Loanwords and Borrowing: Methods
Summary and Keywords
The study of loanwords, and of language contact more generally, is a useful tool in tracing encounters and exchanges between different communities in the past. Loanwords often come in sets related to specific semantic fields, illustrating the nature of exchanged goods and ideas, as well as the nature of contacts between those communities, for instance, economic exchanges or political dominance. Examples include the adoption of new crops and subsistence techniques, in both ancient and recent periods, and the strong Arabic influence in multiple domains on the Swahili language. Loanwords are but one outcome of language contact. More intense contacts can lead to structural borrowing; to convergence between nonaffiliated languages, resulting in linguistic areas; and to language shift. The languages of so-called pygmy hunter-gatherers are a notorious example of people abandoning their own language in favor of a new one.
To identify loanwords and to distinguish them from inherited vocabulary, it is necessary to apply the comparative linguistic method. Irregular sound correspondences and morphological traits, and a continuous distribution across linguistic boundaries are indicative of borrowing. The possibility of semantic analysis and the presence of cognates in related languages may confirm the identity of the donor language. The identification of loanwords suffers from a few drawbacks, however. Some sounds have not changed for centuries or even millennia, preventing the distinction between loans and inherited words. Or loanwords may have become integrated in the phoneme inventory of the recipient language, giving the impression of regular sound correspondences. But even if loans can be recognized as such, the donor language cannot always be traced. Finally, it must be said that the study of loanwords attains the best results when it is based on well-annotated data, with detailed semantic description and a list of regular sound correspondences and adequate classification at hand.
Languages are constantly shaped by the life trajectories of their speakers. The original base, already subjected to linguistic changes, is further altered by extralinguistic factors, such as sociocultural changes, technological innovations, and encounters with other speech communities. In the African context, the effects of contact with Europeans and Arabs are well known and well documented. But contact is an intrinsic part of history. Events of initial colonization, meaning migrations into unoccupied or sparsely occupied lands, by definition has been limited in number. In most cases, migrations or trade routes lead to contact situations, each translated differently in language. Linguistic contact thus also played a role some centuries and even millennia ago. The study of loanwords, and of language contact more generally, is highly informative of the history of the respective speakers.
History through the Lens of Loanwords and Language Contact
Several periods involving human contact in Africa have received attention from historians and linguists.1 During the last few centuries, contacts with Europeans have had far-reaching consequences in both the material and immaterial culture of the continent. African languages contain many loanwords related to European introductions, whether new political structures, Christianity, technologies, or tools. European traders and colonizers often acted as intermediaries, shipping goods from the Old World to the New World, and vice versa. These exchanges also involved crops that in the 21st century are integral elements in many cuisines around the world. American tomatoes, maize, cassava, sweet potatoes, chilies, and various types of fruit can be found all over sub-Saharan Africa. The names for these crops often indicate their recent arrival on the continent and sometimes reveal their exact origins in the Americas or the itineraries of dispersal through the African interior, or both. Names for cassava in Kongo languages, for instance, have clear Brazilian origins, specifically, Tupinambá, with Portuguese acting as intermediary:
Tupinambá mani-oca ‘cassava’ > Portuguese mandioca > northern Kongo mayáka > southern Kongo madyóko
The distribution of the two Kongo loans suggests that the word was borrowed twice, independently from the other. Each instance of borrowing followed a different trade route into the hinterland.2 A similar north-south pattern can be observed in the introduction of other crops of American origin in this same region, with loanwords mostly being used in the south. Because the south had been home to the Kongo Kingdom, the higher use of loanwords in the south probably owes to the more intense contacts with the Portuguese.3 In many cases, however, coastal peoples gave local names to the new crops, blurring their true origins. Still, the new African names traveled farther and were borrowed by communities living far into the interior of the continent. Tracking these loans illuminates the trade routes along which the crops were dispersed from the coasts to the interior. Many fine examples can be cited from the northeastern bend of the Congo River, which was a crossroads of multiple trade routes. The So, for instance, use a Kongo word for “peanuts,” ngúbà, indicating that the crop and its name were brought up the Congo River. Lokele and Foma peoples, who live farther east, apply a Swahili word to the same crop: kàlángà. In contrast to their neighbors, the Lokele and Foma acquired peanuts through trade with Swahili caravans traveling from the eastern coast of the continent.4
The import and impact of American crops in Africa is impressive. The reverse trajectory was less pervasive. Still, several loanwords of African origins can be found on American and even European shores. In English colonies, for instance, sesame was known as bene, a word of West African origins, in, for example, Manding and Bambara. But borrowing especially occurred from languages of West-Central Africa. The word for “sesame” in Jamaican English, Gullah, and the French-based creole from French Guiana is wangila, as it is in Kimbundu and East-Kongo languages. The famous Louisiana stew gumbo takes its name from the word for okra in West-Central African languages such as Yaka and Nsong. And peanuts, a crop of American origins, owes its name and presence in many parts of the New World, including the southern United States, to the intermediary of Kongo slaves, as can be seen in the words goober and pinda. The latter then again traveled from the Dutch Caribbean to the Dutch-speaking part of Europe.5
A second well-studied case is the East-African coast, where the principle language, Swahili, is emblematic of cultural contact. Aside from the fact that East Africa hosts speech communities from three major language families (Nilo-Saharan, Afroasiatic, and Niger-Congo, more specifically, the respective subgroups Nilotic, Cushitic, and Bantu) and some language isolates (Hadza, Sandawe), the coasts have been settings of contacts with many non-African peoples, as well. Written documents from Greek and possibly Roman authors describe East-African coasts from the 1st century ad onward, but Egyptian and Persian traders visited the region long before that time.6 Ever since Swahili culture developed on the East-African coasts during the first millennium ad, it has been influenced by “oriental” elements through contacts with Persian, Iranian, and Indian peoples.7 Most famous, however, are the contacts with Arabic-speaking peoples, who established settlements along the Swahili Coast from the 10th century ad onward. When the Portuguese first arrived at the Swahili Coast in 1498, they encountered a multicultural environment that included people of local African descent and people with “oriental” origins. Portuguese, and during colonialism also British, features were then brought into the mix. Swahili contains traces of all these contacts—mostly loanwords, but also structural elements. It is the Arabic component, however, that receives most attention, which has led to assumptions that the Swahili language resulted from contacts with Arabs. The existence of Swahili literature in Arabic script surely reinforces that idea. The claim, however, is unjustified. Swahili is clearly an African language, a member of the Sabaki subgroup of East Bantu. The influence of the Arabic language is mostly seen in loanwords, which can mainly be categorized into the semantic fields of “jurisprudence, trade, religion, nonindigenous flora, and maritime affairs.”8 These semantic fields correspond to domains in which the Arabs did indeed introduce knowledge into the region—notably: Islam; foods like adesi “lentils” and kashata “cakes, biscuits”; and, at the seaside, the sight of dhows offshore. Applying the principles of comparative linguistics, Nurse and Spear concluded that these loanwords were not affected by specific sound changes and therefore must postdate ad 1500. The absence of large numbers of these loanwords in the non-Zanzibari dialects confirms the late date of their borrowing. The aforementioned Swahili literature also dates to the last few centuries. Still, Arabic influence has a longer history, as is exemplified by vocabulary cited in 12th-century travel accounts and the reconstruction of Arab loans into Proto-Swahili, for instance, ngamia “camel.” The linguistic evidence merely indicates that the contacts in the earlier periods were not as intense as was previously thought.
Especially when one travels back in time to the period not documented by European or Arab texts, the study of contact by means of comparative linguistics becomes essential to unveil parts of that history. Many historians have applied the Words-and-Things method to study the early history of sub-Saharan Africa; this method not only allows for the identifying of inherited cultural traits but also the tracking of borrowed crops, technologies, and concepts. A well-studied case is the agricultural exchange between speakers of Central Sudanic, Eastern Sudanic, Southern Cushitic, and Bantu languages in the Great Lakes region, with a particular focus on cereal cultivation and cattle-keeping. Based on loanword evidence, Christopher Ehret and David Schoenbrun have concluded that East-Bantu speech communities learned to cultivate and prepare pearl millet from Central and Eastern Sudanic communities.9 Early East-Bantu speakers also acquired knowledge of cattle, but it was only when daughter communities, such as “Great Lakes Bantu,” came into contact with Southern Cushitic communities, who specialized in herding, that livestock gained in importance and also became appreciated for secondary products such as milk and blood. The same daughter community, “Great Lakes Bantu,” also adopted a new cereal, finger millet, from speakers of Sudanic languages. Linguists have re-analyzed and adjusted some of this evidence. Koen Bostoen, for instance, proposed origins for the pearl millet term in West-Nilotic, an Eastern Sudanic subgroup.10 Birgit Ricquier has demonstrated that at least one of the presumed loanwords concerning cereal preparations is actually of Bantu origins, suggesting that East-Bantu communities integrated new crops into their inherited culinary practices. Nevertheless, the widespread though dispersed occurrence of gari/gali-like loanwords displays the link with either Nilo-Saharan or Afroasiatic speech communities.11
Another case in which loanwords are used to trace the introduction of an important starch food in Central Africa concerns plantains. An osculant series of words, meaning words that are similar in form and meaning, can be reconstructed for Bantu languages with the semantic value “plantain”: *-gòndò ~ *-gòndè ~ *-gò, or the same but with initial consonant *k.12 The series led Malcolm Guthrie to the conclusion that plantains were known to the first Bantu speakers. Jan Vansina, on the other hand, stresses the impossibility of reconstructing one form into Proto-West-Bantu, let alone Proto-Bantu, and suggests an early loanword scenario.13 Plantains would have been introduced from the Upper Nile and further distributed to the west. Vansina’s interpretation is based on archaeobotanical evidence. Other archaeobotanists disagree and propose that they were introduced at the West-Central African coasts.14 The loanword evidence on the trajectories of plantain introduction, however, remains inconclusive.
Borrowing, of course, is not the only outcome of contact between speech communities. In Central Africa, early Bantuphone communities came into contact with hunter-gatherers. Central African hunter-gatherers speak languages that are closely related to the languages of their neighbors and thus belong to three different language groups: Bantu and Ubangi (both Niger-Congo) and Central Sudanic (Nilo-Saharan). The degree of affiliation with neighboring languages varies between dialectal differences, on the one hand, and the absence of mutual intelligibility, on the other, as is the case for Aka, Baka, and Asua. Scholars have previously attempted to retrace the original “Pygmy language,” but genetic evidence demonstrates that the differentiation of these hunter-gatherer communities from other populations occurred too far back in time, some 60,000 to 90,000 BP, making the hypothesis of common language inheritance highly unlikely.15 Still, some hunter-gatherer communities may have spoken the same language in a more recent past, as is evidenced by a large set of common vocabulary between Aka and Baka, who currently speak languages belonging to different language groups—Bantu and Ubangi, respectively. Central African hunter-gatherers live in a system of economic exchange with neighboring farmers. Their need for communication explains why language shifts occurred in the past. In some cases, the language of Central African hunter-gatherers does not correspond to that of their neighbors but to people living farther afield, which is suggestive of past migrations.
Identification of Loanwords
Ehret compared languages to archives, each containing “many thousands of individual artifacts of the past.”16 Following Ferdinand de Saussure, the link between a given word and its referent—that is, the concept it refers to in the real world—is arbitrary. The phonological shape of a word in a given language cannot be predicted. A tree, for instance, has no intrinsic quality that would determine its name.17 Therefore, a tree can be called ǹti in Kikongo and arbre in French. There is no evident link, either, between the semantic value of a word and its referent. The Kikongo word, for example, can be applied to any long wooden object—not just to a tree but also a branch, a stick, or a pole, and so on. The meaning of the French word, in contrast, is limited to a particular type of flora. The lexicon of a language thus offers a specific impression of reality. Given the arbitrary relations between phonological shape, semantic value, and referent, a word with a similar form, meaning, and referent that is shared by two or more languages is the result of shared history. If the word can be reconstructed in the relevant proto-language, it is evidence of shared inheritance. If, however, the word appears to have a longer history in one language, it is likely to have been borrowed into the other. Historians can use this information to study the past. It must be observed, however, that “linguistic artifacts” are only a reflection of reality and therefore always constitute indirect evidence.
To distinguish inherited lexicon from borrowed lexicon, the comparative linguistic method must be applied, in particular the Words-and-Things approach. Loanwords can be identified formally, by tracing irregularities in the sound correspondences or the morphological characteristics, or both, of the word in the recipient language. For instance, the Swahili words daktari “doctor” and maktaba “library” can easily be recognized as loanwords considering that the combination /kt/ is not part of the phoneme inventory in Swahili or the related Bantu languages. Maktaba is also similar to the Swahili kitabu “book.” The morphological link between the two immediately reveals their Arabic origins, since Arabic words generally consist of three consonants, and prefixes, suffixes, and infixes change their meaning. Another criterion is the possibility of semantic analysis of the word in the donor language. In this case, kitabu and maktaba are clear derivations from Arabic ktb “to write.” It must be also noted that words can be borrowed in a sort of chain reaction: one language may borrow a term in the role of recipient language, and then subsequently act as donor language and transfer the term to a new recipient language. Swahili daktari is a loan from English doctor, but its form reveals that Indic languages played the role of intermediary.18
Other linguistic criteria for the identification of loanwords have to do with geography and language groups. The donor language can be traced through cognates in related languages. Getting back to the maktaba and kitabu example, the verb ktb is also found with many derivations in Hebrew, proving the direction of borrowing from Arabic to Swahili. When one looks at the recipient languages, loanwords tend to display a continuous geographical distribution that crosses linguistic boundaries, in contrast to inherited words, which usually display a dispersed distribution, with distant dots on the map of a given language family or subgroup. Taking another example, words referring to “porridge” that can be reconstructed as *NP9-kímà occur all over the Bantu domain.19 Combined with the evidence of regular sound correspondences, they clearly concern an ancient inherited word. Words with the form fufu “cassava flour porridge,” in contrast, occur in a continuous region following the Congo River and crossing boundaries between Bantu subgroups, thereby proving its loanword status.20
Making Sense of Loanwords
Loanwords inform us about past contacts between specific speech communities, both through the identification of the donor language and through the routes of dispersal. The porridge example reflects trade along the Congo River between speakers of different Bantu languages, which allowed for the dispersal of a new dish—namely, cassava flour porridge. Most of all, of course, loanwords provide insights into the history of the respective referents. They speak, among other things, of technological innovations, the introduction of new ingredients, and the spread of novel ideologies. Loanwords often come in sets that are linked to particular semantic domains. Many African languages contain words of Amerindian origins referring to New World crops, which were adopted through European intermediaries. And as mentioned, many Arabic loans in Swahili are situated in the semantic fields of religion, jurisprudence, and trade. Focusing on semantic fields makes it possible to identify layers of cultural contact. For Rumanyo, for instance, the lexicon for subsistence strategies can be divided in loanword sets, each from different origins:
Native Rumanyo (not borrowed)
ndjóvu “elephant, gen.”
nkungúru “elephant bull”
muncúngu “water lily spec.”
shikémbe “elephant calf”
ligcù “gut of fish”
lirândo “elephant trail”
ligcuvancé “sardine (adult fish)”
hámbo “cattle pen, kraal”
farama “farm (European style)” < English
mahêndjere “fresh milk after milking”
katófuru “potato” < German
lihênga “beestings, first milk of a cow”
háraka “rake” < German, Afrikaans
-kámura “to furrow” < Afrikaans
muhará “rope for leading a cow”
fáikisha “barrel” < Afrikaans21
The loanword evidence here indicates a layered history of newly adopted subsistence strategies: from a hunter-gatherer economy with expertise in elephant hunting to the exploitation of aquatic resources, cattle-keeping, and, ultimately, agriculture.
The semantic fields of adopted loanwords are indicative of the nature of the contacts that occurred between the concerned speech communities. The loanwords from Kwangali are mainly linked to male activities, especially cattle-keeping; this, according to Möhlig, is linguistic proof of intermarriage between Rumanyo-speaking women and Kwangali men.22 A more frequent type of contact leading to the adoption of loanwords is trade, which may vary between occasional contacts, at one end of the spectrum, and a system of economic symbiosis, at the other end of the spectrum. Other contact situations include those of religious and political dominance. Arabic loans in Swahili may be cited here; but think also of the many loans of European origins concerning political structures and Christianity.
Of course, history needs dates or, at least, a relative chronology. When were the loanwords adopted, and when did the introductions become part of the recipient community’s material or immaterial culture? Most often, the adoption of loanwords is dated using extralinguistic evidence. It is evident that loans regarding the Columbian Exchange must postdate the first African-European contacts, for instance, not earlier than 1482 for the West-Central African coasts. Links can also be forwarded between the pattern of loanword distribution and historical records of trade routes through the continent or zones of linguistic influence. When there is a long-standing history of language contact, as is the case with Swahili and Arabic, caution is needed. Because of the large number of words from Arabic origins in Swahili, it is popularly assumed that Arabic loans were adopted into Swahili early. However, most such words are specialized vocabulary, and the linguistic evidence from sound shifts suggests that they date to the last few centuries. To be precise, Arabic loans have not been affected by sound changes, except by the loss of /l/, a late sound shift that was in progress during the 18th and 19th centuries, as can be seen in the literature from that period.23 Obtaining a relative chronology using sound correspondences requires thorough historical-linguistic research, which, sadly, is not available for the majority of African languages. In the case of ancient loanwords, relative dating can be obtained by linking loans to linguistic subgroups and even reconstructing a proto-form in the relevant ancestor language.24 Attributing Eastern Sudanic and Cushitic loans to either East Bantu or Great Lakes Bantu thus allows for a relative chronology of introductions linked to agriculture and livestock keeping.25
Constraints and Limitations
The identification of loanwords depends on the quality of the linguistic data, the available knowledge on sound correspondences, and insights into language classification. If there are issues with any one of those three components, it complicates the comparative work. Concerning the data, the collected lexicon is ideally notated with all phonological and tonal information, and with a detailed description of its semantic value and all possible referents. The best way to retrieve such data is through fieldwork or by the use of detailed dictionaries. However, the study of early history requires comparing vocabulary in a large set of languages, rendering fieldwork for each language impossible. The solution is to work with dictionaries, glossaries, word lists, and other documents that contain lexical data. Tone and semantic detail—valuable information that may be necessary to determine a word’s origins—are often lacking in such sources.
A second issue concerns regular sound correspondences established by previous authors. Guthrie’s Comparative Bantu, for instance, is regularly consulted by historians and historical linguists. Still, Guthrie doesn’t include all Bantu languages. Moreover, he doesn’t list all sound correspondences for the languages that are included. This can be overcome by following the principles of comparative linguistics and establishing sound correspondences by looking at a wider sample of vocabulary. Of course, this adds extra steps to the research process and required the availability of sufficient comparative data for the languages concerned.
Finally, the interpretation of the results is shaped by the available insights on the historical relations between the languages under study. How to classify language groups like Bantu is the subject of an ongoing debate. As a consequence, integrating the results in one language tree or the other may lead to different historical narratives. Moreover, classifications of smaller subgroups may not even be available. Some historical works show the investment of the respective scholars in the linguistic history of the communities under study.26 However, time constraints force most studies to rely on the available literature on sound shifts and classifications, which means that they are subject to the aforementioned drawbacks.
The study of loanwords also suffers from some limitations linked to the nature of linguistic change and contact. When the donor language and the recipient language belong to different language families, as in the Swahili examples, identifying loanwords is a rather straightforward matter considering the very different phoneme and morpheme inventories of each language family. When borrowing occurs between languages of the same language family, or even the same language subgroup, it is harder to distinguish between cognates and loans. Möhlig offers the example of Rumanyo vocabulary displaying the sound correspondence Proto-Bantu *c > h, whereas many other terms are the outcome of *c > sh.27 Words such as mahângu “millet,” hûpa “calabash,” and hâmbo “cattle-pen” can be linked to reconstructions for earlier stages in Bantu history (*-cángú, *N-cúpà, and *N-cámbò, respectively) and, at first glance, look to be inherited into Rumanyo.28 Given that the sound change *c > h is not regular in Rumanyo and seems reserved mostly to vocabulary in the lexical fields of cattle-keeping and agriculture, Möhlig concludes that the relevant words are loans from another Bantu language, namely, Kwangali.
Still, sound correspondences do not always provide the answer. In a fairly homogenous language group such as Bantu, some sound shifts are identical in many of the individual languages. Sometimes, there is no sound shift at all. Proto-Bantu nasals, for instance, have not changed much in most of the Bantu languages and, except for the 7- to 5-vowel merger, the same holds true for vowels. The word unga “flour” is identical in many East-Bantu languages or is semi-identical when different forms of the noun prefix are taken into account—vunga in Tikuu and ubunga in Bemba, for example.29 Considering the distribution of the noun, a reconstruction could be proposed for Proto–East Bantu with the form *NP14-ʊ̀ngà. However, the distribution also largely corresponds to the region where Swahili is used as a lingua franca, and several dictionary authors consider the word to be a loan from Swahili. Tonal information could be helpful here, because loans from the nontonal language Swahili regularly appear with a high-low tone pattern in the recipient tonal language. Unfortunately, dictionaries too often lack tone notations. In short, there is no conclusive linguistic evidence in this case to determine whether unga is a recent loan or an inherited term.
Sound correspondences can also be blurred by the tendency of languages to integrate foreign words in their own phoneme inventory. Standard Swahili prescribes the pronunciation of Arabic loanwords with the original sounds. But many Swahili dialects and other languages at the East-African coasts substitute more familiar sounds for the Arabic sounds:
/q/ > /k/
/ɣ/ > /g/
/θ/ > /s/
/ð/ > /z/30
Such phone substitution often cannot be distinguished from a regular sound change. Ha ubhugari “mush” can thus be regarded as a regular reflex of the early East-Bantu word reconstructed as *NP14-gàdɩ̀, but it could equally be a loan from Swahili ugali. The Ha word integrated the second consonant of the noun stem, but also the form of the prefix. In Bantu languages, loanwords are easily integrated morphologically. Consider Swahili kitabu “book,” whose first syllable is reinterpreted as a nominal prefix of noun class 7, thus forming its plural in the corresponding noun class 8: vitabu “books.”
The opposite may occur as well—namely, sounds may be added to the phoneme inventory of a given language because of extensive borrowing. The most famous case of this are the click sounds originating from Khoisan languages that now form an integral part of southern Bantu languages. Rumanyo, for instance, counts a large number of lexemes with click sounds, presumably borrowed from an unknown and now extinct Khoisan language. Of these loans, 20 percent can be categorized into the semantic field of aquatic life, which reinforces the loanword claim. Some examples:
“sardine (adult fish)”31
An important asset of loanwords is the possibility to trace the origins of a word and, consequently, its referent. The above-mentioned examples illustrate the adoption of tools, crops, techniques, and so on, by one or several speech communities from identifiable source communities. However, loanwords can often be distinguished from inherited vocabulary without a clear indication of a donor language, and hence a community. For instance, Northern Kongo kipáki “wooden board on which cassava is kneaded” and kípúúti “cassava dough” display the plosive /p/ instead of the regular, expected /h/ or /v/. The loans kipáki and kípúúti are clearly related to the adoption of cassava preparations, but for now, no origins can be formulated for the two nouns.
More Types of Borrowing and Language Contact
Loanwords constitute but one outcome of contact between speech communities. On the lexical level, contact may also result in loan translations or calques. A loan translation is a literal translation of a foreign concept, for instance, Swahili poa “be(come) cool” from the English cool, or maneno ya kuazima “borrowed words” from English loanwords. Other lexical outcomes are loan blends and correspondence mimicry. Loan blends are words that combine borrowed and inherited elements, for instance, kijipicha “thumbnail,” which combines the Swahili diminutive ki-ji- with the English borrowing picha “picture.” Correspondence mimicry refers to the phonological adaptation of loanwords based on the sound correspondences between the languages concerned as perceived by the speakers, for instance, the insertion of /l/ in Nyakyusa itala “lamp, torch” from Swahili taa after the regularly corresponding Nyakyusa -zala and Swahili zaa “to bear a child.”32 Structural borrowing involves the borrowing of phonological and grammatical features. Lodhi, for instance, observes that the Arabic influence in Swahili includes structural elements, as in the use of conjunctions and prepositions.33 Long-term contact between languages can also lead to convergence. A linguistic area, or “Sprachbund,” is one in which languages belonging to different language groups or families display common features as a result of convergence. In East Africa, the Tanzanian Rift Valley forms a linguistic area, where Sandawe, Hadza, and Southern Cushitic, Nilotic and Bantu languages share linguistic features.34 More dramatic consequences of language contact include language shift, as in the case of Central African hunter-gatherer languages, and the creation of new languages—namely, pidgins and creoles.
Discussion of the Literature
The use of linguistic elements as historical evidence to study the past of Central and Southern Africa began in the 1960s with the work of Christopher Ehret. Ehret’s first articles focused on language contact in East Africa.35 Other historical studies with a lot of loanword evidence include Mary Allen McMaster on the Uele Region, and David Schoenbrun on the Great Lakes.36 Most of the historical research that applies the Words-and-Things approach, however, addresses both inherited and borrowed words. For instance, the study by Edna Fields-Black, a rare example of applying linguistic methods to the history of West Africa, addresses both inherited, innovated, and borrowed lexicon to investigate the origins of rice cultivation in coastal Guinea.37 Linguistic studies that are of interest to historians include Lodhi’s study of oriental loanwords in Swahili; Nurse’s work on the effects of language contact on Daiso and Ilwana; Nurse and Hinnebusch on the history of Sabaki languages, which also contains loanword lists, and several volumes of the journal Sprache und Geschichte in Afrika in which language contact is the central theme.38 These are just a few titles, since linguistic studies covering language contact are numerous. Regarding the theme of language shift, the “pygmy” case has been addressed by Bahuchet since the 1980s, and Klieman used linguistic evidence in her study of historical relations between Bantuphone peoples and hunter-gatherers in the Central African rainforest.39
The comparative method necessarily involves comparing data in multiple languages. Aside from fieldwork, the data can be obtained from dictionaries, word lists, and the like. In Europe, important library collections for African languages can be consulted at the Royal Museum for Central Africa (Tervuren, Belgium), the African Studies Centre, Leiden (the Netherlands), the SOAS University of London (United Kingdom), and the Goethe University Frankfurt (Germany), among other institutes. In the United States are the Melville J. Herskovits Library of African Studies at Northwestern University in Evanston (Illinois) and the libraries of the Indiana University Bloomington and the University of California, Los Angeles, among others. Language materials can also be found online, for instance, on the website Comparative Bantu OnLine Dictionary, which also includes the database Tanzanian Language Survey.40 The lists of basic vocabulary in selected Bantu languages collected for the classification by Bastin et al. are available on the website of the Royal Museum for Central Africa.41
Linguistic studies that may be helpful for the identification of loanwords include works on sound correspondences and reconstructions. For Bantu languages, for instance, the work of Malcolm Guthrie is foundational, but sound correspondences are also available for specific Bantu languages or language groups, such as Kongo languages, Sabaki, languages of zone F, and several other East-Bantu subgroups.42 These works include lexical reconstructions. The largest collection of reconstructions for the Bantu languages can be found in the online database Bantu Lexical Reconstructions 3.43 Reconstructions can be used to eliminate loanword hypotheses in cases of doubt or to establish sound correspondences. An article by Ricquier from 2017 summarizes the Words-and-Things approach and the comparative historical linguistic methods and fieldwork protocols for working with African languages.44
Bahuchet, Serge. “Changing Language, Remaining Pygmy.” Human Biology 84, no. 1 (2012): article 9.Find this resource:
Dimmendaal, Gerrit J. Historical Linguistics and the Comparative Study of African Languages. Amsterdam: John Benjamins, 2011.Find this resource:
Ehret, Christopher. An African Classical Age: Eastern and Southern Africa in World History, 1000 B.C. to A.D. 400. Charlottesville: University of Virginia Press, 1998.Find this resource:
Guthrie, Malcolm. Comparative Bantu: An Introduction to the Comparative Linguistics and Prehistory of the Bantu Languages. 4 vols. Farnborough, UK: Gregg Press, 1967–1971.Find this resource:
Lodhi, Abdulaziz Y. Oriental Influences in Swahili: A Study in Language and Culture Contacts. Gothenburg, Sweden: Acta Universitas Gothoburgensis, 2000.Find this resource:
Möhlig, Wilhelm J. G., ed. Cultural Change in the Prehistory of Arid Africa: Perspectives of Archaeology and Linguistics. Sprache und Geschichte in Afrika 18. Cologne: Rüdiger Köppe Verlag, 2007.Find this resource:
Möhlig, Wilhelm J. G., Frank Seidel, and Marc Seifert, eds. Language Contact, Language Change and History Based on Language Sources in Africa. Sprache und Geschichte in Afrika 20. Cologne: Rüdiger Köppe Verlag, 2009.Find this resource:
Nurse, Derek, ed. Historical Language Contact in Africa. Sprache und Geschichte in Afrika 16/17. Cologne: Rüdiger Köppe Verlag, 2001.Find this resource:
Nurse, Derek, and Thomas J. Hinnebusch. Swahili and Sabaki: A Linguistic History. Berkeley: University of California Press, 1993.Find this resource:
Nurse, Derek, and Thomas Spear. The Swahili: Reconstructing the History and Language of an African Society, 800–1500. Philadelphia: University of Pennsylvania Press, 1985.Find this resource:
Pasch, Helma. Linguistische Aspekte der Verbreitung lateinamerikanischer Nutzpflanzen in Afrika. Cologne: Universität zu Köln, 1980.Find this resource:
Ricquier, Birgit. “The History of Porridge in Bantuphone Africa, with Words as Main Ingredients.” Afriques 5 (2014).Find this resource:
Ricquier, Birgit. “Kongo Cuisine and the Middle Passage in Terms of Peanuts.” In Kongo across the Waters. Edited by Susan Cooksey, Robin Poynor, and Hein Vanhee, 258–263. Gainesville: University Press of Florida, 2013.Find this resource:
Ricquier, Birgit. “The ‘Words and Things’ Method.” In Field Manual for African Archaeology. Edited Alexandre Livingstone Smith, Els Cornelissen, Olivier P. Gosselain, and Scott MacEachern, 261–263. Tervuren, Belgium: Royal Museum for Central Africa, 2017.Find this resource:
Schoenbrun, David L. A Green Place, A Good Place: Agrarian Change, Gender and Social Identity in the Great Lakes Region to the 15th Century. Portsmouth, NH: Heinemann, 1998.Find this resource:
Schoenbrun, David L. “We Are What We Eat: Ancient Agriculture between the Great Lakes.” Journal of African History 34, no. 1 (March 1993): 1–31.Find this resource:
(1.) This article is not an exhaustive list of historical works using loanword evidence, or of the literature on each of the topics covered, but it includes some fine examples of historical conclusions based on the linguistic evidence of language contact, some of which are discussed in detail. The examples stem mostly, though not exclusively, from the field of expertise of the author—namely, the food history of Bantu-speaking communities, a lexical domain in which borrowing has occurred extensively throughout history.
(2.) Birgit Ricquier, “Porridge Deconstructed: A Comparative Linguistic Approach to the History of Staple Starch Food Preparations in Bantuphone Africa” (PhD diss., Université Libre de Bruxelles, 2013); and Birgit Ricquier, “A Foodie’s Guide to Kongo Language History: Early Events, North versus South, and the Innovative West,” Africana Linguistica 22 (2016): 107–146.
(3.) Ricquier, “Foodie’s Guide.”
(4.) These results were forwarded by Birgit Ricquier and Odette Ambouroue, “Congo River Crossroads Cuisine: A Blend of Western and Eastern Words” (5th International Conference on Bantu Languages “Bantu 5,” Paris, 2013). More names and the dispersal routes of American crops can be found in Helma Pasch, Linguistische Aspekte der Verbreitung lateinamerikanischer Nutzpflanzen in Afrika (Cologne: Universität zu Köln, 1980).
(5.) See Judith A. Carney and Richard Nicholas Rosomoff, In the Shadow of Slavery: Africa’s Botanical Legacy in the Atlantic World (Berkeley: University of California Press, 2009); and Birgit Ricquier, “Kongo Cuisine and the Middle Passage in Terms of Peanuts,” in Kongo across the Waters, ed. Susan Cooksey, Robin Poynor, and Hein Vanhee (Gainesville: University Press of Florida, 2013), 258–263.
(6.) Abdulaziz Y. Lodhi, Oriental Influences in Swahili: A Study in Language and Culture Contacts (Gothenburg: Acta Universitas Gothoburgensis, 2000), 50. See also Abdul Sheriff, “The Swahili in the African and Indian Ocean Worlds to c. 1500,” The Oxford Research Encyclopedia of African History (December 2017).
(7.) Adria LaViolette, “Swahili Cosmopolitanism in Africa and the Indian Ocean World, A.D. 600–1500,” Archaeologies: Journal of the World Archaeological Congress 4, no. 1 (2008): 24–49; Matthew Pawlowicz, “Modelling the Swahili Past: The Archaeology of Mikindani in Southern Coastal Tanzania,” Azania: Archaeological Research in Africa 47, no. 4 (2012): 489; and Derek Nurse and Thomas Spear, The Swahili: Reconstructing the History and Language of an African Society, 800–1500 (Philadelphia: University of Pennsylvania Press, 1985), 31.
(8.) Nurse and Spear, Swahili, 15.
(9.) Christopher Ehret, “Cattle-Keeping and Milking in Eastern and Southeastern African History: The Linguistic Evidence,” African History 8 (1967): 1–17; Ehret, “Agricultural History in Central and Southern Africa, ca. 1000 B.C. to A.D. 500,” Transafrican Journal of History 4, no. 1–2 (1974): 1–25; Ehret, An African Classical Age: Eastern and Southern Africa in World History 1000 B.C. to A.D.400 (Charlottesville: University of Virginia Press, 1998); David L. Schoenbrun, “We Are What We Eat: Ancient Agriculture between the Great Lakes,” Journal of African History 34 (1993): 1–31; and Schoenbrun, A Green Place, A Good Place: Agrarian Change, Gender and Social Identity in the Great Lakes Region to the 15th Century (Portsmouth, NH: Heinemann, 1998).
(10.) Koen Bostoen, “Pearl Millet in Early Bantu Speech Communities in Central Africa: A Reconsideration of the Lexical Evidence,” Afrika und Übersee 89 (2006–2007): 183–213.
(11.) Birgit Ricquier and Koen Bostoen, “Stirring Up the Porridge: How Early Bantu Speakers Prepared Their Cereals,” in Windows on the African Past: Current Approaches to African Archaeobotany, ed. A. G. Fahmy, S. Kahlheber, and A. C. D’Andrea, 209–224 Proceedings of the 6th International Workshop on African Archaeobotany, Cairo (Frankfurt: Africa Magna Verlag, 2011); and Birgit Ricquier, “The History of Porridge in Bantuphone Africa, with Words as Main Ingredients,” Afriques 5 (2014).
(12.) Gérard Philippson and Serge Bahuchet, “Cultivated Crops and Bantu Migrations in Central and Eastern Africa: A Linguistic Approach,” Azania (The Growth of Farming Communities in Africa from the Equator Southwards) 29–30 (1994–1995): 103–120; Edmond De Langhe, Rony L. Swennen, and D. Vuylsteke, “Plantain in the Early Bant World,” Azania: The Growth of Farming Communities in Africa from the Equator Southwards 29–30 (1994–1995): 147–160. The series is reconstructed with initial consonant *k in Malcolm Guthrie, Comparative Bantu: An Introduction to the Comparative Linguistics and Prehistory of the Bantu Languages, 4 vols. (Farnborough, UK: Gregg Press, 1967–1971); Jan Vansina, Paths in the Rainforest: Toward a History of Political Tradition in Equatorial Africa (Madison: University of Wisconsin Press, 1990).
(13.) Vansina, 62–64.
(14.) De Langhe, Swennen, and Vuylsteke, “Plantain”; and Christophe M. Mbida, Hugues Doutrelepont, Luc Vrydaghs, Rony L. Swennen, et al., “First Archaeological Evidence of Banana Cultivation in Central Africa during the Third Millennium before Present,” Vegetation History and Archaeobotany 10, no. 1 (2001): 1–6.
(16.) Christopher Ehret, “Writing African History from Linguistic Evidence,” in Writing African History, ed. J. E. Philips, 86 (Rochester, NY: University of Rochester Press, 2005).
(17.) Exceptions are, of course, onomatopoeias, even if these are also language-specific and susceptible to inheritance and borrowing. Consider, for instance, the word for “sheep” dimeeme, which can be reconstructed into Proto-West-Bantu.
(18.) Lodhi, Oriental Influences, 129. Of course, English doctor is itself a loan from Latin.
(19.) Words with an asterisk * are reconstructions in a proto-language, for instance Proto-Bantu, Proto-East Bantu, and so on. Where possible, not just the noun stem but also the nominal prefix is reconstructed. Here, “NP9-” refers to the nominal prefix of noun class 9.
(21.) Sample from Wilhelm J. G. Möhlig, “Linguistic Evidence of Cultural Change: The Case of the Rumanyo Speaking People in Northern Namibia,” Sprache und Geschichte in Afrika 18 (2007): 135, 144, 148, 150.
(22.) Möhlig, “Linguistic Evidence,” 147.
(23.) Nurse and Spear, Swahili, 15.
(24.) Nurse and Hinnebusch link sound shifts to subgroups and seek correlates for the concerned ancestor communities in archaeology. As such, they obtain dates for the occurrence of the mentioned sound shifts. Derek Nurse and Thomas J. Hinnebusch, Swahili and Sabaki: A Linguistic History, ed. Thomas J. Hinnebusch, with a special addendum by Gérard Philippson (Berkeley: University of California Press, 1993), 648. See also The White Fathers, Bemba-English Dictionary (London: Longmans, Green, 1954).
(25.) Ehret, “Cattle-Keeping”; Ehret, African Classical; Schoenbrun, “We Are What We Eat”; and Schoenbrun, Green Place.
(26.) For instance, Schoenbrun, Green Place; linguistic material and conclusions in David L. Schoenbrun, “Great Lakes Bantu: Classification and Settlement Chronology,” Sprache und Geschichte in Afrika 15 (1994): 91–152; and Schoenbrun, The Historical Reconstruction of Great Lakes Bantu Cultural Vocabulary: Etymologies and Distributions (Cologne: Rüdiger Köppe Verlag, 1997). Another example is Kathryn de Luna, Collecting Food, Cultivating People: Subsistence and Society in Central Africa (New Haven, CT: Yale University Press, 2016).
(27.) Möhlig, “Linguistic Evidence,” 147–148.
(28.) This is Möhlig’s notation. In the database Bantu Lexical Reconstructions 3, the vowels receive a different notation—namely, *-cángʊ́, *N-cʊ́pà, and *N-cámbò.
(29.) Nurse and Hinnebusch, Swahili and Sabaki, 54.
(30.) Lodhi, Oriental Influences, 30.
(31.) Möhlig, “Linguistic Evidence,” 144.
(32.) Examples from Lodhi, Oriental Influences; and Gerrit J. Dimmendaal, Historical Linguistics and the Comparative Study of African Languages (Amsterdam: John Benjamins, 2011).
(33.) Lodhi, Oriental Influences, 99–115.
(34.) Details in Roland Kiessling, Maarten Mous, and Derek Nurse, “The Tanzanian Rift Valley Area,” in A Linguistic Geography of Africa, ed. Bernd Heine and Derek Nurse, 186–227 (Cambridge, UK: Cambridge University Press, 2008).
(35.) Ehret, “Cattle-Keeping”; and Ehret, “Agricultural History.”
(36.) Mary Allen McMaster, “Patterns of Interaction: A Comparative Ethnolinguistic Perspective on the Uele Region of Zaïre ca. 500 B.C. to 1900 A.D.” (PhD diss., University of California, 1988); Schoenbrun, “We Are What We Eat”; Schoenbrun, Historical Reconstruction; and Schoenbrun, Green Place.
(37.) Edda L. Fields-Black, Deep Roots: Rice Farmers in West Africa and the African Diaspora (Bloomington: Indiana University Press, 2008).
(38.) Derek Nurse, ed., Historical Language Contact in Africa, Sprache und Geschichte in Afrika 16/17 (Cologne: Rüdiger Köppe Verlag, 2001); Wilhelm J. G. Möhlig, ed. Cultural Change in the Prehistory of Arid Africa: Perspectives of Archaeology and Linguistics, Sprache und Geschichte in Afrika 18 (Cologne: Rüdiger Köppe Verlag, 2007); Wilhelm J. G. Möhlig, Frank Seidel, and Marc Seifert, eds., Language Contact, Language Change and History Based on Language Sources in Africa, Sprache und Geschichte in Afrika 20 (Cologne: Rüdiger Köppe Verlag, 2009); Lodhi, Oriental Influences in Swahili; Derek Nurse, Inheritance, Contact, and Change in Two East African Languages (Cologne: Rüdiger Köppe Verlag, 2000); and Nurse and Hinnebusch, Swahili.
(39.) Serge Bahuchet and J. M. C. Thomas, “Linguistique et histoire des Pygmées de l’ouest du bassin congolais,” Sprache und Geschichte in Afrika 7, no. 2 (1986): 73–103; Serge Bahuchet, “History of the Inhabitants of the Central African Rain Forest: Perspectives from Comparative Linguistics,” in Tropical Forests, People, and Food: Biocultural Interactions and Applications to Development, ed. C. M. Hladik, 37–54 (Paris: Unesco/Parthenon, 1993); Bahuchet, “Changing Language”; and Kairn Klieman, “The Pygmies Were Our Compass”: Bantu and Batwa in the History of West Central Africa, Early Times to c. 1900 C.E. (Portsmouth, NH: Heinemann, 2003).
(41.) “African Linguistics: Lexicostatistic study Bantu languages,” Royal Museum for Central Africa,; and Yvonne Bastin, André Coupez, and Michael Mann, Continuity and Divergence in the Bantu Languages: Perspectives from a Lexicostatistic Study (Tervuren, Belgium: Royal Museum for Central Africa, 1999).
(42.) Guthrie, Comparative Bantu; Jan Daeleman, “Les réflexes Proto-Bantu en Ntándu (dialecte Kóongo),” in Mélanges de culture et de linguistique africaine, publiés à la mémoire de Leo Stappers, ed. C. E. S. Faïk-Nzuji Madiya, 331–397 (Berlin: Dietrich Reimer Verlag, 1983); Jean-Noël Nguimbi Mabiala, “Phonologie comparative et historique du kongo (groupe H10): Kiyoombi–Civili–Kibeembe–Kihangala–Kisuundi–Cilaadi” (PhD diss., Université Lyon 2, 1999); Nurse and Hinnebusch, Swahili and Sabaki; Balla Masele, “The Linguistic History of SiSuumbwa, KiSukuma and KiNyamweezi in Bantu Zone F” (PhD diss., Memorial University of New Foundland, 2001); Derek Nurse, Classification of the Chaga Dialects: Language and History on Kilimanjaro, the Taita Hills, and the Pare Mountains (Hamburg, Germany: Helmut Buske Verlag, 1979); and Thomas J. Hinnebusch, Derek Nurse, and Martin Mould, Studies in the Classification of Eastern Bantu Languages (Hamburg, Germany: Helmut Buske Verlag, 1981).
(44.) Birgit Ricquier, “The ‘Words and Things’ Method,” in Field Manual for African Archaeology, ed. Alexandre Livingstone Smith, Els Cornelissen, Olivier P. Gosselain, and Scott MacEachern, 261–263 (Tervuren, Belgium: Royal Museum for Central Africa, 2017).