Subjective aspects in the thematic representation of photographic images

Adriana Aparecida Lemos Torres (Departamento de Organização e Tratamento da Informação, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil)
Benildes C.M.S. Maculan (Departamento de Organização e Tratamento da Informação, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil)
Célia da Consolação Dias (Departamento de Organização e Tratamento da Informação, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil)
Gislene Rodrigues da Silva (Departamento de Organização e Tratamento da Informação, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil)

Collection and Curation

ISSN: 2514-9326

Publication date: 1 October 2018



This paper aims to present the results of an analysis about the subjectivity in the process of thematic representation of photographs.


The experiments were applied to students of the Course of Librarianship (Biblioteconomia) of the Universidade Federal de Minas Gerais, during the discipline “thematic representation of images”. The methodology included two methods of analysis: in free form and using four methodologies: the methods of Panofsky (1979), Smit (1996), Shatford (1986) and Manini (2002).


The results showed that subjectivity is always present, even with the use of methodologies.

Research limitations/implications

This paper shows the importance of doing the librarian course and the need to improve and update their methods, standards, techniques and tools to provide subsidies for the normalization of subject analysis as an essential procedure for information retrieval.

Practical implications

It is concluded that thematic representation of images requires clear and detailed indexation policies that can systematize the activity and minimize the effects of the subjectivity involved.

Social implications

The determination of the subject of the document involves many factors. One of them is the cognition of the indexer, which is influenced by his prior knowledge, limitations and biases about the context. However, it must take into account, above all, the social reality of the user, as the main purpose of the representation is the retrieval of information.


This study contributes to reaffirm the importance of the indexing discipline in vocational training in librarianship courses, emphasizing the cognitive aspects to which the indexer will be subject.



Torres, A., Maculan, B., Dias, C. and Silva, G. (2018), "Subjective aspects in the thematic representation of photographic images", Collection and Curation, Vol. 37 No. 4, pp. 151-157.

Download as .RIS



Emerald Publishing Limited

Copyright © 2018, Emerald Publishing Limited

1. Introduction

The photographs are increasingly present in society, be it in the collections of institutions, in stock of images and even in personal collections. To be retrieved by users, these images need to be thematically represented by the information professional.

Thematic representation (or indexation) is an intellectual process that involves the subjectivity of the indexer, since it depends on several factors such as the audience to be served, the professional’s knowledge, the environment and the context in which the activity is inserted. Among the several factors, Dias and Naves (2007) highlight differences in judgment about the use of the most appropriate terminology and the use of different levels of specificity in representation, a condition that is also confirmed by Rubi (2009).

The same subjectivity mentioned by the authors for indexing textual documents can be observed in the indexing of images, because it is also an intellectual process of analysis and, therefore, it is affected by subjectivity.

To minimize this subjectivity, in the literature, there are studies and discussions about the consistency in indexing, which is the agreement on the terms that are used to represent a document, in the same IR, by two or more indexers, or by a single indexer at different moments of analysis (Dias and Naves, 2007). According to the authors, to avoid a variety of terms for the indexing of the same document, called indexing inconsistency, methodologies must be adopted that can systematize the analysis.

Under the thematic representation of photographs, although there are methodologies to guide the analysis (Panofsky, 1979; Shatford, 1994; Smit, 1996; Manini, 2002), it is considered that even when adopted, the subjective aspects inherent in this process will continue to influence the thematic representation of the photos.

In this sense, this paper briefly discusses subjectivity in indexing and presents the results of two exploratory experiments that aimed to evaluate how subjectivity can influence the indexing of photographs, whether it is done in a free way or with the use of an analysis methodology. After this introduction, the rest of the paper is organized as follows: Section 2 presents a brief theoretical review on the subjective aspects in the thematic representation; Section 3 discusses the methodology and procedures for applying indexing activities, results and discussion; Finally, Section 4 presents the final considerations.

2. Subjective aspects in the thematic representation

Chaumier (1988), p. 74 points out that the thematic representation “is an essential operation so that documents can be retrieved from the documentary collection and then respond adequately and effectively to every request or issue of users”.

Thematic representations process comprises two basic steps:

  1. subject analysis, with technical reading, recognition of the essence (theme) of the document (of what it is about, why it was incorporated in the collection, what aspects will interest the user), and the identification and selection of concepts; and

  2. translation, with the conversion of the concepts selected in the first stage, in natural language, to a controlled language (vocabulary control) adopted by the information retrieval system (IR).

The moment of the subject analysis is the most critical of the thematic representation, because, according to Dias and Naves (2007), this stage is a practice in which perceptive and intellectual senses are present, and the indexer has a subjective capacity of interpretation. In this moment, it is the indexer who offers meaning to the document. For this reason, the thematic representation is a process known to be imbued with subjectivity, due to it being carried out by humans, who employ their prior knowledge in form of techniques to facilitate reading and reach their purpose, which is the identification and selection of concepts contained in a document (Rubi, 2009).

Thus, making the subject analysis in iconographic documents to represent it by terms descriptors requires a great effort by the indexer to remain faithful to the purpose of his work. This effort seems larger with the photographs, since according to Felipe (2016), if in a white screen can be found several interpretations that goes from the void to the opportunity, in a photograph then, in which there are several constituents besides the referent, the personal interpretation is even greater Smit (1996) adds that [...] given the inherent polysemy of the image and the unpredictability in relation to its use, it is still essential to develop totally transparent procedures for the end user, so that the latter, having information on the criteria adopted in their treatment, has the conditions to enjoy of the image representation, without being a hostage. (Smit, 1996, p. 35).

On this issue, Rodrigues (2007) reiterates that photography is a copy of the referent, being a copy of something or someone that is reproduced in the form of an image. The author states that the photographic image is polysemic in nature because it is susceptible of countless meanings. The image has a denotative meaning, according to its literal representation, and also a connotative meaning, which refers to what can be interpreted in a figurative and symbolic sense.

Rodrigues (2007) also emphasizes that photography has a look of who produced it, it has a certain degree of illusion, as the photograph is what the photographer wants to show, according to the photographic techniques used in the composition as angle, framing and lighting. In addition to the look of who produced, the image can be interpreted according to the cultural and ideological values of an individual. Bloomer (1990) ensures that many elements, related to the human being affect how one receives and interprets an image, are ethnicity, culture, nationality, gender, values and social class.

3. Methodology

The two experiments of indexing of photographs were carried out with students from the sixth to the eighth period of the Librarianship Course (Biblioteconomia) of the Universidade Federal de Minas Gerais (UFMG), during the discipline “Thematic Representation of Images”, given in 2017. The methodology procedures were applied from two indexing activities:

  1. Activity 1: Thematic representation of two images of the UFMG institutional collection, of free form, without the use of methodologies.

  2. Activity 2: Thematic representation of three images, of free choice of the student, to be represented from the use of four methodologies:

    • iconological Method of Erwin Panofsky (1979);

    • the categories “Who, where, when, how and what” (Smit, 1996);

    • OF and ABOUT method of Sara Shatford (1986); and

    • expression dimension method of Miriam Manini (2002).

3.1 Activity 1: implementation procedures, results and discussions

The first activity was developed with 11 students, divided into six groups (doubles/trios), who analyzed two photographs of UFMG’s institutional collection without definition of indexing methodology, following the following steps:

  • (1st) subject analysis of the photographic document;

  • (2nd) extraction/assignment of terms; and

  • (3rd) record in the template.

The template was used as support and aimed to gather, in a single record, all the information of the object referenced for later analysis, as shown in Figure 1.

The two photographs used in the activity were presented with subtitles created by the teachers of the discipline. Intentionally, the captions presented only basic image information. However, students were given guidance to explore their prior knowledge and, if they felt it necessary, to obtain more information about the photographs and their contents.

In this task, the students inserted the terms indexers in the six fields of the template reserved for such. It should be noted that the number of template index fields (six fields) was only a reference, and the students were advised to insert as many as possible, if necessary. However, it was noted that the number of terms used by the students varied from one to six, both in the individual analysis and in the group analysis.

First, the 11 students analyzed Photograph 1 (Figure 2).

The activity of analysis of Photograph 1 totaled 28 indexing terms, as shown in Table I.

Among the 28 index terms identified by students, there were predominant terms referring to elements of the composition of the photograph (14 terms, which correspond to 50 per cent of the total), such as architecture, trees, sky, lawn, students, building, garden, monument and clouds. Then, the group of descriptors that related to the institution appeared (5 terms, about 18 per cent of the total), such as Rectory, UFMG, Universidade Federal de Minas Gerais. Then, terms appear (3 terms, about 11 per cent of the total) that refer to the concept of Arts (e.g. Modernism). Finally, terms were used to designate:

  • place/space (e.g. Pampulha Campus);

  • to represent elements of the photographic language (“front view” and “day photo”); and

  • to designate related issues as “first designed building” (two terms of each item, about 7 per cent each, totaling 21 per cent of the total).

The diversity of descriptors occurred in the analysis of Photo 1 indicates the unique look of the indexers in the subject analysis and image representation: only two indexers signaled the idea of time and/or clarity (“photo of day”, which was used by a student, and “image of the rectory on sunny day” and “rectory garden on sunny days”, by another student). It was noticed that only one student indicated the angle of the photograph (“frontal view”); only two students mentioned elements of the image, such as “sky,” “clouds” and “trees,” and only one student mentioned the existence of people in the photograph. Thus, it became evident that the indexer analysis selects the elements that, in his view, are the most important in the content of the photograph and evidence them in terms of indexation. Thus, the image analysis passes through the interpretation and the world knowledge of the indexer.

The phrases used by the indexers (in the “Descriptive Identification” field of the template) to retract the contents of Photograph 1 also point out the different emphases used by each student in the representation. One of the students was very succinct in the description of the photograph: “View of the UFMG Rectory building from lawn”. Another indexer emphasized different elements of photography, including the work of art that appears in the image: “View of the UFMG Rectory building, surrounded by lawn and trees in the background. Lighting posts and monument Aleijadinho in front. Image of the Undated UFMG Collection”. And a third student presented a more historical perspective of the building, signaling a deeper knowledge about this element: “Opened in October 1962, a building inaugurated on the eve of the military government. Neoclassical project for the university city. Architect Eduardo Mendes Guimarães.”

The use of this diversity of phrases and indexing terms made it clear that the subjectivity of the indexer and the polysemic nature of photography are variables that determine the result of thematic representation of this type of document, a fact highlighted in the literature by Rodrigues (2007).

Then, the 11 students analyzed Photograph 2 (Figure 3).

A atividade de indexação da Fotografia 2 totalizou 30 termos indexadores, conforme demonstra a Table II.

Among the 30 indexing terms identified by students, the set of terms referring to the institution was predominant (12 terms, 40 per cent of the total). Next comes the set of terms that refers to people (6 terms, 20 per cent of the total), although they retract the same person. Then, terms referring to elements that:

  • Compete the image, such as a library, teacher, book and chair; and

  • Designate actions/events, such as “book launching” and “autographing” (four terms of each item, about 13 per cent each, totaling 26 per cent).

Next, the terms designating spaces/space, such as “reading space” and “Biblioteca Central” (3 terms, 10 per cent of the total), appear. Finally, the term “communication and articulated technologies” appears, which represents the theme of the book that was being published (1 term, about 4 per cent of the total). It should be emphasized that none of the students used indexing terms in the analysis of Photograph 2 that could represent the category “photographic language”, which occurred in the analysis of Photograph 1. Without using methodology in this activity, no student mentioned elements about the specific techniques of photograph, as the angulation, the framing, the luminosity, according to the proposal of Manini.

The content of Photograph 2 was also represented in sentences, presenting different levels of representation, from general to more specific. One student used the following sentence: “Title: Maria Aparecida Moura assumes the Directorate of Information Governance (Diretoria de Governança Informacional - DGI) of UFMG. Author: Foca Lisboa. Collection: UFMG. Date: 2014”. Another student added elements of the place, without specifying the date: “Image of Foca Lisboa by Maria Aparecida Moura in the reading space of the Biblioteca Central of UFMG assuming the Directorate of Information Governance (DGI). Image without date”. A third student has already presented a more detailed level of information: “Maria Aparecida Moura with the copy of the book that talks about the public access of the information: Communication and Articulated Technologies, Maria Aparecia Moura assumes the Directorate of Information Governance (Diretoria de Governança Informacional – DGI) at UFMG (Foca Lisboa). Department created on April 29, 2014 “.It is noteworthy that only one student represented the two events that are retracted in Photograph 2:

  1. The inauguration of the board.

  2. The launch of the book by Professor Maria Aparecida Moura.

Although the caption of the photograph only retracted the Professor’s tenure as director of the Directorate of Information Governance (Diretoria de Governança Informacional - DGI), the name of the book in the photograph would be an important starting point for the research on the event.

Another example that demonstrates the subjectivity of indexers was the variation of the name of Professor Maria Aparecida Moura in the representation. Some students used the term “Cida Moura”, in which way the teacher is treated by the community of the School of Information Science (Escola de Ciência da Informação), including the students of the Librarianship Course (Biblioteconomia), which is the case with indexers.

At the end of Activity 1, it was noticed that the subjectivity in the thematic representation was manifested in the indexation of the two photographs. Because natural language was used, where there were a multiplicity of terms for students to use in the indexing of photographs, it can be observed that many of them were repeated, others were synonyms and some were related by associations or hierarchies.

3.2 Activity 2: Implementation procedures, results and discussions

The second activity of thematic representation of photographs was performed with a group of six students. Each student analyzed three images of their free choice, based on the following criteria: a self-produced photograph, one from an internet image bank and another from their personal or family collection. In this activity, four methodologies were used:

  1. Iconological method of Erwin Panofsky (1979): this method is composed of three levels: pre-iconographic level, which represents what you see in the photo; the iconographic level, which requires a little more knowledge and cultural understanding; and the iconological level, which demands an even greater historical, social and cultural knowledge.

  2. The categories of Smit (1996): the categories “Who, where, when, how and what” encompass core issues for document representation in general, and can also be applied to images.

  3. Shatford’s (1986) method: the levels of this method - OF (generic and specific) and ABOUT – approximate Panofsky’s levels of representation: GENERIC DEGREES relate to pre-iconographic level and OF SPECIFIC at the iconographic level; the ABOUT is related to the iconological level.

  4. Manini’s (2002) Expressive Dimension Method: This method encompasses the way the image is shown through specific photography techniques, such as angulation, framing, luminosity.

The students applied these four methodologies in the three photographs selected from the established criteria and were instructed to describe the facilities and difficulties in the activity. The students’ reports on the application of the different methods in the representation of the same photograph demonstrate particular difficulties of each indexer in relation to the different methods.

For some, it was easier to index according to Panofsky’s methodology, while others declared greater difficulty in using the same method, for example. In this aspect, we also noticed the subjective aspects involved in the thematic representation, even in the choice or preference of the methods. It was evidenced that the perception and knowledge of each student/indexer favor the choice of methods for the representation of photographs, as indicated by one of the indexers:

I had greater difficulty in indexing the Panofsky and Manini methods. The first one requires greater knowledge during a symbolic analysis of the image and the last one due to lack of knowledge in the field of Photography. I had an easy time with Shatford’s method and found the proposal of representing the meaning of the objects photographed more palpable and adequate. The indexing according to the categories “who, when, where, how and what” is interesting in the description of the characteristics of the image, but leaves in the directions that can be inferred from the same. Shatford’s method mixes the description of the information contents of the photos with the symbolic content of the image, particularly what I liked the most because it represents the image in a satisfactory way without requiring too much time from the indexer. Panofsky’s method was the most difficult, because it required a greater technical knowledge of the goal of image production, that is, it required a symbolic analysis not only of the objects of the photo, but of what they meant for a society with a certain culture and in a given context. The Manini method is based on the description of the photographic technique used in the production of the photograph, these data support the user to create the criteria of selection of an image, that is, if he comes across several images that represent the same subject he can choose, for example, the one with the best angle, or framing, or lighting. I do not see much use of the method for generalist institutions, but rather for more specialized image bank users. (INDEXER 6)

It is noticed, by the report, that beyond the knowledge about the subject retracted, thematic representation of photographs requires a more substantial knowledge about the photographic technique.

Student/indexer 1 applied all the methodologies proposed in the activity and one of the photographs was produced by himself and retracted his parents’ house, as shown in Figure 4.

Regarding the image representation experience, Student/Indexer 1 made the following report:

The image had an easy indexing, because it was a photo chosen by me (personal collection) and produced by me, and because it has a very great sentimental value, it brought me homesickness of my land. It was such a thrill to have found this photo in my collection to index it. The indexation was easy to analyze because it was a personal photo collection, where I was the producer of the photo and I know the place in detail. It is easy to index because it contains sentimental and expressive value. (INDEXER 1)

From the report, we can see a strong link between the student/indexer and the analyzed photograph. If, on the one hand, the student considered the indexation easier, due to his knowledge of the topic retracted, on the other hand, evidences the condition of a high degree of subjectivity at the moment of representation, once it brought to the student/indexer memories that perhaps were not, in fact, retracted in the photo, giving it its own values that could create a bias in the representation. This condition confirms the literature of the area that emphasizes that image indexing is an interpretation of the indexer that uses its own cultural and ideological values.

This issue of subjectivity is also explicit in the student’s report/Indexer 5:

All the techniques have a level of difficulty and, more importantly, a level of subjectivity, in some more than others, in the personal photo is even more complicated, because it is an image with great particular significance some information only the photo of who belongs who will be able to identify, when the photo is more generic is more difficult to stick to great details, because the concentration is in more superficial elements. (INDEXER 5)

In the report, it is perceived that the student’s opinion about the difference in indexing images from the personal context and images from other contexts, highlighting the issue of subjectivity in the analysis process and the cognitive and external factors that interfere with it is well understood.

It should be emphasized that the knowledge about the subject retracted in the photo is a determining factor, as it can alter the thematic representation of images. In this sense, a professional/indexer who is more knowledgeable about the content of a photo, or who is an expert in the subject, will be more able to be exhaustive or use coherent terms at the time of indexing. In addition, knowledge about the photographic technique will also be differential.

From the analysis of the two activities, it can be evidenced that several aspects related to subjectivity can influence the thematic representation of photography, as it is capable of having several meanings and interpretations. Thus, to soften this personal interpretation of the indexer, it is necessary to establish indexation policies and criteria to systematize the thematic representation of photographic images, especially as photography has its characteristic to be polysemic.

4. Final considerations

The results confirm the proposition that the thematic representation is influenced by the indexer’s individual cognitive aspect, prior knowledge about the image and their theme, as well as his vocabulary.

In the thematic representation without the use of methodology, the use of informal terms was verified, even when indexing institutional terms or people. It was verified that the ease or difficulty of making a very specific representation is directly related to the degree of knowledge about the content of the photographs of the indexer.

As the photographic image is polysemic in nature and susceptible to many types of comprehension, it is more subject to the subjectivity of the indexer. It was possible to verify that the subjectivity interfered in the results of the thematic representation, because the image analysis process involves an interaction between the information expressed in it and the indexer’s understanding of it. It is this understanding that will allow to elaborate the representation of the image content. This shows that thematic representation of images requires indexing policies that provide detailed criteria for the systematization of the activity, aiming at objectifying it and minimizing the subjectivity involved in it.


Template for indexing in Activity 1

Figure 1

Template for indexing in Activity 1

Photograph of the Activity 1

Figure 2

Photograph of the Activity 1

Photograph 2 of Activity 1

Figure 3

Photograph 2 of Activity 1

Indexing with methodologies according to Indexer 1

Figure 4

Indexing with methodologies according to Indexer 1

Terms and categories of the indexing of the Photograph 1

Category Terms that compete the category Quantity of
of occurrences
Elements of photography composition Architecture
Lawn (2)
Image of students in front of the Rectory
Image of architectural work in front of the Rectory
Image of UFMG’s Rectory Building
Rectory Garden
Rectory garden in sunny day
Rectory’s Monument
Clouds (2)
Rectory building
Rectory’s lawn
14 16
Institution Image of Rectory on sunny day
Rectory (7)
UFMG (4)
UFMG Pampulha Campus
Universidade Federal de Minas Gerais
05 14
Art Tribute to Aleijadinho
Orlando Magalhães Carvalho
03 03
Location/space Pampulha Campus
University Spaces
02 02
History First building designed
Building, responsible, education and chemistry
02 02
Photographic language Photograph of day
Front view
02 02

Terms and categories of Photograph indexing 2

Categories Terms that compete the category Quantity
Of terms
Quantity ofoccurrences
Location/space Reading space of Biblioteca Central da UFMG
Reading Space
Biblioteca Central
03 03
Elements of photography composition Library
Book Social construction of public access to information in Brazil
Padded chairs
04 04
Institution UFMG (3)
Board of directors (2)
Directorate of Information Governance (3)
Directorate of Information Governance UFMG
Directorate of Information Governance DGI
Information Governance
DGI (2)
Coordination of Information Inclusion Policies CPInfo
Universidade Federal de Minas Gerais
Escola de Ciência da Informação ECI
Teachers UFMG
12 18
Ações/eventos Autograph
Book release
Cida Moura giving autographs
Image of Maria Aparecida Moura signing books
04 04
People Maria Aparecida Moura (6)
Cida Moura (3)
Professor Cida Moura
Image Maria Aparecida Moura UFMG
Director of Information Governance UFMG
Image Cida Moura at DGI
06 13
Interpretive terms Articulated communication and technology 01 01
Photographic language 00 00


Bloomer, C.M. (1990), Principles of Visual Perception, 2 ed., Van Nostrand Reinhold, New York, NY.

Chaumier, J. (1988), “Indexação: conceito, etapas, instrumentos”, Revista Brasileira de Biblioteconomia e Documentação, Vol. 21 Nos 1/2, pp. 63-79.

Dias, E.W. and Naves, M.M.L. (2007), Análise de Assunto: teoria e Prática, Thesaurus, Brasília.

Felipe, C.B.M. (2016), “Os aspectos sociocognitivos Para a indexação de fotografias”, Ms dis., Universidade Federal de Pernambuco.

Manini, M.P. (2002), “Análise documentária de fotografias: um referencial de leitura de imagens fotográficas Para fins documentários”, PhD thes, Universidade de São Paulo, São Paulo.

Panofsky, E. (1979), Significado nas Artes Visuais, Ed. Perspectiva, São Paulo.

Rodrigues, R.C. (2007), “Análise e tematização da imagem fotográfica”, Ciência da Informação, Vol. 36 No. 3, pp. 67-76.

Rubi, M.P. (2009), “Os princípios da política de indexação na análise de assunto Para a catalogação: especificidade, exaustividade, revocação e precisão na perspectiva dos catalogadores e usuários”, in Mariângela Spotti Lopes, F (org), A Indexação de Livros: A Percepção de Catalogadores e Usuários de Bibliotecas Universitárias, Cultura Acadêmica Editora, São Paulo, pp. 81-94.

Shatford, S. (1986), “Analysing the subject of a picture: a theoretical approach”, Cataloging and Classification Quartely, Vol. 6 No. 3, pp. 39-62.

Shatford, S. (1994), “Some issues in the indexing of images”, Journal of the American Society for Information Science, Vol. 45 No. 8, pp. 583-588.

Smit, J. (1996), “A representação da imagem”, Informare: Cadernos do Programa de Pós-Graduação em Ciência da Informação, Vol. 2 No. 2, pp. 28-36.

Further reading

Kobashi, N.Y. (1994), “A elaboração de informações documentárias: em busca de uma metodologia”, PhD thes., Universidade de São Paulo, São Paulo.

Panofsky, E. (1986), Estudos de Iconologia: temas Humanísticos na Arte Do Renascimento, Estampa, Lisboa.


The authors thank the agencies Fundação de Amparo à Pesquisa em Minas Gerais (FAPEMIG), to Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) e to Pró-Reitoria de Pós-Graduação (PRPG) of UFMG for their research support.

Corresponding author

Benildes C.M.S. Maculan can be contacted at: