Classification

Classification is the process of assigning elements or units to classes (or types or taxons) according to some criteria. The system of classes into which the units may be classified is called a classification system. Often it is said that the criteria of classification are likeliness: A classification unites like things and it scatters unlike things. Things may, however be like in many different ways, why a classification should unite things from a functional or a pragmatic point of view based on the purpose of the classification.

“The fundamental elements of any classification are its theoretical commitments, basic units and the criteria for ordering these basic units into a classification” (Hull, 1998).

The objects we classify have attributes with values. Attributes may, for example be color or weight. Values may be red or heavy. Classifications are made by considering different attributes and their values. It is important to realize, however, that any set of objects may be classified in an unlimited number of ways.

The units to be classified may be physical objects, persons, processes, ideas, concepts, words etc. We may, for example, classify the objects in a room in furniture, people, potty plants etc. However, in making such a classification we are using concepts such as “furniture” and “plants”. We are often doing so unconscious, without realizing that the way we classify things are reflecting conceptions built into our language and mental conceptual structures. Concepts are thus classifications that we use all the time. Concepts vary from culture to culture or from language to language. A tree is not exactly the same in English and Danish (cf., Hjelmslev, 1943; Buckley, 2001). Our own concepts (and conceptions) seem mostly natural to us, and we often find other cultures concepts unnatural, strange and unfruitful.

As shown below are concepts needed to classify objects. Any classification is “conceptual” in this sense. However we can have concepts for which there exist no objects (e.g. general classes in which all objects are members of more specific classes). In such cases we can have concept classifications and bibliographical classifications- documents can be written about general classes - but we cannot have object classifications: there are no objects that are general for their kind). The classification of words mean to classify them as words, e.g. their grammatical class, their number of letters etc, NOT to classify after their meaning because this is a concept classification.

We classify things, persons etc. in our everyday life, in our professional life, in the sciences. In teaching we classify the issues to be taught according to their difficulty and the range in which they are suitable for being taught. In libraries we classify documents in order to facilitate information retrieval, browsing and other ways of obtaining relevant documents. The criteria of classification (and knowledge organization) are thus deeply related to relevance. It is not the “inherent” qualities of the objects as much as it is the pragmatic criteria that determine the criteria of classification.

One well-known division between kinds of classification systems and ways of classifying is between Aristotelian classification and prototype theory. Prototype classification was defined by experimental psychologist Eleanor Rosch (1978) inspired by Wittgenstein’s later philosophy. “An Aristotelian classification works according to a set of binary characteristics , which the object being classified either presents or does not present. At each level of classification, enough binary features are adduced to place any member of a given population into one, and only one class. . . . A technical classification system operating by binary characteristics is called monothetic if a single set of necessary and sufficient conditions is adduced . . . polythetic if a number of shared characteristics are used. . . . Aristotelian models―monothetic or polythetic―have traditionally informed formal classification theory in a broad range of sciences, including biological systematics, geology, and physics.

Rosch’s (1978) prototype theory argues that, in daily life, our classifications tend to be much fuzzier than we might at first think. We do not deal with a set of binary characteristics when we decide that this thing we are sitting at is a chair. Indeed, it is possible to name a population of objects that people would in general agree to call chairs that have no two binary features in common.” (Bowker, 1998, p. 256). (Wittgenstein's view and thus prototype theory is criticized by Sutcliffe, 1993).

"Any set of entities can be classified in indefinitely many ways. Books can be classified according to author, title, subject matter, and so on. Although these various classifications can be integrated into a single reference system so that any book can be retrieved on a variety of different counts, books as physical objects can be arrayed on library shelves in only one order. Even here, contingent problems arise. For example, very large books may have to be shelved out of order, and expensive first editions may be locked away in separate collections.

More stringent requirements apply to scientific classifications. The ultimate goal for scientific classifications is to group entities so that these classes function in, or facilitate the formation of, scientific laws (see Laws, natural §1). Aristotle divided motion into super- and sub-lunar as well as forced and natural (see Mechanics, Aristotelian). The primary justification of his classification was the system of laws that he was able to generate using it. When Newton introduced his quite different system, his classification replaced Aristotle’s because Newton’s system of laws was more powerful, accurate and inclusive. In general, systems of scientific classification are intimately connected to scientific theories and cannot be evaluated independently of them. Different sorts of theories require different classifications.

One major difference is between structural and historical classifications. The periodic table of physical elements is structural. The elements are individuated and ordered linearly according to their atomic number. Hydrogen comes first, then helium, lithium, and so on. These elements in turn can be arranged hierarchically as metals, rare earths, and so on. Some of these arrangements are perfectly nested; others are not. A more contemporary classification would include reference to subatomic particles and their relations (see Chemistry, philosophical aspects of §4). In general, structural hierarchies do not include very many levels. They are not very deep. Although cosmology is a legitimate area of physics, no one has suggested a historical classification of the physical elements; for example, classifying them in the order in which they appeared after the Big Bang." (Hull, 1998).

Elin K. Jacob finds that classification and categorization are different kinds of processes:

"The Process of classification involves the systematic assignment of entities to groups or classes according to an established set of principles. Classification entails a one-for-one slotting of objects, events, or properties, based upon the apperception of a core of necessary and sufficient characteristics, into mutually exclusive classes within the hierarchical structure imposed by an arbitrary and predetermined ordering of reality. Categorization, on the other hand, refers to the process of dividing the world of experience into groups --or categories--whose members bear some perceived relation of similarity to each other. In contrast to the process of classification, the process of categorization entails neither that membership within a category is determined by the apprehension of a set of definitive characteristics nor that inclusion within one category prohibits membership within any other category. It is this recognition of similarities between otherwise unlike entities and the subsequent identification of categories that permits the individual to discover order within an otherwise complex and chaotic environment". (Jacob, 1991, p. 78).

Jacob's view is however only associating one epistemological view with classification. If we regard, for example, Ereshefsky (2000), we see that the philosophy of classification consist of different "paradigms":

What Jacob describe as "classification" is only one view in classification, essentialism, why her distinction is not correct. What she terms "categorization" have some similarities with clustering (see also category in Epistemological Lifeboat).

Now, we could simply conclude with Dolby and others that library classification continues mainly as a practical matter, that it is by and large devoid of substantive intellectual content, and that it continues merely because of inertia in a field in which classification schemes invented late in the nineteenth century continue to be used (Dolby, 1979, p. 187; Mayt, 1982, pp. 1-48)" (Miksa, 1998, 49).

Schmidt & Wagner (2004) introduce some distinctions between classification and other forms of discriminatory practices:

"The point we want to make is that we have to be quite specific in distinguishing different types of discriminative practice: seeing something, seeing something for what it is as opposed to something else (reflecting on what one is seeing), physically separating things in some regular way, saying that x is C (‘categorizing’ x as C), and classifying x as C according to an inscribed, publicly available classification system. These are radically different practices, involving radically different forms of convention, principles of abstraction, etc." (Schmidt & Wagner, 2004, pp. 45-46).

Bowker, G. C. (1998). The kindness of strangers: Kinds and politics in classification systems. Library Trends, 47(2), 255-292.

Bowker, G. & Star, SL. (2000). Sorting Things Out: Classification and its Consequences. Cambridge, MA: MIT Press.

Burke, P. (2000). A Social History of Knowledge: From Gutenberg to Diderot. Cambridge, UK: Polity. (Chapter 5, Classifying Knowledge: Curricula, Libraries and Encyclopedias, pp. 81-115).

Dolby, R. G. A. (1979). Classification of the sciences: The Nineteenth Century Tradition. IN: Classifications in their social contexts. Ed. by R. F. Ellen & D. Reason. (Pp. 167-193). New York: Academic Press.

Ereshefsky, M. (2000). The Poverty of the Linnaean Hierarchy: A Philosophical Study of Biological Taxonomy. Cambridge: Cambridge University Press.

Feger, H. (2001). Classification: Conceptions in the social sciences. IN: International Encyclopedia of the Social and Behavioral Sciences. (Vol. 3, pp. 1966-1973).

Hjelmslev, L. (1943) Omkring sprogteoriens grundlæggelse. København: B. Lunos bogtrykkeri a/s. (Many later editions and translations, e.g. Prolegomena to a theory of language. translated by Francis J. Whitfield. Baltimore : Waverly Press, 1953).

Hull, D. L. (1998). Taxonomy. IN: Routledge Encyclopedia of Philosophy, Version 1.0, London: Routledge.

Jacob, E. K. (1991). Classification and Categorization: Drawing the Line. In B. H. Kwaśnik & R. Fidel (Eds.), Proceedings of the 2nd Annual ASIS SIG/CR Classification Research Workshop (pp. 63-79). Washington, DC.

Marradi, A. (1990). Classification, Typology, Taxonomy. “Quality and Quantity” XXIV, 2 (may 1990): 129-157. Available at: http://web.archive.org/web/20040705070709/http://www.unibo.edu.ar/marradi/classqq.pdf

Mayr, E. (1982). The growth of biological thought: Diversity, evolution, and inheritance. Cambridge, Mass.: The Belknap Press of Harvard University Press.

Miksa, F. (1998). The DDC, the Universe of Knowledge, and the Post-Modern Library. Albany, NY: Forest Press.

Rosch, E. (1978). Principles of categorization. IN: E. Rosch & B. B. Lloyd (Eds.): Cognition and categorization (pp. 27-48). Hillsdale, NJ: L. Erlbaum Associates.

Schmidt, Kjeld & Wagner, Ina (2004). Ordering systems. Coordinative practices and artifacts in architectural design and planning. Computer Supported Cooperative Work. The Journal of Collaborative Computing, 13(5-6), 349-408. Retrieved 2008-01-21 fra: http://www.itu.dk/~schmidt/papers/ordsys_preprint.pdf

Sokal, Robert R. 1974. Classification: Purposes, Principles, Progress, Prospects. Science, 185 (4157): 1115-1123

Sparck Jones, K. (1970). Some thoughts on classification for retrieval. Journal of Documentation, 26(2), 89-101. Click for 2005 reprint pdf

Sparck Jones, K. (2005). Revisiting classification for retrieval. Journal of Documentation, 61(5), 598-601. [Reply to Hjørland & Nissen Pedersen, 2005] Click for full-text pdf

Sutcliffe, J. P. (1993). Concept, class, and category in the tradition of Aristotle. In: van Mechelen I, Hampton J, Michalski R S, Theuns P (eds.) Categories and Concepts. Academic Press, London, pp. 35-65.

Indenfor informationsvidenskab anvendes betegnelsen "klassifikation" især om særlige typer af *emnedata eller *"information retrieval languages". Et klassifikationssystem i denne snævre, dagligsprogs betydning er eksempelvis Dewey Decimal Classification, i modsætning til f.eks. "verbale emneordssystemer" som *tesauri.

Det, der karakteriserer et klassifikationssystem i denne snævre betydning er, at det foretager en systematisering af et givet emne-univers i klasser og subklasser udfra en "top-down" behandling af det pågældende emne-univers (fra en helhedsbetragtning af emneuniverset til det enkelte begreb f.eks. fra "psykologi" til "børnepsykologi" til "kognitiv udvikling"). Klassifikationssystemer er endvidere karakteriseret ved at være kodede (modsat verbale) ir-sprog, og disse koder er strukturvisende. Tesauri klasssificerer også deres emneunivers, men dette gøres mest udpræget udfra en "bottom-up" (empirisk baseret på konkrete dokumenter) behandling (fra det enkelte begreb til dets over-, under- og relaterede begreber; naturligvis er der overgange, og de to tilgange kan forenes).

Klassifikationssystemer i denne snævre betydning har en stor udbredelse i biblioteker og bibliografier, databaser m.v. De anvendes dels til ordning (*orden), såvel fysisk opstilling af bøger, som til ordning af *dokumentrepræsentationer f.eks. i trykte referatpublikationer. Dels anvendes klassifikationssystemer til genfinding i fastformskataloger og online databaser. Som genfindingssystem er klassifikationssystemer "umoderne", trenden i tiden går tydeligt i retning af verbale emnedata, som anses for mere brugervenlige.

Lancaster (1991, side 126) understreger, at klassifikation i dens bredeste betydning ikke kun skal forstås som ovenfor nævnt, men indbefatter alle former for lagring og genfinding af informationer. Således er "oversættelsen" af et begreb til en notation eller en "label" som et indexeringsord eller -frase en indplacering af dokumentet i en klasse af dokumenter, d.v.s. en klassifikation. Den begrebsmæssige analyse, der finder sted ved såvel "klassifikation" som "indexering" udgør dermed en klassifikation i bred betydning.

Et meget vigtigt biprodukt ved klassifikation er, at man tvinges til at definere sine klasser, d.v.s. klassifikation fremmer begrebsmæssig klarhed, hvilket skyldes den nære sammenhæng mellem klassifikation og definition (Jfr. Jørgensen, 1963, side 192).

Klassifikationssystemer kan opdeles efter form og indhold. Efter form kan man f.eks. skelne mellem *enumerative klassifikationssystemer og *facetterede klassifikationssystemer (samt indexeringssystemer, hvis man taler om klassifikationssystemer i bred forstand). Efter indhold kan man skelne mellem universelle klassifikationssystemer og fagafgrænsede klassifikationssystemer.

Foruden at være en *BDI-faglig disciplin er klassifikation også en filosofisk-logisk-matematisk disciplin og videnskabernes klassifikation udgør et særligt filosofisk problem, der er særdeles relevant for den informationsvidenskabelige teori (Jfr. *"Klassifikation, videnskabernes"). Foruden klassifikation anvendes også begreberne taxonomi, typologi og kategorisering. Nogle forfattere, f.eks. Adams & Adams (1991) skelner mellem klassifikation og taxonomi, hvor en taxonomi udgør en særlig form for klassifikation, der har en specifik hierarkisk egenskab, en klassifikation, i hvilken mindre og mere specifikke klasser eller taxa (ental taxon) er grupperet i større og mere generelle ditto. En typologi er en form for *klassifikation, specielt udformet med henblik på at sortere elementer i gensidigt udelukkende kategorier, der betegnes typer.

Begrebet "kategori" anvendes i lidt forskellige betydninger. Filosofisk taler man om kategorier, grundbegreber. Aristoteles, Hegel og andre store filosoffer opererer med kategorisystemer for alt værende. Mammen (1983, side 306) definerer en kategori som en mængde af genstande, for hvilket det gælder, at det for enhver genstand, der tilhører den, også kan afgøres, at den tilhører den. Denne definition stiller Mammen i modsætning til kognitivisternes "universelle" klasser af genstande alene defineret ved deres almene sensoriske eller funktionelle egenskaber.

Formelt-logisk opererer man ofte med to krav til et klassifikationssystem (Jfr. f.eks. Jørgensen, 1963, side 190):
1) klassifikationssystemet skal være eksklusivt, d.v.s. intet element må være indeholdt i mere end een klasse (klasserne må ikke overlappe hinanden)
2) Klassifikationssystemet skal være exhausivt, d.v.s. klasserne skal udtømme emnet (der må f.eks. ikke være bøger, der ikke kan klassificeres i det pågældende system, såfremt de falder ind under dets område).

Disse formelle betingelser er sjældent opfyldt i virkelighedens verden. Hvis klasserne f.eks. består af fag, overlapper fag som psykologi, psykiatri, sociologi og pædagogik meget. I moderne kognitiv videnskab er man begyndt at vurdere de formelle klassifikationskrav udfra nye synsvinkler. Istedet for at et element enten er eller ikke er medlem af en given klasse, taler man idag om grader af slægtskab og om mere eller mindre typiske medlemmer (f.eks. mere eller mindre typiske medlemmer af klassen "grøntsager"). Også i den moderne "fuzzy logic" har man opgivet det veldefinerede klassebegreb.

Når talen er om klassifikationssystemer som middel til bogopstilling er kravet om ikke-overlappende klasser naturligvis stærkt, idet den samme bog ikke kan placeres flere steder rent fysisk. Det er imidlertid et spørgsmål, hvorvidt disse kriterier ikke er mere skadelige end gavnlige, når det drejer sig om moderne *emnedatasystemer i *elektroniske databaser.

En meget væsentlig skelnen går mellem naturlig og formel klassifikation (behandlet af bl.a. Kedrow, 1976). En naturlig klassifikation afspejler f.eks. en videnskabs alsidige forbindelser til andre videnskaber, mens en formel klassifikation ordner alle videnskaber efter eet enkelt eller ganske få formelle karakteristika.

Et givent *vidensdomæne kan klassificeres udfra forskellige udgangspunkter: Indenfor *BDI området klassificeres *dokumenter. Denne ordning kan kaldes bibliografisk. Selve domænet beskrives ofte i håndbøger og encyclopædier og gives her en fremstillende orden, der bygger på en implicit eller eksplicit metodologisk/videnskabsteoretisk opfattelse af emnet. Denne orden kan kaldes den faglige ordning eller den filosofiske ordning. Til undervisningsbrug ordnes stoffet på en pædagogisk måde, således at elementer, der forudsætter andre elementer sættes efter disse. Det er en pædagogisk klassifikation. Ordningen kan også tage hensyn til fagforeningsmæssige, sociologiske, arbejdskraftsmæssige forhold. Der findes endnu flere udgangspunkter for ordning. Det interessante er naturligvis, i hvilken grad disse forskellige udgangspunkter overlapper med hianden. Dette problem er berørt i artiklen om *vidensrepræsentation.

Man kan stille det spørgsmål, om klassifikation overhovedet er en nødvendig eller frugtbar aktivitet. Dette diskuteres af Rescher (1979), der peger på, at viden om sammenhænge på den strukturelle side af naturlige taksonomier i kognitive systematiseringer udgør et nøgleaspekt af faktuel viden i sig selv.

Her er vi så ved eet af informationsvidenskabens grundlæggende problemer: Systematiseringer opstår og udvikler sig samt ændres løbende efterhånden som den menneskelige erkendelse og den videnskabelige viden udvikles. Som McGarry udtrykker: "Knowledge is a cultural entity and keeps shifting its pattern like a kaleidoscope. An emergense of the new knowledge modifies the structure of the whole. Contrary to H.E.Bliss (1870-1955) there is no permanent order in knowledge. "Pattern is new every moment" said T.S.Eliot (1888-1965), with a poetic vision".

Klassifikationer i *BDI-faglig forstand skal afspejle denne erkendelsesmæssige udvikling. Ethvert forsøg på at etablere præetablerede systemer, som den løbende erkendelsesudvikling presses ind i, er et udtryk for rationalisme og "objektiv idealisme", d.v.s. for det synspunkt, at der findes givne principper og "idéer", der er evige og uforanderlige, som går forud for den videnskabelige erkendelse, eksisterer apriori. Da informationsvidenskaben gerne vil selvstændiggøres fra den enkelte videnskabelige fag ligger der en permanent idealistisk tendens og lurer: Tendensen til at skabe systemer, der ikke afspejler, men udgør alternativer til faglige systemer: til at informationsvidenskaben bliver en særlig "idé-verden" uden tilstrækkelig forbindelse med den virkelige verden. Svaret på dette problem ligger i at arbejde med generaliseringer af vidensstrukturer, d.v.s. at arbejde filosofisk.

Adams, W. Y. & Adams, E. W. (1991). Archaeological typology and practical reality. A dialectical approach to artifact classification and sorting. Cambridge: Cambridge University Press.

Jacob, Elin K.: Classification and Categorization: Drawing the Line. Advances in Classification Research. Vol. 2. Proceedings of the 2.nd. ASIS SIG/CR Classification Research Workshop. Washington, D.C.: American Society for Information Science, 1991, pp. 67-83.

Jørgensen, Jørgen: Psykologi på biologisk grundlag. København: Munksgaard, 1963. 683 sider.

Kedrow, B. M.: Klassifizierung der Wissenschaften. Band 1-2, Köln: Pahl-Rugenstein, 1976. (Oversat fra russisk).[Et tredie bind er omtalt i bd. 2; det vides ikke, om det er udkommet]

Klassifikation der Wissenschaften (i: Europäische Enzyklopädie zu Philosophie und Wissenschaften. Band 1-4. Herausgegeben von Hans Jörg Sandkühler u.a. Hamburg: Felix Meiner Verlag, 1990).

Lancaster, Frederick Wilfred: Indexing and abstracting in theory and practice. London: Library Association, 1991.

McGarry, Kelvin: Epilogue. (i: Knowledge and Communication. Essays on the Information Chain. Ed. by A.J.Meadows. London: Library Association (A Clive Bingley Book): 1991. 164 sider. [Her citeret fra anm. i Int.Classif. 1992, s.41]

Mammen, Jens: Den menneskelige sans. Et essay om psykologiens genstandsområde. København: Dansk Psykologisk Forlag, 1983. 543 sider

Mills, Jack: Library Classification. Journal of Documentation, 1970, 26, 120-160.

Mills, Jack & Vanda Broughton: Bliss Bibliographical Classification. 2. ed. Introduction and Auxilliary Schedules. London: Butterworths, 1977. 209 sider (Specielt kap. 4: Organizing Information and the Role of Bibliographic Classification; Kap. 5: The Structure of a Bibliographical Classification & kap. 6: The Bliss Bibliographical Classification (BC).

Needham, Joseph: Integrative Levels: A Revaluation of the Idea of Progress. (1937 Herbert Spencer Lecture, reprinted in: J.Needham: Time: the Refreshing River. London, 1943.

Ranganathan, S.R.: Prolegomena to library classification. Ed. 3. London: Asia Publ. House, 1967.

Richardson, Ernest Cushing: Classification: Theoretical and Practical. 3.ed. New York: 1930.
Rescher, Nicholas: Cognitive Systematization. A Systems-theoretic approach to a coherent theory of knowledge. Oxford, 1979.

Samurin, E.I.: Geschichte der bibliotekarisch-bibliographischen Klassifikation. Band I-II. München: Verlag Dokumentation, 1967.

Moustgaard, L. (1985). De Dahl-Ankerske kataloger på Universitetsbibliotekets 2. afdeling. En klassifikationsteoretisk analyse af kataloger og emneregister. København: Universitetsbibliotekets 2. afd.

Weitemeyer, M. (1991). Klassifikationsbetænkning. [København: Det kgl. Bibliotek].

Se også *Information retrieval languages; *"Klassifikation, Videnskabernes" *Klynge; *Kundskabsorganisation; *Logik; *Orden.

to carve nature at its joints (Plato)