Do more experienced critics review differently? : How field-specific cultural capital influences the judgments of cultural intermediaries

Abstract

Purpose

Cultural intermediaries define the standards many consumers use when evaluating cultural products. Yet, little research has focused on whether cultural intermediaries may systematically differ from each other with regard to the standards they emphasize. The purpose of this paper is to build on Bourdieu’s theory of cultural production to examine how the type of subfield reviewed and/or the cultural intermediary’s expertise (or “field-specific cultural capital”) affect the standards an intermediary uses.

Design/methodology/approach

This paper employed a computer-aided content analysis of the full corpus of “Rolling Stone” music album reviews (1967-2014).

Findings

Critics with lower field-specific cultural capital reflect the same logic as the subfield they are critiquing. Critics with higher field-specific cultural capital reflect the opposite logic.

Research limitations/implications

Bourdieu was ambivalent about whether cultural intermediaries will reflect the logic of a subfield. Results show that the answer depends on the intermediary’s field-specific cultural capital. The results also reinforce previous findings that individuals with high field-specific cultural capital are more likely to break with the logic of a field.

Practical implications

Not all intermediaries are created equal. Producers and consumers who rely on cultural intermediaries should understand the intermediary’s critical analysis within the context of his/her experience.

Originality/value

This is one of the first studies to examine how a cultural intermediary’s field-specific cultural capital impacts his or her work. The findings are based on a large review sample and include reviewers’ analyses as they developed from having lower to higher field-specific cultural capital.

Citation

Corciolani, M., Grayson, K. and Humphreys, A. (2020), "Do more experienced critics review differently? : How field-specific cultural capital influences the judgments of cultural intermediaries", European Journal of Marketing, Vol. 54 No. 3, pp. 478-510. https://doi.org/10.1108/EJM-01-2019-0095

1. Introduction

When determining the value of a product or service, consumers do not rely only on their personal judgment or on information about the producer. Consumers also frequently consult information from people who are experts in the product or service category. Scholars refer to these experts as “cultural intermediaries” – a term that includes theater critics (Shrum, 1991), advertising agencies (Gurrieri et al., 2016), bartenders (Ocejo, 2012), food marketers (Jamal, 2003), television programming buyers (Kuipers, 2012), fashion models (Lonergan et al., 2018), restaurant guides (Lane, 2019), personal trainers (Maguire, 2008), fundraisers (Banks, 2019), book publishers (Childress, 2012) and music critics (Glynn and Lounsbury, 2005). Although these various professional roles differ from one another in many ways, they all serve the function of helping consumers understand and evaluate what a producer has created. As a result, success or failure in a market often depends on a cultural intermediary’s interpretation and influence (Humphreys and Carpenter, 2018; Shrum, 1991). Some producers therefore carefully consider the potential influence of cultural intermediaries when developing their products and services (Booth and Matic, 2011). But, what drives a cultural intermediary’s appraisals? Do intermediaries tend to apply standards that are similar to the standards used in the domains they are critiquing or do they apply different standards? This is not only an important practical question for companies and producers that rely on the cultural intermediaries’ evaluations but also an interesting theoretical question for those interested in theories of cultural production (Bourdieu,1971/1985, 1993).

This paper proposes that the answers to these questions depend on the level of experience (or “field-specific cultural capital”) a cultural intermediary has. The results of an automated text analysis of music album reviews from a major music magazine show that, when considering the authenticity of artists and their music, critics with low field-specific cultural capital use standards that are similar to the standards of the artists they are critiquing. In contrast, those with high field-specific cultural capital differentiate by applying different standards. These findings shed light on the status patterns that drive critics who, in turn, shape the cultural conversation that drives markets.

2. Theoretical framework: fields, cultural intermediaries and cultural capital

Throughout his career, sociologist Pierre Bourdieu (1969, 1971/1985, 1984, 1992/1995) developed and refined a unified framework for analyzing and understanding cultural products such as films, classical music and popular music. This paper uses Bourdieu’s framework as a lens for understanding and analyzing cultural intermediaries and for guiding the project’s key research questions. In Section 2.1, we review a central element of Bourdieu’s theoretical perspective – the concept of a field. In Section 2.2, we explain that, according to Bourdieu, fields tend to be divided into two subfields, each with different standards. One subfield focuses on popularity and economic success, whereas the other focuses on the status within the subfield. Section 2.3 reviews research on cultural intermediaries – people and organizations who mediate between those who produce cultural products and those who are the target audience for these products. Here we introduce our central research purpose, which is to examine whether cultural intermediaries mirror the standards of the subfield they are interpreting or follow a different logic.

In Section 2.4, we review Bourdieu’s concept of cultural capital and suggest that the answer to the question posed in Section 2.3 may depend on the cultural intermediary’s field-specific cultural capital. Lastly, in Section 2.5 we explain how a cultural intermediary’s standards for a cultural product (and particularly the product’s authenticity) may depend on his or her field-specific cultural capital and the type of subfield whose cultural products the intermediary interprets.

2.1 Bourdieu’s concept of fields of cultural production

Bourdieu developed his framework for understanding and analyzing cultural products because he was dissatisfied with the two well-established approaches that critics tended to use at the time. One approach was to focus on a cultural product’s inherent characteristics, usually in comparison to similar works. Bourdieu thought that these “internal readings” of cultural products ignored the influential historical and social factors that affect the product’s creation (Bourdieu,1986/1993, p. 178). Another approach focused on the influence of social and historical factors, but Bourdieu felt that this “external mode of analysis” did not sufficiently consider the producer’s personal autonomy (Bourdieu,1986/1993, p. 180). Bourdieu therefore sought to develop an analytic perspective that simultaneously incorporated the influence of social/historical influences, personal factors and attributes of the work of art itself (Bourdieu,1986/1993).

Bourdieu’s solution was to introduce the concept of a “field” (Bourdieu and Waquant, 1992). A field is a domain of activity in which people compete with each other for desirable resources. Examples of fields include fashion (Dolbec and Fischer, 2015), religion (McAlexander et al., 2014), advertising (Chávez, 2012), indie consumption (Arsel and Thompson, 2011) and commercial music (Anand and Peterson, 2000). When explaining the concept of a field, Bourdieu and those building on his work sometimes use the metaphor of a game: For instance, they describe those participating in the field as “players” trying to win competitively against each other and who must follow a set of “rules” to “win” the game, where winning means achieving key benefits such as status, material resources and/or social connections (Bourdieu and Waquant, 1992; Cosckuner-Balli and Thompson, 2012; Drumwright and Kamal, 2016; Lonergan et al., 2018). Because different fields have different rules and are differentially likely to lead to particular benefits, each field will naturally attract different players – frequently those who are suitably qualified for competition in the field. Consequently, the players who compete in each field tend to have similar backgrounds (or “habitus”), similar goals (or “stakes”) and a common understanding of the rules of the game (or “doxa”). In Warren and Dinnie’s (2018, p. 303) words:

Fields are formed from networks of social relations; they are competitive environments in which social actors leverage their own habitus to compete for placement – for economic, cultural, social and symbolic power.

The concept of a field, which has been applied to many domains beyond cultural production, solves the analytic problems that Bourdieu identified with previous approaches to analyzing cultural products: A field simultaneously considers not only social and historical factors (because these factors strongly influence the field participants’ habitus and understanding of the field’s doxa) but also personal agency (because people are free to pursue their personal strategies within the doxa) and the characteristics of the works themselves (which are understood in relation to each other and in relation to habitus and doxa).

For example, in February 2019, the female pop singer, Arianna Grande, released a number-one song in the USA. called “Thank U Next.” According to Bourdieu, a full and accurate understanding of this song requires identifying the other players in the field in which Grande is competing, becoming acquainted with the backgrounds and upbringings of these players, characterizing the personal and professional relationships among them and specifying the positions each has taken in the field with their activities – a specification that requires comparing and contrasting the works of art that each player has produced (Bourdieu,1986/1993, pp. 183-184). This approach might reveal, for example, that “Thank U Next” was successful because:

• It is musically similar to previous hit songs by Dua Lipa and Camilla Cabello (other players in the field).

• The song references themes regarding the empowerment of women that are personally important to artists in this field (their backgrounds and upbringing).

• It incorporates lyrics that are more specifically personal than the works by other artists in the field (the characteristics of the works themselves).

2.2 The subfields of cultural production: two logics, two sets of standards

Bourdieu proposed and observed that, as players compete in a field, they are likely to coalesce into two subfields (Hesmondhalgh, 2006), each of which operates according to its own logic (see Table I for a summary of these two subfields’ contrasting logics). An important difference between these two logics is the stakes that are most important to participants (Bourdieu,1971/1985). In the first subfield, participants care most about gaining status (symbolic capital). Bourdieu referred to this subfield as the field of restricted production or the avant-garde. In the second subfield, participants care most about gaining popularity (economic capital). Bourdieu referred to this subfield as the field of large-scale production. For example, in the field of cinema, cultural products from the subfield of large-scale production would include most blockbuster action films, whereas products from the subfield of restricted production would include art-house films, which, although earning critical acclaim, might be shown in only a few movie houses.

Bourdieu observed that a key motivation for those participating in the subfield of restricted production is to reject the logic of the subfield of large-scale production. Consequently, the tensions between the two subfields influence the activities in each. Fully understanding a cultural product therefore requires analyzing not only the players and cultural products produced within a particular subfield but also the players and products in corresponding subfields with different logics. For example, understanding Ariana Grande’s music requires not only comparing it with others in the same field of large-scale production but also with music by artists such as 12th Planet, whose work is categorized in the niche music genre of dubstep and follows the logic of restricted production. According to Bourdieu (1971/1985), aiming for either economic or symbolic capital encourages players in different subfields to focus on different audiences. On the one hand, those competing for popularity (and thus economic capital) care most about acceptance from the “public at large” (p. 17) – an audience comprised of consumers (not producers) of cultural products. Importantly, this audience needs to be sufficiently large so that, if its members choose a cultural product, the product’s creators and distributors can earn significant economic capital. On the other hand, those competing for status (and thus symbolic capital) care most about recognition from other producers in the subfield. Other producers in the subfield are much more familiar with the rules of the game than the consumers are and are themselves competing for status in the subfield. This audience is therefore most appropriate for determining the status of others in the subfield, but is also much narrower than the audience of consumers and potential consumers that the field of large-scale production targets. This explains why Bourdieu refers to this as the subfield of restricted production, and why players who win in this subfield generally cannot achieve the material resources those who win in the subfield of large-scale production earn.

To Bourdieu (1986/1993), the most important difference between the two types of subfields is the players’ autonomy when creating cultural products and influencing the field’s rules. Because the subfield of large-scale production is centrally concerned with eliciting positive reactions from a large audience of consumers, the players producing and disseminating cultural products must pay special attention to these consumers’ tastes and preferences or risk not gaining the stakes that are important to the field, namely, the “conquest of the largest possible market” (Bourdieu,1971/1985, p. 17). The subfield of large-scale production is therefore more “heteronomous” – it is more influenced by the standards and expectations of those outside of the field of production. Consumer expectations for product features and genres are relatively well-established in the subfield of large-scale production, and producers who compete in these fields therefore achieve “success and the corresponding profits by adjusting to pre-existing demand” (Bourdieu, 1993, p. 82) – that is, pre-existing preferences and tastes.

In contrast, those competing in subfields of restricted production are more concerned with influencing others who are competing in the subfield. Therefore, the standards and expectations of those outside of the field are less relevant. Members of the general public are not only incapable of granting the kind of status these cultural producers seek but are also insufficiently informed about how to accurately judge cultural products produced in this subfield. Consequently, the producers in the subfields of restricted production are more “autonomous” – they are “freed from the censorship and auto-censorship consequent on direct confrontation with the public foreign to the profession” (Bourdieu,1971/1985, p. 17). In fact, as Bourdieu (1969, p. 92) argues, those who compete in subfields of restricted production often enhance their status by emphasizing their “claim to independence” and their “indifference to the public.”

Although Bourdieu often discussed the subfields as if they are two separate and distinct types, he also recognized that the factors defining a subfield are continuous rather than binary. “Within a single universe,” he wrote:

[…] one always finds the entire range […] between works produced with reference to the restricted market on the one hand and works determined by a representation of the expectations of the widest possible public (Bourdieu,1971/1985, pp. 29-30).

Whether a subfield is heteronomous or autonomous is therefore more a question of degree than of kind. As Bourdieu observed, even authors who produce in fields of restricted production must consider how audiences outside of the field respond, partly because acceptance from these audiences is potentially problematic to those seeking status in the field of restricted production. “Even the author most indifferent to the lure of success and the least disposed to make concessions to the demands of the public,” Bourdieu (1969, p. 97) wrote, “is surely obliged to take account of the social truth of his work as it is reported back to him.”

This paper compares and contrasts the critical reception of musical products produced in relatively autonomous fields to the reception of products produced in relatively heteronomous fields. More specifically, this paper analyzes the music reviews by critics who write for Rolling Stone magazine, an outlet that focuses on popular and niche music genres. As stated, researchers in sociology and business have referred to such individuals as cultural intermediaries and the next section provides a brief review of this research.

2.3 Cultural intermediaries: definition and research summary

Bourdieu analyzed not only how social factors influence a cultural product’s production, but also how social factors influence its meaning and value after it has been produced. Bourdieu proposed that as part of their effort to understand and appreciate a cultural product and to integrate it into their lives, audiences create public “myths” about the creators and their works. These myths are descriptions of, and narratives about, the work, how it was created and what it symbolizes. These myths are influenced by not only what is happening in the artist’s subfield of cultural production but also what is happening in the fields in which the myth-makers and other audience members are competing. Bourdieu observed that certain professions focus on creating and promoting these public myths. An example that Bourdieu frequently discusses is the art critic, who tends to take responsibility for analyzing and explaining the fine arts, such as literature, painting and sculpture (Bourdieu, 1969, 1971/1985, 1992/1995). Bourdieu also argued that in a consumer marketplace that focuses increasingly on selling symbols and meaning, cultural products such as perfume and clothing – and even cleaning products and breakfast cereals – require their own public myths. Consequently, those producing these products require professionals to develop, translate and transmit these myths to the appropriate audiences – professionals who include, for example, advertising agency executives, property developers, tourism managers and market research experts (Bourdieu, 1969, 1984).

These observations, as well as similar observations by other authors and thinkers, have inspired researchers to develop a stream of research that focuses on cultural intermediaries (Adkins, 2011; Coulter et al., 2003; Jamal, 2003; Lonergan et al., 2018; Shrum, 1991). A cultural intermediary is someone whose role is to mediate between the producer’s needs and the product’s consumer (Cronin, 2004, p. 350; Kobayashi et al., 2018; Wright, 2005). A prototypical cultural intermediary has two key qualities. First, his or her primary professional function is to explain or frame the value and significance of a symbolic good or service to particular consumers or other audiences (Adkins, 2011; Coulter et al., 2003; Durrer and Miles, 2009; Komarova and Velthuis, 2018; Lane, 2019; Lonergan et al., 2018; Maguire and Matthews, 2012; Ocejo, 2012; Shrum, 1991). Second, to serve this function, cultural intermediaries must have a level of knowledge and expertise that is greater – in the relevant areas – than that of their target audience (Durrer and Miles, 2009; Glynn and Lounsbury, 2005; Lonergan et al., 2018; Maguire and Matthews, 2012; Parker et al., 2018; Warren and Dinnie, 2018). In certain markets, structural or cultural factors have prevented formal cultural intermediaries from emerging, in which case producers must shoulder the responsibility for translating their work (Menon, 2019; Whitson et al., 2019). However, formal cultural intermediaries exist in a plethora of markets as diverse as biology, accounting and music (Negus, 2002).

What standards are likely to be most important when a cultural intermediary translates a product or work of art? Bourdieu was at best ambivalent in his response to this question. On the one hand, he sometimes suggested that critics strongly reflected the logic of the fields they analyzed. For example, he described how the critics of works produced by a subfield of restricted production place themselves “unconditionally at the service of the artist […]” and attempt “scrupulously to decipher his intentions, whereas excluding the public of non-producers from the entire business […]” (Bourdieu,1971/1985, p. 18). Bourdieu also described how cultural intermediaries in subfields of large-scale production sometimes mirror their subfields’ logic (Kuipers, 2012; Maguire, 2008; Ocejo, 2012). These intermediaries “sell so well because they believe in what they sell […] [and are] predisposed to collaborate with total conviction” in the dissemination of the subfield’s values (Bourdieu, 1984, p. 365). On the other hand, Bourdieu sometimes suggested that critics are more likely to follow their personal logic rather than the logic of the field that produced a cultural product. For example, Bourdieu argues that the critical response to Flaubert’s “Sentimental Education” was due more to the fields in which the critics were working than to the novel’s attributes or the attributes of the field in which Flaubert was operating (Bourdieu,1992/1995, pp. 100-103). Bourdieu also sometimes referred to the artists’ and critics’ influence on each other as a system of circular causality (Bourdieu,1971/1985, p. 21, 1984, p. 99, 1992/1995, p. 289), whereby audiences always influence the producers to a certain extent and vice versa, and where no original cause can therefore be identified.

In light of Bourdieu’s ambivalence about cultural producers’ influence on cultural intermediaries, this research explores the question of whether or not intermediaries reflect the logic, and thus apply the standards, of the producers in the fields they analyze. More specifically, this research explores the proposition that cultural intermediaries may systematically differ from each other in their work because they operate within their personal fields and will therefore leverage their individual qualities to maximize their success in relation to other intermediaries (Chávez, 2012). In the next section, we identify field-specific cultural capital as a key difference that exists among cultural intermediaries in all fields and which, we argue, will significantly influence the intermediary’s work.

2.4 Field-specific cultural capital affects how a person competes in a subfield

A key factor that differentiates players in a field is their “field-specific,” “localized” or “field-dependent” cultural capital (Cosckuner-Balli and Thompson, 2012; Kates, 2002; McAlexander et al., 2014; McQuarrie et al., 2012; Saatciojlu and Ozanne, 2013). All of these terms refer to resources that a player develops while learning about and competing within a particular field. These resources can include expertise and knowledge, social connections, material resources and/or prestige (Cosckuner-Balli and Thompson, 2012). By definition, field-specific cultural capital is more helpful to players when they are competing in the field where the capital was developed than when competing in other fields (whereas field-independent or general cultural capital is helpful to players across multiple fields).

This paper analyzes how a cultural intermediary’s level of field-specific cultural capital influences his or her work. (We operationalize field-specific cultural capital in terms of experience – how long a person has been playing the role of cultural intermediary, and how much work the person has done in the field.) The influence of a cultural intermediary’s field-specific cultural capital has received some attention from researchers. For example, McQuarrie et al. (2012) show that, as cultural intermediaries in social media gain field-specific cultural capital, they need to renegotiate their relationship with consumers. Also Komarova and Velthuis (2018) show that a cultural intermediary’s success may depend more on the institutional context and economic resources available to the intermediary than on the intermediary’s field-specific cultural capital. Yet, no research to our knowledge has examined how an intermediary’s field-specific cultural capital may influence the standards that he or she applies to cultural works, and whether those standards are similar to or different from the standards of the people producing the cultural works.

Notably, a few studies have demonstrated that a consumer’s level of education and experience in a field – and, thus, his or her field-specific cultural capital – can influence the standards that he or she deems important when making consumption choices in the field. This previous research has suggested that those who have been competing in a field for a relatively short time tend to pursue stereotypic strategies that strongly reflect well-known assumptions about the standards in the subfield. In contrast, those who have been competing for a relatively long time tend to purposefully choose strategies that are more individualistic and which purposefully differ from the new participants’ strategies. For instance, Kates (2002, p. 390) notes that men who are relatively new to the homosexual community are more likely to rely on gay stereotypes to identify other gay men and guide their own behavior, whereas those who have been in the community for a while have a “more refined” approach, which often results in choices that violate the stereotype. As another example, Arsel and Thompson (2011) show that those who have been long-time players in the hipster field tend to make more individualistic choices and, therefore, tend to eschew choices that have become stereotypic of the field (and which those who are relatively new to the field are more likely to select).

These findings paint distinct pictures for those with less versus more field-specific cultural capital [although, see Beunza and Garud (2007) for an alternative perspective]. Those who are relatively new to a field first need to learn how to play according to the field’s rules, which often means making choices that raise as few doubts as possible about whether the player knows the rules and can competently follow them. In contrast, those with greater field-specific cultural capital are more likely to have the confidence, status and sophistication to veer away from standard and expected choices. They know how to play by the rules, but they enhance status by playing with the rules in a way that distinguishes themselves from those with less field-specific cultural capital. Based on these findings, we predict that the standards that cultural intermediaries with less field-specific cultural capital apply are more likely to reflect the logic of the field they are reviewing. In contrast, the standards that cultural intermediaries with more field-specific cultural capital apply are less likely to reflect the logic of the field they are reviewing. In fact, because those with more field-specific cultural capital seek to distinguish themselves from those with less, we predict that relatively high field-specific cultural capital can sometimes lead a cultural intermediary to follow a logic that is opposite to the field he or she is reviewing.

2.5 How field-specific cultural capital and field autonomy may affect a cultural intermediary’s perspective on authenticity

To investigate questions about how field-specific cultural capital (i.e. a cultural intermediary’s experience) might interact with field autonomy (i.e. whether a field is more heteronomous or autonomous), this paper focuses on how music critics discuss a cultural producer’s authenticity. Research has shown that consumers value authenticity (Grayson and Martinec, 2004; Newman and Dhar, 2014; Rose and Wood, 2005) and that the authenticity concept can be generally defined in one of two ways. On one hand, authenticity can mean being true to one’s self. This paper uses Carroll and Wheaton’s term, “moral authenticity,” to refer to this type of authenticity, which is also known as “self authenticity,” “expressive authenticity,” “indexical authenticity” or “existential authenticity” (Dutton, 2003; Grayson and Martinec, 2004; Newman and Smith, 2016; Wang, 1999). On the other hand, authenticity can mean being true to a certain established type or genre. To refer to this kind of authenticity, this paper uses Carroll and Wheaton’s (2009) “type authenticity,” which is also known as “iconic authenticity” or “categorical authenticity” (Grayson and Martinec, 2004; Newman and Smith, 2016). This project focuses on authenticity because Bourdieu offers specific predictions about how the logic of restricted production and the logic of large-scale production are likely to demand different types of authenticity. This in turn allows us to predict what authenticity standards are likely to be most important to producers in each field, and to compare that prediction with the type of authenticity discussed by critics.

Academics define moral authenticity as not responding to standards and listening to your inner voice. In Holt’s (2002, p. 83) words, a product or brand is authentic if it is “perceived as invented and disseminated by parties without an instrumental economic agenda, by people who are intrinsically motivated by their inherent value.” A morally authentic producer is a producer who is “sincere, assumes responsibility for his/her actions and makes explicit value-based choices […] rather than accepting pre-programmed or socially imposed values and actions” (Carroll and Wheaton, 2009, p. 261). According to Bourdieu, artists producing in the fields of restricted production prioritize moral authenticity. In these subfields:

[…] the true subject of the work of art is nothing other than the specifically artistic manner in which the artist grasps the world, those infallible signs of his mastery of his art (Bourdieu,1971/1985, p. 20).

Subfields of restricted production therefore “exclude those artists suspected of submitting to external demands” (Bourdieu,1971/1985, p. 20). Artists in fields of restricted production ignore or purposefully violate established standards – an artistic decision that they expect, and even hope, will result in rejection from those who anticipate and prefer cultural products that adhere to standards:

[T]he structural gap between supply and demand [in the field of restricted production] contributes to the artists’ determination to steep themselves in the search for “originality” (with its concomitant ideology of the misunderstood genius) (Bourdieu,1971/1985, p. 22).

In contrast, Bourdieu expected subfields of large-scale production to produce works that adhere to standards and which therefore focus on type authenticity. These works rely on “immediately accessible technical processes and aesthetic effects, or the systematic exclusion of all potentially controversial themes or those liable to shock this or that section of the public” (Bourdieu,1971/1985, p. 28). As an example, Bourdieu mentions Westerns—films that “have to work within the very strict conventions of a heavily stereotyped genre” and which are:

[…] continually referring back to previous solutions – assumed to be known [by the audience] – in the solutions they provide to canonical problems, and they are continually bordering on pastiche or parody of previous authors, against whom they measure themselves (Bourdieu,1971/1985, p. 30).

Bourdieu (1971/1985, p. 30) also refers to works produced by subfields of large-scale production as being “characterized by tried and proven techniques and an oscillation between plagiarism and parody.” This is type authenticity or “a focus on whether the object meets the criteria for inclusion or membership in a type of genre or category,” and which “presupposes the existence of the […] type or genre, which is a culturally defined classification” (Carroll and Wheaton, 2009, p. 261). As a result of this expectation for adherence to standards, expressions of self – especially those that purposefully ignore or violate standards – are not valued as highly in subfields of large-scale production:

Original experimentation entering the field of large-scale production almost always comes up against the breakdown in communication liable to arise from the use of codes inaccessible to the “mass public” (Bourdieu,1971/1985, p. 32).

Consequently, while the logic of large-scale production prioritizes adherence to standards, it makes little room for expression of self: Works produced for the field of large-scale production are “most often the culmination of transactions and compromises among various categories of agents,” who “use their specific competencies to guarantee a wide variety of cultural interests while simultaneously reactivating the self-censorship engendered by the vast industrial and bureaucratic organizations of cultural production” (Bourdieu,1971/1985, p. 29).

Bourdieu’s predictions about the value of different kinds of authenticity in different subfields of cultural production have been supported in the context of popular music, which is this project’s context. For instance, Peterson’s (1997) work on authenticity in country music supports the idea that subfields of restricted production are more likely to validate expressions of self (i.e. moral authenticity), whereas subfields of large-scale production are more likely to validate adherence to standards (i.e. type authenticity). According to Peterson (1997), this progression is because of the fact that younger, smaller subfields are less likely to have standards that artists and audiences alike accept and institutionalize. Thus, artists producing works in younger subfields cannot rely on type authenticity as a basis for authenticity and can rely only on moral authenticity. In contrast, more established subfields have been market-tested through trial and error and have established the expectations that are required for success. Peterson (1997, p. 223) observes that, from the inception of country music in the early 1920s to its popularity in the late 1990s, what counted as authentic evolved in the music genre. At the genesis of country music, there was:

[…] no clear tradition […] with its own past, its own iconic progenitors, its own institutional delivery system, and its own self-conscious fan community. There was then no shared understanding of what constituted country music as a distinct genre.

However, after years of institutionalization and crystallization of genre expectations, “artists seeking to establish the bona fides of authenticity now have available a set of signifiers that had not been codified in 1953” (Peterson, 1997, p. 255). (Stavraki et al. (2018) offer a similar analysis of how consumers who are more familiar with well-known signifiers interpret art differently than those who are less familiar.)

To summarize, Bourdieu proposed that cultural production industries are likely to be divided into two general kinds of subfields. The first kind – subfields of restricted production – is more autonomous and, thus, more likely to value moral authenticity than type authenticity. The second kind – subfields of large-scale production – is more heteronomous and, thus, more likely to value type authenticity than moral authenticity. Building on our earlier discussion of field-specific cultural capital and standards (where those with lower field-specific cultural capital are more likely to follow a subfields standards), we therefore predict the following:

H1.

For more heteronomous fields, (a) type authenticity is more likely to be discussed by cultural intermediaries with lower (versus higher) field-specific cultural capital and (b) moral authenticity is less likely to be discussed by cultural intermediaries with lower (versus higher) field specific cultural capital.

H2.

For more autonomous fields, (a) type authenticity is less likely to be discussed by cultural intermediaries with lower (versus higher) field-specific cultural capital and (b) moral authenticity is more likely to be discussed by cultural intermediaries with lower (versus higher) field-specific cultural capital.

3. Data and methods

This project uses a quantitative, automated content analysis of critics’ music album reviews. Researchers generally use automated content (or textual) analysis to codify text into groups or categories based on selected criteria, aiming to convert the qualitative data into quantitative measures that can be statistically analyzed (Weber, 1990, p. 18). This approach allows researchers to make “replicable and valid inferences from texts (or other meaningful matter) to the contexts of their use” (Krippendorff, 1989, p. 18). We followed Humphreys and Wang’s (2017) recommended approach for analyses of this nature. According to these authors, automated content analysis must begin with the identification of a research question and, relatedly, the identification of constructs to be examined. Once a researcher has identified the research question and the related constructs, he or she should start collecting the data. Thereafter, the data has to be prepared, unitized and stored. The next step involves the operationalization of constructs, followed by the validation of the instruments adopted to measure them (e.g. dictionaries of keywords). Finally, the data has to be analyzed and interpreted, and the main constructs’ predictive validity must be tested. In the next sections, we describe how we followed these steps.

3.1 Research question and related constructs identification

Our hypotheses center on whether authenticity discussions are a function of reviewer field-specific cultural capital and field autonomy. Thus, for our textual analysis, authenticity was the focal construct and, more specifically, the two types of authenticity (i.e. type authenticity and moral authenticity), which have been a central focus in the literature on authenticity. We anticipate that two other constructs – a reviewer’s field-specific cultural capital and the relative autonomy of the field in which the music was produced – will affect the extent to which type and/or moral authenticity is mentioned in a music review.

3.2 Data collection and data preparation

The full corpus of Rolling Stone album reviews is available on the magazine’s official website, and we therefore used Web Content Extractor to download 4,452 reviews. Rolling Stone is a useful source for this investigation because it provides not only a large review database for analysis but also, owing to its longevity, allows estimating a reviewer’s experience (i.e. his or her field-specific cultural capital) because many reviewers worked for the magazine for several years. Furthermore, because Rolling Stone focuses on popular and niche music genres, it facilitates comparing fields with different levels of autonomy (i.e. more heteronomous versus more autonomous fields). Rolling Stone is also an influential magazine (Frith, 1983) and its output is commonly analyzed in empirical analyses of the popular music market (McLeod, 2001). Also important, the full corpus of Rolling Stone reviews during this time period is available online, allowing a comprehensive analysis of the album reviews. These methods are therefore similar in spirit to Kristensen et al.’s (2019) analysis of the Mad Men TV series’ reviews.

We collected the following data for each review: the review’s text, the review’s date, the artist’s name, the album’s name and the reviewer’s name. The resulting dataset includes the entire population of reviews published from 1967 to 2014. A total of 421 reviewers wrote these reviews. Each of these reviewers published an average of 10.61 reviews and referred to a total of 1,404 artists, who each produced 3.17 music albums on average.

Next, we identified the genres for each album. Using the tags that Last.fm associates with each artist (e.g. American, blues-rock, hard-rock, heavy-metal) and Wikipedia’s “List of Popular Genres,” a research assistant identified one music genre for each album (See Table II for more detailed information). The research assistant collected all the tags that Last.fm associated with each artist and then picked a basic music genre, which could be consistently associated with the specific artist from Wikipedia’s list. For instance, Last.fm might list an artist as being associated with the genres of folk, singer-songwriter, Americana and folk-rock. In this case, Wikipedia classifies all four of these genres as part of the basic music genre, folk. However, in some cases, Last.fm associated artists with multiple music genres. For instance, an artist might be tagged as rock, hard-rock, blues-rock and blues. This artist could be associated with either rock or blues, which, according to Wikipedia, are two distinct music genres. In these cases, we followed Last.fm’s approach, which lists an artist’s most important or definitive tag first and the less definitive tags second. Hence, the artist imagined above would be considered a rock instead of a blues musician because rock was listed first. To test this classification approach’s reliability, a second coder manually coded a random sample of 150 artists (about 10 per cent of the total). The percent agreement between the two coders was 92.00 per cent.

After identifying and collecting the data, we spell-checked the text and created an MS Word macro to segment the reviews into 4,452 separate text files, which were stored on each of the co-authors’ personal computers.

3.3 Operationalization and dictionary validation

This research used a top-down approach to construct definition, which is common for computer-aided content analysis (Humphreys, 2010; Ludwig et al., 2013), and which uses theoretical constructs as a starting point for creating custom dictionaries to be used for textual analysis. This process began with a preliminary analysis of how Rolling Stone reviewers commonly reference type and moral authenticity – constructs identified in previous research as being important and informative. We identified a random sample of 450 reviews (about 10 per cent of the total), stratified by music genres and decades. The first author followed common qualitative procedures to analyze the reviews (Spiggle, 1994), which occupied 416 (double spaced) pages in total.

More specifically, using Atlas.Ti, the first author applied open, axial and theoretical coding procedures to the reviews (Corbin and Strauss, 1990). This analysis enabled us to identify two different variables associated with type authenticity (i.e. genre reference and place reference) and two variables related to moral authenticity (i.e. personal authenticity and marketing language). Genre reference refers to words that indicate a particular type of music, such as jazz, blues or country. Relatedly, place reference refers to words that mention a particular location in association with the music – for example, Memphis which is often associated with blues, Jamaica which is often associated with reggae and ska and Detroit which is often associated with Motown. Words referencing genre and place are therefore more likely to indicate a reviewer’s consideration of whether an artist’s music lives up to expectations for a particular type of music genre. Personal authenticity refers to words mentioning an artist’s sincerity or honesty. Marketing language uses words that reference the opposite – a focus on producing music for profit, generating sales and selling out. Words referencing personal authenticity and marketing language are therefore more likely to indicate a reviewer’s consideration of whether an artist produces music from the heart or for monetary gain.

The dictionary validation stage was next (Humphreys and Wang, 2017). We refined the four dictionaries to minimize the likelihood of false negatives and false positives. First, to ensure that we captured as many of the various ways in which the key concepts could be discussed (and to avoid false negatives), we augmented the original dictionaries with potential synonyms, word stems and tenses (Humphreys and Wang, 2017), as well as additional keywords derived from the related literature on authenticity and popular music (Corciolani, 2014; Grayson and Martinec, 2004; Holt, 2002; Newman and Smith, 2016; Peterson, 1997; Trilling, 1972). Adding these terms to the dictionary increased the likelihood that we would capture a construct when it was discussed. Second, certain words identified via this process had multiple connotations, including connoted meanings that were unrelated to our concepts of interest. Because these could lead to false positives, we tested the dictionaries on the texts and analyzed ten instances at a time to remove words with a high incidence of connoting concepts other than the ones we were interested in (Weber, 2005).

Furthermore, in accordance with Humphreys and Wang (2017), as well as Pennebaker et al. (2007), we asked three external judges to assess whether each proposed dictionary should include each keyword. For instance, should the “place reference” category include the word “Liverpool”? Should the “personal authenticity” category include the word “soulful”? Words remained in the dictionary if at least two coders agreed it should be kept; we removed words only if two judges thought that it should be removed (Humphreys, 2010). This process produced the dictionaries that are described in Table III and which we used for our subsequent analysis.

3.4 Analysis and interpretation

Next, we analyzed the relationships among our key variables. We used the Linguistic Inquiry Word Count (LIWC) software to obtain the frequency of words for each kind of authenticity – type and moral – as a percentage of total words in each review (the descriptive statistics are reported in Table IV). Following similar analyses by Genevsky and Knutson (2015), as well as Ludwig et al. (2013), we used ordinary least squares (OLS) regression to determine whether the interaction between field autonomy and reviewer’s field-specific cultural capital predicted the kind of authenticity discussed in the review.

To measure field autonomy, we counted the number of reviews published in Rolling Stone for a particular genre in a particular year (i.e. genre size). Given Rolling Stone’s importance as an arbiter of mainstream music (Frith, 1983), we can reasonably assume that the genres that reviewers analyze more frequently are more likely to reflect the logic of large-scale production, and the genres least frequently analyzed are more likely to reflect the logic of restricted production. We report the frequency of reviews per genre by decade in Table II.

We estimated the reviewers’ field-specific cultural capital by counting the number of reviews each reviewer wrote and published in Rolling Stone during a specific review’s calendar year (i.e. reviewer experience). As stated, the average number of reviews a reviewer wrote during the time period of our analysis was 10.61 (SD = 33.58), and the range of reviews a single reviewer wrote was 1-366. The average number of reviews in the top quartile was 35.85 (SD = 60.36). Both genre size and reviewer experience are two relative measures; that is, our investigation focuses on whether reviewers with more (versus less) experience treated bigger (versus smaller) music genres or more (versus less) autonomous fields similarly or differently. Because our measures of field-specific cultural capital (i.e. reviewer experience) and field autonomy (i.e. genre size) were not normally distributed, we applied a Box Cox transformation to each.

Because Rolling Stone is not the only outlet where music critics can publish their reviews, it is possible that reviewers who wrote few reviews in Rolling Stone could have been contemporaneously writing extensively elsewhere. To the extent this was the case, the number of reviews published in Rolling Stone might poorly reflect a reviewer’s experience and field-specific cultural capital (for example, if a reviewer wrote extensively in other prominent magazines but not much in Rolling Stone). Therefore, to increase confidence in this measure, we tested the extent to which the reviewers in the Rolling Stone database published reviews in other outlets. We extracted a random sample of 45 Rolling Stone reviewers from our data set, which represents about 10 per cent of the total, stratified according to the number of published reviews. Then, we searched for each reviewer in both the Music Magazine Archive (https://mma-napubcoonline-com) and the Entertainment Industry Magazine Archive (Proquest) databases to count how many reviews each one authored during the same period in other outlets. We then assessed the correlation between the number of reviews written by each reviewer in Rolling Stone and their number of reviews in each of the other two databases. If reviewers publishing infrequently in Rolling Stone were publishing frequently in other outlets we would expect a negative correlation, raising concerns about whether publication frequency in Rolling Stone is a good measure of experience. Our analysis of the entire sample did not show any significant correlations (r_RS-MMA = –0.08, p > 0.10; r_RS-EIA = 0.04, p > 0.10). We also separately estimated the correlations for those who were above and below the median publication frequency in Rolling Stone for this sample. This additional analysis did not show a significant correlation for reviewers above the median (r_RS-MMA = –0.01, p > 0.10; r_RS-EIA = –0.02, p > 0.10), but did show a marginally significant positive correlation for those below the median (r_RS-MMA = 0.38, p = 0.06; r_RS-EIA = 0.39, p = 0.06), suggesting that those writing infrequently in Rolling Stone were also writing infrequently for other outlets. Because this analysis did not produce any significant negative correlations, and did produce a marginally significant positive correlation for reviewers with a lower publication frequency, this increases confidence that number of reviews published in Rolling Stone is a good measure of experience and field-specific cultural capital, especially given the prominence and prestige of Rolling Stone as an outlet for cultural intermediaries.

The regression analyses that we implemented for this research also included several covariates to account for otherwise unexplained variance in our models. Type of artist distinguished among male single artists (566; 40.31 per cent of the total), female single artists (178; 12.68 per cent), male bands (537; 38.25 per cent), female bands (12; 0.85 per cent) and other artists (e.g. male-female bands, various artists’ albums) (111; 7.90 per cent). We measured tags as the number of tags Last.fm associates with a reviewed artist (M = 3.10; SD = 1.15). We then calculated genre diversity as the percentage of music albums associated with more than the average number of tags in a particular music genre per year (M = 1.12; SD = 0.16). We measured artist experience as the number of albums an artist had released at the time of a specific review (M = 2.13; SD = 0.64). Because artist experience and genre diversity were not normally distributed, we applied a Box Cox transformation to each. We measured these covariates, because we suspected that – unconnected to a subfield’s relative autonomy and a reviewer’s field-specific cultural capital – authenticity issues might depend on the type of artist, the artist’s relative experience or the extent to which a genre was clearly or loosely defined. We also controlled for date, measured as the day on which a music album review was released, to account for any variance that might occur in a particular issue; for example, a holiday issue will probably contain more album reviews about boxed sets. Finally, we controlled for word count, i.e. the total number of words of each music album review (M = 375.30; SD = 347.28), and words per sentence (M = 24.15; SD = 5.50), assuming that longer sentences and reviews may be more – or less – likely to include certain kinds of analyses. (We also modeled the effect of artist experience [moderated by reviewer’s field-specific cultural capital] and the effect of genre diversity [moderated by genre size]. We discuss these effects at the end of the findings section.)

3.5 Validation

As Humphreys and Wang (2017) suggest, testing the main constructs’ predictive validity helps to increase confidence that a study’s constructs are measured appropriately, and that the analytic methods are appropriate for capturing construct relationships. One approach that Humphreys and Wang (2017) recommend is a triangulation analysis, which tests the relationship between a study’s key constructs and other constructs that may not be central to the study, but which may be expected to be related to a study’s constructs in accordance with construct definitions or previous theory (Humphreys, 2010; Pennebaker and King, 1999). To the extent that these expected correlations emerge, confidence in the construct validity is enhanced.

Consequently, we ran an analysis to check whether our main dependent variables correlated with the other related constructs. Regarding type authenticity (i.e. genre reference and place reference), we correlated our two variables with the affiliation dictionary developed and validated by Pennebaker et al. (2007). This dictionary was created to capture a person’s fundamental need to feel part of a social group, as defined by McClelland (1987). Because judgments about type authenticity are also about deciding whether something is part of a group, we anticipated a degree of convergence between the two concepts. For example, a few of the keywords included in the affiliation dictionary are “associates,” “belong” and “tradition.” As expected, we found a positive and statistically significant correlation between affiliation and genre reference (r = 0.02; p < 0.05) and between affiliation and place reference (r = 0.10; p < 0.001).

Regarding moral authenticity (i.e. marketing language and personal authenticity), we correlated our two variables with the dictionary developed and validated by Opoku et al. (2006). This dictionary, which is based on Aaker’s (1997) five traits of brand personality, measures, among other things, a brand’s level of sincerity. Opoku et al. (2006, p. 31) refer to sincerity as reflecting an entity’s “true […] nature”, so we expected that this dictionary would correlate with our measures of moral authenticity. As expected, we found a positive and statistically significant correlation between sincerity and personal authenticity (r = 0.05; p < 0.001) and a negative and statistically significant correlation between sincerity and marketing language (r = –0.04; p < 0.01).

4. Results

Using the PROCESS macro for SPSS (Model 1, Hayes, 2013), we estimated four models to test whether the interaction effects between reviewer experience (i.e. field-specific cultural capital) and genre size (i.e. field autonomy), predicted genre reference, place reference, marketing language and personal authenticity (i.e. type authenticity and moral authenticity). The results of the four OLS regressions (see Table V) show that three of the four hypothesized interactions are statistically significant, whereas one is marginally significant. The interaction between reviewer experience and genre size as a predictor of genre reference is negative and marginally significant (ß = –0.005; t = –1.915, p < 0.10), whereas that of place reference is negative and statistically significant (ß = –0.022; t = –2.950, p < 0.01). In contrast, the interaction effect of reviewer experience and genre size on marketing language is significantly positive (ß = 0.009; t = 2.209, p < 0.05), as is that of personal authenticity (ß = 0.013; t = 2.787, p < 0.01).

Figure 1 graphically illustrates the interactions and plots the conditional effects of reviewer experience at various levels of genre size for each dependent variable. We also summarize the results in Figure 2. If reviewers follow the logic of the field that produced a cultural product, we would expect that reviewers of music produced by smaller (i.e. more autonomous) fields would focus less on language related to type authenticity (i.e. genre reference and place reference) and more on language related to moral authenticity (i.e. marketing language and personal authenticity). We would also expect that reviewers of music produced by larger (i.e. more heteronomous) fields would focus more on language related to type authenticity (i.e. genre reference and place reference) and less on language related to moral authenticity (i.e. marketing language and personal authenticity). Our hypotheses predict that reviewers with lower field-specific cultural capital are likely to follow this pattern but that reviewers with higher field-specific cultural capital are likely to do the opposite.

We plot our results in Figure 1. In each of the plots for this figure, the frequency of referencing the relevant language is indicated by the y-axis, and reviewer experience is indicated by the x-axis. Within each plot, lines of different styles (solid, dashed, etc.) represent different genre sizes (i.e. different levels of field autonomy). Because our hypotheses predict differences between reviewers with different experience (i.e. different levels of field-specific cultural capital), moderated by genre size (field autonomy/heteronomy), support for our hypotheses would be indicated by different slopes for the lines in the plots.

The general pattern of results shown in Figure 1 supports our hypotheses. First, consider the effects when genre size was large (high heteronomy). To do this, focus on the small-dashed lines in Figure 1, which represent the effects for the largest genres in our data set. Recall Bourdieu’s prediction that, for these heteronomous genres, type authenticity will be more important and moral authenticity will be less important. We hypothesized that less-experienced reviewers are more likely than more-experienced reviewers to reflect these standards. We would therefore expect a downward-sloping line for type authenticity language (because reviewers with more experience are less likely to use this language) and an upward-sloping line for moral authenticity language (because reviewers with more experience are more likely to use this language). The small-dashed lines are indeed downward-sloping in Figure 1(a) and 1(b) (type-authenticity language) and are upward-sloping in Figure 1(c) and 1(d) (moral-authenticity language).

Now consider the effects when genre size was small (high autonomy). To do this, focus on the solid lines in Figure 1, which represent the effects for the smallest genres in our data set. Recall Bourdieu’s prediction that type authenticity will be less important for autonomous genres and moral authenticity will be more important. Given our hypotheses, we would expect an upward-sloping line for type authenticity language (because reviewers with more experience are more likely to use this language) and a downward-sloping line for moral authenticity language (because reviewers with more experience are less likely to use this language). The solid lines are indeed upward-sloping in Figure 1(a) and 1(b) (type-authenticity language) and are downward-sloping in Figure 1(c) and 1(d) (moral-authenticity language).

While the visual pattern of data in Figure 1 suggests general support for our hypotheses, we performed statistical tests on each model by using the Johnson–Neyman technique (Hayes and Matthes, 2009; Johnson and Neyman, 1936), which calculates the range of the moderator in which the independent variable’s effect on the dependent variable is statistically different from zero. This technique is also called a floodlight analysis (Spiller et al., 2013). The points of transition identified through this method distinguish areas, if any, in which the independent variable’s effect on the dependent variable is statistically distinct (Hayes and Rockwood, 2017).

Our hypotheses were supported for all the associated tests for this analysis, except in the case of genre reference language when genres were smaller. For this dependent variable [Figure 3(a)], there is only one significant Johnson–Neyman region, which corresponds to a genre size > 8.308. This means that, for more heteronomous fields (genre size > 8.308), reviewer experience has a statistically significant and negative effect on genre reference. That is, as predicted by H1a, when the genre size was big, this type-authenticity language was more likely to be discussed by reviewers with lower (versus higher) field-specific cultural capital. However, H2a was not supported for this dependent variable. When genre size was smaller (genre size < 8.308), reviewer experience had no effect on the likelihood of mentioning this type-authenticity variable (See the gray regions in Figure 3). (See the Appendix for the complete table of conditional effects of reviewer experience on the dependent variables at different values of genre size.)

For place reference (a measure of type authenticity), PROCESS identified two Johnson-Neyman significance regions [Figure 3(b)]. For more heteronomous fields (genre size > 6.502), this type-authenticity language was more likely to be discussed by reviewers with lower (versus higher) field-specific cultural capital, supporting H1a. For more autonomous fields (genre size < 4.435), this type-authenticity language was less likely to be discussed by reviewers with lower (versus higher) field-specific cultural capital, supporting H2a.

PROCESS also identified two Johnson-Neyman significance regions for marketing language, which is a measure of moral authenticity [see Figure 3(c)]. For more heteronomous fields (genre size > 6.601), reviewer experience has a positive and statistically significant effect on marketing language. When discussing music produced by more heteronomous fields, this moral-authenticity language was less likely to be discussed by reviewers with lower (versus higher) field-specific cultural capital, supporting H1b. More autonomous fields (genre size < 1.372) exhibited the opposite tendency. In these fields, this moral-authenticity language was more likely to be discussed by reviewers with lower (versus higher) field-specific cultural capital, supporting H2b.

For personal authenticity (also a measure of moral authenticity), PROCESS again identified two Johnson-Neyman significance regions [Figure 3(d)]. More heteronomous fields exhibited a positive, significant effect of reviewer experience on personal authenticity (genre size > 5.241). That is, in these fields, this moral-authenticity language was less likely to be discussed by reviewers with lower (versus higher) field-specific cultural capital, supporting H1b. In contrast, in more autonomous fields (genre size < 1.386), this moral-authenticity language was more likely to be discussed by reviewers with lower (versus higher) field-specific cultural capital, supporting H2b.

While many of this study’s hypotheses were supported by the data, we tested two sets of hypotheses that were only partially supported, and where our interpretation of the results is therefore more speculative. The first set of hypotheses centered on the effect of genre diversity. We anticipated that the likelihood of discussing different types of authenticity would be affected by genre diversity, moderated by genre size. We believed that, when a genre has higher internal diversity, widely agreed-upon stereotypic elements would be less likely, making it more difficult for a reviewer to discuss type authenticity (Mattsson et al., 2010). We therefore expected that more-diverse genres would, relative to less-diverse genres, focus less on type authenticity and more on moral authenticity. We further expected that this tendency would be stronger in larger genres, where the logic of large-scale production and its emphasis on type authenticity would be more disrupted by genre diversity.

Our analysis showed that the genre diversity × genre size interaction was a significant predictor of only two of our dependent variables (genre reference and marketing language), and in ways that only partially supported our hypotheses. An investigation of this interaction showed that, as predicted, the authenticity language used in reviews of music from small (autonomous) genres was unaffected by genre diversity, but that the authenticity language used in reviews of music from large (heteronomous) genres was affected (Figure 4). However, while we had anticipated that diversity would decrease focus on type authenticity and increase focus on moral authenticity, our results indicate a decreased focus on both. Reviews of music from more-diverse genres were, relative to reviews of music from less-diverse genres, less likely to mention genre (type authenticity) and less likely to mention marketing language (moral authenticity).

Details of this analysis are depicted in Figure 4. For genre reference (a measure of type authenticity), PROCESS identified one Johnson–Neyman significance region (Figure 4[c]). This means that, for a genre size higher than 1.391 (see the gray regions in Figure 4), an increase in genre diversity led reviewers to mention significantly less genre references. Similarly, for marketing language (a measure of moral authenticity), one Johnson–Neyman region was identified, corresponding to genre size > 5.134 [Figure 4(d)]. Thus, for larger genres, genre diversity had a negative and statistically significant effect on marketing language. The result for genre reference supports our intuition that the co-existence of different types in more diverse genres makes it less productive or more challenging to discuss authenticity relative to a particular type. However, the result for marketing language does not support our expectation that, when type authenticity is less important, discussions of moral authenticity naturally rise. In fact, it supports the opposite. Although moral authenticity is already relatively unimportant in larger genres, it could be that – to the extent it is important – it is linked to type authenticity. For example, while success in a field of large-scale production such as country music hinges primarily on being authentic to type, it may be additionally helpful if the artist is also perceived to be personally authentic to the type. A musician who follows the standards of bluegrass music may be considered to be high in type authenticity, but his or her authenticity may be enhanced if the musician grew up in a part of the country known for bluegrass and is therefore also playing music that is morally authentic to him or her. However, our results suggest that greater diversity in a field of large-scale production (for example, greater diversity in what counts as traditional bluegrass music) may make it more difficult for reviewers to discuss this version of moral authenticity.

A second set of hypotheses centered on artist experience. Given Marshall’s (2006) observation about the importance of artists establishing a personal connection with their fans (even in fields of large-scale production), we anticipated that, as an artist’s career develops, he or she would increasingly focus on moral authenticity, regardless of whether he or she works in a field of restricted or large-scale production. We also reasoned that it is harder for an artist to build moral authenticity after only one music album and easier to build it over time. Because the central hypothesis for our work is that less experienced reviewers are more likely to assess artists on the artists’ terms (and that more experienced reviewers are less likely to do this), we hypothesized that less experienced reviewers would focus more on moral authenticity when reviewing experienced artists and more on type authenticity when reviewing less experienced artists.

Our analysis showed that the artist experience × reviewer experience interaction was a significant predictor of only two of our dependent variables (place reference and personal authenticity) and in ways that only partially supported our hypotheses. Looking at a plot of the results, the general pattern supports our hypotheses for these two dependent variables. Consider the solid line in Figure 5(a) and 5(b), which represent the least experienced artists in our data set. This line slopes downward with reviewer experience in the plot for mentions of place reference (a measure of type authenticity) and slopes upward with reviewer experience for mentions of personal authenticity (a measure of moral authenticity). The short-dashed lines in Figure 5(a) and 5(b) represent the most experienced artists in our data set. As expected, this line slopes upward with reviewer experience in the plot for mentions of place reference and downward with reviewer experience for mentions of personal authenticity. So, as expected, less experienced reviewers are more likely than more experienced reviewers to mention type authenticity for less experienced artists and moral authenticity for more experienced artists.

Figure 5 also shows that, for place reference, PROCESS identified two Johnson-Neyman significance regions [Figure 5(c)]. The first region falls below an artist experience equal to 1.632, whereas the second region corresponds to an artist experience higher than 2.549 (See the gray regions in Figure 5). This result shows that, for less experienced artists (artist experience < 1.632), reviewers with less experience discussed place references significantly more than reviewers with more experience (i.e. reviewer experience negatively affects place reference). In contrast, for more experienced artists (artist experience > 2.549), reviewers with more experience paid more attention to place reference than less experienced reviewers (i.e. reviewer experience positively affects place reference). However, for personal authenticity, only one Johnson-Neyman region was identified, corresponding to artist experience < 2.173 [Figure 5(d)]. For less expert artists, reviewer experience had a statistically significant and positive effect on personal authenticity. These tests provide partial support for the idea that, as an artist’s career progresses, a reviewer with less experience is more likely to focus on moral authenticity whereas a reviewer with more experience is more likely to focus on type authenticity. And, at a more general level, the patterns reinforce the idea that reviewers with more experience are systematically different from reviewers with less experience.

5. Discussion and implications

Cultural intermediaries explain to target audiences how a cultural product should be understood and evaluated, and therefore play an important role in shaping the reception of these products. However, the nature of an intermediary’s influence depends on the standards that he or she uses when interpreting a cultural product. It is therefore important to understand the factors that influence the selection of these standards. This research tests whether Rolling Stone music critics mirror the authenticity standards that are important to musical artists, or whether they apply different standards. This question is theoretically important because Bourdieu was ambivalent about the answer. In certain writings, Bourdieu suggests that cultural intermediaries try to adhere to the standards that are important to the creator, whereas in others he observes that cultural intermediaries operate according to their personal logic, even if that logic is separate and distinct from the artist’s logic. Our results suggest that both of Bourdieu’s observations are true, depending on the cultural intermediary’s field-specific cultural capital. Figure 2 summarizes our results.

More specifically, the pattern of effects indicates that reviewers with less experience (and thus lower field-specific cultural capital) are more likely to apply the authenticity standards that Bourdieu predicts are important to artists in a particular subfield. For instance, when reviewing music produced by smaller, more autonomous subfields, reviewers with less experience are more likely to focus on moral authenticity and less likely to focus on type authenticity. Furthermore, when reviewing music produced by larger, more heteronomous subfields, these reviewers are more likely to focus on type authenticity and less likely to focus on moral authenticity. However, the pattern is opposite for reviewers with more experience. These reviewers show a tendency to apply authenticity standards that differ from the standards that Bourdieu predicted are important to artists in the subfield. When reviewing music produced by smaller, more autonomous subfields, reviewers with more experience are more likely to focus on type authenticity and less likely to focus on moral authenticity. Moreover, when reviewing music produced by larger (more heteronomous) subfields, these reviewers are more likely to focus on moral authenticity and less likely to focus on type authenticity.

While Bourdieu did not specifically propose that field-specific cultural capital might be the key to unlock his ambivalence about cultural intermediaries, past research on field-specific cultural capital helps explain this result. Fields are social spaces where members compete for status and, as we explained earlier, research has suggested that those with less field-specific cultural capital are more likely to follow expectations, whereas those with more field-specific cultural capital are more likely to distinguish themselves by deviating from expectations. Against this backdrop, it is not surprising that reviewers with less experience would mirror the standards that are important to the artists, whereas those with more experience might focus on the opposite. Although some work has recognized that cultural intermediaries may differ from each other with regard to their stance toward the artist and the consumer (Kuipers, 2012) and that they may adjust their standards when the logic of a field changes (Glynn and Lounsbury, 2005), this research is the first to identify a cultural intermediary’s field-specific cultural capital as a factor that systematically produces differences among cultural intermediaries.

This paper focuses on comparing the standards reflected in the appraisals of cultural intermediaries with the standards of the artists being reviewed. However, an important area for future research, which this paper does not address, is the potential influence of the standards that are important to consumers. Figure 6 illustrates a theoretical space that simultaneously considers the standards held by artists and by consumers, and thus indicates four general positions that a cultural intermediary can take. As the table indicates, a cultural intermediary can take a stance that is – or is not – in line with the artist’s standards and which is – or is not – in line with consumer standards. Given the cultural intermediary’s role as a translator, one might expect them to always adopt standards similar to those of consumers, but research has identified circumstances in which a cultural intermediary’s standards differ from those of consumers (Chávez, 2012; Coulter et al., 2003; Gurrieri et al., 2016; Wright, 2005).

Our research compared the rows in Figure 6, namely, reviewers who apply standards that are either the same as, or different from, the artist. We found that reviewers with lower field-specific cultural capital are more likely to be situated in the first row of the table (applying the same standards as the artist), whereas those with higher field-specific cultural capital are more likely to be situated in the second row of the table (applying different standards). Based on previous research, we have proposed that reviewers with more field-specific cultural capital situate themselves in the second row because they are motivated primarily by a desire to differentiate themselves from reviewers with lower field-specific cultural capital, who are situated in the first row. It could be that – like the Flaubert reviewers characterized by Bourdieu – these reviewers are applying their own internal standards, without regard for the standards held by typical consumers of the product or by typical readers of the review. Such reviewers would be placed in Cell 4 of Figure 6. To the extent that more experienced reviewers exhibit these tendencies, this might suggest that a cultural producer should pay less attention to (and, if possible, try to navigate away from) cultural intermediaries with significant experience as intermediaries. This is because such intermediaries are speaking primarily to an internal audience of other intermediaries, and not to stakeholder groups that are important to the cultural producer.

However, it could be that reviewers with higher field-specific cultural capital are more attuned to consumer standards than those with lower field-specific cultural capital, and that their differentiation from these less-experienced reviewers is because of a desire and ability to better connect with certain target consumers. Such reviewers would be placed in Cell 3 of Table VII. While many cultural producers would bristle in response to reviewers who apply standards that differ from the standards they were considering when they produced their works, some may nonetheless appreciate the benefit of being effectively interpreted for a particular audience. Consider a musician from Cuba who seeks to succeed in the field of restricted production by combining the genre of her home country (salsa) with the genre of her youth (rap). As an artist focusing on the avant-garde, she views her music as being a unique and morally authentic expression of her true self, and is more concerned about what other musicians think of her work than what the general public thinks. This musician may not mind (and may even appreciate) a review that evaluates her music negatively with regard to type authenticity because she knows that her audience eschews type authenticity. This musician may alternatively appreciate a review that evaluates her music positively with regard to type authenticity, especially if the positive review is targeted to customers who care about that standard and therefore results in popularity and sales. Niche artists who become more popular often struggle with concerns about selling out (Corciolani, 2014), but the benefit of greater financial reward is sometimes worth the cost.

A further important step for future research is to examine whether a similar pattern of results can be found among other cultural intermediaries, particularly those in domains other than music. The results of our research support the premise that those who are new to a field tend to follow stereotypic or expected patterns, whereas those who are more expert to a field tend to deviate. A similar dynamic might be expected in fields such as the restaurant market, the cinema market, and the tourism market. However, some research (Beunza and Garud, 2007) has suggested that those who are new to a field may tend to be more iconoclastic instead of more stereotypic. This is not what we found in our research context, but it seems likely that fields differ on this dimension. Therefore, it may be informative to identify fields in which there is a higher expectation that newcomers will exhibit either the one or the other behavior, and to assess the potential impact on the content of cultural intermediaries’ reviews. Another area for future research is to explore how other differences between cultural intermediaries might predict the standards they apply. For example, remembering that Bourdieu’s framework was specifically designed to accommodate how a player’s personal idiosyncrasies might influence how he or she plays the game, future research might examine how personality differences between cultural intermediaries might encourage those with the same field-specific cultural capital to use different strategies and, thus, apply different standards.

Finally, reviews now exist ubiquitously on online platforms such as Amazon, Yelp!, and Hotels.com, in addition to a number of social shopping and technology sites and blogs. Some reviewers even gain status to become known as experts or thought leaders (McQuarrie et al., 2012). Do the dynamics of field-specific cultural capital operate similarly in these very different contexts, where reviewers exist outside of the sanctioned mass media field of magazines? In a world where both writing and reading reviews is common and influential, understanding the ways in which reviewer standards change the nature of the field promises to shed light on the ways in which consumers and producers interpret and enjoy not just artistic products, but perhaps more mundane products such as consumer electronics, cleaning products and building supplies.

Figures

Figure 1.

Interaction effects of reviewer experience and genre size on the four dependent variables

Figure 2.

Summary of findings for reviewers with more (and less) fieldspecific cultural capital

Figure 3.

Conditional effect of reviewer experience on the four dependent variables at different values of genre size (with identification of Johnson–Neyman points)

Figure 4.

Interaction effects of genre diversity and genre size on genre reference and marketing language (with identification of Johnson–Neyman points)

Figure 5.

Interaction effects of reviewer experience and artist experience on place reference and personal authenticity (with identification of Johnson–Neyman points)

Figure 6.

Potential stances for cultural intermediaries

Table I.

Summary of the contrasting logics of two subfields of production

Field attribute	Subfield of restricted production	Subfield of large-scale production
Most important stake for artists	Symbolic capital (status)	Economic capital (popularity)
Most important target audience	Other producers	The public at large (non-producers)
Relative size of target audience	Relatively small	Relatively large
Control granted to artists	Artists are relatively autonomous	Artists are relatively heteronomous
Most important evaluation criterion	Moral authenticity: Is the artist true to him/herself?	Type authenticity: Does the cultural product meet established expectations?

Table II.

Number of reviewed music albums by decade and music genre

Music genre	1960s	1970s	1980s	1990s	2000s	2010s	Total
Avant-garde	2	10	5	12	14	49	92
Blues	2	29	9	15	19	42	116
Caribbean and Latin	0	3	7	6	11	15	42
Comedy	0	0	0	0	0	5	5
Country	5	30	13	16	60	129	253
Electronic	2	13	5	22	41	177	260
Folk	8	78	38	39	59	143	365
Hip-hop	0	0	8	38	79	207	332
Jazz	2	15	4	12	9	33	75
Pop	0	27	27	35	107	157	353
R&B and Soul	8	72	34	39	65	110	328
Rock	47	356	292	283	553	700	2,231
Total	76	633	442	517	1,017	1,767	4,452

Table III.

Dictionaries

Construct	Operation-alization	Definition	Sample words	No. of words	% agreement
Type authenticity	Genre reference	Discussion of how well an artist or his/her music reflects a particular genre (or genres)	Hip hop, metal, rocker	47	95.65
	Place reference	Discussion of how well an artist or his/her music reflects a particular place	California, Caribbean, scene	44	81.82
Moral authenticity	Marketing language	Discussion of marketing, business, or sales	Advertising, muzak, sponsor	31	95.70
	Personal authenticity	Discussion of real emotion, true personal facts, or staying independent, true, and fair	Consistent, honest, sincere	56	83.63
Total				273	86.13

Table IV.

Dependent variables’ descriptive statistics

Variable	Mean	SD	Minimum	Maximum
Genre reference	0.12	0.28	0.00	2.60
Place reference	1.06	0.85	0.00	8.00
Marketing language	0.35	0.47	0.00	5.62
Personal authenticity	0.39	0.51	0.00	6.36

Note:

Mean = (Number of dictionary-related words/Total words of the review) × 100

Table V.

Ordinary least squares regression model coefficients (standard errors in parentheses)

Predictor variables	Type authenticity				Moral authenticity
	Genre reference		Place reference		Marketing language		Personal authenticity
	Coefficient	(SE)	Coefficient	(SE)	Coefficient	(SE)	Coefficient	(SE)
Constant	0.146*	(0.067)	0.846***	(0.203)	0.372**	(0.115)	1.108***	(0.125)
Reviewer experience	0.025****	(0.014)	0.122**	(0.043)	–0.049*	(0.024)	–0.057*	(0.026)
Genre size	0.032***	(0.009)	0.102***	(0.028)	–0.049**	(0.016)	–0.034****	(0.018)
Reviewer experience × Genre size	–0.005****	(0.002)	–0.022**	(0.008)	0.009*	(0.004)	0.013**	(0.005)
Male single artists	0.008	(0.013)	0.025	(0.040)	0.004	(0.022)	0.014	(0.024)
Male bands	–0.026****	(0.014)	0.087*	(0.042)	–0.011	(0.024)	–0.035	(0.026)
Female bands	0.066	(0.048)	0.360*	(0.145)	0.222**	(0.083)	0.107	(0.090)
Other artists	0.004	(0.020)	0.171**	(0.060)	0.026	(0.034)	–0.067****	(0.037)
Date	0.000***	(0.000)	0.000**	(0.000)	0.000***	(0.000)	0.00**	(0.000)
Number of tags	–0.073****	(0.004)	0.014	(0.012)	–0.014*	(0.007)	–0.002	(0.007)
Genre diversity	–0.290***	(0.036)	–0.296**	(0.109)	–0.139*	(0.062)	–0.058	(0.067)
Genre size × Genre diversity	–0.052***	(0.009)	–0.006	(0.029)	–0.048**	(0.016)	–0.004	(0.018)
Artist experience	–0.045*	(0.020)	–0.071	(0.062)	0.048	(0.035)	–0.076*	(0.038)
Artist experience × Reviewer experience	0.012	(0.011)	0.075*	(0.033)	–0.022	(0.019)	–0.059**	(0.021)
Word count	–0.000***	(0.000)	–0.000***	(0.000)	–0.000***	(0.000)	–0.000*	(0.000)
Words per sentence	–0.004***	(0.007)	–0.009***	(0.002)	0.002	(0.001)	–0.002	(0.001)

Notes:

Estimation results are based on the linearized model; significance levels: ****p < 0.10; *p < 0.05; **p < 0.01 and ***p < 0.001. All artist type variables (e.g., male single artists, male bands, etc.) are modeled as dummy variables, except “female single artists”, whose effect is modeled when all of the other artist type dummy variables are zero

Table AI.

Conditional effect of reviewer experience on genre reference at values of genre size

Genre size	Effect	SE	t	p	LLCI	ULCI
0.000	0.025	0.014	1.771	0.077	–0.003	0.052
0.544	0.022	0.013	1.751	0.080	–0.003	0.047
1.087	0.020	0.011	1.725	0.085	–0.003	0.042
1.631	0.017	0.010	1.691	0.091	–0.003	0.037
2.174	0.014	0.009	1.644	0.100	–0.003	0.032
2.718	0.012	0.008	1.576	0.115	–0.003	0.027
3.261	0.009	0.006	1.472	0.141	–0.003	0.022
3.805	0.007	0.005	1.301	0.193	–0.003	0.017
4.348	0.004	0.004	1.001	0.317	–0.004	0.012
4.892	0.002	0.003	0.459	0.646	–0.005	0.008
5.435	–0.001	0.003	–0.374	0.708	–0.007	0.005
5.979	–0.004	0.003	–1.179	0.238	–0.010	0.002
6.522	–0.006	0.004	–1.635	0.102	–0.014	0.001
7.066	–0.009	0.005	–1.834	0.067	–0.018	0.001
7.609	–0.011	0.006	–1.918	0.055	–0.023	0.000
8.153	–0.014	0.007	–1.954	0.051	–0.028	0.000
8.308	–0.015	0.008	–1.960	0.050	–0.030	0.000
8.696	–0.017	0.008	–1.971	0.049	–0.033	0.000
9.240	–0.019	0.010	–1.977	0.048	–0.038	0.000
9.783	–0.022	0.011	–1.979	0.048	–0.044	0.000
10.327	–0.024	0.012	–1.979	0.048	–0.049	0.000
10.870	–0.027	0.014	–1.978	0.048	–0.054	0.000

Table AII.

Conditional effect of reviewer experience on place reference at values of genre size

Genre size	Effect	SE	t	p	LLCI	ULCI
0.000	0.122	0.043	2.874	0.004	0.039	0.206
0.544	0.110	0.039	2.859	0.004	0.035	0.186
1.087	0.098	0.035	2.838	0.005	0.030	0.166
1.631	0.086	0.031	2.809	0.005	0.026	0.146
2.174	0.074	0.027	2.766	0.006	0.022	0.126
2.718	0.062	0.023	2.701	0.007	0.017	0.107
3.261	0.050	0.019	2.593	0.010	0.012	0.087
3.805	0.038	0.016	2.404	0.016	0.007	0.068
4.348	0.025	0.012	2.044	0.041	0.001	0.050
4.435	0.023	0.012	1.960	0.050	0.000	0.047
4.892	0.013	0.010	1.334	0.182	–0.006	0.033
5.435	0.001	0.009	0.132	0.895	–0.016	0.018
5.979	–0.011	0.009	–1.158	0.247	–0.030	0.008
6.502	–0.023	0.012	–1.960	0.050	–0.045	0.000
6.522	–0.023	0.012	–1.983	0.047	–0.046	0.000
7.066	–0.035	0.015	–2.400	0.016	–0.064	–0.006
7.609	–0.047	0.018	–2.610	0.009	–0.083	–0.012
8.153	–0.059	0.022	–2.725	0.006	–0.102	–0.017
8.696	–0.072	0.026	–2.792	0.005	–0.122	–0.021
9.240	–0.084	0.030	–2.835	0.005	–0.142	–0.026
9.783	–0.096	0.033	–2.863	0.004	–0.162	–0.030
10.327	–0.108	0.037	–2.882	0.004	–0.181	–0.035
10.870	–0.120	0.041	–2.896	0.004	–0.201	–0.039

Table AIII.

Conditional effect of reviewer experience on marketing language at values of genre size

Genre size	Effect	SE	t	p	LLCI	ULCI
0.000	–0.049	0.024	–2.035	0.042	–0.097	–0.002
0.544	–0.044	0.022	–2.011	0.044	–0.087	–0.001
1.087	–0.039	0.020	–1.980	0.048	–0.077	0.000
1.372	–0.036	0.018	–1.960	0.050	–0.072	0.000
1.631	–0.034	0.017	–1.940	0.052	–0.068	0.000
2.174	–0.029	0.015	–1.884	0.060	–0.058	0.001
2.718	–0.023	0.013	–1.803	0.071	–0.049	0.002
3.261	–0.018	0.011	–1.680	0.093	–0.040	0.003
3.805	–0.013	0.009	–1.480	0.139	–0.031	0.004
4.348	–0.008	0.007	–1.128	0.259	–0.022	0.006
4.892	–0.003	0.006	–0.496	0.620	–0.014	0.008
5.435	0.002	0.005	0.470	0.638	–0.007	0.012
5.979	0.008	0.005	1.396	0.163	–0.003	0.018
6.522	0.013	0.007	1.915	0.056	0.000	0.026
6.601	0.013	0.007	1.960	0.050	0.000	0.027
7.066	0.018	0.008	2.138	0.033	0.001	0.034
7.609	0.023	0.010	2.231	0.026	0.003	0.043
8.153	0.028	0.012	2.270	0.023	0.004	0.052
8.696	0.033	0.015	2.286	0.022	0.005	0.062
9.240	0.038	0.017	2.292	0.022	0.006	0.071
9.783	0.044	0.019	2.293	0.022	0.006	0.081
10.327	0.049	0.021	2.292	0.022	0.007	0.090
10.870	0.054	0.024	2.289	0.022	0.008	0.100

Table AIV.

Conditional effect of reviewer experience on personal authenticity at values of genre size

Genre size	Effect	SE	t	p	LLCI	ULCI
0.000	–0.057	0.026	–2.178	0.029	–0.109	–0.006
0.544	–0.050	0.024	–2.107	0.035	–0.097	–0.003
1.087	–0.043	0.021	–2.019	0.044	–0.085	–0.001
1.386	–0.039	0.020	–1.960	0.050	–0.078	0.000
1.631	–0.036	0.019	–1.906	0.057	–0.073	0.001
2.174	–0.029	0.016	–1.756	0.079	–0.061	0.003
2.718	–0.022	0.014	–1.551	0.121	–0.050	0.006
3.261	–0.015	0.012	–1.255	0.210	–0.038	0.008
3.805	–0.008	0.010	–0.805	0.421	–0.027	0.011
4.348	–0.001	0.008	–0.090	0.928	–0.016	0.014
4.892	0.006	0.006	1.038	0.300	–0.006	0.018
5.241	0.011	0.006	1.960	0.050	0.000	0.022
5.435	0.013	0.005	2.474	0.013	0.003	0.024
5.979	0.020	0.006	3.510	0.000	0.009	0.032
6.522	0.028	0.007	3.837	0.000	0.013	0.042
7.066	0.035	0.009	3.826	0.000	0.017	0.052
7.609	0.042	0.011	3.727	0.000	0.020	0.064
8.153	0.049	0.013	3.622	0.000	0.022	0.075
8.696	0.056	0.016	3.530	0.000	0.025	0.087
9.240	0.063	0.018	3.452	0.001	0.027	0.099
9.783	0.070	0.021	3.387	0.001	0.029	0.110
10.327	0.077	0.023	3.333	0.001	0.032	0.122
10.870	0.084	0.026	3.287	0.001	0.034	0.134