BOZAL IN THE CONTEXT OF RACIAL PARODY

The enregisterment of Bozal as the voice of (brutish) Blackness is part of a much larger story of racialization of Africans that has unfolded across half a millennium, and the world brought into being during that span. In the nexus of Western epistemic, economic, and moral orders that “invented Africa” and created the racialized category of “Africans” (Appiah Reference Appiah1992; Mudimbe Reference Mudimbe1988; Reference Mudimbe1994), one small but significant contribution has been the circulation of a distinctive, usually parodied “African” voice. From Cervantes, through the mid-nineteenth- to mid-twentieth-century heyday of “blackface” entertainments such as minstrelsy, to modern instantiations of Mock-Ebonics in the United States, there are countless instances in which a voice of “Blackness” has been ventriloquated, caricatured, and then denigrated according to what, borrowing from the blackface comic dialogue in the epigraph, I ironically call a trope of “brutology.” Although I focus narrowly on the trajectory of one such voice, Bozal, in Cuba, in doing so I do not mean to suggest that Cuban Bozal is or has ever been isolated from racializing currents across the Atlantic World (see Hartikainen Reference Hartikainen and Palmié2008, for a Brazilian example). While a fully comparative analysis must wait for another occasion, I will point out some connections that promise deeper insight into the dynamics of racialization at work in not only the Cuban case, but elsewhere as well.

Lately, a variety of contemporary mocking registers have received serious analytical attention. Jane Hill's pioneering work on Mock Spanish laid out the basic premises of subsequent work on other racializing mocking or “mock” forms (Hill Reference Hill1993). Most important is the parodic voicing of mocking forms, in which speakers use an ironic or disparaging stance to distance themselves from identification with a marked social voice they deploy. Related to this are the linguistic features framed as tokens of the disparaged register, which tend to be a few widely recognized forms—shibboleths—that often do not accurately reproduce actual, non-mocking speech practices of the disparaged group. These socially charged tokens are re-contextualized by association with negative, or at least extremely limited, personal or group attributes drawn from widely-circulating racial stereotypes. Ronkin and Karn's (Reference Ronkin and Karn1999) example of Mock African-American Vernacular English (AAVE) in overtly racist web sites shows how stigmatized tokens taken to represent AAVE, such as double negatives, are associated with ugly stereotypes such as lack of education and criminality, to create what most Americans would recognize as an explicitly racist caricature of African Americans. This is in contrast to Hill's Mock-Spanish examples, in which the racializing consequences for Latinos are much less explicit and more deniable, because more or less stylized tokens of Spanish are associated with a light-hearted, humorous, fun-loving, and even cosmopolitan Anglo, English-speaking persona. Nonetheless, the way in which Mock Spanish re-circulates associations between Spanish and a laid-back or lazy, trivialized, party-loving, and even law-breaking persona, often through intertextual chains of implicature, accomplishes much the same racializing work as do more overtly racist mocking forms (Hill Reference Hill1993; Reference Hill, Gal and Woolard2001; Reference Hill2005; Reference Hill2008).

Recently, more complex voicing effects have been described for mocking forms. Chun (Reference Chun, Reyes and Lo2009), for example, explores critical re-appropriations of Mock Asian by an Asian American comedian. Roth-Gordon (Reference Roth-Gordon2011) examines the inevitable leakage across even the most robust attempts to distance a (White) speaker from the (non-White) persona of a mocking form he or she voices. She argues that such disorderly discourses are inherently asymmetrical, since Whites apparently suffer only momentary linguistic “contamination” in using Mock forms, while even non-White speakers who linguistically “discipline” themselves can never lose the taint of their bodily non-Whiteness. In the same issue of Journal of Linguistic Anthropology, Gaudio (Reference Gaudio2011) argues that linguistic appropriations of stigmatized voices where speakers align with, rather than distance themselves from, the marked voice can subvert some of the disparaging associations even while making use of the racializing effects—for example using AAVE forms in association with Nigerian Pidgin to index shared claims to transatlantic Blackness through “broken English.” For that matter, White British youth and young White American hip hop fans have been shown to claim alignment with Blackness through their admiring, rather than mocking, appropriations of “hip” markers of “Black” speech (Cutler Reference Cutler2003; Rampton Reference Rampton1995), and groups construed as non-White, such as Dominican American and Laotian American teens, strategically deploy AAVE and mainstream U.S. English features to negotiate their racialized positionings (Bailey Reference Bailey2001; Bucholtz Reference Bucholtz, Reyes and Lo2009).

Although the rich scholarship on Anglophone, and particularly American blackface minstrelsy has rarely attended to the workings of language, it has provided abundant examples of all of these forms of voicing, albeit sometimes too hastily describing texts such as song lyrics, characterizations, plays, and even advertisements as imitating what is assumed at face value to be Black dialect or West Indian English (e.g., Mahar Reference Mahar1999; Nathan Reference Nathan1962). Eric Lott (Reference Lott1993) has compellingly argued that mid-nineteenth-century American blackface entertainments reveal moments both of social distancing from Blackness, which were part of immigrant efforts to assimilate to the privileges of Whiteness, and of desire for and admiration of Blackness, in a context of extreme racism and racial intimacy. While Mahar (Reference Mahar1999) has critiqued Lott's analysis as giving a psychological reading too far from the social contexts of blackface performance, he has also argued that White performers used blackface, and voice, as a “disguise” and “vehicle” for social critique of dominant views and parodies of a much wider range of social types than just African Americans. Chude-Sokei (Reference Chude-Sokei2006) has developed the theme of disguise and “masking” to describe the use of blackface by Black performers as a medium for heightening and critiquing the “double consciousness” imposed by a racist society. He does so in his biography of the early-twentieth-century Black superstar Egbert Austin Williams, whose blackface virtuosity on the eve of the Harlem Renaissance ensured his erasure from theater and pop culture history. Chude-Sokei's focus on “black on black” minstrelsy highlights both the distinction between and the interpenetration of processes of self-identification and interpellation, subjectivity and subjectification.

The range of uses and effects of racial parody clearly show the significance of metalinguistic awareness in the stylistic choices of speakers and perceptions of their interlocutors—for example, in the very identification of a parodic theatrical style as “the speech of plantation slaves.” The recent body of work on contemporary mocking registers also demonstrates the larger point that language styles cannot neatly be matched to social identifications, and that the normal state of linguistic affairs is complex in its heteroglossia (see Agha Reference Agha2006; Coupland Reference Coupland2007). Meek (Reference Meek2006) gives a telling example of interpellation from a recent Hollywood film, in which one character's Native American identity is crudely established for film audiences in part by another non-Native character's use of Hollywood “Injun” English in addressing him (right before inflicting unspeakable violence against him), and in part by his silent response, construed as an index of stereotypical “Indian stoicism.” When a recognizable linguistic style does come to tightly index a particular social persona, like AAVE and Mock-AAVE index different takes on the meaning of Blackness, or White-invented Hollywood Injun English indexes a stereotyped figure of Native Americans, we must investigate how this happens. But in doing so, we must take care not to collapse representations, mocking forms, and actual speech of members of these groups into each other. We need to look instead for moments of interdiscursivity: citation, imitation, appropriation, and re-appropriation, whether they are acknowledged or not.

The modal use of Bozal in Cuba today to voice the presence of an historical African figure during spirit possession ceremonies is clearly not a mocking form, nor, for that matter, are most of the staged folklore performances of Bozal. In these cases, Bozal does mark a distancing between animator and author/principal speaking roles, but for the purpose of manifesting a spirit's presence in the body of a human host (see Du Bois Reference Du Bois, Hill and Irvine1992; Irvine Reference Irvine and Parkin1982; Keane Reference Keane1997), where the relationship is one of power and respect, not of irony, parody, or disparagement. Nevertheless, by tracing the historical context of these contemporary instantiations of Bozal, I aim to show that Bozal's trajectory includes genealogical links to the quintessentially and overtly mocking form of blackface comedy that complicate its indexical effects in the present. A consideration of the discourse history that gave rise to Bozal raises questions about how we should model interdiscursivity across larger scales of time and space, in light of longstanding questions about the effects of metalinguistic awareness and the “indeterminacy of contextualization” (Cameron Reference Cameron1995; Gumperz Reference Gumperz, Duranti and Goodwin1992; Silverstein Reference Silverstein1981; Reference Silverstein and Duranti2001). Our answers to these questions have implications for how we go about defining a register like Bozal.

BOZAL, METACULTURAL AWARENESS, AND WEBS OF INTERDISCURSIVITY

The earliest canonical Spanish writers in the age of Cervantes were already representing parodied “African” speech, influenced by an emerging Portuguese literary genre of fala de preto (“Black speech”) (Lipski Reference Lipski2005). African “types” joined in the dialogic play of voices in popular comedic theater and song, both secular and religious, at the dawn of the “Atlantic World.” The “father of Spanish theater,” Lope de Rueda (1510–1565), featured Africans speaking a distinctive register in his comic dialogues. Scarcely a generation later, the poet Quevedo (1590–1645) mocked the already established comic formula of representing “the Guinea tongue by changing the r's into l's and the l's into r's, thus Francisco = Flansico” (Stevenson Reference Stevenson1994: 485–86). Numerous popular religious songs of Iberia and the New World colonies, the villancicos of the seventeenth century, were called negros or negrillos because they animated rustic “African” voices. Sor Juana Inés de la Cruz, of late-seventeenth-century New Spain, employed a kit of dialectal markers and invented “Africanisms” that would be as recognizable among twentieth- and twenty-first-century purveyors of Bozalisms as to Cervantes and his era. Her “Africanisms”—expressions like “gurugú” and “he he he cambulé”—mark one moment in a history of characterizing African voices with nonsense words.Footnote ³ Even in making a case for mining these texts as “sources of evidence on earlier Afro-Hispanic speech,” Lipski sums up these Golden Age Bozal characters as “frozen stereotypes” (Reference Lipski2005: 93). Centuries later, twentieth-century Cuban poet Nicolás Guillén employed similar-sounding vocables in poems like “Canto negro” from 1931 (Guillén Reference Guillén1972, vol. 1: 122; Stevenson Reference Stevenson1994):

While Guillén's vocables speak to a specific early-twentieth-century context in which self-consciously national forms of cultural production were drawing upon then-current theories of the “transculturation” of African, European, and other “roots,” they perpetuated what had become a several-century tradition of voicing the “African” as primitive Other. Through the trajectories of colonialism and postcolonialism, including genres of theater, literature, and scholarship, African figures and Bozal speech have circulated widely. Bozal voices continue to circulate not only in Cuba, but also even in sites where other traces of the African diaspora have been otherwise scrubbed from the national historical consciousness. In Uruguay, for example, Montevidean Carnival comparsas(ensembles) of “negros lubolos” feature blackface performers in a nation that defines itself as a product of European immigration, and in contradistinction to Black Uruguayan troupes known as llamadas that specifically perform an alignment to Blackness in Black neighborhoods (Andrews Reference Andrews2007: xiv, 59, passim; Remedi Reference Remedi2004).

Bozal thus emerged among Europeans as a parodic representation of Africans, a usage that has continued. Parodies require a doubling of voices—a ventriloquation, as Bakhtin puts it—where one voice evaluates a co-present voice. The comic dialogue in the epigraph is a good example, especially considering that the playwright himself first performed the role of Aniceto (in blackface, of course). Linguists seeking the “missing Spanish creole” have sought to reconstruct “original” slave speech from today's Bozal and from nineteenth-century sources like comic plays. I assert that they are on the wrong track, and that we must attend to the less obvious evaluative voice, which implies an alternative interdiscursive trajectory, not necessarily exclusive of actual African voices, but problematic for any straightforward reading of those voices in historical Bozal texts.

The analysis I am developing here moves between the micro-discursive focus on communicative events and the longue durée of historical-cultural movement in order to consider the enregisterment of Bozal as a racializing process. The very term “enregisterment” refers to the reflexive social processes through which “cultural models of speech” emerge and can be traced historically (Agha Reference Agha2003; Reference Agha2006). Enregisterment occurs at the interdiscursive level, when events in the stream of discourse become recognizable instances of a social voice indexing a particular social persona. When the indexical link between type of person and speech style becomes tight enough at the metalinguistic level (if not necessarily in practice; see Ochs Reference Ochs, Duranti and Goodwin1992), we can call that speech style a register, whether or not it goes by a particular label—legalese, baby talk, Bozal.

There are two particular issues I want to explore in some detail. The first is the recognition of resemblance across moments of replication that is required for a generic type to become salient as a type. To attend to this issue of iconicity is to raise questions about how the conditions of interdiscursivity are shaped by different degrees or types of metapragmatic awareness of the similarities linking disparate moments of discourse, with or without expressly metacultural awareness, such as objectifying a cultural form to make it available for commentary. Consider the process of tracing out discourse histories to link events across time and genre: What are the ramifications to social actors’ ability to recognize when interdiscursivity occurs? What are the “limits of awareness” in determining that two separate instances of discourse are iconically linked (Silverstein Reference Silverstein1981; Reference Silverstein and Duranti2001; Reference Silverstein2003)?

To speak of processes of entextualization and enregisterment, then, is to posit that a fundamental similarity or set of similarities is recognized in comparing two or more semiotic events. The recognition of a “family resemblance” links disparate discourse events to each other, and, if the entextualization is robust, to an abstract and extractable “type,” such as a recognizable social persona, interactional trope, code, genre, or register (Bauman Reference Bauman1998). That is, the similarity that serves as a basis for an interdiscursive link must be culturally meaningful within some interpretive frame in order to be recognized (Agha Reference Agha2006: 78; Irvine Reference Irvine2005: 24, and others in the same 2005 special issue of Journal of Linguistic Anthropology). It follows that some degree of metapragmatic awareness is required for the recognition of any particular case of interdiscursivity, but we know that metapragmatic function often does not require fully articulated metadiscursive awareness that would allow for conscious reflection on the phenomenon. We also know how fickle, incomplete, and misdirected metadiscursive awareness can be, precisely because it is saturated by ideologies, linguistic and otherwise (Agha Reference Agha2006; Irvine and Gal Reference Irvine, Gal and Kroskrity2000; Irvine Reference Irvine2008; Silverstein Reference Silverstein and Lucy1993; Reference Silverstein2003). For this reason, we can describe many racializing discourses as covert or tacit: their racializing functions in replicating signifiers of racial difference go unrecognized by those participating in their circulation (Dick and Wirtz Reference Dick and Wirtz2011).

There is something fundamentally open-ended about interdiscursivity: iconicities linking two discourse events can be discovered at any moment, and likewise, such similarities can also be disavowed, ignored, or erased, depending upon the chronotopic horizon in which discourse events and types are being compared. Elsewhere I have demonstrated how Cuban signifiers of Blackness in folklore performance index a chronotope of colonial nostalgia that expressly links Blackness in the present to colonial-era figurations of Africans (Wirtz Reference Wirtz2011). Scholars, too, in our role as metaculture specialists, also engage historically-specific metacultural frameworks in our analyses of interdiscursivity: we may find that some interdiscursive links form historical chains, allowing us to reconstruct discourse histories—histories of uptake and circulation, for example, through which particular speech registers emerge and become salient.

But what of the shifting metacultural awareness of those who reproduce features indexing a recognizable register? How are they negotiating—by expanding or narrowing—the inevitable intertextual gaps? I want to focus, for now, on the dynamics connecting interdiscursivity to the creation of event-internal metacultural models of discourse history. Asif Agha (Reference Agha2006: 72–73) discusses how some interactional events formulate “virtual models” of their own socio-historical situation, and points out such models may or may not bear much resemblance to empirically verifiable discourse histories. One such virtual model is the chronotope relegating Blackness to “the timeless past,” wherein emblems such as Bozal speech, slave dress, and even movements and gestures are understood to be unproblematic icons of actual slave speech, dress, movements, and gestures. Just how has this semiotic time warp, linking signs of Blackness to this particular imagined slave persona, become so over-determined?

This brings me to the second issue I wish to explore here: I wish to reconsider the metaphor of “speech chains,” and offer instead the more complex metaphor of interdiscursive “webs.” Speech chains are a powerful idea in a semiotically informed, diachronic study of discourse, and one traceable in linguistic anthropology's own discourse history to Saussure, who provided the famous basic diagram of what he called the “speech circuit” (Reference Saussure and Harris1996: 11) (see Figure 2). He pointed out “the linearity of time” inherent in linguistic signals (ibid.: 70), and introduced a theoretical apparatus for delimiting the semiologically meaningful divisions in sound sequences. This apparatus includes “compar[ing] series of phrases in which the same unit occurs” in order to segment the flow of speech-sound (ibid.: 103). Recognizing similar units separated in time is fundamental to our ability to segment the stream of speech into meaningful chunks. Note in Figure 2 that the basic unit of his “speech circuit” consists of a speaker and receiver, with the two perhaps alternating roles.

Figure 2 Saussure's basic unit, the speech circuit, reproduced from Saussure (Reference Saussure and Harris1996: 11).

The contemporary idea of discourse chains, deriving also from Hilary Putnam and Saul Kripke's work on meaning and reference, allows us to shift scale from the temporality within unfolding interactions to that across interactions comprising longer-term discourses, and even across the longue durée, thereby tracing how semiotic objects—names, for example—are constituted across socio-historical chains of reference. An example of the speech chain model is Asif Agha's (Reference Agha2003) study of the enregisterment of British Received Pronunciation. Figure 3 shows that Agha's diagram is in essence an elaboration of Saussure's diagram, to represent links connecting multiple speaking events. I think the metaphor of a chain is misleading, however, in being too linear for the multiple reticulations that produce what look more like interdiscursive webs. Metaphors matter as models, and the kinds of rich analyses of unfolding discourse that describe multiple cultural movements, indexical orders, and scales of context are better described using the less linear and teleological, more open-ended “web” metaphor.

Figure 3 Asif Agha's speech chain model (after Agha Reference Agha2003: 247, Fig. 2).

But even the metaphor of a “web” has its limitations: it may be too static a structure to capture the open-endedness of real-time contextualization processes, in which new juxtapositions, such as those I will presently propose for spirit possession speech, invoke new orders of indexicality. Another potential replacement metaphor is that of discourse genealogies: ideologies of lineage direct our attention toward some lines of descent over others, but of course, with each subsequent generation, the number of ancestors doubles (two parents, four grandparents, and on). To apply the genealogy metaphor to the Bakhtinian “echoes” of prior speech in each subsequent speech event, this tension between the potential number of earlier speech occurrences that could influence a current speech event and those actually recognized as relevant is precisely what my metaphor of discourse genealogies seeks to highlight.

BOZAL'S ENREGISTERMENT IN CUBA

“Bozal” is derived from the Spanish word for “mouth.” Its semantic range includes “wild” and “untamed,” “ignorant,” and also “dog muzzle” and “bridle bit” as instruments of control. This label, with its polyvalent attention to speech and social control, was applied to African-born slaves (but not Iberian- or creole-born slaves), and it seems to have been in use in Spain and the colonies at least from the sixteenth century (Alonso Reference Alonso1958; Covarrubios Orozco Reference de Covarrubios Orozco2006 [1611]). Early on, the term also came to refer to the way that such Africans would have spoken Spanish, which has been consistently characterized, over hundreds of years and into the present, with terms such as “mixed up,” “corrupted,” “defective,” and “ugly.”Footnote ⁴

An early mention of Bozal in Cuba's historical record is a call by Havana's new bishop in the 1750s for Catholic priests to learn Bozal, if not African languages (cited in Castellanos Reference Castellanos1990: 71). Bozal received increasing metacultural attention with the vast growth in Cuba's slavery economy due to the sugarcane agriculture boom on the island after the 1790s. Whereas in 1792 slaves comprised about 31 percent of the population (87,000 out of 270,000), in 1827 the slave population alone had increased to 287,000, or about 41 percent of the total population of 700,000 (Bergad Reference Bergad2007: 17–18; Bergad, Iglesias García, and Barcia Reference Bergad, García and del Carmen Barcia1995: 39). Over time, a substantial free colored population arose, and by the 1862 census there were some 370,000 slaves and 200,000 free people of color, totaling 44 percent of Cuba's population of 1,400,000 (Bergad Reference Bergad2007: 10; see also Curtin Reference Curtin1969; Kiple Reference Kiple1976; Murray Reference Murray1971). Abolition came late and gradually to Cuba, being completed in 1886.

A catechism published in Havana in 1797, the “Explanation of Christian Doctrine Adjusted to the Capacities of Bozal Blacks,” advised priests to speak slowly and teach African slaves to enunciate words correctly. The author, Havana priest Nicolás Duque de Estrada, refers to “that language that they use without cases, without types, without conjunctions, without agreement, without order” (Laviña Reference Laviña1989: 66, 118; my translation). These characterizations of non-European language are, of course, all familiar enough from countless other cases of European colonial linguistics (for an early critique, see Boas Reference Boas and Stocking1974 [1889]).

Another elite commentary comes from Cuban geographer Estebán Pichardo y Tapia, who in 1836 published a dictionary of Cuban Spanish that included examples of Bozal, which he characterized as consisting of “grammar defects and sound changes” that created a “disfigured Castilian.” Below is a sampling of his textual representation of Bozal, together with my attempt at a more standard Spanish gloss, in the second line, to indicate what is marked as disfigured in his version, and my English translation in the third line. Pichardo y Tapia did not himself provide a gloss of his example, suggesting that he might have thought its denotational meaning to be clear to readers despite the many nonstandard orthographic forms he deploys to represent this speech style. Indeed, his heavy use of eye dialect (textual representations of nonstandard speech) suggests that he expected his audience of literate elites to already recognize Bozal. I indicate my uncertainties with parentheses in the glosses:

1. 1. yo mi ñama Frasico Mandinga, nenglito reburujaoro, crabo musuamo ño Mingué

   yo me llamo Francisco Mandinga, negrito (reburujador?), esclavo mi su amo Don Miguel

   my name is Franscisco Mandinga (out of place?)Footnote ⁵ little black, slave, my masterFootnote ⁶ Don [honorific] Miguel

2. 2. de la Cribanerí, branco como carabon, suña como nan gato,

   de la (carbonería?), blanco como carbón, las uñas como un gato,

   of the (coal yard?), white like charcoal, nails like a cat,

3. 3. poco poco mirá oté cribi papele toro ri toro ri,

   hace poco miro a Usted escribiendo papeles toro ri toro ri [nonsense vocables]

   for a while I watch you writing papers blah-blah-blahFootnote ⁷

4. 4. Frasico dale dinele, non gurbia dinele, e laja cabesa,

   Fransisco le da dinero, (está agobiado?) Footnote ⁸dinero y raja la cabeza,

   Francisco gives him money, the money (is not sufficient?), and breaks his head,

5. 5. e bebe guariente, e coje la cuelo, guanta qui guanta

   y bebo aguardiente, y me coge el cuero, aguanto aquí aguanto

   and I drink cane liquor, and get the whip, I endure here I endureFootnote ⁹

We do not know how Pichardo y Tapia produced this sample, nor from what source or sources, but note that the marked phonological, lexical, and grammatical features of this passage, as represented orthographically anyway, are in keeping with later representations of Bozal. He provides examples of phonological substitutions, including “r” for “l” and “l” for “r” (already noted in Spain, we have seen, and today described as typical of colloquial Cuban speech), phonological elisions, nonstandard verb conjugation, and a few other markedly non-normative syntactic features such as reduplication for emphasis or intensification (e.g., “poco poco” in line 3). Note, too, that this text is presented as a first-person account by a Bozal-speaker, in which markers of Bozal speech co-occur with stereotypical semantic indications of slave existence: his surname “Mandinga” is an ethnonym describing his presumed place of origin in Africa; he seems puzzled by “written papers” and mentions needing money and drinking cane liquor; and if my interpretation is correct, he seems to end on a note of misery. A crucial observation is that Bozal is always tightly linked to a slave persona, to the extent that even these early scholarly renditions of Bozal cannot help but voice it in a projected “first person” voice of a slave, in this case a slave whose self-descriptions exemplify DuBoisian “double consciousness.” I will return to the issue of voicing later, in my discussion of spirit possession, since it contributes to the chronotope of Blackness as “the timeless past among us.”

If this text and its accompanying characterizations represent elite, literate, White recognition of Bozal as African slaves’ distinctive way of speaking, other texts indicate how Bozal was being represented in popular entertainment for middle- and lower-class Whites. Vernacular musical theater, popular literature like comic verse and “costumbrist” (“quaint customs”) scenarios, and newspapers began to represent Bozal through blackface characters on the stage and in print. This is first evident in the “negrito” performances of Covarrubias in 1812–1815, who is described as blackening his face, but whose texts have not survived. The trend accelerated in the 1830s and 1840s thanks to writer and performer Crespo y Borbón and his African persona as Creto Gangá, and again in the late 1860s, with the teatro bufo. Cuban comic theater presenting “negritos” and derived from French, Italian, and Iberian comic theater was well-established long before U.S. blackface performers brought “Jim Crow” and the like on tours to Cuba during the U.S. Civil War (Leal Reference Leal1975: 21). Today we have access to numerous libretti and song lyrics from these vernacular theater genres, as well as advertisements and accounts of performances in newspapers. In these, we see a personification of Bozal speakers in what becomes a set character type, the negrito, with the diminutive ending on negro reflecting condescending White attitudes toward Blacks.

To briefly illustrate the early theatrical Bozal, consider a sample of an influential characterization by Bartolomé José Crespo y Borbón, who famously inhabited his invented alter ego, Creto Gangá, in live performances and in what were considered hilariously illiterate letters that Creto Gangá penned to Havana newspapers during the 1840s. Crespo y Borbón had already published comic dialogues and collections of poetry featuring the interplay of various “Cuban” voices, including Bozal, often under pen names. But once he created Creto Gangá, this became his best-known persona. Mary Cruz (Reference Cruz1974) documents how Creto Gangá's biographical details slowly emerged over numerous publications and newspaper announcements. In the persona of Creto Gangá, Bozal once again voices a projected first-person account by a putative African whose last name is also his ethnonym, although Creto Gangá could represent either a slave or a freedman. Creto Gangá's 1846 poem's title is itself an exemplar of his Bozal (ibid.: 50):

Similar kinds of phonological substitutions and elisions, nonstandard verb conjugation, and other syntactic “errors” of agreement and case are evident in this sample as in others. Note, too, that this character's concerns—with frivolous pursuits of Carnival—are in keeping with negative characterizations of slaves. And finally, note the comedy in this presumed dissonance of social indexicality between the attempted pretentious register of this versification and the Bozalisms throughout, perhaps akin to reciting a Shakespearian sonnet in a Cockney accent.Footnote ¹⁰

At this moment in Cuba's history as a Spanish colony, Cuba's literati and theater were still oriented toward the genres, fashions, and opinions of the metropole—Europe, after all, was clearly the seat of political as well as cultural authority, despite revolutionary movements in the Americas. Unsurprisingly, then, Bozal enregisterment continued along the lines already laid down in Spanish literature. As the logic went, African-born slaves were denigrated in part for their inability to speak properly, which served as proof that they belonged in their low status. Africans in Cuba who aspired to higher economic and social status were correspondingly ridiculed as rising above their station—that was why Creto Gangá's verses and letters to newspapers inspired hilarity among the predominantly White reading public. Things were poised to change, however, with the emergence of a proto-national Cuban identity that would layer a new, rebellious metacultural frame onto the existing one concerning the social significance of Africans, as I will describe.

Two additional trends emerged. First, Cuban vernacular theater produced a whole set of stock character types, standing for Cuba itself, and no marked racial, internal, or foreign group was left out: peasants (“guajiros”), Galicians, Canary Islanders, and even Yanquis were parodied. Initially, Cuban peasants and Bozal Blacks were represented with similar voicing devices—grunts, slurred words, and nonstandard phonology and grammar, and they only became differentiated over time. Second, amid this costumbrist proliferation of what Cuban theater scholars have called Cuba's homegrown “comedia dell’arte” (Leal Reference Leal1975: p.17), the prototypical negrito character itself, emblemized by Creto Gangá, proliferated into distinct types of Black characters who could be played against one another to heighten the comic effects. Innovations included juxtaposing Bozal-speaking “Congo blacks” to negros catedráticos or “educated Blacks,” where this term was meant to convey a humorous oxymoron. The teatro bufo also created scenarios depicting stereotyped mulatas, Black coach-drivers or caleseros, and negros curros, in an urban subculture of free Blacks who resisted White authority through fashion, speech, and economic independence. (These were nineteenth-century predecessors of Harlem “zoot suit” street culture in the United States.)Footnote ¹¹

Amid the proliferation, Bozal Blacks continued to be represented as crude-sounding brutes, for comic effect, but increasingly they became honest, earnest, and authentic brutes, even as audiences that had begun to think of themselves as Cubans could identify with such brutish resistance to the polished hand of Spanish imperial power. To illustrate, consider the interplay of stock character types in an example from the later teatro bufo of the 1860s introduced in my epigraph: Francisco Fernández's one-act “comedy of Cuban customs,” called “Los negros catedráticos,” builds its humor by juxtaposing the humble Congo, José, to the pretentious “educated Blacks” Aniceto and Crispín. The latter envision a perfect match between one's daughter and the other's son, and disdain José as a suitor, until it turns out that José, ill-speaking Bozal though he is, makes a lot of money as a stevedore. Compare these self-descriptions by José and Aniceto, which occur within a few lines of each other, early in the play (Fernandez Reference Fernandez1868: 10):

José (negro congo):

1. 1. Yo só congo trabajaore la muelle. Yo no toma güariente.

   Yo soy congo trabajador de los muelles. Yo no tomo aguardiente.

   I am a Congo dockworker. I do not drink cane liquor.

2. 2. Yo só libre. Yo ganá do peso tuitico lu día.

   Yo soy libre. Yo gano dos pesos todos los días.

   I am free. I earn two pesos every day.

Aniceto (negro catedrático):

1. 1. Escúchame negro hidrógrafo. ¿Tú sabes el crímen que has cometido con tu atrevida petición?

   Listen to me hydrographic black. Do you know the crime you have committed with your cheeky petition?

2. 2. Si no fuera por que soy un hombre instruío y metafísico que reconozco tu inadvertencia

   If it weren’t because I am a well-instructed and metaphysical man who sees your lack of caution

3. 3. é ignorancia, ya te hubiera elevado a las raíces cúbicas.

   and ignorance, I would already have elevated you to the cubic roots.

In this brief juxtaposition, we can see the emergent differentiation of Afro-Cuban voices, where the malapropisms, register infelicities, and silly locutions of the ridiculous, overreaching “Black professor” are counterposed to the steadfast, authentic, dumbed-down Bozal of the “Congo” laborer, who sounds humble, as befits his low social position—what Americans might call an “aw shucks” persona.

Bakhtin's (Reference Bakhtin1981) exploration of the carnivalesque roots of the novel highlights precisely what we see in Cuban teatro bufo, which is a dialogic play of voices that allows for experiments with new social alignments and identifications. Nineteenth-century Cuban theatrical representations of Black speech were double-voiced by White actors and writers interested in representing Bozal and its prototypical speakers in particular ways. As Laurie Frederik (Reference Frederik1998) has argued, building on the work of Rine Leal (Reference Leal1975; Reference Leal1982) and Robin Moore (Reference Moore1997), for vernacular theater the agenda of such actors and writers included reinforcing the social order of slavery and dealing rhetorically with political and social tensions, including those arising from the presence of a large class of free Afro-Cubans in a slave society. Scholars of Cuban vernacular theater agree in characterizing its representations of Black speech as what Jill Lane (Reference Lane2005) calls a “literary register of bozal as discursive blackface” intended as an “inherent parody,” and what Leal (Reference Leal1975: 22) describes as an effort to “aurally define” and ridicule the figure of the Bozal Black.

But it is also noteworthy that theatrical Bozal in the 1860s comes to be the register of the more sympathetic Black character, the one who recognizes, as José says in the sequel play, “¿Tó no só negro?” (“Aren’t we all black?”) (Montes Huidobro Reference Montes Huidobro1987: 11). Theatrical Bozal thus also came to mark the popularization of a notion of Cubanness, one ventriloquated through blackface characters and thus increasingly establishing Afro-Cuban culture as a basis for Cuban national identity (Frederik Reference Frederik1998). At the same time, pro-imperial and pro-slavery political rhetoric was heightening the sense of racial Otherness and danger that Blacks represented (Ferrer Reference Ferrer1999).

Above all, Cuban teatro bufo, with its interspersed dialogues, songs, and verses was, as Matías Montes Huidobro says, “meant to be superficial and subversive through a heightened awareness of language” (Reference Montes Huidobro1987: 48–49, my translation). Indeed, he argues that the Cuban teatro bufo, arising at a moment of increasing political tensions and censorship in the 1860s, amid nascent social awareness of the potential for an independent Cuban nation, can be compared to twentieth-century theater of the absurd in its emphasis on façade, word play, and parody.

Given this analysis, it would stretch credulity to assume that “Bozal” as it appears in these theatrical libretti attempted documentary accuracy in representing how enslaved Africans actually spoke, except insofar as it mocked nonstandard colloquialisms and errors common to many foreign learners of Spanish. Indeed, commentators at the time, such as the anti-slavery writers and activists of the Del Monte group like Félix Tanco, critiqued costumbrist depictions of Bozal as exaggerations and caricatures (Williams Reference Williams1995).Footnote ¹² Afro-Cuban elites of the late nineteenth and early twentieth centuries took pains to distance themselves from such false depictions (Lane Reference Lane2005; Morrison Reference Morrison2000). Lorna Williams (Reference Williams1995) argues that the shifting, often-contested portrayals of White and Black Cubans were tied to changing conceptions of national identity. Language ideologies, and particularly notions about how different social groups spoke Spanish, were certainly intertwined with racial and national imaginaries, creating mutually reinforcing notions about accents and social identifications (Gal and Woolard Reference Gal, Woolard, Gal and Woolard2001: 10). I agree with Lane's (Reference Lane2005: 48) assertion that what was enregistered in these nineteenth-century “literary” representations of Bozal in written text and performance was not the actual speech of Africans struggling to learn Spanish but an increasingly entextualized, imagined speech style associated with the caricatured persona of the African slave. What I am contending here is that the Bozal spoken today was heavily influenced by this discursive history of “blackface” Bozal.

Let me now take a moment to summarize the first strands in the interdiscursive web linking Bozal instances into a discourse history of “brutology,” to again borrow Aniceto's neologism directed at Bozals. I have traced a genealogy of what, for shorthand, I will call scholarly and theatrical metadiscourses offering representations of Bozal. My intention with these labels is to condense a lot of detail in order to focus on interdiscursive relationships across entire genres and contexts. Clearly, both scholarly and theatrical metadiscourses of Bozal could be expanded into their own interdiscursive webs, constituting those discourses as recognizable and even nameable types (e.g., across events, authors, texts, and performances). Moreover, as I have indicated, when we examine these discourses in greater detail, we find each has its own distinctive and often complicated stance toward the Bozal voices ventriloquated through it, very often juxtaposed to other voices in classic Bakhtinian heteroglossia. Each followed its own circuits, and theatrical Bozal likely found a wider audience than did the catechisms or dictionaries.

Both scholarly and theatrical Bozal presupposed and further enregistered a certain figure of the African slave, the alleged original speaker of Bozal, but in neither case is actual slave speech a necessary “source.” The influence of actual slaves’ speech on the Bozal figure represented in these metadiscourses is uncertain, and there is reason to doubt that there was much. But in any case, it was overshadowed from the start by other metacultural concerns driving the producers and audiences of scholarly and theatrical Bozal, including an overwhelming marginalization of Africans in the racial order. Notice, too, that because Bozal as a way of speaking, and the figure of the Bozal to which it was linked, came to have their own metacultural life as objects of discourse, we can see that the specific intentions of various authors and performers are beside the point in tracing uptake and further circulation of Bozal as a type. Lipski's (Reference Lipski2005) and others’ very careful efforts to delineate which historical sources give more or less authentic representations of Afro-Hispanic speech, based on those sources’ apparent intent to parody, denigrate, celebrate, or uplift, do not, and cannot, dispel our doubts about the impact of circulating metacultural models of Bozal on even the most careful ears of the most astute and sympathetic observers of Afro-Cubans. It seems likely that few people in Cuban society were unaware of the social indexicality of Bozal as “the speech of African slaves,” given the pervasiveness of even just scholarly and theatrical instantiations of Bozal in nineteenth-century Cuba, not to mention other possible pathways for Bozal's enregisterment: journalistic, literary, musical, religious, and juridical. The intertextual webs enregistering literary and theatrical Bozal are robustly attested.

In the early years of the Cuban Republic, Afro-Cubans struggled for equal rights as full citizens of a nation whose independence many of them had fought for. Even as the heroic figure of the mambí fighter, the uneducated rural Black ex-slave who fought the Spanish imperial forces armed only with a machete, took on national political importance, the crude and countrified speech attributed to poor, rural Black creoles had become a new target for Bozal (Lipski Reference Lipski2005: 167–70). This expansion meant that Bozal significantly overlapped with vernacular Cuban Spanish but carried all the negative racial associations it had long since accumulated as the speech of Bozal African slaves, and it became a weapon used against Afro-Cuban political claims. Cuba's founding folklorist Fernando Ortiz himself, in the early twentieth century, elaborated an equivalence between Africanness, witchcraft (understood as superstition), and criminality in a series of influential books and articles delineating Afro-Cuban social types and cultural forms, starting with Los Negros Brujos (The black witches) (Ortiz Reference Ortiz1973 [1906]). Its parenthetical subtitle was, “Points for a Study of Criminal Ethnology.” Afro-Cuban elites of the late nineteenth and early twentieth centuries took pains to distance themselves from such false depictions and to instead align themselves with nationalist political rhetoric of racial brotherhood and equality, epitomized by José Martí's writings and enacted as a goal of the multiracial insurgency (Ferrer Reference Ferrer1999; Lane Reference Lane2005; Morrison Reference Morrison2000).

In the struggle over the role of Black citizens in the new Cuban nation, twentieth-century scholars took up the nineteenth-century scholarly and theatrical representations of Bozal as if they were transparent records of slave speech. Ortiz wrote about the negro catedrático as if this character reflected a real social type rather than a theatrical parody (Lane Reference Lane2005: 76–77). His study of the infamous nineteenth-century social type, the negros curros, a resistant subculture of free, urban Blacks, relied on characterizations of their speech by nineteenth-century costumbrist writers such as novelist Cirilo Villaverde. For example, to illustrate the curro's “affected, insolent, and defiant” speech with its “special jargon” and “peculiar pronunciation” mixing up /l/ and /r/ and substituting /i/, he cited an 1848 poem by José V. Betancourt called “El Negro José del Rosario.” As illustrated in the first stanza below, the poem voices the persona of a negro curro using eye dialect such as “nasí” for “nací” (indicating a shift in pronunciation to /s/ instead of Castillian /θ, written as unvoiced “th” in English) and “Manglai” for “Manglar” (/i/ for /r/), as well as terms from urban jargon (Ortiz Reference Ortiz1986: 66–70):

Comic and costumbrist texts thus became the ethnohistorical evidence of later generations (Lane Reference Lane2005: 78–79). Scholarly and theatrical representations of Bozal not only survived the end of slavery and birth of the independent Cuban Republic at the turn of the twentieth century, they outlived African-born Cubans themselves. It is the nineteenth-century legacy of scholarly and theatrical representations of Bozal that most directly influenced twentieth-century literary and theatrical genres that together comprised the many facets of the Afrocubanismo movement in the 1920s and 1930s, including novels, poetry, early folklore studies, song lyrics, and zarzuelas (light comic operas) (Moore Reference Moore1997; Williams Reference Williams1995). Indeed, zarzuelas, often in blackface, were enormously popular right up to the 1959 Cuban Revolution (and they have reemerged in nostalgic revivals of the 1990s) (Moore Reference Moore1997; Thomas Reference Thomas2009). In sum, the enregisterment of Bozal as a metacultural object of attention had taken on a life of its own, with comic theater influencing not only later theater, but also seemingly unrelated genres such as the emerging “criminal ethnology” and “folklore studies.”

Ortiz's sister-in-law Lydia Cabrera contributed groundbreaking folklore scholarship from the 1930s through the 1980s, including influential literary representations of Bozal in her short stories and ethnographic studies (see Castellanos Reference Castellanos and Castellanos1994; Rodríguez-Mangual Reference Rodríguez-Mangual2004). Her work, even more than Ortiz's, has been and continues to be directly taken up by practitioners of Afro-Cuban religion as an ethnographically and religiously authoritative source, constituting an interdiscursive link connecting comic theater typifications and folk religious practices. Her entextualizations of Bozal were ostensibly grounded in how she heard her elderly Afro-Cuban informants speak, most of whom would by then have been at least at one generation's remove from their African-born predecessors. But Cabrera, herself elite and White, would not have been innocent of earlier and cotemporaneous literary and theatrical instantiations of Bozal any more than Ortiz was. Moreover, caught up as she was in Paris’ “primitivist” craze when she lived there during the 1930s, she was certainly aware of the productions of fellow members of the AfroCubanist movement, such as poet Nicolás Guillén and novelist Alejo Carpentier. Cabrera's work attests to her ethnographic sensibilities, but also to her equally excellent literary gifts. It is important to remember, too, that she was initially working without recording technology, and that the youngest of any African-born informants would already have been at least in their seventies by the late 1930s, when she began her ethnographic investigations in earnest. Importation of captive Africans had diminished to a trickle of illegal shipments during the final decades of Cuban slavery, so most of her informants would have been Cuban-born, not “Bozal.”Footnote ¹³ It is dubious, then, to assume that the Bozal-inflected voices through which she conveys the wise and funny stories, and herbal and magical lore, are transparent records of actual African-born Cuban speech (see Lipski Reference Lipski2005: 163–68).Footnote ¹⁴

Cabrera's Bozal examples demonstrate a similar range of features to those evident in the cases I have already presented, as can be seen in a small excerpt from a story recounted in El Monte (Reference Cabrera1993 [1954]: 183), a book that in turn has been mined by creolists seeking evidence of how her elderly informants spoke. These lines are from a song embedded in a story she attributes to an informant called “R. H.”:

Thus far, I have brought the White scholarly and theatrical blackface representations of Bozal into the twentieth century. But what of actual African-born slaves who would have spoken to their children, creating a completely separate circuit for African-inflected Spanish? There are two related metadiscourses of Bozal with sociologically distinct patterns of circulation among Afro-Cuban descendants and cultural heirs of slaves, who also produced (and produce) a fully oral register of Bozal, both as recollections of how elderly African-born relatives spoke in the early decades of the twentieth century (Ortiz López Reference Ortiz López1998), and a register of religious performance. When in trance, many folk religious practitioners, of all racial backgrounds, are possessed by spirits of deceased Africans and African deities like the orichas of Santería, who are often characterized as speaking in Bozal. Bozal thus continues to demand a first-person voicing in which the speaker performs the persona of an African speaking Spanish, whether as deity, free person, slave, or religious authority.

This indexical connection is so powerful that when, during a sociolinguistic interview, the linguist Luis Ortiz López asked a young Cuban man to demonstrate Bozal speech, the man first “imitated a black woman during a spiritual session,” then actually fell into trance, possessed by a muerto (Ortiz López Reference Ortiz López1998; see Wirtz Reference Wirtz2007a). I repeat here part of the sample transcribed by Ortiz López, with my standardized gloss and translation below each line:

1. 1. Niño, tú tá queré que lo negro áa decí cuanto yo vá hacé, si me tá acodá

   Niño, tú quieres que el negro diga en cuanto yo voy a hacer, si estoy de acuerdo

   Kid, you want the Black man to speak as much as I will if I am willing

2. 2. Cuando yo tá vení de lo tierra mío sí poqque yo tá sé negro de nación

   Cuando yo vine desde mi tierra es porque yo soy negro de nación

   When I came from my land it is because I am African-born

The spirit describes himself as a negro de nación, a more polite colonial-era label than “negro bozal” for someone born in Africa. In his speech we see features characteristic of theatrical and scholarly Bozal, including the presence of proto-creole verbal aspect markers similar to those that Germán de Granda identified in Cabrera's transcriptions of her elderly Black informants. Indeed, Cabrera's work has been tremendously influential among folk religious practitioners, and her book El Monte has become something of a religious source-text, alongside the religious notebooks Santería practitioners’ have a tradition of keeping, some of which have been published as manuals (Dianteill and Swearingen Reference Dianteill and Swearingen2003; Duany Reference Duany1988; Rodríguez-Mangual Reference Rodríguez-Mangual2004). Cabrera's writings have become a source of Bozal, as well as other specialized religious language and knowledge, treated by some as a transparent record of orally transmitted knowledge passed from Bozal slaves to their genealogical and cultural heirs. This is but one case illustrating my more general point that some later-appearing discourses in the interdiscursive web now mediate the metacultural interpretation of earlier discourses, such that contemporary representations of Bozal are interdiscursive with multiple antecedents.

Interdiscursive relationships can exist between coeval metadiscourses as well, even when those discourses follow quite different sociological circuits, because even actors in separate social positions are nonetheless aware of each other and of more socially distant discourses, if only in a vague and general way. It seems likely that Lydia Cabrera was influenced by both literary “blackface” Bozal and elderly Afro-Cubans and folk religious practitioners, although she explicitly type-sourced only the latter. Likewise, Afro-Cuban descendants of slaves would have been aware of, and even performed in, consumed, or critiqued some of the ubiquitous “blackface” Bozal that was saturating popular theater, music, and literature.

In religious contexts, Bozalisms and biographical details of an African or slave persona usually co-occur with markers of one or another register of folk religious practice, be it Lucumí used in Santería, or the “Congo” jargon of the Reglas de Palo. The entire combination is often referred to as “lengua” (“tongue”), shorthand for “African tongue” (Wirtz Reference Wirtz2007b). Among religious participants, virtuosic displays of ritual jargon from those working with or possessed by African deities or muertos index religious authority, expertise, and community belonging, while Bozal speech authenticates the presence of deities and spirits. With the exception of Cabrera's studies, these collocations of Bozal and ritual registers are not present in earlier literary-theatrical instantiations of Bozal. Even those who are not religious experts can appreciate the esoteric and foreign African “flavor” of such speech, reminiscent as it is of the invented “Africanisms” of Sor Juana and Nicolás Guillén alike. To give a very brief example, consider the utterance below from a Santería ceremony in 2000, in which an elderly female priest was possessed by the oricha Obatalá and spoke to a small group I was in (the parentheses indicate a response from another participant).

The oricha refers to two of us by pointing and using what appears to be a neologism invented on the spot, “el suyo va el suyo” (literally the denotationally confusing “yours goes yours”) that my Cuban transcribers described as a Bozalism. The oricha continues with nonstandard verb uses, including a present tense, third-person conjugated “está” in place of the first-person conjugation and future marking that would be more standard in colloquial Cuban Spanish. She also uses two words from Santería's ritual jargon, Lucumí, emí (I) and ara (land, body), inflecting the latter with a Lucumí emphatic particle (-o). The overall effect of this utterance, especially in the full interactional context, was of a Bozal African voice struggling to express itself in unfamiliar Spanish, and of indirectness and hidden layers of meaning within a humble and earthy way of speaking.

This and other transcriptions of contemporary ritual performances of spirits and saints I have made, in this way, show the same range of phonological and syntactic features being used to mark Bozalisms, albeit alongside additional discursive features and within an interactional context very different from most of the literary and theatrical examples I have discussed (e.g., Wirtz Reference Wirtz2011; and see below). Today, many such religious and folklore performances draw upon and metaculturally refract earlier “literary” and other contemporary oral instantiations of Bozal.

Analyzing a corpus of ritual speech recorded during Regla de Palo Monte ceremonies in western Cuba, Armin Schwegler (Reference Schwegler, Fuller and Thornburg2006) catalogues Bozal features that appeared, including:

• • morphological reductions (e.g., yo te llama, “I call you”);

• • phonetic alterations (riba instead of [a]rriba, “on top”);

• • grammatical simplifications (e.g., elimination of articles);

• • grammatical overextensions (e.g., use of genderless article lo instead of el or la);

• • reduplications (e.g., tiempo tiempo, “a long time ago”);

• • omissions of syntactic complementizers (e.g., tiempo tiempo ___ [que] te llama);

• • deletion of prepositional “of” (e.g., cuarto ___ [del] fundamento);

• • uninflected ta + INFINITIVE;

• • archaic Spanish-derived vocabulary (e.g., ague, “today”)

The parenthetic examples are Schwegler's, but examples of all of these features have appeared in this article's examples as well. Balga Rodríguez (Reference Balga Rodríguez2000–2003) mentions many of the same features but characterizes them as, for the most part, typical of the mistakes any non-fluent speaker of Spanish might make. We must therefore return to the question of how to define Bozal: there may be characteristic markedly nonstandard linguistic features to expect in a Bozal sample, such as those listed above, but this still begs the question of what exactly makes a segment of speech or text “Bozal,” rather than some other kind of nonstandard or disfluent speech? I do not think we can rely on a decontextualized set of diagnostic features, because what really identifies a text segment as “in Bozal” is that it draws on a repertoire of “nonstandard” linguistic resources to voice a prototypical Bozal speaker: a Bozal African. The circularity should by now be apparent: Bozal is what Bozals spoke, and it reflects their unfamiliarity with Spanish. This circularity should not be surprising in light of the conditions of Bozal's enregisterment as a mocking or parodic form, but it is too seldom acknowledged by those who study Bozal.

Alongside religious and folklore representations of Bozal, there has emerged a new generation of scholarly metadiscourses about Bozal. Linguists studying marginalized Spanish dialects and seeking evidence of possible Spanish creolization (McWhorter Reference McWhorter2000) bring together examples culled from all of the earlier and contemporary representations of Bozal as a single, unitary register that they characterize as the speech of African slaves. Granda (Reference Granda1994) describes two approaches to collecting exemplars of Bozal, one focused on written historical sources, and the other directed toward oral traditions. He and other creolists have, of course, pointed out the caricatures evident in some literary sources, even as they start with the telling assumption that all represented speech of colonial-era Africans belongs to the category of Bozal (Castellanos Reference Castellanos1990; Granda Reference Granda1988; Lipski Reference Lipski2005). That is, they make a secondary distinction by setting off sources deemed authentic or spurious because contaminated by stereotypes, but they still retain Bozal as a totalizing category. Overall, by virtue of finding historical linguistic value in Bozal exemplars, creolists have initiated a whole new, positive metacultural reappraisal of previous Bozal discourses. But in doing so, they also do the metacultural work of arranging neat, linear discourse chains that point back to actual slave speech as the original “baptismal event” creating Bozal.

We should be wary of assuming that the enregisterment of Bozal as it has come down to us tells us much at all about how African-born slaves once spoke Spanish, and whether their speech could even be collectively characterized as a unitary register (except retroactively). Instead, my analysis here speaks volumes about how they were viewed, and how these simultaneously over-determined and condensed representations of Blackness, and particularly of Africanness, have served a variety of ideological ends over the decades and centuries. It shows how claims to represent Black speech authentically (usually staked on having heard actual slave or ex-slave speech) have, at some nodes of the interdiscursive web, authorized those ideological ends. There are reasons for caution: Balga Rodríguez's (Reference Balga Rodríguez2000–2003) linguistic study of costumbrist and bufo Bozal samples concludes that there is much variation, very little evidence of creolizing features, and a great deal of evidence that authors were deliberately creating comedic and derogatory effects.

Twentieth-century examples of Bozal include textual and oral metadiscourses and representations. We can note the consistency in how the register is characterized, even as elite scholarly metadiscourses have shifted their stance from early denigration of Black speech to today's “neutral” linguistic account of the degree of phonological and morphosyntactic deviation from unmarked colloquial Cuban speech. But what comes out of this web of distinct nodes of metalinguistic activity is a clear sense of a unitary register under the label “Bozal.” I have roughed out an enormously detailed story in tracing links between nineteenth-century literary “blackface” caricatures, twentieth-century folklore studies, and present-day uses of Bozal in popular religious ceremonies and linguistic analysis, in order to focus here on the broadest interdiscursive connections across these genres. That is because it is at this macro-level, I have argued, that Bozal is enregistered, enabling it to transcend the ebb and flow of popularity, and thus replication, of discourses instantiating it. For example, the relative decline of blackface vernacular entertainment has not prevented exaggerated visual and aural cues of blackface racial mockery from reappearing as authenticating signs of historical African presence in folklore and religious ritual performances.

The genealogy I have traced shows that both uptake and metacultural reevaluation are central to enregisterment, such that a robust and durable African slave persona is created and tightly linked to a recognizably “disfigured” way of speaking that unites all of the disparate discourses in the web I have described as instances of a unitary type called “Bozal.” This genealogy is an elaboration of what Judith Irvine describes as “triadic interdiscursivity,” where “discourse A is linked to discourse B, but also each of these is linked to (say) person P” (Reference Irvine2005: 78). For the case of Bozal, the figure of the African-born slave, rather than any specific set of linguistic features, can be envisioned as the central axis around which enregisterment of Bozal has occurred, as it is always implicated in Bozal voicings.