silubreinons bokos:

hvan anakunnand izwis raihtaba?

Some thoughts on ‘Codex argenteus: Lingua gotorum aut lingua gotica’ (language of the Goths or Gothic language) by Graeme Davis [ http://www.shakespeare.uk.net/journal/1_3/davis1_3.html ].

There is a central contradiction between the quality of the manuscript of the Codex Argenteus and what appears at first site to be a linguistic weakness of the text preserved. Wulfila's Bible is a word-for-word gloss of the Greek original. [...] The syntactic incongruities of the Codex Argenteus may be examined through an examination of the problems that the artificial word-order would have caused a reader or listener.

As is well known, the Gothic Bible can’t be used indiscriminately as a source for early Germanic syntax. It displays independence from the original in certain respects, such as case selection (Klein 1992), but generally replicates the idiom of the original language (Greek) more faithfully than Luther’s Bible (Greiner 1992), or indeed most modern renderings. Its word order, in particular, matches that of the Greek quite rigidly (Friedrichsen 1939, p. 1, Streitberg 1920, § 284), albeit with certain regular exceptions, such as the initial placement of the conjunction iþ ‘but’.

But while such literalness might be counted a weakness by our standards, we can’t assume that the Gothic Bible would have been rated a failure, on these grounds, by the standards of its own day. Word-for-word Biblical translation was the norm at this time, a feature the Gothic Bible shares with its contemporaries (Stutz 1966, p. 47-48). Some have speculated that alien diction may even have been considered appropriately mysterious and poetic, satisfying the demand for a text that was cryptic, oracular and in need of expert interpretation from the clergy (Friedrichsen 1926, p. 21, Greiner 1992, p. 97, Anderson 1938, p. 130). Be that as it may, it’s ironic that what two contemporary critics whose words survive objected to was the overly loose rendering, as they saw it, of the Gothic Bible, which for them strayed beyond the “claims of idiom” and took “unwarranted liberties” with the sacred text (Friedrichsen 1939, pp. 268-272).

Just how alien though would the Gothic Bible have sounded to native speakers? Only by close comparison of the relevant texts can we hope to approach an answer, and even then, the nature of the evidence will limit what conclusions we can draw, there being no native speakers to comment on the matter.

Davis quotes the Lord’s Prayer from the Gothic Bible (correctly except that ne briggas should read ni briggais), but all that he offers for comparison is an Old English translation from Latin. Of course, what is really needed here, to assess the independence of the Gothic idiom, is the corresponding passage from the original Greek.

The exact Greek version that the Gothic translation was based on is unknown, although its type can be narrowed down; see, for example, Friedrichsen 1939, pp. 4-5. Some passages show further influence from pre-Vulgate Latin versions of the Bible, and from exegetical Latin commentaries (Friedrichsen 1939, p. 214). These Latin influences, scriptural and exegetic, are particularly strong in the Epistles: “The difference between the extant Ambrosian text and Streitberg’s ideal original stands in sharp contrast with the almost exact correspondence of the primitive Argentean text of St. Matthew with its Greek original” (Friedrichsen 1939, p. 1). A number of different texts are available online and worth consulting. For the Gothic Bible with interlinear Greek text of the Nestlé-Alland edition, see [ http://www.wulfila.be/gothic/browse/text/?book=1&chapter=6 ], Mt 6:9-13. Compare also [ http://www.greeknewtestament.com/B40C006.htm#V9 ]. So, is the Gothic a word-for-word gloss? Almost. What can we conclude?

1) The Greek word-order differs significantly from that which would have been expected in Gothic (as for example shown by the Old English).

It can’t be assumed that the same rules governed 4th century Gothic word order as governed 10th century Old English word order, any more than word order in an Old English text can be judged unnatural because it fails to conform to the rules of continental Germanic dialects of the same period or those of Old Norse from the 13th century. The old Germanic languages share many features, but are clearly not identical in all respects. If late Old English syntax can differ significantly from that of its near contemporaries and geographical neighbours, we can expect it to differ yet more from a language some hundreds of years older, from the other end of the Germanic speaking world.

Nevertheless, comparison with Greek versions does show that the word order of the source is largely replicated in the Gothic translation, albeit with some regular and very frequent differences.

It seems reasonable then to imagine that the word order and idiom of such a translation would have often contrasted with what was most natural in Gothic, and this is implied by the existence of distinct patterns found only in those instances where the translator was forced to use more than one Gothic word to translate a single Greek word. It doesn’t necessarily follow that “problems of comprehension would have been enormous” throughout. This is logically conceivable, but remains to be established. We can imagine that such a style risked impeding comprehension, at least on occasion, but to what degree? Comprehension isn’t a simple issue in a religious text where devotees may differ over subtle matters of interpretation, literal and especially metaphorical, and where groups of specialists each claim to have the one true interpretation. Some parts of the Bible are much simpler in style than others, e.g. the narrative passages and plainly told parables of the synoptic Gospels, as against the metaphysical argumentation of the Epistles, which evidently presented many difficulties of interpretation even in the original Greek. We can also imagine that with the original, or any translation, comprehension would grow with familiarity. But is it likely that anyone would go to the extent of creating a largely incomprehensible translation, or gloss, or that such a text would find widespread acceptance and be preserved in such a state for centuries by a literate community with close connections to Greek and later Latin scholars?

As mentioned above, there’s much evidence of later revision of certain parts of the text, so it seems that Gothic scholars were prepared to alter what they felt was inadequate. If the original translation had been such a failure in their eyes, we might expect them to have made whatever changes they considered necessary. Observe also that the Gothic manuscripts contain marginal glosses, but only occasional ones, perhaps suggesting that this was a text in need of only occasional clarification.

Finally, we should bear in mind the flexibility of word order in the early Germanic languages compared to modern ones, as permitted by their robust morphology (Gabelentz & Löbe 1832, § 288), and as exploited to much greater extremes than this in some native genres of poetry.

None of this proves that no native Gothic speaker could ever have felt confusion when confronted with any part of the text, or even that many Goths might have been baffled by much in it, but it ought to make us wary of assuming the latter case, on the basis of our far-from-perfect knowledge of Gothic word order and idiom, or of imagining that we know better than Gothic speakers themselves what constituted acceptable language in their own Bible.

2) Vocabulary is contrived. For example thiudinassus for kingdom (4 syllables) when a perfectly good two syllable word exists, reiki.

Simply counting syllables in two words can’t tell someone unfamiliar with Gothic (namely you and me and all people now living) that one of those words would have sounded ‘contrived’ in that particular context, and that the other wouldn’t. For that we need context, and the only context we have is the rest of the surviving Gothic Bible fragments where reiki invariably translates Greek ἀρχή (ARHH) ‘power, principality, rule’, e.g. “delivered him into the power and authority of the governor” (L 20:20), while thiudinassus renders Greek βασιλείας (BASILEIAS) ‘kingdom, realm’, except for one instance where it translates ἡγεμονία (hHGEMONIA) ‘reign’ (L 3:1), namely that of the Roman Emperor Tiberius. thiudinassus is a common word in the Gothic corpus and, with that one exception, is used everywhere as here. That cognates of reiki might be appropriate words to express the concept ‘kingdom’ in other Germanic languages hundreds of years and miles distant, doesn’t necessarily mean that reiki was appropriate here, or that thiudinassus wasn’t. I see no reason to suppose that translator and subsequent scribes and Gothic-speaking scholars of failure to chose a suitable word (or to replace or gloss an inadequate one) in this instance.

3) Subordination not natural. The first line needs a main clause followed by a relative clause; the Gothic has two main clauses.

The first line contains no complete main clause in any of the versions under discussion: Gothic, Greek, Latin or Old English. Indeed the first line of the Gothic has no verbs at all, but rather a Determiner Phrase (or Noun Phrase, in more traditional terminology, and according to some theorists), atta unsar, followed by a second Determiner Phrase, þu in himinam, modifying, and in apposition to, the first. This supposedly un-Germanic second part can be paralleled in other old Germanic languages, for example, Old Norse: þá uppi við fjöllin ‘those up in the mountains’ (Gunnlaugs saga 5); kaupmenn af hafi komna ‘merchants [who have] come from the sea’ (Þórðar saga hreðu 2), see Faarlund p. 85, and is of course quite acceptable in some contexts and registers of Modern English, which seems to be Davis’s touchstone.

Oddly enough, Davis doesn’t mention the one point in this line where the Gothic does diverge from the Greek idiom. It may be significant that there are found elsewhere in the Gothic Bible vocatives translated in exactly the same way as this line, with the same divergence from the Greek original, namely with a personal pronoun in place of the Greek definite article, e.g. Mk 9:25, Mt 11:23, Mt 25:41. In a meticulously literal translation like this, such systematic differences stand out. Note that Old Norse, by contrast, does use the definite article before an apposed phrase modifying a vocative: inn fráni ormr, þú gørðir frœs mikla ‘gleaming serpent, you made a great hissing’ (Fáfnismál 19); far þú, hinn góði Urðarköttr, ok hjálp oss við ‘go, good Urðarköttr, and help us’ (Finnboga saga ramma 7); gakk þú hingat, inn mikli maðr, á hólminn ok bersk við mik ‘come here, big man, and fight a duel with me’ (Egils saga 65), which goes to show the danger of making generalisations from a comparison of just two small texts in two languages, and of assuming that what is natural in one Germanic language must be natural in Gothic too.

That said, there are indications that a subordinate clause would be the norm in Gothic in at least some instances where it wasn’t in New Testament Greek. For an example of this, see line 7 (also not mentioned by Davis), where a subordinate clause in the Gothic does indeed contrast with a Greek Noun Phrase. Greek: ‘let us off from the debts of us’. Gothic: ‘let us off from whatever we may be debtors of (i.e. let us off what we owe)’.

5) Germanic theme-rheme order is infringed (eg line 2) because of gloss.

There is no universal “Germanic theme-rheme order” that demands one rigid order of constituents in such a sentence. If this isn’t apparent from the difference between Old and Modern English, compare Old Norse: helgisk nafn þitt ‘hallowed be thy name’ (Hómilíubók); veri þá allir búnir mínir menn at berjask ‘let all my men be ready to fight’ (Magnús saga góða); hjálpi mér svá nú Freyr ok Njörðr ok hinn almáttki áss ‘so help me Frey and Njord and the almighty god’ (Þorsteins þáttr uxafóts 1), as against: en auðræði þín veri í eilífri glatan með þér sjálfum ‘but may your wealth be in eternal perdition along with your self’ (Margrétar saga); guð hjálpi mér ‘God help me’ (Njáls saga 170), etc. And even if there was such a rule that applied to all the better-known Germanic languages, we couldn’t take it for granted that this rule would also apply to Gothic till we’ve looked at the evidence, which is as follows.

Where Gothic is forced to use two words to translate a single Greek verb, it regularly places the optative verb after the object, except with negatives—thus unlike the OE version printed here for this line!—so it may well be that the Greek model has indeed influenced the order in that respect. Again, this illustrates the dangers of drawing conclusions from a superficial comparison of an Old English text without regard to Greek, the source language for the Gothic translation. Not only can we be led into mistaking natural Gothic idiom for unnatural translation artefact, but the same reasoning can lead us to mistake translation artefact for natural Gothic idiom.

6) Verb forms appear wrong. Qimai (line 3) suggests an imperfective sense (which makes a nonsense of the meaning) - it should be perfectivised. Similarly wairthai (line 4)

Not so. Prefixes such as ga- can fulfil a perfectivising role only when attached to a simplex verb which is imperfective. Although ga- can be prefixed to those few Gothic simplex verbs which are perfective, it always in such cases, of necessity, has a different function. As Elizabeth Leiss points out, it would be meaningless to perfectivise an already perfective verb (Leiss 2000, p. 121). The same goes for other prefixes, such as us-, which may convey some additional meaning besides having the grammatical function of perfectivising the verb. Even Wilhelm Streitberg, who pioneered the idea that Gothic verbs are strictly marked for aspect, more or less as in the Slavic languages (except for the lack of a Germanic iterative class), noted that the verbs qiman ‘come’ and wairthan ‘become’ are inherently perfective in meaning (Streitberg 1920, § 296). And indeed, looking at the rest of the Gothic corpus, we see that adding a prefix, ga- or us-, to these particular roots changes the meaning in ways unrelated to aspect: ga-qiman ‘to come together’, us-qiman ‘kill’.

Some have questioned Streitberg’s views on aspect generally. Philip Scherer has presented detailed evidence for a looser system, with many verb forms indifferent to aspect, regardless of whether they’re prefixed with ga-, some other prefix, or lack a prefix (Scherer 1954). Scherer concluded that there isn’t any formal aspectual system in Gothic at all, although some verbs may be inherently perfective or imperfective by virtue of their meaning. Coleman has characterised the situation as lying “between the extremes of Streitberg and Scherer” (Coleman 1996, § 2.2.4), concluding that “[t]hough not strictly systematic, the perfective aspect of the ga- forms could still be activated in appropriate contexts” (Coleman 1996, § 2.2.5). Stutz considers Streitberg’s system not wrong in principle, while conceding that it may require certain modifications (Stutz 1966, p. 49). Leiss is in no doubt that Gothic possessed a formal aspectual system, differing from that of Slavic mainly due to the existence in Gothic of a semantically neutral perfectivising prefix ga-, and the consequent lack of imperfectivising devices (Leiss 2000, pp. 120-128). More generally, Erwin Roedder, who accepted a perfectivising role for ga- in some compounds, saw no evidence for ineptness here—quite the contrary: “Painstaking accuracy at an epoch of the most incisive dissentions in ecclesiastic dogma was the first and foremost requirement for the translator. [...] It is questionable whether most of the Greek laity [...] would without elucidation from the clergy have understood the Greek text any better than the Goth was able to understand Wulfila’s” (Roedder 1937). On the independence of the verbal prefixes from the Greek model, especially ga-, see Rice (1932). Incidentally, the cognate prefix appears here in some continental West Germanic translations and is absent in others.

Since no qualms are shown elsewhere about translating a simple Greek verb with a Gothic verb prefixed with ga-, and since wairthan and qiman are ubiquitously common perfective verbs and are are therefore never found with any ‘perfectivising’ prefix, I see no grounds for declaring these forms ‘wrong’. On the other hand, an accumulating or a lethal kingdom of heaven would be very wrong indeed.

7) Very clumsy translation of give us this day our daily bread : loaf of the day of ours give to-day to us. Two different words are used for day (though the Greek uses only one).

What, in this context, is ‘clumsiness’? How is it measured? The two English versions offered by Davis aren’t particularly literal translations of the Greek and Gothic respectively, and don’t compare like with like. That’s to say, Davis’s pseudo-Greek version is arbitrarily composed in natural English, whereas his pseudo-Gothic version is made artificially arcane. How many words of one language are needed to translate a particular word of another depends on the structure of those languages. Languages often differ if how many words they need to express an idea, so we can’t interpret every instance of a different number of words used in a translation as a stylistic choice on the translator’s part. Greek has a word ἐπιούσιον (EPIOUSION), which is believed to signify ‘daily, necessary(-to-live-on), enough(-for-today/tomorrow)’, and a word σήμερον (SHMERON) ‘today’. Gothic has a word for ‘day’ daga (dative singular), a word for ‘regular’ sinteinan, and a word for ‘this’ or ‘to-’ himma. Is the fact that English happens to say ‘hallowed be’ or nowadays ‘may...be holy’ grounds for calling any translator ‘clumsy’ who might chose this natural idiom to translate a single Greek word? How else could the prayer be expressed? Is the Old English translation, with its sy gehalgod, more ‘clumsy’ than the Gothic, with its weihnai, in this respect? Of course not.

By Davis’s way of reasoning, the ‘artificiality’ of the Codex Argenteus would be unfalsifiable (beyond discussion), since he supports his proposition indiscriminately both with instances where Gothic matches Greek, and with those where it differs. For a description of Gothic syntax, the meaningful distinction to be made is between possible native features (revealed only where the translation differs from the presumed source text in a regular way that can’t be accounted for otherwise), and features that are imitated from the source language, or which can be explained as the result of other influences, such as pre-Vulgate Latin versions of the Bible, Latin and Greek exegetical commentaries, or the influence of parallel passages of the Gothic Bible, which themselves may be accounted for by the above-mentioned factors (Friedrichsen 1939, Streitberg 1920, § 234.2, Stutz 1966, p. 48).

The only meaningful measure is objective comparison of the relevant texts. The results of such study don’t depend on personal judgements about elegance or ‘clumsiness’. Modern English speakers can’t rely on their feel for what sounds right in Modern English as a basis to make such a judgement about another language, least of all a long-extinct one, so little attested as Gothic.

e) Much of the vocabulary of the Codex Argenteus is in the form of loan words from Greek.

This is a misleading statement. The vast majority of the vocabulary is inherited from Germanic, far more than in any Modern English Bible I’ve seen. Of 64 Greek words (including Hebrew loans) examined by Friedrichen which were adopted by the Latin Vulgate, only 28 appear in the Gothic Gospels (Friedrichsen 1926, p. 37, Heather et al. 1991, pp. 148-53); the rest were translated with native Gothic words. As Dunstan comments, reviewing Friedrichsen’s work in Modern Language Review 22:2 (1927), we have no means of knowing for sure how many of the loans from Greek which do appear in the Gothic Bible were borrowings on the part of the translators, and how many were adopted before the translation, but “The Gothic version seems to avoid the introduction of Greek words as far as possible.”

Greek lexical input comes mainly in the form of proper names, as in any Bible, and a small number of technical terms associated with the Christian or Jewish religions, some of which are originally Aramaic. In the Lord’s Prayer, for example, there appears only one loanword from New Testament Greek, amen, which is itself a loanword in Greek, limited to this context of marking the end of a prayer. In addition, there might be one naturalised loanword, atta ‘father’, whose immediate source is hard to judge since it’s widely found in many languages.

The Gothic translation shows great freedom in its choice of vocabulary items and prefixes (Rice 1932), and follows its own rules of case selection (Klein 1992). The Gothic divides many lexical fields along quite different lines to the Greek, losing some distinctions, but introducing others not present in the original (Stutz 1966, p. 78). For example, there’s nothing in Gothic corresponding to the Greek tendency to use πρὸς (PROS) ‘to’ with animate complements, εἰς (EIS) ‘to’ with inanimate ones: iddja du imma = ERQETAI PROS AUTON ‘went to him’ (J 6:5); galaiþ du garda seinamma = APHLQEN EIS TON OIKON AUTON ‘went home’ (L 1:23). Likewise, there’s nothing in Greek to motivate the Gothic rule that verbs of motion should be used with the preposition du ‘to’ except for qiman ‘to come’, which patterns like verbs of stasis, being used with the preposition at ‘to’ (or with verbs of stasis ‘at’): iddjedun du imma = HRCETO PROS AUTON ‘came to him’ (Mk 2:13); qemun at imma = HLQON PROS AUTON ‘came to him’ (J 10:41).

The extent to which the translation may contain calques coined for the occasion is unknown, but these would presumably be confined to novel concepts, and unlikely to play much role, if any, in the Lord’s Prayer with its simple diction.

References

Anderson, G. K. (1938) ‘Some notes on Gothic syntax’, Germanic Review 13, 130-138.

Coleman, Robert (1996) ‘Exponents of futurity in Gothic’, Transactions of the Philological Society 94:l, 1-29 [ http://www3.interscience.wiley.com/journal/120146493/abstract ].

Faarlund, Jan Terje (2004) The Syntax of Old Norse. OUP.

Friedrichsen, George Washington Salisbury (1926) The Gothic Version of the Gospels: A Study of its Style and Textual History. London: Oxford University Press.

Friedrichsen, George Washington Salisbury (1939) The Gothic Version of the Epistles: A Study of its Style and Textual History. London: Oxford University Press.

Gabelentz, Hans Georg Cronon von der & Löbe, Julius. 1832. Ulfilas. Leipzig: Brockhaus.

Greiner, Paul (1992) ‘Tempted by original syntax: Luther, Wulfila and the Greek New Testament’. In: On Germanic Linguistics: Issues and Methods. Trends in Linguistics: Studies and Monographs 68, Irmengard Rauch, Gerald F. Carr, Robert L. Kyes, 97-108. Berlin, Walter de Gruyter.

Heather, Peter J. and Matthews, John Frederick and Matthews, John (1991) The Goths in the Fourth Century. Liverpool University Press.

Klein, Jared S. (1992) ‘On the idiomatic nature of the Gothic New Testament: A comparative study of prepositional usage in Gothic and New Testament Greek’, Transactions of the Philological Society 90:1, 1-80 [ http://www3.interscience.wiley.com/journal/119984567/abstract?CRETRY=1&SRETRY=0 ].

Leiss, Elizabeth (2000) Artikel und Aspekt: Die grammatischen Muster von Definitheit. Studia Linguistica Germanica. Berlin: Walter de Gruyter.

Marchand, James W (1973) The Sounds and Phonemes of Wulfila’s Gothic. Mouton: The Hague.

Rice, Alan Lake (1932) ‘Language Dissertation No. 11: Gothic Prepositional Compounds in Their Relation to Their Greek Originals’, Language, 8:4, 7-142 [ http://links.jstor.org/sici?sici=0097-8507%28193212%298%3A4%3C7%3AGPCITR%3E2.0.CO%3B2-9 ].

Roedder, Edwin (1937) ‘Gothic gasaihwan: A study in Germanic synonyms’, PMLA, 52:3, 613-624 [ http://links.jstor.org/sici?sici=0030-8129%28193709%2952%3A3%3C613%3AGGASIG%3E2.0.CO%3B2-L ].

Scherer, Philip (1954) ‘Aspect in Gothic’, Language, 30:2, 211-223 [ http://links.jstor.org/sici?sici=0097-8507%28195404%2F06%2930%3A2%3C211%3AAIG%3E2.0.CO%3B2-4 ].

Streitberg (1920) Gotisches Elementarbuch, Heidelberg: Winter.

Stutz, Elfriede (1966) Gotische Literaturdenkmäler. Stuttgart: J.B. Metzlersche Verlagsbuchhandlung.