1The term social epistemology, as highlighted by Bouvier and Conein (2007), corresponds to a theory about the social dimension of knowledge, which considers that what we know stems not only from direct sources but also from indirect sources whose reliability derives from a notion of socially constructed or agreed authority. Thus, social epistemology features a contextual, critical-interpretative nature and, as a result, is strongly permeated by language, and an object of interest for different disciplines in Social Sciences such as Economics, Anthropology, Social Psychology and Sociology, among others. As Hjorland (2018: 636) points out, “social epistemological theories provide a better foundation for indexing compared to individualist epistemological theories”, especially because “indexing is always influenced by socio-cultural contexts and that these contexts need to be considered in indexing practice and theory, and that criteria for classification and indexing must be found in pragmatic criteria”.
2Based heavily on the so-called “indirect sources of knowledge” such as testimonies, social epistemology attaches special emphasis to arguments of authority, as well as to relationships of trust and epistemic dependence. It also focuses on the ways in which these arguments are constructed, and the dynamics that underlie them, which is why it establishes significant dialogues with argumentation theory (Hardvig, 1985; Goldman, 1994, 1999).
3Given its contextual nature in time and space, social epistemology focuses special attention on interactive relationships (such as those of cooperation and collaboration), similarly to scientific environment, addressing issues related to mutual collaborations, group knowledge and to epistemic communities, the latter strongly associated with the social division of labor (Nelson, 1993; Goldman, 2004).
4Defined by Schmitt (1994: 1) as “the conceptual and normative study of the relevance of social relations, roles, interests, and institutions to knowledge”, social epistemology analyzes the role of social positions in accessing knowledge from the study of everyday ways of thinking in societies as well as the epistemological norms established among social actors.
5Social epistemology, therefore, is opposed to an individualist approach to the theory of knowledge, as it starts from the idea of knowledge as a social product in which different subjects act in relations of epistemic dependence, whether among individuals or among groups or institutions (Bouvier & Conein, 2007).
6More specifically in the field of Information Science, social epistemology was firstly explored by Margaret Egan and Jesse Shera (Egan & Shera, 1952), prominent supporters and protagonists of the Chicago School (Richardson, 1982). They argued that all information activities and, in particular, the documentary processes, must be permeated by the study of how society articulates itself to generate knowledge. As these authors point out, social epistemology consists of “the study of intellectual life from that of the individual to an inquiry into the means by which society, nation, or culture achieves an understanding relationship with the totality of the environment, and its focus will be upon the production, flow, integration, and consumption of all forms of communication throughout the entire social pattern. (Egan & Shera, 1952: 132). From this conception, more concrete questions emerge, such as: “how a particular text comes to be regarded or selected as a piece of knowledge?”; and “how, given a set of historical circumstances and interpretations a given text is thought to be knowledge?” (Froehlich, 1989: 308-309). To do so, it primarily considers contexts in such a way that users come to be understood as representatives of different cultures and knowledge domains which, in turn, are permeated by values and idiosyncrasies.
7And it is precisely in this contextual dimension that themes such as domain analysis (Hjorland & Albrechtsen, 1995; Hjorland, 2002, 2017; Tennis, 2003; Smiraglia, 2012), metatheory (Ritzer, 2001; Tennis, 2008; Tognoli, 2013); and epistemic communities (Holzer & Marx, 1979; Haas, 1989, 1992; Adler & Haas, 1992; Meyer & Molineux-Hodgeson, 2010; Evangelista, 2021) take place within knowledge organization and representation.
8This contextual nature, in turn, presupposes, an organic approach, since a given piece of knowledge to be represented stems from a set of documentary instances that, if considered separately, no longer make sense for the context that generated them. As highlighted by Froehlich (1989: 308) “... that whatever we call knowledge comes out of a cultural, historical horizon and goes back into a cultural, historical horizon”.
9In the context of knowledge organization and representation, the analysis and indexing of films constitutes, within the scope of the so-called “non-book resources”, one of the most complex challenges for the organization of film collections and film information retrieval. This challenge is due to the dynamic visual and sound elements, thematic content, and various narrative, descriptive, and argumentative structures that have a strong semiotic nature, and it applies to all types of film collections including cinematheque collections, personal archives, and libraries.
- 1 It is important to highlight that the term "documentation", for the purposes of this paper, is not (...)
10Thus, the present paper aims to discuss indexing perspectives of film documentation1, from the perspective of social epistemology, utilizing an experience with collections related to Brazilian Cinema Novo.
11The eminently organic nature of the film activity, whereby the film itself, as a document, encapsulates the entirety of the documentary context that precedes, accompanies and succeeds it, and without which the indexing process risks oversimplification and reductionism.
12Thus, four aspects deserve special attention in the film indexing process, so that one can contemplate this organic relationship established in this documentary context, as shown below. The first aspect lies in the process of producing the film and the documents generated during this process; the second is due to the investigative experience of film indexing through facet analysis; the third is related to the identification of which access points are most present on certain websites for the retrieval of digital films; and the fourth and final aspect examines whether any evidence of indexing principles is present in the tagging of digital objects.
13This film-making process allows to contextualize and anticipate evidence in relation to producing meaning in a film, in the context of dynamics of indexing considering the film’s family tree. Thus, it is necessary to study the documents produced for and by the film to be able to detect, through them, information that gives the indexer parameters for the indexer-analysis, regarding the generating-documents, and also some aspects of the film. It must be considered, however, that the indexing of documents related to the film does not replace the indexing of the film itself - since the object film brings together elements that characterize it as a unique element in its audiovisual narrative grammar - but significantly complements its meaning and its contextualization (Cordeiro, 2000).
- 2 When we approach cinematographic archives, some sensitive differences arise regarding their constit (...)
14In this integrative perspective, the collective work is contemplated in which several activities, phases and procedures take place in parallel and simultaneously by different professionals to make the film in such a way that the documentary records show the planning, development and dissemination of the cinematographic work. These various activities (pre-production, production and post-production) result in a significant number of technical documents (technical analysis of the script, storyboard, schedules, shooting plans, camera and sound bulletins, photographs and many others) of different natures and supports, such as, for example, photographs of scenes and everything else that is generated for the dissemination of the film. In short, there is an organic principle of indexing that considers the rescue of the film’s family tree2.
15This research experience was carried out in the collections of two Brazilian Cinema Novo filmmakers – Nelson Pereira dos Santos, and Helena Solberg. It revealed that film documentation has an intrinsic and more complete organicity in comparison with archival documents, as it incorporates the elements that precede and follow the film. Thus, indexing serves not only to information search and retrieve, but also as an element for building memory that is portable in space and permanent in time (Smit & Barreto, 2002).
16It is worth noting that the collections of these two filmmakers are representative of a generation of artists involved with the Brazilian Cinema Novo movement. The movement influenced cinematographic cultural thought from the 1960s onwards. Despite various technical obstacles to filmmaking, these artists sought to produce works that critically reflected the reality of the time, taking into account its various social, political and cultural aspects. Thus, a Brazilian cinematographic aesthetic and language was sought, initially strongly influenced by Italian neo-realism and, consequently, by the Nouvelle Vague” and, more closely, by the Brazilian independent cinema of the 1950s” (Carvalho, 2006: 290).
17For Jean-Claude Bernardet (1978), Cinema Novo represented a “cultural avant-garde”. As pointed out by Carvalho (2006: 290), it sought to “answer fundamental questions Brazilian cinema faced at that time: what should Brazilian cinema express and how could it be done without equipment, funding or an exhibition circuit. Consequently, the films were permeated by a significant dose of radicalism and violence, with aggressive images and themes that mirrored the turbulent years of the 1960s in a developing country. Such aspects, employed as a “creative strategy, came to define the general characteristics of Cinema Novo, whose emergence is associated with a new way of experiencing life and cinema [...]. (Carvalho, 2006: 290).
18In this context of Cinema Novo’s history, mostly featuring male voices, it is essential to mention female voices such as Helena Solberg and Norma Bahia Pontes, among others. Specifically considering the work of filmmaker Helena Solberg, the longest-serving Brazilian film director with a 57-year-long career, the knowledge organization approach begins with an analysis and representation of her personal archives. These archives include both the documentation generated for and by her cinematographic works and records pertinent to her private life, which provide valuable context for interpreting and understanding her production. In this case, we examine work activities in the singularities of the various actions involved with cinema, specifically film production, and resulting document records, as well as the impact of personal life on professional activities. (Cordeiro & Tedesco, 2023).
19To this end, the family tree methodology offers a significant contribution to the organization, archival treatment, and access to personal records in a rich and complex collection of texts, images, and audiovisual materials that narrate the history of Brazilian cinema along an international trajectory. This narrative is based on four critical phases of the filmmaker’s production: feminism, politics, rapprochement with Brazil, and within Brazil. (Cordeiro & Tedesco, 2023).
20When approaching the cinematographic archives, some sensitive differences are presented regarding their constitution, documentary types and custody unit. These documents form collections in personal archives or institutional archives. However, a common and constantly observed aspect resides in the fragmentation and absence of documents referring to the process of viability, production and post-production of a film.
21The concepts and methodologies of Archival Science, like those in all sciences, have been developed with a supposed universal subject in mind. However, we have long understood that this universal subject embodies class, gender, race, sexual orientation, and geographic location. While the tradition that predates its constitution as a science is crucial and should not be disregarded, it is pertinent to highlight the discipline’s challenges in addressing matters outside the public sphere.
22The second aspect involves creating a set of facets. These facets include elements such as gender; temporal frame record; time hook; historical reference; represented themes; narrative structure; nature of the representation; relevant sequences; thematic approximations of the chosen sequence(s); and synopsis. The conceptualization and inclusion of information in conceptual fields were specially influenced by film theory. (Cordeiro & Amâncio, 2005).
23In this sense, the research experience was conducted applying the facets to a set of 100 national and international fiction films. In the following phase, the same facets were used to index 16 films by filmmaker Nelson Pereira dos Santos. We decided to use the cinematographic work of feature films scripted and/or directed by the filmmaker influenced by the following factors for selecting the sample films: the fact that the filmmaker’s production is vast and heterogeneous in terms of genre, dramaturgy and themes and they portray the production of Brazilian cinema in different periods. For example, regarding the properly cinematographic aspects of the filmmaker’s extensive work: its relationship with literature/adaptations (Vidas Secas, Memórias do Cárcere); with ethnography and history (Como Era Gostoso Meu Francês), with ethnic issues (Tenda dos Milagres), with the Latin American melodramatic tradition (Cinema de Lágrimas) (Cordeiro & Amancio, 2005).
24The categories established for indexing fiction films have proven to be viable and consistent. They provide a reliable framework for film analysis, enabling efficient retrieval of content in information services. Such a proposal could be structured into a computerized model suitable for a web-based film information service environment.
25The third aspect we mention is related to the identification of which access points are most present in certain websites for digital film retrieval. For this purpose, three scenarios are especially important: a) institutions focused on custody, memory and preservation (such as libraries, archives, cinematheques); b) service-focused vendors (i.e., streaming service platforms; c) sharing-focused services (e.g., IMDB). The guidelines provided to users regarding the use of keywords, controlled vocabularies, criteria applied in the architecture and categorization of items or access points, and in the elaboration of synopses were also analyzed on these websites. (La Barre & Cordeiro, 2016). It is noteworthy that this configuration in scenarios has a pedagogical aspect for analyzing and discussing the theme, given the vast visual and audiovisual information universe on the web, with its unique particularities.
26With regard to the scenarios analyzed on web-based film environments, the distinct institutional objectives of the three studied scenarios are reaffirmed. The first scenario illustrates an institutions’ concern with maintaining collections for the preservation of a country’s filmographic memory. Special care is given to the description and representation of item in the collection. The second and third scenarios expose companies focused on the commercialization of services and products, with their survival depending on who uses them. It is noteworthy that the survival of a service such as YouTube is dependent on the content creator, who both produces and consumes the information (videos). While it is true that many of the accessed videos are certainly not watched in their entirety, it is curious to see how Netflix, reliant on its subscribers to maintain the service, robustly attempts to learn and expand the interest profile of its customers.
27The most present access points in the surveyed websites were identified as: summary, date, title, associated persons, genre, rights, color, format, keyword and subject. The examined sites were IMDB (US/UK+), Moving Image Archive (US), Porta Curtas (Brazil), BBC Motion Gallery (UK) (La Barre & Cordeiro, 2012). However, such access points can be expanded to ensure greater representation of the studied universe, based on the following facets:
-
Action: movement or events that propel the story forward.
-
Associated person: persons associated with film director, actor, cinematographer, etc.
-
Character: performer of action in film.
-
Commentary: audience narrative of a qualitative nature “I loved this film.” Features include intensity of discourse and identification with or distance from plot.
-
Emotion: descriptions of emotive aspects of film.
-
Feature: aspect of film that is not related to genre such as subtitles or special formats availability (streaming).
-
Genre: class or type of film that shares common, predictable or distinctive artistic and thematic elements.
-
Object: props or other aspects of the mise-en-scène
-
Relation: connections between film and other entities such as books.
-
Time: indications of chronology in movie, or when film was produced (La Barre & Cordeiro, 2012: 248).
- 3 Considering tags as subject access points, then we can “evaluate which access points are important (...)
28The fourth and final aspect entails the practice of naming “digital cultural objects” (Foster & Rafferty, 2016), which involves examining whether indexing principles are present in how digital objects are named by tags, such as in content publications on entertainment websites and social networks. At the same time, it is worth considering the cognitive strategies netizens apply in this process of digital objects representation (tagging)3. This aspect agrees with current concerns with tagging on the web and discussions that compare professional indexing with social tagging (Tennis, 2006; Kipp, 2011; Peters et al., 2011).
29Regarding the various aspects of tagging audiovisual information, Jörgensen (2016) underscores that; a) user-assigned tags vary greatly in word formation, description levels and semantic meaning; b) image search is most commonly conducted at the basic level of everyday language (such as “apple”), which is neither the most abstract nor general term (such as “fruit”); c) the lack of a standardized tagging policy among organizations poses a challenge for both researchers and internet users; d) the organization or communication of “tagging”, whether for oneself or for others (family and friends, audience)- are motivating factors for taggers; and e) tags fulfill multiple functions for internet users.
30In this tagging trend, there has been a noticeable surge in studies related to emotional indexing, considering that the Internet environment has encouraged the free assignment of words to information items, with immediate participation of Internet users. Rafferty (2016: 10) points out that “the movement to digitize, collect, access and produce cultural documents on and through the web has led to a greater interest in developing approaches to indexing or ’tagging’ and emotion seeking”.
31Given the above, we have observed that research is advancing in understanding the practices of tagging and searching for images and audiovisual content by emotion (distinguishing the emotions present in the informational resource and those evoked by the user), and the motivations and intentions of the users in engaging with institutional collections of “digital cultural objects”.
32The research experience with representative film collections from the Brazilian Cinema Novo movement, particularly given its strong ties to social issues, revealed that the analysis and representation of this documentary universe requires: a) acknowledging the contextual and organic nature of this universe in such a way that the processes of knowledge organization and representation must extend beyond the product itself – the film – to encompass all related documentation, which precedes, accompanies and succeeds it, often from personal and institutional archives. Without accounting for such organicity, the representation suffers from impoverishment due to its incompleteness.
33A second aspect to highlight, especially concerning the facets of representing knowledge that materializes as a social product, is the need for these to reflect not only the established epistemic dependencies between subjects throughout this process but also that they were documented in different ways and at different times.