Is It English or Engelsk?—part 3
The previous two GeoCurrents posts have examined the much-tout paper by Jan Terje Faarlund and Joseph Emonds purporting to show that English is a Scandinavian rather than West Germanic language. I have argued that the constructions identified by Faarlund and Emonds as Scandinavian imports developed internally to English, although the presence of large numbers of Norse-speaking Vikings, especially in northern England, played an important role in precipitating some of these changes. Thus, Faarlund’s statement that “it is highly irregular to borrow the syntax and structure from one language and use it in another language” is erroneous. In fact, languages in close contact over a long period often do swap grammar as well as words.
In her excellent response on the LanguageLog, Sally Thomason mentions several well-researched examples of grammatical borrowing. One of them, originally described by Gumperz and Wilson (1971), involves the languages spoken in Kupwar, a village Maharashtra, India in the border area between Indic languages (Urdu, Marathi) in the north and Dravidian languages (Kannada, Telugu) in the south. Kannada-speakers form the majority and are socially dominant in Kupwar, but Marathi is the regional language and is used for socially neutral intergroup communication. The local variety of Urdu spoken in Kupwar has borrowed from both Kannada, a Dravidian language, and Marathi, another Indic language. According to Thomason, “the changes include adoption of an inclusive/exclusive ‘we’ distinction, subject-verb agreement rules in four different constructions, word order features, and about a dozen other features”. The grammatical borrowing is not uni-directional: Kupwar variety of Kannada has borrowed from Urdu as well. For example, while Standard Kannada does not use a copula be in sentences like ‘This house is yours’, Kupwar Kannada does:
|‘This house is yours.’|
|‘This house is yours.’|
Another striking case, originally reported by Andrei Malchukov in 2002, involves the Tungusic language Evenki, which has borrowed a volitional mood suffix and an entire set of personal endings from the Turkic language Yakut. Bernd Heine and Tania Kuteva in Language Contact and Grammatical Change describe the use of interrogative pronouns as relative clause markers among younger speakers of Tariana, a North Arawak language of northwestern Brazil, modeled on a Portuguese constructions (pp. 2-3).
As Thomason notes in her LanguageLog post, the most frequently borrowed type of syntactic feature is word order, which is particularly relevant to argument on hand, as Faarlund and Emonds identify three English word-order structures as contributions from Old Norse. Their claim that English must be an off-shoot of Old Norse because such structures cannot be borrowed from one language to another is not corroborated by facts. For example, in my book Languages of the World: An Introduction (pp. 171-172), I discuss languages in the Toricelli family in Papua New Guinea, which present an exception to the generalization that Papuan languages are verb-final. Instead, Torricelli languages have the Subject-Verb-Object word order, as do Austronesian languages spoken in coastal areas adjacent to where the Torricelli languages are to be found.
Another example of a borrowed word order pattern, also discussed in Languages of the World: An Introduction (pp. 246-247)is found in Yiddish, a Germanic language whose speakers migrated into Slavic-speaking lands in the 13th century. Through long-term coexistence with Slavic neighbors—Polish, Russian, Ukrainian, Belarusian—and extensive bilingualism, Yiddish picked up not only Slavic words and morphemes (e.g. the noun-forming suffix -nik, see image on the left), but also grammatical properties. One such property is the possibility of placing more than one question word at the beginning of a question. In English, as in other Germanic languages, only one question word in a multiple question can appear in the beginning of the sentence. All other question words must appear where their respective answers would appear if the sentence were declarative rather than a question. So if several people bought various items for a potluck party, we can ask Who bought what? One question word, who, appears in the beginning of the questions, while the other one, what, which questions the object, appears post-verbally, that is exactly where the object would appear in a declarative sentence. Placing both question words in the beginning results in an ungrammatical string *Who what bought? (linguists mark ungrammaticality by * placed before the ungrammatical string). In Slavic languages, such as Russian or Polish, all question words must appear in the beginning of a question. The appropriate way to formulate the abovementioned question in Russian is Kto čto kupil?, that is literally ‘Who what bought?’. The literal translation of the English question—*Kto kupil čto? —is as ungrammatical in Russian as the Russian order is in English. Interestingly, such multiple questions in Yiddish reveal both its Germanic roots and its contact with Slavic languages, as it allows two equally acceptable ways to ask a multiple question. One possibility—illustrating the Germanic pattern—is to place only one question word in the beginning of the question, in which case the other question word must appear before the non-finite lexical verb, if there is one: Ver hot vos gekoyft? is literally ‘who has what bought?’. The other possibility, just as acceptable in Yiddish, is to place all question words in the beginning: Ver vos hot gekoyft? (literally, ‘who what has bought?’). The option is a Slavic-influenced pattern.
But one does not have to look at exotic languages for examples of grammatical borrowing. English itself bears traces of grammatical structures adopted from other languages. For example, the use of do in questions and negative sentences (e.g. Do you speak a Scandinavian language? and I do not speak Norwegian) is thought to have come from Celtic languages (for a more detailed discussion of this issue, check out John McWhorter’s book Our Magnificent Bastard Tongue: The Untold History of English). French grammatical influence is visible in compounds like attorney-general and court-martial, where—contrary to the normal English pattern—the adjectival element appears after rather than before the noun. (Many English speakers are confused by these atypically ordered compounds, especially when it comes to the plural forms: is the plural attorney-generals, with the plural -s at the end is in regular English compounds, or attorneys-general, with the -s in the middle, but attached to the noun element?) The suffix -ee in words like employee and lessee is of the French origin as well (linguists appear to be particularly fond of this suffix, inventing such words as possessee, kissee, lovee, etc.). All in all, English vocabulary and grammar are based on a mix of words and structures from a variety of sources, Old Norse being just one of them.
All these examples reveal a common pattern: “the probability of such [grammatical] influence naturally varies with the degree of intimacy that exists between the speakers of two languages”, as Albert С. Baugh and Thomas Cable put it (A History of the English Language, 5th edition, p. 104). In other words, grammatical borrowing occurs when the two (or more) groups find themselves in prolonged, intense, and intimate contact. When it comes to Scandinavians in England, the popular image of the Vikings as engaged exclusively in “rapine and plundering”, as an excerpt from the Anglo-Saxon chronicle dated 793 CE describes them, is appropriate for the first half-century of the Viking presence on the British soil, from their initial raids on the monasteries in Lindisfarne and Jarrow in 793 and 794 till the mid 9th century. In the 10th and early 11th centuries, “large bands of marauders [continued to] march and countermarch across England, carrying hardship and devastation” (Baugh and Cable, p. 95), but at the same time Scandinavian farmers “intermarried with the English, adopted many of their customs, and entered into the everyday life of the community” (ibid). An important aspect of this amalgamation of the two peoples was the adoption of Christianity by many of the Scandinavians, as evidenced by “the large number of Scandinavian names found not only among monks and abbots, priests and bishops, but also among those who gave land to monasteries and endowed churches” (ibid). A prime example is the inscription on the Kirkdale sundial, discussed in the previous GeoCurrents post, which mentions two priests with Scandinavian names, Hawarth and Brand, as well as the benefactor who rebuilt the church, Orm, son of Gamal. As discussed in an earlier GeoCurrents post, the everyday nature of many Norse loanwords, such as dirt, freckle, skirt and window, suggests that they made their way into English “through the give-and-take of everyday life” (Baugh and Cable, p. 100). For a period of time, English adopted military and legal vocabulary elements from Scandinavian, including scegþ ‘vessel’ and hūsting ‘assembly’, but such words were later supplanted by loanwords from Norman French (although “hustings” survives as a word meaning “electioneering circuit”). Yet the many everyday Norse loanwords survived the Norman invasion and continue to be part of the English vocabulary.
Faarlund’s claim that “it is highly irregular to borrow the syntax and structure from one language and use it in another language” not only runs contrary to facts, but also puts Faarlund himself in a bind: if languages are unable to borrow grammar, how is it that Modern English features many grammatical elements that trace to Old English and its West Germanic roots? He states that “English… is virtually unaffected by Old English”, but this is patently false. As discussed above, Modern English features many structural patterns that are direct descendants of Old English structures that developed in accordance with the internal logic of the language. As difficult as it is to quantify the Norse influence on English, studies such as Terrence Kaufman’s survey of phonological and morphological features (see Language Contact, Creolization, and Genetic Linguistics, edited by Sally Thomason and Terrence Kaufman), led Thomason to conclude that
“no more than 20% of the total set of comparable structural features of the most Norsified English dialects came from Norse. Even if Faarlund’s percentage for the syntax turns out to be higher, his syntactic Norse features are unlikely to raise the overall percentage of Norse-origin structures to an unusually high level, compared to other instances of structural diffusion in intense contact situations.”
Given the preponderance of Saxon elements in English, it is easier to account for the few Scandinavian borrowings into an otherwise West Germanic language than to explain the numerous West Germanic borrowings into a putatively Scandinavian tongue. So, as Thomason puts it, “It’s English, not Engelsk”.
Gumperz, John J. and Robert Wilson (1971) “Convergence and creolization: a case from the Indo-Aryan/Dravidian border”. In: Dell Hymes (ed.) Pidginization and Creolization of Languages: Proceedings of a Conference Held at the University of the West Indies Mona, Jamaica, April 1968. Cambridge University Press. Pp. 151-168.
« Is It English or Engelsk?—part 2
When Did Roma Leave India?—New Discovery... »