Print advertising in anglophone and francophone Canada from a critical discourse analytical point of view: establishing different relations between the producer and viewer of advertisement images

Objective and Rationale

This article explores whether marketers use different strategies for the visual design of magazine advertisements when addressing English and French Canadians; in particular, whether marketers create different relations with these two language groups.

In general, there are two reasons for analyzing advertisements, particularly their visual design, from an intercultural perspective. First, from a media sociological point of view, Maigret (2003: 244) contends that

Les contenus des médias de masse peuvent être vus comme des traces des interactions qui les ont constitués, comme des plis condensant les rapports sociaux, les logiques d’action et les mouvements culturels. (The contents of mass media can be viewed as traces of the interactions of which they are composed, as folds that condense social relationships, the logic of actions, and cultural movements.) ¹

In other words, the contents of mass media, and in particular, as in the present study, of advertisements, reflect social relations and culture-bound social practices.

Moreover, in order to conduct a cross-cultural analysis, advertisements are ideal objects of research since advertisements are the most important carriers of cultural values in public space (Richards et al., 2000: 1). Furthermore, in cultural analysis the goal is, generally, to detect ‘regularities and recurrent patterns, dominant associations which will be shared by many people within a given cultural context’ (p. 141). Therefore, I have chosen advertisements in consumer magazines to produce a cross-cultural analysis of a corpus of advertisements.

Second, De Mooij (2005: 9) states that, despite the increasing importance of visual communication over verbal representations in advertising, research usually focuses on the copy used, which neglects the importance of the visual components of advertisements. The rising importance of pictorial elements over verbal elements in consumer magazine advertisements has been confirmed by McQuarrie and Phillips (2008: 103). They carried out a content analysis of the style of magazine advertisements focusing on the nine editions of Which Ad Pulled Best? from 1969 to 2002. They found that, especially after the 1990s, the emphasis shifted from the verbal to the visual. As a reason for this development, McQuarrie and Phillips (2008: 96) maintain that ‘while the older advertisements assume an attentive reader, the more recent advertisements presume a visually oriented, casually browsing viewer’ who seeks ‘visual entertainment’ (emphases in original).

Particularly in advertisements for consumer products, such as for perfumes, food and clothes, which appear typically in women’s magazines, imagery is most important. This is due to the fact that this product category is frequently associated with low-involvement, which is why little verbal information on the product is required (Schweiger and Schrattenecker, 2001: 207). Generally, low-involvement purchases are defined as those that show little relevance to the consumer, are not perceived as risky, and result in a limited level of information processing (Schiffman and Kanuk, 2010: 229). Hence, I have paid particular attention to the strategies applied by marketers in relation to the visual design of the advertisements to be analyzed.

The comprehension of these strategies requires, among other factors, the ability to produce and understand the grammar of visual design in today’s society. Kress and Van Leeuwen (2006: 3) use the term ‘visual literacy’ to express this skill. In other words, knowledge of the basic rules is becoming a prerequisite for performing successfully, that is, to encode and decode visual representations correctly, in the workplace, in public communication, art, science, and in many other domains (pp. 2–3).

For the empirical analysis of the visual design of my corpus I have used the Reading Images: Grammar of Visual Design model developed by Kress and Van Leeuwen (2006), which is described in some detail below. Several researchers have applied this model to a variety of fields. Pederson (2002), for example, attempts to reveal the manipulative character of advertisements and Heidinger (2006) presents a pictorial analysis of images in wellness brochures. Furthermore, Watts (2004) demonstrates the suitability of Kress and Van Leeuwen’s model to design templates for internal corporate communication.

Although Forceville (1999: 163) appreciates Kress and Van Leeuwen’s innovative approach in providing a tool for analyzing visuals, he doubts the general applicability of their grammar of images across various genres. He argues that ‘while grammaticality in language is relatively stable across text-genres, the grammar of images may well be considerably more genre-dependent’ (p. 172). Moreover, Forceville (1999: 173) stresses that Kress and Van Leeuwen’s model needs empirical testing. Hence, I have limited the corpus analyzed to magazine advertisements and have conducted an empirical analysis of the corpus of advertisements.

Data and Methods Applied

To answer the research question, I performed a corpus analysis of advertisements drawn from two Canadian magazines by employing a synchronic approach. The basis for selecting an appropriate magazine was the Print Measurement Bureau (PMB) (2003) report, which is published annually. It conducts research on the readership of more than 115 publications and consumer usage of more than 2,500 products and brands by using an annual sample of 24,000 interviewees (PMB, 2011). PMB (2003) gives detailed information on circulation and readership according to various criteria such as sex, education, region and the language most often spoken.

The criteria for selecting a particular magazine were the following:

Among all the Canadian magazines, some are published in both official languages, with the magazines Chatelaine and Châtelaine (Chatelaine, 2011) having the highest circulation. Choosing the magazine with the highest circulation increases the chances that large international companies will place advertisements in it. Hence, since these companies have the necessary resources, they are more likely to adapt their advertising strategies. As a consequence, the corpus consists of English and French issues of the monthly women’s magazines Chatelaine and Châtelaine published by Rogers Publishing Limited, a division of Rogers Media Inc. (Rogers, 2011). Chatelaine was first published in March 1928 (Korinek, 2000: 3) and is ‘Canada’s national women’s magazine’ (p. 6). In October 1960, the first issue of its French counterpart – Châtelaine – appeared (Des Rivières, 1992: 11). Moreover, the readers’ demographic profile of both Chatelaine and Châtelaine are very similar in terms of age group (between 35 and 49), marital status (married or living together with a partner), family size (more than 3 people), and income (between $50,000 and $74,900), according to the PMB (2003) report.

For my empirical analysis I applied the technique of content analysis to a corpus of advertisements. According to Krippendorff (2004: 18), ‘content analysis is a research technique for making replicable and valid inferences from texts (or other meaningful matter) to the contexts of their use’. In this definition, the expression ‘or other meaningful matter’ implies that the use of content analysis is not limited to written material, but may also be applied to texts, such as images, sounds, signs, symbols, or works of art.

From all the issues available, I collected 13 issues of Chatelaine (January to December plus an extra Halloween issue in October) and 12 issues of Châtelaine (January to December) over the year of 2006. I chose a whole year in order to ensure the largest possible variety of products. The reason is that, in the summer, advertisements for particular products such as ice creams may appear whereas, in the winter, for instance, decorations for the Christmas season may be advertised. From all the issues of Chatelaine and Châtelaine I recorded a total of 2,280 advertisements. To avoid duplication, completely identical advertisements, both visually and verbally, which were placed several times in a number of issues of the same language, that is, English or French, were only counted as a single one. Consequently, the total number of advertisements analyzed decreased to 863 in Chatelaine and 713 in Châtelaine.

As a next step, I determined variables, based on Kress and Van Leeuwen’s Reading Images: Grammar of Visual Design model, together with mutually exclusive and exhaustive values (Bell, 2001: 15–16). In the following, the coding frame was pre-tested on parts of the corpus of advertisements to find out whether it needed to be revised. This was the case in few particular cases. For example, where the visual showed a person before and after the use of a certain cosmetic product, the person was counted twice. The reason was that the repetition of a person’s representation in the same advertisement contributes to the overall impression the viewer gets of the advertisement, which is, for example, the occurrence of several demands or offers. Moreover, with respect to gaze, social distance and point of view, a variable that was based on the main person or persons depicted in the advertisement was added. This proved to be necessary to take into account that the person or persons usually shown in the centre of an advertisement – and hence the most salient ones – has or have a prominent communicative function.

Subsequently, the intercoderreliability with four other coders was ascertained. For that purpose, I calculated the percentage of agreement between my coding and that of the other coders. As a basis for the coding, I selected the advertisements of the first half of both the English and French issues of January and July. Whereas coders 1 and 2 looked at the English issues, coders 3 and 4 analyzed the French issues. The January and July magazines were selected to provide as high a diversity of products as possible, taking into account that the types of products advertised may vary according to the seasons of the year. The four coders were trained carefully to ensure the accurate application of the variables.

The average percentage of agreement between my assessments and those of the other four coders in relation to the 17 variables of the coding frame was 0.83; this is in compliance with the minimum level of agreement of 80 per cent demanded by Riffe et al. (1998: 128). In particular, for each variable, the number of cases was determined where my judgement and that of both the two coders of the English advertisements and the two coders of the French advertisements were the same; then this number was put in relation to the total number of advertisements examined.

After coding all the advertisements of the corpus, the software SPSS was used to generate cross-tabulation tables and to carry out a Pearson chi-square test on the data. Now let us take a closer look at the constituents of Kress and Van Leeuwen’s Reading Images: Grammar of Visual Design model (2006) which served as a basis for my empirical analysis.

The Reading Images: Grammar of Visual Design Model

Kress and Van Leeuwen’s visual grammar model refers to realizations of visual communication in Western cultures. The reason for this is that communication is always socially constructed and uses the resources available in these cultures (Kress and Van Leeuwen, 2006: 3).

Kress and Van Leeuwen (2006: 2) argue that ‘like linguistic structures, visual structures point to particular interpretations of experience and forms of social interaction.’ Moreover, they hold that ‘meanings belong to culture, rather than to specific semiotic modes.’ There are things that can be expressed either verbally or visually, whereas other things only use verbal codes, and again for others only visual codes are possible. However, generally speaking, the way things are conveyed is culturally and historically bound.

In addition, even if things appear to have the same meaning when expressed in a verbal or visual mode, this meaning is communicated in a different way. Whereas in language certain word classes or semantic structures are used to express an idea, in visual communication, for example, colour or compositional structures help to convey the message (p. 2).

Kress and Van Leeuwen created ‘a descriptive framework that can be used as a tool for visual analysis’ (pp. 13–14). This tool not only has practical implications but can also be used to make critical judgements. It can be applied, for example, in the classroom or in a business context for designing and critiquing maps, charts, diagrams, pictures and various forms of layout to optimize efficiency in teaching and presenting. Moreover, their concept makes critical assessment of visual communication possible and thus also contributes to a critical discourse analysis that goes beyond the analysis of only verbal representations (p. 14).

In general, every semiotic act includes two kinds of participants. First, there are the represented participants, who or which are the subject of communication, for example, people, places and things. Second, there are the interactive participants, who are the people involved in the act of communication, that is, the speaker and listener, the writer and reader, or the image-maker and the recipient of the image (pp. 47–48).

So, with respect to images, there are, on the one hand, represented participants, that is, the people, the places, and things shown in the image, and, on the other hand, interactive participants who are the producers and the viewers of the image establishing a communication process with each other. By and large, we can distinguish between three types of relationships (p. 114):

(1) relations between represented participants; (2) relations between interactive and represented participants (the interactive participants’ attitudes towards the represented participants); and (3) relations between interactive participants (the things interactive participants do to or for each other through images).

In the following, I will refer to relations between the image-producer and the recipient, determined by how people are represented, for which Kress and Van Leeuwen have provided a detailed analysis in their model.

Representation and Interaction: Establishing Relations between Image-Producer and Recipient

When drawing a map to give directions or explaining a diagram to somebody, both the producer and viewer of the image are present and interact directly with each other. In contrast, referring to an image in a magazine advertisement, usually neither the producer, nor the viewer, enters into personal contact.

Whereas the recipient normally has only a vague idea of the image-producer and the production process, the image-producer does not know the audience, though he or she attempts to define the target group as precisely as possible in order to avoid waste coverage. What is unique about this situation is that ‘there is a disjunction between the context of production and the context of reception’ (Kress and Van Leeuwen, 2006: 114). An exception would be graffiti artists since they respond to billboards directly by defacing them, albeit with sometimes a considerable time lag from placing the billboard (pp. 114–115).

Nevertheless, according to Kress and Van Leeuwen, the image-producer and the viewer share, with regard to the image, ‘a knowledge of the communicative resources that allow its articulation and understanding, a knowledge of the way social interactions and social relations can be encoded in images’ (p. 115). With their Reading Images: Grammar of Visual Design model they try to disclose these cultural competencies (p. 116).

Gaze

We have to distinguish between two types of pictures whose communicative functions are quite different, namely ‘demands’ and ‘offers’. Kress and Van Leeuwen (2006: 117) define a ‘demand’ as a picture where represented participants look at the viewer; contact is established, at least on an imaginary level. These represented participants are usually humans, animals, or quasi-human beings such as cartoon figures (p. 118).

The functions of a ‘demand’ are twofold. Firstly, it addresses the viewer directly, and secondly, the image-producer does something to the recipient through the image. The gaze of the represented participant, and also his or her gesture, if any, demand that the recipient creates an imaginary relationship. The kind of relationship, in turn, depends on, for example, the facial expression and gesture of the represented participant; a smiling, staring or frowning face as well as a prompting, aggressive or defensive gesture are intended to arouse different reactions on the part of the viewer (pp. 117–118). This is demonstrated in Figure 1 where the model looks directly at the viewer with a big smile on her face. She establishes contact with the viewer and tries to create the viewer’s favourable attitude towards her and, what is more important, the perfume.

OPEN IN VIEWER

Figure 1.

Estée Lauder (Chatelaine, May 2006: 17).

In contrast, there are also pictures which address the viewer indirectly, where no contact is established, and where ‘the viewer’s role is that of an invisible onlooker’ (Kress and Van Leeuwen, 2006: 119). This type of pictures refers to images depicting human or quasi-human participants who do not look directly at the viewer. Such pictures are termed ‘offers’; they ‘offer’ the represented participants to the viewer as items of information, objects of contemplation, impersonally, as though they were specimens in a display case’. Figure 2 shows an example of an ‘offer’. The model does not look at the viewer. Hence, she addresses the reader indirectly and no relation is made with the viewer. What seems to be most essential is to show the effects of the makeup on the model’s face.

OPEN IN VIEWER

Figure 2.

Covergirl (Chatelaine, February 2006: 169).

So, whenever an advertiser wants to show people in an advertisement, it has to be in the form of a ‘demand’ or an ‘offer’. This choice may be limited by the various ‘pictorial genres’. While, for example, not only the anchorman or anchorwoman on television, but also the model in a magazine, are supposed to look at the viewer, an interviewee on television, or an actor in a movie, usually do not direct their glance at the viewer (Kress and Van Leeuwen, 2006: 120).

Perspective

Besides taking the decision between demand and offer, as well as selecting a certain distance, one also has to choose a particular angle, or in both literal and figurative terms, a particular point of view. The image-maker can express his or her subjective attitude, which is usually socially formed, towards the represented participants, human or non-human, by opting for a certain point of view (Kress and Van Leeuwen, 2006: 129).

Horizontal angle and the difference between involvement and detachment

With reference to the perspective chosen, let us first consider the horizontal angle. According to Kress and Van Leeuwen (2006: 134), in general, the frontal plane of the image-producer and that of the represented participants can be either in parallel or divergent from each other. Consequently, one can choose ‘either a frontal or an oblique point of view’ (p. 135). It goes without saying that different subjects of a photo may be shown from different angles along the horizontal axis.

Generally speaking, if a subject of an image is depicted from the frontal angle, this denotes that the subject is part of the world of the image-producer, and hence the viewer. In contrast, if a subject is shown from an oblique angle, it means that it is not part of the image producer’s and viewer’s world. In other words, the frontal angle represents the interactive participants’ involvement and the oblique angle their detachment from the represented participants (pp. 136–137).

Figure 3 is an example of a person shown from a frontal point of view. The model in the advertisement for a skin cleansing product is depicted from the frontal angle, which means that she is part of the viewer’s world. In addition, she acts like a disc jockey. Therefore, both the visual design and the clubbing atmosphere encourage potential customers, especially young ones who often fight against skin impurities, to relate to the model and thus create favourable attitudes towards the product.

OPEN IN VIEWER

Figure 3.

Noxzema (Chatelaine, March 2006: 33).

In contrast, Figure 4 shows a person from an oblique point of view. The frontal plane of the model shown in the advertisement diverges from that of the viewer. Therefore, she is not part of the viewer’s world. Nevertheless, she looks directly at the viewers and establishes contact with them.

OPEN IN VIEWER

Figure 4.

Clairol: Herbal Essences (Châtelaine, August 2006: 111).

Vertical angle and the exercise of symbolic power

Along the vertical axis one can also select different angles. According to Kress and Van Leeuwen (2006: 140), represented participants can be seen from a high or low level. In the first case, the viewer seems to have power over the represented participant while, in the second case, the represented participant appears to have power over the viewer. If the picture is at eye level, then the point of view is one of equality and there is no power difference involved.

Models in magazine advertisements are usually shown from a low angle. They look down on viewers as if they were exerting power over them. This symbolic power derives from the fact that they demonstrate to us the latest fashion items, most prestigious perfumes, or most efficient shampoos. They somehow dictate what to buy in order to stay ‘trendy’ (Kress and Van Leeuwen, 2006: 140; Messaris, 1997: 38–40).

Note that in the advertisement for Merle Norman (see Figure 5), which follows the before–after format, the woman depicted in the small picture in the lower right-hand corner is at eye level with the viewer. However, in the large picture the woman is seen from a low angle. In other words, after visiting the beauty salon and wearing make-up applied by cosmeticians she shows her superiority to the recipient; the red roses next to her shoulder may symbolize that she is now desired by men. Last but not least, the woman in the large picture is depicted from an oblique angle, that is, she is not part of the viewer’s world. Nevertheless, she represents a role model the viewer may measure up to after going to the cosmetic studio promoted in the advertisement.

OPEN IN VIEWER

Figure 5.

Merle Norman (Chatelaine, December 2006: 26).

Results of the Statistical Analysis

Generally, we can conclude that two variables are related if the significance value (p) is small enough, that is, if it is below 0.05 (Field, 2009: 696–697). Moreover, in order to make the results accurate, the expected frequencies, which, generally, are based on the assumption that the proportions are equal across samples, need to be greater than 5 in each cell (p. 690). In the following cases where variables showed significant differences between the languages, both requirements were met.

The analysis below presents the variables chosen in some detail. The analysis for each variable is followed by the conclusions which can be drawn from the findings.

Perspective related to the vertical axis: symbolic power

As mentioned above, regarding the vertical axis, people can be depicted from a low, eye-level or high angle, according to the symbolic power relation between the person shown and the viewer (Kress and Van Leeuwen, 2006: 38–40). If people are presented from a low angle, they exert power over the viewer; if they are shown from a high angle, the viewer, conversely, exercises power over them; and if they are at eye level with the viewer, the power relation is represented as equal. Taking only those advertisements into consideration which depict people, the results show a correlation between the perspective and language.

First, the results corroborate Kress and Van Leeuwen’s (2006) theory that models in magazine advertisements usually look down on the viewer, that is, are shown from a low angle, because, when adding up the number of the advertisements of both languages, 719 of them show people from a low angle, compared to only 129 representing people from a high angle, and 203 at eye level with the viewer. The predominance of the low angle perspective is even more pronounced in the case of the French advertisements than in the English ones. Thus, in critical discourse analytical terms, the large majority of advertisements show models who exert symbolic power over the recipient.

For the remaining cases of advertisements where people are depicted from a high angle, or at eye level with the viewer, and therefore deviate from the most frequently found low angle perspective, we see from Table 1 that these two perspectives are more common in the English advertisements than in the French advertisements. Thus, the English advertisements show a higher variety of symbolic power relationships than the French ones.

OPEN IN VIEWER

Table 1.

Overview of statistical results.

Variables	Values	Comparison of frequencies
		in English ads	in French ads
Social distance from which the main person or persons is or are shown	Intimate distance	Higher
	Close personal distance		Higher
	Far personal distance		Higher
	Close social distance	Higher
	Far social distance		Higher
	Public distance		Higher
	Social distance not precisely identifiable	Higher
Horizontal angle: Perspective from which the main person or persons is or are shown	Oblique point of view		Higher
	Perspective is not precisely identifiable	Higher
Vertical axis: Perspective from which people are shown	Low angle		Higher
	Eye level	Higher
	High angle	Higher

Conclusions, Limitations and Directions for Further Research

Table 1 sums up the results of the statistical analysis with respect to the relations between the image producer and viewer.

No significant associations were identified between the language and the relative frequency of demands and offers, that is, whether the person depicted looks directly at the viewer or not. Nevertheless, considering the whole corpus of advertisements analyzed, the total number of main persons who enter into contact with the viewer is much higher than the total number of those who do not do so. This may derive from the fact that a large portion of advertisements in women’s magazines promote beauty and personal care products, and therefore require ‘personal contact’ to be established with the reader, for example, to demonstrate the results of the product. These findings corroborate Kress and Van Leeuwen’s (2006: 120) claim that the use of demands and offers appears to largely depend on the genre images appear in.

Furthermore, it emerged that English advertisements do not determine the social distance from which the main person is shown as much as French advertisements do. Thus, the social relations between the represented person, and hence the image-producer, and the recipients are identified to a relatively less extent. French advertisements also seem to show a higher variety of social distances, which leads to an overall greater variety in the depiction of body details and contexts, that is, backgrounds. These results may indicate the cultural differences between anglophone and francophone Canadians. In general, French Canadian culture shows a higher degree of social distances than English Canadian culture. In addition, English advertisements employ an intimate distance (only face or head are shown) and a close social distance (whole figure without space around it) more frequently than French advertisements. By contrast, the latter make comparatively greater use of a close personal (head and shoulders) and far personal distance (from the waist up), as well as a far social (whole figure with space around it) and public distance (torso of at least four or five people). The differences in public distance are of particular importance since these reflect the generally greater group-orientation of French Canadians compared to English Canadians.

In addition to the above, several other findings with respect to Kress and Van Leeuwen’s model were made. As regards the horizontal axis, the proportion of English advertisements in which the point of view from which the main person or persons is or are shown cannot be exactly determined is significantly higher than among French ones. Therefore, in comparison to these, English advertisements seem to avoid not only defining the social distance, but also determining whether the person or persons depicted is or are part of the viewer’s world or not. What is more, considering all the advertisements of the corpus, the results support Kress and Van Leeuwen’s (2006: 140) as well as Messaris’ (1997: 38–40) claim that models in magazine advertisements usually look down on the viewer. In other words, from the point of view of critical discourse analysis, models exert power over the viewer by giving directions about, for instance, which hair shampoo or facial cream to use, or which snacks to eat.

In view of the above, we can conclude that the corpus of visuals drawn from the two Canadian consumer magazines shows the ‘regularities and recurrent patterns’ (Richards et al., 2000: 141) of English and French Canadian social practice. In particular, the visual design of the advertisements analyzed reflects French Canadian society, which is much more characterized by social distances than English Canadian society.

Of course, this study has its limitations. First, it only examines advertisements in a particular media class, i.e. magazines, and in two particular media vehicles (Pickton and Broderick, 2005), i.e. Chatelaine and Châtelaine, with both addressing a specific set of target groups. Second, it considers solely English and French Canadian recipients.

The advertisements studied here themselves also offer other research opportunities. On the one hand, one option would be to re-focus attention from the represented participants in the Chatelaine and Châtelaine advertisements to the interactive participants involved in the communication process, to use the terminology of Kress and Van Leeuwen (2006: 47–48). Specifically, one might take a look at the producers of the advertisements, that is, advertising agencies or companies which produce in-house advertisements themselves, and attempt to identify the criteria they apply in devising their advertisements. For that purpose, qualitative interviews with the marketing experts involved may be conducted.

On the other hand, the recipients of the magazine advertisements could be examined. Here, a possible question would be whether anglophone and francophone readers receive the various visual designs of advertisements differently. To this end, advertisements could be deliberately manipulated; the same advertisement whose main person represents on one occasion a demand, and at another time an offer, could be shown to the subjects in order to investigate differences in cognitive processing and emotional response. Similarly, the main person’s social distance, or the perspectives from which he or she is depicted, could be manipulated. As research method, a survey could be carried out among respondents from both language groups.

Last but not least, Forceville (1999: 171–173) points out an important shortcoming of Kress and Van Leeuwen’s model; he criticizes the fact that Kress and Van Leeuwen (1996) do not sufficiently take into account how the visual is anchored on the verbal text and how the visual relates to other images. In other words, they do not state how (con)text can be integrated in the analysis of visuals. To overcome this methodological weakness, Forceville suggests considering text-external context in the widest sense, that is, including context beyond the accompanying verbal and visual representations. For that purpose, not only the intentions of the producer of the visual, partly determined by the special genre it falls into, but also the recipients’ identity need to be taken into consideration.