‘A Thing Like God’: Re-Reading Gothic Philippians 2.6–8

Abstract

The Gothic translation of Phil 2.6–8 differs from the Greek in three ways: it says that Christ did not think it robbery to be ‘like God’; it breaks the parallelism between the ‘form of God’ and ‘form of a slave’; and it states explicitly that Christ was obedient ‘to the Father’. Scholars have focused exclusively on the first element, crediting it to the Homoian ‘Arian’ prejudices of the translator, Wulfila, or to his opposition to modalist tendencies in pro-Nicene thought of the 340s. Neither interpretation is satisfactory, the first because the Gothic displays no generalised Homoian bias, the second on philological grounds. When the passage is viewed as a whole, an explanation can be found in the history of exegesis. Homoian churchmen, who followed a theology close to the elderly Wulfila’s, seem to have construed ἁρπαγμός (Gothic wulwa, English ‘robbery’) as res rapienda, in the typology developed by N.T. Wright. Christ did not ‘seize’ equality with God. Incompatible with this view, the Gothic is a better fit for res retinenda (Christ did not ‘hold fast’ his divine status). In an ancient analogue to modern ‘functional equivalence’, it is representing the meaning of the text, as agreed among Greek exegetes, on the translation’s surface. Just why Wulfila did this remains obscure: certainly to clarify the passage’s Christology, but possibly also to head off misinterpretation in his Gothic context. Either way, the Gothic text shows a more flexible approach to translation than scholarship, still focused on stereotyped ‘Arianism’ and lexical equivalence, has yet recognised.

5 Let that, then, be thought among you, which also [was] in Christ Jesus,

6 who, being in God-splendour, did not think it robbery for himself to be (a thing) like God/to be in like fashion to God,

7 but emptied his own self, taking the aspect of a servant, becoming (8) in the likeness of men, and, being found in appearance as a man,

8 humbled himself, becoming obedient to the Father unto ….

Gothic Phil 2.6-8¹

1. Introduction

Three verses remain, almost complete, from the Gothic translation of the Christological hymn in Phil 2.6–11.² The Gothic version is based on a Byzantine text that includes other (primarily ‘Western’) readings,³ and these verses, like the rest, are of ancillary interest for New Testament textual criticism. For the history of theology and of exegesis, they are more significant. The Gothic version is not simply an early example of something that has, over the last century, become common: a rendering of the Bible (here with attendant invention of a new script) into a language without a prior literary tradition. Preserved in several manuscripts, all fragmentary and mostly of the New Testament, it is an important textual remnant of a now extinct branch of Christianity.

According to fifth-century church historians, the translator was a bishop of mixed Gothic and Cappadocian descent.⁴ His name is recorded as ‘Ulfila(s)’ or Οὐλφίλας; scholars commonly use the presumed underlying Gothic form, ‘Wulfila’.⁵ Ordained by Eusebius of Nicomedia, the most important of Arius’ allies, he was linked from the beginning to the so-called ‘Arians’.⁶ At a synod held in Constantinople in 360, he endorsed a creed, promulgated in 359 following double councils at Ariminum (Rimini) in Italy and Seleucia in Isauria, that declared the Son to be ‘like (ὅμοιος) the Father’ and forbade theological use of the term οὐσία.⁷ This ‘Homoian’ creed would remain the profession of the imperially recognised church until the council of Constantinople in 381.⁸ Thereafter, it was used by churches that consciously rejected Nicaea and asserted the ontological subordination of the Son, as a secondary God, to the Father, and of the Spirit, ‘neither God nor our God’ (as the elderly Wulfila is said to have professed), to the Son.⁹ Wulfila’s own, Christian Goths had settled on the Roman side of the Danube, following persecution in the 340s.¹⁰ His translation was probably taken up by other Germanic-speaking groups, and was presumably still in use within the Homoian churches of the Western Roman Empire’s successor states down to the adoption of the Niceno-Constantinopolitan creed by the last ‘Arian’ kingdom, the Spanish Visigoths, in 589.¹¹ The Gothic Bible and the history of Homoian Christianity are inextricably intertwined.

Since the first edition of the text of Gothic Philippians in 1835, one verse – indeed, one word – has seemed proof that Wulfila’s translation was shaped by his theological prejudices. Christ, according to Gothic Phil 2.6, ‘thought it not robbery’ to be like God, rather than equal to God. Apart from strictly linguistic questions, that one word (galeiko, in the Gothic) has absorbed virtually all scholarly attention to the passage. It has been explained in two ways. The traditional view finds in galeiko a direct reflection of the Homoian creed of 359/360. Wulfila’s translation was a disingenuous misrepresentation, meant to deceive Gothic converts. The alternative, developed by an eminent historian of the Gothic church, the late Knut Schäferdiek, found the explanation, rather, in the accepted orthodoxy of the 340s and early 350s. In avoiding closer analogues of the Greek ἴσα, Wulfila meant, like contemporary Eastern synods, only to deny that the Son was identical with the Father or that both were modes of the One God.

Neither interpretation, we will argue, does justice to the Gothic translation. As is clear from the provisional translation offered above, the Gothic differs markedly from the Greek, in ways that extend well beyond the choice of a word apparently meaning ‘like’ or ‘alike’ to represent the Greek ἴσα. Taken as a whole, the extant verses certainly do offer a theological retouching of the Christological hymn. The main theme remains Christ’s humility in the incarnation, but the Gothic shifts the emphasis from his status or office, qua servant, onto his ontology, qua divinity in the flesh. While the passage’s Christological impact is increased, the status of Christ’s divinity is left ambiguous. However, the specific arguments that have been used to connect the passage either to Homoian theology or to the disputes of the 340s do not persuade. Though the ambiguity does run contrary to pro-Nicene interpretations of the hymn, it also stands oblique to the interpretation adopted by later, Latin-speaking Homoians otherwise in close agreement with Wulfila’s theology. The translation is therefore rooted, we argue, not in the polemics of the ‘Arian’ controversy but in a Greek exegetical tradition of wide appeal. What motivated Wulfila to introduce it cannot be definitively reconstructed, but the cause may lie in whatever forgotten nuance of Gothic usage led him to make an even more glaring alteration to the Greek, hitherto seemingly unnoticed by scholars: a break in the parallel between the ‘form of God’ in Phil 2.6 and the ‘form of a servant’ in 2.7.

2. Philippians 2.6–8: Gothic vs. Greek

6 saei in gudaskaunein wisands ni wulwa rahnida wisan sik galeiko guda,

ὃς ἐν μορφῇ θεοῦ ὑπάρχων οὐχ ἁρπαγμὸν ἡγήσατο τὸ εἶναι ἴσα θεῷ,

7 ak sik silban uslausida, wlitskalkis nimands, in galeikja manne waurþans, jah manaulja bigitans swe manna.

ἀλλ’ ἑαυτὸν ἐκένωσεν μορφὴν δούλου λαβών, ἐν ὁμοιώματι ἀνθρώπων γενόμενος˙ (8) καὶ σχήματι εὑρεθεὶς ὡς ἄνθρωπος

8 gahaunida sik silban, waurþans ufhausjands attin und ….

ἐταπείνωσεν ἑαυτὸν γενόμενος ὑπήκοος μέχρι θανάτου, θανάτου δὲ σταυροῦ.

The Gothic deviates in four noteworthy ways from the Greek (taken here from NA28, but without significant variation across the manuscript tradition).¹² Each is underlined above.

1) The three words ἐν μορφῇ θεοῦ are rendered by in and the nominal compound gudaskaunein (accusative case). As in later Germanic languages, nominal compounding was productive in Gothic where conceptually and lexically useful.¹³ Coining or deploying gudaskaunei* may suggest that its referent formed a unitary concept in the translator’s mind. More significant is the fact that the clearly parallel μορφῇ θεοῦ and μορφὴν δούλου (discussed below) are given disparate treatments which eliminate the parallelism.

A further potential deviation from the Greek is in the shade of meaning of the second element of the compound. Gudaskaunei* (in the unattested nominative) combines the Gothic guda-, ‘God’, with an independently unattested *skaunei, cognate with English ‘shine’, German Schein, and other early Germanic words meaning ‘appearance’, and also (at least by connotation), ‘glory, splendour’.¹⁴ While we cannot know for certain that the latter sense was present in Gothic, it is etymologically probable. Furthermore, the skaun- stem appears again in Philippians at 3.21 in another compound, ibnaskaunjamma (dative singular of ibnaskauns*). In that passage, Paul foretells the Saviour’s change of our bodies to be σύμμορφον τῷ σώματι τῆς δόξης αὐτοῦ ‘conformed to the body of his glory’. Given that the two uses of -skaun- in Gothic are in contexts of divinity and glory, it is highly probable that the connotation or even denotation of ‘shining, splendour’ was present in Gothic as in other Germanic languages.

2) In Phil 2.6, the key phrase τὸ εἶναι ἴσα θεῷ is rendered as wisan sik galeiko guda, ‘himself to be galeiko to God’. Sik, the reflexive pronoun, needs no explanation: it is just the subject of the accusative-infinitive construction with wisan (the Latin Vulgate, similarly, uses esse se aequalem). galeiko is more problematic. It must derive from the adjective galeiks,¹⁵ cognate to English ‘like’, ‘alike’, German gleich, and related to leik, ‘body’, itself cognate to English ‘lich’ (as in ‘lichyard’), German Leiche. To all appearances, it is an adverb, but the formulation ‘himself to be similarly to God’ is no more grammatical in Gothic than in English,¹⁶ and so galeiko should perhaps be read as a neuter singular weak-declension adjective. If so, the translator was attempting a grammatical structure analogous to the Greek neuter plural ἴσα (and possibly trying to retain syntactic ambiguity, granted adverbial usage of Greek neuter plurals).

The potential boundaries of the semantics of galeiko may be traced by cognates. Context naturally helps delimit the precise connotations of any instance of the cognate terms, but, in general, ‘like’ is a vague word: without ruling out close resemblance, it requires only a limited similarity between two things under comparison. They will necessarily display rather more similarity if they can be described as ‘alike’, and more still if they can be described in German as gleich (a word that can be used to denote mathematical sameness, and so overlapping with English ‘equal’ in a way that ‘like’ and ‘alike’ do not). Likeness often implies equality, but it only rarely entails it and never without considerable contextual information. Thus, while the Gothic verse is a palpably careful translation, balancing the stylistic flourish allowed by compounding with a close representation of an oddity (as it can seem to English usage, too) of the Greek, its result is a watering down – though not an overt contradiction – of the Greek’s reference to equality.

3) In the next verse, Paul speaks of Christ’s assumption of the μορφὴν δούλου. The Gothic duly renders ‘form of a servant’ as wlit skalkis – not just with two words, as opposed to the one word used for μορφῇ θεοῦ, but with a word for ‘face’ or ‘appearance’ (wlits). This word otherwise translates πρόσωπον (as at Mark 14.65) or ὄψις (as at John 11.44) and is entirely unrelated to *skaunei; it has cognates of similar meaning in other Germanic languages, including Old Frisian wlite ‘appearance, face’, Old English wlite ‘appearance, shine’ and Old Saxon wliti ‘shape, appearance, shine’¹⁷. As yet a third equivalent of μορφή is attested (farwa, from a presumptive farws* in Mark 16.12, with cognates meaning ‘appearance, colour’, as in Modern German Farbe), μορφῇ clearly did not have a single, natural correspondent in Gothic. The alternation within one sentence is remarkable nonetheless.

4) In Phil 2.8, finally, an extra word appears. Christ is said not simply to have become ‘obedient unto death’, but to have become obedient ‘to the Father’ (attin) ‘unto’ (the preposition und, after which the extant text cuts off).

The combined effect of these alterations is to shift the theological significance of the original. The Greek sets up a parallel between the ‘form of God’ (μορφῇ θεοῦ) and the ‘form of a servant’ (μορφὴν δούλου). That parallel is removed in Gothic, split between gudaskaunein and wlit skalkis. That could reflect simple variatio, but it inevitably also suggests that divinity and servanthood are ontologically disanalogous statuses for Christ. In its place, a new parallel is set up, between Christ’s being ‘like God’ (galeiko guda ∼ ἴσα θεῷ) and assuming the ‘likeness of men’ (galeikja manne = ὁμοιώματι ἀνθρώπων). Compared to these translatorial moves, the addition of attin is easy to explain. Many Greek and Latin writers added a clarificatory patri or πατρί to the passage, too.¹⁸ In Gothic, it might well have helped (for all we know) to prevent the implication that Christ was ‘obedient to death’.¹⁹ Either way, it contributes an unmistakable theological nuance.²⁰ Christ set aside, refused to exploit or refused to grasp after a divine status that is left blurred; he yielded his obedience, however, specifically to the ‘Father’.

The duelling parallelisms in v. 6 and 7 can only be handled in depth, and we will turn to them next. It is remarkable, however, that the Gothic betrays no anxiety about the word that has most exercised modern exegetes of the very same verse whose alteration has most troubled Gothicists. Ἁρπαγμός is rare. First attested in this passage, it is used afterward almost exclusively in reference to it. Does it mean that Christ did not think his (real) equality ‘booty’ he had illicitly won (res rapta, according to the exegetical classification developed by N.T. Wright)? Was equality something he had but refused to ‘cling to’ (res retinenda)? Was it something he did not have and refused to ‘snatch after’ (res rapienda)? Or is the point of the verse in fact that Christ refused to use his real equality as grounds for ‘plundering’ in the fashion of too many human potentates (raptus: one element in Wright’s own view)?²¹ In the Gothic, ἁρπαγμός simply becomes wulwa, a derivative of the verb wilwan, which can mean ‘to rob’ (as at Mark 3.27, where wilwan and the compound diswilwan render διαρπάζω with different objects), but is also used for the attempt to ‘seize’ Jesus at John 6.15 (a sense elsewhere borne by the compound frawilwan, as for example at Matt 11.12; John 10.12, 28–9). Like the English ‘robbery’, wulwa therefore captures well the polysemy of the Greek. As we will see later on, what we make of wulwa will indeed matter for the interpretation of the passage. The translation itself, however, appears to leave its significance appropriately open.

3. Christ Was ‘Like God’? The Problem of galeiko guda

For as long as scholars have been reading Gothic Philippians, the use of galeiko for ἴσα has seemed a theological smoking gun. Elsewhere in the New Testament, ἴσος and related words are rendered using ibns or compounds of sama.²² Sama is cognate with English ‘same’, ibns with English ‘even’ and German eben.²³ The former is commonly used to render compounds with σύν, as indeed in Phil 2.2, where σύμψυχοι, τὸ ἓν – or perhaps the widely attested variant τὸ αὐτὸ – φρονοῦντες becomes samasaiwalai, samafraþjai (‘same-souled, same-thinking’).²⁴ Gothic ibns, too, could denote a close enough likeness to render σύμμορφον in Phil 3.21, which, as noted above, is translated with the adjectival compound ibnaskaunjamma (dative singular of ibnaskauns*).

Though ὁμοίως is routinely rendered with samaleiko (a compound of sama and leiks), ὅμοιος and its derivatives are rendered with galeiks or related forms.²⁵ One instance, as we have seen, even occurs within the same sentence as galeiko guda. The pattern implied by biblical usage is confirmed by our most significant extra-biblical Gothic text. Eight palimpsest leaves survive from the so-called Skeireins, a Gothic commentary on John.²⁶ Overlap with Greek catena-entries show it to have been derived from the John commentary by Theodore of Heraclea, an ally of Eusebius of Nicomedia, and so of Wulfila.²⁷ The Skeireins contrasts galeiks three times with ibns or the apparently synonymous ibnaleiks.²⁸ One instance mirrors idiomatic pairings of ὅμοιος and ἴσος.²⁹ The other two infer from John 5.23 and 17.23 that we are ‘to render similar (galeiks) but not equal (ibns) honour’ to the Son as to the Father, and that the Father has ‘similar (galeiks) but not equivalent (ibnaleiks) love’ for the Son’s disciples as for the Son.³⁰ Only Phil 2.6 deviates from the general pattern, and precisely in a statement that could be taken (depending on the significance of ἁρπαγμός) to put Christ ‘on a level’ with the Father. Multiple renderings of other words are known in the Gothic New Testament. They sometimes reflect stylistic variatio but might also indicate that multiple translators were at work.³¹ Here, however, we have a consistent pattern broken only in one place. Ἴσος is being assimilated to ὅμοιος. The only question is why.

To many scholars, the explanation has seemed obvious.³² The creed of 359/360 made Christ ὅμοιος (and no more than ὅμοιος) to the Father. Gothic Phil 2.6 offered a literally Homoian rendering, one perhaps meant for outright deception of Goths who had no other access to the scriptures.³³ Though it echoes an ancient understanding of the origins of ‘barbarian’ heresy,³⁴ this simple explanation cannot be right. The extant parts of the Gothic Bible neither suppress difficult passages nor warp others in a subordinationist direction. Examples that have been proffered are thoroughly unconvincing. To stress father-son lineages in Jesus’ genealogy in Luke 3, for example, or his subjection, until the beginning of his ministry, to Joseph bears no specially subordinationist implication.³⁵ On the other hand, no passage offers a neater prooftext for the ontological equality of Father and Son than John 10.30, and here the Gothic heightens their unity by rendering the verb with a rare dual: ἐγὼ καὶ ὁ πατὴρ (μου) ἕν ἐσμεν becomes ik jah atta meins ain siju, ‘I and my Father are-[both] one.’³⁶ That verse is representative of a reasonably competent translation. Sometimes mannered, sometimes nuanced, sometimes frankly misguided, the Gothic bears no resemblance to a systematically subordinationist version such as the Watch Tower Bible and Tract Society’s New World Translation. ³⁷

Mere theological bias cannot account for the rendering of galeiks in Phil 2.6. The Gothic church historian Knut Schäferdiek found a subtler explanation: modern readers have misunderstood Gothic semantics.³⁸ Again the Skeireins is key. Like other members of the Eusebian alliance, Theodore of Heraclea was staunchly opposed to the pro-Nicene firebrand Marcellus of Ancyra, whom they accused of holding to a modalist theology that made the Logos a temporary ‘expansion’ of an absolutely singular God.³⁹ When the Skeireins declared that the Son was owed an honour galeiks rather than ibns to the Father’s honour, the commentator was asserting that the honour owed the Son was numerically distinct from the honour owed the Father, not necessarily that one honour was less than the other. Ibns therefore implied identity, not equality, and galeiks was equivalent to the German gleich.⁴⁰ Wulfila chose, likewise, to render ἴσα with galeiko so as to avoid the unduly modalist implications of ibns. The Gothic translation of Phil 2.6 finds its explanation in theological party politics not of the 360s but of the 340s. Not a wilful misrepresentation, it is an attempt to head off a position that seemed, to most Eastern bishops, patent heresy (and was rejected by later pro-Nicene consensus, too).

This conclusion cannot stand, either. To begin with, the semantic argument does not follow from the theological data. Theodore participated in councils that assigned different grades of honour to Father and Son.⁴¹ Eusebius of Caesarea, whose late rebuttal to Marcellus is broadly representative of the Eastern mainstream, inferred from John 5.22–3 that the Son is to be honoured ‘in nearly equal fashion’ or ‘in about the same degree’ (παραπλησίως) to the Father.⁴² That terminology asserts a genuine closeness of Son to Father: he has been ‘honoured’, as Eusebius had earlier put it, ‘with the deity of Paternal glory’.⁴³ Without implying any slight against the Son, Eusebius’ wording still did not place the divine persons exactly on a level. For this current of Greek theology, gradation and multiplicity, of honour as well as divine hypostases, still went hand-in-hand.⁴⁴ Ibns, therefore, could still very well mean ‘equal’. Even more importantly, Schäferdiek’s arguments applied as much to the putative Greek original of the Skeireins as to the extant Gothic fragments. If ibns bore modalist implications, then so did ἴσος. In rendering Phil 2.6 with galeiko, Wulfila would, under Schäferdiek’s interpretation of the semantics, still have been doing what Schäferdiek hoped to show that he had not: imposing his party’s preconceived interpretations on a Pauline text that he himself knew to contradict them.

Schäferdiek’s defence of Wulfila collapses back into the position he opposed. It still points to an essential principle: whatever is going on in Gothic Phil 2.6, the translation must reflect contemporary exegesis. We have no sign of a pervasive effort to adjust the biblical text to subordinationist sympathies, and indeed we would hardly expect such adjustment of Wulfila. He was an active theologian and teacher, or so we are told by the main source for his life, a eulogistic letter, written shortly after his death in 383, by a former student, Auxentius of Durostorum.⁴⁵ A disputant in Greek, Latin, and Gothic,⁴⁶ Wulfila will have been familiar with current exegetical work. Indeed, the Skeireins itself could well be his. He will have known how to derive a plausible (in his view, orthodox) meaning from passages difficult for readers who shared his theological sensibilities, just as pro-Nicenes of all eras have been able to harmonise John 14.28 or Prov 8.22 with the Son’s eternal equality. What is true of John 10.23 must undoubtedly have been true of Phil 2.6, too. There is therefore little motivation for intentional distortion of meaning, and mistranslation is less useful a paradigm than functional equivalence.

The use of galeiko for ἴσα is, after all, only one of several oddities within the Christological hymn. That suggests that Wulfila was trying to convey the passage’s meaning, while deviating from its literal wording. What, then, do we expect a man like him – a subordinationist theologian of the mid-fourth century – to have made of the opening of the Christological hymn?

4. Res rapienda and Res retinenda: Gothic Phil 2.6-7 and Subordinationist Theologies

The term ἁρπαγμός, as usual, is the hinge on which interpretation turns. We would expect theologians convinced of the Son’s inferiority to find, in the celebration of Christ’s humble obedience and his refusal to treat equality with God as ‘robbery’, a proof that he had not set himself equal to his begetter. This appears indeed to have been the standard view among adherents of the Homoian creed in the late fourth and fifth centuries. In Greek, the evidence is weak: just John Chrysostom’s indication that a res rapienda interpretation of Phil 2.6 was a customary ‘Arian’ talking point, and a few vaguer parallels.⁴⁷ Two Latin writers seem, however, to take res rapienda for granted. One, whose fragmentary work is preserved in palimpsest, quotes an ally of Arius: Athanasius of Anazarbus, one of very few recorded to have defended the ex nihilo creation of the Son.⁴⁸ According to the later Latin writer, who appears to be endorsing his views, this Athanasius had written, ‘The Son does not raise himself against the Father, nor does he think that there are equal things with God; but he yields to his Father and confesses, teaching all, “The Father is greater.”’⁴⁹ The greatness of the Father consists in his properly divine attributes (perpetual existence, aseity, etc.). The Son, therefore, is fully aware of the Father’s superiority, and, if the original wording is faithfully reflected in translation (neque putat paria esse), οὐχ ἁρπαγμὸν ἡγήσατο is really just a circumlocution for οὐχ ἡγήσατο.⁵⁰ So great is the gap in status that a seizure of equality is, quite literally, unthinkable.

Res rapienda also appears to be the view adopted by the Latin Homoian disputant we know best. Augustine’s last in-person sparring partner, Maximinus, had accompanied Gothic-speaking Roman troops to North Africa in 427. He was very likely the author, some years later, of the work that preserves Auxentius’ letter on Wulfila and praises Arius himself for his orthodoxy.⁵¹ In the Christological hymn, quoted at the outset of his long disquisition, Maximinus found proof of the Son’s subordination.⁵² Augustine, however, had argued that the non rapinam (οὐχ ἁρπαγμὸν) of Phil 2.6 proved that Christ was equal to the Father by nature.⁵³ Maximinus returns to the point, quoting Phil 2.6 several sections into his disquisition. ‘This the blessed apostle Paul has taught us, that he did not seize it, nor do we say so.’⁵⁴ The wording is compressed, even ironical, and, though Maximinus insists a little later that he has given his party’s opinion on the passage,⁵⁵ he never actually explains how Phil 2.6 is to be interpreted. It was therefore left to Augustine to tease out the logic in his two-book rebuttal to Maximinus’ speech. ‘If’, Augustine asks, ‘you confess “the form of God”, why don’t you openly confess the Son of God equal to God?’ Maximinus, he insinuates, could not turn to his advantage the apostle’s words in Phil 2.6, and so he said:

‘That he did not seize it, nor do we say so’, as if ‘He did not seize’ meant ‘he did not have (equality to God)’; and so the statement, He did not deem it robbery to be equal to God, is tantamount to saying, ‘He did not deem that equality with God should be seized, inasmuch as it was alien to him’ – for a robber is a usurper of another’s property – as if the Son was unwilling, though he could, to seize it.⁵⁶

Here, Augustine must have the right of it. The equality of the Trinitarian persons is constantly opposed by the Latin non-Nicenes, and Maximinus is no exception, though he dwells chiefly on the status of the Holy Spirit.⁵⁷ Maximinus can only mean that the Son did not aspire to an equality he did not possess.

The doctrines of the Latin Homoians neatly match those Auxentius attributes to Wulfila.⁵⁸ It is therefore striking how poor a fit res rapienda is for Gothic Phil 2.6. As we saw, the Gothic version eliminates the parallelism between μορφῇ θεοῦ and μορφὴν δούλου but replaces it with a parallelism between galeiko guda and galeikja manne. If Wulfila really does have Paul say that Christ ‘did not seize for himself likeness to God’ yet did take on ‘likeness to men’, then Paul would be denying that the Son is like the Father at all. That view is improbable from a sincere Homoian. Claiming to relay Wulfila’s own theology, Auxentius speaks of the Father ‘making’ (fecit) and ‘creating’ (crea|uit) the Son, but insists that Father and Son alone deserve the title ‘God’ and that each is creator, while all other beings (including the Holy Spirit) are creaturae.⁵⁹ The Christ of the Latin Homoians is unlike the Father in important ways, but, as God and creator in a lesser, derived sense, he is still both divine and like the Father in a way absolutely nothing else is.

Though galeiko guda softens the implications of ἴσα θεῷ, it can only be construed in a way that attributes divine status (in whatever degree) to Christ. The new parallelism with galeikja manne puts a decisive shift at the incarnation, and so implies that wulwa (ἁρπαγμός) is to be taken in the sense that Wright terms res retinenda. That Christ had ‘let go’ of divine glory was a venerable interpretation. An early example stresses Christ’s refusal to promote himself at the Father’s expense: the martyrs of Vienne and Lyons, killed after riots in 177, imitated Christ to such a degree ‘that, though they existed in such glory … they did not proclaim themselves martyrs’.⁶⁰ The point of Phil 2.6, by implication, is that Christ was divine yet refused to promote his own divinity. That view shades over into res rapienda, but of the prerogatives or glory, not the attributes, of the Father: a view with echoes both in the firmly subordinationist Novatian and in Nicaea’s Latin champion, Hilary of Poitiers.⁶¹ For Origen in his commentary on John, Christ’s refusal to think equality with God ἁρπαγμός was expressed in his willingness ‘to become a slave for the salvation of the world’.⁶² That refusal displayed the love that led him to associate with sinners, ‘to descend as far as death on behalf of the impious’.⁶³ Origen’s follower Eusebius of Caesarea, squaring off against Marcellus of Ancyra, found in the passage proof of Christ’s pre-existence and distinctness from the Father.⁶⁴

Neither Origen nor Eusebius is concerned, in these passages, to specify just in what way Christ was ἴσα θεῷ. Phil 2 served not to prove Christ’s strict equality with God but his divinity as such. Theologians of the late 350s and 360s developed that thought further, in ways strikingly parallel to Gothic Phil 2.6. In 358, the ‘Homoiousian’ Basil of Ancyra, defending Christ’s likeness to the Father ‘in essence’, held that Christ’s existing ‘in the form of God’ and being ‘equal’ to God meant that he had ‘the properties of the deity’.⁶⁵ In 366, a similar view was advanced by a sometime ally of Valens of Mursa and Ursacius of Singidunum, the leading Western proponents of the Homoian formula. Accused of going beyond the 359/360 councils by asserting the Son’s likeness to the Father ‘in all things, except Unbegottenness, Germinius of Sirmium doubled down.⁶⁶ ‘Who would not perceive’, Germinius asked, ‘that, just as our flesh was true in Christ according to “the form of a servant”, so also the divinity of the Father in the Son was true “in the form of God”?’⁶⁷

Neither Germinius nor Basil is asserting Christ’s equality with the Father. Aequalitas and ἰσότης are absent, and in fact Basil and his allies concluded from John 5.19 that the Father acts ‘sovereignly’ (αὐθεντικῶς), but the Son ‘servantly’ (ὑπουργικῶς) – a doctrine that Basil had already inferred from the lack of articles in ἐν μορφῇ θεοῦ and ἴσα θεῷ.⁶⁸ They nonetheless found in Phil 2.6 a statement that Father and Son were, in all respects not touching on Fatherhood and Sonship, alike. The interpretations advanced by the ‘Homoiousians’ and by Germinius represent the converse of a famous pro-Nicene argument. For Hilary of Poitiers, to predicate similitudo naturae of Father and Son was to declare their equality.⁶⁹ Germinus avoids the language of ‘nature’, but otherwise inverts the proposition. When Paul called Christ ‘equal’ to the Father, he was according him homologous attributes. Phil 2.6 was about the Son’s divinity, yet still allowed the distinctness that the most radical pro-Nicenes seemed to deny.

Both Basil and Germinius accused Valens and Ursacius – like their fellow Eusebian, Wulfila, from the Danube region – of wanting to hold the Son dissimilar, at least in certain attributes, to the Father.⁷⁰ That is the flipside to the profession of ‘similarity’ that Auxentius credits to Wulfila. In an echo of the Homoian creed, Wulfila taught ‘that the Son was similar to his Father … according to the Scriptures and tradition’, but his teaching, as Auxentius describes it, underscored the differences between Father and Son.⁷¹ Thus, Auxentius lists off a massive string of attributes proper to the Father alone and, while the Son’s titles are lacunose, only deus and auctor, among those that remain, is shared with the Father. Even the perfectly scriptural dominus is the Son’s alone.⁷² As divine creator of the universe, Christ is like the Father, who is his own, even more truly divine, creator.⁷³ On the fine points, he is distinct.

This, as we have said, is much the same teaching we find in the later Latin Homoians, and that seems in Maximinus’ hands to undergird a res rapienda interpretation. It was, however, hardly an innovation of the 380s. Athanasius of Anazarbus had advanced a similar reading decades earlier, as had Novatian long before the Arian controversy began.⁷⁴ It clearly was an exegetical possibility all along, but only happens to emerge relatively late in texts related to the controversy. The res retinenda interpretation may by the same token have remained ‘live’, especially among Greek-speaking Homoians who could read Origen and Eusebius. It is nonetheless possible that we are seeing a trace of the sectarian hardening of the Homoian movement, as their creed, once upheld by emperors, became the confession of a separatist, minority church, following the pro-Nicene settlement in 381.⁷⁵ Certainly, Wulfila’s teaching, as relayed by Auxentius, looks like a product of recent controversy, with its intense stress on the inferiority of the Holy Spirit and desire to exclude positions, focused on the Son’s equality or close likeness, that the Homoian creed had not explicitly ruled out.⁷⁶ If so, Gothic Phil 2.6 will likely be a relict of a time when Wulfila held to a more generous view of the Son’s divinity, akin, maybe, to that of the late Eusebius of Caesarea. Otherwise, it shows him transmitting for Gothic-speaking posterity an interpretation of the passage that was firmly traditional, but out of step with his private theological opinion. Either way, it owes no more to the staunch subordinationism of the Latin Homoian mainstream than it does to Nicene belief in the Son’s full equality. What we are seeing is a longstanding Greek exegesis crystallised into the Gothic text.

5. Conclusion: Functional Equivalence in Gothic

Why, then, did Wulfila (or an assistant) choose to remodel the text so extensively? A desire to communicate basic theological tenets more clearly was surely part of the motivation. Though the Gothic version cannot be taken in a crassly ‘Arian’ sense, it does look like an attempt to tie the passage more neatly to what will have seemed, in Wulfila’s day, the central themes of Christian theology: in particular, the fact that Christ is (in whatever sense) genuinely both human and divine, while still distinct from the Father. The question, then, is why this doctrine would have seemed in need of clarification. Here, we would suggest that the other aspects of the passage are significant. In immediate context, the deployment of galeiko guda looks like an attempt to retain and realign the parallelism deleted from the translation of ἐν μορφῇ θεοῦ and μορφὴν δούλου. We know we are missing vital pieces of the puzzle: virtually everything, in fact, about traditional Gothic culture, including its pre-Christian religion and extra-biblical semantics.⁷⁷ Instead of the question, so salient in modern exegesis, of what to make of ἁρπαγμός and Christ’s ‘equality’, could the starting point actually have been the problem of divine ‘form’ or ‘shape’?

We have an expression for the ‘shape’ of God in a text intended for public reading to an uneducated, still-converting people out of reach of established, Graeco-Roman apologetic. We also have a reference to the ‘servility’ of Christ, parallel in Greek to the ‘shape’ of God but set apart from it in the Gothic. It seems unlikely that *skaunei, connoting ‘splendour’, could have been linked with skalkis. Could Wulfila have had a good reason to avoid linking wlits with guda, as well? Perhaps wlits, implying ‘face’, seemed to make the μορφὴ θεοῦ into a literal idol. Less excusably, could Wulfila simply have wanted to soften the shocking juxtaposition of Almighty God and servitude, out of concern that it would repulse his Gothic audience? Another possibility, in keeping with the exegetical sophistication of the rendering as a whole, would relate to the widespread understanding of the Old Testament theophanies as the work of the Son. Might the translator have been trying to make it clear that the Son’s divinity, though not strictly invisible, was still immeasurably grander and more radiant than Christ’s human ‘aspect’?

The intention behind the translation is shrouded in the same obscurity that cloaks the rest of early Gothic culture. It is clear, however, that Wulfila’s reasons for translating the passage thus must have lain in some interplay between Greek theological consensus and Gothic semantics. Gothic Phil 2.6–8 is not an overtly pro-Homoian or anti-Marcellan text. It is a representation, in an ancient counterpart to modern ‘functionally equivalent’ translation, of the interpretation at which influential Greek exegetes had long since arrived. Though a complete resolution of this passage’s puzzles is impossible, this realisation bears two important consequences, one for fourth-century church history and its extensive modern historiography, the other for the interplay between Germanic philology and study of the New Testament. Scholars have learned to question the traditional label ‘Arian’, which homogenises dramatically different theological views. ‘Homoian’ has become its customary replacement, for the phase of the controversy after 359/360, but it, too, bears only limited utility in theological analysis. Properly a descriptor of a compromise creed and not – at least before the 380s – of a coherent, doctrinally united faction of bishops, ‘Homoian’, like ‘Arian’, offers no more than an approximate placement, among the available theological options, of a given churchman’s position.⁷⁸ As the examples of Germinius, Maximinus, and the Wulfila of the Gothic Bible reveal, adherents of the Homoian creed could take markedly different views on such weighty prooftexts as Phil 2.6, and remain responsive to prior exegetical tradition as they did so. The second consequence promises to reshape the modern reading of the Gothic Bible. Wulfila’s version has long been interrogated, without deeper consideration, for traces of ‘Arianism’. The Gothic Christological hymn illustrates a much more dynamic interplay among ancient exegesis, theology and translation practices. Our conclusion as well as the methodology by which we have reached it invite new and deeper attention to Wulfila’s exegetical and translational methods throughout the extant fragments. It also undercuts the assumption, as inveterate as the focus on ‘Arian’ theology, that he rendered the Greek through strict lexical equivalence. As this passage reveals, the Gothic can also represent the best exegetical insights of the Greek theological tradition, up through Wulfila’s own day.