The Greek SourceS of The GoThic BiBle TranSlaTion

Almost all of what we know about the structure and properties of Gothic comes from the Gothic translation of the New Testament from Greek. No analysis of Gothic syntax is therefore feasible without reference to the Greek original. This is problematic, however, as the autograph that was used in translating the Bible into Gothic does not exist, and the choice of the Greek edition of the New Testament for comparative study is a matter of debate. The article argues that, in spite of the general structural affinity of the Gothic text to the Greek, the numerous observed deviations from the Greek represent authentic properties of Gothic—it has been argued in the literature, based on such deviations, that Gothic is an SOV language. A comparison of the Gothic Bible and different versions of the Greek New Testament gives a taxonomy of structural and linguistic differences. Based on this, I argue that the correct version of the Greek Bible to use when analysing the structural properties of Gothic is the Byzantine text form, represented by the Majority Text of the New Testament.

not much is known about the circumstances of Wulfila's life, but the few facts that are known, handed down to us by ancient historians Philosturgius, sozomen, Theodoret, Auxentius and others, can help to reconstruct a reasonably clear general picture. Wulfila was not an ethnic Goth, and is usually said to have belonged to a cappadocian family, who were taken captive and carried away following a Gothic raid. he was, however, a Goth by upbringing who spoke Gothic as a native language 4 , and was educated in christianity as well as the classical languages, especially Greek. it is possible that he obtained his education in constantinople, although nothing is known definitely about his schooling. At some time around 341, or possibly earlier, he was ordained Bishop of the Goths, but his missionary work among the Goths may have started well before his episcopal office-Wulfila is generally credited with converting Goths to christianity. This daunting task was achieved, among other things, by providing his nation with christian literature in the form of a Gothic translation of the new testament. significantly, Wulfila himself was a devout Arian, and an Arian form of christianity was what his countrymen came to embrace.
the Gothic BiBle As A tRAnslAtion: MotiVAtinG A linGuistic study Gothic is the earliest attested Germanic language, a statement that rests not only on the physical nature and relative historical value of the surviving manuscript, but one that can be corroborated on the basis of linguistic evidence, particularly the various linguistic archaisms that are unattested elsewhere in Germanic. not all evidence is equally trustworthy in terms of its representativeness however, and its reliability can be roughly graded on a cline of authenticity, with phonology at the top and syntax at the bottom. Morphology occupies an intermediate position, with its different aspects testifying to different degrees of representativeness. so while Gothic inflectional morphology is generally more authentic, the authenticity of numerous aspects of derivational morphology is less certain.
This linguistic ambivalence of Gothic is due to the fact that most (if not all) of attested Gothic evidence is a translation. This is a well-known fact, and while the precise extent to which Gothic imitates Greek (including correspondences and deviations on every level of grammar) is yet to be established, the philological nature of the translation has often been discussed. Wright (1954, 380) places the Greek original in constantinople as the eastern (syrian/Antiochian) version of the new testament, and the syrian/Antiochian types are also advocated by Metzger (1977, 384-385), who provides a number of further references in support of this view.
While there is no doubt about the basic source text of the Gothic translation being Greek 5 , it has also been conventionally assumed that the Gothic text has been influenced by old latin (pre-Vulgate latin translations of the Bible) 6 . The way in which elements of latin have entered the Gothic text is not entirely clear-it is important to note that this view is directionally biased in its a priori assumption that there was one or several latin versions that have lent aspects of their language to Gothic. Various assumptions have been made on the relation between latin and Gothic, ranging from Wulfila himself referring to an old latin version (i.e. in line with the above-mentioned directional bias) to certain latin versions having possibly been brought into line with Gothic. it is more than likely, however, that the latinate element in the Gothic Bible is all a question of textual transmission, and in particular the history of the surviving Gothic manuscripts 7 , none of which can be attributed to the hand of Wulfila. Thus Metzger (1977, 386) concludes that any old latin elements observed in the Gothic version are post-Wulfilian.
Wright (op. cit.) cites streitberg (1919/2000), who attempted a reconstruction of the Greek text, as well as offering an explanation for why Gothic differs from Greek: any deviations from the original are due to the influence of latin scriptures or parallel passages from a range of Greek texts in existence before the fixing of the canon. While this may be the case to some degree, it would be a mistake to overlook language-internal factors in explaining the differences of a translation from the original. indeed, as will be argued later, some aspects of grammar which define the Gothic text as different from Greek are unique archa-5 including fragments of the Gothic old testament, which are believed to have been trans-including fragments of the Gothic old testament, which are believed to have been translated from a lucianic text. 6 for an overview of basic ideas on the latin component in the Gothic Bible, see Metzger (1977, 385-386). lampe (1969,(346)(347)(348)(349)(350)(351)(352)(353)(354)(355)(356)(357)(358)(359)(360) offers an especially useful account on the various possible sources of latinate influence. for a dedicated account of how latinate influences are reflected in the use of prepositions in Gothic, see yoshioka (1986). 7 There is no doubt about the origin of the Codex Argenteus, which was produced in italy. ic properties of Gothic which cannot be accounted for in terms of any external factors. At any rate, the above-mentioned linguistic ambivalence of Gothic is all a question of the extent to which different linguistic phenomena can survive in translation. As a result, phonology is the most unaffected level, while any 'creative' aspect of the language (including derivational morphology and syntax) is likely to follow the original.
Thus the notion of Wulfila's Bible being a translation should not be underestimated: a fairly superficial glance at Gothic and Greek suffices to make it clear that, rather than being a translation in the conventional sense of the word, Gothic actually glosses the original to the extent that the result is an accurate structural calque of the Greek new testament. Any structural affinity, however extensive, is not absolute, and an examination of the translation reveals a number of internal peculiarities and variations. While these language-internal variations may be linguistically useful insofar as they may represent genuine archaic features significant for internal reconstruction, some structural variations may actually represent certain ad-hoc 'innovations' which are the diametrical opposites of true archaisms. These innovations are calques proper and should be understood as violating the structural conventions of the language. The question which naturally arises at this point is whether there is room for Gothic to display any authentic characteristics, at least with regard to morphology and syntax, and whether these so-called 'archaisms' are not all a result of calque-induced sporadic variation in glossing.
in a recent article, Kejdan (2006, 149-151), building on historical and philological investigations by simonetti (1976) and Gusmani (1968), attempts to answer this question by appealing to the stylistic and literary canons of Wulfila's time, and in particular, the environment in which Wulfila developed as a literary creator. Wulfila was born into a barbaric environment of Goths, but received his education and made a clerical career in constantinople. he was a devoted Arian who may be rightfully placed at the top of Arian intelligentsia, which has various ramifications, including-importantly-an ideological adherence to the canons of Atticism. Kejdan argues that, as an Arian Attic orator, Wulfila honoured the Attic cult of antiquity by developing for Gothic-a barbaric language with no literary or philosophical tradition-an elevated literary style.
Kejdan fails to clarify, however, in what way the assumptions that underlie this argument constitute a strong case to analyse the structure of Gothic or how such generalisations follow from the data; nor is it clear how this would help explain instances of attested linguistic variation, either translation-induced or internal. As a person who undertook the task of translating the Bible into a barbaric language, Wulfila was, beyond doubt, an innovator. The linguistic reality of the Gothic Bible, however, is a near-wholesale importation of Greek presented in Gothic guise. Thus, while the cross-cultural circumstances of Wulfila's life certainly illuminate his motivations for preaching the Gospel in the language of his nation, the Arian or Attic argument does not present a real challenge to comparative evidence. Wulfila's slavish adherence to the original of the new testament is evident upon comparison of his rendering with the Greek text. it can be explained as having been motivated by a compulsion to reproduce as faithfully as possible what he ultimately perceived to be a word of God.
let us now examine some instances of grammatical difference between Greek and Gothic in an attempt to motivate a linguistic study of Gothic syntax as a valuable resource for our understanding of early Germanic. consider the following sentence from luke 5:28: one obvious shared trait of the two above sentences is their structural equivalence, suggesting that what the translator was aiming at primarily was not to stray from the sequence of elements as they were presented in the original. however, there are at least two clear instances of deviation here: one morphosyntactic and the other lexical with syntactic ramifications. first, unlike the Greek counterpart, the Gothic verb bileiþan 'to leave' seems to require a dative (rather than an accusative) complement; second, the Gothic translation manifests a clear case of lexical insufficiency: what is a single lexeme in Greek (i.e. the verb 'to follow') is represented by two lexemes in Gothic. however, in spite of this, Gothic does not violate the principle of structural equivalence to Greek as any instances of such lexical 'excesses' are contained within the 'slot' assigned for the lexeme in the original. in other words, while slot-internal syntax is enriched in Gothic, it does not have an impact on the overall structural arrangement of the sentence 8 .
jasanoff (2004/2008, 904-905) points out another, and a more useful, instance of lexical insufficiency. in some instances, what is a single verb in Greek is rendered in Gothic by means of a verb and an object, with the object pre-posed to the verb: wrakos winnand 'will suffer persecution'). The opposite word order is observed when the object is a pronoun. such instances of lexical insufficiency are useful because they provide an insight into the syntactic type of Gothic. What is perhaps even more important is that both deviations from Greek in case assignment and lexical insufficiency constitute phenomena that are almost certainly manifestations of authentic Gothic usage rather than externally-influenced or internally-induced parallels.
The taxonomy of how Gothic differs from Greek does not end here. one other grammatical feature particularly relevant for the purposes of the current study is the use of the double adjective inflection, a characteristic unique to Germanic. Although inconsistencies and variations both in the syntactic use and semantic motivation of one inflection over the other are frequent, the use of the adjective in Gothic provides some very strong evidence for the reconstruction of aspects of the proto-language.
to sum up, in spite of the fact that the Gothic text is a faithful copy of the Greek new testament, the linguistic significance of various morphological and syntactic deviations from the original should not be overlooked, and presents a valuable source of evidence on the structure of Gothic and, by extension, early Germanic. While some instances of such deviations may seem more 'gothicky' than others, the number and variety of Gothic deviations from the Greek new testament provide a good launch-pad for the study of various aspects of Gothic morphology and syntax.
the PRoBleM of GReeK BiBlicAl texts: histoRy And context Perhaps the most acute problem in the study of Bible history and transfer is the fact that there is no single original text, and the thousands of manuscripts of the Greek new testament that exist are all copies, with most if not all being copies of copies. Well over 5,000 Greek Bible manuscripts exist, and although this impressive number certainly speaks in favour of a reliable attestation of the new testament, the problem with it is that no two manuscripts are likely to be entirely identical.
Most of this vast number of manuscripts are generally attributed to the Byzantine family (also known as the Byzantine text type), and although there exists some degree of variation between any two manuscripts, most of the variant readings are of a minor kind. hence Metzger (1992, 186-206) attempts a typology of the variant readings in terms of unintentional changes (including errors which result from faulty eyesight, hearing, errors of the mind, errors of judgement and the like) as well as various intentional changes (see also Royse (1995) for various related problems). The numerous surviving manuscripts have been juxtaposed, and with inconsistencies and errors in the different manuscripts thus eliminated, have resulted in the Majority text, which, as its name suggests, stands for the most faithful and statistically representative version of the Byzantine text type. Thus, the Majority text version does not coincide with any one surviving manuscript text, but rather aims to represent the entire family in as consistent a way as possible.
one other version closely related to the Majority text is the Textus Receptus (the Received text). like the Majority text, it does not duplicate any single attestation of the Greek new testament but is derived from a small number of Byzantine manuscripts-hence the similarity between the Majority and Received texts. however, even though the two versions are close in many ways, a considerable number of differences exist between them.
The authorship of the Received text is conventionally attributed to the 15 th /16 th -century dutch humanist and theologian desiderius erasmus (Aland and Aland 1987, 3-6), who used for his version of the new testament a handful of Byzantine manuscripts of the 12 th /13 th centuries (far from the best manuscripts of the type). Thus, while the Received text is clearly irrelevant for our investigation of the Gothic Bible, it is a significant work in that it served as the basis for the translation of the new testament into various european languages at the time of the Reformation, including the english authorised King james Version and the German luther Bible.
in seeking to determine what more accurately represents the original scripture, textual critics have invested time and effort in assessing the authenticity and relative value of the different text types, and the one believed by many to be closer to the autograph of the new testament is the Alexandrian text type. Reliance on this text type has resulted in the development of what is known as the critical text, produced in the 19 th century by the cambridge professors B. f. Westcott and f. j. A. hort. The standard modern equivalent of the received text, which incorporates the readings of Westcott and hort, is the 27 th edition of nestle- Aland (nestle et al. 2001), also known as Novum Testamentum Graece. nestle-Aland's new testament has enjoyed a very wide circulation, and while widely regarded as the standard text for scholarly study, it has also served as the basis for numerous translations of the Bible carried out throughout the 20 th century.
The chief motivation for favouring the Alexandrian text type is its age, with some manuscripts going back to the third century A. d. (the autograph(s) of the new testament-none of which survive-are assumed to date from the first century A. d.). The Byzantine family texts, by contrast, are post-fourth century. nevertheless, many text critics have produced various arguments against relying on the Alexandrian type, ranging from inadequate manuscript copying practices in egypt (which is where the Alexandrian manuscripts come from), to the quantitative predominance of the Byzantine text type as well as its wide distribution and acceptance. Among the staunchest critics of adherence to the Alexandrian text and proponents of the Byzantine version are the text critics Robinson and Pierpont (2005: iii-xv, 533-586), who argue in favour of the Byzantine Priority hypothesis and suggest that the Alexandrian text type is a result of 'an early localised recensional attempt to purge and purify the alterations found among the Western manuscripts.' This, combined with the use of manuscripts of inadequate quality, had the effect of corrupting the text: while various Western influences failed to be corrected, the original readings were compromised. The outcome is 'a textual patchwork that within numerous verses finds no support among any extant document, even over relatively short segments of scripture' (2005, iv). The Byzantine text form, Robinson and Pierpont argue, circumvents many such issues and is a consistent representation of what was, from the fourth century and onwards, the dominant version of the scripture; the combined testimony of a plurality of manuscripts is a more plausible representation of the archetype than any minority or isolated resource, however old.
Whereas attempting to defend the primacy of either text type is outside the scope of this article, knowledge of the context of new testament origins and text criticism is significant in assessing the nature of the Gothic Bible as well as the resources conventionally used for comparative study. This warrants a brief consideration of one other version of the Greek new testament, streitberg (1919/2000)-a version that, although virtually unknown to Bible scholars or classical linguists, has served as a readily available source of the Greek new testament for Germanists.
in addition to providing what is a classic edition of the Gothic text, streitberg supplies parallel passages of Greek. similar to the Byzantine and Alexandrian text type-based versions of the new testament, streitberg's version does not derive from any single manuscript, but is rather a hybrid reconstruction of his own. The problem with it is that is does not coincide with either the critical or Majority texts, and even though it has emerged from my comparison of the streitberg text with nestle-Aland and Robinson and Pierpont that streitberg largely relies on the Byzantine version, there is no absolute correspondence between the Majority text and his proposed reconstruction, with various kinds of deviations observed. The point is that streitberg was not a Bible scholar, and his motivations and methods for positing his own version of the Greek new testament should therefore be approached with great caution-more so because any observed differences between the Majority text and streitberg favour the shape of the Gothic text 9 . coMPARAtiVe scholARshiP on Gothic And GReeK Problems with using streitberg's Greek text have been noted in the literature. in her recent book, ferraresi (2005, 2, 5) relies on streitberg's edition throughout, but seems aware of its problematic nature (at least with regard to differences between streitberg and the critical text). As a result, she elects to supply readings from nestle-Aland where they are different from streitberg's. This may still be seen to be a questionable tactic as it automatically assumes (whether consciously or not) the primacy of streitberg's Greek rendition over the Majority text, which is accepted as standard in biblical scholarship. streitberg's edition seems to have had the most impact on the comparative study of Gothic and Greek throughout the century, with potential ramifications 9 Aland and Aland (1987, 206) suggest that any differences between streitberg's Gothic and Greek are arbitrary, which is a consequence of his hypothetical reconstruction. They also propose that a reconstruction of aspects of the Greek new testament based entirely on what is observed in Gothic would be a useful addition to biblical scholarship as Gothic might thus help shed light on the status of the mid-fourth-century Koine text. Wegner (1999) expresses reservations in this regard due to supposed coptic influences on the Gothic translation.
for research results. Among the studies based on his edition of the Greek text (whether streitberg is acknowledged explicitly or is taken for granted without comment) are Metlen (1932) on participial forms; Rice (1932) on prepositional compounds; Klein (1992b) 10 on aspects of syntax including interrogativity, complex sentences, tense, mood and diathesis; Klein (1992a) on the use of prepositions; Berard (1993) on the configurational properties of Gothic; coleman (1996) on means of expressing futurity and fertig (2000) on zero subjects.
it must be noted that most of these studies are on aspects of language structure which are outside the domain of pure word order variation, and any results are thus minimally affected by the use of streitberg. in fact, the same is likely to be the case if nestle-Aland is followed for Greek evidence where word-order is outside the scholar's focus. it is therefore unsurprising that fertig (2000, 4) finds hardly any difference between streitberg and Marshall (1974) in this regard, the latter edition adopting nestle's text. streitberg also seems to be the preferred text in dawson (2000)-again, with no explicit explanation of her motivations, which is regrettable given her use of examples that differ between streitberg and the critical text (for example Mark 14:47, john 18:10, Matthew 8:32, Mark 15:45).
A diametrically opposite stance is taken by Thomason (2008, 373-374), who heavily advocates the critical text: '[t]he nestle & Aland edition should have been made the primary consulting source for the Greek new testament [...]'. Thomason's holistic adherence to the critical text is also reflected in her Phd thesis on prepositional systems (Thomason 2006), where the Greek component relies on nestle-Aland's text entirely. nestle-Aland has also been used as a source of Greek evidence in Goetting's (2007) more recent study on complex verbs with pleonastic prepositions, though no justification is given for the preference.
A discussion of a comparative study of Gothic and Greek calls for considering another closely related issue, and namely the question of the idiomatic nature of attested Gothic. two mutually exclusive trains of thought are at cross-roads here: the idealists and the agnostics. The former view is perhaps best represented by curme (1911), who argues extensively for Gothic as entirely idiomatic. At the centre of his argument is the notion that the structure of Gothic naturally conforms to that of new testament Greek, and the structural affinity of the Gothic and Greek texts is therefore unsurprising. This structural equivalence, he claims, is a result of the two languages being 'in the same state of historical development' (1911,158). Although the limitations of such an a priori approach along several dimensions are apparent (if not comical) in view of the vast differences between the grammatical structures of Germanic and Greek, views similar to curme's can still occasionally be heard in the academic community.
The agnostic view of the usefulness of the Gothic data, and namely the opinion that any judgements on the grammatical structure of Gothic are either impossible or questionable, has been a common cliché in Germanic studies. Bennett (1980, 127) thus writes with reference to both the skeireins extract and the main Gothic text: 'Both documents, in fact, contain so many Greek syntactic features that they are all but useless for the study of Germanic syntax. ' Metlen (1932), who, incidentally, provides a useful critique of curme's (1911) article, comes to a conclusion similar to Bennett's based on a comparison of Gothic and Greek participial forms: Gothic present participles imitate Greek ones qualitatively and quantitatively, with little or no authentic idiom observed in the Gothic Bible.
As is often the case, the truth probably lies somewhere in between the extremes, and the view advocated here is similar to that adopted in most of the other above-mentioned studies of aspects of Gothic grammatical structure: while Gothic undoubtedly presents in itself a close copy of the Greek new testament, the intrinsic differences between the two languages result in Gothic deviating from the source text in various ways; these variations serve as a useful point of reference for a careful assessment of the features that may represent authentic attestations of Gothic.
Which GReeK BiBle is ReleVAnt foR Gothicists?
As has been suggested above, there does not seem to be a consensus among Germanists as to which Greek version of the Bible to use for comparative purposes. While it has been traditional to rely on the 'reverse' reconstruction of the Greek text by streitberg (1919/2000), such reliance is arguably unsound given that, having been reconstructed to match the Gothic text where no other editions of the Greek Bible do, streitberg's text prejudices the objectivity of comparative study.
in spite of the fact that there is no such thing as an autograph of the Greek Bible or indeed a manuscript that Wulfila used in the process of translation, it is possible to conjecture an answer to the question of which Greek Bible compares to the Gothic text most favourably. After all, the choice is between only two versions recognized by current biblical scholarship: the critical text, represented by nestle- Aland's 27 th edition (2001), and the Majority text, represented by Robinson and Pierpont's (2005) edition.
While the Greek texts are, largely, very similar, an examination of my Gothic evidence against the two versions of the Greek text has revealed a striking picture of differences between the Greek versions as well as different degrees of propinquity to the Gothic version. The following are the three most frequent types of difference between the critical text and the Majority text. The above illustrates clearly that Mark 7:8 as presented in the Majority text contains a clause which is not found in the critical text. like the Majority text, the Gothic Bible incorporates the clause in identical wording.

(ii) Word order
syntactic differences between the Greek versions include various permutations of word order, from differences between the ordering elements within a noun phrase to more radical restructuring at clause level. compare the following fragment from Mark 11:18 across the three versions: Whereas the above Greek examples share the same elements, they are different in the arrangement of the coordinated nouns archiereis and grammateis. unlike the critical text, the Majority text has the two words in inverse order, but the same order as the one found in the Greek Bible. one difference between the Gothic and Greek is that Gothic has an attributive adjective in the final noun phrase. This is, however, no great innovation, as it merely spells out what in Greek is a compound (archi-'first, chief ' and iereos 'priest'). What may be of some significance for the study of Gothic syntax though is that our Gothic translator chooses to place the attribute in post-position to the head noun! As such patterns may represent authentic Gothic syntax free from Greek influence, they merit separate thorough examination, even if the translation of Greek compounds into Gothic is outside the scope of this article.

(iii) Word forms
This category represents various differences between the Greek texts on lexeme level, including the occasional use of different lexemes or, sometimes, different morphological forms of the same lexeme. consider the following fragment from luke 9:35 across the three versions: The difference between the critical and Majority texts above is in the use of post-nominal attributes. Whereas the critical text has the participle eklelegmenos, the Majority text has the adjective agapetos. Again, Gothic here presents evidence of siding with the Majority text as it translates the attribute as liuba 'beloved'.
conclusions The current investigation is by no means an attempt at comparative textual criticism of the Bible-a world of scholarship in its own right-nor does it purport to present anything like a comprehensive picture of the differences between Greek biblical texts. however, my comparative examination of Gothic adjectives against Greek has highlighted certain differences as being more frequent and significant and, moreover, has suggested that whatever the type of difference between the Greek texts, Gothic clearly tends to adhere to the Majority version. Accordingly, where the critical text is different from Gothic, the Majority text may be found to endorse the Gothic version; however, where the Majority text strays from Gothic, there will be no consensus between Gothic and the critical text. of the three basic types of difference, word order differences are by far the most frequent ones, followed by omissions and differences on lexemic level.
Beyond any doubt, there are instances of Gothic deviating from Greek in various ways, and it is these instances that should be examined in an effort to establish what may represent authentic characteristics of Gothic grammatical structure. interestingly, most instances of such Gothic deviations from Greek occur where the Greek texts agree, which lends further support for the idea of our translated Gothic displaying authentic gothicisms.
Whatever the nature and ramifications of these deviations, one thing appears to be certain: Gothic follows the Majority text much more closely than the critical text, and the former should thus be used as a point of reference in the comparative study of Gothic and Greek.