A Reconsideration of the Early Greek-Arabic Translation Phenomenon*

Why the translation movement again?
The Greek-Arabic translation 'phenomenon' is, even after more than a century of constant scholarly attention, still a controversial subject. The scope and character of the transmission of texts were conditioned by a number of factors, such as the cultural and political background of source texts and the contemporary cultural and political situation of both translators and their audience on the one hand and the language barrier on the other hand, a formidable obstacle in view of the substantial structural differences between the main languages involved in the process: Greek, Syriac and Arabic. Whether for lack of interest or lack of appreciation, these variables have received less attention than one might assume. With research often focused on the potential value of the translated material as a substitute for lost Greek writings or with Hellenized Islamic philosophy, one of the disciplines benefiting most from the translation phenomenon, dismissed as derivative and therefore unoriginal, the cultural and linguistic issues involved in the process of translation have frequently been left unconsidered.

   A conspicuous feature of earlier research on the translation phenomenon is the reliance on a slim body of evidence for sweeping claims on the subject of translation techniques. The small amount of information culled from Khalíl ibn Aybak as-Safadí (d. 1363) and Hunayn ibn Isháq (d. 873) has by virtue of the sheer frequency of repetition been promoted to a holy writ for research. Fortunately, at least the former source has been exposed as unfounded; Hunayn's remarks on his own methods are yet to be put into perspective: he offers an account of his own work, applicable perhaps to his own colleagues, but probably not to every translator working after him. As unfounded as any extension of his methods to later translators is the accompanying demotion of his predecessors to mere dilettantes. Islamic philosophy, too, has received its share of criticism; in an intellectual climate steeped in Eurocentric notions about the superiority and universality of the Greek philosophical tradition, the translations and Islamic philosophy alike have been interpreted as a step, albeit sterile and unproductive, on the way from Greek thought to contemporary forms of philosophical expression.

   Parallel to an essentialist account of philosophy, invalidating philosophical traditions that cannot be readily pressed into the mould of Greek thought, most research into the translation phenomenon suffers from a flawed concept of both language and translation. The na•ve view of translation as the transfer of a basically immutable content from one linguistic medium into another is insufficient even in the case of closely-related languages; in the case of Greek, Syriac and Arabic, such a view seriously distorts any treatment of translation problems. And even though translation problems caused by the linguistic differences between these languages are discussed on numerous occasions, they are seldom, if ever, elaborated into a contribution towards the better understanding of both translations and the philosophical works inspired by them. Behind the deficient understanding of translation in evidence in most writings on the translation phenomenon lurks a static and essentialist understanding of language. Unfounded assertions on the merits and potential of a language to express a given set of ideas betray a lack of understanding of the basic flexibility of linguistic systems to accommodate and express concepts derived from differing linguistic and cultural contexts. At least for the older strand of research exemplified by Walzer, Rosenthal and Badawí, a text-centred approach could be responsible for the deficiencies of their understanding of the translation phenomenon: basing themselves predominantly on extant translations isolated from their cultural 'environment', they often fail to take into account the evidence for cultural contacts and transfers preceding the translation movement and subtly informing a whole range of fields, e.g. jurisprudence and theology. The impact of Hellenistic science and philosophy stretches farther back than the beginning of a recognisable translation 'movement' (cf. Kunitzsch 1975, 1976; Versteegh 1980).

Text linguistics as a tool for translation assessment
One possible method for a more nuanced appreciation of the translation phenomenon can be found in the theoretical approaches developed in translation studies. Most of its exponents draw from both practical experience in translation and their grasp of relevant research in literary theory and literary criticism - the former often acting as a check on the more efflorescent speculations of the latter. Any explanation of the evolution of translation methodology and assessment of the degree of modifications in form and content imposed by structural differences between the languages in question can only benefit from taking into account the insights offered by translation theory. Explanations in purely philological terms are clearly insufficient to account for the findings: based purely on terminological and syntactical analyses and therefore restricted to the semantic and syntactic level, they ignore basic issues of text structure and cannot, in view of the substantial structural differences between Greek and Arabic, provide a sufficient basis for an assessment of a text.

   The translation phenomenon had an impact not only on the host culture, but on the form and content of the translated texts as well. On the linguistic side, the structural differences between source and target languages need to be assessed and related to the corresponding translation methodology devised to deal with them. Based on the results derived in this first stage, source texts and translations need to be compared in terms of their textual equivalence and the restrictions and modifications imposed on the target language versions by virtue of the linguistic difficulties the translators had to face. Equipped with the theoretical tools developed in translation studies in the last two decades - the specific theoretical framework will be discussed below -, I propose to analyse originals and translations and offer an account of and explanations for modifications on the linguistic level.

The history of the translation phenomenon in the Islamic world
Eminently relevant both for the historiography of the translation movement and the assessment of its interaction with Islamic cultural phenomena is the chronology of the earliest translations and the amount and efficacy of oral communication taking place between different linguistic and cultural groups in the former Byzantine and Persian territories conquered by Muslim armies. From a static, text-centred approach exemplified by Walzer, Rosenthal and Badawí, research has developed towards a view stressing non-textual transmission and the influence of transcultural diffusion in the early history of Islamic expansion (e.g. Jadaane 1967, Kunitzsch 1975, Versteegh 1980). With the shift of focus away from extant translated texts, the chronological relationship between philosophy and translations has become once again a matter of debate: have translations brought about the 'invention' of Islamic philosophy or has any preexisting philosophical tradition provoked the transfer of relevant Greek material into Arabic? Or grew both the translation phenomenon and Islamic philosophy out of the same intellectual impulse? On the theological front, Islamic polemicists had to face a barrage of arguments against Islamic doctrines brought forward by Christian theologians steeped in a tradition of philosophical and logical training that had been kept alive in Syriac monasteries and schools. Their desire for argumentative training manifested itself early on and might have had an influence on the actual course of the translation movement.

   The role of the Syriac translation tradition must also figure prominently. Many commentators have remarked on the existence of Syriac intermediaries for Arabic translations of Greek philosophical texts. Another, even more obvious avenue for Syriac influence on the character of Arabic versions is the linguistic and cultural background of many of the translators themselves: the majority of them were Syriac native speakers and went through the Syriac educational 'system'.

The cultural circumstances of the translation phenomenon
While the question of intentional change might be a moot one on the linguistic level - translators were probably forced to have recourse to their problem-solving strategies rather than consciously choosing them -, the translations carried out by the circle of researchers around the philosopher Ya'qúb ibn Isháq al-Kindí (d. after 870) offer numerous examples of a second type of modification, i.e. conscious additions, amplification and deletions. To explain these interventions, the intellectual, political and social situation at the time of translation has to be studied. The translations are obviously a product of their time. Authors had to enable their audience to understand the texts not only linguistically, i.e. their vocabulary and literary form, but also textually, to grasp concepts originating from a different cultural context. The removal of references to polytheism in translated texts is only one, albeit particularly obvious, example of cultural factors at work. Successive translations of specific Greek texts bear witness to the fact that the relativity of translations to their historical context necessitated a new rendering of the source text with changing literary, cultural and intellectual standards and changing tastes and demands on the part of the reader.

   Of particular interest in this respect are two crucial political and theological developments: Firstly, the accession of the 'Abbásid dynasty to power in the year 750 coincided with the beginning of a systematization of translation activities, culminating in the work of whole groups of translators, exemplified by the 'circle of al-Kindí' and, shortly afterwards, Hunayn ibn Isháq and his associates. Several theories have been put forward to explain this shift, stressing e.g. narrow-minded 'Arabism' on the part of the Umayyad rulers as a factor for the apparent lack of translation activity during their reign (Rosenthal 1965, Endress 1989.2) or the role of translations as an important part of the 'Abbásids' strategy to justify their claim to the caliphate (Gutas 1998). Secondly, the relation between contemporary theological debates and disputes on the one hand and the translation phenomenon and the philosophical speculation based on it on the other hand has to be part of any explanation of the motivation for the translation movement, the choice of texts or the stake different social and political groups had in the translation movement as a whole. As is evident e.g. in al-Kindí's Risálat fí l-falsafat al-úlá with its conscious use of expressions laden with religious connotations, philosophers and theologians strove to uphold their right to define key terms and concepts: translations and philosophical texts spawned by the translation phenomenon can be better understood by taking into account the subtext of the continuous debate between philosophy, theology and law for the interpretation of the Islamic revelation.

The essentials: methodology and sources
An account of the Greek-Arabic translation phenomenon should incorporate the findings which these considerations produce and the results of a linguistic analysis of relevant texts. I have based my decisions about the material included in the sample of texts to be studied on the following considerations: The early exponents of the translation movement, pioneers in their field and, most importantly, continually obliged to make crucial terminological and methodological decisions, formed the backdrop for the later peak of the translation phenomenon represented by the work of Hunayn ibn Isháq and his associates. They are a natural starting point for the observation of the formative period of philosophical terminology in Arabic. The philosopher al-Kindí in turn used these early translations in his own philosophical speculations; he was an active participant in the development of the philosophical terminology and very much part of the political, theological and social conflicts accompanying the early translation movement. His texts provide ample evidence for his personal synthesis of Hellenistic and Islamic cultural elements into a system of philosophical thought, for the terminology his ideas have been couched in and the use to which Greek philosophical concepts were put in contemporary theological and political debates. They offer a glimpse into the dialectic of disparate cultural traditions that was Islamic civilization in his time.

   As a starting point, Aristotle's Metaphysics Book Lambda has been chosen. Due to their inclusion in Muhammad ibn Ahmad ibn Rushd's (d. 1126) Commentary on the Metaphysics, we have at our disposal two translations of parts of the book. The earlier one, probably commissioned by al-Kindí himself, was produced by a certain Ustáth, one of the more productive members of the circle of translators working alongside and in cooperation with the philosopher. The second translation is ascribed to Hunayn ibn Isháq's son Isháq ibn Hunayn (d. 910) and gives an impression of the state which philosophical terminology and translation methodology had reached by the time of Hunayn. To verify and expand the evidence from metaphysical texts, passages taken from other extant translations ascribed to members of the 'Kindí-circle' will be included, namely Ibn Ná'ima al-Himsí's rendering of the Pseudo-Aristotelian Theology of Aristotle, Yahyá ibn al-Bitríq's extant translations of scientific and logical works and the rendering of Aristotle's De Anima previously attributed to Hunayn ibn Isháq. One important aim is to isolate terminological and methodological solutions which exponents of the early translation movement devised while translating philosophical and especially metaphysical texts. The longest and, in terms of its subject matter, most suitable Kindian treatise to study his creative use of the material he commissioned from translators is his Risálat fí l-falsafat al-úlá. The influence of both the Metaphysics Book Lambda and the Theology of Aristotle can be traced in the speculative theology he introduces to his audience - an audience in all probability not used to this type of text. He seems to consciously borrow theologically suggestive terms and concepts and embeds them into his text and is therefore an important source for the 'application' of Greek philosophical ideas in an Islamic religious, literary and intellectual context.

The cultural and historical dimension of the texts
The comparison of the texts aims to trace modifications of content between source and target texts and to uncover evidence for the historical relativity of the translations, identifying them as important evidence for ideological conflicts of the time. The integration of translated concepts into areas of Islamic culture can be illustrated with al-Kindí's speculation in relation to contemporary theological debates: his vocabulary mirrors the struggle for terms and their interpretation between theology and philosophy.

   By following several Greek terms through their development from the original text - sometimes through Syriac intermediaries - to different Arabic renderings, it will become clear that both vocabulary and the underlying ideas changed with their frame of reference (the respective source and target cultures) their field of application. The texts quite literally could not mean to al-Kindí what they had meant to students of Aristotelian and Neoplatonic philosophy in pre-Islamic Alexandria, let alone to Aristotle himself. This will be borne out by an analysis of the Greek and Arabic versions of the Metaphysics Book Lambda on the basis of the theoretical framework of contrastive text linguistics.

The linguistic content of the texts
Any given translation is the result of a translator's confrontation with a number of factors and conditions. These include the structures and potential of the source and target languages and the classification of reality they enforce; the linguistic, stylistic and translational norms of the translator; the structural features of the source text; the translation tradition the translation forms a part of; the physical conditions of the translator's work. Even though linguistics cannot in itself fully explain features of a translation, it can offer valuable, if not indispensable contributions towards such an explanation.

   One of the promising models for linguistic analysis, straddling the borders between translation theory, contrastive linguistics and discourse analysis, is the 'contrastive-discourse' model proposed by Basil Hatim and Ian Mason (1990, 1991, Mason 1994, Hatim 1997) which aims at recording and analysing the various translation decisions a disciplined reading of a text suggests. From the pre-reading stage, the familiarity with the author's background, the historical situation and the influence of the chosen medium for the meaning of a text, the translator proceeds to the actual text processing, the development of an idea of the text's contents and its relevance and tests it against the evidence provided by textual elements encountered in the reading of the text. This 'global patterning' of a text, expressed through textual genre, discourse type and text type, informs both isolated elements of a text and its overall compositional structure; in different languages, all three categories of 'global' textual organization are regulated in their usage by linguistic restrictions that have to be taken into account. Moreover, they interact with the local structures, i.e. the local semantic, syntactic and textual decisions a translator has to make in accordance with his own linguistic experience. Finally, at the micro-level of texture, linguistic differences between languages loom large and demand decisions in order to reconcile cross-cultural and trans-linguistic differences.

   In practice, the proposed analysis would involve the identification of the genres, discourse and text types of the texts and evidence of their influence on textual details such as terminology, syntax etc. and the correspondence - or lack thereof - between higher-level structures and their semantic and syntactical expressions. On this basis, a first assessment of equivalence could be undertaken. Incongruities between higher-level (discourse, text type) and lower-level (texture) organization could then be identified and examined as indicators for possible corresponding shifts in underlying ideological orientation.

Combining the evidence: the results
With the set of cultural and linguistic findings established in the course of the investigation, I hope to be able to comment on the following complexes of questions: Firstly, on the linguistic level, the degree of equality between source texts and translations, i.e. their adequacy in terms of transmitting a specific message and the effectiveness of the translational strategies employed by the translators. Secondly, on the cultural level, the impact of the translation phenomenon, especially the philosophical concepts introduced by it, on areas of Islamic culture such as theology and Islamic law and the often-disputed originality of the philosophical speculation based on the translated texts and, conversely, the influence of contemporary political and social factors in the target culture on the translated content. Thirdly, the importance of this specific case of cross-cultural transfer for translation theory and the results that can be derived by the application of a particular linguistic model on a group of texts, supported and supplemented by an appreciation of the cultural and intellectual context.

   On the evidence of this and other translations, it should become clear that the transfer of concepts and ideas even between structurally unrelated languages is possible and has been successfully undertaken. Sceptical claims to the contrary, founded on the confusion between universal translatability, i.e. the unqualified possibility of adequate translation of texts between two languages, and universal expressibility, the possibility to express the content of a source language text with the linguistic instruments provided by a target language, have to be rejected. Equally misplaced are assertions from postmodernist commentators stressing the cumulative effect of linguistic incompatibilities between languages to the point of dissolving the link between source and target texts in favour of an unrestricted equality of coexisting 'originals'. The results of this thesis will provide enough evidence to demonstrate that even in the case of apparent incompatibilities such as those between Greek and Arabic, the link between source and target text remains strong enough to label the relation between the texts as 'translation' rather than 're-creation' or mere 'inspiration'.

   Moreover, though successfully transferred from Greek into Arabic, metaphysical concepts were subjected to modifications due to the specific cultural and intellectual differences between source and target cultures and the exigencies of transcending the language barrier(s). Having successfully overcome these obstacles, the transferred concepts entered into and transformed contemporary discourses in philosophy, theology and other fields. Finally, the data offered in this study should demonstrate sufficiently that linguistic, cultural and historical relativity affects both source and target texts and changes their relation into one of equality in terms of their originality and fertility to engender discourse rather than subordination and derivation - without severing the link between them.

* Throughout the proposal, I am using expressions such as 'Islam', 'Greek/Islamic philosophy', 'the West' etc. as if to posit them as monolithic cultural or intellectual entities devoid of conflict, diversity and change. Far from intending to gloss over the existence of numerous different interpretations of each of these cultural phenomena, incessant intellectual developments and the parallel existence of diverging and conflicting schools of thought inside them, I have decided to adopt this shorthand due to restrictions of time and space.