Painter, Ann F., Ed.

Classification: A Definition5forces of stratification, languages of kinship became formalized, precisely indicating the levels of relationship betweenbranches of a family and its individual members. If an analogycan be made, it would be this: from a perceived difference(based on artificial criteria) between societal members, terms ofaddress in kinship languages formalized these distinctions, giving notational relationship between these individuals. If sometribes locked these patterns in too rigidly, then at what costwould personal initiative have to suffer in order to break thedeadlock of these expectations? Then again, with forces of diffusion and dispersion so widespread through this aged world.what changes were made to the set of these kinship orders?Could societies change the basis of kinship expectations without changing their stratified orders? This point, perhaps imaginary, is made to demonstrate the complexity of ordering andchanging the universe of man's perceptions. Our classification,as a philosophy and a practice. stands in the same proportionof difficulty as these anthropological phenomena.In this paper. I wish to illustrate classification as a process ofnaming and ordering this universe, but not merely an activitysolely directed to some objective world of knowledge. The historical process of stratification-classification has been one-ofadvancing knowledge as our understanding of natural and -Mimi& orders has increasedto give new relationships a rightfuland accurate place in the scheme of things. We have had tocompare these changes to hierarchical orders previously constructed. In this way. changes in our classified orders came notin scattered bits, or bytes. as it were. but as alternatives to thehierarchy of established factsand hence to knowledge itselfA carefully stratified order moves continuous time into separateepochs. thus, the extrapolation of time and circumstance wasgiven definite boundaries Within each epoch, alterations couldbe observed in terms of that specific time period, with each onehaving its own level of development. contrasted with otherages. ordered and classified according to these distinctionsThis was our heritage of intellectual classification, changing inits sophistication as our accumulation of facts increased Theorder it created became the foundation for comprehending theuniverse of knowledgeWhy do men have difficulty introducing radical change intotheir classified orders? Why has the existence of intuitive leaps10

Classification: A Definition2been such an important device for accomplishing thesechanges? After hypothesizing and experimenting. sifting thestudied relationships carefully, imagining through these concepts the new possible orders for them. after all this. how oftendo the gaps still exist? Man cannot consciously finish the job.Tile intuitive leapthat process of comprehension so slightlybeyond the conscious worldaccomplishes the extraordinaryfeat of interpolating these facts into the order that was notquite within reach of the thinker. With intuition. old classifications are destroyed and new ones created. These new classifications of phenomena. discovered in hard thinking and timelyserendipity. have their own language, classes. their distinguishing characteristics from former classifications and their pointsof duplication. In some cases.the terms gaining access tothese classifications aro absoluteaccurately part of the neworder itself, unrelated, in its essentials, to previous classes andterms. In other cases, the terms of reference will be fuzzy. questionable, almost belonging to one or another class and to several different classifications. In gaining the best access to thehierarchy, how can one be sure of the accuracy of terms? In anabsolute order there is no confusion. In separating a homogeneous world, however, how can man's order duplicatea natural order with the t ame degree of perfection? Therefore.no order can be absolute. it is only a temporarily derived stagefor viewing the accumulation of facts to date. Yet. as a processspecial with man. is it not fascinating to recollect that our progress as a species was so fast. accurate, and unstoppable because we had gamed control of such a power as classification?If I have not explicitly defined anything yet. you may see thatthe difficulty rests with trying fo pinpoint an activity so pervasive in man's growth and history. One of the large conflicts inthinking of our library classification. in contradiitinction to theintellectual process I have described. is in deciding what classification really represents. When we invoke a Dewey Decimalnumber. are we seeing a pattern. a piece of the classified order.falling Into our comprehension? Are we handling a representative from that order in the form of a document in which part ofthat classified order shall be revealed? Or. is it merely a placereserved for such a representative? The differences betweenthese ideas can yield three definitions.Historically. it may be said that classification is a process andan act of ordering and differentiating the universe, yet classifi-1.1

Classification: A Definition1cation also mirrors philosophy. has its own terminology and isan organization of places to store things, be they ideas ordocuments. The mixture of these elements has caused a reignof confusion as to what purpose classification should serve.has been wellThe Baconian influence on library ciassiticailis ification fromdocumented. Bacon's era extended libraryreverse (thenthe "art- of making philosophical charts of t, of collectedat its apogee) to a conjtinction, on a cosmic Ibits of a developed process of mental classificati6n. One respects Bacon for his ability to do this so beautifully, comprehensively and lastingly. Several hundred years later, throughinterminable changes. influences, and practices, the Baconianuniverse met Melvil Dewey. and therein one of the more important historical events of man's history of classificationwefind the first significant and lasting admixture of philosophicalperception and the rather mundane practice of storing documents in libraries. In this encounter, a question was createdthat has not been solved. What is classification? Is it the practice of philosophical differentiat:on I have been describing orthe art of accurately storing documents with a mind toward effective retrieval of related pieces of information?1iDewey lifted a sagging world of shelvers and card catalog makers, and gave librarians a chance to participate in the comprehension of rarity and beautythe worlds of philosophy, themirrors of man's universe. determined through these specialized perceptive and cognitive abilities Henceforth. whenwe hear about the classical debates of where to place certaindocuments in the Decimal Classification. it is not merely that aquestion of location is being argued. Indeed, one has the feeling that the interlocutors were questioning the inherent orderof the universe itself. Why else would these practical storers ofinformation give so much heat to the argument? Even though itwas dealing with the documents of knowledge. the welding ofphilosophy into the Decimal Classification made it a processwhereby the universe was divided into identifiable classes.further subdivided by these perceived differences, and given anappropriate notation for retrieving the documents of this orderThis may be a simplification, but I am illustrating classification.not giving a manual for the Dewey Decimal Classification(DOC). Later elaborations of this process found more enumeralive schemes, in which this universe was explicitly developedfor document retrieval and the collation of related materialsThe intent was to integrate materials as they were being stored12

LClassification: A Definition8An ironic fact about the DDC is that it has managed to persistso long, undergoing numerous reordering of its scheme and yetremaining representative of our changing world, cohesive andcontemporary. It is this duality of purpose that has given it somany problems. Perhaps it is being made to do too much?Perhaps it is only feasible for a limited point in time?,In beautiful opposition to the philosophy and practice of DDC,we have a modern science of classification which exploresanother extreme. It is called subject analysis, faceted classification, developing in its glory as a computerized operation. 1E31ginning with a universe of discrete facts, ideas or subjects, itseeks initially to abolish formerly perceived classes, and substitute for the old method of finitely breaking down the universe,one which minutely orders these facts, bringing related aspectsof documents together into classes which represent the cooccurrence of,terms as analyzed in the documents themselvesIt is a process of building up the classification from these facts,or facets, without seeking to create a complete universe. Withthis method of classification, philosophy has been returned tophilosophers. Computerized classification offers an opportunityto relate conceptually the documents of knowledge much moreprecisely. The classifications are not stable, but change frequently with the reordering of subjects in these documents. Indeed, one Would wonder whether this is really classification, asit seems so antithetical to the progress of mental classificationwith which I began this paper.These classifications are interesting in principle, but significantcosts well have to be assumed to perfect their development astools of classification. It is generally assumed the faceting canwork for certain small classes of documents, but with generallibrary collections they would be useless. In some experiments,classes created by computer algorithms had to be combinedwith classes from a traditional classification in order to reducethe amount of irrelevant materials retrieved. Even where computers can create such classificaticns, other practices must beappendedlike in-depth indexing systems, independent notational systems for storing documents and the great interpretiveinvolvement of librarians.Are there any commonalities between these contrasted classifications that would allow us to define classification? Yes, one!In both cases classification must become an ordering scheme1

.Classification: A Definitionfor locating these documents. At all costs, whatever scheme weinvent or use, the classification must be able to perform effi-ciently at this central tasklocating the document for the userModern classificationists are attempting to take this qualification and to say. "If machines can do it faster, though not morecomprehensively, and on an average perform as well, then ourmethods must be equally as valid. Not better, but at least asvalid." For whatever slight cost advantage is given by this perceived equality, these classification systems will be developedfor special collections, perhaps one day for general library useIt must be said again, therefore, that whatever the method, thefinite dispersion of the universe or the monolithic creation of aworld of discrete subjects, we still have to store and retrievethese documents. I think, now, that my definition of classification is apparent. it is the activity of storing documents for retrieval. No order is complete without a basis for distinguishingand differentiating the documents, but the locating and storingfunctionthe notational devicemust be indepehdent of thatscheme. I could conceive that, given the ability of successfullystoring and locating documents, any scheme in the futuremight be adaptable to that purpose. So, in the end, I have offered a rather unstartling definition of classification. Is it reflective and worthy of my argument?tThe problem is that I am forced to uphold it, but I do not fullybelieve it. Classification was initially described as a mental process of ordering the universe, and we have taken our librarypractice and reduced it to a mere act of storing and locatingdocuments in a collection of materials. On one hand we canspeak of classification in the highest sensethat which hierarchically orders the universeallowing us to proceed fromconcept experience to concept experience, revising ourcategories, but building up our knowledge as a consistent attempted representation of the universal order. On the other, wemerely speak of our library classification as a shuffling device,devoid of that presence which exists when the two are combined, as in the original encounter of the-Dewey Decimal Classification and Baconian philosophy. Yet, our age has a newphilosophy, and its herald, the computer, allows for quick, subtle and efficient manipulations of ideas, facts and subjects. If,therefore, my definition of classification is unsubstantive, I feelit must remain so. We have a different age upon us, and whathave existed as inherent mental processes are changing, offer-14

Classification: A Definition10ing us new and different experiences and practices. The humanprocess of classification and its substantiation in the library willmirror these influences. It is only a question of whether we willfind the same kind of welding in library classification that willgive us a new and unique opportunity to classify our documents and store them in the same mode, or some kind of dispirited shuffling system that is efficient, but lacking in humandynamism.15.

The Historical Context:Traditional Classification Since 1950Gordon StevensonIntroductionTwenty-five years ago, when librarians in the United Statesspoke of classification, they were usually referring to twospecific library classifications: the Dewey Decimal Classification(DDC) and the Library of Congress Classification (LC). Thehabit of confusing the general idea of library classification withthe possibilities and limitations of DDC and LC had beencharacteristic of United States librarians for generations. Aclear distinction was not made between general principles ofthe nature, structure and uses of library classification and theapplication of these principles in specific systems. This approach to classification enshrined DDC and LC somewhere,near the center of librarianship. Even today, our problem is, notso much classification as what we think classification is andhow we think about it. The way we thought about classificationaround 1950 was such as to give DDC and LC a legitimacy andpermanency of the sort usually reserved for religious texts andsacred rituals. Unfortunately, this approach is still found to agreat extent today; and though DDC and LC seem to be ovenmore inextricably embedded in United States librarianship thanever, it is now necessary to identify these two systems'as "traditional \library classifications." They must also be identified asgeneral classifications," because they know no subject limitations.DDC and LC are traditional in an historical sense bebause theirroots are deep in the past, and in a practical sensbecausethey are used by librarians today in essentially the game waythey Were used when they were introduced before 'aild shortlyafter 1900. They a e also traditional because of their structures.Gordon Stevenson Is Issociate Professor, School of Library and Information Science, State University of New York, Albany.1611

The Historical Coritex14'12They are internally structured with mutually exclusive, enumerated classes that are arrived at by a logical process of divisionthat proceeds from broad concepts and. disciplines to ever narrower and more specific subclasses. Since 1900, these systemshave changed; but most changes have been quite superficial interms of classificatory techniques. New classes have been introduced, finer subdivisions have been made, and old classeshave been rearranged. But the traditional systems employ nobasic structural or classificatory device that was not known before 1900.In 1950, Jesse Shera critically evaluated the traditional classification schemes and the principles on which they are based.' Indoing this, he succinctly defined their parameters and clarifiedthe difference between traditional and nontraditional systems.In the meantime, we have learned a lot about classification andits theoretical and practical foundations. The past several decades have seen a more intense examination of the foundationsof classification than any other period since the last quarter ofthe nineteenth century. The results of these investigations, experiments, philosophical speculations and theories havecreated an ever-widening gap between the traditional systemsand the newer, modern systems. It is the purpose of the presentreview to consider the two traditional systems in their historicalcontext and to comment on the idea of general, as opposed tospecial, library classification. With DDC and LC, we are dealingwith two dinosaurs that one would have thought could not survive into the second half of the twentieth century. They wouldappear to be relics of the past, and their survivalindeed, theircontinuing vitalityraises important questions about the natureand uses of classification by librarians in the United States. It isthe contention of the author that it is impossible to understandthe condition of classification in the United States today or tospeculate intelligently about its future without an historicalperspective.Classification Around 1950.General library classification as we knew It around 1950 was aproduct of decisions made around 1900. Expectations aboutthe contributions of classification to subject control and access, ideas about the structure of classification systems, andgeneral agreement about whEat constituted a proper subject17

The Historical Context13catalog had long since been formalized and incorporated intothe conventional wisdom of librarianship. A key historical eventof almost unprecedented importance in the history of subjectaccess was the rise of the dictionary catalog and the subsequent disappearance of the classified catalog from UnitedStates libraries. After that happened, the way we thought aboutclassification and its uses changed fundamentally. By 1950,most librarians in the United States were not quite sure what aclassified catalog was for or how it was different from an alphabetical subject heading catalog. Why this happened and itslong-range impact on both classification and subject access arehistorical questions which have never been answered Added tothis fundamental change in the use of classification was thephenomenal dispersal of DDC and later LC. All competing systems were swept aside and these two became such monumental edifices that they have never been seriously challenged inthe United States. By the time Bliss published the final volumeof his Bibliographic Classification in 1953,2 hardly anyone in theUnited States took his work seriously. It is very possible that theBibliographic Classification was a better classification thanboth DDC and LC, but it was published too late to have anypractical impact in the United States.As late as 1950, many, if not most, library schools in the UnitedStates taught all students the DDC system and saved LC forthose hardy-students who went on to take ''advanced cataloging." It did not occur to anyone that there might be an alternateto DDC and LC. Most of what we knew

marized. With the theoretical framework provided, the Dewey Decimal Classification, Library of Congress Classification and the Universal Decimal Classification are examined. And finally, there are two papers on the futureautomatic classification. and res