The dialect of São João da Chapada: Possible remains of a mining language in Minas Gerais, Brazil

3.1 Data selection

The lexical database in this study is based on 149 ^[4] different words or expressions presented in the “Vocabulary for the Creole of São João” (Machado Filho 1943: 117–133). The book also includes a second list (Machado Filho 1943: 134–138), with words and expressions from other regions of Brazil that were not considered in the present study. Limiting the analysis to the vocabulary from the district of São João da Chapada makes it possible to relate local demographic, historical, and linguistic data to shed light on the origins of the local African components. This limitation is also motivated by the interest in the social functions of this regional lexicon, which can be related to the origins of its population, the sociocultural context, and the mining activities in the area.

The study focuses on the distribution of the lexical items in semantic domains and word classes, as well as on their etymologies. A similar analysis of an Afro-Brazilian variety called Cupópia was performed by Álvarez López and Jon-And (2017), inspired by the methodological approach adopted in the articles included in a volume edited by Bartens and Baker (2012) about transplanted European languages, which present data from 17 creole languages and contact varieties involving European and African languages.

3.2 Semantic domains

In order to discuss the classification of this variety as “mining language”, the analysis includes a comparison with other mining languages, as well as with various Afro-Brazilian varieties that may share part of their lexicon with this one. This includes the identification of specific words that can be associated to mining activities as opposed to the lexical items identified as part of a basic vocabulary based on Swadesh’s (1971) classical list of basic concepts for comparative linguistic studies.

Since Baker (2012: 275–277) acknowledged difficulties classifying lexical items into the 12 semantic fields that he suggested, an alternative classification of 8 semantic domains is proposed. The semantic domains are: universe and creation (animals, nature, etc.), person (people, body, etc.), language and thought (emotions, communication, etc.), social behavior (religion, relationships, etc.), daily life (home, food, etc.), work and occupation (work activities, tools), physical actions (posture, movements, etc.), and states (quantity, quality, time, etc.). These domains are clearly specified with subdomains in Fieldworks Language Explorer software used to organize the lexical database (http://fieldworks.sil.org/flex/). Such classification left no “unclassified” items and makes it possible to compare the results with those of an analysis based on the same methodology regarding another Afro-Brazilian variety (Álvarez López and Jon-And 2017).

3.3 Etymologies

The methodology for the etymological study aims at verifying the etyma proposed by Machado Filho (1943) and Simões (2014), and focuses on the relation between linguistic and demographic data. It also addresses the conclusions about sociocultural context that can be drawn from the analysis of the distribution of the lexicon in different semantic domains and word classes.

Machado Filho (1943) does not systematically mention the origin of the words listed in his study. Simões, who gathered data in the same region in 2011 for his Master’s dissertation, compared his findings with those presented in Machado Filho, his primary source. Simões (2014) findings show that most of the words documented by Filho had also been registered in other Afro-Brazilian communities and are of Bantu origin. No lexical items on the list could be traced back to West Africa and Peixoto’s work. Therefore, Simões (2014) limited his etymological analysis to the use of three dictionaries of Umbundu, three of Kimbundu, and one of Kimbundu and Kikongo, the languages that are usually mentioned as the sources of Bantu-derived lexical items in Brazil.

The contribution of the present etymological analysis will be to increase the number of dictionaries and secondary sources in order to verify the reliability of Machado’s work by confirming the existence of the words, to verify their meanings in other sources, and to corroborate the etymologies proposed by both Machado Filho (1943) and Simões (2014). The additional sources for Africanisms in Brazil are Angenot et al. (2013) and Álvarez López (2004). Angenot et al. (2013) is a glossary with 3,704 entries, and it is limited to Bantu-derived words found in 112 different studies. Álvarez López (2004) includes a glossary of approximately 700 African-derived words and expressions, both Bantu and non-Bantu, gathered in the state of Bahia, but is not one of the sources included in Angenot et al. (2013). The additional dictionaries are Bentley (1895, Kikongo), Johnston (1919, comparative Bantu), Laman (1936, Kikongo), and da Matta (1893, Kimbundu). These are early sources for African languages, published at the end of the nineteenth century or the beginning of the twentieth century, which is important, given that we are discussing African-derived words introduced between 1730 and 1930.

Unfortunately, there are still few secondary sources with diachronic data relating to the population of the relevant region around Diamantina; nevertheless, given that Simões (2014) mainly relies on data from the entire region of Minas Gerais, the aim is to add relevant available historical and demographic information from the municipality of Diamantina.

3.4 Word classes

Regarding the analysis of the distribution of the lexicon in word classes, the aim of this aspect of the study is to offer information about the origins of this variety. Earlier studies suggest that nouns are more easily borrowed and exchanged than terms in other word classes, and that they usually predominate among the African-derived words in contact varieties similar to the one in focus (Bartens and Baker 2012). Moreover, the etymologies of lexical items other than nouns can reveal which languages were involved during the period of formation of a linguistic variety, and such results can be related to demographic data from the region. The word classes distinguished in this study are: nouns, verbs, adjective/adverbs, pronouns, and other (one interjection for which it was impossible to identify the etymon). Some of the words may pertain to more than one word class, for example, in cases in which a word is used both as a noun and as a verb.

We also know that nouns tend to be borrowed particularly when new or unknown objects and concepts need to be named (Thomason and Kaufman 1988: 77). In this case, basic vocabulary should be less likely to be borrowed than more culture-specific African vocabulary or specific words and expressions that may be associated with mining activities.

4.1 Distribution of words in semantic domains

Figure 1 shows that 76% (113/149) of the words belong to the domains that have to do with universe and creation (nature and animals), daily life (home and food), social behavior (social activities and relationships), and work and occupation. This means that this vocabulary is not limited to mining terms.

Figure 1:

Distribution of the lexicon in semantic domains.

4.2 Basic vocabulary

With respect to the list of 100 basic vocabulary items compiled by Swadesh (1971), 16% (24/149) of the words can be classified as basic vocabulary, that is, not culture-specific words that exist in individual languages:

Although the basic vocabulary is normally not borrowed lexicon, convincing African etymologies have been identified for 22 of the 24 items that belong to the core vocabulary. These lexical items are distributed in seven of the eight semantic domains: universe and creation, person, social behavior, daily life, work and occupation, physical actions, and states. For acrepú, the etymology was not identified, and for maravir the etymology given by Simões (2014: 123) was not considered convincing (oluali, olwili, meaning Planet Earth). All these lexical items (or items with similar forms and meanings) have been attested in other Afro-Brazilian communities.

4.3 Specific mining vocabulary

The words that may be related to specific mining activities show similar results if compared to the basic vocabulary: a total of 18% (27/149), among which 16 have been identified as being of possible African origin. This result is not surprising as the expertise of Africans was important in the mines, and some mining techniques, such as the use of wooden mining bowls, may have been of West African origin (Boxer 1964: 184; Costa 2007: 252). Available descriptions of mining activities in Minas Gerais during the eighteenth and nineteenth century were consulted in order to identify the vocabulary that could be associated with such activities as panning for gold and diamonds, for example:

The original process of washing and panning for gold was of the simplest. The sole instrument required was the bateia, a large swallow, cone-shaped basin,^[5] made of wood or metal, which the miner held with both hands. The gravelly subsoil mixed with quartz (cascalho) which contained the gold particles was placed in the bateia with a quantity of water just sufficient to cover it. […] The cascalho was not usually on the surface of the soil, but at varying depths below a layer of sand, earth or clay. (Boxer 1964: 38)

On the heap of cascalhão, at equal distances, are placed three high chairs for the officers or overseers. After they are seated, the negroes enter the troughs, each provided with a rake of a peculiar form and short handle with which he rakes into the trough about fifty or eighty pounds weight of cascalhão. The water being then let in upon it, the cascalhão is spread abroad and continually raked up to the head of the trough, so as to be kept in constant motion. […] When a negro finds one [diamond], he immediately stands upright and claps his hands, then extends them holding the gem between his fore-finger and thumb; an overseer receives it from him, and deposits it in a gamella or bowl suspended from the center of the structure, half full of water. (Mawe 1812: 223–224)

There are also descriptions of mining activities and the vocabulary used to describe it in dictionaries elsewhere. Two examples are given here from the U.S.:

There are two ways to mine – by hand or by machine. The old hardrock miners (now nearly extinct) were either single jackers or double jackers. The first held his own steel in one hand and hammered with the other. In double jacking, one man held the jack while the other hammered. (Moore 1926: 87)

[…] mud, the call for a driller to indicate that the water poured into the drill hole has now formed mud and that the hammering on the drill should stop […]. (Davidson 1929: 146)

The following terms were identified as being related to mining activities:

The mining vocabulary is found in two semantic domains: “work and occupation” and “universe and creation”. The lexical items ‘gold’, ‘diamond’, ‘black mud’, ‘earth’, ‘gravel’, and ‘dust’ were classified as universe/creation in Figure 1. With the addition of these words, the domain of work/occupation represents almost a fifth (18%) of the total vocabulary (and the domain of universe/creation decreases to 21%).

Convincing etymologies were not found in the consulted dictionaries for the following items: arapossé-arapossí, jambá, jombô, maravir, ongá, orongoia, oropungo, orovanga, orumbá, ovicaiá, rubudú. Of these, jambá is the only word that has been attested in other sources; Vogt and Fry (1996) registered this item in the nearby community of Milho Verde with the meaning of ‘diamond’, and Simões (2014) proposes a Bantu etymology based on the word ‘elephant’. However, the etymon was discarded in this study since its relation to gold was not explained (maybe through the value of ivory as compared to gold). This means that convincing possible etyma have not yet been identified for 11 of the words related to mining activities.

Unexpectedly, while 16 items of this group were also found in other Afro-Brazilian communities with similar or related meanings; 11 items were not found in other sources with the same meaning (arapossé-arapossí, arengá, barundo, comboêro, jombô, ndimba, onjundo, oropungo, orumbá, ovicaiá, rubudú), or were only attested in nearby communities by Simões (2014), indicating that their use might be restricted to a few communities in the mining district. This indication is reinforced by a study of the vocabulary of small-scale diamond mining in the adjacent Brazilian state of Goiás, which does not include any of the items registered by Machado Filho (1943) in São João da Chapada (Jerônimo 2014).

4.4 The African-derived lexicon

Among the 17 languages examined in Bartens and Baker (2012), the average proportion of nouns in the African-derived lexicon is 74%. The results of the present study for the words of proposed African origin are: 85% (105/124) nouns, 8% (10/124) verbs, 7% (9/124) adjectives and adverbs, and 2% (2/124) pronouns.

Table 1 shows that vocabulary of shared possible origin represents 61% of all African-derived nouns, 80% of the African-derived verbs, and 33.5% of the African-derived adjectives. As 55% of adjectives and adverbs, and both pronouns, were found only in Umbundu dictionaries, and the etymologies reported as shared include Umbundu, this language stands out, although less than 50% of the verbs and nouns were identified only as Umbundu.

The relatively high proportions of Umbundu etyma among adjectives and adverbs together with the two pronouns indicate that this language may have played a more dominant role than other African languages when those languages were still spoken in the region. Baker’s hypothesis is that groups with higher demographic weight in the earliest periods of colonization contribute more words in general and more lexical items that belong to word classes other than noun in particular (Baker 2012: 283). In sum: among the words of African origins there are mostly nouns, as expected, and Umbundu stands out in all word classes.

Table 2 shows that the African-derived vocabulary is distributed among different semantic domains and is not restricted to specific activities associated with African cultures or slavery. The results are similar to the numbers in Figure 1 for the whole vocabulary: a majority of the lexical items are in the domain of universe/creation, and the lowest number is in language/thought.

Summing up, 17% of all the words were identified as core lexicon, and 18% as being related to work/occupation or mining activities. At the same time, almost 50% of the African-derived lexical items belong to the semantic domains universe/creation and domestic/daily life.

4.5 Etymologies and demography

With the gold rush, enslaved Africans replaced indigenous labor force at the end of the seventeenth century, among other reasons, because Africans had experience with gold extraction. Consequently, Minas Gerais became the region with the largest contingent of slaves in Brazil (Costa 2007: 252).

Apart from Africans and their descendants, “many ‘poor whites’ engaged in prospecting and mining on their own”, and there were also numerous Indigenous slaves (Boxer 1964: 167, 184). In the eighteenth century, most immigrants from Portugal living in Minas were young unmarried men from the regions of Minho and Douro (Boxer 1964: 163–164). This is probably why the vocabulary includes words of African origin, but also words with no relation to Africa.

According to Costa (2007: 252), there were 120,000 inhabitants in Minas Gerais in 1720. Slaves from West Africa were the most numerous between 1714 and 1740, “closely followed by slaves from Angola [probably Kimbundu speakers] and Benguela [probably Umbundu speakers]” (Boxer 1964: 175). Although West Africans were the majority in Minas Gerais during the eighteenth century, Brazilian-born slaves and speakers of Bantu languages from West Central Africa dominated in the nineteenth century (Luna 1980). Unfortunately, we lack detailed historical studies about the population of the municipality of Diamantina, but for the region of Minas Gerais, this means that speakers of Umbundu were present, as well as speakers of other Bantu languages. In 1767, there were 126,603 slaves in the captaincy, a number that represented 60.7% of its total population (Resende 2006: no page).

From a more local perspective, we know that in 1830, slaves represented 36% of the total population in the traditional mining centers in Minas Gerais (Luna and Klein 2004: 10), and in Diamantina there were as many as 7,800 slaves out of 15,200 inhabitants in 1853 (Costa 1998: 104). This means that half of the population was still composed of slaves three years after the legal slave trade had ended in 1850. The death rate of slaves was high, and their reproduction rate low (Boxer 1964: 175), probably because they lived under unfavorable conditions imposed by the slave system in the mining region. Furthermore, in Diamantina, African-born slaves predominated longer than elsewhere in Minas due to the activities of the royal diamond monopoly (Bergad 1999: 129).

In rural communities in this district, such as São João da Chapada, is it likely that newly imported African slaves who spoke Bantu languages had only limited access to Portuguese spoken as a first language unless they were house slaves. After the abolition of slavery in 1888, some of the former mining slaves probably ended up living in communities similar to the six maroon communities (called quilombos, a term with Bantu origins) situated around São João da Chapada and described by Machado Filho (1943: 54–56). At that time, a class society in which Africans and their descendants were marginalized was established – especially since there were no measures taken to integrate former slaves into Brazilian society (Álvarez López 2004: 12–13). As a result of those colonial and postcolonial social structures, there are today approximately 3,000 more or less marginalized communities throughout Brazil (many of them in Minas Gerais) that can be classified as maroon, or quilombolas, in accordance with the Presidential Decree 4887 from November 20, 2003. However, runaway slaves did not found most such communities, as many of them emerged only after 1888, when freed slaves did not have anywhere to go. Several Afro-Brazilian varieties have emerged in similar communities (see, for example, Lucchesi et al. 2009; Vogt and Fry 1996).

4.6 Comparing with mining languages and other Afro-Brazilian varieties

As has been the case in researching other mining communities in different parts of the world, in Minas Gerais we are dealing with work songs and an innovative, specialized vocabulary related to small-scale mining technologies in multilingual, male-dominated communities where people of various origins gather to exploit minerals (see Muysken this volume). As expected, based on research on mining languages, the vocabulary relies on the languages originally used by members of the work force.

However, if we look at Afro-Brazilian multilingual and male-dominated settings where mining is not part of everyday life, we will also find slaves, work songs, and small-scale technologies in everyday life. Sugar-cane plantations may serve as an example in this regard. At the same time, Afro-Brazilian communities in the State of Minas Gerais have been studied because of their use of “special languages” with vocabularies similar to the one registered by Machado Filho: Língua do Negro da Costa in Tabatinga (175 words, Queiroz 1998) and Calunga in Patrocínio (305 words, Byrd [2012]). These varieties are used as secret in-group codes in which the specific vocabulary replaces part of the lexicon in in-group conversations in Portuguese.

In the case of São João da Chapada, Machado Filho (1943) affirms that a “creole language” was spoken, but his data does not provide more than a number of songs and a glossary. In light of the research that has been accomplished in pidgin and creoles studies from the 1980s onward, this glossary falls short of the linguistic evidence necessary for proving the existence of a creole language. His statement can be related to the debates on the origins of innovative features in Brazilian Portuguese, a discussion that has been largely marked by the polarization involving mainly two hypotheses. The first is the so-called “hypothesis of internal drift”, according to which changes attested in Brazil are solely derived from a natural evolution of European Portuguese (Naro and Scherre 2007). The second is the “contact hypothesis”, in which creolization and the role of African languages have received special attention, given that Africans and their descendants came to compose a significant part of the Brazilian population between the seventeenth and nineteenth centuries (Lucchesi et al. 2009). This issue is still the topic of an ongoing debate, but during the last 30 years a significant number of studies have supported the contact hypothesis, both in empirical studies on varieties defined as “Afro-Brazilian” and in emerging studies of Portuguese in Africa.

One of the studies on Afro-Brazilian varieties focuses on the community of Cafundó, situated in the neighboring state of São Paulo. The variety called Cupópia can be compared to the dialect of São João da Chapada, and has probably survived until today because of its prevailing function as a secret in-group code, or marker of social distance with out-groups, used in everyday life (160 words, see Álvarez López and Jon-And [2017], Vogt and Fry [1996]). The results of the lexical study of Cupópia are similar to those obtained in the analysis of the dialect of São João da Chapada. The basic vocabulary of Cupópia represents 28% (compared to 17% in São João da Chapada), and the African-derived words 71% (compared to 83%), and, out of these, there are 83% nouns (compared to 85%), 11% verbs (compared to 8%), 7% adjectives and adverbs (compared to 7%), and 1% other (one greeting compared to 2%, represented by two pronouns). Regarding the distribution in semantic domains, it is similar in the two vocabularies, as shown in Table 3; the differences are mainly that Cupópia has fewer words in the domains “universe and creation” and “social behavior” and more words classified under “person”.

For Cupópia, the distribution in semantic domains was tentatively explained by the fact that it has the function of a secret code used to signal distance between speakers and out-groups, and is used in everyday life by members of in-groups as an identity marker (Álvarez López and Jon-And 2017). They probably needed basic words to use Cupópia as a secret code in everyday life. This could also be the case for the dialect of São João da Chapada, since historical sources reveal that mining slaves may have needed such a code to perform daily activities.

Etymologically, the results show that approximately half of the nouns and verbs were of shared origin in both vocabularies, but whereas Kimbundu dominated in Cupópia for adjectives and adverbs, Umbundu dominated in São João da Chapada. In the case of Cupópia, the detailed study of the genealogies in the community by Robert Slenes (Vogt and Fry 1996: Ch. 2) showed a higher demographic weight historically for speakers of Kimbundu. Unfortunately, we still lack detailed information concerning the origins of the African population in São João da Chapada and its surroundings.

4.7 Secrecy and strategies of resistance in the mining region

In sociocultural and historical settings similar to the one described by Machado Filho, the political protest in work songs and the value of secret codes as strategies of resistance have been emphasized:

[…] unlike the songs of lumberjacks and other occupational groups in the Midwest, coal-mining songs represent mass protest; there is a growing passion for unionism that doesn’t show up in many of the songs of other occupational groups. Miners’ songs tell of greedy bosses, long hours, dangers of work, poor working conditions and disasters. In addition, miners lived with their families close to the shafts, so miners’ songs reveal a domestic life that doesn’t appear in songs of lumberjacks and other occupational groups. (Baker 2015: 794)

In Minas Gerais, the work songs, or vissungos, were performed to accompany different phases of the mining work: in the morning, during lunch break, in the afternoon, when the boss was arriving at the field, to pray the Lord’s Prayer, and to carry the dead to the closest cemetery (Machado Filho 1943: 61–62). There were also special songs addressed to “fine” individuals who arrived at the mining site, demanding that they be offered something.

Nevertheless, as society has changed, the social functions of the mining songs have also changed. Nowadays, the vissungos are not necessarily related to mining work, but they represent a cultural heritage, and sometimes resistance, in the same way as African-American work songs in the U.S. or coal-mining songs in Europe do. We also know that several vissungo masters were rivals; they had their own choirs and did not teach their songs to outsiders. According to Andrarde (2012: 338–339), vissungos can be understood as a secret language in the remains of maroon communities. They embody the transmission of knowledge through a whole set of expressive oral resources.

A concrete example of secrecy and resistance in this mining sociocultural context is the vissungo in which a son tells his father that he found a diamond, and the father tells him to hide it in the gravel and be quiet (Machado Filho 1943: 75). With the use of the specific vocabulary, the song becomes unintelligible for an overseer who speaks only Portuguese.

In addition, travelers who visited Brazil during the nineteenth century comment on the problem of slaves stealing gold and diamonds, maybe in order to buy their freedom. Mawe (1812: 223–224) affirms that “[m]any precautions are taken to prevent the negroes from embezzling diamonds”. Furthermore, in his study of the Golden Age of Brazil, Boxer (1964: 217) states: “the authorities were convinced that the mulatas and Negresses who sold food and drinks to the miners were chief intermediaries in the smuggling game”, which means that this did not only happen at an individual level, but in a more or less organized way. Mining slaves and their allies might therefore have needed a secret code in order to be able to keep some of the minerals that they worked so hard to find for someone else’s profit. Another illustrative example of organized resistance in the area is the Serro slave rebellion in 1864, in which more than four hundred mining slaves gathered, as well as slaves from the farms and the cities of Serro and Diamantina, in order to obtain their freedom with the support from maroon communities and black freedmen (Mota 2006).