Searching for Swedish LGBTQI fiction: challenges and solutions

Abstract

Purpose

The purpose of this study is to investigate the needs of potential end-users of a database dedicated to Swedish lesbian, gay, bisexual, transgender, queer, and intersex (LGBTQI) literature (e.g. prose, poetry, drama, graphic novels/comics, and illustrated books), in order to inform the development of a database, search interface functionalities, and an LGBTQI thesaurus for fiction.

Design/methodology/approach

A web questionnaire was distributed in autumn 2021 to potential end-users. The questions covered people's reasons for reading LGBTQI fiction, ways of finding LGBTQI fiction, experience of searching for LGBTQI fiction, usual search elements applied, latest search for LGBTQI fiction, desired subjects to search for, and ideal search functionalities.

Findings

The 101 completed questionnaires showed that most respondents found relevant literature through social media or friends and that most obtained copies of literature from a library. Regarding desirable search functionalities, most respondents would like to see suggestions for related terms to support broader search results (i.e. higher recall). Many also wanted search support that would enable retrieving more specific results based on narrower terms when too many results are retrieved (i.e. higher precision). Over half would also appreciate the option to browse by hierarchically arranged subjects.

Originality/value

This study is the first to show how readers of LGBTQI fiction in Sweden search for and obtain relevant literature. The authors have identified end-user needs that can inform the development of a new database and a thesaurus dedicated to LGBTQI fiction.

Citation

Golub, K., Bergenmar, J. and Humlesjö, S. (2022), "Searching for Swedish LGBTQI fiction: challenges and solutions", Journal of Documentation, Vol. 78 No. 7, pp. 464-484. https://doi.org/10.1108/JD-06-2022-0138

Introduction

While full-text information retrieval may work well for certain search tasks in some contexts, it does little to support the subject searching of fiction. Looking for themes or topics of fiction is challenging in systems relying exclusively on full-text retrieval methods, since works of fiction are characterized by language that is often intentionally metaphorical and themes may be abstract rather than manifest. The problem is exacerbated in lesbian, gay, bisexual, transgender, queer, and intersex (LGBTQI) fiction as LGBTQI themes are not always openly expressed in texts. They can be subtly present through metaphors and intertextuality and expressed in contested historical language.

To alleviate the challenge of LGBTQI fiction searching, the three-year Queerlit project, 2021–2023 (https://www.gu.se/en/research/queerlit-data-base), aims to help the researcher and the general public interested in these topics by building a dedicated database with an advanced search interface as well as an information retrieval thesaurus to support the subject searching of Swedish LGBTQI fiction. This paper investigates the needs of potential end-users in order to inform these developments.

End-user needs were investigated via a web-based questionnaire comprising 40 closed and open questions of which 15 were required. The survey was sent to relevant mailing lists and advertised in associated social media channels in the last week of November 2021. As of the survey closing date, 20 December 2021, 108 responses had been received.

The remainder of the paper is structured as follows. “Background” provides the broader context of the article, building on related research. “Methodology” describes the methods of data collection, sampling, and analysis. Then, in “Results,” the results are reported and discussed together with the main implications for the development of the database, search functionalities, and thesaurus. Finally, “Conclusion” presents suggestions for future research as well as final remarks.

Background

Importance and characteristics of LGBTQI fiction

Queer literature scholars have emphasized that fiction is part of the collective history of LGBTQI people; it is an important medium for knowledge and self-awareness (e.g. Wallace, 2016; Heede, 2015; Love, 2007) as well as part of meaning-making processes in which people seek information to help them understand their lives and how to live them differently (Ruthven, 2019; Huttunen and Kortelainen, 2021). This is particularly relevant to the coming-of-age experiences of LGBTQI people (Adler, 2015, p. 479; Banks, 2009), since adolescence has been culturally scripted as the period in life when heterosexual identities are developed. Similarly, Liming (2007), Pecosksie (2012), and Rothbauer (2013) demonstrated the importance of literature as support for identity formation and the communication of sexual identities, particularly during the early stages of understanding one's sexual identity. Often the first encounter with transgender identity is a serendipitous one through fiction or movies (Pohjanen and Terttu, 2016).

And yet there are few overviews of LGBTQI literary fiction. Gay male literature is the only research field where larger overviews are found (Koponen, 1993; Cart and Jenkins, 2006), while studies of lesbian literature have often focused on particular themes or works (e.g. Faderman, 1981; Zimmerman, 1990; Castle, 1993; Griffin, 1993; Smith, 1997; Johnston, 2007). As for trans and intersex literature, these are fields that have only recently started to develop, and it is only during the last five or six years that scholars have started to outline transgender and intersex literary canons (e.g. Amato, 2016; Koch, 2017; Chess, 2016), although specific themes have been analyzed since the 1990s (e.g. Garber, 1992).

What we do know about LGBTQI fictional literature is that specific LGBTQI themes may have been present, but less overtly represented, at least until the 1990s, when LGBTQI themes started to become more overtly manifested (Cart and Jenkins, 2006). The LGBTQI spectrum has been unevenly represented, with gay male experiences clearly overrepresented (Cart and Jenkins, 2006; Hicks and Kerrigan, 2020). Research on young adult (YA) literature shows a dominance of tragic stories of LGBTQI characters, at least until the 1990s; however, even fiction describing LGBTQI identities in a more inclusive way often presents characters “whose existence is a struggle” (Banks, 2009, p. 35). Transgendered characters appear less often and in less central roles than do lesbian, gay or bisexual characters, while their portrayal is less positive (Adams and Pierce, 2013; Waite, 2013). All these findings about LGBTQI fictional literature are relevant to understanding the survey results presented later in “Results.”

Indexing fiction

The subject indexing of fiction in libraries is usually limited to genre, complemented with facets of time and place (Saarti, 1999). However, users' complex information needs often cover “a combination of different aspects, such as specific genres or plot elements, engagement or novelty” (Bogers and Petras, 2017), and widely used controlled vocabularies do not normally address such information needs well (Bogers and Petras, 2017). It has thus been advocated that more subjective aspects of literary fiction known as “appeal characteristics” should be represented (Saricks and Wyatt, 2019). Appeal characteristics such as pacing, characterization, storyline, frame/setting, tone, and language/style are identified to help users find works like ones they have already enjoyed. However, there are no widely available subject indexing systems for fiction that support the indexing of those aspects.

The subject indexing of fiction entails great complexity, as fiction is “innately subjective; themes are hidden, motifs are subtle, and sometimes the story itself is a representation of an idea, philosophy, or lesson” (Viola, 2020). A methodologically difficult question is how to separate “denotative,” or factual, elements from “connotative” elements (Saarti, 2019), that is, elements interpretively derived from the text. This is an especially acute challenge for fiction featuring same-sex desire, where “homosexuality often appears through implication” (Campbell, 2013, p. 295).

Developing alternative or complementary approaches to the professional subject indexing of fiction has thus been pertinent. In the last two decades, solutions have mostly been found in social tagging services, as automatic options remain largely unsuccessful in texts characterized by subjectivity, metaphoricity, and hidden and subtle motifs. Despite the recognized disadvantages of social tagging, Rafferty (2018) concluded that although such tagging may underperform established subject indexing systems, it will still “complement, enrich, and … enhance conventional retrieval systems” (p. 510). These findings were reiterated by Adler (2009) for transgender subjects and by Bates and Rowley (2010) for LGBTQ as well as racial and ethnic minorities. Bates and Rowley (2010) concluded that library catalogs were lagging the community vernaculars, for example, using “transsexual” instead of the community-preferred term “transgender,” even after the term “transgender people” had been added to the Library of Congress Subject Headings (LCSH), and omitting the term “queer,” which has been widely applied to works on LibraryThing that convey criticism of heteronormativity. However, while user tags on LibraryThing were found to provide more relevant and accurate access for users searching for LGBTQ literature, this presents a problem for British users, since this folksonomy largely reflects an “Americentric worldview” (Bates and Rowley, 2010, p. 443). This difference in language use is of course a potential problem for users in the rest of Europe and other parts of the world.

Indexing LGBTQI fiction

General controlled vocabularies do not represent LGBTQI themes well. There is a long history of critique of biases in controlled vocabularies (see, e.g. Olson, 2002, for an overview) and their failure to “respectfully organize library materials about social groups and identities that lack social and political power” (Drabinski, 2013, p. 95). A well-known example is the pathologizing of homosexuality or trans identities present in systems, such as LCSH, criticized early on by Berman (1971/1993) in the 1970s. More recently, Edge (2019) noted that LCSH includes terms representing LGBTQI individuals that these individuals would use to describe themselves, while terms such as “heterosexual” have been lacking.

Furthermore, Edge (2019) showed that another problem in providing adequate terms for LGBTQI actions and identities is the changing vocabularies for these and ambiguous uses of certain terms such as “gay.” The term “queer” as a concept signifies fluid identities and resistance to definition, so a queer indexing practice must allow for changeability and rely on collective understandings rooted in the community (Keilty, 2009). Both Edge (2019) and Rawson (2009) also raised the question of changeable identities. Is it, for example, correct to apply the term “transgender” to a cisgender (i.e. non-trans) woman passing as a man, and not identifying as transgender?

Queerlit project

Most Swedish libraries use Swedish Subject Headings (SAO) for the subject indexing of literature (both factual and fictional). SAO comprises slightly fewer than 40,000 terms (Libris, 2022). The representation of terms for LGBTQI people is typically broad – for example, the term “transgendered people” is there but not more specific terms such as “trans men,” “trans women,” or “transitioning.” Depending on the definition, 80–150 subject headings could be described as referring to LGBTQI subjects. Fictional literature in general is indexed using only a few terms: in 2020 the average was 3.7 terms (National Library of Sweden, 2022).

For children's and YA literature, another indexing system is used, the Children's Subject Heading List (The Swedish Institute for Children’s Books, 2022). In this system, terms describing LGBTQI are also rather few and general, comprising only the following: “asexuality,” “bisexuality,” “homosexuality,” “homosexual parents,” “non-binary,” “coming out,” “queer,” “rainbow families,” and “transgender persons.”

To address the challenges that LGBTQI literary fiction is too sparingly thematically described and that SAO LGBTQI terms are too broad and lacking in specificity, the Swedish Queerlit project aims to create a sub-database of the Swedish union catalog Libris (https://libris.kb.se) covering Swedish LGBTQI fiction, primarily for scholars but also for the general public. In addition to the joint LIBRIS interface, the Queerlit sub-database will be searchable through a separate interface supporting advanced search functionalities based on the dedicated LGBTQI thesaurus also developed as part of the project. This thesaurus, the Queer Literature Indexing Thesaurus (QLIT), is largely based on the English-language Homosaurus Vocabulary (https://homosaurus.org/v3) used by the Digital Transgender Archive (DTA) (https://www.digitaltransgenderarchive.net), among others. Homosaurus was developed to cover a range of varied information resources and its first version dates to 1997 (for more about Homosaurus, see Watson et al., 2021; Cifor and Rawson, 2022).

The QLIT thesaurus comprises selected terms from Homosaurus translated to Swedish. Of the 1,186 Homosaurus terms in 2021 when the work began, we selected those that had been used to index at least one work of fiction in databases using Homosaurus, in this case the Internationaal Homo/Lesbisch Informatiecentrum en Archief (IHLIA) and DTA databases. Those terms were enriched with narrower and broader terms. The QLIT thesaurus was further adapted for works of fiction and to reflect the Scandinavian context. This entailed: adding terms for the indigenous Sami people; adjusting judicial terms to reflect Swedish legislation; adding terms suggested by end-users in the project's workshops; and adding common symbolic concepts in queer fiction such as “mirrors” (in prose) and “rainbows” (in picture books).

The QLIT thesaurus today comprises 848 terms of which 91 are unique to Queerlit, while the remaining 757 have exact or close matches in the Homosaurus. The thesaurus and the search functionalities are further informed by the needs of potential users, including subject experts, librarians, and the general public. The remainder of this paper describes the study surveying needs of potential users.

Methodology

Purpose and aims

In order to help inform developments related to a dedicated LGBTQI fiction database and search interface in Sweden, the study aims to determine the needs of potential end users. Specific research questions were the following:

RQ1.

Why do people read LGBTQI fiction?

RQ2.

How do people find out about which LGBTQI fiction titles to read?

RQ3.

How happy are people with online searching for LGBTQI fiction?

RQ4.

What search functionalities in online search systems would improve the search experience?

RQ5.

What specific subject terms would users like to see supported in search systems/thesauri?

Since potential end users involve both general public as readers and researchers concerned with LGBTQI fiction, a web-based questionnaire was used in this first research stage of the Queerlit project in order to collect a significant number of responses from heterogenous end users via electronic channels.

Data collection

The web questionnaire comprised 40 open and closed questions, of which 15 were required. It included a description of what the survey was intended for, an estimate of 10–15 min needed to complete it, a statement that it was anonymous, and a statement that submitting the survey simultaneously indicated consent to participation. The background questions targeted age, highest education level, and frequency of reading LGTBQI fiction. The main section entitled “Reading and searching for LGBTQI fiction” started by defining LGBTQI fiction as fictional works with LGBTQI themes, characters, or motifs. Here questions were asked about: people's reasons for reading LGBTQI fiction, how they find out about what LGBTQI fiction to read (e.g. via friends, libraries, and social media), where they find copies of the desired works, the format that they normally read (e.g. paper, e-books, and audio), how they look for LGBTQI fiction on the internet (e.g. library catalogs and bookstores), their experience of searching for LGBTQI fiction in general, common search facets (e.g. author and subject), their latest search for LGBTQI fiction, desirable subjects to search for, as well as desirable search functionalities.

The survey was administered via Google Forms, chosen based on the team's earlier experience and its lack of cost. However, this limited the question structuring, resulting in a questionnaire that was lengthier than needed; this was due to the need to create a separate open question to allow participants to explain why they chose “Other” in a closed question. Initial versions of the questionnaire were pilot tested for clarity, usability of the data collected, and duration with several people outside the project. As a result, the questions were slightly revised for greater clarity.

Sample

The final version was distributed via various channels in the last week of November 2021, with input collected by 20 December 2021. These channels included:

1. Social media: the Queerlit project's Instagram and Facebook presence; a Swedish literature Facebook group (Litteraturgäris); KvinnSam's Facebook page (KvinnSam is the Swedish National Resource Library for Gender Studies, also a main project partner);

2. Faculties and departments of project partners, including the six colleagues in the project advisory board (from the fields of LGBTQ literary studies and library and information studies);

3. Relevant Swedish mailing lists and email contacts of institutions and organizations such as the LGBTQI fiction network (Bögbibblan), Lesbian Power (Lesbisk makt), the Queer Movement Archive and Library (Queerrörelsens arkiv och bibliotek), the Swedish Archive for Queer Moving Images, RFSL (Swedish Federation for Lesbian, Gay, Bisexual and Transgender Rights), the Female Academics Association (Kvinnliga Akademikers Förening), the National Research School in History, the Center for Gender Medicine (CfG), and a range of research centers on gender studies at Swedish Universities; and

4. Relevant Danish and Norwegian mailing lists and email contacts of institutions and organizations (as the three languages are understood across their borders) such as the Danish Center for Research on Women and Gender (KVINFO) and the Norwegian archive for queer history.

Data analysis

From the end of November to 20 December 2021, in total 108 replies were collected. Data analysis was conducted using Google Forms' built-in data representation: the summary overview and the spreadsheet of complete data. Content analysis of open-ended questions was first conducted by one author of this paper and the initial results were reviewed by each of the other two authors, in turn, for several iterations, until final agreement was reached.

Results

Note: the quotations from participants are translated into English.

Demographics

Of the 108 anonymous respondents, most (81 or 75%) were in the younger age range of 18–39 years; of these, about half (42) were 30–39 years old and the other half (39) were 18–29 years old. Older age ranges were represented by fewer participants: 16 were 40–49 years old (14.8%), eight 50–59 years old (7.4%), two 60–69 years old (1.9%), and one 70 or more years old (0.9%). Considering that this was a web-based questionnaire distributed via mailing lists and social media, more participants in the younger age ranges would be expected, as these age groups are more present online. The results of the study reflect the information needs related to LGBTQI fiction characteristics of adults under the age of 50 years, while other age groups remain underexamined.

The majority of participants (84 or 77.5%) had completed university education. More specifically, 36 participants (33.3%) had a Bachelor's degree as their highest educational achievement, 11 (10.2%) a four-year Master's degree (Swedish “magister”), 30 (27.8%) a five-year Master's degree, and seven (6.5%) a doctoral degree. About a fifth of the participants had so far completed high school as their highest educational level (22 or 20.4%) and two had completed elementary school (1.9%). The high proportion of people with completed university education in the sample was much higher than the estimate of 29% for the Swedish general population in 2020 (Statistikmyndigheten (SCB), 2021). This is likely because the survey was disseminated via channels visited by people who tend to read a lot, which, at least here, also coincided with higher educational achievements.

Reading frequency

About half of the respondents read LGBTQI fiction several times per year (53 or 49.1%). Next were more avid readers: 28 (25.9%) read LGBTQI fiction a few times per month, 16 (14.8%) every week, and four (3.7%) daily. Seven participants never read LGBTQI fiction, giving several reasons for this:

1. Three indicated that they would like to read LGBTQI fiction but cannot find books easily, which also shows the difficulty of finding LGBTQI fiction;

2. Three preferred to choose books based on other features such as how interesting the story is, rather than author's or characters' sexual orientation; and

3. One indicated lack of personal interest in recurring themes of injustice, discrimination, and difficulties in coming out in a heteronormative society.

The seven participants who never read LGBTQI fiction automatically left the survey at that point and were thus excluded from further results.

Reasons for reading

The remaining 101 respondents were then asked about why they chose to read fiction with LGBTQI themes. The results revealed the following most popular reasons:

1. For pleasure reading in free time (86 respondents or 85.1%);

2. To find experiences or characters they can identify with (82 or 81.2%);

3. To find subjects and themes they can identify with (74 or 73.3%); and

4. Interested in LGBTQI questions (71 or 70.3%).

Less popular but still significant were the following reasons:

1. To better understand LGBTQI persons' experiences (31 respondents or 30.7%);

2. To find positive role models and ways of living (36 or 35.6%); and

3. For work (e.g. research) (22 or 21.8%).

Seven respondents (6.9%) indicated that they read LGBTQI fiction because they had happened on it, not because they actively searched for it; only one of these seven replies was from a person who indicated that as the only option, while all others showed some interest in LGBTQI themes, i.e. chose other of the above reasons as well. In total, 10 respondents also indicated other reasons, and some specified them further: to enjoy a good book that may happen to have LGBTQI themes (mentioned by 2); to immerse oneself in LGBTQI culture (1); to feel “queer romance” (1); and being a member of book circles that often read such books (1). Other stated reasons included reading professionally (as a librarian) (1) and independent research outside regular employment (1).

Information channels

The next multiple-choice question, about what information channels respondents used to identify what LGBTQI fiction to read, showed that most used either: social media such as TikTok, Reddit, Facebook, Goodreads, and LibraryThing (mentioned by 77 or 76.2%), and/or b) friends and acquaintances (73 or 72.3%). A considerable number relied on newspapers or magazines (31 or 30.7%), the library (e.g. via public displays or author events) (29 or 28.7%), and bookstores (16 or 15.8%). In total, five (5%) were not actively looking for LGBTQI fiction, likely related to reasons stated in response to the previous question: they enjoy a good book that may happen to have LGBTQI themes or are in reading clubs that choose the books. As many as 22 respondents (21.8%) had other channels, too. It should be recalled that social media, especially Instagram and Facebook, innately also represent the information channels of people such as friends and acquaintances: as one person wrote, “most often friends write on Instagram and Facebook.” The exception may be TikTok, which is a social media platform that does not rely on personal profiles and where “interpersonal connections are downplayed” and, therefore, “creative interaction is also prioritized over discursive interaction” (Zulli and Zulli, 2020, p. 2).

Regarding libraries, most used municipal public libraries (9 people). Related to this library type, some specifically mentioned the following: events featuring LGBTQI authors (4, 2 of whom learned about them through their library's Facebook page), catalog search (3), reading tips (2), and a rainbow shelf, a physical shelf or section at the library where LGBTQI material is gathered (2 mentions, 1 specifically in the city of Umeå, which seems particularly active in this area). A university library was mentioned by two, one of whom specifically referred to its databases, and their high-school libraries were mentioned by two, one of whom specifically referred to the library's book display. One person also used the RFSL library.

Regarding magazines and newspapers, the following were mentioned: the combined online and print magazine QX (2 people), specializing in LGBTQI topics; Ottar (1), which is the RFSU's (the Swedish organization for sexual and reproductive health and rights) largest magazine on LGBTQI themes, including politics; morning newspapers with both national and local coverage, such as Dagens Nyheter (1), Göteborgs Posten (1), and Svenska Dagbladet (1); the largest evening papers Expressen (1) and Aftonbladet (1); and Svensk Bokhandel, a magazine for professionals in book publishing and sales (1). The choice of general newspapers indicates a certain level of coverage of LGBTQI fiction by the mainstream media in Sweden.

Regarding bookstores, both online and physical bookstores were used, as well as general and specialized ones. Adlibris (used by 3), Akademibokhandeln (2), and Bokus (1) are the largest general bookstores in Sweden; Adlibris and Bokus are primarily online, while Akademibokhandeln is the dominant physical chain. One person also mentioned the physical chain Pocket Shop, covering different subjects, and one mentioned diverse physical bookstores in Sweden and abroad while traveling. Of the specialized bookstores, the Science Fiction Bokhandeln, which is both an online store and a physical chain, was chosen by two, and Hallongrottan in Stockholm, a physical bookstore, was used in the past by one person – it no longer operates but used to focus on feminist and LGBTQI literature. Two used Page 28, an LGBTQI bookstore and cultural center in Malmö, also operating an online webstore, and the New York-based Bureau of General Services—Queer Division, which also performs the dual function of a cultural center and a bookstore with an online version. Two mentioned Storytel, a streaming service for e-books and audio books.

Regarding other information channels not addressed in the previous question, 12 respondents relied on Internet searches, 10 of whom specified Google, using search queries such as “LGBTQ novels” or “queer YA.” An additional two mentioned specific websites for lesbian fiction and one that used to list Swedish LGBT fiction (https://www.janmagnusson.se). Two also referred to Autostraddle (a website for the LGBTQI community) and one to Archive of Our Own (AO3), a fan-run web archive for fanfiction. Three followed publishers with dedicated LGBTQI lists.

One respondent also mentioned relying on Swedish public radio, likely through cultural news. Another would ask bookstore staff for guidance and recommendations. One person responded to reading challenges such as those available via Book Riot. In the librarian role, one participant mentioned looking in all possible channels to acquire books by diverse authors and in different genres.

Several referred to finding information through reading and following up on references (mentioned by 5), including from watching documentaries (1), reading more books by the same author (1), and via the publications of researchers writing on LGBTQI topics (3). Information could also be found serendipitously: one person described finding out about Samuel R. Delany via research on Foucault (Delany is an American science-fiction author exploring queer themes, also known for his autobiographical accounts of his life as a gay person of color as well as a writer with dyslexia). One person generally “keeps an open eye,” writing down recommendations on LGBTQI books.

Information channels on the internet

Looking more deeply into the internet channels specifically, the majority used social media as the dominant online information sources – 72 participants (71.3%). An optional question about which specific social media showed that the most common were:

1. Instagram (23)—specific mentions were made of Instagram's forums and book clubs (1 respondent), Bookstagram (1), and Instagram pages such as Jennifers Lesbiska Boktips (Jennifer's Lesbian Book Recommendations) (1) and the Bureau of General Services—Queer Division's Instagram (1);

2. Services for reading recommendations and personal reading tracking (20), primarily Goodreads (19) and Storygraph (1); and

3. Facebook (15) – two specifically mentioned the Litteraturgäris page and two Lesbiska bokklubben (Lesbian Bookclub).

Microblogs were used by seven respondents, specifically Twitter (used by 5) and Tumblr (2); YouTube was used by three, one of whom specifically used the “BookTube” sub-community; Tiktok was used by three and Reddit by one. One person commented that on social media they followed publishers with a pronounced LGBTQI “agenda” and authors who have previously written LGBTQI books who then may promote their own work as well as works by others. One informant found LGBTQI literature by following illustrators who have designed covers for queer books.

Just under half of the respondents consulted online library catalogs – 45 or 44.6%; of these, 18 respondents (17.8%) used Libris, in which we intend to create the Queerlit sub-database, while 27 or 26.7% consulted other library catalogs. Search systems at university libraries such as discovery services were used by nine participants (8.9%). Regarding specific online library catalogs, references were made to municipal public library catalogs (used by 17), specifically of municipal libraries in Gothenburg (3), Malmö (3), Stockholm (3), Lund (2), Hörby (1), Höör (1), and Mölndal (1). Two also used Danish public libraries, with one specifically mentioning bibliotek.dk, their union catalog. In addition, specific university library search systems mentioned were Lunds University Library (1), Södertörns University Library (1), as well as OneSearch (1), which is a discovery service. One person commented that they used Libris only for known-item searching (“When I already know about the work or author”), indicating the problem of subject searching in the union catalog, a topic addressed further in other questions below.

Many also used online bookstores (41 or 40.6%). Of online bookstores, the most commonly used were general subject ones, such as Adlibris (used by 15), Bokus (9), and Akademibokhandeln (3). Two people mentioned using the Science Fiction bookstore (SF Bokhandeln) and one an unspecified online antiquarian bookstore. Three respondents mentioned Storytel, a streaming service for audiobooks and e-books.

Twenty-four respondents (23.8%) also used other online channels. Ten people resorted to recommendations from publishers, relevant magazines, and libraries: one got recommendations from their municipal library website and another from the RFSL library; one librarian relied on library acquisition channels; one person relied on publishers' catalogs, one on the magazine Svensk bokhandel, which features newly published books twice per year, one on LAMBDA, which gives a prestigious award for LGBTQI literature called the Lambda Literary Award, and one on Autostraddle web magazine; and one person resorted to Kvinnohöjden feminist course literature. Also interesting were blogs (used by 6), specifically Tekoppenstankar (2) and Lesbiska bokbloggen (1), both Swedish blogs about LGBTQI literature. Other popular channels were websites focusing on LGBTQI literature (3), as well as those mentioned earlier, such as the website of Page 28, a Swedish LGBTQI bookstore and venue for queer culture (3), the website, social media page, and podcast Bögbibblan (2), and the podcast and website Book Riot (1). Some used more general websites such as Internet Archive (1), Wikipedia (1), and Listchallenges.com for reading challenges (1) and/or looked for information on Google (7).

Obtaining book copies

The next multiple-choice question was about how to obtain copies of books. Most got books from a library (72 or 71.3%) and/or a bookstore (70 or 69.3%). About a third got books from friends and acquaintances (35 or 34.7%) as well as from streaming services (31 or 30.7%). A good proportion of the respondents also used antiquarian bookstores and flea markets (24 or 23.8%). The 11 (10.9%) participants who also used other ways to acquire books used the following:

1. Downloaded free e-books (used by 3; one explained that authors often make their e-books freely available, especially fantasy and sci-fi books with queer characters);

2. Online antiquarian bookstores (2);

3. AO3 for fanfiction and original work by fan authors (1);

4. Kindle (1);

5. Internet Archive (1);

6. Internet (1); and

7. Work (2), i.e. through advance copies from publishers or via library work.

The preferred book format was the printed book (used by 94 or 93.1% participants). About a third used audiobooks from streaming services (32 or 31.7%). Some also read e-books in apps, either on tablets (13 or 12.9%) or mobile phones (14 or 13.9%); 10 (10%) read e-books via Internet browsers.

Online searching experience

The following questions are about online searching experiences in relation to LGBTQI fiction. The first one examines general satisfaction with using typical online information channels when searching for LGBTQI fiction. On a scale ranging from 1 to 5, where 1 stands for completely unsatisfactory and 5 for completely satisfactory, most respondents (66 or 65.3%) chose 3 for neither satisfactory nor unsatisfactory. Nineteen respondents (18.8%) considered their searching experiences relatively satisfactory (4 on the scale) and 12 (11.9%) rather unsatisfactory (point 2). Only four respondents (4%) were completely satisfied (5 on the scale), while none were completely unsatisfied (1 on the scale). However, there were indications that some experiences, such as unsuccessful attempts at subject searching in library catalogs, made some people abandon a particular service altogether (see below). Also, in the very first questions of the survey we heard about three people who no longer actively searched for LGBTQI fiction because they had previously been unable to find it.

When asked to explain their responses about online searching, the reasons varied. Most replies cited dissatisfaction with online searching for LGBTQI fiction. The most common reason (mentioned by 23) for this dissatisfaction was that subject searching in library catalogs does not work well or, as one respondent wrote, “it is hard to search for LGBTQI works systematically.” Most believed the reasons for this difficulty were:

1. Too few books with LGBTQI keywords (5), for example:

   • “Books with LGBTQI themes that I know of don’t get retrieved”;

   • “Older literature and non-Swedish literature can be difficult to identify”;

   • “Few identified LGBTQI works are in the library catalog; unless the main theme is obvious from title or review, it won’t be classified as such”;

   • “Library catalogs or bookstore webpages rarely provide information that a book has an LGBTQI theme, unless that's the whole point of the plot and it is printed clearly in the back text or the like. Side characters and more subtle storylines are not specified in the metadata or information text.” This specific problem was identified earlier by De la Tierra (2008).

2. Indexing or indexing systems that are too shallow or too general (6), for example:

   • “Subtler themes or supporting characters are not indexed”;

   • “Overly general subject keywords and titles seldom reveal what the book is about”;

   • “I tried to find fiction about lesbian/queer divorces and did not find a good search, the same with queer literature where friendship is in focus, and about queer families with a different type of narrative than the usual ‘have children’ or ‘be a rainbow family/star family’ [‘Star family’ is wordplay on the Swedish term for nuclear family, with the star family in comparison being a family with stepparents, stepsiblings, or other types of family members.] in a heteronormative environment.”

However, one person also mentioned the opposite problem, that of too exhaustive indexing: “Overly broad searches result in books that just barely touch on LGBTQI themes.”

1. Not all libraries use LGBTQI keywords (2).

2. User-unfriendly interfaces (mentioned by 2), especially of municipal library databases (1): One respondent noted that it is hard to find LGBTQI literature using subject keywords; one must, therefore, rely on browsing an online “rainbow shelf,” which is hard to scroll through because of the poor interface (this comment also indicates the importance of hierarchical subject browsing; see, e.g. Koch et al., 2006).

Another respondent described actively searching the library catalog a few times but not being successful – “[It was] as if information was missing”; two others said they usually did not find even known items in library catalogs and thus preferred bookstores.

Many commented that searching did not work well in bookstores or on social media (mentioned by 22), and more specifically on social media because LGBTQI categories were missing (3); when social media such as Goodreads, Instagram, and blogs do use LGBTQI keywords, some considered them too general, making it hard to find certain topics (5), as was noted above concerning library catalogs:

1. “Overly general subject keywords (LGBTQI) and especially no more unusual subjects such as poly, asexuality, intersex”;

2. “Too much focus on identities/sexualities and too little on what takes place”;

3. “Little help from LGBTQ + tags—LGBTQ + tags don’t tell me anything. If I want a romance between queer women, a search for LGBTQ + or queer or romance will usually give me results about, for example, gay men” and

4. “LGBTQI [literature] is often for young adults, and it is hard to find [literature] for other age groups, especially if combined with fantasy.”

Two respondents also thought that searching was too complex, including one who found it hard to find the right tags in Goodreads or on Instagram, indicating a need for subject browsing.

Bookstores and Storytel were also considered hard to search because of a lack of subject keywords, meaning that only known-item searching worked (mentioned by 8). Google was considered hard to search based on facets or ways in which a book is about LGBTQI themes (1).

As many as seven respondents thought that there were too few LGBTQI fiction works, especially titles for adults (most are for youth) (mentioned by 2) and, moreover, that the quality of what they did find was not that good (2). One also wrote that it was easier to find books abroad than in Sweden. Another one commented that it was hard to find books in languages other than English and Swedish (Anglophone literature dominates other literature as well).

Of those who were happy with their online searching experience, one was satisfied because they thought that AO3 worked well with good specific tags and two appreciated Goodreads for the same reason, although AO3 and Goodreads use social tagging services similar to those criticized by others (see above). Two were happy because known-item searching in library catalogs works well, another one was a librarian who knows how to search for information, and a fourth one, although unable to find LGBTQI literature as a category, was happy with the other literature found.

Common metadata elements

As to the types of elements searched for, the most common were searches by author (used by 71 or 70.3%) and title (60 or 59.4%), while half of participants conducted subject searching (51 or 50.5%). As seen above, some readers do not actively search for subjects but instead follow social media for information, using libraries, bookstores, and streaming services only when they know what specific item to look for. The dominance of known-item searching could partly be due to the low-quality subject searching (and indexing) or subject browsing available in many online library catalogs, bookstores, and social media systems.

About a third of respondents searched for genres such as crime fiction, novels, and poetry (34 or 33.7%), and about as many combined two or more aspects (39 or 38.6%). Some used other ways of searching (8 or 7.9%) such as: year of publication; recommendations based on liked books (“On Goodreads I search for a book I like, and then the site recommends similar books”); or specific subject facets such as identity (e.g. intersex, asexuality/ace [“ace” is short for asexual], as well as a combination of ideological point of departure [e.g. socialism and Marxism] and genre [e.g. science fiction and fantasy]).

Their latest search

To capture the context of their search tasks, we asked respondents about their latest search for LGBTQI literature by subject – the reason for the search and then the subject search terms. Most (59) searched for LGBTQI fiction by subject for pleasure, citing specific reasons such as:

1. “I was intrigued by a specific event in history and wanted to find out more about how LGBTQ people experienced these events. It was only for pleasure reading, to be better read.”

2. “I was looking for horror from an LGBTQ perspective and preferably in the sci-fi genre. It was important that the author was queer but preferably not a cisgendered man. The purpose was to find something both a little thrilling and stimulating to listen to in the dark.”

3. “I wanted the sequel to a historical lesbian romance I had read and liked.”

4. “Pleasure—looked for a fictional work with characters with trans experience/gender in relation to sexuality (being a lesbian and how it plays a role in identity), found nothing.”

5. “I was looking for a book with a Christmas theme for fun reading. Wanted to find something that was not just romance focused.”

6. “The last time I was looking for a book with LGBTQ characters and relationships was when I was looking for porn stories to read as entertainment. Then I used TikTok's tagging system to find recommendations, and then specifically looked for users on ‘Booktok’ that I already follow and whose recommendations I had followed earlier. Then I searched for the titles that came up on the authors' own pages, Storytel, as well as pirate download pages for e-books.”

The following three comments indicate the need to include a “feel-good” facet in subject indexing:

1. “Feelgood novel.”

2. “Wanted to find a feelgood book with an LGBTQ theme, preferably love. Thinking like average straight romcom but queer. I was so tired of all the depressing LGBTQ stories and think that a lot of YA (especially fantasy) is poorly written. It was for light reading.”

3. “I was trying to find a new novel to read. It was quite difficult. I usually want more information than I am given. I usually look for fantasy and science fiction with romance between queer women (and a happy ending).”

Seven combined pleasure with another purpose, such as:

1. Pleasure and work (2 respondents), including one who is an acquisitions librarian;

2. Pleasure and studies (“I was looking for fiction on sexuality, both for pleasure and related to my studies”);

3. Pleasure and information (2) and

4. Pleasure and knowledge of the 1950s.

Four respondents also listed finding literature about experiences they can relate to as the purpose of their search, including:

1. “To get away from the heteronorms that are in everything in general and that are not something I can relate to.”

2. “To find experiences I can relate to, specifically to fill up on good experiences.”

Three searched for information purposes, including:

1. “Curious to find out more about an author's works.”

2. “I wanted to know who wrote the book, when it was published, and a brief description of what it is about. Both genre, themes, and what LGBTQ + people are represented (is the book about a trans person or a lesbian person, for example).”

Six respondents searched for research purposes (e.g. for a BA thesis, for studies, and to edit a Wikipedia entry), and four indicated work-related searching (e.g. to set up a book display at one's work). One searched on behalf of another person.

Their latest search: common subject terms

Of subject search terms used in their most recent online search, the most common, either alone or in combination, were terms representing LGBTQI identities such as: HBTQ (in English: LGBTQ) (mentioned by 29), including HBTQ+ (LGBTQ+) (1), HBTQI (LGBTQI) (2), LGBT (2), LGBT(Q+) (1), HBTQIA+ (LGBTQIA+) (1), Lhbtqia (1), and Lgbtqia+ (1). The first three are Swedish acronyms for LGBTQI identities, with H standing for homosexual, B for bisexual, T for transgender, and Q for queer. Acronyms including I for intersex and A for asexual are also used in Swedish, and + is similarly used to indicate inclusiveness. Other common search terms were:

• Lesbian (21),

• Queer (14), gay (5),

• Homosexual and related word forms (e.g. homosexuality) (5)

• Non-binary (4),

• Trans/transpersoner (i.e. transgender people) (3),

• Young adult (YA) gay (2),

• Intersex (2),

• Asexuality and related word forms (e.g. asexualitet, asexuality, and asexuell) (1),

• Bi/bisexuell (1),

• Butch (1),

• Same-sex (1),

• Transmasculine (1), and

• Två mammor (two moms), indicating rainbow families (1).

All of these terms are already present in some form or variant in Homosaurus and QLIT, indicating that LGBTQI identities are indeed well represented. Of other specific LGBTQI-related foci that should also be covered by the thesaurus, the respondents used the following in their recent searches:

• Love (Swedish, kärlek) (2),

• Romance (2), and

• One mention of each of the following: Antinous (1), bondage, discipline (or domination), sadism (or submission), masochism (BDSM) (1), divorce (1), erotica (1), family (1), gender (1), happy (1), mental illness (1), parenting (1), Pride (1), sexual assault (1), queerecology (1), queer oral history (1), and transformative justice (1).

All these terms except queerecology (used by 1), transformative justice (1), and Antinous (1) are included in the Homosaurus. Queerecology is a perspective of ecology studies and not primarily a focus of fiction and is thus not represented in QLIT; the same holds for transformative justice. Antinous was the lover of Emperor Hadrian, and QLIT does not have specific persons as subject headings. Finally, the following subject terms used by the respondents represent other facets that we believe should be covered by general subject heading compilations such as SAO: anorexia (1), Christmas (1), dark (1), and tense (1). However, terms such as “dark” and “tense,” while possibly describing what users are looking for in fiction, are not normally included in established controlled vocabularies. More common are genre terms found in subject headings, of which our participants used the following: sci-fi (and synonyms such as science fiction) (5), fantasy (3), urban fantasy (1), graphic novels (1), horror (1), and YA (1). Also, one person used the term “Spanish”: since it was a descriptor of a preferred language for fiction, rather than a subject for indexing, it would also be excluded from the thesaurus.

Their latest search: satisfaction with results

When it comes to satisfaction with the results of their latest search, the results were similar to those concerning general satisfaction reported above. On a scale ranging from 1 to 5, where 1 stands for completely unsatisfied and 5 for completely satisfied, most respondents were somewhere in between – neither unsatisfied nor satisfied (point 3 on the scale), chosen by 44 (43.6%), followed by partly satisfied (point 4 on the scale), chosen by 22 (21.8%), with the remainder split between totally satisfied and partly unsatisfied (14, or 13.9%, in each category). Some respondents (7 or 6.9%) were completely unsatisfied. An optional question asked what would have made the search experience better for the respondents; of the 40 replies, most mentioned too few results. Two main reasons were given for this by the participants. One cited reason was a lack of keywords in search systems (across the web) that would represent LGBTQI identities and perspectives (confirming the results reported above). Two comments were about the need for help with searching: these respondents were unsure whether they were using the right keywords or wanted to know how to search better and required general information literacy and user-interface support. A browsing tree could help address these problems, as users would learn what keywords there are and what thematic coverage is as well as how it is structured so they can better find their way. Another cited problem was a general lack of desired literature: literature covering topics other than just romance and love relationships, more positive and mood-lifting books, illustrated books with two mums, and more English books in Swedish libraries (many from Goodreads but not in Libris). Another mentioned problem was the difficulty of identifying self-published works that are normally not acquired or indexed by libraries.

Desirable search terms

Finally, what search words would respondents like to search for? Most were related to LGBTQI identities: trans or transgender (mentioned by 18), including transperson (transgender person) (1), transman (1), and transmasculine (1); lesbian (17), including lesbianism (2), older lesbians (1), and women (1); queer (13), including queer+ (1); non-binary (7); bisexual (7), including bi (3); HBTQ (7), including HBTQI (7), HBTQIA+ (1), LHBTQIA+ (1), per each individual letter (3), and over history (1); intersex (3); homosexual (3), including homo (1) and homosexuality (1); pan (4), including pansexual (2) and pansexuality (1); gay men (2); femme (3); butch (2); agender (1); genderfluid (1); and rainbow family (1).

Similarly, closely related search words capture LGBTQI themes: queer friendship, LGBTQ phobia, coming out as lesbian, lesbian love, lesbian adult love, lesbian young adult love, lesbian erotica, lesbian fiction, lesbian pulp fiction, characters with different sexual preferences, lesbian and children/family, lesbian history, lesbian Sweden, lesbian and selected content, lesbian 1800s/1900s, queer theory, gender theory, same-sex love stories, same-sex parenting, queer parenting, two mums, elderly LGBTQ-people history, same-sex, LGBTQI with warning for sexual violence, LGBTQ canon, coming out, lesbian queer friendship, queer platonic, queer in one's 30–40s, miscarriage, pregnancy, madness, and obsession.

These terms, although not all variants of them, are generally included in the QLIT thesaurus. For example, trans person and transmasculine are included, but not trans man; lesbian (lesbiska) is an included term, but not lesbianism (lesbiskhet). Combinations with more general terms, such as terms pertaining to time periods, geographical areas, and genres, are also searchable in the Queerlit database, since titles in the database are also indexed with SAO terms, in addition to the QLIT terms.

Some terms related to the LGBTQI subject area that were mentioned by the respondents are included in the QLIT thesaurus. Titles in the Queerlit database will also use search terms from SAO.

Several extensive comments point to the need for combined searching:

1. “Lesbian characters, love, poetry, sci-fi … Generally, to be able to search for the words that are usually found in a literary database, but in a database with a focus on literature about LGBTQ + people. So, if I'm looking for ‘sci-fi’ + ‘love’ I will find romantic sci-fi books with LGBTQ + characters.”

2. “To search within genres for specific relationship formations. In the fanfiction archive AO3, you, for instance, have the option of searching a tag, fandom, genre, category, etc., that indicates a romance plot between women. Additionally, I can tag ‘happy ending’ or ‘no death.’ Having a similar option when searching for published work would be helpful.”

3. “The most important thing for me when it comes to large search services/catalogs is the ability to combine searches. I am interested in reading books with LGBTQ characters, but read almost exclusively genre literature, such as science fiction and fantasy. Many pages only allow searches on either genre or theme, so I often have to go into LGBTQ tags/categories and manually find the books that are in the genres I am looking for. It would be much easier if it were possible to combine the search terms as filters.”

Desirable search functionalities

What search functionalities would end-users like to see in online subject searching across web services? While all options could be selected, most respondents (88 or 87.1%) chose the one that would allow the use of related terms for broader searches to get more results (i.e. higher recall). Similarly, many (71 or 70.3%) would also like to see an option to get fewer, more specific results based on narrower terms when too many results are retrieved (i.e. higher precision). More than half (60 or 59.4%) would also appreciate what many search services lack today, likely influenced by Google's simple search-box interface, and that is the option to browse by hierarchically arranged subjects (as one would browse library shelves in a physical library); this is very useful to get an overview of the collection and to find the right search terms. Similarly, about half (48 or 47.5%) also considered the option of word sense disambiguation important in order to get only those results related to the specific meaning of the search term sought. We also learned of one participant's desire to find books by the same author, although on another subject.

Conclusion

This study aimed to investigate the needs of potential end-users of a database dedicated to LGBTQI fictional literature in Sweden, in order to inform the development of the database, search interface functionalities, and an LGBTQI thesaurus for fiction. A web questionnaire was distributed in autumn 2021 to potential end-users. In total, 108 questionnaires were completed by participants who were mostly 18–50 years old and had completed university education. Of those, 101 participants read LGBTQI fiction, 53 doing so several times per year, 28 a few times per month, 16 every week, and four daily. The seven who never read LGBTQI fiction were excluded from further analysis. Most participants chose to read fiction with LGBTQI themes for pleasure in their free time, to find experiences or characters they could identify with, to find subjects and themes they could identify with, and because of their interest in LGBTQI questions. The majority found information about what to read via social media and/or their social networks. The relevant social media were dominated by Instagram, Facebook, and Goodreads. Just under half consulted online library catalogs or online bookstores. Most obtained copies of works from a library and/or bookstore, and the preferred format was a printed book.

Most were neither satisfied nor unsatisfied with their online searching for LGBTQI fiction. Their dissatisfaction seems largely attributable to too few books with LGBTQI keywords, to indexing or indexing systems too shallow or too general to represent LGBTQI identities in sufficient depth, and to a lack of overview of the collections (hierarchical subject browsing would address this). Also, some would like to see more variety in LGBTQI fiction, confirming the previous research reviewed in the “Background” section. It is still hard to find books on specific LGBTQI identities: for example, books with trans themes or characters are harder to find than books about male same-sex relations. Weariness with books with unhappy endings is also evident in the respondents' desire to be able to search for “happy endings” or “no death.” This is evidence that LGBTQI people have been, and in some cases continue to be, fictionally represented as a group vulnerable to tragic experiences.

Most common are searches by author (used by 71 or 70.3%) and title (60 or 59.4%), while half of participants conducted subject searching (51 or 50.5%). The dominance of known-item searching could partly be due to the low-quality subject searching (and indexing) or subject browsing available in many online library catalogs, bookstores, and social media systems.

Of the subject search terms used in the respondents' most recent online search, the most common, either alone or in combination, are terms representing LGBTQI identities, which are already included in both QLIT and Homosaurus in some form or variant, although all variants should of course be supported in the search interface to ensure successful searching. Also common are genre terms found in general subject headings. Terms such as “dark” and “tense,” while possibly describing what users are generally looking for in fiction, are not normally included in established controlled vocabularies; however, this is outside the scope of this project and is a matter for further research.

Users would like to see the following search functionalities: broader, related-term searching to get more results (i.e. higher recall); an option to get fewer, more specific results based on narrower terms when too many results are retrieved (i.e. higher precision); the option to browse by hierarchically arranged subjects, as one would browse library shelves in a physical library; as well as word sense disambiguation to get only those results related to the specific meaning of the search term sought.

That most users do not use libraries to find LGBTQI fiction but instead turn to different kinds of social media implies insufficient library services. There is both a need to develop more precise topical descriptions, and to capture topics with broader terms, which supports the decision to build QLIT as a hierarchical structure that includes broader top terms, narrower and more precise terms, and related terms. The need for alternative wording, reflecting different vocabularies used by the respondents, provides further support for the decision to use alternative terms, or altLabels, in QLIT, thus fulfilling the need to search for a concept using a variety of terms and word forms.

The questionnaire also provides valuable information for the development of the external interface in the Queerlit project. The result shows that interface is an increasingly important tool for people used to commercial interfaces with faceted search options. The respondents want an interface that supports combinations of LGBTQI topics with other search parameters, such as genre and time periods. Furthermore, there is a need to develop interfaces where topical terms can also be easily used by those who do not normally use subject terms in searches in library catalogs, for example by making the terms more visible as topics to choose from and combine them in a search query.Further research entails implementation of the search interface and testing it with potential end-users in follow-up interviews. The QLIT thesaurus will also be continuously updated with relevant terms and term variants or forms. As mentioned earlier, themes in fiction not specifically related to the LGBTQI context should also be better covered by SAO, and work on that is planned.

Funding: This research was supported by Riksbankens jubileumsfond (grant no. IN20-0013).