Chapter 4

Model of Description and Task of Analysis

 

4.1. Introduction

Having discussed a concrete example of a model of content description* applied to a corpus of audiovisual texts* in Chapter 3, in this chapter we shall evoke more general consequences relating, on the one hand, to the hierarchical and syntagmatic structure of any model of description and, on the other hand, to the analytical task devoted to the textual object.

In section 4.2, we shall very briefly discuss the main constituents of a model of description.

In section 4.3, we shall give a general overview of the canonic syntagmatic order of a model of description which, at the same time, determines every concrete analytical scenario. By “syntagmatic order”, we mean the structure of succession of the main sequences according to which the process of analyzing an audiovisual text is carried out.

In section 4.4, we give a more detailed discussion of the distinction – which, to us, appears crucial – between four hierarchical levels for correctly dealing with and modeling the task of analysis. The distinction to which we refer is drawn between:

i) the type of analysis;

ii) the task of analysis making up a type of analysis;

iii) the procedure of description according to which a task of analysis is carried out; and finally,

iv) the concrete activity or activities of description which make up a procedure of description.

Finally, in section 4.5, we shall give a more particular account of the type of analysis called content analysis (of an audiovisual text or corpus) – a type of analysis which will serve us as a reference in this book.

4.2. The structural organization of a model of audiovisual content description

A model of description of the audiovisual content has a structure, a canonic structural organization which conditions and facilitates the construction of new models and local modifications to pre-existing models. This canonic organization is shown in Figure 4.1, and is based around 6 central components:

1) All models of thematic description are made up of one or a small set of thematically and functionally specialized fields of analysis, including, for instance, the field of identification and explicitation of a knowledge object thematized in the audiovisual text, the field of geographical and historical contextualization of the object being thematized, the field of discourse production from the object being thematized, etc. (for further information, see Chapter 3).

2) A model of description is made up of conceptual terms (concepts), part of the meta-lexicon of conceptual terms which forms an essential component of the metalinguistic system upon which all models of description are ultimately founded (in our case, the metalinguistic system known as “ASW”, standing for Audiovisual Semiotic Workshop). We shall distinguish between two meta-lexicons (two conceptual vocabularies): a) the meta-lexicon containing the conceptual terms which represent the objects of analysis of the universe of discourse* ASW, and b) the meta-lexicon containing the conceptual terms which represent the procedures of analysis of the textual object.¹

3) A model of description is a hierarchical structure within which we can distinguish different levels of grouping and integration of conceptual terms. In our approach, we distinguish two levels – the level of the schemas* of definition of the conceptual terms, and the level of the sequences* of description of the conceptual terms: sequence and schema are defined in three ways: as thematic units (they “take care of” a specific aspect of the object of analysis), functional units (they form part of one or other of the fields of analysis identified in Chapter 3) and finally, hierarchical units (the two units form part of two different levels of integration).²

4) A model of description is equipped with analytical tools for each conceptual term and/or each group of terms including, e.g.:

– tools of indexation per se of a knowledge object thematized in an audiovisual text in the form, e.g. of its designation, its denomination, its identification in the guise of keywords, and so on;

– tools of textual annotation of an object being analyzed;

– extremely important tools of referencing an object to be analyzed using thesauruses, terminologies or standards;

– tools of geographical and chronological referencing of a knowledge object thematized in an audiovisual text.

These tools are brought together in the form of a library of schemas of indexation* which we shall present in greater detail in Chapter 14.

5) The conceptual terms defining, on the one hand, the object of analysis and, on the other, the procedures of analysis, form a configuration, i.e. an internal structure of reciprocal positioning and hierarchical integration into a “whole” which is the model of description*.

6) Finally, a model of description is equipped with a set of instructions which form its rhetorical level and which aid and guide the analyst in his task of analysis of the content of an audiovisual text.

4.3. The canonic syntagmatic order of a form of description

A form of description of an audiovisual text (or a specific part thereof) always begins with the sequence (or sometimes, sequences) of description reserved for the referential description* of the knowledge object thematized in the audiovisual text (see Figure 4.2). In our example developed in Chapter 3, the analyst is first invited to describe the civilization spoken about in the audiovisual text being analyzed, and then to specify what type of cultural formation is in question. As Figure 3.2 (Chapter 3) shows, in our particular case, it is a question of the Chavín civilization (a Native South American civilization which reached its peak around the last few centuries B.C.) and its highly remarkable material culture. Here, we are using a relatively simple model of description, but one which demonstrates the huge advantage of using an explicit metalanguage (a descriptive ontology) which facilitates all sorts of practical applications, discussed in [STO 11a; STO 11b]. Later on, we shall see that this basic referential model of description can give way, if necessary or desired, to far more complex models of description but which also require very specialized analytical skills and more time to carry out the analysis.

The referential sequence of description is always followed by the sequence(s) reserved for the description of the space-time context of the domain of knowledge being dealt with in an audiovisual text (see Figure 4.2). As a general rule, the sequence(s) of spatial localization (geographic, geopolitical, etc.) come before the sequence(s) of temporal localization (historic, etc.).

Going back to our example, developed in Chapter 3, it indeed includes two sequences of geographic localization. The first (Figure 3.3) invites the analyst to provide information relating to the global location (of the Chavín civilization), i.e. relating to the region in question on the American continent. In our particular case, this is the <Andes Cordillera> in <South America>. If applicable, the analyst can specify a more circumscribed geographical territory in the region identified. In our case, the analyst specifies that we are talking about the <Cordillera Blanca>. In addition, he can classify the geographical territory in question using a specialized thesaurus. In our case, he specifies that the Cordillera Blanca is a territory of the type <Altiplano> (high plateau).

Once the information relating to the geographical location has been furnished, the analyst can also provide information of a geopolitical and administrative nature: about the country in question, the province (district, canton, etc. depending on the terminology used), the commune, etc. (see Chapter 3, Figure 3.4). In our case, the analyst adds that the administrative-territorial framework needed to locate the subject in question is the village <Chavín de Huantar> situated in the province of <Ancash> in <Peru>.

It should be underlined that the analyst is free to add any information necessary for a spatial (geographic and/or administrative-territorial) localization of the subject. In order to record a model of description in the ASW system’s database of descriptions [LEM 11a], the only sequence which must be filled in is that reserved for referential description.

Temporal localization, in turn, can be reduced to as simple a sequence as that shown in Figure 3.5. This sequence invites the analyst to provide a temporal localization (of the Chavín civilization) in terms of centuries. In our case, it is a period situated roughly between the 7^th and 4^th Centuries B.C. These centuries are identified (as Figure 3.5 shows) using a microthesaurus specialized in the enumeration of periods of (human) history in centuries. If the centuries are identified to localize a knowledge object form a recognized era, the analyst can then provide the name of that era (see Figure 3.5). There is still a more developed library of sequences of temporal and historical localization which the analyst can use if need be (we shall come back to this later, in Chapter 5).

As Figure 4.2 shows, in third position after the sequences for the referential description and the space-time context, we find the sequence(s) reserved for discourse analysis of the domain of knowledge in question. To return to our example developed in Chapter 3, this analysis is concentrated on the two questions of discursive thematization and the authorial point of view which prevails in the treatment of the thematized knowledge object (see Figure 3.6). Here, the analyst specifies that it is (above all) the historical origins of Chavín technical culture between the 7^th and 4^th Centuries B.C. which are dealt with in the audiovisual text being analyzed. He also adds that this question of the historical origins is dealt with by the author both in the form of references to his own research and of references to the research of other authors. The specialized micro-thesauruses* made available to the analyst to explicitize the discourse production from the specific topic <technical culture of the Chavín civilization between the 7^th and 4^th centuries B.C.> constitute an attempt to classify the ways in which to approach and process (mediatize) the objects of the domain of knowledge to which the audiovisual text being analyzed refers to, through and in the discourse. We shall come back to this point in Chapter 15, which is dedicated to presenting the ASW thesaurus*.

The fourth position is occupied, as Figure 4.2 shows, by the sequence(s) reserved for the description of the audiovisual and/or verbal mise-en-scène (the expression) of the subject dealt with in an audiovisual text. This option is not manifest in the form we are using as an example, i.e. the form serving as an interface for describing audiovisual texts thematizing the domain of civilizations and their cultures on American soil.

Finally, as shown in Figure 4.2, the fifth and final position is filled by the sequence(s) reserved for expressing the analyst’s own point of view, his interpretation, his expertise. In our example (Figure 3.6), this is a very simple sequence which is reduced to the option offered to the analyst to add to his analysis of the domain of knowledge <Technical culture of the Chavín civilization> in the form of comments, additional information, etc. This syntagmatic position can be enriched by a whole library of more specialized sequences enabling the analyst to produce almost any kind of statements, comments and interpretations.

An important point which we have not yet mentioned relates to bringing the analysis of a specific topic (such as that of Chavín technical culture dealt with in this-or-that segment of a select video) closer to the different standards and norms used in the context of digital libraries and archives, distance teaching and learning, etc.

However, one possibility (and there are others!) of bringing a concrete analysis closer to the ASW style is to set aside a designated sequence for entering information required by a norm or standard (such as Dublin Core, OAI, LOMFR, etc.) or by a reference schema of description (such as the Semantic Elements Schema of the European library Europeana). We have developed examples of this bringing together – particularly with the norm LOMFR³ (an adaptation of the LOM standard for the French education system) – but we have not yet had the opportunity to implement this approach more systematically.

Let us stress once more the fact that the syntactic structure shown in Figure 4.2 is canonic for all models of analysis of audiovisual corpora. Certain models have sequences of analysis for each of these five stages; others only have sequences for one stage or another.

The simplest models, which are closest to the standard forms of description/ indexation, are those which only have sequences for referential analysis and/or analysis of the space-time context. Other forms, however, possess a very elaborate structure, exploiting to the full the richness of the semiotic approach to the analysis of audiovisual corpora.

Thus, we can see that the canonic structure in Figure 4.2 remains crucial for the definition of a model of description and for its development in the guise of an interactive working form.

4.4. Types of analysis, analytical tasks, procedures of description and activities of description

As we have already mentioned, we can see that in general (see Chapter 1 as well as [STO 11a and STO 11b]) the analysis of audiovisual corpora which, for instance, make up the collection of an archive, constitutes one of the important activities – in the workflow – which characterizes the working process* of compiling, broadcasting and preserving cultural or scientific heritage.

The analysis itself may take the form of a “simple” indexation of an audiovisual corpus; it may also take the shape of a genuine expert assessment (in relation to archive sciences, the reference here is to the diplomatic study of documents), or manifest itself in the form of an action of such-and-such an interpretative school or approach. In other words, it may be carried out in pursuit of very different objectives and by actors (analysts) with very varied skill sets.

Going back to our discussion in section 2.3 about the different types of analysis developed and practiced as part of our research on audiovisual archives, Figure 4.3 now offers an overall and structured vision of this crucial activity. We shall refer to this figure throughout the book.

The work of analysis may, as Figure 4.3 shows, be comprehended according to 4 main levels. The most general level is that which defines the type of analysis. The definition of the type of analysis influences the selection of the appropriate tasks of analysis (2^nd level in Figure 4.3); a task of analysis, in turn, is made up of one or more procedures of analysis (3^rd level in Figure 4.3); and finally, a procedure of analysis is itself made up of one or more activities of analysis (4^th level in Figure 4.3).

As we have already explained in Chapter 2, we consider an analysis of a text or corpus to always and necessarily be linked to a certain point of view (the analyst’s) as regards the text to be analyzed, and to a certain interest in the text, or indeed to an objective, a goal which the analysis sets out to fulfill. It is utterly futile to attempt to reduce all these points of view, interests and objectives to a single analytical framework. It is better to allow the analyst the choice of carrying out this-or-that analysis, to carry out an analysis of a textual object as he sees fit. Without wishing to suggest that the ASW metalinguistic resources covers all imaginable desires or needs of analysts, we do believe that these resources are flexible enough to fulfill a fairly large number of such desires and needs.

This is a direct consequence of the fact that we distinguish between four hierarchical levels to approach the task of analysis. At the first level, as Figure 4.3 shows and as we have briefly seen in Chapter 2⁴, we distinguish various types of analysis. One type is mainly concerned with a particular layer of the audiovisual text. This is the case, for instance, with audiovisual description* stricto sensu which is mainly interested, e.g., in pinpointing the visual or sound shots. Another type is thematic description* which is primarily interested in the thematization of a domain of reference in an audiovisual text or corpus.

Again, certain other types of analysis can be distinguished in relation to the question of whether we have to take account of the audiovisual text in its entirety, or just of this-or-that passage, this-or-that sequence.

Still more types of analysis can be distinguished in terms of the level of specialization envisaged, i.e. in terms of whether an analysis should more closely resemble a basic description (e.g. with no more than the 15 elements of description which make up the Dublin Core standard) or is extended in the sense of a genuine intellectual assessment of the textual object in question.

Choosing one type of analysis over another (level 1 in Figure 4.3) necessarily entails selecting one or more specific tasks of analysis (level 2) and the appropriate procedures of analysis (level 3) for each task. Thus, the paratextual description* type of analysis is aimed solely – as indeed the name suggests – a Dublin Core-type analysis of the audiovisual object; i.e. an analysis carried out by way of a task (Figure 4.3) known as description of the formal identity of the object. The model of description for that task is thus made up of a “long” sequence with objects of analysis specified according to the Dublin Core standard such as [Author], [Title], [Rights], [Description], etc. In order to fill in each of these objects of analysis, a procedure of description (or sometimes two) is associated with them.

As we shall see in Chapters 9 and 10 in particular, we distinguish between two basic procedures: free description* and controlled description*:

– free description signifies that it is the analyst who, in accordance with his own habits and preferences (particularly linguistic), chooses a conceptual term (or a set of conceptual terms) representing the object of analysis;

– controlled description, on the other hand, indicates that the analyst uses a thesaurus to carry out his description.

Each analytical procedure is, in turn, defined by one or more activities of analysis*. As we shall see once more in Chapter 14, we have a rich descriptive ontology at our disposal for describing activities of analysis of the textual object, which by using we are able to produce a very nuanced and fine specification of the procedures and tasks of analysis needed to carry out a given type of analysis.

This means that a particular procedure of analysis such as that of free description can vary enormously depending on the specificity of the object of analysis and the desired level of precision of an analysis. Thus, it is very different from having to designate a conceptual object using a verbal expression, illustrate it with a visual icon or indeed, if applicable, indicating its geographical, temporal or “conventional” coordinates (e.g. in the form of an address). All these eventualities mean that the same type of analytical procedure can vary greatly from one type of analysis to another, from one task to another, and so on.

This means that a particular procedure of analysis such as that of free description can vary enormously depending on the specificity of the object of analysis and the desired level of precision of an analysis. Thus, it is very different from having to designate a conceptual object using a verbal expression, illustrate it with a visual icon or indeed, if applicable, indicating its geographical, temporal or &ldquo;conventional&rdquo; coordinates (e.g. in the form of an address). All these eventualities mean that the same type of analytical procedure can vary greatly from one type of analysis to another, from one task to another, and so on.Let us highlight, therefore, that the choice of a type of analysis entails a selection of relevant tasks of analysis and procedures of analysis appropriate both for the specificity of the object of analysis and for the framework of the type of analysis chosen. The ASW metalinguistic system takes account of these variations, even though the current version of the ASW Studio software only partially integrates it .⁵

4.5. Particular tasks in analyzing the content of an audiovisual corpus

Let us now look more specifically at the type of analysis which we are most interested in here: thematic description* (of the content) of an audiovisual text or corpus. Figure 4.4 singles out the five main tasks of analysis relating to the content of an audiovisual text or corpus:

1. Analysis of the domain of reference (or expertise): analysis of the referential content, of the topical structure stricto sensu (task 1.1 in Figure 4.4).

2. Analysis of the referential contextualization (of the spatial and temporal localization) of the knowledge object thematized in the audiovisual text (task 1.2 in Figure 4.4).

3. Analysis of the discourse held about the knowledge object in question (task 2 in Figure 4.4).

4. Analysis of the mise en scène or audiovisual expression of the discourse held about the knowledge object in question (task 3 in Figure 4.4).

5. “Free” meta-textual commentary by the analyst (task 4 in Figure 4.4).

This being so, the analyst is not obliged to carry out all of the tasks identified above when performing a concrete analysis of the content of an audiovisual text or corpus. He is not obliged to furnish a minute description of the discourse production about a specific topic; nor is he obliged to provide more explicit information as regards his personal point of view every time he analyzes a specific topic.

Indeed, as Figure 4.4 suggests, with the different tasks identified, we have a sort of hierarchy of specialization in the thematic description, which covers basic thematic indexation of a domain of knowledge as well as the various interpretations of it, through its treatment as a discursive and audiovisual object:

1. In terms of basic thematic indexation, the analyst contents himself with explicitizing and identifying the domain(s) of knowledge thematized in an audiovisual text. Identification or explicitation may, as we shall see, be done either freely – it is the analyst who designates the text, names it, describes it, etc. verbally, visually (e.g. using icons) or acoustically (e.g. using jingles)) – or in a controlled manner (using a verbal or visual thesaurus).

2. Basic thematic indexation may be complemented by an analysis of spatial and/or temporal localization of the domain of knowledge. Again, this analysis is performed using procedures of free and/or controlled description.

3. In the context of discourse analysis per se, the basic thematic indexation of a domain of knowledge is complemented by an assessment of the discursive framing of the domain thematized in the audiovisual text and of the development of that domain.

4. Furthermore, the analysis of the discourse production may be complemented by an analysis of the verbal expression of the topic (e.g. of the terminology, the sociolinguistic registers, etc. with videos containing testimonies, interviews, etc.); it may also be complemented by an analysis of the audiovisual expression of the topic – e.g. when the thematization of a domain of knowledge (of a place, a person, an activity, etc.) is done essentially using visual techniques (visual framing, camera panning, visual shots, etc.), or sound-based techniques.

5. Finally, in the context of a professional assessment or a personal interpretation, thematic indexation may also be enriched by explanatory annotations (comments, etc.).

The possibility offered to the analyst to choose between these five approaches (or a combination of them) enables account to be taken of the specific interests and pragmatic constraints which necessarily influence the work of analyzing the content of an audiovisual text or corpus (time, human resources, etc.).

4.6. Concluding remarks

In the next part of this book (Part 2: Tasks in Analyzing an Audiovisual Corpus), we shall give a more detailed discussion of some of the tasks identified in Figure 4.4 which are particularly important for any content analysis. Thus, we shall discuss in greater detail:

– referential description* of a knowledge object (Chapter 6);

– analysis of the referential contextualization of a knowledge object (Chapter 7).

However, we shall discuss neither the analysis of the audiovisual mise en scène nor that of the verbal expression of the knowledge object.

In Part 2 of this book, we shall, more specifically, discuss the procedures of description. Finally, the fourth part will be given over to a discussion of the objects of analysis (i.e. the objects with which a concrete work of analysis is “concerned”), and also the activities and tools specific to the analysis.

---

1 For a more in-depth discussion of the ASW meta-lexicon of conceptual terms, see Chapters 12, 13 and 14 of this book.

2 For a more in-depth discussion of the two units, sequence and schema, see Chapters 15 and 16 of this book.

3 See: http://www.lom-fr.fr/.

4 For further information, please see [STO 11a].

5 The current version of the ASW Description Workshop enables the analyst to choose between different types of analysis, but this choice does not entail an automatic selection of relevant tasks and procedures. The analyst has to follow a “simple” methodological guide in order to perform his analysis appropriately.