Albums in the Attic  

An Investigation of Photographic Metadata  

In the late nineteen-eighties, my father owned a camera which would  imprint the time of the photographic moment on the film. Every print  would bear a sequence of faded, red numbers on its edge. Perhaps the  camera had a fault, or perhaps my father never bothered to change the  batteries, but the imprinted time would invariably be incorrect. Those  numbers had their own logic, one that, to this day, I haven’t quite managed  to understand. But, as a child I would marvel at the manner in which even  that incorrect inscription would guide the way in which the photograph  was perceived. Those faded red numbers on the bottom-right corner of the  photographic paper would enter the image into a series that would  potentially enable me to understand the sequence in which the  photographs were taken. This was my first encounter with metadata, even  if it was one without realisation. It is now common to expect the digital  camera to record a host of information from shutter speed to the geo location of the photograph. These machine-recorded data guide the way in  which the image is perceived. Metadata not only shapes the way the object  is understood but also aids in its search and retrieval within digital  repositories. The metadata of each individual object affects the experience  of the entire collection. Metadata, then, provides a significant way in which  descriptions can be provided for objects that are not self-describing. This 1 article contends with, specifically, the possible ways of articulating the  material aspects of the photographic image. It is argued that beyond the  

These are objects that cannot articulate any information about themselves. Unlike 1 the printed book which reveals the author, the date, the publication house etc.  (inherent in the structure of the book), visual objects or pieces of music for instance,  cannot reveal any information about their nature. So, they are not ‘self-describing.’ 

content image (the indexical quality of the photograph), contextual clues  derived from inspections of its physical properties are important for a  coherent understanding of the object. Following an examination of the  physical nature of the photograph, its presentational forms are inspected;  the content image, the material existence, and the presentational forms of  the photograph, together, construct a consistent body of knowledge that  affords a deeper understanding of the object. An accurate digitisation of  the physical object into a digital image (specifically for research purposes),  then, must present both the content image and the physicality of the object  with sincerity. While current digitisation techniques have found accurate  methods of copying the content image, the description of its materiality  remains a challenge. If the physicality of the photograph is central to its  understanding, this article inspects the possibilities of providing the digital  referent with the material information. It presents an examination of both  the materiality of the photographic image (and its transformation into a  digital object), and the means through which the presentational forms of  the original may be inscribed in the digital referent.  

3.1 Photographic materiality  

While the image imprinted on photographic paper is important, it would  be an error to assume that the photograph may only be understood as an  image; scholarly study of the photographic image, specifically with the  ‘material turn’ in anthropology and cultural studies, has attempted to 2 inspect the photograph as a physical object. This ‘material turn’ expresses  the complexity of the social existence of objects and allows the  investigation of the photograph as a physical object whose understanding Material cultural analysis, from an anthropological position, questions the 2 assumed superiority of language over other forms of expression (Miller 1987: 9). In  the context of photography, Elizabeth Edwards (2002: 67) suggests the need to  ‘break, conceptually, the dominance of the image content and look at the physical  attributes of the photograph.’ is augmented by its form. Photographs have specific modes of circulation,  production, and consumption, and their inspection has potential beyond  the critiques of representation alone. The inherent bias of the indexical  qualities of the image over its material properties may overlook the social  and cultural contexts within which the photograph is born and used. ‘[A]s  material objects whose “currency” and “value” arise in certain distinct and  historically specific social practices’ (Tagg 1988), the stability of  photographic meaning remains a descriptive challenge. Arjun Appadurai  (1986) explains how objects and the practices within which they are  embedded are interwoven, cannot be read independently of each other. He  writes (1986: 5), ‘things have no meaning apart from those that human  transactions, attributions, and motivations endow them with.’ Geoffrey  Batchen (1997: 6) contends that the meanings of photographic objects are  dependent on the context in which they find themselves at any particular  instance. He argues that, ‘[a] photograph can mean one thing in one  context and something else entirely in another’ and that the identity of the  object is contingent on its ‘use’ in the physical world. Gillian Rose (2010:  18) echoes similar sentiments when she says, ‘what people do with  photographs is not an optional analytical extra; it is fundamental to  exploring photographs’ effects in the world.’ The examination of  photographs, as material objects, appears to present two inter-related  concerns: the first discusses the ‘plasticity of the image itself, its chemistry,  the paper it is printed on, the toning, [and] the resulting surface variations’  (Edwards and Hart 2004). The rationale here is that the choices made in  the making of photographs are rarely random. As Schwartz (1996: 58)  expounds, ‘the choice of ambrotype over paper print implies a desire for  uniqueness, the use of platinum over silver gelatine intimates an  awareness of status, the use of gold toning a desire for permanence’. The  second concern is that of the presentational forms — cartes de visite,  

cabinet cards, albums, mounts and frames are intrinsically linked with 3 photographs and have constituted a major consumer market since the  twentieth century, especially after the Kodak revolution of the early 4 twentieth century (Edwards and Hart, 2004).  

Digital objects (both born-digital and digitised artefacts) present a  conceptual challenge: Joanna Sassoon contends that ‘[b]y the direct  conversion of light into a digital format to create a stable image  “photographs” that only exist in the digital form can be seen in one context  as a truer version of photography (writing with light) than those that  require the creation of a physical intermediary to view the image in a  material form’ (2004). The central rationale of Sassoon’s argument takes  premise on the assumption that digital objects have no material existence.  This reveals a fallacy in our inspection of digital objects: the digital  artefact, without question, is a material object. To observe this materiality,  a separation between content and carrier of the object is required. The  photographic image is printed on paper, and this paper is the carrier of the  content image. Similarly the content of the book is carried in the physical  form of the book — the paper, glue, and ink that hold it together. In the  physical object (as opposed to its digital counterpart) the content and the  carrier are closely inter-linked to the point where their separation is  difficult. However, the carrier may change at different moments of time,  

 Photographs were, and still are, used in the production of many curiosities. For 3 instance, see Susan Legêne’s (2004) account of playing cards illustrated with  photographs (though the production of these were in the first half of the twentieth  century).  

 The technological changes, specifically with the introduction of the Kodak 4 camera, resulted in the device being smaller, portable, and cheaper. This allowed  enthusiasts to practice the form and the portability of the camera allowed them to  take landscapes and views with relative ease. This, to some extent, accounts for the  larger range of photographic subjects in the early twentieth century. See Mia  Fineman’s ‘Kodak and the Rise of Amateur Photography.’ Heilbrunn Timeline of Art  History. New York: The Metropolitan Museum of Art, 2000. Web. 21 November  2013. 

which may provide the object with different contexts: consider an image  that was first printed on photographic paper, then printed in a newspaper,  and then, perhaps, in a coffee shop book, displayed on a bookshelf, never  quite read. The different material existences of the image provide contexts 5 that locate it within different points in history. The digital object is,  similarly, carried by electronic circuits — a complex creation of technology  that is able to house a large amount of data within a tiny space. That the  digital object has materiality, then, is undeniable. The problem, perhaps,  in identifying this materiality lies in the degrees of separation between the  circuits and the perception of the object. To view an electronic image, a  screen — an enabler — is required. The experience of the object then is  governed by an intermediary system. The dislocation between the carrier  of the object and the experience of the object is perhaps the source of the  material fallacy. If it is acceptable to think of this separation between the  content and the carrier, it is possible to argue that the digital object is  merely one iteration of a different carrier. The digitised image, then, is the  original image, in a new material form. However, it must also be  recognised, that the materiality of the digitised object does not mirror that  of the original photograph. Within our current technological limitations,  the only possible method of attributing the material specifics of the  original photograph to the digital referent is through textual means. The  metadata of the digital object relates the materiality of the original. The  process of digitisation focuses on the content image, converting the  original image into a series of pixels, while the carrier of the photograph is 6 The circulation of the photograph moving between carriers and put into new uses, 5 derive new contexts. Brent Harris (1998), for instance, in his investigation of  photographs from Africa (1850-1950) provides a strong argument for examining the  circulation of the photograph. His examination reveals the contested meaning at  various stages of the photograph’s production, circulation, and consumption. 

 The process of digitisation, from a computational perspective, is quite fascinating. 6 A researched description, specifically investigating the process, may be found in  Jähne, 1995. 

described through the metadata of the object. Sassoon’s investigation of  the process of digitising photographs addresses some major concerns. She  writes (2004):  

[I]t is no longer an accepted canon that a photograph is merely a  print on paper, nor is it a simple and uncomplicated translation  between reality and its mechanical representation. […]the  digitising process can no longer be seen as merely changing the  physical state of a photograph from the material to the pixel. If a  photograph can be seen as a more complex object than simply an  image, digitising can be seen as more than simply a transformation  of state, or a transliteration of tones. The process of digitising  involves a more complex cultural process of translation — or a  change between forms of representation. 

What Sassoon identifies in this transformation is the complexity that lies  in articulating details about both the content image and the carrier, even if  this distinction is not explicitly made. The metaphor of translation, or even  transliteration, that she uses, is perhaps not entirely convincing; the  transformation from physical to digital is a re-configuration of state, one  that is consistent with any other similar transformation — from the shellac  disc to the audio tape or from celluloid to the video cassette. Translation,  even if the word is used loosely in Sassoon’s work, indicates an inherent  change in content and carrier. As discussed, digitisation (or effective  digitisation) is focused on preserving an accurate content image, even if an  accurate representation of materiality is beyond its purview. What the 7 

 Beyond the content image of the photograph, its physical peculiarities are also 7 recorded. The back of the photograph, any tears or marks, and any inscriptions are  also captured with the indexical image (this depends, though, on institutional  policy or individual motivations). The tactile response of the photograph or its smell  are more difficult to record. They may be described, but an inevitable bias is very  likely.  

word ‘translation’ does point towards is, in fact, the tension between  original and the copy, in the accurate representation of the source. What  she stresses on repeatedly in her essay is the importance of cultural  contexts within which these material objects reside. Any transformation of  state, whether physical or digital, creates a new context for the object; the  digital state has its own context, while our attempt is to record the context  of the source within the digitised.  

All digitisation activities are carried out for some specific purpose. The  purpose might be merely to record the content image for the purposes of  the collection. As we have already discussed, the desire of the collector is  central to the affect of the collection. For research collections, however, the  need to record more than the content seems to be necessary. The focus of  digitisation, then, cannot be only on a singular aspect of the photograph —  the indexical quality of the image. The relationship between the negative  and the paper forms of the photograph may reveal visible clues about its  technical origins. The physical dimensions of the photograph may reveal  details about the kind of camera used, the negative size, and a possible  date of production. The tonal ranges, and the photographic texture may  indicate the type of process used to develop the image. All these point to  the social existence of the object. Viewed on computer screens of different  sizes and different calibrations, it is easy to lose sense of the proportions  and tonal hues of the original image. Digitisation effectively alters the way 8 the photograph is viewed; using a loupe to magnify the details of the  content image, we are far readily aware of the material aspects of the  

 Enablers, for instance computer monitors, present their own problems. The 8 calibration of the screen (brightness, colour fidelity) along with its physical  dimensions change the perception of the digitised object. In most digital  collections, the digital referent does not possess a reference point against which it  can be compared to derive an understanding of the original. Cultural institutions  follow strict rules when digitising objects; unfortunately, they have no control over  the enabler and the perception of object, for the viewer, may be fundamentally  different from what its natural form is. 

photograph than that of the digital referent. The use of keyboards and  mice might allow new interactions, and the ability to magnify to extreme  points, but they fail to offer the tactile response of the physical object. The  fundamental difference in viewing the digital object, brought about  through mediation (the computer screen) might lead to questions about  the perception of the object. Beyond the material peculiarities, the imprint  of time on the object — the dirt and the tear, the slow fading of chemicals  on paper — is evidence of the lives of the photograph, about the  provenance of the image. Paratextual material such as captions, scribbles 9 on the back of the photograph, hand-colouring of black and white images,  are invaluable to its contextual understanding. While research on  advanced techniques of image representations point to the future of  digitisation, it is rarely available for average consumer use. How are we 10 to record these elements — the material, the contextual, and the  paratextual aspects of the image — with the range of technology afforded  to us?  

3.2 The Photograph and its Presentational Forms  

The photograph has multiple lives; it exists within socio-cultural contexts,  and to understand it, the content image must be seen in conjunction with  

 Roland Barthes, in his seminal work titled Camera Lucida: Reflections on 9 Photography, begins by describing the the material aspects of the image. He writes,  ‘The photograph was very old. The corners were blunted from having been pasted  into an album, the sepia print had faded, and the picture just managed to show two  children standing together at the end of a wooden bridge in a glassed-in  conservatory, what was called a Winter Garden back then’ (1981: 67). The imprint  of time on the photograph is clear from his description; the material contextualises  the index. 

There is a lot of research on object recognition in images. Vinyal et al.’s (2015) 10 paper shows how complex images are provided a dynamically-generated caption.  The work uses computer vision and natural language processing to form a complete  image description. Their real-world applications haven’t reached the average  consumer. Also, these are only able to articulate what is in the photographic frame. 

its material form. Since the inception of photography, photographic  images have been used in a variety of contexts, and have been presented in  a multitude of ways; the carrier often determines the use the content image  is put to. Whether preserved in photo-albums (arranged thematically or  sequentially), sold as postcards for the curious, or published as  documentary evidence the presentational form of the photograph weighs  heavy on the readings of the image. Presentational forms, in particular,  guide the way in which photographs were used after their inception, and  also the way they were understood. It is, here, important to distinguish  between the carrier and the presentational form: the carrier is always  material, while the presentational form is ideological. The materiality  guides the technical production of the image, and bears the imprint of time  on it. The presentational form reveals the social, political, cultural, and  religious contexts within which the photograph is used. The ideological  and the material may exist in the same form.  

To elaborate on the affect of presentational forms, and to see more clearly  the separation between the carrier and the presentational form, it may be  prudent to inspect the photographic album. The photographic paper is  attached to the album; the album is the presentational form that is imbued  with ideology. Photo albums are unique cultural artefacts in the sense that  they look beyond the content and the creator of the photographs to the  creator of the object in which the photographs reside. Glenn Willumson  (2004), in an essay on the displaced materiality of photographic objects  writes:  

The performance of thumbing through the photographs, selecting  and sequencing, and gluing them into an album breaks the bond of  the materiality of the photograph from its links in commerce and  mass production. In choosing, sequencing, organising and  

captioning the photographs for the album, the person responsible  transforms the meaning of selected images into an intensely  individualistic expression. At the moment of creation, the photo  album is a personal artefact, a record of people and events that are  rich with biography and personal memory. 

Photographic albums are embedded with the traces of their owners and  their practices. The family album, for example, is immersed in ‘mnemonic  frameworks’ (Langford 2001). It is argued that while textual  communication (captions, titles etc.) may aid the reading of photographs,  oral communication plays an important role in the experience of family  albums. The oral stories surrounding the family albums are  ‘performatively intertwined’ with the photographs (Edwards, 2009: 39)  and thus are construed within the etiquettes of social interactions. From  this, it can be observed that presentational forms bring photographs into  use, immersing them within social practices and cultural contexts. The  travel album — a form so ubiquitous in the nineteenth century that they  are sometimes invisible’ (Nordstrom, 2004) — exist as a presentational  form that provides layered subjects for examination. They bear testimony  to the use, and consequently the context, of the photographic images.  Though the examples in this thesis are mostly derived from South Asian  contexts, the nature of photography and photographic presentation is  remarkably similar is other spatial contexts. Nordstrom’s study of the  Tupper scrapbooks, specifically the material aspects of it, highlights the  distinction between the photographs and the album as an object. Made by  hand, the albums are visually unremarkable in comparison to other  surviving examples from the time. What is remarkable, however, is the  precision and detail with which extended captions have been afforded to  each photograph. The presentational form has its own history, its own  purpose, its own intent: the way a series of photographs are collected and  

presented in an album can relate a tale on its own accord. At this point, an  important question arises: how do we record the material information  within the existing structures that digital technology offer us? To  coherently articulate the context and the history of the photograph, we  must, first, be able to create the distinction between the indexical and the  material at the level of its description. The challenge here is not merely to  describe the photograph, but to delineate the separation in a form that is  comprehensible to the computer; this separation must happen at the level  of the metadata of the photograph. 

3.3 Metadata of the Digitised Photograph  

Recent writings on metadata have focused on the more functional aspects  of its creation and usage. While examining learning objects and e-print  communities of practice, Barton, Currier, and Hey (2004: 5-20) point out  the lack of formal investigation of the metadata creation process. While  some collection professionals, specifically those working in libraries have  written about descriptive practices (Caplan 2003, Haynes 2004, Cole and  Foulonneau 2007, Taylor and Joudrey 2008), metadata content standards  (Roe 2005, Hillman 2015, Baca 2006) and the use of metadata in digital  projects (Kenney and Rieger 2000, Stielow 2003, Hughes 2004) the  literature has not yet developed to the point where it affords a complete  picture (Park and Tosaka, 2013: 110). A number of existing surveys  attempt to provide an overview of metadata practices (Ma 2007, Smith Yoshimura 2007, Park and Tosaka 2013). Surveys and analyses of the work  done by cataloguing and metadata professionals, while very important  work in terms of understanding the usage of metadata across institutions,  betrays the problems of such approaches. As Ma (2007) points out, the  limited number of responses (in comparison to the vast amount of  digitised resources), and sometimes, the veracity of the responses prove  

problematic in gaining a firm perspective on the issue. The attribution of  quality metadata is a process that requires significant time and effort. The  results of these surveys are, usually, not very surprising: In Ma’s study of  the use of metadata schemata, the overwhelming use of Machine-Readable  Cataloguing (MARC), a favourite in libraries, followed by Encoded  Archival Description (EAD) is what can be expected considering digitised  verbal resources far out-number non-verbal ones. This is confirmed by  Smith-Yoshimura (2007: 27-29) and Park and Tosaka (2013: 111) who  present remarkably similar results regarding the use of metadata  schemata. For digitised resources on the web, Unqualified Dublin Core  (DC) is the most popular. If the previous studies had only considered  libraries in the United States of America, Palmer, Zavalina, and Mustafoff’s  study (2007) attempts to widen the research lens to include both research  and non-research institutions. They note an overwhelming presence of  Dublin Core, with almost half or more of the digital collections using it  alone or in combination with other schemata. The criteria in selecting  specific metadata schemata are derived from collection-specific  considerations of the type of resources, the nature of the collections, and  the needs of primary users and communities. Existing technological  infrastructure, encompassing digital collection or asset management  software, archival management software, institutional repository software,  integrated library systems, and union catalogs also greatly affect the  selection process. The skills and the knowledge of metadata professionals  and the expertise of staff also are significant factors in understanding  current practices in the use of metadata schemata and controlled  vocabularies for subject access across digital repositories and collections  (Ma 2007: 21-28; Park and Tosaka 2013: 113).  

Surveys carried out mostly within archival institutions and digital libraries  do not contend with the very large number of digital collections produced  

by individuals who are untrained in preservation methods. Melissa Terras  (2010) refers to the growing trend in the creation of amateur online  museums, archives, and collections, as an example of how individual  endeavour may influence traditional memory institutions in creating  useful, interesting repositories. These collections, perhaps not always built  along the best practices guidelines, still provide a valuable resource when  inspecting our cultural heritage. The digitisation of private collections are 11 increasingly becoming more popular. The experience of metadata lies not  with the creator, but with the user of the digital resource. Considerations  on metadata usage, then, needs also to be made from the perspective of the  user who may not be well versed with the protocols and vocabularies  practiced within the institutions. Kathleen Fear’s study, an excellent  investigation of metadata from the user’s perspective, discusses the value  of Dublin Core in digitised image collections, in an attempt to understand  the use of metadata. The study, conducted through surveys, focus groups,  and search and usability testing, tries to identify the nature of information  that non-expert users rely on when searching for images and to locate the  vocabularies that best express that information. The study of metadata  then has to contend not only with the semantics of expression (with  regards to interoperability), but also with a certain lucidity and familiarity  that may engage the user (Fear, 2010).  

Visual resources are inherently different from textual ones. Their needs,  the modes of their description, and the way they need to be conceptualised  require a different standard to the long established (and very popular)  MARC format. To describe the complexity of visual resources Categories  for the Description of Works of Art (CDWA) affords a comprehensive  

With the advancement of technology, it has become easy to commit analog 11 material to the digital space — specifically private collections. These, when  committed to the digital space, might not necessarily adhere to the guidelines set by  institutions, but provide resources that can be explored for both scholarly and  private pursuits.  

guideline and framework. It has 532 categories and subcategories to 12 include every possible aspect of a visual object. The exhaustive nature of  CDWA requires a careful identification of what needs to be described. In  other words, collections built using CDWA display the intent of the curator  at the very outset. The inherent difference between visual objects and  printed verbal resources is that the data related to visual objects, like  photographs, is notoriously difficult to articulate and often contradictory.  A work of art, like a painting, might have been attributed to one painter at  a certain point in time, and another at a later date; it may have multiple  titles or multiple dates assigned to it. CDWA, built specifically for art  objects, was designed to accommodate these issues (Baca 2002: 3).  

The curation of historical photographs prove a difficult challenge in this  respect. In a pragmatic or even manageable sense, one might think of the  metadata sheet as the space for setting out the physical and verifiable  properties of a digital file. This is certainly so; but the lack of  accompanying information about photographs acquired from private  collections continually presents difficulties. One, particularly, is with the  date of recordings: even if some collectors were assiduous providing  details of dates and places, others were content with the slightest of  notings on the physical carriers. Of course in their original location, it was  almost invariably the collector’s memory that would supply a whole range  of details, but in the passage from collection to digital repository, the lack  of recorded information results in irretrievable gaps. The problem is  compounded when we consider that some of these markings on physical  carriers might have been misremembered or misattributed: questions on  the veracity of information stored within metadata is substantial.  

 For the enthusiast (as opposed to the expert) there exists a CDWA Lite schema 12 with 19 categories (there are further sub-categories for these) to describe a work of  art (‘CDWA Lite: Specification for an XML Schema for Contributing Records via the  OAI Harvesting Protocol’). 

Knowledge of the object, then, becomes paramount in our attempts to  create this envelope of tags and markers that we call metadata. The  process of curation also has to grapple with the circulation of the historical  photograph. The same photograph may appear as an image in the  newspaper, or it may also be used as the symbol of a political movement,  or, perhaps, it is displayed on the walls of an art gallery. The social life of  the historical photograph, then, is determined by the circulation of the  image. The use of the photograph — its ideological presentational form —  determines its context, and in turn renders new layers of meaning to the  same image. How are we, as curators of digital collections, to gain an  understanding of the object if its nature is continuously changing? The  little that we can articulate about the photographic image is derived from  an understanding of the history of the object. How we proceed to classify  the image, place it within numerous other photographs is dependent on  how we recognise the photographer, the period, the photographic plate,  photographic process and the photographic context. The history of the  object is vital in our attempts to place it within the structures of a digital  archive. A problem that is persistent in issues of digital curation of  photographs is related to the provenance of the caption. For the archival  image, the caption may be attributed by the curator at the time it is  committed to the digital archive or it may have been inserted at some  intermediary point in its history. This intermediary caption is of archival  value as it may locate the image within the larger discourse of its history;  the provenance of the caption determines its archival significance. If the  caption, then, is of archival value does it merit a separate level of curation?  

The manner in which we write the metadata — the vocabulary — conveys  our knowledge of the object as series of facts: the name of the  photographer, the date when the photograph was taken etc. The tradition  of equating knowledge with facts has exists from a philosophical and  

scientific perspective that can be dated as far back as Aristotle. This view  was augmented through the renaissance and enlightenment in order to  systemise knowledge. The expression of knowledge becomes fundamental  in our attempts to understand and locate the digital object. Traditionally,  the efforts to represent knowledge were largely seen as an attempt to  manage collections of facts relating to the physical world. The  contemporary interest in ontologies can be seen to originate within this  tradition and can be taken as an extension of this monolithic view of  knowledge. This view on knowledge has been argued over the centuries:  Bacon and Locke can be seen to consider knowledge as a single system of  beliefs to which new concepts are added. This view would be challenged by  Quine who would consider knowledge to be like a ‘field of force’, which  impinged on experience only along the edges (Quine 1963: 42). However,  what can be agreed on, is that a body of formally represented knowledge is  based on a conceptualisation: the objects, concepts, and other entities that  are assumed to exist in some area of interest and the relationships that  hold among them (Genesereth & Nilsson 1987: 9). A conceptualisation is  an abstract, simplified view of the world that we wish to represent for some  purpose. Every knowledge base, knowledge-based system, or knowledge level agent is committed to some conceptualisation, explicitly or implicitly.  We use common ontologies to describe ontological commitments for a set  of agents so that they can communicate about a domain of discourse  without necessarily operating on a globally-shared theory. Pragmatically,  a common ontology defines the vocabulary with which queries and  assertions are exchanged among agents. Ontological commitments are  agreements to use the shared vocabulary in a coherent and consistent  manner. The agents sharing a vocabulary need not share a knowledge  base; each knows things the other does not, and an agent that commits to  an ontology is not required to answer all queries that can be formulated in  the shared vocabulary. Thus, we can assume that ‘concepts’ are the key  

building blocks and that we manipulate these concepts with words.  Ontologies are dependent on human language to represent the world. It is  here that we face the first and perhaps the most significant challenge in  order to achieve a shared understanding of the humanities. 

With reference to ontologies, at the very outset, we are faced with two  distinct issues – a problem of metaphysics and a problem of semiotics. The  philosophical investigation of ontology seeks to find the necessary building  blocks of the world, their properties and their inter-relationships. A  starting point could be found in Brentano’s notion of intentionality and  ‘objects of consciousness’ (Brentano 1973: 127-128). An ontology must  make clear what the nature, necessary conditions, and properties of these  objects could be. Formal ontologies combines this goal with a use of logic  that is intended to ensure rigour and axiomatizability of postulated results.  

The general programme of ontology relies on it being possible to uncover  properties that could not fail to be as they are for the world to be as it is.  Existing ontologies have been concerned with the organisation and the  structuring of human knowledge of reality rather than with reality itself.  However, to engage with an ontology at a level deeper than this – with  specific focus on the conceptual framework – it needs to be  epistemologically adequate. Some form of accepted constraints on  modelling decisions agreement over conceptual ontology construction is  required. The main issue with creating these constraints is, of course, in  defining the required ontological level. Since this level has to include  accounts of basic objects and basic relations independently of our  knowledge of them, it is necessary for the account to define how such  objects and relations may be put together in order to reveal an  understanding of the world. As argued by Heidegger (Heidegger, 1962  1927), and later by Schutz (1966: 82) Wittgenstein (Wittgenstein, 1953),  

and others, the world of human being is essentially committed and inter subjective. That is, the world which human beings have access to is already  organised ontologically in inter-subjective terms of human interest.  Creating a committed view of the world from a ‘God’s eye-view’ neutral  perspective of necessity appears to be extremely difficult.  

 The semiotic problem (Bateman 1993: 5) is derived from a non-theoretical  understanding of language that hinders an appropriate construction of  ontologies. The underlying conception of language is that it places an  emphasis on the world as a source of its decisions concerning ontology  construction without a prior analysis of what is meant by the world. It  compounds the problem by driving attention away from natural language  as it is inadequate and restricted. The relationship between a sign and its  meaning is only arbitrary for the most trivial of possible sign-types – that  between linguistic form and phonetic substance. A semiotically richer view  can capture the fact that more complex signs are strongly and non arbitrarily related to their social purpose (Hodge and Kress 1988: 82).  

With these considerations we can posit that ontology or knowledge  representation is a surrogate standing for the objects and relations outside  in the world. The fidelity of the representation depends on what the  ontology captures from the real thing and what it omits. Perfect fidelity is  impossible. A simplistic view would say that an ontology is a model of the  world which can be used to reason about it. One of the major claims made  in favour of ontologies is that can facilitate the interchange of knowledge  between agents, or the reuse in different systems. However if each  ontology is modelled around an imperfect universe, knowledge sharing  would increase or compound errors which were not visible in the initial  use of the ontology. Again, an ontology is a set of ontological  commitments. The choice of ontology is also a ‘decision about how and  

what to see in the world’ (Davis et al., 1993). This is unavoidable when we  consider that representations are imperfect; however, at the same time,  the purpose-built ontology has its advantages as it focuses on what is  relevant or interesting within the boundaries of the domain. These choices  allow us to cope with the overwhelming complexity and detail of the world.  Consequently, the content of the representation provides a particular  perspective on the world. The way a knowledge representation is conceived  reflects a particular insight or understanding in human reasoning. The  selection of any of the available representational technologies commits one  to the fundamental views on the nature of intelligent reasoning and  consequently very different goals and definitions of successes. An ontology  must allow for computational processing, and consequently issues of  computational efficiency will inevitably arise. Since all ontologies depend  on a propositional view of knowledge in order to begin to be  computationally tractable, already a very restricted view of what it is  possible to represent has arisen. All forms of knowledge representation  including ontologies are both media of expression for human beings and  ways for us to communicate with machines in order to tell them about the  world.  

The criticism levelled at ontologies focuses on the fact that they are  unsuited to the world of applications once they get beyond a certain level  of complexity. While some ontologies are acceptable there is always a  trade-off between expressivity, usability and accuracy. Further arguments  can be made (on a more pragmatic level) about the difficulty of  maintaining ontologies and reify a particular point of view of the domain  knowledge. Ontologies can be seen to be struggling to keep pace with the  dynamic, complex world of knowledge bodies and knowledge-sharing. One  of the most basic issues facing the users and developers of ontologies is its  degree of complexity. Folksonomies are comparatively easier to use and  

maintain while offering a flexible and personalised perspective; however  their use is limited due to two reasons – (a) their quality of concepts  involved does not match that of ontologies and (b) their reliability cannot  be compared to that of an ontology. On the other hand, formal ontologies,  such as the Descriptive Ontology for Linguistic and Cognitive Engineering  (DOLCE), the General Formal Ontology (GFO), the Web Ontology  Language (OWL), or the Resource Description Framework (RDF), require  specialised knowledge to build and use them, and are more challenging to  maintain. They are also more rigid than the ubiquitous folksonomies and  thesauri, and less adaptable to changing applications and user perspectives  (Brewster et al. 2007: 563-568).  

The realisation of the obvious technical challenge of creating both human  readable and machine readable data is heightened when we consider that  over the last few years there has been a steady increase in localised  digitisation projects by individuals often unaffiliated to archival and  preservation institutions. The digitisation and curation of collections  outside the walls of the archival institution bring with it its own set of  challenges. Projects carried out without specific funding and done with  consumer-level instruments, are often discouraged by the archival  community. Katrina Dean (2014: 172), in a defence of traditional archival  practices, contends that the practical problems of creating adequate  descriptives for digital objects for the purposes of creating digital  repositories are intensified by the digital economy. She argues that  ‘shifting content from public domains to commercial ownership, either  through the domination of commercially generated content or the  commercial ownership of and exploitation of third party and user generated content’ reflects poorly on the value of information and  heightens its fragility. The act of publishing archival documents online,  thus bringing them out of a controlled and moderated public sphere and  

into a digital economy brings with it issues of intellectual property, and  consequently ownership of cultural property in general. The level of  descriptive, technical, and administrative metadata required to manage,  preserve, and discover digitised collections generally exceeds the level of  metadata required to manage and make accessible their physical  counterparts. Increased standards of explicit evidential value, and  compliance required to bring archival collections online merely reflect on  the inadequacies of meeting those standards in the traditional archive, in  the first place. Traditional archives, for Dean, are about relationships; for  their evidence and informational value to be fully explored, the objects  must reveal relationships between contexts and records, and among  sources. While in the traditional archive, these relationships were implicit  in the architecture of the buildings, in the storage configurations, within  the old registers and the administrative files, and in the curatorial  knowledge, its digital counterpart is still searching for those connections.  ‘Short of digitising whole collections and transposing these contexts into  metadata, it seems unlikely that collections in their present configurations  will be transmitted into the future knowledge economy’ (172). While Dean  speaks about digital repositories created by institutions and by individuals  in the same breadth, there is a difference: the value of archival objects  digital and curated to preservation standards lies in the expectation that  they would stand the test of time. Projects done in individual capacities,  mostly out of interest or curiosity, have a different approach and a  different value. These projects accumulate private collections, objects that  would otherwise have not entered the archival institutions. Most of these  are freely available and resources that may be explored for research  purposes. These collections often do not meet the required standards of  preservation, or the quality that an institution would demand: their value  lies singularly in their availability.  

The creation of digital collections requires more than the availability of  scanners, cameras, and a knowledge of metadata schemata. The primary  problem of creating digital resources, specifically with historical artefacts,  lies in the process of their curation. Beyond the technical challenges of  metadata attribution, curatorial expertise is required to foster some  understanding of the digital object. This is faced severely in the case of  historical objects, specifically non-textual objects, which require a degree  of curatorial more expertise to interpret. The ubiquitous nature of digital  objects have made metadata a primary concern for all engaged in work in  the digital space. To aid identification of the object, the machine requires  textual markers; this become more necessary for non-textual objects where  the machine cannot adequately look through the content of the object. The  framework of metadata is vital to the creation of digital repositories.  Beyond search and retrieval, metadata guides the conceptual  understanding of an object within the collection.  

22,362 thoughts on “Albums in the Attic