MASTER’S THESIS IN LIBRARIES AND INFORMATION SCIENCE

FACULTY OF LIBRARIANSHIP, INFORMATION EDUCATION AND IT

Examining the limitations of classification and indexing

from a librarian perspective, with a particular view onto
library representations of marginalised communities

Anuka Ramischwili-Schäfer

© Anuka Ramischwili-Schäfer
Partial or full copying and distribution of the material in this thesis without permission is forbidden.
English title: Examining the limitations of classification and indexing from a librarian
perspective, with a particular view onto library representations of
marginalised communities

Author(s): Anuka Ramischwili-Schäfer

Completed: September 2022

Abstract:

This thesis investigates the inconsistencies in classification and indexing

present in the University of the Arts Library catalogue, and examines how
these relate to the perspectives of qualified librarians through the use of
qualitative interviews. We surmise that effective representation of items
relating to marginalised communities, within the language of classification
and indexing, has a great impact on the findability of these items.
Representation of marginalised groups within knowledge institutions
proves to be ever more significant in aiding democracy, dialogue and
equitable access to information; therefore, it is significant to scrutinise,
critique and propose development of this representation.

Keywords:

classification, indexing, librarian, perspective, cataloguing, neoliberalism,

higher education, marginalised communities, representation

2
Table of Contents
Table of Contents 3
1. Introduction 5
1.1 Defining terms 9
1.2 Establishing our research questions 10

2. Literature review 11
2.1 Examination of the librarian profession 12
2.2 Racism, inequality and bias in libraries 14
2.3 Subject access and Library of Congress Subject Headings 15

3. Theoretical framework 19
4. Method 22
4.1 Observing the space and staff 22
4.2 Observations of University of the Arts London library catalogue and libguide 23
4.3 Context and execution of interviews 25
4.4 Interview preparation 25
4.5 Interview guide 27
4.6 Alternative methods 28

5. Contextual chapter on University of the Arts London library 28

6. Results and analysis 30
6.1 Results and analysis - the physical library space and staff engagement 31
6.2 Results and analysis - the UAL library guidelines 32
6.3 Results and analysis of catalogue observations 34
6.4 Interview results and analysis 41

7. Discussion 49
7.1 Relating of findings back to research questions 49
7.2 Limitations 54
7.3 Appraisal of methodology and findings 54
7.4 Scope and potential solutions 55

8. Conclusion 58
References 61
Appendix 67
A: Consent forms 67
B: Interview guide 67
C: Follow up questions 69
D: observation schedule 69
E: observation questions 70

3
4
1. Introduction
Libraries are knowledge institutions housing and providing access to information: within these
settings, librarians act as arbitres of space and information, making key decisions on how to
organise collections in order to make them accessible to the user population. Amongst librarian
tasks are included the processes of indexing and classification, particularly significant in that
they form a large part in making items searchable and findable, both on a library catalogue and
within the physical space. Indexing refers to the representation of books or stock within the
library catalogue through, for example, the referencing of common frameworks such as Library
of Congress Subject Headings (LCSH), the application of keywords and the importation of tables
of contents; meanwhile classification refers to the exact placement of items within the sequence
of a collection, using the Dewey Decimal system or other systems that attribute shelfmarks
according to a thematic order. In order to perform these tasks, librarians must make decisions
around what language to use in order to represent items, how to decide on genre or formal
attributes, how to condense the contents of an item concisely, how items relate to the rest of the
collection, whether or not certain actions might uplift or put items under erasure, they must note
differences and similarities between items, and so forth. Training, experience, time and
availability, job requirements, cultural context, referencing of institutional guidelines, bias, and
many other factors contribute to a librarians’ decisions.

As someone from the Caucasus region, with experience in researching the culture as well as
examining Western representations thereof, the initial spark to examine classification and
indexing on a deeper level sprung from my own particular interest in understanding how
materials relating to the Caucasus, in the context of libraries, sit behind an access barrier of
language and misrepresentation.

The Caucasus as a case study is interesting as it is a region outside of Europe which is

under-represented in the literature and media produced within the Global North, and presents a
set of marginalised ethnic minorities that suffer from miss-representation. In the English
language, descriptive vocabulary relating to the Caucasus is limited, as is common in the
representation of many marginalised groups (Block, 2020). Thus, we can utilise representations
of the Caucasus to understand short-comings within library language and illustrate the
importance of rigorous portrayals of marginalised communities within knowledge institutions.

Throughout my years of study I have faced many problems when constructing search queries and
accessing relevant materials. I have run into problems when conducting basic searches on

5
Figure 1: screenshot of a Google search performed on a cleared browser

information regarding the Caucasus region. A simple Google search, starting off with the phrase
‘is the Caucasus’, as depicted in Figure 1, produces auto-fill suggestions that are indicative of the
misconceptions, in knowledge institutions and wider culture, around the Caucasus. The word
‘Caucasian’ has cycled through different meanings and is now used, in the West, to refer to a
constructed idea of ‘white’ people, or people of European origin - although the history of this
meaning is racist and contested. This means people of the Caucasus are unlikely to find accurate
information when utilising this term. Similarly, the word ‘Georgia’, as in the Caucasus country,
has multiple meanings in English, making it difficult to retrieve accurate results pertaining to this
particular word, when using search engines or catalogues.

I began to realise that there are particular barriers, specific to libraries, in the ways in which
classification and indexing enable as well as hinder access to materials. Using the Caucasus as a
starting point, we will investigate classification and indexing, as it is presented in the University
of the Arts London library catalogue, and examine the ways in which librarians make key
decisions in terms of library practice that impact how users are able to access materials and
traverse the online catalogue. We can draw on the IFLA Working Group standards to understand
the significance of investigating classification and indexing as actions that create access to
library materials for the average user:

Subject access results from subject indexing and classification activities. Subject
indexing is needed because publication titles and tables of contents do not always offer
sufficient information; sometimes they even contain misleading terms. Subject indexing
is the basis for finding relevant information successfully by offering synonym search
terms and standardising natural, ambiguous language. It places the content of resources in
relationship to other resources with similar content. Today’s enormous amount of
published information can be reduced and categorised using one of several available
indexing methods. Thus, the publishing output becomes organised into more manageable

6
 units that are more readable, selectable, and that can be searched more precisely. (Jahns,
2012, p. 2)

University of the Arts London (UAL) is a higher education institution, comprising six colleges
which used to exist as individual polytechnics. Each site houses a respective library space, but all
are tied together - there is a singular library catalogue and students have access to all sites
equally (Smithers & Curtis, 2003). In 2020, in the wake of the death of George Floyd in the US
city of Minneapolis, the UK experienced a wave of solidarity Black Lives Matter protests,
especially in cities with more ethnic diversity, such as London and Bristol. One of the knock-on
effects was the pressure on institutions and organisations to demonstrate their support for
anti-racist efforts: UAL implemented a communications strategy as part of a public relations
campaign (arts.ac.uk), making several statements on Twitter and Instagram. The institution then
faced backlash for purportedly using the moment of these protests to paint itself as anti-racist,
without taking concrete action, with student grassroots campaigns springing up in response
(Marcelline, 2020).

We begin to hone into the research problem: namely, there is an issue of neoliberal and
managerial ideologies leading to the appropriation of anti-racist or inclusivity campaigns to paint
institutions as ‘on-side’, but only for monetary gain or advertising purposes, ignoring the need to
undertake clear anti-racist work in order to serve and represent marginalised communities. Mehra
et. al. (2011) find that libraries are seen to be ‘centres of inclusion’, and that these should reflect
contemporary developments in knowledge production and organisation, such as the necessary
serving of local communities and marginalised groups. If UAL seeks at all to act in service to its
local communities, there is much work to be done, beyond implementing branding campaigns
around diversity and inclusion. Over the years there have been protests over UAL’s hand in
accelerating processes of gentrification, through its partnership with firms like Delancey, to
extend its campuses and house a larger student body of paying customers (James, 2018). For
example, arguably, the University’s role in regenerating the Elephant and Castle area drives up
rent prices and pushes out 150 business owners of ethnic minority backgrounds (James, 2018), in
contradiction to its statements surrounding the support of marginalised communities and students
(arts.ac.uk). Bates and Rowley (2011) note that it was Tony Blair’s New Labour government
who were keen to frame public libraries as arbitres of social inclusion and community support (p.
432); however, this same government introduced steep higher education fee rises and the
dissolution of monetary support for the education of low income students - who are also more
likely to be of ethnic minority backgrounds - leading to great historic changes in equity of
accessibility to educational institutions (Williams, 2017).

As highlighted by Kimberly Black (2018), it is a distinctly crucial time to be working towards

inclusion of marginalised groups in knowledge institutions: ‘[t]he diversity of librarianship and
other information access professions is of particular concern because of the structure and nature
of a twenty-first century world’. As highlighted by the political unrest of 2020, academic
institutions in the Global North are faced with a point of realisation, where it is necessary to

7
uncover the violences enforced on a vast political and cultural scale by dominant powers onto
marginalised groups, which has led to the under-developement of the Global South as well as a
structurally racist class system in the capitalist world. In knowledge institutions these inequalities
materialise in terms of the weighting and prioritisation of cultural materials relating to the white
European standard. As Hope Olson (2001) explains, ‘libraries, like other institutions, reflect the
marginalizations and exclusions of the society they serve’ (p. 639). It becomes clear that there
are strong political structural forces moving against the tides of progressive development - with
repercussions to the understanding of library naming processes and information representation
traditions inherent to librarianship. Meanwhile, ‘full participation in modern life hinges on the
ability of the individual and groups to access and control the means of information production,
communication and transformation’ (Black, 2018). Writing for Chartered Institute of Library and
Information Professionals (CILIP), the professional body for librarians in the UK, John Vincent
(2015) states that libraries are significant in maintaining communities, claiming that,

When people’s daily lives are disrupted, they tend to look for places of sanctuary and
sense. Places in which they can feel safe and from whose vantage point they can
understand the things that are happening around them. [...] Libraries can help us be more
resilient and better-prepared, whether personally or as a community.

Therefore, we establish the necessity to take a closer look at classification and indexing, as
carried out by librarians at the University of the Arts London, against the backdrop of austerity
and neoliberal ideological acceleration in the UK and the involvement of academic institutions
therein. In 2022 we are at an important historical point, where the UK is facing the new
government of Liz Truss and dealing with drastic hikes to the cost of living (Hinsliff, 2022).
Higher education has already seen extreme cuts to spending and subsidies, leading to the
introduction of steep fees and exclusion of working class communities. It is likely that the new
government will implement further measures that will impact education; therefore, at this
moment, it is useful to examine the state of the librarian profession and attempt to measure ways
in which ideology, colonial knowledge structures and other hierarchies exert force on the duties
of librarianship.

Because I work in the arts field I am interested in examining arts libraries. Further, these types of
institutions will hold many ethnographic materials and will thus deal with problems of language
from an already colonial-influenced problematic standpoint (Singer-Baefsky, 2021). We will
undertake an investigation of the University of the Arts London library catalogue, available at
libsearch.arts.ac.uk, and make note of observations: we will utilise queries relating to the
Caucasus in order to focus our research. We will conduct observations of the physical library
space at Camberwell College of Arts, one of the six UAL sites, and observe staff behaviour. We
will then utilise this information to inform a set of interview questions that we will pose to a
handful of librarians who have training and experience in the processes of classification and
indexing. The aim is to paint a picture of the constraints and limitations of language in the
understanding and representation of items, utilising a view onto Caucasus-specific items to

8
illustrate how under-represented communities are at a particular disadvantage and are affected by
these language and classification limitations.

1.1 Defining terms

Before we hone in on the catalogue we will be investigating, let us explore some definitions of
terms that we will be utilising. The American Library Association Glossary of Library and
Information Science explains librarianship as ‘[t]he profession devoted to applying theory and
technology to the creation, selection, organization, [sic] management, preservation,
dissemination, and utilization [sic] of collections of information in all formats’ (Levin-Clark &
Carter, 2013). When we speak of under-represented groups or marginalised communities we
refer to this guide, written by Janice Hamlet (2013) for The SAGE Encyclopedia of
Communication Research Methods:

Underrepresented groups are nondominant groups such as people of color; people with
disabilities; people from a lower socioeconomic status; people who are gay, lesbian,
bisexual, and transgendered [sic]; people of a nondominant religion; and retirees. These
groups signify a distinct area of research, including in communication studies, both as
subjects for investigation and as creators of innovative research perspectives and
methodological approaches about them. Challenges to the traditional research paradigms,
texts, and theories used to explain the experiences of people of color or that regard all
human experiences as universal prompted such methodologies and theoretical
approaches.

In library science, studying under-represented groups lies in contrast to the inherent or

presupposed study of dominant cultures of the Global North or the ‘West’ - i.e. a culture that is
constructed to already privilege and uphold white European values and heteronormativity. We
are conducting this study with a particular focus on classification and indexing, which involves
performing key decisions around categorising, excluding and ranking. Further, we are using as a
case study the representation, within the library catalogue, of peoples and cultures of the
Caucasus region - this area of the world is geographically contested, in that it lies on the borders
of Europe and Asia, comprising the three main states of Georgia, Armenia and Azerbaijan, as
well as a diverse range of smaller ethnic groups. To be realistic with our scope, we will focus on
the more dominant representations of the three main ethno states. We will then expand our view
onto the representation of marginalised communities generally.

In terms of library-specific terminology, let us explore the terms ‘classification’ and ‘indexing’.
As Satija (2000) outlines, classification is the ‘systematic arrangement of catalogue or index
entries, in the manner most useful to those who are seeking [...] a definite piece of information’
(p. 223): this definition also implies the existence of the complexities of classification. We are
using the Caucasus as a case study or sample to explore how classification and indexing pose
limitations, especially for under-represented groups. In this respect, and to put it in Satija’s

9
words, we question who the audiences are who seek information - and how are they aware of this
information availability? Knowlton (2005) raises similar concerns regarding the automatic
imagination by institutions of their average readers and how to reciprocate their needs (p. 124).

For the purposes of our study, classification refers to the shelf placement of an item, where the
ascribed number relates to the content of the item. The catalogue we are investigating uses the
Dewey Decimal System, where a number between 0 and 1000 is attributed to an item, where
each number corresponds to a topic. According to the Guidelines for Subject Access in National
Bibliographies IFLA, the Dewey Decimal Classification system is a ‘general knowledge
organization [sic] tool that is continuously revised to keep pace with knowledge. The system is
further extended through number building, interoperable translations and association with
categorized [sic] content’ (Jahns, 2012, p. 24). To take an example from the University of the
Arts London library catalogue, we can see that the 700s as a general section encompasses art and
fine art: ‘709’ specifically houses fine and decorative arts, and at the university, many books on
Picasso are specifically located at 709.445PIC. The library user faces a shelf of organised books,
broken down within the 700s, rather than all 700s housing a disorganised mass of anything
relating to art.

Meanwhile, indexing is the process of describing the contents of an item, such as through
keywords and subject headings. The Library of Congress Subject Headings (LCSH) is a common
scheme that is referenced to perform indexing - University of the Arts London library also
references LCSH as recommended in local guidelines.

1.2 Establishing our research questions

This brings us to the formation of our research questions. We are interested in the classification
and indexing of items within the University of the Arts London library catalogue. We are
interested in the classification of Caucasus-related items; how these are placed within the
University of the Arts London (UAL) library, what keywords have been assigned to them
through indexing processes and references to the Library of Congress Subject Headings, and the
ability of users to gain access to these items. We utilise the Caucasus example as a case study to
understand how librarians at UAL deal with creating representations of items relating to
marginalised communities on the library catalogue, as well as gaining a clearer understanding of
the consistencies between librarians’ varying ways of classifying and indexing any items.
We explored the ways in which neoliberal ideologies, especially in the UK, put Universities and
higher education institutions in harm’s way. Librarians are seen to be mediators of the
connection between knowledge and communities (Mehra et. al., 2011) . Therefore, the profession
holds a significant amount of responsibility in being knowledgeable on colonial, racist, unequal
structures that enable the erasure of under-represented communities, and how to take these into
consideration in order to make information held within a library and its catalogue accessible.

10
RQ1: To what extent are librarians at the University of the Arts London library consistent in
their methods of classification and indexing?

RQ2: In what ways do librarians at UAL deal with the classification, indexing and
representation of items relating to marginalised communities?

2. Literature review
In order to better contextualise the constraints and limitations of the University of Arts London
library catalogue, as well as critique librarians’ processes of classification and indexing, we will
examine previous research and literature and take into consideration other similar studies around
library catalogue efficacy and classification and indexing. We will continue to work with a
particular angle that takes into account the representation of marginalised communities within
libraries.

Libraries house information and knowledge and make these available to different groups of
people. Public libraries perhaps have the strongest goal to reach as many groups of users as
possible, or at least those represented in their local communities, and provide them with free or
low cost access to books and print items, as well as basic information technology. University
libraries, on the other hand, present a semi-public semi-private set of institutions, depending on
location, and also endeavour to provide access to a selection of materials to a certain user group
based on the subjects of academic study.

For many years conventional thinking has tended to emphasize [sic] the collections of
libraries over their societal or community-based roles. Many perceive libraries as
collections of things (especially books), or tend to place information processes (selecting,
collecting, organizing [sic], preserving, providing access to information) at the center
[sic] of how they define libraries. Yet when David Lankes and colleagues (Lankes,
Silverstein and Nicholson, 2007) describe the library as a ‘facilitator of conversations’
they are bringing forward – and reframing for the digital age – equally important
assumptions underlying the perception of libraries as trusted social institutions that are
vital to democracies, open inquiry and the advancement of knowledge and culture.
(Calhoun, 2014, p. 140)

As outlined by Calhoun (2014), libraries enable conversations, as well as performing key

processes in organising collections and making choices during these processes. We can argue that
in order to provoke rich conversations, libraries also need to be aware of their audiences,
represent these within their stock, and have an awareness in describing respective stock in order
to make it available to corresponding communities and researchers. Calhoun also invokes the
ideas of David Lankes and colleagues, who believe libraries to be vital to democracy and

11
advancements in knowledge and culture: therefore, we can note the significance in libraries
seeking to reduce access barriers to the collections they hold.

Our literature review will take on a thematic structure, clustering papers into themes of the
following: examining the librarian profession; understanding racism, inequality and bias within
library structures; and understanding the limits of the Library of Congress Subject Headings and
subject access. It was not possible to find papers that match exactly to our research and goals,
which indicates a need to perform this research. Simultaneously, this brings up the need to utilise
and review studies under varying themes, to then take them into account collectively as a basis or
contrast for our research.

2.1 Examination of the librarian profession

We have already made reference to the American Library Association definition of librarianship
as a profession that applies ‘theory and technology to the creation, selection, organization, [sic]
management, preservation, dissemination, and utilization [sic] of collections of information in all
formats’ (Levin-Clark & Carter, 2013). Theory is acquired through a librarianship qualification,
usually at graduate level within an institution of higher education, and includes the exploration
and understanding of the use of classification and indexing to organise and represent library
materials.

As Hope Olson (2001) documents, there is a ‘pervasive belief among information scientists that
in order to create an overriding unity in language the diversity and the subjectivity of language
need to be standardized. [sic] [...] Gathering the items depends upon always naming a topic in
the same way- hence the justification of controlled vocabulary: it allows for one-stop shopping’
(p. 640). In order to create a systematised understanding or representation of thousands of library
stock within a library, librarians use categories to represent these; categorising entails loss of
nuance. Olson calls this a universality/diversity binary, where the systems ‘actually hide their
exclusions under the guise of neutrality’ (p. 640). Olson also considers classification,
cataloguing, and library practices as the basis of how a knowledge institution can function;
therefore, non-mainstream information, especially that stemming from or relating to
under-represented groups, is heavily impacted.
Peter Ingwersen (1999), of the Royal School of Library and Information Science in Denmark,
considers libraries and knowledge institutions to ‘contribute to the education, leisure, and thus to
the socialisation processes of our societies’ (p. 11). Therefore, we can perhaps contemplate the
view that librarians constitute key players in the ways our societies access knowledge, imagine
themselves, see themselves represented, and relate to each other: and thus, there is significance
in examining librarians’ processes, biases, and reliance on guidance. As stated by the IFLA
Working Group, ‘[p]roviding universal and all encompassing public access to information is one
of the main activities of librarians’ (Jahns, 2012, p. 2). Librarians are involved in the acquisition
of materials, the handling of purchasing budgets, the presentation of items through shelf

12
placement and displays, cataloguing, classification, indexing, the advising of users, the
subscription to and dissemination of e-resources and databases, and more.

Librarians are also known for their ability to discern reliable sources and use conscious critique
in order to assess information. Many criticisms of library practices come reflexively from within
the library professions - librarians are forced to work according to guidelines, for the sake of
consistency, legibility, ethics, and so forth; therefore, they know the procedures well and are fit to
make criticisms. Steven Knowlton (2005) points out that it is librarians themselves who have
called out Library of Congress Subject Headings for decades, critiquing the structure for holding
and reproducing language biases and actively undermining marginalised communities (p. 125).

Heather Seminelli (2016) seeks to examine the role of the librarian as a professional, and looks
into stereotypes and perceptions of this job in order to understand whether librarians meet the
academic standards to qualify as professionals (pp. 63-69). This paper is useful to us in that it
reviews previous research on the librarian profession and illustrates that there is still a lack in this
area. Further, whilst Seminelli’s aims do not align with our research, the definitions provided by
their article, in particular in relation to the American Librarian Association Glossary, are useful
for us to set some parameters, in terms of defining librarianship and noting the need for further
research into the profession.

There has been other research done into the consistency of library duties - there seems to be an
interest in this type of research particularly in the Global South, perhaps because adherence to
world-wide standards is a newer phenomenon and becomes significant as a marker in the stages
of development of infrastructure and public support systems. Further, there may exist a stronger
awareness of history and colonialism, as a recent or current political and structural force, that
may enhance the amount of reflexivity and evaluation that occurs within different professions at
knowledge institutions.

For example, in their paper on the appraisal of librarians’ skills, researchers Dalhatu U. Jibril,
Godwin Amidu and Sani Suleiman Adamu (2021), each based at one of three respective
Nigerian Universities, utilise a close-ended structured questionnaire to gather data, which is then
analysed using descriptive statistics. Ultimately, the study finds that librarians are mostly aware
and well trained in cataloguing and classifying processes, but struggle to carry out their duties
due to issues with electrical power supply, under-subscription of serials, and with the cost of soft-
and hardware (Jibril et. al., 2021).

There are limitations to the paper in that it is at times poorly written and difficult to follow.
Further, the researchers argue that classification and cataloguing are more significant to analyse
than indexing and abstracting of concepts due to their modern role in online catalogues (Jibril et.
al. 2021, p. 50). There is space for them to discuss cultural and institutional biases that might
impede on librarians’ ability to conduct these processes, but these are not discussed. The
researchers also utilise a definition of cataloguing which leaves room for a lot of interpretation,

13
referring to it as the ‘correct’ representation of a library item. There is no discussion of
objectivity or bias behind determining ‘correctness’ (Jibril et. al., 2021, p. 50).

In discussing their motivations, Jibril et. al. (2021) critique the role of the internet in improving
cataloguing for librarians; however, there is no exploration into the downsides. This perhaps sets
up a study that has a uniformity in belief of how online cataloguing and classification should
occur, giving little space to librarians to answer and provide critique. They study the extent to
which librarians follow their duties, rather than why, which might offer up better solutions. As
the goal is to appraise librarians’ skills, there is a focus on how librarians adhere to conducting
the online work, rather than a greater investigation into perhaps how and why limitations exist.
Perhaps a more ethnographic methodology could have enabled researchers to observe librarians’
daily duties and set out better recommendations. Instead, they deal with quantitative data
analysis.

In reference to the research questions we can perhaps observe a mis-match between the
researchers’ goals and their actual interests; they claim to appraise librarians’ awareness of
classification and cataloguing, but it seems the actual interest is to measure the impact of the
internet on library cataloguing.

The study attempts to fill a gap: previous studies, appraising the skills of librarians in the age of
online cataloguing and classification, do exist, however, however, not in the context of these
three universities. This perspective is interesting in that there is a need to broaden research in the
field of library and information science in the Global South. In that respect it is also useful to
understand limitations lesser known to the Global North, such as power outages which may
detract from a librarians’ time to focus on representing relevant communities through their
library processes.

2.2 Racism, inequality and bias in libraries

With the development of academia after the second half of the 20th century, in particular in
regards to gender studies, Black studies, Cultural studies, race studies and other academic and
cultural phenomena that endeavour to uncover and work against systems of oppression, the
educational field has been called upon to become ever more self-reflective and critical. However,
in order to meet demands on equity and diversity on a systemic and institutional level,
undertaking work in diversity and inclusion can result in tick-box exercises to appear progressive
and on-board - it can become a branding exercise rather than a constructive process towards
inclusion. As Sadler and Bourg (2015) set out, libraries, as part of academic institutions, are not
neutral. In order to respond to calls for inclusivity, many academic institutions, including the
University of the Arts London, partake in ‘decolonising’ exercises. Whilst involved actors may
have worthwhile intentions, many academics and librarians dismiss these efforts as performative
and reductive (Evans, 2020). Whilst there are certain undertakings, such as that of Wilson
(2022), at London School of Economics, which measure the percentage of library stock

14
published in the Global North and attempt to present solutions towards diversification, a lot of
the work of ‘decolonising’ tends towards the superficial. UAL presents an annual ‘training’
where librarians attend the same 2 hour lecture on the importance of inclusion, but there are no
resulting concrete guidelines or recommendations on how to incorporate this ethos into the daily
practise of librarianship (arts.ac.uk). In fact, outside of UAL, there is a lot of material published
on the ways in which libraries reflect the inequalities of society, and plenty of pointed
suggestions on the types of advocacy, research or examination that could be executed in order to
affect change.

In Power to Name, Hope Olson (2001) explores the ways in which members of marginalised
communities may struggle to engage with library materials: ‘library users seeking material on
topics outside of a traditional mainstream will meet with frustration in finding nothing, or they
will find something but miss important relevant materials. Effective searching for marginalized
[sic] topics will require greater ingenuity and serendipity than searching for mainstream topics’
(p. 639). Olson mirrors our hypothesis and experience of using library catalogues and search
engines to find non-mainstream information on Georgia and the Caucasus, proclaiming that
libraries reproduce the same exclusions present in the society that they are a part of. Similarly,
Melissa Adler (2017) pushes this idea of library exclusion even further , asserting that
classification through the Dewey Decimal system concretely follows the lines and borders of
racism that were used in law to segregate people by race in 15th-20th century USA. Adler (2017)
also uses Google as an example of a search engine running on algorithms with embedded racism
- a search for ‘three black teenagers’ results in a set of mug shots, while the same for ‘white
teenagers’ generates happy smiling faces.

As we will explore further on in regards to subject access, there are many documented
short-comings, relating to unequal representations and definitions, within the Library of
Congress Subject Headings, as originally constructed by Charles Cutter (Olson, 2000). LCSH,
referenced in the processes of indexing, goes hand in hand with the Dewey Decimal
Classification, first established by Melvil Dewey in 1876. Olson (2000), as explained by Sadler
and Bourg (2015), notes certain archaic patterns within DDC, such as the classification of
pregnancy as ‘disease’ and lynching under ‘law enforcement’. Adler (2017) points out that
patterns, racialised assumptions and histories of classification run so deep and complex that it
can be incredibly difficult to disentangle this web and identify biases (p. 5). Therefore, we can
argue for the continued vigorous critique of historical library systems still in use today.

2.3 Subject access and Library of Congress Subject Headings

The Library of Congress Subject Headings (LCSH) were first designed, on the basis of Charles
Cutter’s proposals, as a system to organise the materials within the Library of Congress itself:
now, LCSH is a tool of subject access that is used in most US libraries as well as in countless
libraries globally (Chan & Hodges, 2000). A 2016 controversy over the use of the term ‘illegal
alien’ as a subject heading was led by students who called for this loaded term to be replaced

15
with one that did not imply negative connotations relating to undocumented migrants. The
Library of Congress promised to remove the term but failed to do so, which ignited a larger
critique into LCSH and ongoing biases, as explored by Grace Lo (2019). The Library of
Congress makes its records available through WorldCat, used by thousands of libraries to share
records and enable consistency (Lo, 2019). University of the Arts London also utilises WorldCat
to import records, or uses it in the second-hand form when downloading records from the Oasis
system, which also provides LCSH indexing on the items available for purchase. As of March
2022 there are 382,713 authority records - records for standardised terms - amongst LCSH, with
4000 new headings added each year, and the Library of Congress stating that all are under
constant review (Lo, 2019). It is possible to submit a proposal for review of a subject heading,
but the library notes that a certain level of expertise is needed to understand the system, such as
librarianship training. The institution estimates a turn around of 9-10 weeks for each proposal to
undergo approval - case studies suggest this can take up to two years (Lo, 2019).

A 1998 study looked into the accessibility of LCSH, measuring the extent to which librarians and
the public were able to engage with it - specifically because the Library of Congress declares that
its main aims include being approachable to the public. The study, by Drabenstott et. al. (1999),
determined that 52% of reference librarians, 55% of technical services librarians, and 39% of
general adults were able to make sense of the meanings of subject headings and their
subdivisions. We can surmise that there may exist certain limitations in vocabulary, identification
with certain terminology, and confusing subdivisions that lead to barriers in understanding
LCSH. Ultimately, Lo (2019) comes to the conclusion that the most operable method for
progressing Library of Congress Subject Headings would be to support the critique thereof by
librarians, who have the ability to educate patrons, colleagues, academics and other librarians on
the functions and limitations of LCSH. Analogously, there is interest in examining the roles of
librarians in indexing and classification processes, and surveying their perspectives on the
short-comings - or strengths - of LCSH.

There are varying studies that attempt to tackle parts or abstractions of the Library of Congress
Subject Headings or deal with subject access in regards to a distinctive library catalogue. Marcia
Bates (1986) provides an explanatory overview of previous research into subject access. Bates
makes note of one study that concludes that due to low use by students of subject headings it
would save time to get rid of these altogether - rather than further investigating why use may be
low, eg. a lack of detail. It is interesting for us to see how librarians view the Library of Congress
Subject Headings and whether they are critical of and interested in developing these.

Bates (1986) mentions a second study by The Council on Library Resources, which found that in
the 1980s, 59% of catalogue access was through subject access, meaning users identified a
particular subject on the catalogue, selecting this, leading them to the relevant listed subject
materials. Bates explores further studies that critique the LCSH during a time that saw a shift
from traditional card to online catalogues: two studies by Kaske and Sanders and Larson and
Graham illustrate further that subject headings are incredibly useful and used if signposted well,

16
and that users’ general criticisms are based around the need for specificity and more in depth
keywords. As we will explore in the Results section, the UAL catalogue, while offering subject
guides pertaining to University courses, does not effectively signpost subject headings or specific
access to materials via subject headings.

In 2012, members of the International Federation of Library Associations and Institutions

(IFLA), the ‘leading international body representing the interests of library and information
services and their users’, met at the annual IFLA World Library and Information Congress as
part of the working group formed (in 2003) to create guidelines for subject access in national
bibliographies (Jahns, 2012). The working group discusses the collective national bibliographies,
which constitute ‘cumulated records of a nation’s publishing output’, and seeks to provide
guidance on how these can utilise indexing and classification to make their bibliographies
accessible to users.

National bibliographies do not only verify authors, titles and ISBNs, but also select the
number of publications in a specific domain, identify changes in the publishing industry,
and identify prominent topics or language pluralism. National bibliographies provide a
key to a country’s publication landscape, for example in science or in literature. Finally,
we can even learn about the importance of books and other media within a society.
(Jahns, 2012, p. 1)

For the sake of consistency, it is highly useful that bodies such as the IFLA exist and meet
annually. However, we are also reminded of Hope Olson’s (2001) critique on controlled
vocabularies and naming information practices. The library scholar utilises the term ‘naming
information’ for all librarian methods used to represent documents within a library and
catalogue, and goes on to critique ‘controlled vocabulary’ - the creation of a universal language
for representing documents. Whilst controlled vocabulary enables consistency between and
within libraries, and allows a user to traverse different libraries with the same language, Olson
(2001) argues that this also creates a false reductive binary between universality and diversity,
losing words to refer to marginalised communities and information outside the mainstream. An
accepted universality also creates the presumption that universal language is inherently objective,
which could lead to complacency, where library workers simply go through the motions without
much awareness. Olson’s article is also useful in its critique of naming information as they focus
heavily on Dewey Decimal System and Library of Congress Subject Headings, which are both in
use at the University of the Arts London library.

Certain limitations within the Library of Congress exist, ironically, within attempts to expand the
subject headings. For example, there is a newer category that is dedicated to ‘gifted women’;
there is no specific category for ‘gifted men’ - instead, LC assumes that it is the norm to assume
men are those who are gifted (Olson, 2001). Therefore, searching for books about gifted women,
one has to already take certain steps to note the gender category and look for this specification
under women. Olson notes that the ‘women’ subject heading has far more precise particularities

17
than that of ‘men’, where ‘many of these terms draw attention to women as exceptions to a male
norm’. In a similar regard, in order to find information on Black women, a user must search for
such; meanwhile, ‘White women’ is not a subheading under ‘women’ - rather, ‘White’ is again
assumed as the norm (Drabinski, 2012). Therefore, while there are constructive developments,
they still function in a system that prioritises certains normative positions.

The noted developments in LCSH are assessed by Steve Knowlton (2005), in a paper that takes
into consideration the criticisms of Library of Congress Subject Headings, as presented by
Sanford Berman in 1971, and measures the extent to which Library of Congress has accordingly
implemented changes, 34 years on. Knowlton remarks that certain historical facts pertaining to
LCSH still hold true, especially in terms of language:

By utilizing [sic] the language and perspective of a particular group of readers, rather
than seeking a more neutral set of terms, LCSH can make materials hard to find for other
users, stigmatize [sic] certain groups of people with inaccurate or demeaning labels, and
create the impression that certain points of view are normal and others unusual.
(Knowlton, 2005, p. 125)

Knowlton’s paper lists a set of keywords that indicates the relevance of this work to our study,
including the terms Sanford Berman, bias, subject headings, library of congress subject headings.
The paper relates to our research, in that Knowlton conducts an appraisal of the Library of
Congress Subject Headings. However, the research is specific to LCSH, and while it does survey
the opinions of librarians regarding their work practices, it does not measure specific instances of
bias nor does it examine or critique librarians and their practices specifically.

The article concludes that Berman’s recommendations were taken into consideration by the
institution itself, with 39% of the headings he criticised having been altered exactly as he
suggested. 24% of headings were also changed, but not in exact accordance with Berman’s
suggestions (2005). Knowlton is impressed with the findings, concluding that the main
remaining biases in LCSH pertain to religious headings and matters of opinion. Perhaps we are
then spurred to wonder about the position of both Berman and Knowlton - Berman is a radical
American librarian, and Knowlton, also an American librarian, is known for his teachings on
History and African American studies. Whilst they both take radical critical viewpoints, there is
still concern that their assessment of LCSH is America-centric, especially as LCSH is referenced
globally and will more than likely continue to hold bias pertaining to Western or Global North
standpoints.

The discussion of Knowlton’s findings is remarkably small, providing little detail regarding
persisting biases. As part of the process, researchers found the subject headings that Berman had
criticised and then examined whether these had been updated: here, there existed the potential to
perform their own analysis of the other parts of the LCSH. This would be especially pertinent, as
much time had passed since Berman’s demands and language, history and culture would have

18
developed greatly: new limitations would be identifiable. As Grace Lo (2019) states, ‘[b]ecause
language is so fluid, what is accepted as neutral rhetoric one day may be outdated in under a
decade [...] This quandary is particularly tricky with identity terms, since people often differ in
how they view themselves and may seek to reclaim pejorative terms’. Of course, an investigation
into contemporary biases across LCSH would have been incredibly costly and time-consuming.

Further research examines directly the efficacy of LCSH as utilised by certain library catalogues
and is useful as an example for us to foreground our research. Lancaster et. al. take an approach
different to ours, examining subject access at University of Illinois, in that they begin by
contacting professors who regularly compile reading lists on their respective subjects - in order
to collate a bibliography of works that are regarded as obvious references relating to certain
subjects (1991). However, they are unable to maintain contact with said experts and instead defer
to encyclopaedias. They then run subject terms through the catalogue - if the search query
retrieves items on their collated bibliography, the search is successful. There exist certain
limitations: estimating the relevance of an item to a search query involves bias. Further, the users
conducting these searches were actually library and information science students who already
possessed a higher level knowledge of utilising library catalogues. Adding to that, the study does
not take into account the percentage of relevant retrieved results as part of all of the results,
meaning many irrelevant items may be retrieved but not accounted for.
For our purposes, the study is useful in its contextualisation of previous research into how to
improve the capability of library catalogues; this includes the introduction of improved subject
headings, the extension of bibliographic records, the adding of searchable elements, the use of
search aids, and systems that limit searches and improve precision.

Taking into consideration this LCSH evaluation, we return to Olson (2001), who ultimately
determines that whilst there are many problems with cataloguing and classifications, library
structures are malleable and open to change - conducting further research into libraries and
attempting to uncover solutions will create change.

3. Theoretical framework
As already exemplified, the research we conduct will have a heavy basis in the work performed
by Hope Olson, drawing on the scholar in order to further understand hierarchies in knowledge
organisation and consequences for information access. Olson (2001) illustrates that there is a
large body of research that demonstrates the vast amount of biases presented within ‘naming
information for retrieval’ (p. 639). Note that Olson uses the term ‘naming information’ to refer to
the ‘creation of document representations’, which is relevant for us as regards classification and
indexing. Emily Drabinski (2012) makes similar - arguably more extreme - proclamations,
stating that library classification uses the same language of the powers that be, reflecting and
reproducing hierarchies. Therefore, there exists also a gap to be filled in the necessity to

19
understand the problems of naming, and the need to shed light on possible solutions. I will, with
this investigation, focus on items relating to the South Caucasus, in order to narrow my research
and deal with the three major ethno-states of Georgia, Armenia and Azerbaijan. However, more
precisely, as I go on to explain, this point of view will act as a lens or case-study in order to
inform a wider investigation into language and libraries.

We can use the example of the multiplicity of the word ‘Caucasian’ to surmise that there will be
many further issues relating to homonymic, metonymic, polysemous and loaded terms. This will
ring particularly true for marginalised communities that are more likely to be represented through
less hegemonic vocabularies. As Hope Olson (2001) outlines, ‘libraries, like other institutions,
reflect the marginalisations and exclusions of the society they serve’, and therefore ‘library users
seeking material on topics outside of a traditional mainstream will meet with frustration in
finding nothing, or they will find something but miss important relevant materials’ (p. 639).
Critiquing librarianship from within is significant in understanding how the ‘scientist
simultaneously constructs and contains nature’ (Olson, 2001, p. 639).

Race is constructed and understood differently all over the world, but through this academic
investigation we are heavily affected by the Western standpoint and this is what we need to pull
into focus - with a deep awareness that we are investigating the limitations of a type of
professional library-specific language that describes communities ‘under erasure’. We must also
take into account that we are conducting this investigation in English, for a Swedish university
degree, with our attention pointed at a British University; there are multiple layers at play that we
do not have the scope to fully explore, but noting them will aid us in establishing our standpoint
and limitations. Britain has one of the strongest colonial histories of any contemporary state.
Accordingly, there is a large and influential population of people with migration backgrounds
and ancestral histories that were affected by racial oppression of the British Empire - these power
imbalances exist into the present day as structural racism that works to exclude on the basis of
race as a cog in the capitalist class-based system (King, 1990). Universities and other educational
institutions function as part of this system, as does the English language; we bump into language
limitations, related to coloniality or white-washed uses, when performing research within a
library context. Drawing on Hope Olson’s research in the field of libraries and inclusion we will
explore why it is significant for library materials to be accessed by many different groups of
people, not just an imagined generic white European public.

My hypothesis is that users face difficulties with language when constructing search queries
relating to the Caucasus - and these difficulties are also heavily reinforced in the classification
and indexing process. Similarly, users face difficulties when searching for materials relating to
many other marginalised communities - there are many documented examples, such as the
multiplicitous and colonial word “Indian”, which can refer to people from South Asia, the
Caribbean as well as North America, including with derogatory intention or ignorance (Patel,
2016).

20
 In earlier research (Olson, 1994) I documented a pervasive belief among information
scientists that in order to create an overriding unity in language the diversity and the
subjectivity of language need to be standardized [sic]. Librarians call such a constructed
universal language a controlled vocabulary. To achieve subject access, representations of
documents having the same or a similar subject are gathered within the context of a
catalog [sic] or index. Gathering the items depends upon always naming a topic in the
same way- hence the justification of controlled vocabulary: it allows for one-stop
shopping. Additionally, using the same system across libraries is economical. For
example, the Library of Congress (LC) has sold standardized catalog [sic] records since
1901, saving cataloging labor [sic] in thousands of individual libraries. Yet in imposing
controlled vocabulary we construct both a limited system for the representation of
information and a universality/diversity binary opposition. Our systems seem transparent
in Henri Lefvebre's use of the term - they appear unbiased and universally applicable -
but they actually hide their exclusions under the guise of neutrality. Not surprisingly, this
fundamental presumption on which our practice rests disproportionately affects access to
information outside of the cultural mainstream and about groups marginalized [sic] in our
societies. (Olson, 2001, p. 640)

As argued by Olson, controlled vocabulary creates a dichotomy where there is a tug and pull
between universalisation and diversification. Items that relate to dominant cultural phenomena
may be more easily found and accessed through their representation via a standardised language.
However, as our study will examine, what happens where items fall outside the grasps of this
standardisation? To what extent are librarians aware of less-dominant vocabularies? To what
extent do librarians, as knowledge access gatekeepers, take into consideration the meaning of
terms applied through indexing and the definitions or borders of certain shelf sections through
classification?

Olson (2001) also performs a critique of indexing, particularly in relation to the Library of
Congress and its history. The scholar points out that the LCSH are based off of the earlier work
of American librarian Charles Cutter, who worked with an assumption of a singular, general
‘public’ that, according to him, was universally understood as one homogenous entity.

Cutter’s initial premise that the voice of a singular public is dominant invokes the binary
opposition of public/private. As Gayatri Spivak notes, the public sphere is woven of the
elements of the private sphere, and, therefore, the public is a construct of the private - “it
is the weave, or texture, of public activity”. (Olson 2001, p. 642)

Cutter’s view is highly subjective but purports to be scientific; moreover, ‘Cutter’s singular
public is not “all members of the community”. It is a particular part of humanity that shares
cultural, social or political interests’, excluding large sections of society (Olson, 2001, p. 643).
Further, even though Cutter was active in the United States of the 1800s, his work still dominates
library practise in the US today and is a standard in many other countries, if not in the least

21
through the prevailing use of the iteration that is the Library of Congress Subject Headings
(Olson, 2001, p. 640).

To ground the theory of Olson in the current political context of the UK, and particular policies
relating to higher education institutions, we use the lens of neoliberalism: this helps us
understand the restructuring, outsourcing and centralising processes occurring within
Universities.

This policy agenda is openly and explicitly demanding that universities develop specific
capacities in the next generation of workers, such as entrepreneurialism and a competitive
spirit, to reproduce neoliberal capitalist relations of production and an ideological agenda
for and in education. (Maisuria & Cole, 2017, p. 605)
Neoliberal policies are implemented by universities to undertake education as a means of
business, turning students into customers expecting a service for their money. For libraries, this
can mean that tasks previously undertaken collectively are centralised and given to one team. It
can mean the joining together of several groups of workers to form one large team, in the name
of ‘efficiency’, so that ‘superfluous’ staff can be made redundant (Maisuria and Cole, 2017). At
universities, neoliberal ideology often entails the outsourcing of certain staff, such as teaching
aids or cleaners, so that contractual responsibility lies with an outside body, making jobs more
precarious and redundancy more easily carried out. Such large restructuring campaigns can result
in a streamlining that changes the way jobs are fulfilled: arguably, priorities that already lie at the
fringes are more likely to fall to the wayside, meaning we can assume processes that are carried
out insufficiently to represent marginalised communities will become more exclusionary. It is
interesting for us to note what sort of themes emerge from our interviews and to measure how
these lay alongside the consideration of a neoliberal ideology.

4. Method
The core part of our method is to perform interviews and use a qualitative analysis; moreover, to
foreground these interviews, we also conduct three different types of observations, which help
establish gaps or inconsistencies in the ways the UAL library operates and help inform our
interview guide. We notate observations of the physical library space at Camberwell College of
Arts and examine the behaviour of staff; then we mark down observations of the UAL library
catalogue and run some test queries.

4.1 Observing the space and staff

When observing the physical library space at Camberwell College of Arts we make note of the
appearance, layout, and any inconsistencies and patterns. Then we also perform a participant
observation of staff behaviour: this is possible due to my work at the library, which allows me to
act as an observing researcher. I have agreed with managers that I can make note of library

22
duties, and all colleagues are aware of my study. Colleagues are not, however, aware of when
and how I will observe the carrying out of daily tasks: this makes it possible for me to act as a
covert participant as observer, meaning I will continue to work as a library assistant so my
presence does not obstruct the function of the employees and I will be able to make the most
authentic observations possible (Bryman, 2004, p. 446). Appendix D shows the observation
schedule that I utilise to guide my observations of colleague behaviour: these take place between
April and June 2022 and are recorded in a document on the University of Borås Google drive at
the end of my three daily desk shifts (Bryman, 2004, p. 272). Appendix E shows the questions I
pose to myself in order to observe the physical space: observations are noted in the same manner
as they are for Appendix D. We perform coding in order to understand themes amongst the
observations; however, this is not done as thoroughly as perhaps a structured observational
method would require, due to the fact that this observational data is necessary to build our
interviews, which make up the largest part of our data collection (Bryman, 2004, p. 275).

4.2 Observations of University of the Arts London library catalogue

and libguide
In order to ultimately build our interview questions and set these into action, we will notate
observations on the University of the Arts London library catalogue. We seek to understand the
constraints within the catalogue and make note of certain inconsistencies that will have been
generated by librarians, cataloguing teams as well as external actors such as distribution
companies. To make these observations, we will perform a few searches, describe the results and
investigate these further in the Results section. We will impose certain limitations so that the data
we work with is manageable. For example, we will note the number of pages of results, but
specifically observe just the first page - this also maps onto how student patrons are likely to
utilise the free-standing OPAC computers available at each library site, with the expectation to
obtain relevant library items within the first page of a set of results.

We perform searches on https://libsearch.arts.ac.uk/ with short queries such as ‘Caucasus’, ‘Art

from the Caucasus’, ‘Caucasian art’ and ‘Caucasian artists’, etc. As noted previously, when using
the term ‘Caucasian’ the goal is to arrive at a list of results that refer to people of the Caucasus,
rather than people of European descent. We will use this definition to determine relevance of
retrieved results. We take into account the first page of results for each search and compare them
to each other - for example, we compare the results for ‘Caucasus art’ with ‘Caucasian art’ and
determine the differences in relevance.

We will also notate observations on the University of the Arts London libguide, which is
recommended for use by librarians in order to perform, amongst other things, classification and
indexing duties. We will utilise a qualitative data analysis, as recommended by Bryman (p. 557),
in order to review and analyse the libguide, which exists in the public domain as provided by the
University library. Specifically, we will use a thematic analysis to investigate guidelines - and

23
ultimately, when analysing interview results, we will compare respondents’ answers to the set of
guidelines. We utilise the Caucasus as a frame of reference in order to discern specific themes -
for example, we will probe the extent to which the Caucasus, and related topics, are
recommended for classification and indexing in the guides.

24
 4.3 Context and execution of interviews
We will be using a qualitative method in order to explore our research questions and measure the
consistencies and differences between UAL librarians’ techniques of classification and indexing,
awareness and use of guidelines, specific handling of the classification and indexing of items
relating to marginalised communities, as well as observe their perceptions of Library of Congress
subject headings and any linguistic or cultural short-comings affecting classification and
indexing in the librarianship context.

As Bryman (2004) states, quantitative interviews mirror the opinions of the interviewer - by
contrast, we are interested in the particular perspectives of each librarian interviewee, and
therefore utilise a qualitative approach (p. 469). Qualitative data also provides us with richness
and depth: we are not counting, in quantitative terms, the number of times the term ‘Caucasus’
appears as a keyword in the UAL catalogue, nor are we comparing this, in ratio form, to the
prevalence of the correct use of ‘Caucasian’. Such a calculation would be flawed due to the
complexities of the catalogue and the lack of scope to retrieve this data. Instead, we are
interested in how perceptions and understandings of representing marginalised communities may
affect different librarians’ decisions when performing classification and indexing duties
(Bryman, 2004, p. 564).

We will conduct semi-structured interviews in person on site at the University, breaking the
group of five librarians into sets of three in accordance with their availability, so that we still
benefit from the in-depth and dynamic answers that result from a group setting. Interview
transcripts will be saved on the University of Borås Google Drive server, kept confidential in
accordance with consent forms. We will then annotate, code and categorise interview answers
and analyse these, in relation to our research questions.

4.4 Interview preparation

As Bryman relays, ‘[i]n qualitative interviewing, there is much greater interest in the
interviewee’s point of view; in quantitative research, the interview reflects the researcher’s
concerns’ (Bryman, 2004, p. 469). I am interested in the librarians’ points of view because these
have a great impact on the way they perform their duties. Utilising open-ended questions as well
as the group setting will enable interviewees to ‘ramble’ or ‘go off tangent’, which, Bryman
argues, ‘gives insight into what the interviewee sees as relevant and important’ (2004, p. 469).
Establishing librarians’ foci speaks to the biases and values they lean on when making decisions
in classification, indexing and cataloguing. Bryman (2004) suggests that interviewers ask
themselves what is puzzling about a situation and why it is necessary to speak with interview
subjects; from observing the UAL library catalogue functions and using Caucasus items as a case
study, there are inconsistencies in how books about Georgian, Armenian, Azerbaijani and other
Caucasus artists are classified or are assigned keywords in the indexing process (p. 473).
Therefore, it is useful to understand how librarians decide to place items at a certain Dewey

25
shelfmark or assign certain keywords. While writing the interview guidelines and performing
interviews, we must continuously return to our research questions whilst taking into account the
perspectives of the interviewees, resulting in a semi-structured interview with a vigorous set of
results (Bryman, 2004, p. 473).

We perform three sets of interviews, with the intention being to hold them in a group setting in
order to provoke productive conversation between respondents - two interviews are conducted in
groups of two, while the last occurs individually due to time constraints. I had wanted to conduct
one big group interview but this was impossible due to the schedules of respondents.

In this particular research setting we do not have the time to perform an ethnographic study
where we observe librarian activities over a long period of time. We are able to make spatial
observations, make note of staffing duties and behaviours, and make note of in-house guidelines.
Undertaking interviews in a semi-structured fashion allows for a relatively quick process -
including preparation, the interview itself, transcription, coding and analysis - with possibility
and room for varying answers. On the other hand, a structured interview would allow us to
standardise the data: in this instance we are interested in providing space for interviewees’
thoughts and tangents in order to gain more insight into librarian responsibilities and
experiences. We acknowledge that each librarian has their own unique experience, and giving
them the space to involve their perspective in explaining their librarian work allows us to
measure the work they do against their background.

We also need to consider follow-up questions - for example, interviewees may not have thought
much about terms such as ‘Caucasian’ and ‘Georgian’. So we can structure follow-up questions,
building in some contextual information:

Caucasian is currently often used to connote white or European, even though this word
stems from a racist misinterpretation of the original word that refers to the mountain
region on the western edge of the Asian continent. How do you think this double meaning
could create problems both in classification and in the use of library catalogues?

We can use the Caucasus as an example scenario and measure librarians’ responses to dealing
with potentially new subject matter.

In the interview guide we cement - although loosely and with space for changes, as this is indeed
a semi-structured interview - our interview questions. As Professor Graham R. Gibbs (2011)
notes, it is important to build a good rapport with your interviewee so that the interview has a
certain flow to it and there is space for tangents and new directions (Research Interviewing Part
1: Interviews and the Interview Society). We must also make sure they know they can retrieve
consent at any point. According to Bryman’s recommendations, we ensure the interviews happen
in a quiet setting and that we are aware of the interviewee’s workplace situation (2004, p. 472).
We will utilise Google Docs’ ‘voice typing’ tool during interviews, followed by a close listening

26
of recordings and subsequent transcription. Then I re-listen to interviews and correct any
mistakes in the transcript. During the interview process I listen carefully and think of suitable
follow-up questions, as well as give space to librarians to feel able to naturally respond to each
other. Interview transcripts will be stored as Google Docs on the Hogskolan i Borås Drive and all
names will be anonymised. Further, the consent forms give certain leeway: interviewees consent
to the asking of follow-up questions but are made aware that they can withdraw their consent at
any moment (Bryman, 2004, p. 134).

Getting in contact with the final set of respondents took about two months of back and forth
between myself, cataloguing colleagues at LCC and then my colleagues on site at Camberwell. I
also had to liaise with colleagues’ managers in order to establish a useful date and setting for the
interviews. In the lead up to interviews I made sure that participants were aware of who I am and
what my intentions are, and that these interviews make up part of the methodology for my
Masters dissertation at University of Borås. I made sure everyone had my email address and felt
welcome to contact me with any questions.

4.5 Interview guide

The interview guide can be found in the appendix, section B. We refer to Bryman (2004) in order
to prepare this guide:

The researcher has a list of questions or fairly specific topics to be covered, often referred
to as an interview guide, but the interviewee has a great deal of leeway in how to reply.
Questions may not follow on exactly in the way outlined on the schedule. Questions that
are not included in the guide may be asked as the interviewer picks up on things said by
interviewees. But, by and large, all the questions will be asked and a similar wording will
be used from interviewee to interviewee. (p. 479)

The aim is to ask each interviewee the same set of questions, with space to branch out. Once all
interviews have been completed we will code the responses, determine which areas need
development, and then produce a set of follow-up questions in order to fill any gaps. Interview
questions are set up in an intentional order, where early questions are broad and allow
respondents to shape the theme. For example, the first proper question is, ‘At the point of
purchasing items, what does the classification and indexing process look like for you?’ The
question does not ask for opinions on limitations or constraints to these processes. The next
question asks about limitations around classification and indexing, but does not suggest what
these might be, so that the librarian can present the answer from their own perspective.
Therefore, it becomes interesting to examine what is at the forefront of librarians’ concerns when
performing their duties - perhaps they will bring up the representation of marginalised
communities on their own accord, or perhaps they will not, which tells us a lot. Later questions
are more precise, asking specifically about the Caucasus as an example of items relating to
marginalised communities, or utilising ‘Middle East’ as an example to consider for classification

27
and indexing - regarding the latter, the question does not imply that there is anything wrong with
this term, but rather asks the librarians’ opinions. Questions become more pointed, in order to
both investigate librarians’ handling of materials relating to marginalised communities and to fill
any gaps that may have been unanswered through the earlier, broader questions.

4.6 Alternative methods

If there were more time and resources we could delve deeper into the University of the Arts
Library catalogue, utilising quantitative methods to measure differences. By involving a team of
cataloguers, we could draw out a clearer picture of the entire stock, taking note of all items
relating to the Caucasus, and illustrating user statistics by looking at the staff-end of the Koha
open software to see if there might be any correlation between keywords, descriptive titles and
the number of times items have been circulated. If we had the time we could also examine the
Dewey Decimal system, noting down the specification of decimal point for different sections and
comparing circulation history to determine whether there may exist any correlation: for example,
if certain sections are given their own detailed Dewey placement it may be easier to differentiate
the section from the rest of the stock and thus easier to access individual items.

Camberwell College of Arts Library contains certain problematic areas, such as the ‘African Art’
section, where authors are not attributed lettered Dewey places - instead, they all fall together
under the number 709.96, meaning the library patron has to look at the title of each book instead
of relying on a precise spine label. This sort of examination lies outside of our reaches, but could
be a good alternative method, if resources were available. Further, such calculations would still
only comprise a set of circumstantial evidence that may hint at certain correlations but never
point at exact causation.

Overall, performing a scientific analysis of language, utilising quantitative methods, would rely
on immense labour going into defining terms and proving ‘truth’; therefore, it is more productive
to gauge perceptions and opinions.

5. Contextual chapter on University of the Arts

London library
The online library catalogue at the centre of this study is that of the University of the Arts
London, a university comprising six physical campuses - Camberwell College of Arts, Central
Saint Martins, Chelsea College of Arts, London College of Communication, London College of
Fashion, Wimbledon College of Art - formerly a group of unrelated polytechnics, focusing on
the arts and fashion (Smithers & Curtis, 2003). There are six physical libraries tied to each site:
students based at any site have access to all six. The library is not a digital library, although it
does hold an increasing number of ebooks as well as online subscriptions to serials.

28
To give some context regarding the labour hierarchies, employees in grades 3 and 4 are qualified
librarians and thus hold and undertake cataloguing and indexing responsibilities. I contacted an
assistant academic librarian at London College of Communications - this is another of the six
sites at UAL and is known to house the most cataloguing staff. The teams from the six UAL sites
will contact them from across London when we encounter issues on the catalogue. They advised
me that there has been recent restructuring at the university which means that local librarians
perform classification and indexing at the point of purchase of items, and that LCC processes
items upon arrival but does not have much power in regards to classification. Therefore, I
decided to interview local librarians at my work site in Peckham, at Camberwell College of Arts.

In order to understand the context of the University, its staff and its users, it is helpful to examine
certain demographics. According to the UAL public-facing website, the student population is
made up of 33% people with ‘Black, Asian and Ethnic Minority’ backgrounds (arts.ac.uk);
therefore, it makes sense to assume that these students would expect to find reading materials
relating to their cultures available in the library. Further, UAL is situated in London, of which
only 43.4% of its inhabitants identify as White British, compared to 84.8% of England and Wales
(ons.gov.uk). Meanwhile, only about 15% of the University’s staff body comprises people of
colour - there is no information on the exact distribution, other than that there is a disparity in
pay, where white staff earn significantly higher. Staff are also library users, with 3,800 staff out
of 8,400 holding library accounts. Library staff are at the lower grade end, so it is possible that
they comprise a higher level of people of colour; further, qualified librarians are at grades 3 and
4, so slightly higher than the lowest paid, with an entry barrier of librarianship qualification, and
a dominance in the profession by white women. However, this data on librarian demographics is
merely anecdotal - it affects librarian perceptions, as this is the information passed around the
sites, but it is not necessarily in accordance with exact statistics. Library staff receive annual
training on diversity and equality and are involved in initiatives, such as the Decolonise Zine,
that promote an awareness of race and the representation of marginalised communities
(https://decolonisingtheartscurriculum.myblog.arts.ac.uk/). Therefore, we can surmise that the
UAL student body consists of a large population of people of colour, deserving proper
information access; at the same time, diversity is lower amongst staff, yet equality and diversity
are within the periphery of UAL professionals and affect the work they put out.

The UAL library catalogue covers all six colleges and runs on the open source Koha software.
As noted by Steven Knowlton (2005), it is significant to point out that historically open source
software originates from the United States, and thus much of this imported software
automatically references the American Library of Congress Subject Headings (LCSH), resulting
in a certain embedded bias (p. 124).

Critics contend that certain subject headings, particularly those that are used to identify
groups of people, perpetuate “the exclusionary cultural supremacy of the mainstream
patriarchal, Euro-settler culture” (Olson, 2000). In a word, many subject headings exhibit

29
 “bias”: that is, they use language that shows a prejudice in favor [sic] of particular points
of view, and against others. (Knowlton, 2005, p. 124)

The Koha software was developed in New Zealand, another Anglophone country, and also
references the LCSH. Additionally, the library at the University provides academic materials for
students, so the contents of the shelves is often determined by the reading lists constructed by
University tutors. Therefore, there is a larger system that determines who and what is represented
within the library. It is up to librarians to decide how items are represented, as well as decide to
make interventions and influence decisions on purchasing and classification.

Through our interviews with librarians we examine further the use of subject headings, with
UAL Library referencing the LCSH. ‘The Library of Congress is the largest library in the world,
with millions of books, recordings, photographs, newspapers, maps and manuscripts in its
collections. The Library is the main research arm of the U.S. Congress and the home of the U.S.
Copyright Office’ (loc.gov/about/). Therefore it is noteworthy that this UK university library is
potentially affected by any bias that trickles down, perhaps in regard to what gets classified how
and for what reasons. As part of the Literature Review we examine previous research into the
history, strengths and weaknesses of LCSH.

For the library user, the libsearch.arts.ac.uk homepage offers up two main options - a library
search function, and, adjustable through the selection of a button, a search of subscription
journals. We focus on the first option, i.e. items - both physical and in the form of ebooks - that
are housed within the library itself. While there are six colleges and six corresponding libraries,
we refer to the general library in the singular, as the six sites are accessible to all university
students and function under the same single library catalogue. Reflecting the courses that the
university offers, the library focuses on the arts, including ceramics, painting, drawing,
illustration, communication, film, fashion, graphic design, spatial design, architecture, sound art,
fashion and business, curation, animation, and more. When speaking of a physical site we focus
on Camberwell College of Arts Library - the interviews we perform also focus on a group of
librarians based at the Camberwell site.

6. Results and analysis

We will analyse results in comparison with the 2012 IFLA Working Group recommendations, as
well as in conjunction with Hope Olson’s writings and all previously mentioned research into the
librarian profession, racism within library structures, and the limitations of subject access and
Library of Congress Subject Headings. We will also reference the UAL libguides on
classification and cataloguing, and determine the extent to which librarians work in accordance
with these prescriptions. Our results and analysis are presented in tandem in this chapter.

30
 6.1 Results and analysis - the physical library space and staff
engagement
Whilst constructing the methodology of this study, I have examined the employment structure of
library workers at the University of the Arts London Library - this helps pinpoint who is most
useful to interview, how duties are split up, as well as giving some insight into who holds the
power to undertake which tasks.

Whilst our study focuses on librarians and their duties, it is also useful to make note of the
behaviours of shelvers and library assistants (grade 2s), especially in terms of how these interact
with the stock - which, after all, has been organised by librarians. The specific college site we are
examining, Camberwell College of Arts, is attended by 10 to 14 staff on any given day. There is
one grade 2 position that became vacant in 2016 and was never filled; there is a second grade 3
position that has also been empty for 8 years. There are three grade 4 staff members, meaning
these hold line-management duties, but only two of them are qualified librarians: one is in charge
of customer service, the other is in charge of collections. Amongst the group of grade 3 qualified
librarians, two are part-time while one works every day: the three are distributed across customer
service and collections duties and are supposed to perform only one daily information desk duty,
but due to short-staffing this has increased to three or four. It is very possible that other grade 3
librarian duties fall by the wayside in order to accommodate the lack of staff. This would include
the ability of librarians to perform duties such as classification and indexing as thoroughly as
possible.

In terms of classification and the Dewey Decimal System, I predict that interviews will illustrate
certain constraints, in that a physical section on the scale (a segment of numbers) is represented
by a set of words, and these can sometimes be outdated or otherwise controversial. Often it is
about context. Through my observations I have encountered books on gay or queer themes that
were once classified and put into the ‘mental illness’ section of a library sequence - Camberwell
staff have identified a few items classified like this and have worked to correct the mistake. In
this context, it is not the phrase ‘mental illness’ that is problematic, but the naming of gay people
as belonging there. Working at The University of the Arts London Libraries I have come across
Caucasus-related items that are classified in ‘Middle East’ or ‘USSR’.

When observing the physical space, the first thing that stood out was the distribution of time
afforded to each section of the library. One duty undertaken by shelvers is the tidying of the
shelves. There is a timesheet with the instruction to continue shelf-tidying where the previous
shelver has left off, but only after observing the collection and tidying the immediately obvious
messiest sections: the shelver makes note of these outlier sections in a separate column on the
spreadsheet. Examining this document, certain numbers jump out over and over again, such as
the 760s (Illustration) and the 770s (Photography). Notably, there is little specification within the
shelfmarks of the Photography section, with many items ascribed numbers with no or only one
decimal point - this likely leads to mess, necessitating extra attention. Illustration does not appear

31
to have any sub-categories below ‘Illustration’, with all items thrown in under 760 or 760.1,
followed by alphabetisation by authors’ surname: during shelving hours, there is usually always
a shelver in this section.

Relating back to earlier statements regarding the particular placement of Caucasus items, we can
also observe the behaviour of library assistants within the Soviet Union Art section. One of their
daily duties includes running the ‘reservations’ procedure on Koha, printing out a list of items
requested by users across the six sites, and then picking these from the shelves before sending
them out via courier. Scanning the field notes, it becomes evident, thematically, that library
assistants regularly complain that certain reserved items appear on the list continuously, due to
the fact that staff members are unable to locate these. Library assistants complain in particular
about books about very famous artists, such as Picasso, because the respective shelves often hold
hundreds of books that are not specified in any further detail to illustrate differences amongst all
books on that one artist. Searching for a specific book on Picasso is often futile, as the title is
likely to be ‘Picasso’, meaning staff must remove each corresponding book from the shelf and
check the barcode to find the matching item. This prolonged process takes up a lot of time and
detracts from other duties, so library assistants are prone to skip such items. A pattern of similar
complaints emerges, where library assistants criticise sections with little shelfmark specification,
such as the African Art section and the Soviet Union section. There are frequent conversations in
the office about the need to extend specification of shelfmarks, especially for items relating to
marginalised communities; library assistants deal with stock on a daily basis, so at least at
Camberwell, within the two month observation period, they repeatedly voice pronouncements
over the distinct lack of specification for items on Black art, African art, Soviet art, post-Soviet
art, Illustration, Photography, Asian art, and more.

6.2 Results and analysis - the UAL library guidelines

This subchapter scrutinises the UAL libguides for their recommendations on classification and
indexing: it acts as an introduction to the guides, as further examination and comparison with
catalogue records and interviewee responses occurs later on in chapters 6.3 and 6.4.

Figure 2: screenshot of UAL classification libguide menu

The UAL libguide on classification exists within a tab alongside 11 other guides, as illustrated by
Figure 2. There is no singular tab for indexing - instead, this is referred to under ‘Cataloguing
guidance’, and briefly under other tabs. The lack of clearer guidance on indexing is

32
disappointing. Generally, the guidelines are difficult to locate and there is no institutional portal
or database for librarians to locate all the tools necessary for their job roles. There is often talk in
the office over where to find which libguide - although, within the observational time period of
spring 2022, no observations have been made regarding the placement of libguides pertaining to
classification and indexing.

Interestingly, the University guidance specific to the Camberwell site states the following:
‘709.495: Include all countries of former Soviet union and Baltic states’ (arts.ac.libguides.com).
There is no explanation for the rationale. This would designate the grouping together of art books
relating to 13 countries spanning Europe, Eurasia and Central Asia, entailing an inherent political
decision regarding the contemporary commonalities within this vast region. Western countries,
on the other hand, receive specifications; for example, within groupings of French artists there
are breakdowns in terms of the different areas of France.

Further, seeing as some Caucasus items are classified under ‘Middle East’ or ‘Persia’ and not
under ‘Soviet Union’, it seems that there is a lot of interpretation and personal knowledge bias
involved in deciding where an item might go. As illustrated in the previous sub-chapter, the
709.495 section is visually in quite a disarray, with little differentiation or use of artist surname
suffixes for the Dewey number shelfmark. Further, this section uses outdated techniques:

It is customary at Camberwell to print the shelfmark as a number that goes into maximum four
decimal spaces and three letters for the artist’s surname, eg. 709.7351BOU - this lies in contrast
to what the libguide states, namely to ‘[a]void using more than nine digits after the decimal
point’, while always including author or artist’s surname in the suffix (see Figure 3). Therefore, it
seems that Camberwell constructs shorter shelfmarks. Further, the shelfmark of 709.495 diverges
from both of these recommended techniques: there are many instances where items are marked
as 709.495SOVIETUNION or 709.495USSR - physical paper shelfmarks do not always
correspond with the catalogue record, either. Further, while still far from reality, the University
Library has the goal of increasing specificity and findability by always including a lettered suffix
in a shelfmark. However, this particular section consists

Figure 3: screenshot from UAL libguide, exert on suffix guidance

33
of a large group of items classified only as ‘709.495’. This is an inconsistency that takes the
section out of line with the rest of the collection - it is also visibly noticeable, as the section is
often messier and harder for shelvers to tidy.

6.3 Results and analysis of catalogue observations

Items do not exist in their full form on a library catalogue; instead, the catalogue reproduces
short-hand representations of items in order to direct users to the actual book, magazine or DVD.
This representation is created by a librarian, who makes key decisions, which are heavily
influenced by the professional’s perspective, education and biases.

There are disparities between the ways in which items are classified and indexed on
libsearch.arts.ac.uk and the reasons they may or may not be returned as the result of a search
query. This is apparent with many search queries - we can utilise our Caucasus examples as a
case study. Searching for ‘Caucasus’ yields more relevant results than searching for the
descriptive ‘Caucasian’, yet indexing does not seem to take much account of the official library
of congress heading ‘South Caucasus’ or perhaps ‘Georgia’ or ‘Armenia’ or ‘Azerbaijan’.

When utilising the University of the Arts London catalogue and searching for the word
‘Caucasus’ we retrieve two pages of results. The first page appears to offer up a relevant list of
20 items, as pictured in Figure 4 - relevant in that they all reference themes that do pertain to arts
and the Caucasus region.

34
Figure 4: screenshot of libsearch.arts.ac.uk result list for ‘Caucasus’

The catalogue does not function better - retrieve many relevant results - the more search words
the user inputs. It seems to work better with very short, broad searches. There is the option to try
the ‘advanced search’, pictured in Figure 5; however, this is neither displayed well nor does it
function much better than the normal search.

Figure 5: screenshot of libsearch.arts.ac.uk homepage

35
There seems to be a tension between titles and subject terms. It is useful when a word in a query
corresponds directly to the title of a retrieved item. But how likely is a search to return items that
are on the searched subject but do not describe this in their title? In conversations at work I have
noticed that staff will often speak of ‘keywords’ - upon closer inspection, these are not tagged
words embedded within the listing of an item. In reality, ‘keywords’ on the UAL catalogue are
simply words that appear in the title of the item, or, if the item is lucky enough to have a more
detailed listing, in the ‘title notes’, as seen with the item in Figure 6.

Figure 6: screenshot of catalogue record on libsearch.arts.ac.uk

There is a problem of the catalogue retrieving items where the title of the item might be
indicative of the contents but the indexing of the item diverges. For example, if something has
the word ‘Caucasian’ in the title then it will be retrieved if you are searching for ‘Caucasian’,
such as Caucasian Chalk Circle, regardless of any indexing process. If something does not have
that word in the title and is classified as ‘Middle east’ even though it is indeed about the
Caucasus, but your search query is for ‘Caucasian’ (as opposed to our initial search for
‘Caucasus’), then you will not retrieve this item unless the academic support librarian has
undergone some indexing and for instance included the table of contents in the item page, such
as the item in Figure 7.

36
Figure 7: table of contents tab in item record on libsearch.arts.ac.uk

The catalogue search function has two main options: searching the catalogue for in-house items
and searching articles plus for journal subscriptions. We are focusing on the first option. A search
for ‘Caucasus’ produces a total of two pages of results, the first page comprising of the following
titles:

Iran and the Caucasus,

Interfax: Central Asia and Caucasus business weekly,

Languages of the Caucasus,

Caucasus survey,

Folk designs from the Caucasus and for weaving and needlework,

Legends of the Caucasus,

Stars of the Caucasus,

37
 Berg encyclopedia of world dress and fashion. Volume 9, East Europe, Russia and the
Caucasus,

Madder red : a history of luxury and trade ; plant dyes and pigments in world commerce
and art,

The Sochi project : an atlas of war and tourism in the Caucasus,

Berg encyclopedia of world dress and fashion. Volume 9, East Europe, Russia, and the
Caucasus,

Madder red : a history of luxury and trade ; plant dyes and pigments in world commerce
and art,

The Sochi project : an atlas of war and tourism in the Caucasus,

Berg encyclopedia of world dress and fashion. Volume 9, East Europe, Russia, and the
Caucasus,

Kilims : the art of tapestry weaving in Anatolia, the Caucasus and Persia,

Before the revolution: a 1909 recording expedition to the Caucasus and Central Asia by
the Gramophone Company,

The political machine: Assembling sovereignty in the bronze age caucasus,

Nomadic visions : tribal weavings from Persia and the Caucasus,

The Sochi project : an atlas of war and tourism in the Caucasus,

Journalism,

Hostage,

Sumak bags of Northwest Persia & Transcaucasia,

Islamic arms and armor in the Metropolitan Museum of Art

We can make certain observations. Very few of these items are classified under ‘Caucasus’ or
‘South Caucasus’. Instead they are under categories such as ‘Middle East’ or ‘Russia and former
Soviet Republics’. Upon clicking on an item the user is led to the book’s page, which displays
the subject heading, author or authors, publishing details, a tab that specifies the classification or

38
placement on the bookshelf, a second tab ‘title notes’ which is often left empty, and a third for
previous or later editions. The book Stars of the Caucasus actually has a more thorough indexing
- it has been attributed keywords under ‘Embroidery’, and then under the sub-category of
‘Azerbaijan’. It is general knowledge amongst students and staff at UAL library that the majority
of our stock will fall into the 700s of the Dewey Decimal System, as this is where most art items
are situated: then, it is common practice to sub-categorise items by the country or region in
which the artist practices or the country subject of the item. In this list of results the
sub-categorisation by country is less pervasive than one may expect - especially if the user is
familiar with UAL libraries in person and has browsed the shelves in person, which makes items
less findable or distinguishable from the rest of the stock.

We are reminded of the IFLA Working Group guidelines referenced in the introduction section,
which formed part of our reasoning to focus on classification and indexing - these guidelines
state that ‘[s]ubject access results from subject indexing and classification activities’ (Jahns,
2012, p. 2); therefore, indexing and classification are key to a librarians’ duties in making stock
accessible to users. The guidelines illustrate how indexing can solve problems of ambiguity, can
offer up synonyms or similar keywords, and can make links to other relevant materials.
However, the UAL catalogue does not sufficiently utilise indexing to avoid such problems.
Further, if we look at the UAL libguide for library resources and systems, this provides a
document with guidelines for ‘basic book cataloguing’, which relies heavily on importation of
records. Librarians across UAL’s six sites purchase books through Oasis; the London College of
Communication central cataloguing team then creates catalogue records by importing
information from Oasis. A default recommendation in the libguide seems to be to rely heavily on
importation as well as personal bias. For example, there are details on how to structure the title,
ISBN and shelfmark; however, it is up to the cataloguer to decide whether any details on
bibliography or item summary are relevant enough to import into the catalogue record, and they
make this decision based off of the original Oasis record without being recommended to look for
further indications as to subject headings that may have been missed out by Oasis. There is much
bias in determining this relevance, especially as cataloguing staff are not necessarily trained
librarians and cataloguing training is optional not mandatory. There is even a recommendation
that subject headings should be deleted if there is the potential for confused keyword searching -
meanwhile, our catalogue observations show that keywords are lacking and there is a far greater
problem of lacking information rather than overcrowded catalogue records.

While there is an option to ‘browse’ the shelves, providing a picture of other related materials
according to grouped classification, this is only available per listing, as exemplified in Figures 8
and 9.

39
Figure 8: ‘Browse shelf’ option in libsearch.arts.ac.uk

Figure 9: drop down section of visualised ‘browse shelf’ option on libsearch.arts.ac.uk

The user cannot input a shelfmark code in order to traverse a certain area of the library. Instead,
one already has to have in mind a particular item, or find an item on a related theme and
manually click through the ‘browse shelf’ option. This creates further distance between the user
and the classification system. Further, it is possible to click on a subject heading on an item
listing. However, one then faces a long list of items that are not presented in any sort of order.
Their classification shelfmarks are not in numerical order, so it is difficult to gain a sense of
orientation. It is useful that the user has this option, but it does not seem developed enough.
The item Berg encyclopedia of world dress and fashion Volume 9, East Europe, Russia, and the
Caucasus has a far more specific set of keywords, shown in Figure 10.

Figure 10: subject and keyword specification for Berg encyclopedia of world dress and fashion Volume 9,
East Europe, Russia, and the Caucasus

While it is an ebook and therefore does not have a certain place on the shelf, it has been assigned
keywords ‘Clothing and dress’, then ‘Caucasus’, then ‘Encyclopaedias’. Further, the responsible
subject librarian has added six other grouped subjects and sub-categories. Interestingly, two of
these groupings include ‘Russia federation’ whilst another specifies ‘Eastern Europe’. Therefore,

40
this item has more of a chance to be picked up through a search query, while it is simultaneously
marked as belonging in the contested subject area of ‘Eastern Europe’. The space under the tab
‘title notes’ is left blank - which in fact is common for most items on our retrieved list. As an
experienced user of this library I can say that this is disappointing; many of our items will have
sufficient information under title notes, often referencing the blurb of the book or other
information that points to the contents. Perhaps the title notes sections are less developed due to
these items being less popular and belonging to more marginalised groups. These lacking areas
can highlight the disparity in cataloguing that occurs when librarians hold different levels of
expertise regarding different academic topics.

As mentioned by Andrea Cuna and Gabriele Angeli (2020), “[n]ovice users [of library
catalogues] are typically unaware of or unfamiliar with the MAchine-Readable Cataloging
(MARC) structure, subject headings and classification numbers underpinning subject access
points in library catalogues and expect that these tools work exactly the same way as
general-purpose Web search engines do” (p. 506). Typing longer phrased queries into the
libsearch catalogue does not yield good results - the user is often met with the message that ‘no
results were found’. There is also no implied truncation, meaning that misspellings of authors
and titles yield no results and the catalogue does not offer up alternative spellings. There is,
however, explicit truncation: typing the beginning of a vocabulary term results in a drop down of
suggested search terms. However, even if search terms are suggested, it does not guarantee that a
search for these yields any results. This is slightly confusing, as a library user would most likely
assume that any suggestions are given on the basis of items relating to that term would be
available within the stock. We can speculate that items that belong to an under-represented
culture or group may be more likely to be connected to keywords that are not native to the
English language. In our specific investigation we are considering the word Caucasus and
Caucasian - in a larger study specific to the Caucasus region we may also deal with more specific
ethnic groups such as Kartli, Armenian, Assyrian, Circassian, Adyghe, Chechen, Daghestani, and
so forth. Point being, both English mother-tongue speakers as well as people from within these
marginalised groups may struggle to automatically produce the correct spelling in search queries
related to these non-dominant cultures, and the UAL library catalogue would not offer up
suggestions.

6.4 Interview results and analysis

In order to process our results, we conduct the process of coding, beginning with an annotation
of the interviewee’s answers, writing notes in the margins, and highlighting recurring themes or
sections that stand out, as recommended by Bryman (2004, p. 298). We then pinpoint the most
prevalent themes and display them in a chart, pasting corresponding interviewee answers under
as many categories as necessary. As Bryman sets out, theory on this topic has been developed
mostly by writers involved with grounded theory. However, we will not utilise the latter and
instead proceed with a method of thematic analysis that utilises coding to categorise and
investigate collected data. We will pick a second approach commonly used in grounded theory,

41
in that we code data once over, identify gaps, and accordingly pose a set of follow up questions,
which we code again. Grounded theory would necessitate a repetition of these processes until no
further knowledge could be gleaned (Bryman, 2004, p. 569), but we are constrained by a lack of
time. It is also possible to utilise a method of narrative analysis for qualitative data (Bryman,
2004, p. 577); however, we did not conduct an experiment with our interviewees where we could
have measured their change of perceptions over time, so this is not a suitable method.

It is this flexibility—the fact that [thematic analysis] can be deployed in such different
contexts—that probably accounts for its popularity, in spite of the absence of a great deal
of codification of its core procedures. (Bryman, 2004, p. 581)

Following annotation of our transcripts, we begin to mark words that stand out - actions and
nouns, for example, such as ‘feminism’, ‘consult’, ‘decision’, ‘inexact’, ‘problematic language’,
‘classification by country’, ‘Global North’, ‘perspective’, ‘publishers’, ‘students’. This helps us
determine similarities and differences between the varying interviewees’ answers. We also label
concepts, opinions, processes and anything else we deem relevant: ‘specification’,
‘independence’, ‘Eurocentrism’, ‘decolonising’, ‘outsourcing’, ‘centralised teams’, ‘institutional
bias’, ‘hierarchy’, ‘multiplicity’, ‘point of purchase’, ‘automation’, ‘cataloguing rights’, ‘broad
searches’, ‘Google’. According to Bryman (2004), ‘relevance’ can be determined by something
constituting a recurring theme, through the interviewee voicing its significance, or by an outlier
or surprising statement (p. 396). Other themes or statements that would be relevant are those that
relate back to our research questions and those that relate to the research presented in our
literature review. We then group code items under categories and present them. Bryman also
notes that a downfall of coding can be the loss of narrative and context - as we are working
digitally we have the space to paste out long answers per each theme as necessary.

It is interesting that interview respondents bring up themes around anti-racism and libraries,
sexism and classism in classification processes, the problems of language in indexing, and other
limitations without being prompted. The interview guide is set up to query them on limitations to
indexing and classification, without hinting at specific answers. Therefore we can infer that when
thinking about limitations, rather than just noting time or money constraints, librarians do hold at
the forefront ideas around diversity and inclusion. When queried on limitations to classification,
Librarian 4 answered the following:

Within our library, we use a very old version of Dewey. Lots of problems with Dewey
itself. Lots of the things that I purchase don't even exist within that scheme as subjects.
(Librarian 4, 2022)

After performing interviews I found that each respondent brought up slightly different themes,
and I found it useful to get a wider understanding by asking follow-up questions - as mentioned,
the repetition of seeking further data and coding this again is common in grounded theory,
though we only perform the process a second time (Bryman, 2004, p. 385). As the original

42
consent forms allow for follow-up and contact, I then asked for supplementary written consent
from librarians, via email, and posed this secondary set of questions via email, receiving
individual responses from each librarian. These follow up questions were answered not as a
group and only in written form. For example, I noticed that only one librarian brought up
guidelines during our interview. In my follow-up I asked the group whether they utilise
guidelines when performing classification. I also asked original questions a second time, where I
had initially received insufficient answers. For example, I asked where exactly they might place
a book on art from the Caucasus, because the answers I had received around this topic were quite
vague.

The first theme around classification that emerged at the beginning of the interviews was that of
temporality - librarians apply classification immediately at point of purchase, without having
seen a physical or digital copy of the book. Librarians noted that they often rely on the title of a
book to gauge its contents, and subsequently create a Dewey shelfmark.

[O]ften books have weird titles that are not really what the book is about [...] You think
the title means something. But then you read the blurb and it’s not about that, it’s more
about this other thing (Librarian 1, 2022)

This seems to be quite an inexact process. They criticised modern blurb writing, stating that most
information available online regarding items acts as branding material rather than description,
and that librarians will sometimes simply ‘do their best’ at guessing at the contents of items in
order to decide on classification. It was also noted that librarians will examine previously
classified items on similar themes when purchasing new items - this may aid with inconsistency,
but might also mean that bias in classifying one item will cause bias in the classification of other
items, setting off a knock-on effect.

At UAL library it seems to be most common for a grade 3 or 4 librarian - assistant and academic
support - to decide on the Dewey Decimal shelfmark at the point of purchase, whilst utilising the
Oasis system to place an order.

I would apply classification at the point of purchase, it’s part of the system with Oasis,
you have to put the number in. If it’s something very straight forward that’s already in the
system then I’ll apply the number. Or if it’s one of those things that doesn’t fit in then I
might ask colleagues or go with what Oasis recommends. (Librarian 1, 2022)

So yeah we put the classification when we buy the book, because they won’t let you put it
through without it, and, but I, like [Librarian 2], if it’s quite straightforward we put it with
other books we already have. Sometimes it’s tricky so you have to make a judgement call
and sometimes I also ask others what they would do, and put it where it’s more useful for
a student to find. (Librarian 2, 2022)

43
Librarian 5 provides a critique, pointing out that,

Traditionally classification is set up to be done by the person who's also cataloguing the
book. So you're supposed to have the book in your hands and you're supposed to be able
to look at it and see the content. So sometimes trying to assign a number to a book that I
am only seeing online in the shop, it’s confusing - you end up having to summarise your
idea of the book through reviews. (2022)

Oasis will usually supply a list of recommended placements and many librarians will select the
one that makes most sense in the context of an art school library.

I suppose priority is to place the items where they should be in the Dewey decimal
system, within the library. And adhere to what's already in this place. (Librarian 4, 2022)

There are some discrepancies in respondents’ descriptions on the use of Oasis - whilst all
librarians mention it, Librarian 1 specifically insists that is it is concretely necessary to adhere to
the system’s recommendations:

We classify them in terms of where they go. But we don't we don't choose or make our
own sections, the classification and heading already comes with the purchased book.
(Librarian 1, 2022)

According to librarians, and described in length by Librarian 5, this ease of semi-automation is

rather new - there used to exist a physical central team that conducted more work in
classification. Now, following restructuring, there remains a cataloguing team at the London
College of Communication (LCC) UAL site, but the staff are not likely to classify items - books
will arrive digitally pre-classified by the librarian at point of purchase, receive catalogue entries
created by the LCC team, and be mailed out to the respective UAL college. Further, these staff
are also unlikely to undertake much indexing, which is why so many online catalogue records are
sparse and lacking keywords. Importation of table of contents into the catalogue occurs
occasionally but leaves much to be desired.

Tied into classification upon purchase is the fact that UAL divides its stock up between subject
librarians. According to the libguide,

At UAL each college library uses varying Dewey Decimal Classification (DDC)
schemes; ranging from edition 16 to 23. Although the classification of some subjects are
similar, the majority will differ from site to site. Each college has its own subject
specialisations and therefore some subjects will be more comprehensive than others.
Books from our main supplier arrive with classification (and spine labels - these are
selected at the ordering stage by college librarians). (University of the Arts Library
Resources and Systems, 2022)

44
This means that librarians are able to specialise in the courses they support and work on
improving and maintaining the corresponding literature. The libguide itself states that the Dewey
Decimal System is used inconsistently across the six sites, in order to suit the differences in stock
resulting from the differences in academic subjects. In interviews, librarians noted these
differences and their benefits, but were also concerned that these cause gaps in specificity and
accessibility. After asking about the use of guidelines in a set of follow-up questions, librarians
also mentioned the need for more detailed guidelines to explain and justify the differences
between the classification of stock at the six sites. Librarians’ specialisation is tied to the
University course that uses each section of stock - so the Architecture Subject Librarian works
hand in hand with the Architecture Course. Calhoun (2014) mentions that ‘[a]cademics
demonstrate what they want by what they use’ (p. 114) - we can predict that it is likely librarians
will develop a narrow habit of purchasing stock, according to the knowledge of the academic.
Some librarians appeared to think more independently, while others were reliant on course
leaders to acquire recommended purchases. We did not query any academics so we are not able
to measure the ways in which professors and tutors decide to create reading lists.

A downside to the way stock is split up and handled by respective subject libraries is that it is
easy, even from a user perspective, to notice stock differences. For example, the Camberwell
College of Arts library houses an illustration section. The associated librarians do not break
down the shelfmark of 760 into smaller sections - books are placed in order of author surname
with no further attribution. Meanwhile, all other sections follow the UAL guideline of
specification by artist’s or author’s country of work.

[I]n my sections with graphic novels and illustration we don’t have countries or regions.
We don't have that sort of way of classifying. (Librarian 1, 2022)

[S]ome librarians say they don’t deal with countries at all and that means huge sections
are just very random. Illustration is one big chunk and not by country. Comic books about
the holocaust will be right next to something about superman - 760.8, generic. [...] I
would be expecting the illustration section, for example, to also be subdivided into
countries and places. Because some people would be looking for illustrators from a
particular country or from a particular area of the world, probably an underrepresented
place - those identifiers aren't on there, how are they gonna know. (Librarian 4, 2022)

Librarians pointed out that there are difficulties in determining country placement because many
artists or writers will have a diasporic or migration background, or be affected by war or political
circumstances in ways that will result in involuntary or multiple countries of work.

However, librarians still follow this guideline, enabling the breakdown of the shelfmark into
decimal points, so the item becomes more findable on the shelves for a patron seeking the
specific shelfmark. The illustration section, on the other hand, is pointedly chaotic, making

45
shelving and shelf tidying more difficult, increasing the volume of necessary labour, and creating
a vague disorganised mass of illustration books. Shelvers are often sent to tidy this section - there
will be other reasons for mess, such as interested users, but different Subject Librarians did
single-out the illustration section.

There was some notable disgruntlement over the outsourcing of classification, with the term
appearing multiple times in interviews, where staff pointed out the benefit of being on site daily,
amongst the stock and student users, with an understanding of the art university’s specialisations;
meanwhile, classification is suggested by Oasis, who do not offer contact between purchasing
individuals and any humans who structure the system’s classification process.

Traditionally classification is set up to be done by the person who's also cataloguing the
book. So you're supposed to have the book in your hands and you're supposed to be able
to look at it and see the content. So sometimes trying to assign a number to a book that I
am only seeing online in the shop, it’s confusing - you end up having to summarise your
idea of the book through reviews. (Librarian 5, 2022)

Further, students were active in the discussion over words relating to trans people - items under
contention include those classified under ‘transvestite’, while words like ‘transsexual’ and
‘transgender’ were also mentioned. Librarians did not think that thoughtful decisions on
classification could be left to unqualified staff or even algorithms at Oasis.

Our [classification] is based on very, very old Dewey. [...] One of our trans books still
came under sexual perversions. (Librarian 3, 2022)

A 1992 study by Cousins, as referenced earlier, suggests that deepening the detail of subject
content, such as through importation of tables of contents, improves retrieval performance in
library catalogues (p. 292). This brings up questions regarding UAL’s procedures - librarians
express that they do not perform much indexing or that it is not within their labour division.
Further, our observations into the catalogue illustrate that indexing is presented sparsely and with
much room for development, especially for items with less hegemonic content.

Librarians express further concern over indexing, invoking the theme of automation, pointing out
that indexing is mostly carried out automatically, when records are imported from Oasis, and that
any extra work is done by the cataloguing team at LCC, which does not comprise trained
librarians. Librarians will frequently notice mistakes on records but are not easily able to amend
these. Librarians mentioned that during their education they studied topics such as classification
and indexing, and were made to think critically on knowledge development and dissemination;
therefore, it would make sense for them to also partake in indexing. In regards to the libguide,
information on indexing is more difficult to find than perhaps cataloguing or classification - the
IFLA recommendations note the benefit of institutional in-house manuals on subject heading

46
creation and application (Jahns, 2012, p. 3). Whilst the libguide includes a tab on cataloguing,
information regarding indexing found here is vague, pointing librarians to general Library of
Congress guidelines, basic instructions on creating catalogue records for books, instructions on
using MARC, and site-specific recommendations for London College of Fashion. Instructions
produced in house are clear on the formatting of catalogue records but lacking in terms of subject
heading designations - page 15 of the ‘Basic books (UAL standards)’ (2022) found under the
‘Cataloguing’ tab recommends that librarians link tags to Library of Congress Subject Headings,
but there is no information on how to choose the appropriate heading - arguably out of line with
IFLA recommendations (University of the Arts London Library Learning Resources).

As previously intimated, management has recently enforced a new restriction where most
librarians are no longer able to edit records, meaning they also cannot add detail to item records.
Staff now have to contact colleagues at other sites and are forced to have to justify their decisions
and trawl through bureaucracy to make changes, slowly working upwards through the chain of
command. A common theme in interviews was the criticism of keywords: librarians notice that
many item records are very empty, with no importation of the table of contents, meaning that
Koha relies on words within the item title as keywords. This is consistent with our previous
observations of the catalogue, with findings pointing to empty records. This makes many items
less findable and is especially a problem where the title is not descriptive of contents at all. This
links to our observations of sparse records, as well as our analysis of the libguide which does not
push cataloguers to perform thorough indexing - so it is no surprise that librarians mention this
topic, but it is useful to note that they are aware of it. Further, in line with recent changes,
librarians no longer receive physical books once they arrive on site (sent from the cataloguing
team); instead, items directly enter the stock and become available to users, meaning librarians
feel they do not have the chance to check items and records. Librarians said it is as though items
disappear into the greater system with no accountability.

In terms of honing in on perceptions and dealings with Caucasus items, librarians generally
stated that they would seek help in the office in order to understand how to classify Caucasus
related items.

I would probably [...] speak to other people, and look at what else we have that might be
about something similar. (Librarian 5, 2022)

Sometimes it’s tricky so you have to make a judgement call and sometimes I also ask
others what they would do. (Librarian 1, 2022)

If I am unsure, I’ll talk to people in the office. There was an architecture book on African
architects and I ended up asking Librarian 3 about it. (Librarian 2, 2022)

It was briefly mentioned that UAL library guidelines need to be updated to better reflect
post-Soviet countries and cultures. When queried on their reliance on other staff, librarians

47
answered that they are most likely to confer amongst the Camberwell office. They are not likely
to contact colleagues at London College of Communication’s ASM office. However, guidance
does ask that this is generally done when librarians are unsure of placement or when an item
arrives without classification. Perhaps there are limitations in the fact that there is physical
distance between the six UAL sites: colleagues may tend to speak only to staff they are familiar
with. There is also no staff list with position details, to make it easy to contact relevant people. It
is common that there is confusion in the office with people wondering whom to call. The act of
walking down the corridor to speak to a cataloguing colleague is only possible for librarians
based at LCC. One may also assume that a librarian might seek advice from their direct superior,
who may then communicate directly with ASM staff. However, there is inconsistency in how
different grade 4 line managers willingly liaise with other sites.

If we compare interview results to the Guidelines for Subject Access in National Bibliographies
developed by the IFLA working group, it is noteworthy that UAL does have the previously
examined libguide dedicated to librarian processes with a specific tab for classification. It is
interesting that the general recommendation is to not go beyond 9 digits after the decimal point,
when Camberwell staff specifically avoid higher than 4. This difference is evident between sites,
especially when we process intersite items; whilst such long numbers would help with
specification, Camberwell librarians are very aware of accessibility and print disabilities and
strive to make items as findable as possible for the average user.

We try not to use long numbers here, for accessibility reasons, but some of the other sites
use very, very long numbers and I wouldn't do that. So I wouldn't copy that. LCC and
LCF do that. Yeah, it's just about an item being findable. (Librarian 4, 2022)

Reference to this libguide - or other - was scarce amongst interviewees.

In response to a different question librarians had spoken of the need to perform Google searches
in order to gain a better understanding on topics that are difficult to search on the UAL library
catalogue: in student 1-to-1 sessions librarians will often do a preliminary search on Google to
understand a topic better, find related books, and then cross-reference the UAL library catalogue
to check if there is availability. They did not mention this process when queried on the meaning
of and classification of books containing in the title words such as ‘Caucasus’, ‘Caucasian’,
‘Georgia’, ‘Georgian’, ‘Armenia’ or ‘Middle East’.

Wilson’s (2022) contemporary examination of London School of Economics library collections

references the constraints surrounding publishing, noticing that librarians become increasingly
prone to purchasing from the same places - this has an immediate impact on material relating to
marginalised groups, as this is often printed by ‘smaller, niche publishers’ (p. 228). Relatedly,
interviewees critiqued reliance on large publishing houses, noticing that stock is often Euro- and
US-centric, and that even efforts to diversify ultimately end with the inclusion of marginalised
communities as defined only by and within the Global North.

48
 Doesn't help that a very large amount of our database content comes from America. I'm
not saying that the white European content is much better. (Librarian 4, 2022)

I have a beef against publishers at the moment because pretty much 90%, I would say, of
what's published seems to be American publishers. And the rest is British. (Librarian 1,
2022)

Librarians also noted that dealing with stock will continue to be a process that is in-flux -
language is changing and exactly defining items is not a quantifiable science. At the same time,
UAL holds a collection that is based around the arts - librarians noted that art itself is difficult to
define, especially as practitioners will partake in varying disciplines. Tied into this ongoing
process, interviewees indicated problems with raising concerns and a lack of accountability as
provided by the institution.

Libraries used to be immovable, now we are thinking in terms of being able to change
stuff. But again, if you don't have institutional support affecting those changes locally it
can be really hard. (Librarian 5, 2022)

Librarians are able to pin-point many issues in classification, which they have power over as
long as items are within their section of responsibility. However, issues with indexing are far
more difficult to change, not only due to labour divisions and restrictions on catalogue editing
permissions, but because librarians will face negative consequences in the forms of reprimands
or increased workloads. As Sadler and Bourg (2015) explore, in order for software to be
inclusive, and for search engines to provide a feminist interaction, there needs to be a ‘concern
for pluralism’, where questions over what to include or deem relevant are left open to
interpretation (p. 1). If librarians are not regularly involved in processes of indexing, and these
are hemmed in due to structural employment hierarchies, then who will be responsible for the
appraisal of keywords or depth of catalogue records?

7. Discussion

7.1 Relating of findings back to research questions

RQ1: To what extent are librarians at the University of the Arts London library consistent in
their methods of classification and indexing?

This question is useful in that it is specific about examining UAL, but remains open enough, so
that it is possible to examine the different processes utilised by librarians to carry out
classification and indexing. In terms of themes explored in the introduction, when laying out our
research problem, the open nature of the question enables also the exploration of why there are

49
inconsistencies in classification and indexing. This question of why ultimately points towards an
understanding of how libraries and knowledge institutions might improve their working
environment in order to strengthen practices of librarianship. We can examine the situation at
University of the Arts London and posit this as a case study for more widely critiquing other UK
institutions.

When answering this question, it was noticeably difficult to obtain library science sources on
language. There was more information on marginalised communities and how these are
represented in library spaces and catalogues - Olson, Black, Drabinski all examine racism,
exclusions and inequity in the stacks. Studies critiquing Library of Congress Subject headings
and the Dewey Decimal System are also available, however specificity is lacking. Finding
materials on the Caucasus, explicitly, was basically impossible; therefore, this study hopes to
begin to fill these gaps.

As explored in our theoretical framework, Hope Olson presents the contradiction of

standardisation versus diversity: standardisation, for example in regards to LCSH or the Dewey
Decimal System, can allow for equal ease of access to materials, but erase nuances, especially
when representing items relating to marginalised communities (p. 640). University of the Arts
London comprises six colleges and six physical library sites under one umbrella. We can surmise
that, beyond classification differences noted in the guidelines, there are many more
inconsistencies existing between how each site completes librarian processes. While our study
thoroughly critiques the inconsistencies observed in the library catalogue, Camberwell College
of Arts library space, and those presented within interviewee responses, it also presents future
scope to compare the six sites to each other and argue for differences or standardisation.

One particular librarian’s response stood out; Librarian 4 pointed out that they are not necessarily
intently interested in classification and indexing. Rather, their intention is to make the collection
as accessible as possible to all users. This points to wider questions around scope - classification
and indexing are two complex procedures that make up part of librarian duties, but there is far
more that can be assessed.

Hope Olson (2001) points out that cataloguing belongs to the profession of librarianship and is as
such done privately, away from the public; with the restructuring of librarian positions at UAL,
the distance between the public and the catalogue is even further obscured (p. 643). Librarians
discussed being able to pick up on changes in language and discourse more easily - than perhaps
off-site actors - by interacting with students. For example, students had pointed out that Korean
books were subsumed into the Chinese section, and that there was no appropriate space for the
Korean items; librarians reacted by building a new number into the UAL use of Dewey.

And certainly in terms of classification [...] - only two years ago here at UAL, we noticed
that there was no number for Korean art. There was literally Korean art pulled into
Chinese art. That was what it actually said. In the scheme. And we had students asking,

50
 why are Korean art books at the same place as the Chinese art books. And so we made up
a number for ourselves. But that's what I mean about that lack of nuance built into the
system. If you look at our sections on European or Western art, it’s very detailed. You
know, there'll be sub regions of France art and a number for that. Whereas, outside of
Europe and North America, that granular detail is not there. And I think that feeds into
this issue of who we are and who we are doing things for. Having space for different
things that might have the same terminology but mean different things, should still be
represented well in the catalogue. (Librarian 5, 2022)

Interviewee responses make clear that librarians have grievances with certain facts: respondents
complain about restructuring across the sites leading to the centralising of cataloguing duties;
similarly there are complains that the central cataloguing team does not hold the same
qualifications; respondents are unhappy with the lack of consistent catalogue editing rights
across the staff body; respondents complain about the more recent procedure of importing
classification from Oasis at point of purchase. As we explore in the Results and Analysis chapter,
interviewed librarians make note of the convoluted process of purchasing, classifying, and
sending a book for cataloguing, that has resulted from the restructuring of library jobs. We can
analyse all of these findings through the lens of neoliberalism - centralising, restructuring and
outsourcing are key components to this ideology, which has been the recent dominant political
force in the UK since the appointment of the New Labour government in 1997, but was
implemented first by Margaret Thatcher and is in full swing with the currently in-power
Conservative rule (Maisuria and Cole, 2017). Arguably, the practices mentioned will only
continue to be enforced, with the ongoing rule of the party, which will affect higher education
and therefore university libraries.

RQ2: In what ways do librarians at UAL deal with the classification, indexing and
representation of items relating to marginalised communities?

The manner in which the interview guide corresponds to research question 2 enabled us to
observe whether or not librarians would bring up themes relating to the suppression of
marginalised communities on their own accord; it was possible to establish that such themes are
immediately present in librarians’ decision making when performing classification and indexing.

There were definitive themes around anti-racist thinking and thoughtful decision making
amongst interviewees’ answers. As mentioned in regard to the first research question, librarians
do not uniformly follow the UAL guideline to contact London College of Communications
cataloguing staff when confronting difficult classification choices.

Comparing the situation around classification and cataloguing at UAL with the Nigerian
university libraries examined by Jibril et. al., librarians at this London institution are at a further
stage where internet reliance has been dealt with for longer; one of the consequences is that UAL

51
librarians no longer have sole responsibility over classification, and instead the employment
structure and responsibilities have been divided. Jibril et. al. conclude that librarians at the
institutions they study are well-trained in classification and cataloguing - through our study, we
then pose the question to what extent will this change once institutions are used to internet usage
and begin to implement stronger divisions in the duties carried out by librarians?

We are reminded of one particular librarian response, which highlights the greater importance of
research question 2: Librarian 4 spoke of the underlying necessity of a library and its librarians to
effectively give access of its collections to users, implying that if classification and indexing do
not fulfil this goal, then their focus will be on other methods. We explored Emily Drabinski’s
(2012) critique of LCSH, and further, their conclusion that Stanford Berman did not go far
enough in proposing change:

Berman did not take issue with the fundamentals of library classifications. The goal of
library classifications - to bring human knowledge together under a single unifying,
universalizing structure and language -was central to Berman's point. (Drabinski, 2012, p.
3)

While Drabinski’s proclamations point to a continued need for the critique of LCSH and
indexing practices, and Librarian 4’s remarks illustrate the need to better understand obstacles to
accessibility, we can discern that research question 2 handles these broader problems, with
classification and indexing at the University of Arts London providing a transferable example.

Inconsistencies in interviewee responses, specifically regarding instances such as the

specification of artists or authors in terms of their country of work, link back to how use of the
local guidelines is imbalanced and not equally recommended by managers. Meanwhile, the
collective invoking of themes around racism and sexism, as well as the criticism of restructuring,
illustrate that staff are well aware of institutional diversity and inclusion campaigns and demand
more. All librarians were weary of the University’s efforts to implement anti-racist methods, in
particular the decolonising project and annual diversity and inclusion training. In measuring
consistencies, we were able to establish a general feeling of disgruntlement that we can arguably
attribute to the consequences of neoliberal ideological application. This in turn opens up a wider
scope; perhaps there is room to clearly lay out neoliberal actions within University settings and
measure the effect on library representations of marginalised communities. Perhaps researchers
akin to Jibril et. al. (2021) can perform a narrative analysis of institutions newly dealing with
neoliberal restructuring and compare librarians perceptions before and after the changes.

The 1998 study by Drabenstott et. al. found that 52% of queried librarians were able to make
sense of Library of Congress subject headings. Our study indicates that UAL librarians have an
understanding of indexing, and the libguide offers references to LCSH standards; however,
indexing rights are not afforded equally to each librarian, and some of it falls on unqualified
centralised staff. In relation to Drabenstott’s study, Grace Lo (2019) presents the

52
recommendation that librarians should be given power to continue to critique LCSH, in order to
allow for the best possible improvements. Again, we can argue that UAL librarians are able to
think critically, however, it is the persistent restructuring policies and lack of allocated time that
hinder the realisation of progress and therefore critical engagement with indexing.

53
 7.2 Limitations
Looking back, I can identify points where I could have improved my interview questions in order
to produce more specific answers. If I had the time I could have also dedicated more effort to
staging an entire set of follow-up interviews. For example, when asked about guidelines in the
first set of interviews, most librarians were vague and could not pinpoint anything specific that
comes from the institution of UAL itself. Through my follow-up questions I then learned that
there are a set of guidelines - although these are limited. If I had put more time into structuring
my interview questions, perhaps I would have stumbled onto this information earlier, and we
could have performed a deeper investigation into the guidelines. However, this instance is still
significant in the way it played out: I chose to conduct a semi-structured interview, and lean on
the benefits of provoking fluid conversation in order to glean a wider set of information from
interviewees. Secondly, the fact that librarians are not themselves clear on what guidelines to use
is a highly significant fact in and of itself; this illustrates a non-cohesive approach to
librarianship, a lack of clear guidelines, and unclear job specifications which all lead to an
inconsistent approach to classification and general librarianship duties.

We also must acknowledge that this research has taken place in the English language, from the
context of a University library situated in the United Kingdom. Results would vary greatly had
we worked from the perspective of a Global South country, especially in understanding
representations of marginalised communities. Covid-19 negatively impacted our scope, making it
harder to contact librarians, as well as sucking up time.

Further, our pool of respondents is located in London which has a population that is only 43.4%
white British (ons.gov.uk). Therefore, respondents are already used to dealing with a certain type
of public. Simultaneously, UAL does not offer up specific opportunities for local students and
therefore holds a disproportionately white and middle class student population - whilst their
education would be positively impacted by librarians’ efforts, and there would be positive
repercussions with their accessing of knowledge, this points to wider systemic problems
regarding who enters the University and therefore receives library access.

7.3 Appraisal of methodology and findings

As Bryman (2004) outlines, qualitative data is useful in terms of the richness and depth the data
provides, but this is also a hurdle - it takes a lot of time and labour to sift through large amounts
of data and it can be more difficult to pick up on patterns (p. 564). Conducting interviews in
order to understand the ways in which UAL librarians perform classification and indexing duties
was successful in that we were able to establish a general picture of the inconsistencies,
techniques and themes present in these processes. We were able to gauge librarians’ perspectives
and the ways in which they utilised previous experience to inform decision making. However,
we may have been able to formulate more productive questions if we had performed a lengthier
observational investigation into the UAL library catalogue itself.

54
Some of the structural issues raised by respondents, such as the recent centralising of cataloguing
staff or the removal of editing and therefore indexing rights for librarians, had direct
consequences on the extent to which we were able to answer research questions. For example,
we were not able to glean much information on indexing. However, this in itself illustrates that
UAL librarians do not necessarily have the agency to be critical when classifying and indexing
items. We again invoke the conclusions of Grace Lo, who deems that institutions must support
librarians to be critical, in order to push the development of indexing systems such as LCSH
(2019); if UAL librarians are not even afforded power over indexing, and if many do not have
editing rights, then how should they make proposals on the evolution thereof? Our findings also
run in line with the work of Hope Olson, exemplifying that there is a need for making
diversification a key basis to the work of librarianship: our catalogue observations laid clear the
many areas needing improvement. There is a mismatch in the way items relating to more
hegemonic themes are represented, versus the more bare records of items of more marginalised
subject matter. Further, UAL librarians do not have access to Global South publishers, they do
not have the time to conduct critical work, and their guidelines are not effective.

In terms of following a method of thematic analysis, our process of coding interviews by theme
and generating follow-up questions was useful. Had there been more time on my side, as well as
availability from respondents, we would have been able to replicate this reflective process
multiple times in order to elicit further answers. Further, it was fruitful to conduct group
interviews, as this allowed respondents to piggy-back and expand their ideas (Bryman, 2004, p.
471). Perhaps it would have been more useful to hold one large group interview, allowing
everyone access to each other’s perspectives and enabling extensive conversation. During the
process of follow-up questions we ensured that each respondent received the same selection of
questions and was afforded equal amounts of time to answer.

Having completed the process, we can put forward certain recommendations for future research
into librarianship at UAL. The University has recently removed the mandatory criteria of
librarianship qualification from the job description of entry level positions, meaning the
institution will rely more heavily on training processes for new staff. This would open up the
potential to perform interviews of new librarians at various stages of training, and then conduct
narrative analysis in order to probe the ways in which training affects librarians’ perceptions on
classification, indexing and library representations of marginalised communities (Bryman, 2004,
p. 582).

7.4 Scope and potential solutions

Arguably, examining the root of librarianship, namely the way it is taught, can provide insight
into problems within the field as well as potential solutions. In their study Integrating diversity
across the LIS curriculum, researchers Bharat Mehra, Hope Olson and Suzana Ahmad (2011)

55
pose a set of questions to respondents - students and educators who are part of the LIS field -
through an online questionnaire, the goal being to establish the mechanisms needed to introduce
the concept of diversity into library practices, as well as distinguish best practices for online
implementation. Respondents describe the structure of library employment that positively
enables inclusion of diverse staff - hiring students at early stages of their careers enables these to
progress further into employment within library science. Drawing from an already diverse
student body thus enables libraries to also diversify their staff body. Similarly to the goals of our
study, this US-based study attempts to illustrate the knowledge of library and information science
practitioners around diversity and inclusion.

Researchers sent their online questionnaire to 333 LIS instructors, receiving a response from 38
(Mehra et. al., 2011). This is quite a contrast to our study, which speaks to five qualified
librarians who practise librarianship as part of their daily work - additionally, we are conducting
open interviews rather than close-ended written questionnaires. The benefit of our methodology
is that we are able to undertake detailed conversations with librarians and establish long-lasting
contact. These conversations, while loosely structured, offer up the possibility for myriad
answers and directions. Further, as some of our interviews constitute group conversations, we are
able to provoke supplementary points of conversation, with librarians building off of each other.
Whilst Mehra et. al. are a group of three professional researchers, our study is limited in that it is
part of a Master's thesis with a far smaller scope.

Compared to the previously discussed study by Jibril et. al. (2021), Mehra et. al. (2011) measure
a less quantifiable approach to librarianship. This perhaps allows for a wider set of results. There
is sufficient background research into the field, with researchers noting that they seek to fill a gap
presented by a lack of studies into diversification as part of improvement in librarianship. This
gap is also evident in our research. Mehra et. al.’s work diverges from our intentions in that it is
also interested in diversity in library and information curriculum as well as the practices of
library educators. In contrast, we seek to understand only librarians’ understanding of
representing marginalised communities; therefore, Mehra et. al. provide research that nonetheless
is significant in understanding the structures and hierarchies of knowledge that lead a librarian
towards certain decisions.

Another option in developing the efficacy of library catalogues exists in the implementation of
user tagging, where digital library catalogue entries present users with the possibility to submit
suggested tags. Opening up tagging to the general public or to a library’s population of patrons
sets up the potential for the infinite accumulation of disorganised data, whilst also diversifying
item descriptions. UAL library catalogue does not offer this interactive function - proposing this
is not within our scope, but it is useful to assess the benefits and predict how our research into
the limitations of classification and indexing in the UAL library catalogue can lay groundwork
for future user tagging projects. Landry et. al. (2011) probe the Stockholm Public Library after
its administering of user tagging in its new 2008 website. Their study finds that most tags relate
to subject headings, which is likely due to the fact that the majority of tags are applied by

56
librarians. Here we are reminded of a different study by Cousins from 1992, which finds that
improving the depth and detail on catalogue records increases retrievability. Potentially, user
tagging may provide greater depth to records. The study by Landre et. al. (2011) is not able to
engage public users and therefore the researchers do not arrive at conclusions on the types of
words users might utilise for tags - potentially this points to a lack of sufficient communication
between catalogue and users. If marginalised communities are to interact with library catalogues
and provide more in depth, specialised tags, as an alternative to subject headings, then these
communities need to be reached in the first place. This links back to our exploration of Mehra et.
al.’s (2011) study, which hints at ways in which libraries can integrate diversity within the LIS
curriculum and thus the librarian profession. A study by Jo Bates and Jennifer Rowley,
conducted in 2011, notes that the use of user tagging will still involve the replication of dominant
discourses within the alternative naming practice - to work around these biases, a University
library would have to actively engage users from marginalised backgrounds. Librarians conduct
induction sessions at the start of each academic year which are mandatory for all foundation,
undergraduate and postgraduate students. Perhaps there is room to incorporate an exercise of
ad-hoc user-tagging - or at least suggestions - into part of the curriculum.

In terms of art libraries specifically, as the University of the Arts London library constitutes,
Singer-Baefsky (2021) notes that the classification of art is problematic in its overlapping nature.
Most art is classified in the 700s of the DDC, with Library of Congress placing various practices
under the ‘fine art’ subject heading; however, LC places certain art under ‘Philosophy,
Psychology and Religion’, due to its divergence from art of the Global North. Already, we can
identify certain orientalist bias (Said, 1978), with Singer-Baefsky (2021) stating that ‘the
challenge facing art libraries comes at the intersection of classification and cataloging [sic]’ (p.
179). The archivist goes on to recognise that a large amount of personal and institutional bias is
involved in making classification and indexing decisions and putting together catalogue records.
The Archives for Black Lives in Philadelphia (A4BLiP) published recommendations in 2019 for
predominantly white institutions to work on incorporating an anti-racist approach to knowledge
categorisation, suggesting the addition of context to pre-existing metadata descriptions and the
offering of alternatives to LCSH. The project also recommends maintaining item records as they
are acquired from the creators, who are more likely to utilise terms relating to their communities.
Of course, this would have to mean that institutions acquire items from places outside of the
larger distributors, which is difficult and certainly unlikely at UAL. Singer-Baefsky (2021) also
acknowledges the difficulty in overriding an entire collection, demonstrating that many
institutions opt to introduce new classification schemes for new acquisitions - this would
complicate systems and require more effort on the part of librarians to educate their users.
However, if we return to Grace Lo’s (2019) assertions regarding the power of librarians, perhaps
all solutions towards the development of indexing point in the direction of a deeper involvement
by librarians - both with stock as well as library patrons. UAL librarians spoke of under-staffing,
lack of time, as well as ignorance, as factors that impede their ability to make progressive
changes within classification and indexing.

57
Therefore, we might logically conclude it necessary to dissolve these obstacles. Unfortunately,
the University of the Arts London library exists frozen within the hierarchies of the larger
academic institution, where the majority of budget goes towards the salaries of management,
who play no role in performing librarianship processes. A second large chunk of the budget goes
towards student experience - arguably, progressing ease of information access constitutes part of
the development of student experience. We are again reminded of Adler’s (2017) examination of
racism within library structures: they express that problems of racism are structural and cannot
be attributed to individuals, even if the power of indexing and classification lies in the hands of
librarians - or in this case, in stilted, inconsistent ways between different staff bodies. Adler’s
assertions expose a contradiction with those of Lo: issues of racism are not the fault of
individuals, however, it is precisely up to individual librarians to enact change. Perhaps we can
arrive at a new set of recommendations; if large structural redistribution is not imminently
possible, then let us lean into the support of librarians. UAL librarians have access to
organisations such as Chartered Institute of Library and Information Professionals (CILIP), a
professional body for librarians in the UK, which provides an educational space for librarians via
its annual conferences, as well as Art Libraries Society (ARLIS). CILIP proposes to increase the
inclusion of people of colour on its panels to 20% (cilip.org.uk). This is not particularly high,
especially when compared with the demographics of London and other UK metropoles.
However, efforts are noted, and point to, at the very least, an awareness of the need for diversity
and inclusion.

Perhaps another solution specific to the library at Camberwell College of Arts exists in its
Special Collections (SC), which houses a selection of materials that are unique to the
Camberwell site. Calhoun (2014) notes that the increase of digital content in libraries is leading
to a homogenisation of library collections, which therefore necessitates heightened development
of any special collections (p. 119). Our study did not take into account the Camberwell archive;
however, this view is useful to set out further recommendations for any future research into or
development of UAL Library procedures. At the same time, we can argue that consideration of
our study would be significant for the careful implementation of classification and indexing of
SC items.

8. Conclusion
In this moment in history where serious work on identity is being overtaken by the capitalist
commodification of identity, it is truly significant to think in terms of knowledge institutions and
their link to processes of identification (Hall, 2013, p. 3). If identification is ever-changing then
the users of libraries are constantly looking to identify with changing resources, and the language
they use to make this identification is ever-changing. There is no ultimate answer to the problems
that arise in classification and indexing, other than to continue to be vigilant as a library
professional in learning new vocabularies and striving to close the gap between user and stock.
We relate back to Bates’ and Rowley’s (2011) assertions that there is no fixed, ultimate method

58
for naming or categorising something, and that libraries are vehicles for society to think and
understand difference. Therefore, our findings do not point to a definitive solution, but rather the
need for continuous review.

Overall, our findings align with previous research, illustrating the weaknesses in University of
the Arts London catalogue records and establishing that there exist inconsistencies in how
librarians perform classification and indexing duties. Grace Lo’s (2019) ultimate
recommendations, regarding the development of Library of Congress Subject Headings as well
as general indexing traditions, rest heavily on the support of librarians and their educational and
critical duties in deconstructing systems of naming: in this regard, we are able to conclude that
UAL librarians are self-reflective and aware of the need for change and dynamism in the library
field. However, we also observe that librarians are not necessarily afforded the powers to enact
this change. In line with employment trends in UK librarianship, UAL librarians earn lower
wages and do not have the time or monetary motivation to input extra efforts.

Calhoun (2014) notes that libraries will buy what their user population demands (p. 116). UAL
Library is not responsible for the recruitment of new students, so it does not have an impact on
the diversity of the student body. Interview respondents noted that University of the Arts London
has been bypassing its ethnicity quota by offering more places to higher-paying, wealthier
international students, instead of offering places to young British students from minority
backgrounds. This will decrease the diversity of the student body - if the library is to reflect this,
then claims for library stock diversification and critical approaches to librarianship may be less
likely to be heard by management. Meanwhile, the diversity of higher education students across
the UK is falling, which is likely to be mirrored at UAL. Further structural issues will have
repercussions on University libraries: A DiversityUK report from 2019 found that only 56% of
Black students will achieve a 2:1, compared to 80% of white students, resulting in lower
retention and recruitment rates for Black students as well as less academic engagement. Enabling
equal access to education, accountability practices, safe educational spaces, continued and
effective support of marginalised students, academic inclusion of subject matter pertinent to
students of colour, representation of staff of colour within academia (which is at 2% for Black
people in the UK), as well as the building of academic spaces within the vicinity of and for
marginalised communities, are all key in creating any link between marginalised groups and
libraries (Diversity UK, 2019).

Our research is notable in that there are no records of other previous investigations into the
University of the Arts Library. Our research could lay the groundwork for further assessments
and recommendations towards the development of indexing and classification at UAL. There
exists more hope from the ground up, in terms of the malleability of libraries. As Satija (2000)
notes, library classification as an academic discipline is very young and has much space to grow
(p. 221). Hope Olson (2001) supports these claims, recommending further research into the ways
in which libraries represent and serve marginalised communities.

59
60
References

Adler, M. (2017). Classification Along the Color Line: Excavating Racism in the Stacks. Journal
of Critical Library and Information Studies, 1(1). https://doi.org/10.24242/jclis.v1i1.17

Ang, I. (2015) Stuart Hall and the Tension between Academic and Intellectual Work.
International Journal of Cultural Studies, 19(1), 29-41
https://doi.org/10.1177%2F1367877915599609

Antracoli, A. A., Berdini, A., Bolding, K., Charlton, F., Ferrara, A., Johnson, V., Rawdon, K.
(2019). Archives for Black Lives in Philadelphia Anti-Racist Description Resources.
Archives for Black Lives in Philadelphia.
https://archivesforblacklives.files.wordpress.com/2019/10/ardr_final.pdf

Baker D. & Evans W. (2013). A Handbook of Digital Library Economics: Operations,

Collections and Services. Chandos Publishing.

Bates, J. & Rowley, J. (2011). Social reproduction and exclusion in subject indexing: A
comparison of public library OPACs and LibraryThing folksonomy. Journal of
Documentation, 67(3), 431-448.
https://doi-org.lib.costello.pub.hb.se/10.1108/00220411111124532

Bates, M. J. (1986). Subject Access in Online Catalogus: A Design Model. Journal of the
American Society for Information Science, 37(1), 357-376.

Berman, S. (1969). Chauvinistic Headings. Library Journal. 94(4), 695-705.

Black, K. (2018). Recruiting Under-represented Groups to Librarianship. Multicultural Learning

and Teaching, 13(1). https://doi.org/10.1515/mlt-2017-0004

Block, S. (2020). Erasure, Misrepresentation and Confusion: Investigating JSTOR Topics on

Women’s and Race Histories. Digital Humanities Quarterly, 14(1).

Bowker, G. C. & Star, S. L. (1999). Sorting Things Out: Classification and Its Consequences.
MIT Press.

Bryman, A. (2004). Social Research Methods (2nd Edition). Oxford University Press.

Calhoun, K. (2014). Exploring digital libraries: foundations, practice, prospects. Facet

Publishing.

61
Caswell, M., Migoni, A. A., Geraci, N., Cifor, M., (Eds.). (2017). ‘To Be Able to Imagine
Otherwise’: community archives and the importance of representation. Archives and
Records, 38(1). DOI: 10.1080/23257962.2016.1260445

Chan, L. M. & Hodges, T. (2000). Entering the Millennium: A New Century for LCSH.
Cataloguing & Classification Quarterly, 29(1-2), 225-234.
https://doi.org/10.1300/J104v29n01_16

Cousins, S. (1992). Enhancing subject access to OPACS: controlled vocabulary vs natural

language. Journal of Documentation, 48(3), 291‐309. https://doi.org/10.1108/eb026899

Crawford, J. C., Thom, L. C. and Powlees, J. A. (1993). A survey of subject access to academic
library catalogues in Great Britain. Journal of Librarianship and Information Science,
25(2), 85-93. https://doi.org/10.1177%2F096100069302500205

Cuna, A. & Angeli, G. (2020). Improving the effectiveness of subject facets in library catalogs
and beyond: a MARC-based semiautomated approach. Library Hi Tech, 39(2), 506-532.
https://doi.org/10.1108/LHT-07-2019-0132

Diversity UK (2019). Universities must do more to tackle ethnic disparity.

https://diversityuk.org/universities-must-do-more-to-tackle-ethnic-disparity/

Drabenstott, K. M., Simcox, S. & Williams, M. (1999). Do Librarians Understand the Subject
Headings in Library Catalogs? Library Resources and Technical Services, 43(3),
140-160. https://doi.org/10.5860/lrts.43n3.140

Drabinski, E. (2012). Teaching the radical catalogue. In K. R. Roberto (Ed), Radical

Cataloguing: Essays at the Front (pp. 198-205). McFarland and Company.

Edwards, J. (2016). “Reflecting the World Increasingly Made Right”: From Response to Action
in Public Libraries. Progressive Librarian, 45(1), 87-94.

Evans, J. (2020). Commodifying diversity: The Danger of Racial Capitalism on Student Growth
in Higher Education. Dean James E. McLeod Freshman Writing Prize, Washington
University. REMAKE.

Frank, E. & Paynter, G. W. (2003). Predicting Library of Congress classifications from Library
of Congress subject headings. Journal of the American Society for Information Science
and Technology, 55(3), 314-227. https://doi.org/10.1002/asi.10360

62
Furner, J. (2008). User tagging of library resources: toward a framework for system evaluation.
International Cataloging and Bibliographic Control, 37(3), 47‐51.

Gibbs, G. R. (2011, November 21). Research Interviewing Part 1: Interviews and the Interview
Society [Video]. Youtube.
https://www.youtube.com/watch?v=9yRgBS2JmXU&ab_channel=GrahamRGibbs

Hall, S. (2013). Representation (2nd ed). Sage.

Harman, S. (2016). Stuart Hall: Re-reading Cultural Identity, Diaspora, and Film. Howard
Journal of Communications, 27(2), 112-129. DOI: 10.1080/10646175.2016.1148651

Hinsliff, G. (2022, September 5). A prime minister must get the big calls right. On the cost of
living crisis, Truss got it badly wrong. The Guardian.
https://www.theguardian.com/commentisfree/2022/sep/05/big-calls-right-cost-of-living-c
risis-liz-truss

Ingwersen, P. (1999). The Role of Libraries and Librarians in Organising Digital Information. De
Gruyter Saur, 40(1), 11-15. https://doi.org/10.1515/libr.1999.49.1.11

Jahns, Y. (Ed.). (2012). Guidelines for Subject Access in National Bibliographies. Walter de
Gruyter.
https://ebookcentral.proquest.com/lib/boras-ebooks/detail.action?docID=893766.

James, S. (2018). Universities are becoming the acceptable face of gentrification. The
Independent.
https://www.independent.co.uk/voices/university-gentrification-ual-ucl-delancy-lendlease
-acceptable-face-a8179816.html.

Jibril, D. U., Amidu G. & Adamu, S. S. (2021). Appraisal of Librarians’ Skills and Awareness on
Online Cataloguing and Classification Tools in University Libraries in Nasarawa State.
International Journal of Library and Information Science Studies, 7(4), 49-60.
https://ssrn.com/abstract=3917636

King, A. (1990). Urbanism, Colonialism, and the World-Economy. Routledge.

https://doi.org/10.4324/9781315715506

Knowlton, S. A. (2005). Three Decades Since Prejudices and Antipathies: A Study of Changes
in the Library of Congress Subject Headings. Cataloguing & Classification Quarterly,
40(2), 123-145. https://doi.org/10.1300/J104v40n02_08

63
Kua, E. (2004). Non-Western Languages and Literatures in the Dewey Decimal Classification
Scheme. Libri, 54(4), 256-265.

Lancaster, F. W., Connell, T. H., Bishop, N. & McCowan, S. (1991). Identifying Barriers to
Effective Subject Access in Library Catalogs. Library Resources and Technical Services,
35(1), 377-391.

Landry, P., Bultrini, L., O’Neill, E. T. and Roe, S. K. (Eds.). (2011). Subject Access : Preparing
for the Future. Walter De Gruyter.

Levin-Clark, M. & Carter, T. M. (2013). ALA Glossary of Library and Information

Science, 4th ed. ALA Editions.

Library of Congress. (2022, April). Library of Congress Subject Headings Introduction,

https://www.loc.gov/aba/publications/FreeLCSH/LCSH44-Main-intro.pdf

Library of Congress. (2022, 10 June). Library of Congress Subject Headings PDF Files.
https://www.loc.gov/aba/publications/FreeLCSH/freelcsh.html#About

Lo, G. (2019). “Aliens” vs. Catalogers: Bias in the Library of Congress Subject Heading. Legal
Reference Services Quarterly, 38(4), 170-196.
https://doi.org/10.1080/0270319X.2019.1696069

Maisuria, A. and Cole, M. (2017). The neoliberalization of higher education in England: An

alternative is possible. SAGE Publications, 15(5), 602-619.
doi.org/10.1177/1478210317719792

Marcelline, M. (2020, July 14). The art school students exposing university racism on Instagram.
Dazed Digital.
https://www.dazeddigital.com/life-culture/article/49791/1/art-school-students-glasgow-lo
ndon-exposing-racism-uk-university-instagram

Matusiak, K. (2006). Towards user centred indexing. OCLC Systems & Services: International
digital library perspectives, 22(4), 283-298. https://doi.org/10.1108/10650750610706998

Mehra, B., Olson, H. A. & Ahmad, S. (2011). Integrating diversity across the LIS curriculum:
An exploratory study of instructors’ perceptions and practices online. IFLA Journal,
37(1), 39-51. https://doi-org.lib.costello.pub.hb.se/10.1177%2F0340035210396781

Moulaison, H. L. (2010). Subject access to materials in online library catalogues: discovery of

Moroccan publications. The Journal of North African Studies, 15(3), 385-397.
https://doi.org/10.1080/13629380902727544

64
Olson, H. A. (2000). Difference, Culture and Change: The Untapped Potential of LCSH.
Cataloguing & Classification Quarterly, 29(1-2), 53-71.
https://doi-org.lib.costello.pub.hb.se/10.1300/J104v29n01_04

Olson, H. A. (2001). The Power to Name: Representation in Library Catalogues. Signs, 26(3),
639-688. http://www.jstor.org/stable/3175535

Patel, S. (2016). Complicating the Tale of “Two Indians”: Mapping ‘South Asian’ Complicity in
White Settler Colonialism Along the Axis of Caste and Anti-Blackness. Theory & Event,
19(4). muse.jhu.edu/article/633278.

Sadler, B. and Bourg, C. (2015). Feminism and the Future of Library Discovery. Code4lib
Journal. 28(1), 1-5.

Said, E. (1978). Orientalism. Routledge & Kegan Paul.

Satija, M. P. (2000). Library Classification: An Essay in Terminology. Knowledge Organization,

27(4), 221-229.

Seminelli, H. (2016). Librarian as professional. Voices of the Future, 71(1), 63-69

https://doi.org/10.1080/0361526X.2016.1168667

Singer-Baefsky, J. (2021). Why Have There Been No Great Art Libraries: The Role of Radical
Cataloging in the Reassessment of Art History. Art Documentation: Journal of the Art
Libraries Society of North America, 40(2). https://doi.org/10.1086/716734

Smithers, R. & Curtis, P. (2003). Art world welcomes Institute change. The Guardian.
https://www.theguardian.com/education/2003/jul/15/highereducation.administration

Templeton, E., Love, B., Davis, B. H. & Davis Jr., M. (2016). Illusion of Inclusion: University
Policies that Perpetuate Exclusion of Students of Color. Journal Committed to Social
Change on Race and Ethnicity, 2(1), 88-115.
https://doi.org/10.15763/issn.2642-2387.2016.2.1.87-115

University of the Arts London Library Resources and Systems (UAL LRS) (2022). Library
Guides: Library Resources & Systems: Classification. University of the Arts London.
https://arts.ac.libguides.com/RandS/Classification

Vincent, J. (2015). The role of libraries in times of crisis.

https://www.cilip.org.uk/news/news.asp?id=482443&hhSearchTerms=%22marginalised+
and+communities%22

65
Wildemuth, B. M. (ed.) (2016). Applications of social research methods to questions in
information and library science. Libraries Unlimited.

Williams, M. (2017). How students were misled and lied to for 20 years. Channel 4.
https://www.channel4.com/news/factcheck/factcheck-how-students-have-been-misled-an
d-lied-to-for-20-years

Wilson, K. (2022). Decolonising library collections: contemporary issues, practical solutions and
examples from LSE. In J. Crilly & R. Everitt (Eds.), Narrative Expansions (pp. 225-250).
Cambridge University Press https://doi.org/10.29085/9781783304998

Xie, I. & Matusiak, K. K. (2016). Discovery digital libraries: theory and practice. Elsevier.

Zaid, Y. (2009). The Use Of The Internet For Cataloguing And Classification. The Information
Technologist, 5(2), 31-35. http://dx.doi.org/10.4314/ict.v5i2.32024

66
Appendix

A: Consent forms
​
Student's name Number of Collection instrument/method
consents
Anuka Ramischwili-Schäfer 5 interview recordings (face-to-face)

B: Interview guide

The aim will be to draw out a picture of how librarians deal with issues of classification and
indexing, and then use this information to understand any limitations. You have signed a
consent form, please know you can withdraw consent at any moment. All information will be
anonymised and the recording will only serve my memory and will not be published
anywhere.

The interest is to look at what the problems are and how you identify them, rather than
necessarily fixing them.

Is English your mother tongue? Did you receive librarian training in English?

At the point of purchasing items, what does the classification and indexing process look like
for you?

What limitations do you see in classification and indexing?

Do you take into account developments in language, when deciding how to classify and index
an item?

Do you have any opinions or immediate thoughts that come to mind regarding the term
‘Caucasian’?

67
Caucasian is currently often used to connote white or European, even though this word stems
from a racist misinterpretation of the original word that refers to the mountain region on the
western edge of the Asian continent. So, the place where Georgia, Armenia, Azerbaijan, and
various smaller unrecognised groups are situated. How do you think this double meaning of
the word ‘Caucasian’ could create problems both in classification and in the use of library
catalogues?

Do you have any thoughts on the term ‘Middle East’?

When searching our library catalogue I noticed that many items that are about the Caucasus are
classified as Middle Eastern. This is a contested way of placing the Caucasus. When
performing classification, how do librarians refer to each other or other sources when deciding
how to place an item?

Are there any immediate issues you see with the library of congress subject headings?

How do you think it is important to take into account perspective when conducting
classification, and the fact that we are situated in a Global North country?

I’ve noticed that when searching according to subject headings, these can be very large, and
the user is faced with a vast amount of items. What can librarians do to tackle such large
subject headings?

Off the top of your head, if you came across a book about an artist from the Caucasus, where
would you classify this?

Would you check with another colleague?

I am aware that one common approach for classifying artists, in the case of our university
library, is to place them amongst the stock geographically, according to the country where they
worked. In terms of marginalised communities I come to think of diaspora. And regarding the
Caucasus, I come to think of the fact that there will have been many artists and writers who
would have fled the Soviet Union and worked elsewhere, but they still identify with the
country they left. How can librarians deal with these problems of movement?

68
Do you have any thoughts on how to deal with classification of diasporic communities? How
can we take into account their multiple points of identity?

In what ways do you think classification can help or hinder a user from accessing items?

Ok we are coming to a close now. Do you have any questions for me? If you want to contact
me about this interview please feel free to do so, I have given you my details. I will stop the
recording now.

C: Follow up questions

1. Off the top of your head, if you came across a book about artists from the Caucasus,
where would you classify this?

2. Would you check with another colleague?

3. How do you think it is important to take into account perspective when conducting
classification, and the fact that we are situated in a Global North country?

4. Do you have any thoughts on how to deal with classification of diasporic

communities? How can we take into account their multiple points of identity?

D: observation schedule
What do library assistants say about picking reservations?

Are there any parts of the Dewey sequence that are more difficult to traverse for staff
members?

Do library assistants complain about specific parts of the Dewey sequence?

69
Do shelvers spend pointedly more or less time tidying any specific part of the sequence?

How do library assistants interact with students when these pose difficult questions about the
collection?

When sharing desk duties with librarians, what observations can you make on their
behaviours?

Does anything else stand out during the daily library assistant activities?

E: observation questions

● What areas of the library are very tidy?

● What areas of the library are very messy?
● Are there certain patterns that you can observe about the space?
● Are there areas of the stacks that seem to be more busy with students than others?
● What is the state of the shelvers timesheet?

70