Discriminating an Accent, Enacting a Race (and Vice Versa): Perception and Representation of Phonic Variability on the Caribbean Coast of Colombia

Abstract

We examine the metalinguistic, psychosocial, and identity-related indexical indications of (speaking about) the other, as provided by 36 Santa Marta inhabitants (Blestel 2022). Santa Marta is a city on Colombia’s septentrional coast, populated primarily by individuals of multiracial heritage, descended from Indigenous peoples, African slaves, and Spaniards. We show that the established main axis of differentiation (Gal and Irvine 2019) and associated judgments are adjacent, sometimes very explicitly, to racializing ideologies that more broadly traverse Colombian society as a whole. We maintain that these metapragmatic discourses, far from being anecdotal, indicate semiotic discrimination processes at work in sustaining or even renewing a system of values passed down from the colonial era.

[Accent discrimination] is so commonly accepted, so widely perceived as appropriate, that it must be seen as the last back door to discrimination. And the door stands wide open.

—Rosina Lippi-Green (2012, 74)

Conducting an investigation on the subject of such a “loose” category of people as that defined by an accent would be surprising,¹ given the degree to which, for lack of “ontological autonomy” (Candea 2021, 19), the definition of this study subject relies primarily on how others perceive specific distinguishing pronunciation traits.² Yet we believe the importance given to audience perceptions in defining the accent deserves focused attention for at least two reasons. The first is that what’s true for the “accent” has to do more broadly with a generalized mechanism in speech, considering that all acts of speaking are none other than, first and foremost, and from a phenomenological point of view, the “controlled production of an acoustic disturbance” that has meaning only if it is “recognizable and interpretable … by the subjects involved” (Bottineau 2012, 74). To use the words of C. S. Peirce, “A sign does not function as a sign unless it be understood as a sign” (Peirce MS 599, 32, quoted in Parmentier 1994, 4).³ This lability in interpretation then encourages us to focus the analysis on the allocutor, who is the subject both perceiving and therefore also interpreting the locution. The second reason we believe the “accent” category is relevant comes from the “privileged denomination of linguistic otherness” (Larrivée 2009, 84). This “otherness” isn’t granted: the differential perceptions are a result of discrimination processes in a given field (phonic and/or social) that must be acknowledged so long as they are informed socially and ideologically. By the expression discrimination processes, we are referring both to the act of separating or distinguishing two or more beings or things based on certain distinctive criteria or characteristics as well as to distinguishing a person or a group based on this (or these) perceived difference(s) and treating them differently (sometimes poorly). In this sense, we assimilate this cognitive operation with what Maturana means by the concept of “distinction” from an enactive perspective, that is, a building-block cognitive operation that makes up one’s “reality”:

The fundamental operation that an observer can perform is an operation of distinction, the specification of an entity by operationally cleaving it from a background. Furthermore, that which results from an operation of distinction and can thus be distinguished, is a thing with the properties that the operation of distinction specifies, and which exists in the space that these properties establish. Reality, therefore, is the domain of things, and, in this sense, that which can be distinguished is real. Thus stated, there is no question about what reality is: it is a domain specified by the operations of the observer.

(Maturana 1978, 55)

It seems to us that this fundamental operation of oriented distinction is what contributes to constructing the contrastive schemas—or axes of differentiation (Gal and Irvine 2019; Gal 2021)—which consist in drawing oppositions between two or more indexicalities. Thus, for these authors, any sign conceived as first maintaining a relationship of contiguity, copresence, or causality with its referent can end up being considered as iconic, that is, a sign maintaining with its referent a relation of resemblance. This is what the authors call the process of rhematization: “By rhematization, contrasting qualities perceived in the signs are taken to be like, to resemble qualitative contrasts in what the signs are taken to index—person-types, actions, objects” (Gal 2021, 137).

As we will see, Santa Marta is a city on the northern coast of Colombia populated mainly by individuals of multiracial ancestry, descendants of Indigenous peoples, African slaves, and Spaniards, and is representative of the Colombian coastal settlement type (as opposed to the Andean region, which is perceived as richer and more ethnically White). The (supra)segmental, psychosocial, and identity-related indexical indications about (speaking about) the other, as provided by the Samarios (i.e., Santa Marta inhabitants) questioned, were crossed against the postures they adopted with regards to the three preceding types of indications. On the basis of this examination, we show that this established axis of differentiation (coast versus Andes) and associated judgments are adjacent, sometimes very explicitly, to racializing ideologies that more broadly traverse Colombian society as a whole. Indeed, the same axis distinguishes educated people from those who are not, people from cold climates from those from hot climates, people who talk softly from those who talk loudly, people who are gay, partygoers, and outspoken from those who are serious, calm, but also a little sneaky, and so on. This same pattern operates at the national level (Andean region versus coastal region) and also at the regional level, with the inhabitants of Santa Marta generally perceiving themselves as more calm, serious, and measured than the inhabitants of the southwestern Caribbean coast (the city of Cartagena, for instance). Thus, by systematically recording these indications among the people questioned, we show that they are part of the same coherent and recursive pattern: certain linguistic forms are systematically associated with certain geographical areas, identities, skin colors, climates, characters, and so on. In this sense, we understand the term psychosocial as covering what Agha calls “characterological labels,” that is, “characterological labels and discourses that identify speakers in terms of the mental, aesthetic and class attributes” (2003, 233).

By examining the metapragmatic discourses used by the perceiving and interpreting inhabitants of Santa Marta to explain what they consider to be the accent (acento or cantadito) of the other, we therefore intend to bring to light those processes of essentializing or naturalizing the association between certain linguistic forms and certain identity or “psychosocial” attributes that replicates at different scales via the same oppositional scheme under the fractal recursivity principle (Gal and Irvine 2019; Gal 2021). Insofar as these metapragmatic discourses are symptomatic of the perceptive and social discrimination of some semiotic practices—those relating to the “accent” in this case, but also to other practices—we maintain that this is an instance of a social process that participates in sustaining or even renewing a system of values passed down from the colonial era, so that this study could contribute to the “raciolinguistic” perspective that Rosa and Flores (2017) willingly named, and that “analyzes the ongoing rearticulation of colonial distinctions between populations and modes of communication that come to be positioned as more or less normatively European” (Rosa 2019, 5).⁴

Santa Marta, between Contrasts and Multiracial Caribbean Heritage

In order to understand properly the schema of contrast that, in our view, structures the metapragmatic discourses of the Samarios whom we interviewed, it is necessary to situate the settlement of Santa Marta in its sociohistorical and economic context. Indeed, the social construct that is race plays a major role in the socioeconomic and political organization of the colonial period. This organization still has consequences today, and also persists—this is what we want to show—in the metapragmatic discourses that contribute to nourishing those ongoing disparities.

The mixed-race capital of the department of Magdalena, Santa Marta (population 499,192, according to the 2018 census) is located between the Sierra Nevada mountain range and the Caribbean Sea and is the third-largest urban city of Colombia’s Caribbean region, after Barranquilla and Cartagena (see fig. 1).

Figure 1. Political map of Colombia (adapted from IGAC 1999)

The city is representative of Colombia’s Caribbean coast. First inhabited by Native populations—the city was founded on land inhabited by native Taironas, in the Chibchan family,⁵ even though Wayuu and Chimila incursions also occurred frequently in the city (Viloria de la Hoz 2015)—labor imported from Africa was gradually and continuously introduced on such a large scale that slavery in Santa Marta came to surpass that of the slave port of Cartagena on several occasions throughout the eighteenth and nineteenth centuries (Bénéï 2011, 116). The migrant waves of the nineteenth and twentieth centuries continued to build this multifaceted landscape: the rise of banana farming in the late nineteenth century attracted newcomers from Bogotá, Valledupar, and the departments of Bolívar, Antioquia, and La Guajira (Viloria de la Hoz 2008). The same era saw the arrival of Jews, Syro-Lebanese, and Arabs (Igirio Gamero 2008, 303), on the Caribbean coast—mainly in Barranquilla but also, to a lesser extent, in Cartagena and Santa Marta (see also Viloria de la Hoz 2003). In modern times, the most recent national census, in 2018, indicates that the majority of immigrants come from Venezuela,⁶ that 8.42 percent of the Magdalena population describes itself as “Raizal, Palenquero, black, mulatto, Afro-Colombian, or of African heritage,” and 1.66 percent as “indigenous” (DANE 2018).⁷

It is important to note here that this characteristic settlement of the Caribbean coast contrasts profoundly with what happened in the inland part of the country, especially in what is commonly called the “Andean region” (see fig. 2).

Figure 2. Natural regions of Colombia (adapted from IGAC 2012)

Indeed, Colombia presents deep regional sociopolitical and economic contrasts that historiography partially attributes to the long-term consequences of the way society was organized at the end of the colonial era. The thinking is that the regions, like the Caribbean coast, that were home to the most exclusive historical institutions (slavery, encomienda, mita, limited access to land and education) still have the country’s highest rates of poverty even today.⁸ As a result, even though the Spanish conquistadors settled first on the Caribbean coast, namely, in the port cities of Santa Marta and Cartagena de Indias, they were quickly driven inland by English pirate attacks; hostile natives in the backcountry; and farming, climate, and health conditions. They moved into Andean territory, where the climate was more favorable and the land more fertile (Wade 1993, 55). The Andean highlands were densely populated by numerous groups, with a political organization that facilitated the exploitation of Indigenous labor. On the other hand, the coastal lowlands were less heavily populated and were not as politically organized, which resulted in higher resistance to domination (Harris 1974). This is why, in the lowlands, as well as in regions where gold mines created a demand for workers,⁹ African slaves were forced to do this hard work (Wade 1993, 54–55).¹⁰ However, it must be pointed out that these initial regional contrasts, which we can see were already structured relative to race, must be considered not alone but in addition to the subsequent racial mixing, which did not play out in the same manner in the various regions (Wade 1993, 2020). For example, while the presence of gold mines in the Andean Antioquia region had resulted in the recruitment of slaves in comparable or even higher numbers than those of the Caribbean coast, racial mixing in this region took on such a dimension that the descendants of these slaves are no longer as “visible” in today’s racial landscape: “Thus, for example, the large numbers of black slaves in Antioquia mixed to such an extent that the region’s black heritage is today obvious only in certain lowland mining districts. The slave component of the Caribbean coastal region, proportionally less of the total than in Antioquia, has engendered a much more apparent contemporary black presence, although it is most obvious in a quite narrow belt along the littoral itself and along the Magdalena River” (Wade 1993, 56).

These are the historic and structural differences that explain how a certain “regionalization of race” arose (Wade 2020) in the country, insofar as Colombians now identify the Andean inlands with White mestizos (often called Cachacos¹¹), or even with unmixed Whites, in the case of the Paisa region,¹² where ethnic identity has even become a “myth of racial purity and lack of black and Indian heritage” (Wade 1993, 66), despite the aforementioned historical data. The inland part of the country is often opposed to the coastal areas.¹³ Located on the Pacific coast, the department of Chocó is clearly identified as a “black” region (Friedemann 1974; Restrepo 2011, 2013). The Caribbean coast also has pockets of “black” populations (Palenque de San Basilio is undoubtedly the most well-known), giving its cities and their social structures more contrast, with the (darkest) Blacks most often living at the bottom of the social ladder (Cunin 2000, 2004; Wade 2020). Finally, the least densely populated departments in the Amazon region (in light and dark green on fig. 2) must also be considered. Today they are perceived as being primarily Indigenous, as confirmed by the number of inhabitants who identified as such in the most recent census.¹⁴

Prompting the Metapragmatic Discourses of the Samarios

In this national landscape marked by regional disparities and a certain “Andean-centrism” (Soler Castillo and Pardo Abril 2009, 132), linguistic studies are no exception in that, to date, they have focused more on the Cachaca variant of Spanish, to the detriment of the Costeña variant (Orozco 2009, 96–97).¹⁵ The dialectal areas of Colombia are customarily distinguished by essentially phonic traits (Flórez 1961; Montes Giraldo 1982; Ruiz Vásquez 2020) and dialectologists as a whole agree on the existence of a Costeña variant (also referred to as Caribbean Colombian), which includes the dialectal variants of the Atlantic coast, itself including the subregion of Santa Marta (Montes Giraldo 1982, 46; and see Blestel [2022] for a history of the various dialectal classifications of Costeño Spanish). Nevertheless, few studies have examined how this variability is perceived, or even discussed, by Caribbean Colombians themselves, with the exception of studies by Salazar Caro (2019) on the beliefs of the inhabitants of Montería about the Spanish spoken in their city and Quinn (2019, 2021) on the linguistic representation of Caribbean people in popular Colombian soap operas. This article therefore assumes a position as part of a larger study of the sociophonetic variability of Spanish in the city of Santa Marta, and its reception by various members of society. With this in mind, we sought to produce a corpus of data that brings together a sample of speakers’ phonic productions from men and women, of varying ages and levels of education. The 36 interviews used in this article and published in Blestel (2022) amount to just a fragment—and the starting point—of the body of data that were gathered for this research on production, variation, and sociophonetic reception in Santa Marta, which also requires complementary studies on production and perception, as well as ethnographic observations that are still ongoing at the time of this writing. The 36 people questioned here are distinguished by the fact that they did not claim to belong to any explicit category recognized by the Political Constitution of Colombia of 1991, nor did they claim other national origins, but that they position themselves as mestizo Samarios. Outside this specific corpus, we also questioned other groups who, for their part, identified as being Indigenous, Venezuelan migrants, or descendants of Africans, but, as we said, we are not using this material in this article. In addition to providing a basis for the study of phonic variability, this corpus has allowed us to formulate hypotheses about the reception of this phonic variability relative to other varieties.¹⁶ It is for this reason that every interview consisted of obtaining metapragmatic data from a number of questions on the Spanish language, the objective of which was to identify which varieties of Spanish are perceived as different and how respondents value different varieties of the language on a national, regional, and local scale (see appendix). The investigation method was built on the protocol proposed by the LIAS (Linguistic Identity and Attitudes in Spanish-Speaking Latin America) group and presented in Chiquito and Quesada Pacheco (2014). We added a few questions to obtain responses on a more local scale, as opposed to all Spanish-speaking countries combined. Our goal was to record respondent’s opinions about variants of the Spanish language on the Caribbean coast of Colombia in general, and those variants specific to the city of Santa Marta and its neighborhoods. These 36 interviews resulted in nearly 27 hours of recordings, averaging 43 minutes each. The vast majority of the interviews took place at the University of Magdalena and were conducted by two foreign investigators, both of whom identified and were perceived as White. The first investigator, who holds a doctorate in linguistics, is an Australian national. When the interviews were being conducted, he had been working as a professor at Magdalena University for eight years. The second investigator, the author of this article, also has a doctorate in linguistics and had been in Colombia on a research trip for nearly two months at the time the interviews were conducted. As shown in the presentation of the corpus (Blestel 2022), the identity of the investigators and the location of the interviews could constitute a source of bias. Indeed, it is possible that respondents’ answers were colored by what they thought we were expecting to hear and that the people questioned were apprehensive. Despite doing everything we could to forestall false beliefs and to reassure the Samarios questioned as much as possible, these aspects must be considered and pondered when examining the data presented hereafter. We must specify that, while we followed the order in the various sections of the questionnaire, we left the discussions and comments that emerged during this situation unstructured, which sometimes resulted in relatively long interviews (up to 1 hour 20 minutes), which were even more rich and varied.

After graphic transcription and sequential division, the data were analyzed both quantitatively, by processing the answers to each section of the questionnaire in a Microsoft Excel file, and qualitatively, by labeling and identifying what seemed to us to be recurring themes and topics of interest in the Nvivo 12 digital qualitative analysis software. As portended in this citation by Williams, according to whom “a definition of language is always, implicitly or explicitly, a definition of human beings in the world” (1977, 21, quoted in Woolard and Schieffelin 1994, 56), it quickly became apparent that obtaining answers on language practices independent of commentary about the people who follow these practices would be difficult,¹⁷ to the point of being impossible. And yet, for the purposes of the study, given that we expected the comments on these practices to give us clues about the linguistic elements perceived as different (and therefore mentioned), we applied different labels to the clues outlined below:

1. i) Metalinguistic indications, more specifically (supra)segmental, expressed spontaneously. This could apply to the manifestation of segments (“they don’t pronounce s’s”; “they swallow r’s”) or the perception of suprasegmental differences (height, volume, tempo, etc.);

2. ii) Psychosocial indications about the speakers described (laziness, aggressiveness, education, politeness, etc.), whether they were expressed spontaneously or induced by the questions (“That’s how the Cachacos are,” etc.);

3. iii) Indications relating to identity (“the Indians,” “the Blacks,” etc.) and/or where people live (Cartagena, Bogotá, etc.).

Furthermore, we also noted the postures of these Samarios with regards to the three previous points when discursive elements allowed us to do so. A posture of othering could, for example, take the form of using the third-person plural (“that’s how they speak”). Conversely, an inclusive we was evidence of an appropriation (“that’s what we do”). Similarly, belonging and rejection could be expressed in the form of ethical (good/bad), aesthetic (pretty/ugly), or affective (nice/unpleasant) value scales.

By crossing this set of indications, we intended to update the main axis of differentiation that occur as a result of these semiotic practices being understood to categorize participants, objects, or practices while “simultaneously taken to be icons of (resembling) the abstract relationships presumed in the schema” (Gal and Irvine 2019, 123). We attempt to explore the emergence of this schema hereafter.

Regionalized and Polarized (Supra)Segmental Indications

Above, we explained that Colombians eagerly opposed some of the country’s large regions. The most common instance was the polarization between the inland Cachaco and the coastal areas. In this context, we made an initial request in Nvivo for the purpose of crossing the postures expressed by the Samarios—simply in terms of adherence and rejection at first—according to the regions with which these accentual characteristics were associated in the set of answers, that is, without looking at the details of the answers to each of the questions on the form. The objective of this query was to verify whether the regional contrasts inherited from the colonial period had a counterpart in terms of epi- and metalinguistic comments. If so, we wanted to see the nature of these comments.

The postures were chosen among the corpus fragments relative to the (supra)segmental indications specifically (meaning the prosodic and segmental features that can be attributed to the “accent”). These indications therefore include all indications of a prosodic or segmental nature expressed spontaneously by the people questioned. The indications could include tempo (speaking quickly or slowly), pitch (“some speak of an ascending pitch in inland speech”), volume (speaking loudly versus speaking softly), the pronunciation of some consonants or groups of consonants, and so on.¹⁸ The result of this first query is very clear, namely, that comments vary according to region: the regional distribution of the 1,085 (supra)segmental indications described—731 positively (adherence) + 354 negatively (rejection)—shows a clear difference between the Andean inlands¹⁹—where the accent is subject to primarily positive judgments (85.2 percent of all opinions expressed)—and the rest of the country (fig. 3).

Figure 3. Distribution in percentage of indications of adherence versus rejection of (supra)segments spontaneously attributed to four Colombian regions by 36 Samarios questioned.

Clarity, tempo (slow speed), and volume (low) were the spontaneously expressed characteristics most often attributed by Samarios to speech in the inlands.²⁰ In terms of segmental pronunciation, the pronunciation of the /s/ (as aspirate [s] rather than the elision characteristic of the coast) also seems to be salient for the Samarios questioned, as well as being perceived as the correct pronunciation. This can be used to interpret transcript 1, in response to the question, “Which are the three regions, areas, or places in Colombia where you feel people speak better?” (question 9; see app.):²¹

Transcript 1

Here, the question reflects the underlying assumption that there is a “better” way of speaking and that a relevant scale to hierarchize this is “regions, areas, or places.” The answer can actually be interpreted as if the interviewee has difficulties aligning with this assumption, since he stops on the term mejor ‘better’ and hesitates. One might think that this term, unlike the expression “correctly” in question 20, which we deal with below, is not clear, since what is best suited to the interview is perhaps the local variant in Santa Marta, the one he speaks. However, what is interesting here is that this respondent finally continues with what seems to be the “best” according to academic standards by referring to the notion of “completeness” in the sense that no segments would be muted. This notion and that of the adequacy of inland speech with normative graphemics, owing to the fallacious idea that it represents “authentic Spanish,” were very often expressed, directly or indirectly, in the responses collected: “pronuncian todas las letras” (they pronounce all the letters) versus “se comen las letras” (they swallow the letters).

Conversely, the region of Amazonia acts as a counterpoint in figure 1, in that 80 percent of the characteristics attributed to the speech of locutors in this region were subject to negative commentary. The reasons invoked were vaguer: the “accent” in general but also the exolinguism of the region’s inhabitants, which was considered to mechanically prevent them from speaking a clear enough version of Spanish to merit respect.

However, these two initial observations do not seem to point to a racial correlation of the same kind as that associated with the vast regions,²² in that, here, the Pacific and Atlantic coasts seem to enjoy more well-balanced distributions in terms of adherence and rejection. At the very most, at this stage, we can assert that the Andean-centrism mentioned above is operating as well, which is not very surprising in terms of phonic norms. Indeed, there is a firm consensus that, even though it may sometimes be perceived as unpleasant or annoying, the Cachaco accent is the norm. This norm is interpreted as being evidenced by segmental pronunciations that comply with current graphemic standards, in combination with low volume and moderate speed. The answers isolated from some of the questions on the form point in the same direction: 57.5 percent of answers to the aforementioned question 9 (“Which are the three regions, areas, or places in Colombia where you feel people speak better?”) were inland regions of the country, as opposed to 38.4 percent for departments on the Atlantic coast, finishing in second place.²³ Similarly, to question 20, which clearly tends this time toward the elicitation of a normative variety by asking who spoke most “correctly,” 58.3 percent answered the inlands, as opposed to 7.5 percent for the Atlantic coast, which also finishes in second place. For question 34 (“Which way of speaking do you associate with high economic resources?”), the results were again 51 percent for the inlands and just 14.3 percent for the coast, which even then was in second place. This last result shows that the Samarios associate the Andean norm with economic power, which is hardly surprising given the socioeconomic history that we have exposed above.

However, we suggest examining the data in greater detail. In figure 4, the Pacific coast seems to enjoy, as previously stated, a distribution leaning more toward adherence: 79.9 percent versus 23.1 percent rejection. Nevertheless, if the results are analyzed by department, the data appear in a new light. A very clear trend emerged from these figures: 90 percent of the suprasegmental characteristics attributed to Chocó were subject to negative comments, while the Cauca department presented a diametrically opposed situation. What is it about the speech of Chocoanos that causes it to be so poorly perceived?

Figure 4. Distribution in percentage of spontaneous indications of adherence versus rejection of (supra)segments attributed to four departments on Colombia’s Pacific coast by 36 Samarios questioned.

The “Beating” Speech of Those of African Descent

The distinguishing characteristics of Chocó that prompt rejection are segmental in nature: the Chocoanos’ speech is “tangled” (enredado, 1M1S2), they omit segments to the point of having “atrophied” (atrofiado, 2M2S1) the language, which makes it sound “weird” (raro, 1F3S2) and difficult to understand. Above all, the majority of comments about Chocó mention the fact that they speak in a “beating” manner (golpeado). This expression, which is by far the most frequently used metapragmatic label to describe the speech of the other (it occurred in the answers of 32 out of 36 Samarios questioned), attracts attention not only because of the violent connotations of the verb golpear ‘to beat, to hit’ itself but also because of the analogical process on which it is based. Indeed, the expression hablar golpeado refers, particularly on the Caribbean coast, to a very specific type of segmental pronunciation: that of consonant gemination in the occlusives, which is a result of the regressive assimilation of the liquid /ɾ/ and /l/ in the syllable coda (e.g., caldero ‘cauldron’, pronounced [kad.ˈdeɾo]; cartón ‘cardboard’, pronounced [kat.ˈton]). At any rate, this is what was described—in their own words, of course—by the locutors questioned and asked to provide details about what they meant by this expression during the interviews. Such gemination of the occlusives, perceived as “beating,” is associated not only with inhabitants of Chocó but also, to an even greater extent, with the inhabitants of the southeastern portion of the Caribbean coast—primarily in the departments of Bolívar, Córdoba, and Sucre, which some of the locutors questioned associated explicitly with Chocó, because they are also populated by people perceived as having African ancestry.²⁴ This is what can be observed, for example, in transcript 2.²⁵

Transcript 2

Cristo Rey, a neighborhood on the outskirts of Santa Marta, is known for having been home to many people who had often been displaced for economic reasons from departments on the southwest Caribbean coast, particularly the department of Bolívar (whose capital is Cartagena; see fig. 1). This neighborhood is perceived as being Black and “happy,” apparently, if the title of this article from the El informador newspaper is to be believed: “CRISTO REY, el barrio de la ‘alegría’” (González Córdoba 2016). The article happens to offer a very instructive snapshot of some of the (apparently) positive stereotypes that are assigned to Afrodescendants (or those associated with them) in the city. This association between the linguistic varieties of three locales, Chocó, Cartagena, and Cristo Rey, that are geographically distant but that the respondent brings together by virtue of the African origins of their settlement sheds new light on the nature of the axis of differentiation that we seek to bring to light. Thus, if we examine which are the spontaneously cited locales when it comes to designating the “incorrect” varieties, these are also the areas that appear among the answers to question 21, which asks Samarios which people they think speak the most “incorrectly.” The three most frequently cited departments (from 54 answers provided in all) are Bolívar (29.6 percent), Córdoba (7.4 percent), and Chocó (5.2 percent).

That this precise type of assimilation—assimilation of implosive consonants happens to be very common in Romance languages as a whole—should be attributed to contact with African languages not spoken by the locutors of interest for several hundred years remains to be proven. Of note here is that, for whatever reason, (i) the Samarios notice this assimilation when it affects the rhotics and never mention the other assimilations, even though they are extremely frequent in Santa Marta (when the implosive is an occlusive, for example, the pronunciation of [et.sat.to] for /eksakto/ is a common pronunciation among Samarios themselves, but it doesn’t seem to be noticed and, at any rate, is not mentioned); (ii) this assimilation is disqualified and glossed with a very semantically loaded metaterm (beaten); (iii) this assimilation is explicitly attributed, by some locutors, to the African origins of the locutors who employ it.²⁶

Transcript 3

Furthermore, this differential perception, as well as the way it is rhematized, may be the manifestation of a perspective unique to Samarios as a whole, to whom historians and anthropologists attribute an erasure or even an oversight of the presence of Black people in the city (Rey Sinning 2002; Bénéï 2011). Perceiving and then commenting and disqualifying these assimilations of rhotics may be an additional (and, of course, unconscious) way of dissociating themselves from their African origins, in a whitening strategy also described on the Caribbean coast (Wade 1993; Cunin 2004).

… versus the “Melodic” Speech of the Paisas?

Conversely, while the inlands, particularly Bogotá, appear to be the national standard—which does not obviate the existence of a certain standard and a feeling of local loyalty in some situations, which we cannot discuss in this article—it is surprising to examine, by contrast, the way Samarios treat Paisa speech. The distribution of opinions expressed on (supra)segments attributed to them reaches 91.4 percent adherence, but that is not the only surprising element. Paisa speech is also the variant most often described as “beautiful” (bonita, hermosa) on an aesthetic scale, with 45 percent of this type of comment attributed to the Paisa region as compared with just half as much for Santa Marta and the capital. Similarly, responses to question 29 (“If you could change your accent, which is the region whose accent you would choose?”) also showed a preference for the Paisa region, which appeared spontaneously (answers were not restricted to multiple choices) in 33.3 percent of the 36 answers provided, as compared with 16.7 percent for Spain and 5.6 percent for Bogotá. More simply, on question 8, regarding which variant of Colombian Spanish was the most pleasing to them, the Samarios once again named the Paisa region in 21.8 percent of the 78 answers, with Santa Marta coming in second, at just 14.1 percent. As we can see, the questions that, beyond the norm, sought to elicit the aesthetic and affective preferences of the Samarios with regard to the varieties spoken in Colombia lead to a very robust tendency: the Paisa variety is their favorite variety, even before the local one.

(Supra)segmentally, the majority of justifications expressed were more or less the same as those for the inlands in general: Paisas correctly pronounce full words at a moderate speed and volume. In addition, their cantadito (tone, or literally “little song”) was often evoked, as well as the region’s characteristic vocal lengthening ([aː], in transcript 4).

Transcript 4

However, as opposed to what happens for Indigenous people or those of African descent, nothing was ever mentioned about the Paisa race (White, or perceived as such; see discussion above), which we can deem an unmarked standard, as described by Bucholtz (2011) in the context of her study in the United States.

Phonic Practices Serving an Ideological Edification

That a specific consonantic assimilation (attributed to those of African descent) or a vocalic lengthening (Paisa) elicit such passionate responses is a legitimate source of surprise for the linguist aware of the degree to which the phonic features of languages are varied, changing, and affected by multiple causes and factors. But the way these segments are remarked upon by the individuals interrogated indicates that by becoming audible, perceivable, and even remarkable signs, they are—no doubt unconsciously—invested with a role to play in the implacable mechanics of ideological constructions. There is no doubt that the majority of Colombians who perceive, then comment on, the “beaten speech” of a given person do not necessarily have ill intentions toward people of African descent. And the Samarios were just as often surprised when their preference for the Paisa accent was revealed to them when we shared the results of the investigation. But beyond the individualities, here we have touched on the inertia of value scales and ideological structures passed down from the colonial era. The metapragmatic labels of these structures are the symptoms: these differential perceptions are the result of a “particular mode of hearing and seeing” that is none other than the “effect of a regime of social power, occurring at a particular historical conjuncture, that enables, regulates, and proliferates sensory as well as other domains of experience” (Inoue 2003, 157).

Golpeado ‘beaten’, atrofiado ‘atrophied’, rajado ‘choppy’, and triturado ‘ground up’ are not neutral expressions to describe speech: they bear a violence incurred upon the language that happens to be attributed to Blackness, or to regions perceived as being Black or Indigenous.²⁷ The numbers in figure 1 can therefore be reconsidered in a new light: it is possible that the negative judgments attributed to accents in the Amazon region must be reconsidered in light of racializing ideologies. The same applies to speakers the Nariño region in figure 2, whom the Samarios disqualify as speaking too slowly (even though that is part of what’s valued in Bogotá and Medellín). The region is undoubtedly perceived as having a strong racialized presence. Let us add that they are also mocked as village idiots and stereotyped as such in the corpus.

On the other hand, inhabitants of Medellín and Bogotá are described as speaking so suave ‘softly’, so hermoso ‘beautifully’, so recatado ‘demurely’. Furthermore, these differences in treatment are to be considered relative to the psychosocial characteristics attributed to the people who speak these variants: while seriousness, courtesy, and cultivation are readily attributed to the Cachacos, the Atlantic coast wins the prize for “vulgarity,” “lack of education,” and the best sense of humor. We cannot unpack those stereotypes here due to space limitations, but they also feed into the same differential schemas. These schemas are reiterated on the local scale and any exceptions to the rule are quite simply erased and not remarked upon. Hearing, noticing, valuing, and then naming the accent of the other is never neutral. Our aim was to shed light on one of the many mechanisms of co-naturalization of social, regional, and racial disparities that contribute to regenerating the structures of society. The subject matter presented here represents an initial step forward, which must be confirmed by acoustic analyses and perception tests.