1For Pierre Bourdieu, the tastes and lifestyle of an individual are in relative unity because they come from the same “generative schemes.” Aesthetic dispositions that are expressed in a particular cultural domain are transferable from this domain to another in such a way that “a typical structure of preferences for and knowledge about painting is likely to be linked to a similar structure of knowledge and tastes in music, or even jazz or cinema” (Bourdieu, Darbel & Schnapper 1969: 101). This idea of the unity and coherence of tastes underpinned the construction of the questionnaire used to establish the “anatomy of taste” (Bourdieu & de Saint-Martin 1976) which was later included in Distinction (Bourdieu 1984a). In view of this, the hierarchical space of taste has a homologous relationship to the space of social positions: the consumption choices of an individual will focus on goods and works occupying most often a homologous position in the space of taste with that occupied in the space of social positions.

2In the wake of Distinction, debates in the sociology of cultural practices have focused in particular on the question of the rise of cultural eclecticism, namely the fact that the upper classes consume and appreciate symbolic goods and works that are far apart or even at opposite ends of the cultural hierarchy (Peterson & Kern 1996). Studies that aim to reproduce the analysis of Distinction anywhere other than France and at any other point in time by remaining as close as possible to the approach of Bourdieu, in particular by combining different types of taste indicator, are generally critical of this idea of rising eclecticism (see Prieur, Skjott-Larsen & Rosenlund 2008 and, for a synthesis, Duval 2010). Within this literature there are different ways of measuring and objectifying taste, and indicators are constructed on the basis of different hypotheses so that differences in results are the product of these methodological differences. Quantitative research documenting a rise in eclecticism is most often based on statistical analysis using aesthetic genre classifications. Such classifications are then distributed throughout the analysis using scales of legitimacy, and its use and the rationale behind its constitution rarely discussed. Each genre is defined as being very legitimate, or having average or little legitimacy. These scales, when they are not constructed a priori, are constructed by inferring the degree of legitimacy of a genre or practice using the social position of practitioners or enthusiasts of this practice according to the hypothesis of homology between a hierarchy of cultural practices and a hierarchy of social positions (Robette & Roueff 2014).

3Genre is therefore considered a reliable indicator not only of taste—which cannot be observed as such it itself—but also of a positioned taste—“classified and classifying”—in the space of tastes and practices, in such a way that it is possible to describe an individual who claims to like both jazz and rap as “omnivorous” or eclectic. However, criticism of genres as imprecise “portmanteau words” is not new (see DiMaggio 1987).

4Genres hide a field or subfield of production behind a general category. This field or subfield is structured around poles (restricted audience versus the general public, symbolic versus commercial recognition, etc.), and in it artists occupy specific positions relative to one another (Bourdieu 1996; Duval 2016). While the use of genres can be relevant and defended when it comes to ascertaining and studying an affinity between a genre and a social group (that is to say two aggregated groups, of artists and works on the one hand, and of individuals on the other), it appears much more problematic when studying the tastes or cultural profiles of individuals. It is possible to state a preference for rock, pop, or easy-listening and chanson, and to like very diversely consecrated or recognized works or artists within these genres.

5Based on analysis of the answers given by a population of French academics to three open questions about their favourite musicians, writers, and painters, this article will examine the relative homogeneity of academic taste and the permanence of unity and relative coherence of tastes based on co-citations and artist combinations in these three fields, studied using network analysis. Do the theories of eclecticism and dissonance stand up to analysis when we move away from genre classifications assigned by researchers between legitimate, average, and popular in favour of questions that relate to a preference for certain artists? Does the preponderant place given to musical taste in sociological literature not lead to over-generalizations based on a specific cultural sphere? Is taste shaped by logics other than genre, as captured by the co-citations of different artists? How should we differentiate between artists within the space of taste and thus determine the existence or not of relative unity? What possible major cultural gaps have been observed?

6There’s a dual purpose to surveying a population of academics. Level of education is the biggest determining factor of taste (Robette & Roueff 2017): by studying a small population whose members are highly qualified, the effect of the qualification obtained is neutralized and allows us to better observe the other principles of determination and differentiation of taste. Being part of a profession with a high volume of cultural capital, academics are also a focus of cultural legitimacy; they embody the cultural pole of the dominant class. They therefore constitute a privileged point of observation of the transformations of cultural legitimacy.

7First, the article discusses different ways of measuring taste and their legitimacy in use in sociological literature, showing how the method employed here is original in its use of open questions relating to artist names and in its analysis of occurrence grids. The second section focuses, in view of this, on the artists cited by this population of academics. It shows that within this profession of individuals whose qualification level is both high and identical, “cultural diversity” is entirely relative. Indeed, the list of artists cited testifies to the significance that scholarly culture continues to hold. However, this is not exclusive and such preferences are accompanied by more contemporary artists who occupy intermediate positions in their respective fields of production. Each sphere of taste has its own structure; the co-occurrences operating between literary authors are not governed by the same logic as that of musicians or painters on account of the different structures that govern the sphere and their respective field of production. Be that as it may—and this forms the focus of the third section—it is possible for tastes in music, literature, and art to intersect, and to obtain categories of taste that are relatively coherent from the point of view of the position of the different artists in their respective spaces of taste. We will see that these taste categories are linked to the social characteristics, other than level of education, of the respondents. Finally, we observe that "cultural transgressions”—that is, listing among cited preferences artists from different spheres that are distant from one another in terms of legitimacy—occur principally in musical tastes. These transgressions consist in the expression of a taste for musicians who have little legitimacy in combination with very legitimate tastes in literature and art.

Methodology

1. Surveying Tastes in Culture

1. 1. Asking about Preferences without Genre Classifications

9In the context of studying the relative homogeneity of academic taste, using preference questions seemed most appropriate. While the population studied—non-Parisian academics—has a common profession and is homogeneous as regards its educational level, the living conditions of its members are diverse (place of residence, family background, age, see Gabrysiak 2021). Asking about preferences rather than practices also allows us to account for the differences between Paris and other agglomerations. The most prestigious institutions, which represent the top of the academic field, are historically concentrated in Paris (Bourdieu 1984b). From a cultural point of view, the inhabitants of inner Paris benefit from a wider range of cultural activities and have unique cultural behaviours (which are not simply explained by their socio-economic characteristics and the high concentration of managers and intellectual professions), particularly with regard to attendance at live performances, which is higher than the national average (Lombardo & Wolff 2020: 22). This uniqueness of practices is not reflected in preferences. It is still necessary to determine which questions to ask in order to better understand these preferences.

10Most often in surveys, preferences are gleaned via questions relating to favourite genres, whether of music (the most studied), literature, or film. This is particularly the case in research into issues of eclecticism, omnivorousness, and dissonance, from the pioneering work of Richard Peterson (Peterson & Kern 1996; Peterson 2004) to the more contemporary work of Bernard Lahire (2008), to which I will return, or that of Tak Wing Chan and John H. Goldthrope (2007), and Taru Lindblom (2022). However, genre classifications appear problematic in several respects, and some researchers have hypothesized that the emergence of the figure of the omnivore may simply camouflage the current difficulty for researchers in drawing a clear boundary between genres and the relevance (or erasure) of such boundaries (Bellavance, Valex & de Verdalle 2006: 142).

1. 2. How to Approach Open-Ended Questions about Taste Statistically?

14While I deemed the use of open questions on preferences for artists to be an appropriate indicator of taste, it presupposes knowing which statistical methods to apply to these artist co-citations and how to situate the artists mentioned in relation to one another.

15Several recent works have been based on analysis of answers given to one or more open or semi-open questions. These include the last three films seen by a sample of 18–20-year-olds (Boyadjian 2021), favourite musicians or groups from a pre-coded list of genres (Vlegels & Lievens 2017), and pieces of music recently heard and enjoyed by a population of 15–25 year olds (Azam et al. 2018). Answers obtained to open questions are treated and organized differently according to the stated object and hypotheses of each study. Investigating the persistence of the concept of cultural capital in explaining the formation of taste, Boyadjian classifies the films cited according to whether they have obtained a label of “arthouse” or “discovery and research” or no label from the Association française des cinémas d’art et d'essai [the French arthouse film association]. Rather than using the social position of respondents as the basis for his legitimacy scale, he uses a professional association recognized within the cultural field studied. He then establishes correlations between these different categories of film seen and the social profiles of respondents.

16For their part, Azam et al. use network analysis to establish clusters of respondents (“music profiles” which combine social profile and ways of listening to and discovering music) and artists (the most cited artists together). The linking of these clusters allows them to establish a typology of “musical realms.” Finally, Vlegels and Lievens also use network analysis to form and relate clusters of music groups and clusters of listeners. They focus on links between clusters in order to establish different patterns of omnivorousness.

17Having read this work, the use of network analysis seemed particularly suitable to deal with questions about the names of preferred artists. Network analysis allows a “from the ground-up” approach (Lamont 2010), requiring no recoding or pre-formed categories. The main names and their links appear in order to characterize style, the unifying principle of taste, a posteriori without supposing genre to be a unifying principle. Doing without genre classifications does not mean abandoning analysis in terms of cultural legitimacy and transferability of dispositions altogether. The limitation of the articles by Vlegels and Lievens and by Azam et al. is that, without genre, the authors also eschew the question of scales of legitimacy and, with it, the hypothesis of homology between the space of consumption and that of the social positions that underlie it.

18Adopting Boyadjian’s approach to the cultural legitimacy of artists via forms of recognition specific to the field of production considered could provide a solution for situating artists in relation to one another. It would then be necessary to use the names of writers, musicians, and painters to determine the degree of legitimacy of each, and then to observe to what extent legitimate writers are cited with legitimate musicians, and so on. In practice, however, there is a threefold limitation in that it is difficult to find equivalents to the labels found in the field of film in other spheres such as music and literature which are not based on a structured network like that of cinemas in the field of film. It would probably be necessary to combine several indicators such as prizes and awards for several hundred artists, which seems insurmountable. Moreover, the conversion of this type of indicator into degrees of legitimacy is not straightforward. There is also the problem of temporality. In the case of Boyadjian’s investigation, label classification works because the films were released relatively shortly before the investigation. However, the degree of recognition and legitimacy of the works is subject to variation as their respective cultural fields evolve.

19Faced with numerous artist names from different spheres (music, literature, and painting), and over an extended time scale, it seems impossible to find simple and unique criteria per sphere to situate them precisely on a scale of legitimacy. In order to study the logic of co-citations, different artist criteria for differentiation were taken into account first of all: years of practice; nationality and language; genre or movement; publishing houses for writers; museums and exhibition spaces for painters; main prizes received; and the sociological literature on the fields of production and consumption of these different spheres.

20After having reviewed the list of all artists cited and noting the low value granted to “popular” and avant-garde artists, in the next section I will explore patterns of association between artists in the same sphere. Literature and the space of taste that is linked to it is different from the space of musical taste, which is itself different from that of taste in art. I will then examine transfers and links between these different spheres, and relate them to the social properties of respondents before, in the final section, focusing on the individual level to study how cultural transgressions materialize.

2. Academics’ Tastes in Literature, Music, and Art

21A total of 183 academics in literature, humanities & social science, and science responded in full to the questionnaire circulated. The distribution of responses varies depending on the cultural sphere considered: while 44% of respondents cited five musicians or music groups, only 22% cited five painters (Table 1). The simple fact of expressing a taste in art is therefore rarer (and more “classifying”) than the expression of a taste in music. The number of artists cited would have been 915 for each domain if all respondents had cited five unique artists (artists cited only once). Although this theoretical maximum was a long way from being reached, the majority of artists cited in all questionnaires were only cited once (Table 2). Nevertheless, the number of respondents who cited five unique artists for any field is marginal (less than 2%), and almost all respondents cited at least one artist who was cited twice or more. Painting has: the fewest unique artists cited; the fewest individuals who cited at least one artist and at least five artists; and artists with the highest occurrences. Indeed, the most cited artist, all fields combined, is Claude Monet (36 occurrences), followed by Pablo Picasso (28 occurrences), and Vincent Van Gogh (25 occurrences).

 

 

 

Table 1. Distribution of number of artists cited by sphere by each respondent

Interpretation: 25 (14%) of respondents did not cite any writers; 61 (33%) cited 5.

Table 2. Total number of artists by cultural sphere

Number of artists cited by sphere	Number	Of which mentioned only once	% artists with only 1 occurrence
Writers	288	198	69%
Musicians	325	214	69%
Painters	122	71	58%

Interpretation: 288 names of writers were cited, of which 198 (69%) were cited only once.

2. 1. A Relative Cultural Diversity

23The list of the numerous artists cited testifies to a certain diversity, whether it be in the artists’ era, nationality, or their presence or not in school curricula. However, this diversity remains limited, and the list of artists is also illustrated by its absences, its voids, by regions and fields not represented in the tastes of the academics interviewed. Whole sections are absent: there is practically no rap or metal, and even less metal described as “extreme,” which implies the persistence of a rejection by the highest categories of these musical genres (Bryson 1996). Nor do we find authors of children’s or young adult literature – the consumption of which has increased sharply in recent years, even among adults (Donnat 2011). On the more commercial side, many of the best-sellers are absent. Only three (Stromae, Daft Punk, and Bob Marley) of the twenty most cited music artists in Azam et al.’s survey (2018: 348) make marginal appearances in the responses to my questionnaire. Maître Gims, Avicii, and Beyoncé are not cited at all. Nor do the usual best-sellers in literature appear: Guillaume Musso is quoted—only twice—but not Marc Lévy or Katherine Pancol. The only author correlated with the working classes in Aalborg (Denmark)’s social space survey, Danielle Steel, does not appear either (Prieur et al. 2008). Finally, there is also under-representation of avant-garde artists. From contemporary artistic movements in painting to more experimental music and with the exception of a few writers of the “Nouveau Roman” genre (Claude Simon), these academics’ tastes remain generally rather conservative in that they are orientated towards works that have long been consecrated.

24We therefore see relative diversity of preference expressed in terms of genre, period, and position in cultural hierarchies: the most popular works, in the sense that they are associated in terms of consumption with the working class, have a marginal presence, just like avant-garde artists. It remains to be seen exactly how this relative diversity is organized, by studying links between artists within each cultural sphere studied, then by bringing together these three spheres to ascertain the level of coherence of the different realms of taste thus created.

2. 2. Very Structured Tastes

25How should we process this set of responses in which some respondents gave fifteen names of artists, others only two, and where the low numbers associated with each artist make it difficult to use many of the usual statistical methods? And in turn, how can the links between different artists be analysed sociologically?

26As discussed above, network analysis provides a good method with which to process this type of data. I first put artist names and co-occurrences in the form of an adjacency matrix before making graphs using Gephi software. Artists are represented by circles (“nodes”), the size of which is proportional to their total number of citations. When two artists are cited together (by the same respondent), a link appears between them, and the more often they are cited together, the thicker the link. These networks were then spatialized. The principle is simple: nodes repel one another and links bring them closer together. Thus, two unconnected nodes repel each other when those united by many links are close. Artists are grouped together using the Louvain method (Blondel et al. 2008). This is a method of partitioning nodes, to ensure that there are the maximum links within each group and the minimum between different groups. Music, literature, and painting were processed separately, since each sphere appears to have its own specific logic in terms of organizing tastes. I maintained a threshold of two for the artists. For an artist to be present, he or she must be cited at least twice (in other words by two different individuals).

27Visualization and spatialization in the form of network analysis of the tastes of academics in literature (figure 1), music (figure 2), and art (figure 3) show different patterns of association from one sphere to another. Tastes are organized differently according to the sphere, which is itself differentiated by its structure. Literature is not the same as music or art, and each has its own methods of discovery and consumption of works.

28The space of literary tastes constructed here (figure 1) reveals principles of aggregation and opposition at work within the field of literary production in France and its particularities. One of its characteristics is its historical link with the educational institution. More than in any other artistic sphere, the norms of literary legitimacy are instilled in the entire educated population in France by the educational institution, built around a national language. This closeness can be seen in the size of group 4 and its links with groups 3, 5, and 1, which include classic and scholarly writers as well as more contemporary, but consecrated—and predominantly French—writers. There are a high number of Nobel prizewinners among these writers—both French, in Sartre, Camus, Simon, and Modiano, and foreign, in Faulkner, García Márquez, and Hemingway. Another characteristic of the French literary field is the significance of publishing houses, and among them some dominant publishers. The most classic writers as well as more contemporary but recognized writers are published by Gallimard. Namely, a well-established publisher in the field with a long history, and one that has accumulated a high volume of symbolic and financial capital (Bourdieu 1999).

30The production, dissemination, and consecration of music is less institutionalized than that of literature: we cannot easily refer to a field of production with well-established poles. Contrary to literary tastes that show an opposition between general and genre fiction, the graph of tastes in music places composers of classical and highbrow music at its centre (figure 2). The centrality of these composers is partly deceptive since it primarily results from the prevalence of Mozart, who is the most cited composer (as was the case for the questionnaire circulated in Flanders and analysed by Vlegels and Lievens [2017]). This “classic” and highbrow group has strong ties with French singers from the late 1950s to the 1970s known as à texte (such as Jacques Brel or Barbara).

31In addition to these composers and singer, several groups are organized according to the genre and era of the artists they contain. There is a time continuum at the bottom of the chart, from rock bands from the 1960s and 1970s (The Rolling Stones, The Beatles, Neil Young, and Led Zeppelin, for example) to more recent bands on the right of the figure, through the 1980s and 1990s (Bob Marley, Nirvana, and Noir Désir), up to the 2000s (Muse, Coldplay, Linkin Park, and so on). This continuum is made up of alignment between artists of the same generation (or those who had significant success at the same time), but also of porosity. This transition from one generation to the next is marked by the progressive weakening of the links between artists: none of the rock bands of the 1960s and 1970s from the bottom of the graph has been cited with those of the 2000s. Time can be a factor of legitimacy and capital accumulation. The oldest rock music has gradually become part of heritage (Le Guern 2012), not only through its “museumization,” but also thanks to the sometimes posthumous recognition that reissues allow. The ageing and displacement of bands within cultural hierarchies also takes place at the hand of the public, who continue to listen while occupying different and higher positions in the social space.

32There were not enough painters sufficiently cited to establish valid groups (figure 3). Painting does not owe its dissemination or the consecration of artists to the educational institution (Moulin 1976). Unlike music, it probably does not play a part in the daily life of the majority of respondents. We therefore see tastes in art which are highly concentrated in a few artists, mainly Monet, Van Gogh and Picasso. The “symbolic revolution” carried out by Manet in the second half of the nineteenth-century rendered painters linked to the Academy obsolete, and led to the advent of a new taste. Probably more immediately than Edouard Manet, Claude Monet established himself as an average and safe taste in painting, after having sparked the ire of art critics during his lifetime (ibid.: 403). Beyond Monet, among the painters linked to him are the Impressionists and their successors (Gauguin) who, when the Impressionists themselves had for all intents and purposes become into the new academicism, came to replace them (Bourdieu & Charle 2013: 573). These artistic tastes correspond to the collection of the Musée d’Orsay, which opened in 1986 to exhibit works from the second half of the nineteenth-century and which has a large collection of Impressionist and Post-Impressionist painters.

33Nevertheless, we also find older painters; these are generally representative of the highest-ranked genre of past centuries—religious or historical painting (mainly De Vinci, but also Bosch). The painting category also accommodates more contemporary artists, with the presence of Andy Warhol, Pierre Soulages, Jean-Michel Basquiat and Gerhard Richter, among others.

34It now remains to be seen how these different tastes overlap, occur side by side, or do not coincide when viewed together. Do those who listen to Bach also read Musso?

3. Coherence and Transgressions

35In order to best analyse this question of the coherence of tastes and obtain data relating to the transferability of aesthetic dispositions, I built a two-mode network using Pajek software. This time, it was no longer a question of linking artists to one another, but of linking individuals to artists and then grouping individuals who tended to cite the same artist combinations. The objective is to observe the social composition of the different clusters and link the socio-demographic properties of respondents to the groups of artists cited. This approach is similar to the use of correspondence analysis (CA), which is often used in the sociology of cultural practices in that the socio-demographic properties of individuals do not play a role in the construction of the taste clusters. But unlike CA, it links individuals more than it shows relationships between different positions relative to each other within a theoretical space (Duval 2012).

3. 1. Tastes and Clusters

36Network analysis reveals nine clusters grouping a variable number of individuals and artists (Table 3). Since the numbers are low, only the most salient social characteristics appear. The naming of these clusters took into account both the association patterns detailed in the previous section and the social characteristics of the individuals associated with them. A group of tastes in literature does not always strictly correspond to a group of tastes in art or music and vice versa. The different structures of content and consumption prevent a perfect correspondence of tastes. Some groups are also defined more by one or two spheres than by another: group 7, that of “contemporary success” is defined first by music and then by the presence of Monet. Affinities are established nonetheless and, like the diversity of artists, the mix of groups is relative.

 

 

 

 

 

Table 3. Clusters linking artists and individuals

Classes	Numb. artists	Numb. individuals	Main artists	Salient social characteristics	Affinity clusters
Group 1 Imaginaries	35	26	J.R.R. Tolkien, Klimt, S. King, Bosch, Pink Floyd, Asimov, Boticelli, Lovecraft, Magritte, Metallica, Linkin Park, Iron Maiden, AC/DC…	Masculine (88,5% men), Men aged 45-54 years old (50% of the group, and 50% of the 45–54 years old men), scientific disciplines.	Group 4, Group 2, Group 7
Group 2 Classic legitimacy	23	23	Van Gogh, Mozart, E. Zola, Bach, F. Dostoïevski, J. Brel, G. Brassens, Klee, Barbara, Cezanne, G. de Maupassant, Brueghel, Beethoven…	Socially undifferentiated. Concentration of “inactive mother”.	Group 4, Group 5, Group 7, Group 3
Group 3 Newly ennobled	30	20	Miro, A. Camus, D. Bowie, Turner, Renaud, Caravage, Noir désir, Soulages, Chopin, M. Yourcenar, J. Kerouac, J.-P. Sartre…	Young (80,5% of less than 45 years old).	Group 2 and 4
Group 4 The average taste	31	20	Picasso, Queen, Matisse, A. Bashung, F. Vargas, K. Follett, H. Mankell, Gauguin, H. Murakami, A. Indridason, Christine and the Queens, Friant, O. Adam, A. Gavalda, A. Souchon…	Rather feminine. None of the youngest (25–34). Midlde-class/working-class parents.	Group 2
Group 5 The sense of Distinction	31	15	Chagall, M. Kundera, Modigliani, Vermeer, G. Garcia Marquez, Schubert, G. Flaubert, Bacon, Bonnard, C. Simon… Her ewe find “established” authors from abroad: Buzzati, Fante, Mishima, Zweig… Only one rock band : Led Zeppelin.	60% of the group has at least a parent who is a teacher/an academic 0% of father OR mother working-class.	Group 2, Group 9, Group 7
Group 6 Rock/jazz classics and noir fiction	17	11	Kandinsky, The Rolling Stones, H. Coben, J. Irving, D. Simmon, L. Armstrong, N. Simone, Dali, A. Christie, B. Werber, Rothko, The Beatles…	Scientific disciplines.	Group 2, Group 4
Group 7 Contemporary success	24	19	Monet, Muse, Manet, U2, J.-J. Goldman, Coldplay, Calogero, de la Tour, A. Nothomb, P. Lemaitre, C. Lackberg, D. Pennac, J. Dicker…	Mostly female “Recent” academic disciplines: Physical education and sport science, education sciences, information and communication studies	Group 4, Group 2, Group 1
Group 8 Nonconformists artists	26	11	Radiohead, Renoir, The Doors, B. Vian, Neil Young, L. Tolstoï, M. Duras, Nirvana, Philip K. Dick, R. Gary, The Cure, Arthur H, Basquiat…	Upper-class parents	The most closed-off
Group 9 History and heritage	20	10	H. de Balzac, U. Eco, De Vinci, Courbet, V. Hugo, Delacroix, J. Austen, S. Gainsbourg…		The lest closed-off: Group 2-1-4-7…

Source: author’s database.

38In addition to their artists, these clusters are defined and distinguished by the social characteristics of the respondents who form part of them. The structure of the tastes of the academics studied is determined by combining their social background, gender, age, and generation. Some taste groups appear to be built primarily around the gender of respondents (1,7), others of the same age (3), sometimes both (4), as well as social background (5, 8) and academic discipline (1,6,7). Others are undifferentiated in terms of age and gender. Being part of a specific class fraction, that of academics, does not have a unique and uniform effect on all members. Professional socializations modify or reinforce prior cultural socializations, according to the tastes that are valued within the university sub-space they occupy: group 1 is predominantly male and composed of academics in scientific disciplines (where female representation is low). These academics therefore evolve on a daily basis in a masculine peer group in which classic scholarly culture likely plays no role. Group 2, that of classic legitimacy, is socially undifferentiated. We may surmise that this is fairly typical of an academic group as a class fraction with strong institutionalized cultural capital. Group 5, “the sense of distinction” stands out both for its high concentration of academics who themselves have at least one teacher parent and for the complete absence of academics from a working-class background. The longer history of cultural capital in this group means that members tend to choose works that are less accessible or less well known than the classics; their choices include writers who are less linked to the school system than those chosen by the newly ennobled, constituting a form of “free culture” (Bourdieu & Passeron 1964).

39The clusters are not closed off, and artists who are in separate clusters are nevertheless regularly cited together. The affinity between clusters was calculated by cross-referencing the cluster of individuals and that of the artists they cited. The aim was to track how many individuals in a given cluster cited artists from another cluster to establish which clusters had the most links between one another. In fact, two clusters have links with almost all the others: that of “classic legitimacy” and that of “average taste.” ‘Classic’ culture is prevalent in all tastes, principally via its three main nodes: Mozart, Van Gogh, and Zola.

40Different taste clusters—which are porous to varying degrees—coexist, therefore, within the academic population. These clusters have a certain coherence, both in aggregation of artists (whose positions, when related to their respective, variously structured fields, can be considered quasi-analogous) and individuals (for whom social background, gender, age, and discipline act together to determine tastes). The occupational group has undergone major morphological transformations in recent decades: from 8,131 members in 1960, or 0.03% of the active population, to 89,400 in 2005-2006, or 0.4% of the active population (Musselin 2008: 3). This strong influx of new members has opened the door to new tastes, as has happened in the wider “managers and higher intellectual professions” group (Coulangeon 2010). At the same time, internal balances and power relations between disciplines have evolved. The relative decline of literature and humanities has contributed to the reversal of a form of classic exclusivism, favouring instead the importation of genres such as rock within classic tastes. However, while such importation has taken place, this has happened in group 5, “the sense of distinction,” via the classics of the genres in question (Led Zeppelin is the only rock band linked to this group). This indicates a form of “enlightened eclecticism” at work within a fraction of the non-Parisian academics studied, which links artists at the frontier of scholarly culture or on their way to crossing it with this classic culture (Coulangeon 2003). This is consistent with the maintenance of class boundaries, which is visible in the absence of artists associated with the working class. This is consistent with observations established in aspects of the lifestyles of the upper classes other than aesthetic taste (for example, on the culinary dislikes of American progressive elites, see Tissot 2013).

41The joint analysis of the tastes and social profile of respondents suggests academics to be part of a profession affected by various power relations at work in the general population (social background, sex, age, discipline) as opposed to a whole “eclectic”, “omnivorous” or “dissonant” class fraction. It is therefore necessary to take into account the interactions between these different power relations in order to establish a scale of legitimacy of tastes (Noues, Robette & Roueff 2021). Since this is an academic population, the most dominated tastes of the scale constructed by Robette and Roueff, which was based on a general population survey (linking romantic literature to female respondents, for example), are absent from the clusters presented. The multidimensional structure of tastes implies a non-linear scale, ranging from “the sense of distinction” cluster, whose members have strong inherited cultural capital, to that of “contemporary success,” whose members are more likely to be women and in disciplines with less seniority within the academic field. We observe that the gender differentiation of tastes declines with the rise on the “scale of legitimacy” (ibid.: 15), while less legitimate tastes are located on the side of female tastes.

3. 2. Great Culture, Small Transgressions

42Having established these clusters and explained the logics of artist-respondent combinations, the question of coherence and, hypothetically, the transposability of taste dispositions meets a final obstacle which requires us to return our focus to the individual level. Taste clusters not being mutually exclusive, the majority of respondents cited artists belonging to several clusters. The absence of a taste cluster composed exclusively of less consecrated or popular artists owing to the presence of legitimate artists in many clusters negates any attempt to associate a degree of legitimacy with a cluster and then carry out a series of counts in order to establish individual “profiles.” It is nevertheless possible to answer the question of distribution, at the individual level, of these multi-group citations. To do this, I proceeded in two ways: on the one hand, I noted artists cited by respondents who combine artists from the most legitimate groups (“classic legitimacy” and “the sense of distinction”) with other artists from the least legitimate groups; and on the other hand I observed artists who are never combined. Which types of artists are combined or, conversely, never cited together? What great cultural gaps emerge thanks to these observations?

43We can see relative cultural “transgressions”: co-citations of very legitimate artists with others who are less so. If such transgressions were merely the result of the multiple secondary socializations of each individual, they would be unpredictable or to some extent disordered. However, these transgressions obey rules: there are transgressions that are acceptable and those that are ostensibly not.

44“Transgression”—or great cultural gaps—happens, in fact, through music. In other words, while we find groups from “contemporary success” within the musical sphere combined with very legitimate artists, we do not find composers other than Mozart combined with writers of “average” literature. For example, one respondent combined U2 with Kundera and García Márquez. We see David Guetta cited with Houellebecq and Dostoevsky, Linkin Park with Sartre, and there is even a co-citation of Lady Gaga and Flaubert. Painters from the most legitimate groups are also combined with music groups with less legitimacy: French punk band Les Wampas are combined with Chagall, and Coldplay with Soulages. But Bach, Chopin, Wagner, Schubert, Schumann, Rachmaninov, Mahler, Debussy, Bartok, and Dvorak are only ever combined with very legitimate writers. The only exception to this is science fiction and fantasy classics: one interviewee cited Bach and Mahler with Lovecraft; another Bartok and Debussy with Tolkien.

Conclusion

47The morphology of academics’ tastes shows that “ascetic aristocratism”—a disposition that is best expressed for Bourdieu by this class fraction in the 1960s—is fairly resistant to the transformations of the group and the spaces of production. Its considerable relationship to culture materializes in preferences for older composers, writers, and artists. This canonical corpus of works and artists evolves carefully, with the addition of more contemporary artists but those whom have already been strongly consecrated. While other cultural dispositions such as alternative or average taste are now expressed within the group, hybridization occurs in two ways: 1) by mixing “classics” with more popular or commercial tastes (relative to other tastes present) for new entrants; and 2) by importing music tastes that are further from the canons of classic legitimacy associated with highbrow writers or painters for academics occupying established positions. The question that remains, with regard to the decline of scholarly distinction seen primarily within the upper classes (Coulangeon 2021)—by which academics appear to be less affected—is that of the symbolic influence of these academic tastes beyond the group boundaries. Classic tastes prevail throughout this fraction of the upper classes, but do they continue to impose themselves as legitimate outside it? Faced with competing upper fractions who are also increasingly graduates, whose “elites”, even the academics, are turning away from public sector professions (Bataille & Falcon 2022), the question deserves to be asked.