William Stetson Merrill and bricolage for information studies

Introduction

In 1914 the Committee on Code for Classifiers of the American Library Association (ALA) issued 200 copies, in mimeograph form, of A Code for Classifiers: A Collection of Data Compiled for the Use of the Committee By William Stetson Merrill, Chairman (Merrill, 1914). The 1914 mimeographed document was the precursor to two editions of the code that ALA published (Merrill, 1928, 1939). Merrill discussed the idea for compiling a code for classifiers in 1911 at the ALA Conference in Pasadena (Merrill, 1911) and in 1912 when he presented two lectures at the University of Illinois. The lectures were also published in the Library Journal (Merrill, 1912a, b). Subsequently, Merrill was appointed chair of the special committee of the ALA that included now famed classificationists J.C.M. Hanson, University of Chicago Library; Charles Martel, Library of Congress, Washington; and other prominent librarians of the time such as Phineas L. Windsor, University of Illinois Library; Urbana, J.C. Bay; John Crerar Library (now the University of Chicago Library); Walter C. Biscoe, New York State Library, Albany; and W.P. Cutter, Library of the Engineering Societies, New York (Merrill, 1914, p. 2). The committee was charged to collect data “relating to a proposed code for classifiers” (Merrill, 1912a, b, p. 3). Copies of the mimeograph, the product of the committee's deliberations, which was essentially a list of the principles that Merrill was using at the Newberry Library, were circulated and “detailed criticism” solicited (Merrill, 1914, p. 3). The feedback from various libraries and librarians around the country, and from an ALA survey (Bostwick and Seymour, 1926) was incorporated into a revised version of the mimeographed code that ALA published, 14 years later, and with a different sub‐title, Code for Classifiers: Principles Governing the Consistent Placing of Books in a System of Classification (Merrill, 1928). A second edition of the code was also published 11 years later in 1939 and the most recent re‐printing of the code appears to have been 1969 (Merrill, 1969). Figures 1‐4 are excerpts from the various editions of the code giving guidance on classifying a book/information resource about historic houses. They are explained in detail later.

Many general principles for library classifiers that are the basis of the code and still taught today – aboutness, intent of the author, class of reader for whom the book is intended, and subject vs topic distinctions – are explained in early published lectures (Merrill, 1912a, b). Four schemes of classification are used in American libraries of the time (Dewey's decimal classification (DDC), Library of Congress classification (LCC), Cutter's expansive classification (EC), and Brown's subject classification (BCS). Merrill provides examples of specific titles when discussing classifying problems such as complex topics, coordinate topics, unrelated topics, bias and influence relations among topics. Ernest Richardson and James Brown are summarized to provide a list of the general characteristics of books and the subject characteristics that may be used for classifying. Most important however, is Merrill's insistence on the difference between the art of classifying and the science of classification. He categorizes himself as a practical classifier, and paves the way for theoretical classificationists like Bliss, Mills, and Ranganathan. Thus while classification theorists Bliss (1933, pp. 134‐144), Ranganathan (1998, p. 81), Mills (1968, pp. 158‐159), and Sayers (1926) were all aware of it, the extent of the code's influence on their own thoughts and works needs to be explored further. In brief, Bliss makes the same distinction between classifying and classification that Merrill highlights in his code but cites Merrill's code in the context of the need for an index to a classification scheme (p. 134) as well as in a subsequent discussion on codes for classifying (p. 144 ff). Mills opens his chapter on practical classification and indexing by using Merrill's definition:

The art of assigning books to their proper places in a system of classification (p. 158).

Ranganathan mentions the code as proof of his argument that classification consistency cannot be achieved between libraries (p. 81).

In more recent times, Merrill's code has been thought to represent US “national consensus” about classifying (Olson and Boll, 2001, p. 62). The code was also translated into Japanese (Merrill, 1928) and Spanish (Merrill, 1958) and used in libraries and library schools inside and outside the US (Fellows, 1914; Mann, 1928, 1929; Maltby, 1972, p. 136; Dutta, 1962, pp. 241‐252). However, today, in the US, both Merrill and his code are not really well known. Few recent writers researching in this area or foundational texts even mention the code or its creator. Paul Dunkin, who published Cataloging USA in 1969, the same year the code was last printed, described it as “a collection of rules for traditional American library classification with some discussion of each” and dismisses it as “out of date” (Dunkin, 1969, p. xx). Yet its creator focused on issues involved in the application of theoretical classification systems to actual library materials, issues that to this day continue to challenge original catalogers, metadata librarians, and information systems designers. Like the scholar‐librarians he knew and worked with – Charles Ammi Cutter and William Frederick Poole – Merrill epitomized the professional who understood the art and the science of not only library administration and service but also information organization, its tools such as bibliography, cataloging, classification, and indexing, and was not unaware of the bibliographical control problems presented by new materials and the growth of knowledge. Consequently, Merrill's career offers the best potential for current practitioners of library and information science if he is understood as an early bridge figure in what would come to be called the documentation and later the information science movement; a bricoleur who was able to use the tools at hand and draw on humanistic and scientific approaches to solving problems of knowledge organization. The rest of the paper highlights the key aspects of Merrill's career and life together with excerpts from his writings, correspondence, and reminisces, followed by a discussion of bricolage and it's relevance today.

Merrill, Poole and Cutter

William Stetson Merrill was born on the 16 January in 1866 in Newton, Massachusetts. He “claimed descent from the Mayflower Pilgrims” (Merrill, 1954, p. 28). The young Merrill entered Harvard in 1884. While at Harvard, he worked as a student assistant librarian and William Lane, Assistant Librarian at Harvard College, was a family friend. Lane recommended Merrill for a position at the Newberry Library and accordingly, Merrill met with Poole, Librarian, in December 1888. He took up his new position at the Newberry in June 1889 and remained there for almost 41 years, until his retirement in 1930 (Merrill, 1954, 1955; Krummel, 1966; Krummel and Williamson, 1978). Thus, Merrill did not belong to the emerging cadre of those professionally trained in the new library school that Dewey had recently set up. Rather, he learned library management as an apprentice, interested in bibliography, indexing, and classifying. That is, he learned library work by doing it.

In the early days of the Newberry Library (1889‐1930), Merrill worked alongside the elder statesman of American librarianship, indexing innovator and pioneer of the modern public library movement William Frederick Poole (Williamson, 1963, p. 6), bibliographers Charles Evans and George Watson Cole (who became a life‐long friend), cataloger Charles Nelson, and classificationists Charles Martel and J.C.M. Hanson (Merrill, 1954, 1955). Merrill was deeply influenced by Poole and Cutter (Librarian, Forbes Library, Northampton, Massachusetts). Until 1894, when Poole died, Merrill was in close contact with Poole; his desk was right outside Poole's office. Poole's Index, the precursor to Reader's Guide and Poole's ideas of classification (as theory) vs classifying (as application) informed his early training at Newberry (Merrill, 1954, p. 15). While he does not seem to have subscribed to Poole's criticism of the Dewey system of subject subordination as “procrustean” (p. 16) he did follow in other matters. He compiled the Archeological Institute of America's Index to Publications, 1879‐1889 (Merrill, 1891), and served as the ALA editor of the Index to Periodicals (issued on cards) from 1913‐1931 (Merrill, 1916a, b). He was involved in the founding of the Bibliographical Society of Chicago, the precursor to the Bibliographical Society of America, and prepared and read a paper on “General and national bibliographies” at one of it's early meetings, in April 1900 (Merrill, 1900).

Merrill's correspondence with Cutter lasted from 1895 until Cutter's death in 1903 (Cutter, 1837). He was active in proposing modifications to the expansive classification scheme; for example, in 1895 Cutter wrote to him:

You want to have a mark for the classes now marked FA‐FZ, called “allied studies”, which shall have only one letter, F and distinguish the studies by figures. If this had been thought of in the beginning it would have been best to use FO for Allied Studies; but now FO is already taken for periods.

When John Vance Cheney became Newberry Librarian in 1894 and brought along his assistant at San Francisco Public Library, Alexander J. Rudolph, who had invented the Rudolph indexer (an alternate to the card catalog), Merrill was instrumental in getting the Rudolph and thereby, the Newberry Library and the indexer, to adopt Cutter's expansive classification scheme rather than the Dewey scheme (Merrill, 1955, p. 11). Rudolph did not like the Cutter notation of letters and wanted decimals and this led to the use of Cutter's classification with decimals at the Newberry. Merrill describes how this came about:

The first letters of the Cutter notation designated main divisions of human knowledge: A – Generalities, B – Philosophy, C – Christianity, D – Church History; and so on, to Z – Bibliography. Subdivisions and topics were designated by further letters. Instead of doing so for subdivisions, why might not the Newberry use numbers: A1, A2, B1, B2 which would be treated as decimals, admitting indefinite enlargement: A11 to A19, A111 to A199. Initial 0 would be used for forms like periodicals, B 07 – philosophical periodicals. A decimal point would separate the class number from the author number, also a decimal (Merrill, 1955, pp. 11‐12).

This beginning is somewhat similar to how the Merrill Book Numbers came bout too – as a result of a practical need for alphabeting books (Merrill, 1912a, b). Library schools such as the Syracuse University Library School (Sibley, 1909) and Illinois Library School (Sharp, 1896) show both requests and acknowledgement of receipt of these numbers. In his memoirs, Merrill describes how these numbers were developed:

Cutter‐Sanborn numbers would be used in classes, like biography or works of individual authors in literature, where elaboration of the author number was called for. For bibliography I devised the employment of lower case letters, corresponding with the capitals. Thus, if F 8962 is for history of Chicago, f 8962 is for a list of books on it. The decimal authors numbers were to be taken from a table that I drew up: AA to ZZ – 01 to 99. These so‐called Merrill numbers have since been used for alphabeting by decimal numbers in other libraries (Merrill, 1955, p. 12).

Merrill's milieu

Bricoleurs and bricolage

Apprenticeship, adaptability, and collaboration are characteristics of the bricoleur that Merrill's career thus far has demonstrated. However, it is also in the approach that he took to solve the knowledge organization problems of the day that we can best discuss Merrill as a bricoleur who has much to offer information studies today. As a bricoleur, Merrill was a thinker tinkerer; he focused on the immediate objects and materials at hand to fashion solutions for problems faced. Levi‐Strauss explains that when he presented a bricoleur in a dichotomous category and compared it with the engineer he was trying to explain that “there is no gap between the way so‐called primitive peoples think and the way we do” (Eribon, 1991, p. 110). Very simply, in the binary view primitive minds are made up of bricoleurs and domesticated minds are made up of engineers. The differences are often explained in the types of data and materials each group uses; in bricolage the engagement is with the materials at hand (improvisational assemblage to solve an immediate problem) and reflects humanistic values while the scientific or engineering mind engages with the abstract, the representations of the material and proceeds in a more formal, often deterministic way. In creating tools such as the decimal book numbers in an attempt to solve the shelf organization problems the Newberry collections faced, Merrill exhibited his tendency to bricolage; Cutter tables were available, but they did not solve all the problems and hence the creation of Merrill book numbers. Interestingly, Maltby and Sayers in discussing Merrill's book numbers, which were essentially author marks, refer to Merrill as the “great American classifier” but also correctly consign author marks to a comparatively unimportant part of classification theory (Maltby, 1975, p. 90). Yet, to both librarians and end‐users locating the book on the shelf is a very important and much‐needed service while the general purpose of classification is to place books where they will be permanently useful (Merrill, 1928, p. 2).

Similarly, Merrill used the intent of the author as the primary principle that should guide classifying; he was simply borrowing a key bibliographical method and principle to solve a classifying problem (how should the classifier determine what the matter in the book was about?). In describing the indexing problems in classifying, he was making connections between the work he did as ALA's editor for the Periodical cards program whereby a set of cooperating libraries agreed to index articles for a selected set of journals on cards, ALA/Wilson printed the cards and these were purchased and used in libraries in card catalogs to provide access to journal articles (Merrill, 1916a).

A bricoleur just does not borrow tools; he also moves between scientific and “other” modes of thinking. Besides a proclivity for statistical reports of work productivity, and the use of logic and collaboration to integrate critical other views into the code, Merrill's capacity for such shifts is illustrated by his role in the debunking of the librarian's hoax (Wiegand, 1979). He reasoned scientifically and proved to be one of the earliest librarians to penetrate the truth behind the librarian's hoax perpetuated by John Cotton Dana and Edmund Pearson (Merrill, 1910). Many of his administrative reports are filled with quantitative evidence in support of his arguments, productivity, or concern for efficiency and systematization (Merrill, 1895‐1918). As has been noted earlier, Merrill was also key in the adaptation of Rudolph indexer, a mechanical alternative to the card catalog at the Newberry. Here again, Merrill did not devise a new scheme, he skillfully modified Cutter's expansive classification for the indexer by integrating decimals with the expanded alphabetical classes. Levi‐Strauss, the originator of the bricoleur concept, acknowledges that the bricoleur/engineer division is false as the same person can exhibit both modalities of thinking (Eribon, 1991, p. 110). Thus, what is important for modern‐day information scientists and librarians is the integrative thinking and tinkering with materials that the bricoleur engages upon. Merrill's attention to detail, his skill in solving problems using the materials at hand, his desire to integrate the fast diverging arts of indexing, classifying, and bibliography, tolerance for diversity, and his ability to switch between systematic and holistic ways of thinking are the important elements that mark him as a bricoleur and that we need to investigate further, if we are to understand the exact potential of bricolage for knowledge organization and other problems in library and information science.

Williamson wrote in his biography of William Frederick Poole, founding librarian of the Newberry library who heralded the public library movement in America, and was the epitome of the scholar‐librarian:

In the long perspective of American library experience, it is to be questioned whether the turn to close and exact rules and procedures did not bring, along with the apparent benefits, a rigidity and an elaborateness that cost more than the advantages were worth (Williamson, 1963, p. 183).

Williamson was referring to Poole's “tolerance for diversity” in the technical procedures that individual libraries must follow in order to serve their readers. In this, Poole was quite unlike Dewey; he was not convinced of the need for detailed enumerative classification schemes like the decimal classification that Dewey was promoting. Nor was he convinced that all libraries must follow the same, standard practices. In his 1886 address as President of the ALA he said:

Methods which are adopted for one library are not necessarily adapted for another where conditions are different. In 1889, Dewey was defeated by the combined power of Cutter, Fletcher, Winsor, and Poole in a final showdown on the question of whether or not the ALA should officially endorse specific systems, procedures, and rules. Dewey was intent upon forcing recognition of “the one best way” (his way) of performing any library job (Garrison, 2003, p. 144).

To Dewey's argument that the ALA constitution charged members “to reach conclusions” Winsor countered that this did not mean the association could endorse “any principles of action or usage.” The use of the decimal classification and standardization in library work became inevitable with Dewey's election as president in 1890. But, the questions remain: was Poole right? Are our library cataloging, classification, and indexing systems rigid and elaborate? What have libraries lost in their adherence to rigid, elaborate systems of organization? If Merrill's code had been kept up, could it be used as the basis for a switching language today? Did it serve as the basis for the development of the Library of Congress subject cataloging manual? These and other questions such as the influence of the code on classification theorists remain to be investigated.

Conclusion

The focus on library standardization, the rush to automation, and the increasing tendency to centralization, appear to have inadvertently fostered a neglect of intellectual diversity. Unfortunately, the integrative and bricolage sorts of solutions that humanists who were documentalists like Otlet (Rayward, 1997) and librarian bibliographers and classifiers, like Merrill tried to follow, have also been ignored. In the process, we lost valuable opportunities to provide flexible, humanistic, customized and universal solutions to the problems of knowledge organization and retrieval. What Merrill's career suggests is that the current understanding of the multiple traditions and heritages arising from documentation, bibliography, indexing, cataloging, and classification as separate trajectories, is misguided; a view that has been supported by recent theorists and researchers in library and information science (Budd, 1995; Hjorland, 2004), as well as in information systems (Ciborra, 1999; Avgerou et al., 2004). Merrill's attempts to integrate these traditions should be studied and emulated as a model for information studies as he represents a time (which may not have passed) when an information scientist is a bricoleur who draws on librarian, documentalist, bibliographer, classifier, classificationist, indexer, technology, and systems traditions to solve information problems in ways that humanize technology and augment us.