THE PALENQUERO LANGUAGE: REVITALIZATION, TEACHING, AND LEARNING PALENQUERO AS L2

Palenquero is an Afro-Iberian creole language spoken in San Basilio de Palenque, a village of around 4,000 residents in northern Colombia. There are also older speakers among the Palenquero diaspora in Cartagena and Barranquilla. The community was formed toward the middle of the seventeenth century (Navarrete, Reference Navarrete2008) by maroons (enslaved Africans who escaped bondage) from the Spanish port of Cartagena de Indias, being the sole survivor of what was once a large number of palenques “fortified communities” in Colombia. The Palenquero language apparently emerged around the turn of the eighteenth century, and well into the twentieth century all members of the community spoke Lengua natively (whether or not together with L1 Spanish is difficult to ascertain). Most of the Palenquero lexicon is closely cognate with or identical to local vernacular Spanish (Cásseres Estrada, Reference Cásseres Estrada2005), but Spanish and Palenquero are not mutually intelligible, although a few short phrases might be understood by Spanish speakers. Palenquero shares with Spanish basic head-initial phrase-structure patterns, including predominantly SVO word order, prepositions, and postnominal adjective placement (Moñino, Reference Moñino, Moñino and Schwegler2002, Reference Moñino, Maglia and Moñino2014; Schwegler, Reference Schwegler, Michaelis, Maurer, Haspelmath and Huber2013a, Reference Schwegler, Michaelis, Maurer, Haspelmath and Huber2013b; Schwegler & Green, Reference Schwegler, Green, Holm and Patrick2007 and the references therein). Like other creole languages, Palenquero has no grammatical gender (modifiers are derived from the Spanish masculine form), and verbs do not inflect for person and number but rather combine with preverbal tense/mood/aspect particles. Specific to Palenquero, negation is typically clause-final rather than preverbal, there are no preverbal object clitics as in Spanish, a single set of pronouns is used for all argument positions, and possession is expressed by postposing the possessor. Another striking departure from Spanish, and the focus of the present study, is the use of the prenominal pluralizing particle ma.

Beginning around the middle of the twentieth century, Palenqueros began to work outside of the community, and were mocked for their language. As a consequence, Palenquero families increasingly avoided using the language with their children, and warned them not to use Lengua outside of the community (Arrázola, Reference Arrázola1970; Hernández Cassiani et al., Reference Hernández Cassiani, Guerrero and Pérez Palomino2008). By the time that intense research began in the early 1980s, Palenquero had the characteristics of an endangered language and several scholars predicted its demise within a couple of generations (e.g., Moñino, Reference Moñino, Alès and Chiappino2003; Morton, Reference Morton2005; Pfleiderer, Reference Pfleiderer1998). Matters began to change with increasing visits by scholars and journalists, the UNESCO recognition of Palenque in 2005 as a part of the Intangible Cultural Heritage of Humanity, and the creation of an ethno-education program designed to teach the Palenquero language in the community’s schools.

The sociolinguistic dynamic is quite complex in a village that can be traversed in less than 15 minutes, and spoken Palenquero is not heard as commonly as in previous generations, but ethnic pride and the interest shown by tourists and scholars has contributed to the increased use of the language by many middle-aged and elderly residents who may once have been reluctant to use Palenquero. However, only a few adolescent or younger residents live in households where the Palenquero language is actively used, and even fewer are routinely addressed in the language. Most young people know at least a handful of the most common lexical items, many of which make their way into the local Spanish vernacular (e.g., moná “child” with Spanish-like gender-inflected diminutives monasito “boy” and monasita “girl,” oriki “dance”, and kombilesa “friend”). The popular hip-hop group Kombilesa Mi (“my friend” in Palenquero) performs rap music in Palenquero, and a few of their lyrics have spread among the younger generation. However, except for the small number of young heritage speakers (typically living in households with a Palenquero-speaking grandmother or great-grandmother), the Palenquero input that young people receive comes almost exclusively from school classes.

There are no externally validated proficiency measures for Palenquero, but studies of individual phenomena reveal considerable consistency of traits that set L2 Palenquero speakers apart from fluent native speakers. This includes use of preverbal particles (Lipski, Reference Lipski2020b), Spanish-like feminine gender agreement (Lipski, Reference Lipski2015, Reference Lipski2018b), placement and processing of negative elements (Lipski, Reference Lipski and Colina2010, Reference Lipski2017, Reference Lipski2018a), use of the pluralizer ma (Lipski, Reference Lipski, File-Muriel and Orozco2012a, Reference Lipski2014), and overall performance with interviews and translation tasks (Lipski, Reference Lipski2020a, Reference Lipski2020b). These observations have proven to be more reliable than self-rating or teacher judgments, which nonetheless cannot be completely dismissed. All these data informed the selection of participants in the present study.

Despite considerable enthusiasm, at present (early 2020), Palenquero language instruction—beginning in preschool and continuing through secondary school—occurs only once or twice a week for no more than an hour or two, and is mostly limited to teaching individual words and emblematic phrases. Only beginning in the eleventh grade does the teacher at times speak to students in Palenquero, but there are no grammatical explanations, contrasts with Spanish, or student production of Lengua. The current twelfth-grade teacher (who has only recently returned to teaching Palenquero) does engage the students in interactive exercises and offers some feedback on student performance, although still without any explanation of structures or contrasts with Spanish (Gómez Rodríguez, Reference Gómez Rodríguez2017; also Herrera Llorente & Álvarez Romero, Reference Herrera Llorente and Álvarez Romero2013). Although adult residents of Palenque generally support the language-revitalization efforts, fluent adult speakers almost never use the Palenquero language with young learners, and it is not uncommon to hear discouraging and even disparaging comments on young learners’ purported lack of competence in the creole. This stands in contrast to the enthusiasm with which many residents teach fragments of the language to visitors. Other adults, when queried about the prospects for survival of the Palenquero language, have expressed optimism because young people “have it in their blood” due to their Palenquero heritage. This viewpoint—perhaps even implicitly held by some teachers—embodies the mistaken assumption that the language can be acquired effortlessly just by living in Palenque (the “genetic fallacy” as described by Dauenhauer & Dauenhauer, Reference Dauenhauer, Dauenhauer, Grenoble and Whaley1998, p. 84), although most young people have few opportunities to effectively process or produce the language. Few Palenquero language students take the initiative to practice their language skills with fluent speakers, even at moments when the latter are speaking Palenquero, due to some combination of inhibition, timidity, indifference, or simply unwillingness to leave their own linguistic comfort zone. The result is the seemingly paradoxical existence of two hived-off versions of the Palenquero language within the same community, with only the Palenquero language teachers in a sociolinguistic position to use the language with both groups (but not serving as a linguistic conduit between them). The combination of no explicit grammar instruction, corrective feedback during instruction, or when attempting to use the language with fluent speakers, or meaningful two-way communication with fluent Palenquero speakers creates a “perfect storm” for unchecked miscues and unanticipated divergences from the target language. One such divergence is found in the use of the Palenquero plural marker.

PLURAL MARKING IN PALENQUERO: L1 AND L2 USAGE

Nominal plural marking in Palenquero is normally effected with the prenominal particle ma (possibly of Bantu origin: Moñino, Reference Moñino, Mihatsch and Sokol2007, Reference Moñino2013; Schwegler, Reference Schwegler, Mihatsch and Sokol2007a, Reference Schwegler, Baptista and Guéron2007b, Reference Schwegler, Maglia and Schwegler2012), and only when plurality is distinctive. Although Spanish affixes the nominal plural suffix /-s/ to all elements of a plural noun phrase, including determiners, adjectives, and quantifiers, in Palenquero only a single instance of ma marks nominal plural, as in (1), produced by fluent adult speakers.

Although there is some variation in the use of ma with plural items, particularly when generic reference is intended, fluent Palenquero speakers never use ma unambiguously with singular noun phrases (cf. Cassiani Obeso, Reference Cassiani Obeso2019a, Reference Cassiani Obeso2019b). Young L2 Palenquero speakers, however, routinely use ma with unmistakably singular referents, as well as (usually) with plural referents, which makes it impossible to determine from production data alone whether they recognize a pluralizing function for ma. The hyper-extension of Palenquero ma by L2 speakers has been verified in several production tasks, including picture naming (Lipski, Reference Lipski2015; also encountered by Deibel, Reference Deibel2020; Dussias et al., Reference Dussias, Gullifer, Poepsel, Schwegler, McWhorter and Ströbel2016), written assignments (Lipski Reference Lipski2014, Reference Lipski2020a, Reference Lipski2020b; Pérez Miranda, Reference Pérez Miranda2011; cf. Schwegler, Reference Schwegler, Haboud and Ostler2011), rapid-translation Spanish >> Palenquero (Lipski, Reference Lipski2020a, Reference Lipski2020b), and sociolinguistic interviews (Lipski, Reference Lipski2014, Reference Lipski2020a, Reference Lipski2020b), exemplified in (2)–(5), where nonplural ma is glossed simply as “ma.” Apparent underprocessing of ma emerges in Palenquero >> Spanish translation tasks, as in (6).

The oral and written production data show that L2 Palenquero speakers are aware of the existence of ma, but given the numerous instances in production of ma with unquestionably singular reference, such corpus data are not sufficient to determine the extent to which L2 speakers recognize ma as a plural marker, with overgeneralization to other contexts. At the same time, the disparities in Palenquero-to-Spanish translations exemplified in (6) suggest that when processing Palenquero, L2 speakers may rely heavily on contexts likely to involve plural referents, with ma having comparatively low cue strength. In (6a) for example, a single farm could plausibly be the referent. In (6b), because Palenquero, like Spanish, pluralizes the word for “father” to mean “parents,” a failure to process ma as plural would result in the plausible reading “my father’s house.” In (6c), however, a singular reading for the first noun phrases (“a child” instead of “children”) is pragmatically unlikely in the context of discussing community-wide language loss, while the presence of suto “we” before the second plural noun phrase renders the plural ma redundant, as does the presence of ané “they” in (6d). Such contextual cues may also aid fluent Palenquero speakers in processing, but only as a redundant backup, given that they do not produce examples of ma with unmistakably singular referents. If Palenquero ma is in fact an exclusively plural marker, then encountering ma during processing should prompt anticipation of a following plural noun phrase, even in the absence of facilitating contextual clues. This provides a criterion for empirically examining the online processing of ma by both L1 and L2 Palenquero speakers.

EXAMINING ONLINE PROCESSING: AN EYE-TRACKING EXPERIMENT

In this study, online processing of Palenquero ma was examined utilizing eye-tracking methodology, a technique that is eminently suitable for the Palenque field environment, and that offers the advantage of excellent temporal resolution combined with a high signal-to-noise ratio (i.e., usable with a single presentation of stimulus items, unlike, e.g., electrophysiological techniques such as EEG/ERP). Moreover, eye-tracking can be carried out with highly portable equipment suitable for even the most challenging field settings, and can be employed with naïve and nonliterate participants as well as with languages that have no established written tradition. These features are crucial in working with the participants in the present study

The eye-tracking experiment that targeted processing of ma utilized the visual world paradigm (Huettig et al., Reference Huettig, Rommers and Meyer2011; Tanenhaus & Trueswell, Reference Tanenhaus, Trueswell, Traxler and Gernsbacher2006), in which participants’ gazes are drawn to specific on-screen images in response to auditorily presented stimuli. For each trial, participants saw two images on the screen while listening to a Palenquero sentence “Find {ma} […]” with the target word inserted. The Palenquero plural marker ma was present in all trials in which at least one of the images would evoke a plural response. The fundamental goal was to determine whether hearing the plural marker ma would prompt participants to shift their gaze to a plural image even before hearing the target word. This approach is similar to that used to test the effects of gender marking on determiners (Dussias et al., Reference Dussias, Kroff, Guzzardo Tamargo and Gerfen2013; Lew-Williams & Fernald, Reference Lew-Williams and Fernald2007, Reference Lew-Williams and Fernald2010; Morales et al., Reference Morales, Paolieri, Dussias, Kroff, Gerfen and Bajo2016), in which gender-marked articles triggered participants’ gaze to the appropriately gendered image. The hypothesis was that when given a choice between a singular distractor image and a plural target image, adult speakers’ gaze would begin to shift toward the plural target upon hearing the plural marker ma, that is even before the name of the object. In addition, if presented with a plural target image together with a singular image of an item beginning in ma- (e.g., machete “machete,” marímbula “thumb piano”) or the opposite configuration (singular target in ma- with plural distractor image), adult speakers would experience momentary indecision, with slower gaze fixation to the target than in nonambiguous plural target versus singular distractor cases. However, the time course of gazes to the target image in combinations of singular target + singular distractor, plural target + plural distractor, and singular target + plural distractor should be similar. If younger L2 Palenquero speakers do not acknowledge ma as a plural marker, then the presence of ma should not prompt anticipatory gazes toward a plural target image in the plural target versus singular distractor condition, and singular nouns in ma-, whether as target or distractor, should behave no differently than other items.

RESULTS AND DISCUSSION

As a group, the younger L2 Palenquero speakers’ gaze patterns were more erratic than those produced by fluent Palenquero speakers. Although this observation cannot be quantified, the composite trial-by-trial gaze plots and the gaze-tracking videos produced by Tobii Pro Lab revealed more prestimulus eye movement away from the center of the screen and “exploring” of the images prior to hearing the stimulus items. Whether this is reflective of younger participants’ level of confidence in dealing with Palenquero or simply another manifestation of the youthful fidgeting observed during other tasks cannot be determined from the data. The rate of correct responses (i.e., the gaze ultimately settling on the appropriate image) was slightly lower (92.9%) for younger L2 Palenquero speakers than for fluent adult Palenquero bilinguals (99.4%), but in general participants followed the instructions and correctly identified the target images. Table 1 gives the times at which gazes to the target images became statistically significant.

TABLE 1. Time past carrier phrase offset at which gaze to target image became significant (including 200 ms minimum saccade time)

The data in Table 1 show that adults’ gaze toward a plural target occurred more quickly when the distractor was singular rather than plural, reflecting the advantage provided by the plural marker ma: Welch-t(71.1) = 3.27, p = .002. Younger L2 Palenquero speakers showed no such advantage, and in fact on average took longer with singular distractors than with plural distractors. Also evident is the significantly longer reaction time for adult participants when choosing between a plural target and a singular distractor beginning in ma-, which initially caused confusion with the Palenquero plural marker ma: Welch-t(37.9) = 2.845; p = .007. There was no such distinction for younger L2 Palenquero speakers. Figure 1 illustrates these differences, including the extended vacillation between target and distractor for adult Palenquero speakers. The left dashed line indicates the offset of the stimulus word, and the right dashed line marks the 200 ms minimal latency for significant language-triggered saccades.

FIGURE 1. Adult (left) and younger L2 (right) gaze plots, plural target with singular distractor in ma-.

For adult speakers, gazes to singular targets beginning with ma- combined with plural distractors also took significantly longer than gazes to other singular targets with plural distractors, indicating the momentary confusion caused by a word whose initial syllable is identical to the Palenquero plural marker: Welch-t(22.9) = 3.65, p = .001. Once more, no such difference was observed for younger L2 Palenquero speakers. Figure 2 displays the behavior of this configuration.

FIGURE 2. Adult (left) and younger L2 (right) gaze plots, singular target in ma- with plural distractor.

Both fluent adult Palenquero speakers and younger L2 participants took longer to shift their gaze to plural targets than to singular targets, presumably because although the screen contained a single image in each upper corner, there was more than one object in the plural targets.

GENERAL DISCUSSION

The eye-tracking data obtained from fluent Palenquero speakers and younger second-language learners indicate that many L2 speakers do not process ma as a plural marker, while fluent Palenquero speakers do respond to ma as an indication of an upcoming plural noun phrase. This is consistent with the previously cited production data, in which L2 Palenquero speakers routinely combined ma with indisputably singular nouns.

The distinction between singular and plural is fundamental to all languages, and at first glance, it is surprising that even in the absence of explicit instruction, L2 Palenquero speakers have not fully acquired the morphologically transparent and syntactically simple Palenquero pluralization with ma. A number of factors converge to render this outcome more likely. Clearly relevant is the aforementioned lack of grammatical instruction in Palenquero languages classes, coupled with the fact that most L2 Palenquero speakers rarely use the language with fluent adult bilinguals. Naturalistic second-language acquisition does not normally involve corrective feedback, but to be effective there needs to be more robust input than what L2 Palenquero learners receive outside of the limited classroom exposure. The combination of minimal naturalistic input and neither explicit grammar instruction nor error correction in school further entrenches noncanonical use of ma. L2 Palenquero speakers are aware that ma is a uniquely Palenquero item associated with nouns, but may regard it as simply an optional element whose presence adds authenticity to attempts at producing Palenquero. As a possible analogy, adult Palenquero speakers often detach the imperfective suffix –ba from verbal roots and freely reattach –ba to a variety of other grammatical forms without changing meaning (Lipski, Reference Lipski2012b, Reference Lipski2016), even extending this usage as a verbal disguise known as retahila (Moñino, Reference Moñino, Mihatsch and Sokol2007, pp. 53–54), and L2 speakers may have mistakenly assumed that ma could “float” in a similar fashion.Footnote ²

L1 VARIATION AS L2 MISINTERPRETATION

In many ways, L2 Palenquero speakers’ overextension of ma is similar to the trajectories followed by children who acquire variable phenomena in their native languages, as well as to the learning of artificial languages (Hendricks et al., Reference Hendricks, Miller and Jackson2018; Hudson Kam & Newport, Reference Hudson Kam and Newport2005, Reference Hudson Kam and Newport2009; Samara et al., Reference Samara, Smith, Brown and Wonnacott2017; Wonnacott et al., Reference Wonnacott, Brown and Nation2017).Footnote ⁵ When faced with variable structures in a naturalistic setting, learners frequently attempt to regularize the data, en route to acquiring nativelike probabilistic distributions. This is particularly true when the variability is not constrained by overt sociolinguistic factors such as speaker gender, class, age, and so forth, and may entail overextension of one of the variants. Children more frequently gravitate toward regularization than adult L2 learners (Hudson Kam & Newport, Reference Hudson Kam and Newport2005), although with higher levels of dispersion, adults may also regularize the variable input (Hudson Kam & Newport, Reference Hudson Kam and Newport2009). In Palenque, L2 learners of Palenquero straddle the child-adult categories because some exposure to the language occurs during childhood (e.g., in the largely symbolic preschool and primary school “instruction”), with more intense reinforcement during later adolescence.

Strictly speaking, adult usage of Palenquero ma does not exhibit the type of probabilistic variability that typically results in regularization by L2 learners because ma is used unambiguously by fluent speakers only to express plural, and from the speakers’ standpoint is never truly optional or redundant. However, the fact that ma is not usually present when numbers or other quantifiers signal plurality, nor is it frequent in generic usage (which in Spanish typically involves a plural noun phrase), especially in nonsubject positions, may create the impression of true probabilistic variation. This semantically grounded variation has not been incorporated into L2 Palenquero learners’ use of ma. A similar regularization can be observed in L2 Palenquero speakers’ use of preverbal tense/mood/aspect particles (Lipski, Reference Lipski2020b), which are used consistently and with clearly differentiated meaning by fluent adult speakers, but whose semantic nuances have not been fully acquired by L2 learners, who have overgeneralized configurations not found in fluently spoken Palenquero. In both cases, the lack of corrective feedback during explicit instruction, or frequent communication in Palenquero with fluent adult speakers has resulted in a microlinguistic “bubble” in which these originally inaccurate regularizations can prosper.

Research conducted with artificial languages has produced results similar to those found in child language acquisition, demonstrating that distributional patterns alone can lead to the acquisition of differing grammatical categories (Reeder et al., Reference Reeder, Newport and Aslin2013, Reference Reeder, Newport and Aslin2017). Younger L2 Palenquero speakers, exposed to Palenquero but not receiving any metalinguistic information in language classes nor sufficiently robust input outside of school settings behave in many ways like learners of an artificial language, and may have mapped the Palenquero plural marker ma onto the structural scaffolding of Spanish definite articles, as a more inductive choice than acknowledging a familiar concept (plural morpheme) in an unexpected position (prenominal).

Not all examples of ma produced by L2 Palenquero speakers fit the structural distribution of Spanish definite articles, whether or not ma is used with plural nouns. In examples such as (7), all produced spontaneously by L2 Palenquero speakers, no article or other determiner would be permitted in Spanish in the position occupied by ma.

While not the most common L2 Palenquero usage, the residue of utterances such as (7) is consistent with the (mis-)interpretation of ma as a variably occurring prenominal marker, whether or not occupying the same position as Spanish articles.