College and Research Libraries The Language of the Levels: Reflections on the Communication of Collection Development Policy Ross Atkinson The collection development policy, as a means to express and systematize guidelines for collec- tion building, fulfills three basic functions: the referential, the generative, and the rhetorical. The division of the policy into subject categories, and the use of "collection levels" (such as those defined in the ALA Guidelines for Collection Development) to rank the collection and the collecting effort for each subject, serves these three functions well. More work needs to be done, however, on defining collection levels and the collecting effort to which they refer. These definitions may be more easily achieved if we begin to view the collection levels as desig- nating varying degrees of two opposing collection strategies: inclusion and exclusion. f we understand policies simply as "guides to carrying out an action,' ' 1 then there are nor- mally as many selection policies in a library as there are selectors, for each selector necessarily develops, over time, a set of personalized guidelines upon which to base selection decisions . Such policies are usually vague and unarticulated. On the rare occasions when they are written, they are of necessity expressed from the specialized viewpoint of the individual se- lector. The values upon which such poli- cies are based and the goals toward which they are directed vary, therefore, from one selector to another. It is the task of the collection develop- ment policy to specify, consolidate, coor- dinate, and adjust such separate selection policies in order to promote the develop- ment of a collection that will, as a whole, best respond to the needs of current and future clientele. This task is accomplished by bringing about the translation of the various selection policies into a single lan- guage, making adjustments in the indi- vidual policies to fit the general collection plan, and then stitching these adjusted policies together into a unified document. While the translation and consolidation reduce the disparity among individual se- lection policies, it should be noted that such disparity can never-and probably should never-be eliminated entirely. The separate segments of the collection devel- opment policy must remain the personal responsibilities and products of individual selectors. The primary objective of the collection development policy, therefore, is to unify or focus expression concerning the cur- rent state and future direction of the col- lection. If we are to determine how policy works and how to use policy for purposes of collection planning, we need first to un- derstand its operation as a system of com- munication. This paper is an attempt to move us a step closer to such an under- Ross Atkinson is assistant university librarian for collection development at the University of Iowa, Iowa City, Iowa 52242. An earlier version of this paper was read at the RTSD Collection Management and Development Institute, Trinity University, on May 16, 1985. The author wishes to thank Professor Terrence Brooks, School of Library and Information Science, University of Iowa, for helpful criticism of the paper's initial draft. 140 standing. It begins with an examination of the general functions of collection devel- opment policy, and then turns to a more detailed dissection of the policy's stan- dard structure. POLICY FUNCTIONS Although the functions of collection de- velopment policy can be described or de- fined from a variety of administrative, bibliographical, or epistemological per- spectives, the collection policy as a com- munications device intended to transfer information about the development of a collection fulfills at least three fundamen- tal functions. First, it provides a descrip- tion of the collection's current state, devel- opment, and desired direction. This is the policy's referential function. Second, the policy serves the selector, if only inferen- tially, as a method or instrument to trans- form the collection from its current to its desired condition. This is the policy's gen- erative function. Finally, the policy also acts as an argument that there is a system- atic collection plan in effect, and that such a plan is worth pursuing. This is the pol- icy's rhetorical function. The referential function is primary; the generative and rhetorical functions derive from the refer- ential function. Let us take a closer look at these three functions. Like any document, the collection policy fulfills its referential function through the application of a conventionalized system of signs. A sign is a "cultural unit" that ''is defined inasmuch as it is placed in a sys- tem of other cultural units which are op- posed to it and circumscribe it. " 2 In other words, meaning derives from the relation- ships among signs. Because such relation- ships take place entirely within a system, moreover, that system is, to use Umberto Eco' s cumbersome term, auto-clarificatory, i.e., "capable of checking itself entirely by its own means." 3 The only way to learn the reference or meanings of the signs of which a sign system (such as a language) is composed, therefore, is to make use of that system, contrast its constituent signs with each other, and arrive at an under- standing of how those signs relate among themselves. The core of the standard collection de- velopment policy for larger libraries, and The Language of the Levels 141 increasingly for smaller libraries, 4 consists of a series of subject categories. The cur- rent collection strength, current collecting intensity, and the desired collecting inten- sity for each subject category are then ranked according to a scale of ''collection levels. " 5 These components will be exam- ined in detail when the problem of struc- ture is discussed. The subject categories and collection levels serve as specialized sign systems, the constituent signs of which can only be understood by using those systems and, through such use, contrasting the signs within each system to each other. The referentiality and the ef- fect of the collection policy derive from the use of these systems in conjunction with each other. The collection policy is, there- fore, ultimately a self-validating network of relationships; the key to making, writ- ing, and using collection policy is to un- derstand how its constituent elements in- terrelate. Turning to the generative function, we should recognize that a successful policy is one that supplies the means to generate, over time, a collection with certain desired properties. The policy must also provide the selector with some insight into the method of achieving such a desired collec- tion state. As a consequence, the collec- tion policy must not only refer to the cur- rent and the desired states of the collection but should permit the selector to infer how to transform the collection from the cur- rent to the desired state. This complex and problematic area of collection policy has received little attention. Nevertheless, it is clear that a policy that does not fulfill a generative function will have little effect. Closely linked to the referential and generative functions is the rhetorical func- tion. The purpose of rhetoric is persua- sion. The targets are the three audiences of collection policy identified by Eric Car- penter.6 First, the policy should show fac- ulty and students that the reasons the li- brary contains certain materials and not others are part of a rational, consistent, publicly announced plan. Second, the in- stitutional administration should be led by the policy to recognize that optimum use is being made of materials funding, and that requests for increased funding derive from a process of sustained and system- 142 College & Research Libraries March 1986 atic planning. Third, the library's consor- tia partners should also be moved by the policy to view the collection development operation as stable and reliable, and to ac- cept the possibility of entering into mutu- ally advantageous agreements with clearly defined goals. Within the library the policy fulfills its rhetorical function by demonstrating to selectors that there is in- deed a consciously controlled, librarywide collection development system in effect that defines the parameters of their re- sponsibilities. These three fundamental functions of collection policy are closely related and in- terdependent. The generative function clearly relies upon the rhetorical function, for merely to provide a method is no guar- antee that it will be used. The selector must receive from the policy not simply direction but also the impetus to take that direction into account. Regardless of how thoroughly the selector is convinced of its merit or utility, this direction cannot be followed unless it is intelligible. This is achieved through the network of relation- ships established by the policy's referen- tial function. POLICY STRUCTURE Over the past two decades, as collection development has become a recognized and distinct library operation, an increas- ingly standardized structure for collection policy has evolved. 7 The components of Collection Aspects CURRENT CURRENT DESIRED COLLECTION COLLECTING COLLECTING this structure have been summarized and canonized in the ALA Guidelines, 8 and have been most bly amplified by Charles Osburn. 9 In its standard form the policy includes introductory material on general objectives, divisions of subject responsi- bility, and duplication. A glance at the an- thology of policies compiled by Elizabeth Futas will confirm the varietcr and signifi- cance of such information. 1 Indeed, the ''analysis of mission, clientele and pro- gramatic objectives is a vital prelude to the detailed subject analysis of collection pol- icy.' ' 11 The heart of the policy remains, however, the segmentation of the collec- tion into subjects, and the rating of the quality of each subject segment according to the system of collection levels. 12 It is to this central component of collection policy that attention needs to be directed. The core structure consists of two parts: (a) the matrix that is formed by the inter- section of the subject classes and three col- lection aspects, i.e., current collection strength, current collecting intensity, and desired collecting intensity (see figure 1) and (b) the collection levels that serve as a scale for rating the collection aspects for each subject. Let us begin with an examination of the matrix. Although the collection aspects usually form the vertical columns, and the subject classes the horizontal rows, I have tilted this formation on its side in figure 1 because I feel this enhances our under- ~ N (") ...::t -1-1 -1-1 -1-1 -1-1 (.) (.) (.) (.)