Classifying the humanities
Rick Szostak · Proceedings of the American Society for Information Science and Technology · 2013
Classification in the humanities faces a host of unique challenges which existing classification systems struggle to meet. The purpose of this paper is to explore whether a new approach to classification is able to better address these challenges. The humanities are defined here as the study of works of art (including literature)1—art itself will be defined below. One challenge for the classificationist is that it is necessary both to classify works of scholarship and the works of art themselves. This paper might have been organized in terms of these challenges. Some of these are briefly surveyed in the next section to provide some context for this paper. But since many challenges will have the same solution it is best to organize the paper around solutions. The third section of the paper briefly outlines the nature of the approach to classification pursued here (which has been developed in many other works). Succeeding sections of the paper then address how the key elements of this approach can be applied to the humanities. In each section it will be shown how particular challenges are thus addressed. Universal classification schemes are often developed with the social and/or natural sciences primarily in mind, and then stretched to fit the special needs of the humanities. This is the case here as well, but the inherent flexibility of the new approach will allow much better treatment of the humanities. Nevertheless important adjustments are called for within some sections in order to better address the needs of the humanities. The result is a classification which should significantly enhance both humanities scholarship and the ability of the general public to satisfy their curiosity about works of art. Ørom (2003) speaks of “the “bricolages” of today's classification schemes. The classifications most in use today were developed over a century ago, and have developed slowly over the succeeding decades. The result in the humanities is that the classifications thus include elements of many schools of thought regarding the arts over the last centuries, but do not represent any one coherent view. The challenge, then, is to take a new approach that encompasses these many ways of looking at art. Medaille (2010) notes that theatre artists find navigating large research libraries troublesome. Yet they list ‘seeking inspiration’ to be one of their six main goals. If artists would turn to the literature for inspiration, but find it daunting to do so, then society suffers. And if they find libraries daunting, it must seem that it is not easy for them to find what they need. Inspiration will often come in art in the same way that innovation occurs in science: by juxtaposing previously separate ideas. In art as in science, then, we need a classification system that alerts users to items that might be related to their core area of interest. “The traditional conception of “subject access” posits that the artifact creator encodes a stable and transparent meaning into an object, and the user (or reader, or cataloger) then decodes the meaning while interacting with the artifact … the cataloger then translates that decoded meaning into easily understandable subject headings, with which other users interact, for the purpose, presumably, of finding artifacts that meet their needs. This … overlooks the fact that while many artifacts, like scientific texts, strive for transparency and stable interpretation; there are other types of artifacts, like poetry and fiction, for which numerous interpretations are both valid and supported by the creator's intention. Representing all works using the scientific model ignores some of the most powerful characteristics of those works whose meaning is not fixed or obvious; and in terms of library systems, access then runs the risk of being limited by this representation.” Simply put, the “subject” of a work of art is harder to establish than is the subject of other works.2 And as Winget proceeds to explain, the subject is multi-dimensional. Following Panofsky, she sees three levels. One level simply describes the main elements (woman on horse). Another gives specifics (name of woman).3 The third records cultural significance (e.g. Christian parable). Information scientists have generally thought the third in particular too difficult. But as we shall see existing classification systems do a poor job of representing the first two levels as well. Works of art may capture static elements (as in woman on horse) and/or they may tell a story. They may represent philosophical or cultural ideas. “Subjects include things, places, activities, abstract shapes, decorations, stories, and events from literature, mythology, religion, or history. Philosophical, theoretical, symbolic, and allegorical themes and concepts may be subjects. Subjects of representational (figurative) works may be narrative, meaning that they tell a story or represent an episode in a story. They may also be non-narrative, representing persons, animals, plants, buildings, or objects depicted in portraits, still-lifes, landscapes, genre scenes, architectural drawings, allegories, and so on. Nonrepresentational works also have subject matter, which may include a reference to abstract content, decoration, function, or implied themes or attributes.” Baca et al (2006) notably suggest that this asks too much of a classification system. But we should see whether a new approach to classification can encompass all of these elements. And so far we have dealt only with what a work is about. For works of art, form or style is also important. A work may express a certain artistic style. It may be purely aesthetic in intent or may have practical uses. It may employ particular materials or techniques. And last but not least we may be interested in not just a work's creator – and many works may have multiple creators – but who has owned a work, where it has been exhibited/performed, the time and place it was created and so on. Existing classification schemes cope poorly with these challenges. Dozens of classifications are used in music libraries, suggesting that classification is difficult and existing schemes do it poorly (Lee 2011). The Art and Architecture Thesaurus was developed in large part to compensate for the manifest limitations of existing approaches to classifying art. As for humanities scholarship, this has inevitably struggled to cope with the multidimensionality of works of art. It is limited by the inability of existing classification systems to cope well with most (some would say any) of these dimensions. A further challenge for both scholar and classificationist/classifier is that humanities scholarship is characterized by a multitude of theories and concepts, both of which change through time. Breaking complex concepts – which lend themselves to differing interpretations across disciplines and groups – into basic concepts that carry broadly similar meanings across individuals and groups. These basic concepts generally refer to ‘things’ in the world or relationships among things (see Szostak 2011) Works and ideas are then classified in terms of free combinations of any set of basic concepts Most works and ideas will be classified in terms of a combination of things and relationships. A set of adjectival/adverbial qualifiers is also developed: these can be freely combined with things and relationships. Works will also be classified in terms of theories, methods, or perspectives applied in the work. Fairly small schedules thus allow very detailed and precise classifications of works or ideas (see Szostak 2012a, b) Szostak (2011) argued that ambiguity could be substantially reduced by breaking complex concepts – which are understood differently across disciplines or groups – into their constituent basic concepts which are understood in a broadly similar fashion across groups and disciplines. Conceptual atomism suggests that shared understanding will be most likely for things and relationships that we regularly perceive. I thus argued in favor of a classification of works in terms of combinations of such things and relationships.4 The vast majority of scholarly works address how one or more things affect in a particular way one or more other things. It will be argued below that this is true not only of humanities scholarship but often of the works they study: (woman)(riding)(horse) or (gods)(celebrating). This approach is best facilitated by allowing ‘things’ (women, horses, gods) and ‘relationships’ (riding) to be freely combined in both classification and search. This spares the classificationist from having to enumerate a vast array of combinations, and the user from having to ascertain how a particular combination was treated. The broad outlines of such a classification can be found in Szostak (2012a); the classification of things is treated in Szostak (2011), and of relationships in Szostak (2012d). Szostak (2013) noted that this classification instantiates a ‘web of relationships’ approach, which allows users to readily follow their curiosity from one work to another. They can easily identify works that relate a subject they have just read about to any other. They can also readily identify works that have studied the same relationship in a different context, or applied a particular theory or method in a different context. We will first need to define art itself. Szostak (2003) defined art as that which appealed to our aesthetic sensibility. At least at this point in human understanding we do not understand humanity's aesthetic sense well enough to determine whether a particular artifact is “art.” We must thus, along the lines of literary warrant, accept as art anything that anyone claims to be art. Many artifacts – notably pottery and textiles – are intended to be both useful and aesthetically pleasing. Indeed most of the artifacts created by humans embody some aesthetic elements. Our buildings, bridges, shoes, cutlery, furniture (the list goes on) could all be made much more cheaply if we focused only on their function and not their appearance. Works that address the practical side of such artifacts are usually given quite different subject headings, and shelved in quite different parts of the library, from works that treat their aesthetic side. One of the values of a ‘web of relations’ approach is that it should be much easier for users to move between studies of aesthetics and of practicality. And this in turn could encourage ever-better integration of ‘beauty’ and practicality in our lives. But should ‘china’ or ‘fashion clothing’ themselves be classified as works of art or as items of utility? General practice tends toward the latter, except when a work stresses their aesthetic elements. This is perhaps unfair: a set of china that is put on display in a china cabinet year-round, and only pulled out once a year for a special family dinner, is likely mis-classified as an item that is primarily useful. But we hardly want either classificationists or classifiers to have to decide where particular artifacts (or works about these) fall on a continuum of art to practicality. But if we will follow common practice and (privilege utility over art by) classifying most artifacts as useful artifacts, then it becomes particularly important to be able to readily identify aesthetic treatments of such artifacts. In other words, we have here a diverse set of artifacts (and many works about these) that do not fit neatly into just one hierarchy. And the hierarchies of works of art tend to be quite distant in most libraries from the hierarchies of utility. Such artifacts are inevitably treated poorly in “universal” classifications that are organized around disciplinary silos, and treat engineering and economic outcomes in a quite different fashion from works of art. These artifacts reflect an intersection of utility and aesthetics that can only be captured appropriately by a classification that stresses relationships. Having raised the issue of hierarchy, we can now address how we will classify different types of art. Szostak (2013) argued that much (maybe all) of the debate regarding how best to identify subclasses dissolves if we appreciate the difference between things and relationships. For example it is often suggested that pharmacologists would want a quite different classification of chemicals from what chemists want. But in fact what pharmacologists are interested in causal relationships such as (chemical)(reduce)(blood pressure) and their concerns are thus not just adequately but best addressed through a classification that stresses free combinations of things and relationships. Some would classify art by purpose. But “art as propaganda” is a causal relationship, as indeed is “art enhances understanding” or “art evokes a better world.” Some would classify art by audience, distinguishing fine from popular art. This distinction has long vexed information scientists, for the boundary is unclear and shifting. Much better to treat art as anything thought to be art, and then distinguish works describing that art in terms of what they see as audience or purpose of a work utilizing causal relationships. In particular, masterpieces are works of art thought to be particularly aesthetic. Some would classify art primarily in terms of time and place it was developed. Such distinctions are readily captured by allowing free combination with classifications of time and place. Finally, some would emphasize differences in materials, but again these are best seen as relationships: (art)(made from)(wood). Some would distinguish art in terms of artistic styles. We will make the novel argument below that styles can be treated in a very similar manner to scientific theories. We are left, then, with the classification of art in terms of “type,” the standard avenue of developing subclasses (occasionally replaced by “parts of”) in the Basic Concepts Classification (Szostak 2012a). In the case of art, “type of” primarily captures what is often termed “medium”: Painting Sculpture Collage Cartoon Graphic art rose Poetry (rhyming or not) Theater Film Photography Music Dance The reproducible versus non-reproducible distinction is stressed here because it has a host of implications for how art is produced and appreciated. Though it is a logical distinction, it is one that reflects the importance of particular relationships: it was thus discovered inductively rather than deductively. Several of these subclasses merit further subdivision. At this lower level we also seek subclasses in terms of medium. In the case of music, this would involve distinguishing singing from a variety of musical instruments.5 We would want a notation that allowed easy recognition of when two or three different types of music were combined (and also some notational shortcuts for indicating common combinations of many instruments such as orchestras). And we would want our subclasses to be free of cultural bias: instruments common in non-western music should not be disadvantaged. Poetry can be distinguished by rhyming scheme and rhythm (and again we want cultural fairness such that haiku is treated similarly to sonnet). Painting can be distinguished by: oil versus water, canvas versus fresco. Film likewise can be subdivided in several ways: silent, black and white, 3D, large screen, made-for-television, 360 degrees.6 In all cases we want flexibility to allow new mediums to be recognized. This classification does serve to distinguish art in terms of human senses: auditory, verbal, visual, mixed. It is not explicitly organized in terms of these. In part this is to not ground the classification unnecessarily in an approach where there is scope for dispute: some have claimed that sculpture is really an art of touch. Arguments about how various senses are applied in both production and appreciation of art are best captured through causal relationships. The classification is flexible such that new subclasses can be added. It is fairly standard, except for incorporating literature within art (where it logically belongs). Architecture was included in previous versions of the BCC but is, admittedly, a misfit. The argument above would suggest that we treat buildings primarily as items of utility, and capture their aesthetic elements through causal linkages: Architecture is thus (aesthetics)(applied to)(buildings). And if we were to include architecture then we risk the slippery slope of demands to include pottery and other artifacts that mix utility and art. It is also necessary to classify works of art by artist. Indeed most art scholarship focuses on individual artists (Ørom 2003). This is probably done best separately from the classification above, since artists often produce more than one type of art, and it is then annoying to have to search for them multiple times (Ørom 2003). But each artist should then be connected to the type(s) of art they produce. As I have argued in many places (Szostak 2011, 2012d), most scholarly works, and likely most general works, investigate some causal link(s): how one or more of the things in the world affects one or more others. This is true of humanities scholarship as well. Existing classification systems unnecessarily privilege some causal links over others. There is thus tremendous value in moving toward a system that allows works to be freely classified in terms of any causal relationship. The non-human environment provides both raw materials and inspiration. Art in turn affects both how we perceive nature and how we construct the built environment, Our genetic predisposition generates aesthetic universals (literary theory at times suggests that these do not exist), our senses; and our ability to structure say music; many have posited that humans are evolutionarily selected for art because it teaches, raises the spirit, and creates bonding (note that these are then links from art to yet other phenomena) Cultural elements and aesthetic sense are combined in works of art are generally seen to express certain cultural values or Art is in all Art may cope with cultural As for individual artists likely display and and perhaps so do their And if art does then art what and Art may be both the for art and of art. Art may also reflect and economic As noted above, artistic the of all various The fact that many have artistic suggests a that art can have important structure affects who becomes an artist and who the arts and thus likely the of art. Art can It can thus either encourage or level social the and of art. As noted above, art the of we The importance of art to human is in a classification that does not allow the on and of art to be readily Yet in practice art scholarship has some links more than others. And classification systems in literary thus make it easier to investigate some links than others. Ørom in a of art, is to emphasize what art have in fact Yet very treatment the fact that the of art scholarship has through time. art the of one artist or style on it a at artist time this was by an on the cultural meanings of works of art is the of There was a in economic with This provides a further argument in favor of a logical classification that provides a place for any a scholar might to If we ground our classification in what emphasize today we will be for what they As Ørom (2003) the new art since the has been in and thus poorly by In the section we made a fairly for a classification that can encompass all of the causal links that humanities scholarship might In this section we make a more that the same should be applied to the works of art that they It is useful to to Ørom Ørom (2003) an importance of study in art this at art in and there are and The of and and and The of The the art and the of the works are to the If of art want to study such and if art to works that such then we should to classify works in terms of the themes they how much easier it would be for – or indeed anyone interested in art – to explore how across different times and places any particular if we this The in classifying art is the same as the above with to the scholarship of that we captured by the themes thought to be important at some place and time. and the to classify in terms of a or so The classification of this type would classify and the as of then or or and more “The as the their such and type are subject to often that the are And it is that these are not regularly but need to There are also that these particular are not themselves part of classification and thus users must then works themselves. It would be much better to allow works of art (including to be freely classified in terms of (and themes within our classification There are of practical challenges. There is a of in the of a work, and this for art works that are And the meanings to particular works may change through time. works may be by as quite But ambiguity is a that information must the subject or of a work of art may on be more than the subject of a work of this is not the some works of art are easier than some works of to And of we do to some to classify the subject of art works within our But we use a small set of terms rather than allowing free use of any in our classification (and a work of art might seek to express any or relationship that humans And we classify different types of art in terms of different subjects. is classified as or or not a set of that was are classified as or or The a more precise animals, (which Ørom represent a sense of art for our it is that these are each basic concepts or very combinations rather than very complex If we will to classify art in terms of some not classify works of art in terms of any that they seem to Some works of art may be about a (or perhaps a rather than some If a work to be about or or so be But most works are better in terms of combinations of basic or And works of art will express some causal One of an classification is that a user needs to identify precise of which to search. If is freely from a logically organized then the user need not If they search for they will find of if is in some way a logical of woman the is by woman and an We noted above that humanities scholarship and can thus any classification in literary The same is true for art itself. Art is an where artists on what has but to This will among other things that artists will seek out new that have not been treated We thus need to be able to classify works in terms of any We to artistic from juxtaposing ideas. The same is true of artistic inspiration. artists easy access to the diverse that have been in art might an important of inspiration. As noted above, abstract art will a special It may be to classify such works (and in terms of and does have a for If this is I that BCC classifications of and There is a more classification of that might be out if this useful in classifying These are further address when qualifiers are below. Most of what needs to be about relationships was captured in our of causal relationships. But it to be stressed that what is important about a work of art is often some if a is of a we will not be able to it very well if limited only to the is to classification and scholarship is likewise characterized by relationships. These are as when art or is by cultural They may also be relationships. As we have art is an want to but from what are selected and thus we can usually point to a style of any time and but one that allows and a allow different styles to in the Much of art scholarship focuses on how a particular artist (or of these) was by other artists and/or the and how they in turn other artists and/or the And this of scholarship will be to classify and to search if we do not classify different types of the basic in the BCC are and These can be combined to of further It was noted above that elements of the form and of works of art might be Some of these are best captured through such as links to classifications of and and or to materials or But the Basic Concepts Classification (Szostak a of qualifiers that can be freely combined with any Many of these may be particularly useful in classifying both art and humanities They capture not only elements of form and but also of subject and and This classification of which the is just a has been developed inductively then organized into two and can readily be to if by humanities notes that information scientists have been about classifying works by for over a And a understanding of elements of may be either or with subject (or type of work, an issue addressed further