The formal adaptation of toponyms in Romanian translations from German between 1780 and 1830

Although the translation of foreign proper names into Romanian before its modern stage is insufficiently explored, it is known that the process is characterized (as the whole Romanian writing of the time) by the lack of generally valid norms. The process was influenced by various factors, including: the existence of different alphabets, orthographic and phonetic systems; the influence of the source-text vs. the existence of traditional pronunciation and writing patterns; the translation of a proper name from various languages (French, German, Italian, etc.); the diversity of proper names; the translator’s personality (linguistic knowledge, cultural formation). Starting from the translation strategies adopted by translators, our aim is to analyse the way in which the toponyms from three historical texts translated into Romanian from German in pre-modern stage (1780–1830) were adapted to the formal system of Romanian.


Introduction
Considered the "modern stage" in the history of literary Romanian (Gheție, 1982, p. 65;Munteanu & Țâra, 1983, p. 169, 176), also referred to as the "pre-modern stage" (Piru, 1970, p. 5, referring back to the transition period from the old to the modern literature), the period between 1780-1830/40 represents the beginning of the modernization of Romanian through the translation of texts pertaining to various fields of the written culture (religion, philosophy, history, geograpgy, literature, etc.) and from various modern languages (French, German, Italian, Russian, etc.).A decisive role in the modernization of literary Romanian was played by the representatives of the so-called "Școala Ardeleană" [The School from Transilvania], who put together a series of normative works (grammars, orthography textbooks or lexicons), translated scientific and legal-administrative treatises, as well as many literary works, accounting thus for the main requirements regarding the standardization and modernization of literary Romanian (Gheție, 1982, p. 75).The high number of Romanian translations dating back to the same period were insufficiently investigated from the point of view of translation strategies.Within this context, we aim to analyse the way in which proper names from several German texts were formally adapted (graphically, phonetically and morphologically) to Romanian.The cases in point are three historical texts, briefly presented as follows: a) Descoperirea Américii.O carte foarte folositoare (= Cam.rom.), published in 1816 in Buda, edited by Nicola Nicolau, is the Romanian translation of the first volume of the work Kolumbus oder die Entdekkung von Westindien.Ein angenehmes und nützliches Lesebuch für Kinder und junge Leute by Joachim Heinrich Campe; the work presents the discovery of America, told by a father to his children, in the form of dialogues; after comparing the German editions, we considered the reference text to be the edition published in 1782 in Tübingen (= Cam.germ.)1; b) The history textbook Élémens d'histoire générale by Abbot Claude-François-Xavier Millot, published starting from 1772 in Paris (in two volumes), was translated into Romanian by Ioan Molnar-Piuariu (1749-1815) and published in 1800 in Buda, under the title Istoria universală (= Mil.rom.); its main source was the German version of the work Universalhistorie alter mittler und neuer Zeiten, dating back to 1794 (= Mil.germ.), and attributed to Wilhelm Ernst Christiani2 ; c) The work Alexander I, Kaiser von Russland. Ein Regierungs-und Karaktergemälde (Berlin, 1814) by Johann Daniel Friedrich Rumpf (= Rum.germ.) was published, just like other laudatory portraits of Tzar Alexander I of Russia in order to counteract the Napoleonic myth and raise the prestige of the French emperor's adversary; the translation into Romanian, Arătarea stăpînirei și a caracterului lui Alexandru I, împăratul a toată Rossia (= Rum.rom.), was published in 1815 in Buda, without mentioning the translator's name3 .The texts mentioned above contain various types of toponyms: hydronyms (names of oceans, seas, lakes, rivers, streams), horonyms (names of continents, countries, administrative divisions, regions, provinces), oikonyms (names of habitats: cities, fortresses), oronyms (names of mountains and hills) and nesonyms (names of islands).Starting from the translation strategies adopted by translators, our aim is to analyse the way in which the toponyms from the mentioned texts were adapted to the formal system of Romanian.

Difficulties in the formal adaptation of proper names in pre-modern Romanian
Although the adaptation of proper names in Old and pre-Modern Romanian is insufficiently studied, it is commonly known that the process is characterized (just like the entire Romanian writing of the time) by the lack of generally valid norms4 .In the absence of any norms5 , the oscillations in the adaptation of proper names are determined by the facts described below.
a) The use of different alphabets The source-texts are written in Latin alphabet.At the time, the Romanian language was written mostly in Cyrillic alphabet.We are thus dealing with two different graphic systems, which means: a different number of graphemes, specific diacritics (with or without phonetic value), the existence of graphemes with multiple phonetic values or, on the contrary, graphemes or grapheme clusters that note the same phonetic sequence.The confrontation of these two graphic systems implies difficulties in choosing a way of transposing proper names into Romanian.

b) Different orthographic systems
The languages involved in the translation of the above-mentioned texts have spellings that are based on different principles: the German orthography exhibits etymological tendencies, while the Romanian-Cyrillic spelling is phonetic.
As far as the principle underlying the Romanian spelling is concerned, specialists in the field have divergent opinions.The "classic" opinion is that the Romanian-Cyrillic spelling is phonetic: "Aplicarea alfabetului chirilic în scrierea limbii române, încă înainte de secolul al XVI-lea și pînă la jumătatea secolului trecut, nu s-a efectuat în virtutea unor norme ortografice fixate în scris, care să fi fost obligatorii pentru toți mînuitorii condeiului, așa încît aceștia-scriind-se lăsau conduși de fonetismul pe care îl deprinseseră în vorbirea limbii; așa se explică și faptul [...] că s-a aplicat ortografia fonetică (s.a.) de la începuturile scrisului în limba română" [The application of the Cyrillic alphabet in the Romanian writing, even before the 16 th century, and up to the middle of the last century, was not carried out based on written orthographic norms, mandatory for all the scriveners; this means that, in their writing, they would be led by the phonetics they had learned through the spoken language; this also explains the fact that phonetic spelling has been used since the beginnings of Romanian writing] (Strungaru, 1976, p. 197).In this sense, Stan (2012, p. 7) finds it symptomatic that, in the Lexicon of Buda (1825), Cyrillic notations were meant to indicate the utterance; therefore, those in Latin alphabet represented the etymological option in Ardeal, while those in Cyrillic characters pointed to the phonology of the word.On the contrary, Boerescu (2014, p. 88) argues that the traditional Romanian Cyrillic writing "este, de la începuturi și pînă aproape de 1830 [...], o scriere bazată exclusiv pe tradiția slavonă și pe etimologismul cultural grec, cu insuficiente adaptări la fonetica limbii române" [was, from the beginning and until 1830, a type of writing based exclusively on the Slavonic tradition and the Greek cultural etymology, with insufficient adaptations to the phonetics of Romanian].This opinion differs widely from that of Strungaru (1976), expressed above; the impression of "phonetic" writing would come from the regularization after 1750.Supporting the idea that the Romanian-Cyrillic spelling was traditional and etymologic, Boerescu (2014, p. 88) relies on the research carried out by Gheție & Mareș (1985, p. 153-164), which shows that the words are used in the first Romanian texts with the values they had in the medio-Bulgarian writing.However, the authors distinguish between religious writings with this feature and the documents of the epoch, characterized by a simplifying spelling and a tendency to leave out conventional uses (Gheție & Mareș, 1985, p. 161); this leads thus to a rather phonetic spelling (although the traditional component is still present).Equally, this was even more valid two centuries later, especially in secular texts.In our opinion, the author exaggerates the etymological character: it can only be correlated with the words originating from Slavonic and Greek, not to inherited or Slavic words, whose writing is phonetic, to which a strong traditional component is added.When a word is written in Cyrillic alphabet, the scribe does not intend to obtain a form that is very close to the etymon, but tries to reproduce the pronunciation, within the limits of the graphic system he uses, with its traditional components.The idea regarding the existence of a fundamental difference between writing before and after 1750 is refuted by the texts used as a case in point for this study; we notice words with multiple values, words and diacritics with no phonetic value, or graphic doublets.Certainly, from the first writings in Romanian dating back to the 16 th century, and up to mid-18 th century, orthography has evolved towards the simplification of some rules, especially as regards the graphemes specialization; however, as our texts show, this development is not significant.We also agree with the idea that the double writing from the Lexicon of Buda (1825) points out to the different principles that the two orthographies of the time (Cyrillic and Latin) are relied on.c) Different phonetic systems Each language has phonemes which do not exist in other languages; this means they need to be approximated in the target-language.In the German sources, such phonemes are, for example, the rounded closed front vowel /y/ (represented graphically in German as ü) and the rounded open front vowel /oe/ (represented graphically in German as ö or oe).

d) The diversity of proper names
The translations dating back to this period often refer to realities that were unusual for the Romanian culture of the time (exotic spaces): Cam.rom.talks about South and North America; Mil.rom. is a universal history; Rum.rom.presents aspects of the history, geography and politics of the Russian Empire.Some of the realities contained in these works and their names were, of course, known to the translator and even to the intended target readership; others were unknown, and therefore the translator cannot relate them to any known system.In most cases, the realities a specific work refers to are outside its linguistic sphere.Consequently, instead of rendering the source-language spelling or pronunciation (for which the name is an exonym), the translator might try to render the pronunciation he understands or assumes in the language corresponding to the universe referred to by the text.e) The existence of traditional pronunciation models ( for the well-known proper names) This limits the possibility of translator's innovations.It is what Garde (1974, p. 4) notices when he speaks about the transcription of French proper names into contemporary Russian: there is a conflict between proper names lacking notoriety, unknown, which can be subjected to a transcription according to unitary principles, and the ones that are well-known, which already have a graphic and/or phonetic tradition in the target-language and must be kept accordingly.

The loan
The loan consists of taking the proper name with the graphic and morphological form as such from the source-language.Ballard (2011, p. 26) calls this strategy report, and defines it as "transferul integral al numelui propriu din textul-sursă în textul-țintă" [the complete transfer of a proper name from the sourcetext into the target-text], "gradul zero al traducerii semnificantului" [ground zero of the translation of the signifier], although he further argues that both transliteration and transcription are report strategies (Ballard, 2011, p. 42).
In the case of Cam.rom., there is sometimes an interesting case of "monolingual bigraphism"; the phrase belongs to Boerescu (2014, p. 87), who uses it to define the coexistence of both the Cyrillic and Latin alphabets in Romanian during 1780-1880.In this paper, the phrase is used to define the situations (specific to this text only) in which the author, transcribing a toponym from the German text, reproduces it in both alphabets.There are three types of situations.In one of them, the original toponyms are reproduced only in Latin alphabet (1): (1) Dominika (Cam.germ., 132) -Dominica (Cam.rom.,78)

Transliteration
Transliteration is the strategy by which the target-language writing is faithfully reproduced in the sourcelanguage, so that each sign in the target-text alphabet corresponds to one in the source-text alphabet (see Elman, 1986, p. 29;Grass, 2002, p. 662-663).Obviously, the strong focus on the graphic level will lead to a phonetic imbalance.
During the pre-modern period of Romanian, the use of transliteration as a strategy of transposing proper names from a foreign language into Romanian posed problems related to the general difficulties of using the Cyrillic alphabet.These originated in the fact that the alphabet was not compatible in all aspects with the phonetic structure of Romanian, which implies the presence of several letters with multiple values (ß, ´, √, ™, å), the fact that the same phoneme can be denoted by several letters (with the same phonetic value: /e/ by ™, e, å; /i/ by √, i, ï, etc.), or the existence of non-phonetic letters in some cases (´or ß); problems are amplified by the lack of general rules for the use of the Cyrillic alphabet, which allowed each translator to innovate within the limits of Romanian graphic system.
One of the problems that occur in the case of transliteration from Latin into Cyrillic points to the difference in number of graphemes between the Cyrillic and the Latin alphabets.This means that, in some cases, the translator may choose from the signs that the Cyrillic alphabet offers.In this case, the very definition of this strategy of translation is questioned, since the same Latin sign can be-and it actually is-transcribed in several ways in the Cyrillic alphabet.Transliteration as such can only concern those cases of 1:1 equation, in one sense or the other, which are scarcely represented; in the other cases, we can talk about approximate transliteration or variants of transliteration.Analysing the Latin alphabet transcription and the transliteration of contemporary Russian proper names 8 , Sakhno (2006,  p. 711) shows that transliteration is not always accurate (besides the fact that it is often combined with the transcription of the same name) and delimits the exact transliteration (Cyrillic signs without any direct equivalent in Latin alphabet are often rendered with the help of diacritics, for example, Gorbačёv) from the approximate transliteration (the diacritics are omitted, as in Gorbacev).Moreover, Bagajewa (1993, p. 350-351) notices the inconsistency of transliteration and the difficulties it poses, including here even the exact loan of proper names between languages using the same alphabet (e.g.phonetic problems posed by the existence of different spelling rules) 9 .Therefore, if this problem characterizes the situation in the current language, it is even more so a characteristic of pre-modern Romanian.Consequently, we also included in this category (transliteration) the situations of partial overlapping (graphic differences occur), as long as they did not affect the pronunciation.
Another problem in identifying the cases of transliteration is that of accent.In the source-texts, the accent is not marked (4); the onomastic dictionaries of the epoch did not always mark it (the proper names of people are very poorly represented in lexicographic works), so it is difficult to determine whether accent in the Romanian text corresponds to that in the source-text In the last three examples, the choice of transliteration involves different pronunciations in the targetlanguage compared to the source-text: s when placed between two vowels = /z/ (Kasan); final -er = /ə/ (Hannover); the graphic group -ie-(Niemen) is pronounced in German as a long i.
As regards the capital letters, it should be pointed out that, at that time, it was used both in the case of proper names, and of other words.Each text has its own features, as follows: a) in Cam.rom., capital letters mark proper names, rarely other words (e.g., Căpetenia, Cam.rom., 8; Gheografia, 10; Observații, 13; Crocodilu, 47); b) in Mil.rom., the capitals consistently mark proper names, but they are more often used at the beginning of other types of words than in the other texts (and, as it seems, randomly); c) in Rum.rom., proper names are written in capital letters; in addition to these, capital letters are used with certain common names that designate public functions or that are neologisms (e.g., Sisteme, Rum.rom.,11).

Transcription
In the case of transcription, attention focuses not on the graphic form, but on the phonetic value of proper names.As a general principle, it should render the pronunciation of the original language as faithfully as possible.The principle of fidelity to the pronunciation of the source-language still constitutes the norm in the case of proper names taken from languages with different alphabets, although it is still a desideratum; see, in this respect, Sakhno (2006), who discusses the differences resulting from the transliteration of Russian proper names and the way in which they "betray" the principle of fidelity to pronunciation.
In our texts, the phonetic transcription of proper names can follow the source-language model (the translator reproduces the name as it occurs in the source-language) or a pre-existing model (the transcription in Romanian does not follow the pronunciation of the source-text, but it is consistent with a 8 The internet is the source of the corpus used. 9See also Lehrer's conclusions (1993, p. 400) regarding the impossibility to create a translation algorithm for proper names. 10The table of contents of Mil.rom. is placed before the text and its pages are marked in Roman letters.
pre-existing pattern of letters/clusters of letters pronunciation).Transcription is often combined with transliteration and phonetic adaptation.
In Cam.rom., the transcription (or transliteration) in Cyrillic letters is sometimes doubled by the replication of the name in Latin alphabet (see above, §3.1); this is not a transliteration [see (5g)], as would be expected, but a bizarre transcription of the original pronunciation (qu = kw; initially voiceless G into k; gu = ku).

Transcription by phonetic approximation
The need for phonetic approximation is determined by incompatibilities between the phonetic systems of the languages that come into contact.Faced with a non-existent phoneme in his own language, the translator approximates it.We will now refer to the phonetic changes that sometimes occur in the transcription process.
The phonetic approximations in the texts with a German source (6) refer to some vowels that the Romanian language system does not have: /y/ (the close front rounded vowel, noted as ü or y) and /ɶ/ (mid-open front rounded vowel, noted oe or ö): (6) Südamerika (Cam.germ.,197)Mil.rom.,13).
In reproducing these vowels, the translators oscillate between i, ü and ¨for ü and y (thus, between vowels and a diphthong: Zíd-Amérika, Lübék, Xa¨tí), and between o and e for oe and ö (Bêócïa, Mê' ris).

A particular case of adaptation: phonetic substitution
We do not refer here to substitution as a strategy of translating proper names by replacing them with an already existing Romanian proper name, for example: Împărăția Turcească for Turcia, Crăia Franțuzului for Franța, Țara Ungurească/Pămîntul Unguresc for Ungaria 11 .We rather take into account what Ballard (2011, p. 42-43) calls "phonetic assimilation", more precisely, a special case that derives from the particular situation characterizing the pre-modern Romanian culture at that particular time.The terminology of Ballard seems here inadequate, as it may create confusion if we relate it to the phenomenon of phonetic assimilation.Ballard also relies on the relationship between French and English; in his view, phonetic assimilation means distinct phonetic variants (for French, English, etc.) of the same proper name.Vaxelaire (2006, p. 720) calls this strategy "adaptation".The situation in the texts discussed here is somehow more complicated, because it leads to the emergence of a third idiom in the equation source-language -target-language.The Romanian culture is characterized, both in the old age and at the beginning of the pre-modern period, by the use of a language of culture other than Romanian; for the old age, this was Slavonic; in the 18 th century, with the Phanariot Age, the Slavonic was replaced by Greek to such an extent that the Greek actually became a tool by means of which culture was circulated within the Romanian Principalities, just as in Transylvania the language of culture was mainly Latin.Under these circumstances, the translator often chooses not to reproduce the pronunciation of the original proper name by transcription, but, especially when it is an exonym in relation to the source-language, replaces it with a familiar variant (one that is accepted, as phoneticism shows, through Greek or Latin), or recovers the pronunciation of proper name from the language corresponding to the reference universe of the text.In another subtype, when transcribing an exotic toponym, for which, on grounds of novelty, there is no equivalent in one of his reference linguistic systems (Romanian, Greek, Latin), the translator does not reproduce the pronunciation from the source-language, but transcribes it according to the pronunciation of the form existing in the reference language of the text.
The translator explicitly states that a given toponym may have different pronunciations.Differences arise either from the specific linguistic patterns, or, as in the case of Athos, from the use of different transfer strategies (transcription or transliteration).

Morphological adaptation
Some foreign toponyms entered the Romanian language through several languages (Greek, Italian, French, German etc.), receiving thus various forms.Others were replaced by their Romanian (formal or denominative) correspondents, established by tradition12 (9), as is the case in the translations under study: (9) Aegypten (Cam.germ., 13) -Eghípet (Cam.rom., 10 In the absence of translated proper names corpora prior to the period we are studying, in some cases it is difficult to determine whether a certain form was established by the translator on the basis of the translated source or it was imposed by tradition.

Gender oppositions
Gender is determined based on both the endings of proper names in the source-text (which often varies) and on the strategy of transposition in Romanian: preservation of the original form, adaptation according to the Romanian toponymic system or the reproduction of pronunciation.The source-texts frequently influence the toponyms ending in Romanian translations; accordingly, different formal gender variants are created, such as: Chíprul (masculine) -Țípria (feminine); Mexico (gender ending non-specific to Romanian) -Mexica (feminine) -Mesic (masculine).

Conclusions
A characteristic of toponyms translation in the studied texts is o s c i l l a t i o n, both as regards the choice of the translation/transposition strategy, and in terms of outcome.The choice of a strategy depends on several aspects: linguistic factors (the incompatibilities between the graphic and phonetic systems, as well as the principles governing the spelling of two languages in contact; the existence, in the case of wellknown proper names, of traditional pronunciation patterns); other factors refer to the translated text (its content and the types of proper names it contains) or the translator's personality (linguistic knowledge and opinions, cultural background).The same toponym (including those very well known) can be transferred in several ways, with multiple results, not just from one text to another, but also within the same text.Sometimes, the translator himself (see Mil.rom.)explicitly indicates that there are several adaptation possibilities into Romanian.
As regards the g r a p h i c and p h o n e t i c a d a p t a t i o n, transcription is the most widely used strategy.There are two cases: a) the transcription is directed at reproducing the pronunciation of the proper names in the source-language; b) the translator attempts to reproduce through transcription the pronunciation from a third language (especially in the case of toponyms that are exonyms in relation to the source-language).Generally, the third language could be one of the languages of culture at the time, one that the translators were familiarized with (in the studied texts, Latin, in its medieval pronunciation), or the language of the reference universe of the text (for example, Russian for Rum.rom., which speaks about the realities of the Russian Empire).Given the lack of overlap between the phonetic systems of languages in translation, translators sometimes operate phonetic approximations, attempting to reproduce the missing phoneme from the Romanian system through one that Romanian possesses.
In terms of m o r p h o l o g i c a l a d a p t a t i o n, the inclusion of a given proper name into a specific gender series is influenced by its ending in the source-language (often fluctuating) and by the adopted translation strategy.Thus, the source-language endings can be: a) preserved (by t r a n s l i t e r a t i o n) or removed (usually resulting a consonant ending); b) s u b s t i t u t e d with endings that are specific to the denominative system of Romanian (for example, German horonyms ending in -en substituted with Romanian ending in -ia); c) a d a p t e d (usually by transcribing the pronunciation).The multiple adaptation strategies and means (different languages) through which a foreign toponym is rendered into Romanian have led, in some cases, to the creation of gender oppositions; they are then reflected in specific pattern of case and number inflexion.As regards the case inflexion, in the studied texts, the genitive-dative endings -ei and -ii are in free variation.

Bibliography
This time, the Latin alphabet form is taken by approximate imitation, with the exception of Kwibio, cf.Germ.Quibio, where the transfer strategy is transcription (see again the different segmentation for Porto Santo).The Cyrillic form transliterates here the Latin one from the Romanian text.