles. Gessner intended his “universal library” to furnish a guide to library acquisitions, and he suggested that it could be adapted to serve as a library catalog.

Gessner showed how bibliographic data could be disaggregated from the localized space of the library and assembled to create an independent tool of knowledge. After Gessner a range of publications established norms for the description and classification of books that intersected with the technical practices of libraries. The year 1564 witnessed the publication of the first printed catalog of the semiannual Frankfurt Book Fair. Printed library catalogs furnished another important source of bibliographic information. The earliest of these were the catalogs of the University of Leiden (1595) and the Bodleian Library at Oxford (1605). Book reviews soon became a staple of the new periodical press, while subject bibliographies diffused more specialized bibliographic information. Like Gessner’s “universal library” many works embraced the ideal of a virtual library: Bibliotheca medica, Bibliotheca theologica, and so on.

The Bibliotheca universalis and its successors were the precursors of the “bibliography of bibliographies” that later confounded General Stumm. During the centuries that separated Gessner’s Zurich and Stumm’s Vienna, this informal network of bibliographic information coalesced into a regulated system that seemingly found its natural home in the bowels of the centralized state library. The transition was gradual, spanning almost four hundred years.

`CARDS, STACKS, NOTATION`

From the Renaissance to the *Enlightenment the ideal of a well-ordered library changed little. The visual display of a collection was of paramount importance. Books were arranged by subject order and accompanied by classified catalogs, alphabetical indexes, and other finding aids.

In the nineteenth century increased volumes of new acquisitions placed urgent pressures on the processing of bibliographic data and the physical storage of books. It is estimated that industrial print technologies triggered a sixfold increase in the annual production of books and other materials over the course of the century. In Germany, which dominated the field of scientific and technical publications worldwide, the number of journal and magazine titles increased by a factor of eighteen, effectively doubling every twenty-five years.

Shifting knowledge paradigms together with increased production capacity meant that new acquisitions were a daily occurrence rather than a rare and privileged exception. By the early twentieth century periodical publications played an unprecedented role in the dissemination of knowledge. The triumph of disciplinary knowledge meant that researchers no longer strived to master defined traditions—which libraries had sought to map and mirror—but to create new knowledge.

The crux of the problem lies in the relationship between multiple enterprises: processing a rising tide of bibliographic information; incorporating accelerating rates of new acquisitions into established collections; storing and locating unprecedented numbers of materials on library shelves; and somehow communicating endlessly shifting pools of information to library patrons through catalogs and indexes.

Librarians experimented and theorized in response. The most important architectural innovation was the invention of the book stack. Built of iron or steel, compact book stacks stored rather than displayed books. The first book stack was erected in the new British Museum Library that opened in 1856. Readers were seated in a single vast, circular reading room. The library’s bound catalogs ringed a central reference desk. Rows of reading desks radiated from the catalogs, while reading room shelves displayed some forty thousand reference works. Invisible to readers was a warren of forty-two kilometers of multitiered iron shelves that surrounded the reading room. The stack system was incorporated into the construction of the new Bibliothèque nationale in Paris (1868) and was rapidly advocated by German and American librarians. From Helsinki to Melbourne, stacks furnished the dominant architectural model for libraries across the Western world.

The arrival of the book stack reflected a wider preoccupation with the differentiation of function in the organization of library work and the distribution of library space. Just as books and readers were separated, so too were work spaces: librarian’s offices, acquisitions and cataloging departments, catalog rooms. In tandem with the increased professionalization of librarianship—library schools, cataloging manuals, professional associations, and the like—the reorganization of library space triggered experiments with the storage, manipulation, and display of bibliographic information.

Removing books from public view meant that catalogs now furnished the sole access point to books. This was true not only for readers, but also for librarians—“stack work” was carried out by clerical staff. This opened the door to the end of systematic shelf arrangement by subject. In the British Museum Library books were originally sorted into 650 subject groupings and then shelved by size and order of accession. Accession-order shelving was taken to further extremes elsewhere. Shelving by size and accession order alone was adopted by the new Bibliothèque nationale facility. It was soon introduced into other Parisian libraries, and even enshrined in law. Instructions for university libraries issued by the French Ministry of Education in 1878 mandated that books be arranged by size and order of accession only.

In this system the accession number becomes an integral part of the physical volume, which permanently identifies it. The accessions register was the lynchpin of the enterprise. Much as the book stack warehoused books, the register archived rather than organized information and “mechanically” generated a number that functioned as a stable identifier for an individual volume. The principle became known as the numerus currens (running number) system. The manipulation and display of bibliographic information was an entirely separate endeavor. This was performed by author and subject catalogs that referenced the accession number.

`Figure 3. The New York Public Library (1911) was one of the largest new book palaces of the modern era. This cover illustration from the popular magazine` `Scientific American` `shows the seven-tiered fortress of steel book stacks that fed—and physically supported—the library’s monumental reading room. Stacks, book lifts, pneumatic messaging tubes, and library staff are all in plain view. The central reading room, depicted upper left, is glimpsed through a rear view of the arched portals of the book delivery desk. Image courtesy New York Public Library.`

Though pragmatically compelling, there was resistance to accession-order arrangement. It never really took hold in the English-speaking world. In Germany, where book stacks were enthusiastically adopted, many librarians considered accession-order shelving an act of *epistemological violence. In 1912 Georg Leyh, a leading German academic librarian, launched a concerted attack on what he called “the dogma of systematic arrangement.” Leyh argued that the system served no one. Readers had long been denied access to library shelves, while the clerical staff employed to fetch books used systems of shelf notation without referencing classification systems. Most large libraries in western and central Europe gradually abandoned arrangement by subject order. Books were placed in broad subject groupings upon accession, assigned an accession number (the numerus currens), and varyingly cataloged by author, title, and subject.

The arrival of book stacks intersected with the rise of the card catalog. Recording bibliographical data on paper slips was not a new idea, though it was usually reserved for the preliminary stages of cataloging. Preparatory materials for one of Colón’s catalogs, for example, record information on cut sheets that were subsequently stitched together for safekeeping. By the end of the nineteenth century the use of cards and slips had become the norm. Individual catalog entries were recorded on stiff cards and inserted in drawers in standardized wooden catalog furniture. Nonetheless in many quarters card catalogs were regarded as an administrative efficiency rather than a terminal solution. Library manuals of the late nineteenth century usually advocate the use of cards, often in duplicate and triplicate, for the accessions catalogs and alphabetized author catalogs maintained by library staff. It was still foreseen, however, that the principal public catalog would be a classed subject catalog in bound book form. If printed, all the better. Critics of public card catalogs complained that cards were easily lost or misfiled. It was argued that information stored in tightly packed drawers was difficult to read and unduly limited the number of readers able to consult the catalog. Some found that isolating bibliographic information on cards impeded the visualization of aggregate data afforded by the display of multiple entries on a single page.

Even in Germany, where the ideal of the bound, classified subject catalog or Realkatalog had reigned supreme for over two hundred years, the combined advantages of cards, stacks, and numerus currens notation was gradually recognized. Critics of the bound catalog pointed out that in most libraries a bound subject catalog remained little more than a pipe dream. In 1906 the *philologist Hermann Diels mocked the entire system, referring to the “arcanum” of modern library catalogs and the “bibliographic hieroglyphics” that linked catalogs to one another. By midcentury most large European libraries used a combination of numerus currens shelf notation, closed stacks, and card catalogs.

The Anglo-American library world went in another direction entirely owing to the implementation of “relative” classification systems. The brainchild of Melvil Dewey (1851–1931), the most vocal exponent of the American “library movement,” relative classification systems used a standardized number of subject areas that were assigned either a numeric or alphabetic value. Dewey’s system used ten categories, which were assigned numbers: 001–099 (general), 100–199 (philosophy), and so on. Charles Cutter (1837–1903), a librarian first at Harvard and subsequently at the Boston Athenaeum, created a rival system of “Expansive Classification” using alphabetical codes. The range of codes could expand or contract depending on the size of the collection being classified. A modified version of Cutter’s method became the *Library of Congress (LC) Classification system. Similar systems were adopted in the United Kingdom. In all cases, numeric or alphabetic class codes are followed by forms of numeric or alphabetic notation (sometimes in combination) that allow a book to be assigned a unique alphanumeric identifier and inserted into an appropriate position in its subject class.

Though the relationship between card catalogs and the development of relative classification is unclear, the two were intimately connected. Dewey and Cutter were both card evangelists. It is possible that habituation to the ease of inserting new cards into an alphabetized author catalog without disrupting the catalog’s principal ordering sequence prompted experimentation with subject classification along similar lines. In this regard the truly unique aspect of the American innovation was the application of information management techniques to physical shelf arrangement.

The advantage of the various relative classification systems is that the “mechanically” generated notation does not refer to the fixed location of a book. It instead provides an “address” for a book defined by its relative proximity to other books in the same class. New books could be added without resequencing the entire class. When coupled with the card catalog, the new methods allowed for systemic efficiencies. Cards and books alike could be added at any time without disrupting established order.

Cards, stacks, and notation schemes assumed varied configurations. A version of the Dewey Decimal Classification system—the *Universal Decimal Classification (UDC)—was developed in Europe, where it was often employed in conjunction with the numerus currens shelving system. In other words, UDC was used to index rather than order a library’s holdings. In the English-speaking world the convenience of classified shelf arrangement was not lost on readers. Public libraries gradually acquiesced in allowing readers access to book stacks, and the practice spilled over into university research libraries. Globally, public libraries and specialized research libraries normally adopted some version of decimalized classification and open stacks. Decimal classification has proven especially adaptive to non-Western cultures where Roman lettering does not enjoy “universal” recognition value. Both Korea and Japan, for example, employ decimalized national cataloging codes.

The new library systems had important social implications. The vast reading rooms of the large public book palaces emphatically housed people rather than books. The promise of access to material resources combined with the cultural power of libraries—reinforced through imposing, monumental architecture—made libraries important sites of social interaction that acted as a magnet for researchers across the social science and *humanities disciplines.

Invisible to readers, the work of cataloging proved relentless. Cataloging departments routinely complained of underfunding and staff shortages. Backlogs were endemic. Though card catalogs, numerus currens shelving, and relative classification solved the problem of incorporating new items, the pace of accessions increased relentlessly. The number of physical cards required was enormous. Some manuals recommend at least four catalogs, all kept on cards. An accessions catalog and alphabetical author catalog were kept behind the scenes for library administration. A second author catalog and a subject catalog were normally made available to the public. To boot, author catalogs were sometimes interfiled with alphabetized title cards. Library of Congress and Universal Decimal Classification systems employed faceted subject indexing—the use of multiple entry terms for a single item. In that case two or three cards were needed for the subject catalog alone. Ambitious libraries required as many as eight or nine cards for each volume that entered the library.

There were efforts to reduce the burden of cataloging. The Royal Library in Berlin and the Library of Congress, for example, developed standardized “codes” for generating catalog entries that were adapted by other libraries. Other strategies included collective cataloging and the distribution of printed cards. When the new Library of Congress building opened in 1897, it had eight hundred thousand books and no catalog. In 1901, under pressure from other American libraries, it was decided to make the library a center for the distribution of bibliographic information. Printed catalog cards—nearly seventy thousand a year—were made available for purchase. The Library of Congress coordinated with other card-printing institutions such as Harvard, the John Crerar Library in Chicago, and the information service of the US Department of Agriculture. The circulation of stable forms of bibliographic data laid the groundwork for the development of union catalogs and interlibrary loan networks. In 1913 the general catalog of the books housed at Columbia University libraries contained over two million cards. A companion union catalog included cards from the Library of Congress, numerous federal bureaus, other major American libraries, and of New York City libraries not at Columbia. The Columbia example is by no means exceptional. It was reported that the University of Chicago Library, which employed a complex system of cross-indexing with departmental libraries, needed as many as twenty-eight individual cards for each item that entered the collection. By 1920 the Library of Congress had nearly three thousand subscribers and distributed over six million cards annually. Many European countries had similar schemes for collective cataloging, card distribution, union catalogs, and interlibrary loans.

The system worked in conjunction with new schools of library science, classification schedules, cataloging codes, indexing services, standardized forms of academic citation, and other material and social mechanisms of knowledge production. The goal was to produce a stable container for the dynamic transmission, storage, and retrieval of bibliographic data that could be generated and accessed by multiple users. The midcentury library system was a decentered and open-ended network. In principle it was capable of reproducing itself at any point. The only requirements were cash and labor, both in chronic deficit.

Familiarity with the management of large volumes of data and experience with alphanumeric notation rendered libraries well placed to experiment with early computational technologies. By some estimations catalog systems were on the verge of collapse in many places by the mid-twentieth century. Largely in response to a mounting cataloging crisis, the Library of Congress developed digital records in the 1960s—MARC, or Machine Readable Cataloging standards—that could be read by computers and shared with other libraries. Though computerization was initially adapted for internal systems management and in order to accelerate the production of physical data records, the digital interface was soon made public. By the 1980s the first hands-on experience with computers and digitized information for many people, the present author included, occurred at a public library catalog terminal.

The arrival of the *internet and the digitization of content has had an even more profound impact. Catalogs, books, and articles are available remotely, and libraries now manage vast virtual collections. Digitization has triggered extensive and controversial deaccessioning of physical stock. Meanwhile systems maintenance and digital subscription services loom large in cataloging and acquisitions budgets. Though libraries, books, and librarians have become all but invisible, the “digital revolution” has proven expensive. The need for funding and expertise will not diminish. In 2012 nearly half ($3.4 billion) of the total expenditures of academic libraries in the United States went to salaries and wages. Another $2.8 billion (40 percent) was spent on “information resources”—a combination of print and digital purchases and subscriptions. The number of digital publications acquired was nearly double the number of print publications. Nonetheless the cultural power of libraries remains remarkably strong in the public imagination. Facing the combined pressures of deaccession and depopulation, library spaces have been reimagined to perform new and sometimes unexpected social and cultural functions.

Paul Nelles

`FURTHER READING`

Andrew Abbot, “Library Research Infrastructure for Humanistic and Social Scientific Scholarship in the Twentieth Century,” in Social Knowledge in the Making, edited by Charles Camic, Neil Gross, and Michèle Lamont, 2011; Alistair Black, Simon Pepper, and Kaye Bagshaw, Books, Buildings and Social Engineering: Early Public Libraries in Britain from Past to Present, 2009; John Willis Clark, The Care of Books: An Essay on the Development of Libraries and their Fittings, from the Earliest Times to the End of the Eighteenth Century, 1901; Eric Garberson, “Libraries, Memory and the Space of Knowledge,” Journal of the History of Collections 18, no. 2 (2006): 105–36; David Kaser, The Evolution of the American Academic Library Building, 1997; Catherine Minter, “Systematic or Mechanical Arrangement? Revisiting a Debate in German Library Science, 1790–1914,” Libri 67, no. 3 (2017): 193–203; Felix Reichmann, “The Catalog in European Libraries,” Library Quarterly 34, no. 1 (1964): 34–56; Wayne A. Wiegand, “The ‘Amherst Method’: The Origins of the Dewey Decimal Classification Scheme,” Libraries and Culture 33, no. 2 (1998): 175–94.