Skip to main content

Museums and the Web

An annual conference exploring the social, cultural, design, technological, economic, and organizational issues of culture, science and heritage on-line.

How to Win an Art Prize Without Making a Thing: QR codes, Classifications and Controversy

Susan Cairns, The University of Newcastle, Australia


Museum classifications play a significant role in the construction of knowledge in the museum, and the process of classification is a vital component of museum work; however, this process is little understood by non-museum professionals. I created Classify Me2.0, a museological work of art exhibited at the Watt Space Gallery in Newcastle, Australia in August 2011 as part of the annual Watt Space Art Prize, to demonstrate these different understandings of museum classification. In Classify Me2.0, I assigned classifications to other works of art on display, without the cooperation or participation of the original artists. I then used Quick Response (QR) codes linked to electronic records to display these classifications proximate to the works themselves, making it possible for visitors with enabled smart phones to consider the classification of object in situ with the objects themselves. A growing number of museums have begun to seize upon the possibilities of QR codes linked to electronic records and multimedia as a way to augment in-gallery information. QR codes seem to provide museums with a simple and accessible way of providing additional in situ information about objects. However, access does not necessarily lead to engagement. This paper, therefore, reflects on the issues that Classify Me2.0 brought to light in the Watt Space Gallery, and considers what they could mean for museums more generally. Although Classify Me2.0 was only a single work of art created by a lone artist, and was not a museum-led exhibition, by utilising museological tropes and exploring museum practices, it exposed complexities in museum practice.

 Keywords: classification, QR codes, complexity, engagement, visibility

In 2010, Australian Senator Guy Barnett launched an inquiry into classification of the visual arts, responding to:   

'community concerns' about a range of issues such as availability of restricted publications, sexualisation of children and concerns about the lack of enforcement of classification laws. The inquiry also recognised that it had been 15 years since the NCS [National Classification Scheme] was introduced and that over that time there [had] been huge changes in technology, with the increasing convergence of media raising the question of whether the classification system was being applied to media in a consistent and comprehensive manner. (Ayres 2011)

In particular, the inquiry responded to controversy following a 2008 exhibition of work by Australian artist Bill Henson, whose photographs featured images of naked children, aged 12 and 13. This motivated a specific consideration into “the differing treatment of the defence of 'artistic merit' in relation to child abuse or child pornography offences contained in Commonwealth and state and territory criminal law.” (Senate Legal and Constitutional Affairs References Committee 2011, 83.)

As media concerns at the time reflected, the inquiry led to fear within the arts sector that:

ARTISTS could be forced to have their work classified before being displayed and some work could be blacklisted despite being legal, if recommendations to a federal inquiry into Australia's film and literature classification scheme are accepted. (Frew 2011) 

Had the recommendations been passed, classification was to be made in very strict terms, with art to be classified for age appropriateness, akin to the classification schemes applied to television, movies and computer games. Of particular concern to the arts community was the possibility that ‘artistic merit’ might no longer prove a defense against such a classification, were the subject matter found to be inappropriate according to new standards.

Although ultimately the inquiry into the National Classification Scheme did not recommend the classification of all works of art by the Classification Board, it brought questions about artistic classification to the attention of the Australian arts community. The 2011 Watt Space Art Prize announced that its annual theme was to be Classify Me, asking artists to respond to the notion of classification in art. The intention of the curator, Amy Hill, was to provoke conversation about artistic censorship and notions of appropriateness in art (whilst drawing attention to the often-arbitrary nature of such judgements). However, although it is not always at the centre of hot political debate, art has always been classified. Classify Me2.0 therefore inverted the art prize theme, exploring the process of museum classification from a museological perspective.

In order to do this, the artist examined all works of art on display during the hanging of the exhibition, and then, based on their name and visual appearance, classified them according to the ICONCLASS classification system, as well as attaching free-form digital tags to each entry. Henri van de Waal developed the ICONCLASS system in response to his “ideas for a systematic overview of subjects, themes and motifs in Western art.” (ICONCLASS 2012, It is a formal museum taxonomy composed of 28,000 hierarchically ordered definitions divided into ten main divisions:

0 · Abstract, Non-representational Art
1 · Religion and Magic
2 · Nature
3 · Human Being, Man in General
4 · Society, Civilization, Culture
5 · Abstract Ideas and Concepts
6 · History
7 · Bible
8 · Literature
9 · Classical Mythology and Ancient History.

This iconographic system is a subject-specific classification system, in which each definition is composed of an alphanumeric code and a description of the subject. (ICONCLASS 2012) It utilises a binary or tree-like hierarchical structure to index, catalogue and describe the subjects of art images, and is used by numerous institutions around the world.

In this case, the artistic interpretation of the ICONCLASS classifications was crudely applied, as the artist had only limited time in which to classify works of art, and had undergone no specific training into the nuance of the system. Further, the artist was not equipped with any information about each work of art except the name of the piece and the artist. Therefore any assessment made about the work and its content could only be done visually and with limited understanding, and thus it could never be an exact replication of the way the system is used within the museum context. However, even with these limitations, the exercise provoked conversation from fellow artists within the local arts community, and exposed a gap in understanding about the way taxonomy has been used within the museum to create relationships between objects. 

Each of the artist’s classifications were then uploaded to a website (, alongside traditional caption information (name, materials, size), creating an elementary digital caption. Classify Me2.0 also utilised a less systematic way of describing and categorising knowledge. Each art object entry was 'tagged' with a number of key words that the artist felt to be relevant in some way. By clicking on a tag or a category link on the Classify Me2.0 website, it was possible for users to draw comparisons among diverse works, drawing the works together to create new meaning. The virtual captions were then linked to QR (quick response) codes, which were displayed proximate to the works of art. Visitors with enabled smart phones could read the artist’s classifications of the works in situ with the works of art themselves. The online caption and classification information was shown only using a simple display, utilising a site, in part due to limitations in the artist’s own technical skill, and in part because short in-gallery display time did not warrant greater investment. However, although Classify Me2.0 utilised technology as its primary means, it was at its heart a conceptual art piece, and the primary questions it raised were not about technology, but about the visibility processes in museum classifications. 

Museum knowledge and the invisibility of classification           

The act of naming objects and relationships has always been crucial in the making and institutionalising of museum knowledge. The overriding purpose of any classification system is to be able to define and describe the similarities and differences between things, in order to simplify the constituent relationships, allowing for general statements to be made about classes of objects. (Sokal 1974, 1116) As Michel Foucault writes of taxanomia, “it concerns a fundamental arrangement of knowledge, which orders the knowledge of beings so as to make it possible to represent them in a system of names.” (Foucault 2002, 172) In the museum this process instantly creates self-limiting definitions of what an object is, what it can be, and how it can relate to another object or an idea. The self-limiting nature of classification is illustrated by Mieke Bal, when she discusses the difficulties with defining a concept of collecting, or in fact, of anything. She writes that “Starting an inquiry with a definition of its subject-matter inevitably leads to frustration: either the definition is too narrow and doesn’t cover the whole range of objects; or it is so broad that a lot of other things are covered by it to… knowledge that begins with definitions is very much like knowledge based on collections and classifications of objects.” (Bal 2004, 86) The definition itself, much like any classification system, proves to be self-limiting. In creating a framework for use, the possibilities of what something can and cannot be are pre-determined in ways that may be useful, but that also problematise the creation of future knowledge.  

The strengths and flaws inherent in classification are particularly pertinent in a museum context, given the institution’s role in rationalising and defining scientific, historical and artistic knowledge. Tony Bennett contends that the birth of the public museum paralleled the emergence of new knowledge across diverse fields including geology, biology, anthropology and history, which utilised the arrangement of objects in order to form a totalizing and historicized order of things and people. (Bennett 2003, 96) Classification was thus a tool for categorising, organising and naming all things, in order to understand the world as a singular and related entity. This drive to rationality became important in the Enlightenment as a critical aspect of the scientific method of knowledge production. However, as Srinivasan et al. contend, “the systematizing of the classification and interpretation of objects is still seen as fundamental to the institutional mission of the museum.”(Srinivasan et al. 2009, 268) Further, historian Ken Arnold argues that “taxonomy and classification have for almost three centuries been the most powerful way by which knowledge has been created and then reinterpreted within museums.” (Arnold, 2006, 243)

However, as the response to Classify Me2.0 from artists whose works were classified in the exhibition indicates, the museum’s role in constructing knowledge through taxonomy are little understood by even an art-making population. During the classification process, the curatorial and installation team in the Gallery (mostly artists whose work was included in the exhibition) began to discuss ideas around classification. However, until I (as artist) started classifying works actively in their presence, many had only ever thought about classification in context of censorship. Few were cognisant of the fact that museum objects are always categorised – whether by genre, materiality, era or subject matter (as in the case of ICONCLASS). They had never considered that classifications create proximal relationships between works that might be unrelated, and give works different meanings from those originally intended. Many if not most of the artists in the exhibition had undertaken university level art or museological courses, and yet few had ever considered the role that judgements made by those in control of a museum’s collection could alter the way art was perceived within a museum context, much less the role that museum classifications might take in knowledge creation more generally. The social construction of museum knowledge was thus seemingly invisible to those exhibiting in the exhibition.

As part of a larger practice, this issue becomes particularly pertinent for museums to be aware of as they upload their collections to the Internet. Museum collections online often utilise existing classifications and categorisations when displaying content, in part because the categorisations make comparable or like works more findable.  Computerised museum databases can make it easier for classification schemes to take on multidimensional perspectives, (Sperberg-McQueen 2004) particularly through utilising faceted classifications. This is useful for museum professionals as a way of gaining different perspectives on a collection. As Anne Fahy comments, in New technologies for museum communication, “computers can provide curators with the opportunity to interrogate their collections’ information in more innovative ways.” (Fahy1995, 85) However, these processes of classification and interrogation are not necessarily obvious or visible to the untrained eye. If even a public educated in art theory and Museology display only limited awareness of the process of classification and its relationship to knowledge creation and rationalisation, as was the case in Classify Me2.0, then a more general public without that training will likely be similarly in the dark about such issues.  

In Digital Futures I: Museum Collections, Digital Technologies, and the Cultural Construction of Knowledge, Fiona Cameron discusses the way digital technologies can offer new ways for understanding museum collections and the relationships between museums and audience. Referring to the work of Marshall McLuhan, Fredric Jameson, Lev Manovich and Jean Francois Lyotard, she explores the synergistic connections between theoretical academic ideas, and the computer ontologies that were being applied to collections information. She writes:

Digital technologies have the potential to rewrite the meaning and significance of collections. By promoting polysemic (plural) models for interpreting collections, for instance, they bring into question absolute claims about meaning, enabling alternative and sometimes conflicting interpretations to appear. (Cameron 2003, 327)

While this is no doubt true, and new technologies do offer new possibilities for interpretation and drawing new meanings out of collections, museums cannot ignore the fact that museum audiences may still remain largely in the dark about the biases and limitations inherent in existing classifications. This is consistent with what Geoffrey C. Bowker and Susan Leigh Star suggest, when they argue that “Remarkably for such a central part of our lives, we stand for the most part in formal ignorance of the social and moral order created by these invisible, potent entities.” (Bowker and Star 2000, 3) Whilst it might be theoretically understood within museums that classifications are not neutral, responses from the artists whose work was being classified in Classify Me2.0 indicated that this process is not well understood. Therefore, although digital technologies do hold potential for opening collections up to new interpretations, it is important to remember that audiences may only rarely realise that the interpretation and classification of collections remains subjective, that museum knowledge is indeed socially constructed, and that there are other ways to understand collections that differ from the institutional perspective.

Two specific examples of classification undertaken in the process of creating Classify Me2.0 underscore this point, and illustrate the complexities possible in museum taxonomies. The first involved the classification of a sculpture by artist Nathan Keogh. The small bronze work of art depicted a pregnant girl, with one foot shod in a ballet shoe (in a nod to Degas), and the other limb a silver prosthesis. The prosthetic limb was, for me, easily the defining characteristic of the piece. The sculpture was therefore classified into the ICONCLASS category 31AA414 paralytic - AA - female human figure, a classification that sits under a parent categorisation of 31AA4 disabilities, deformations and monstrosities; diseases – AA – female human figure. Now, although the broad object class is clearly derived from a particular art historical paradigm, when viewed through contemporary eyes, the grouping together of disability with monstrosity was conceptually problematic. Artists and visitors to Classify Me who learned of this broad classification category often reacted quite negatively to the relating of disability with monstrosity. However, the example illustrated the associations between concepts that classification can make in a way that visitors had not previously considered. The act of classification, therefore, became newly perceived by the artists (and myself) as being socially and ethically charged.

A further element of complexity emerged regarding this work of art, however. Although I, as classifier, responded most emphatically to the prosthetic limb on the dancer, a different person with a different background or interests to my own might have classed the same object differently had they considered the girl’s pregnant belly as the most prominent aspect of the work. Such an interpretation from the person charged with the initial classification would have undoubtedly affected the proximate relationships that the piece would have had in any art collection, and illustrates the limitations of any subjective process. Because knowledge itself is created in the museum context through classification, the process of classifying is never unproblematic. 

The classification of a second piece of art, too, raised questions about the restrictions upon future interpretation that the initial application of taxonomy can have. A large purple painting by artist Elizabeth McGregor titled What do you see in me? was classified as 25G21(+1) fruits (+ plants used symbolically), based on an initial viewing. However, upon later discussion with the artist, it became apparent that although the work of art did depict a giant apple hanging upside down, it was also a painting of female sex organs (hence the ambiguous title of the work). The classification of the object as a fruit was both simultaneously accurate and inaccurate. However, to someone with access only to the classifications and not to the art object itself, this plurality of meanings would be lost. Future interpretations of the work could be limited because of the subject-based classification that was originally assigned.

So why is this important? C.M. Sperberg-McQueen states that:

In practice, most classification schemes intended for general use content themselves with representing something less than a perfect image of the intellectual structure of their subject area and attempt with varying success to limit their theoretical assumptions to those most expected users can be expected to assent to. At the extreme, the assumptions underlying a classification scheme may become effectively invisible and thus no longer subject to challenge or rethinking; for purposes of scholarly work, such invisibility is dangerous and should be avoided. (Sperberg-McQueen 2004)

Similarly, the assumptions that underpin any museum classification scheme are likely to be effectively invisible to anyone not trained in their intricacies. Therefore, although museums are increasingly seeking to become open and inclusive in regards to their collections, in part through uploading them to the Internet, it is important to remember that there remains a general invisibility of process at the very base of the classification process.

Ramesh Srinivasan et al. maintain that traditional museum practices such as classification, publication, collecting practices and exhibition are “so deeply embedded in daily museum work that they are rarely questioned.” (Srinivasan et al. 2009, 268) Since classification and comparisons between like objects and classes of objects remains a core way through which knowledge is constructed in the museum, it therefore remains critical that museums:

consistently explore what is left dark by our current classifications (“other” categories) and design classification systems that do not foreclose on rearrangements suggested by new forms of social and natural knowledge. (Bowker and Star 2000, 321)

It is also important to recognise the inherent lack of transparency within the classification process to those outside it, and educate audiences accordingly.

QR codes and engagement

The second critical aspect of current museological practice that Classify Me2.0 utilised was the use of Quick Response (QR) codes to link to the online classifications of the works of art, in situ with the works themselves. QR codes are a type of matrix barcode that can be read by QR readers or enabled smartphones. Although QR codes were created in Japan in 1994, the widespread adoption of smartphones has seen their use skyrocket in recent years. Museums too have started using QR technologies to provide additional context to artefacts and works of art on display. In Classify Me2.0, the QR codes were used to provide a proximate relationship between the art object and its imposed classification, making possible a direct discussion between the two. There was also a computer terminal available in the Gallery space, for people without smartphones to be able to access the work.

Because Classify Me2.0 was displayed as part of a competitive art prize, Watt Space Director Anne McLaughlin decided that no signage was to be included in the Gallery space telling people what the QR codes were, or how to use them (not wanting to disadvantage other entrants). There was a short blurb included in the exhibition program, but no additional information was made available to visitors to the Gallery. The QR codes, therefore, appeared largely inaccessible to Gallery visitors who were not yet familiar with the technology and how to use it. In fact, over the three weeks that the exhibition was on display, only 13 people (out of around 600 visitors) engaged with the QR codes as a way to view the classifications. The actual visits to the Classify Me2.0 website were significantly higher than this, with around 75 people from the local area visiting and interacting with the site. More than half of those visitors came to the site more than once, and visited an average of 5.4 pages whilst there, at a bounce rate of only 33.33%. Once on the site, visitors engaged with the content. However, very few utilised the QR codes as a way to get there.

Obviously, this is only a single example from a one-off work of art displayed in a gallery space. However, it does suggest that the lack of engagement with the QR codes might have been indicative of problems with engaging with the technology itself, rather than the content behind the technology. This is something that Nina Simon recently addressed on her blog:

the biggest issue with how QR codes are deployed in museums is that there's very little information provided about WHY a visitor would want to scan a given code. There's often an object label, a code, and an unwritten mystery about what you'll get when you scan the code. When I visited one contemporary art museum last year, this mystery took on an almost poetic scale. Sometimes, I'd scan a code and get a 10-min video of the artist working on a piece. The next code would take me to someone's website. There was no consistency and few pointers to let me know what I'd get.

QR codes without context are appealing to two audiences: museum geeks and technology geeks.
(Simon 2011,

Anecdotally, from observing visitors in the Watt Space gallery space, I would agree with Simon’s conclusions. Those visitors who did engage with the QR codes via their mobiles were those who were already in the know about the technology. In museums, there is often a persistent argument that content is king. However, even interesting content is not necessarily enough to entice people to engage with technology if the benefits of doing so are not made clear up front.

While QR codes make it possible for museums to link to infinitely more content that can supply contextual information about collections, audiences need to be supplied with both the means (both technical and informational) and the motivation to actually utilise the technology. In The Museum Experience, John Howard Falk and Lynn Diane Dierking identify that “Visitors must be motivated and engaged if they are to learn and want to return. Motivation and engagement are the basic elements of effective education in all settings.” (Falk and Dierking 1992, x) In Classify Me2.0, the technology might have been a conceptually interesting way of linking classifications with works of art, but there was no real call to action for visitors to engage with the QR codes – particularly if they were not already familiar with, and interested in, the technology.

Conversation and controversy

As a work of art Classify Me2.0 was almost entirely non-visual and viscerally unengaging. However, artistic and meta-museological interventions into the museum and gallery space can have significant impact upon museum practice, and the public understandings of the role of museums in creating knowledge and history. American artist Fred Wilson’s exhibition Mining the Museum at the Maryland Historical Society in Baltimore dug into the collection of the Society, installing objects in a way that brought to light the museum’s complicity in the writing of the history of race.

By excavating the site of institutional racism and retrieving forgotten African-American artifacts and heroes, Fred Wilson's "Mining the Museum" brings to light a history and cultural presence that have been buried beneath layers of neglect and deliberate exclusion… The objects chosen, and the sly twists of Wilson's juxtapositions, call attention to the biases that normally underlie historical exhibitions, thus subverting and shattering them. In this way, "Mining the Museum" takes a place alongside other works that have focused on the meaning of the gallery or museum as both a formal space and an ideological construct. (Wilson and Halle 1993, 170)

Meta-museological works of art can provide significant insight into museum practices, by offering critical new ways to think about the implications of the way museums construct knowledge. However, as works of art they can be difficult for audiences to understand.

At this point, it might be worth mentioning that Classify Me2.0 was actually the winning work of art in the $1500 Watt Space Prize, and that all three judges (prominent artists from around Australia) spoke at length in support of the work at the opening. Their positive reaction, however, ran counter to popular response amongst the local arts community. The arts reviewer of the local newspaper refused to name the artist or the work, and to discuss specific content or the concept of the work. (Private correspondence 2011) Because Classify Me2.0 was not a “thing” – an art object – it was difficult for audiences to accept as the winner of an art prize. However, as a conversation piece designed to shed light on the complex nature of classification and its role in knowledge production in a museum context, Classify Me2.0 was both effective and provocative. As the in-gallery reactions to Classify Me2.0 from other artists and members of the local arts community indicated, few had ever considered the role that museums played in classifying works of art and what that could mean for the creation of meaning more generally.

This lack of common understanding into the role of classification in the making of museum knowledge should have implications for museums as they upload their collections to the Internet. The theoretical underpinning of a museum’s classification system might be well known internally within its own walls. This does not mean, however, that the general public will be aware of, or understand, the consequences of these decisions upon the construction of knowledge either specific to that collection, or more generally. As David Bearman argues, “Collectively museums hold the universe of all objects and ideas and all their relations.” (Bearman 2009, 35) Classification is remains a significant way through which these relationships are made and expressed. However, museum processes for knowledge construction are often invisible to the general public – an issue that becomes even more pertinent as museums make their collections available online via the Internet.


Classify Me2.0 was a single work of art/Museology, and so the conclusions that can be drawn from it are limited. However, the conversations with artists and other museum visitors that arose out of its execution raised real questions about the visibility of processes behind museum classifications. This is a pertinent issue for museums, because classification has played a constructive role in creating museum knowledge, and knowledge more generally. It is therefore important that museums remain cognisant of the lack of awareness of these processes, even as they upload their collections to the Internet and make them more publicly able to be viewed.

Similarly, it is important for museums to consider how to engage audiences in new technologies, when those technologies are being used to deliver museum content. Although the QR codes utilised in Classify Me2.0 were not well sign-posted, making engagement difficult from the start, there was also no real reason for audiences to engage with them, beyond mere curiosity. This, too, should serve as a reminder that supplying interesting content is not necessarily enough. The content needs to be delivered in a way that makes engagement enticing.

Classify Me2.0 was created for Classify Me - the 2011 UoN Services Ltd. Annual Art Prize. Classify Me was on display in Watt Space Gallery, Newcastle 24 August – 11 September 2011.


Arnold, K. Cabinets for the Curious: Looking Back at Early English Museums. Hants & Burlington: Ashgate Publishing, 2006.

Ayres, R. "Arts Law: Report of Senate Inquiry into the National Classification Scheme." Arts Law Centre of Australia,

Bal, M. "Telling Objects: A Narrative Perspective on Collecting." In Grasping the World: The Idea of the Museum, edited by Donald Preziosi and Claire Farago, 84-102. Hants, UK & Burlington, VT: Ashgate Publishing Company, 2004.

Bearman, D. "Representing Museum Knowledge." In Museum Informatics: People, Information, and Technology in Museums, edited by Paul F. Marty and Katherine Jones, 35-57. London, UK & New York, NY.: Routledge, 2009.

Bennett, T. The Birth of the Museum. Edited by T. Bennett, J. Craik, I. Hunter, C. Mercer and D. Williamson, Culture: Policies and Politics. London and New York: Routledge, 1995, 2003.

Bowker, G. C., and S. L. Star. Sorting Things Out: Classification and Its Consequences: MIT Press, 2000

Cameron, F. "Digital Futures I: Museum Collections, Digital Technologies, and the Cultural Construction of Knowledge." Curator: The Museum Journal 46, no. 3 (2003): 325-40.

Falk, J. H., and L. D. Dierking. The Museum Experience. Edited by Mellen Candage. Reprinted 2002. vols. Ann Arbor, MI.: Howells House, 1992.

Fahey, A. "New Technologies for Museum Communication." In Museum, Media, Message, edited by Eilean Hooper-Greenhill, 82-96. London, UK & New York, NY: Routledge, 1995.

Foucault, M. The Order of Things. English edition first published 1970, reprinted 2002 ed. London, UK & New York, NY.: Routledge, 2002.

Frew, W. "Warning: This Art Debate May Contain Adult Themes." The Courier, 18 April 2011 2011.

ICONCLASS. “Home – Iconclass.” Copyright 2012. Accessed online 28 January 2011.

ICONCLASS. “What is Iconclass? – Iconclass.” Copyright 2012. Accessed online 28 January 2011.

ICONCLASS. “History of Iconclass – Iconclass.” Copyright 2012. Accessed online 28 January 2011.

Senate Legal and Constitutional Affairs References Committee. "Review of the National Classification Scheme: Achieving the Right Balance." edited by Senate Legal and Constitutional Affairs References Committee, 216. Canberra, 2011.

Simon, N. “QR Codes and Visitor Motivation: Tell Them What They'll Get with that Shiny Gadget. Museum2.0, 2 August 2011. http;// Accessed online 26 August 2011.

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

Srinivasan, R., R. Boast, J. Furner & K. M. Becvar (2009): Digital Museums and Diverse Cultural Knowledges: Moving Past the Traditional Catalog, The Information Society, 25:4, 265-278

Sperberg-McQueen, C. M. "Classification and Its Structures." In A Companion to Digital Humanities, edited by Susan Schreibman, Ray Siemans and John Unsworth: Blackwell Publishing, 2004.

Wilson, F., and H. Halle. "Mining the Museum." Grand Street 44 (1993): 151-72.