Galen's highly influential treatise On the Affected Parts (Περὶ τῶν πεπονθότων τόπων, often referred to by its Latin title De locis affectis, hereafter indicated with the abbreviation De loc. aff.) is currently being critically edited by the Corpus Medicorum Graecorum at the Berlin-Brandenburg Academy of Sciences.Footnote 1 Over the last decade, a team of scholars, including the present authors as well as the late and lamented Aḥmad ʿEtmān, have worked on producing a critical edition of the Arabic translation of this text, and their efforts are now drawing to a close.Footnote 2 Here we present new insights into how this Arabic translation relates to the Greek textual tradition.
GREEK AND ARABIC MANUSCRIPTS
De loc. aff. survives in sixteen Greek manuscripts, as discussed in detail by Florian Gärtner, both in a preliminary article and in his recent edition of Books 1–2.Footnote 3 Excluding the apographa, Gärtner's conclusion about how the manuscripts relate to each other can be summarized as follows. The oldest manuscript, which often represents an independent branch of the tradition, is Laurentianus Gr. 74.30, thought to have been written in the twelfth century; it is designated as F. The other branch of the tradition, designated as ε, includes several older manuscripts. The most important such manuscript is Laurentianus Gr. 74.16 (L), a composite manuscript, consisting of two parts, La and Lb and dating to the late twelfth or early thirteenth century. The largest subgroup of ε is ζ, which includes La and η; η in turn comprises three manuscripts: C (Canonicianus Gr. 44), Q (Ambrosianus Gr. 679) and N (Parisinus Gr. 2332). Lb derives directly from ε, whereas La is included in ζ. For the present argument, it is important to bear in mind that F often has independent readings from the rest of the tradition, which is represented by ζ and its descendants. Gärtner assumes an archetype ω, from which both F and ζ ultimately descend.
The Arabic translation of De loc. aff. is preserved in twelve manuscripts.Footnote 4 For the purpose of the present discussion, this article bases its arguments primarily on an edition established through the collation of the following three manuscripts:
Cairo, Egyptian National Library (Dār al-Kutub), MS Taymūr ṭibb 311 (C1)
Madrid, Escorial, MS 799 árabe (E1)
London, Wellcome Library, MS Arabic Or. 14a (W2)
Given that the Arabic manuscript tradition is heavily contaminated, it is difficult to establish a stemma. Yet, the tradition can be divided into two main groups, the first represented by C1 and E1, and the second represented by W2. In general, the readings of the first group are significantly more accurate, whereas the second group shares many obvious mistakes and lacunae. Still, there are several instances in which the second group offers a correct reading that is absent in the first group. Therefore, the three manuscripts C1, E1 and W2 offer the most authoritative readings and are cited systematically. Moreover, C1 often contains variant readings and corrections in the margins (designated as C1marg), and it is highly probable that it was collated against E1 or an antecendent. W2, the representative of the second group, contains many minor mistakes, scribal errors and omissions, and therefore must be used with caution. In editing the extracts below, the primary aim was always to restore the text to what was originally produced by the Arabic translator. This required the systematic comparison between the source and target texts.
A quarter of a century ago, Ivan Garofalo produced an article on the relationship between the Arabic translation of De loc. aff. and the Greek original.Footnote 5 Garofalo based his study on W2, which often preserved an inferior Arabic text, and on Kühn's often faulty Greek printing, as no critical text was available at the time. This has led to some problems: his list of seven ‘superior readings against the whole Greek tradition (“lezioni poziori … contro tutta la tradizione greca”)’, for instance, does not offer any reliable textual evidence: the alleged readings are due to mistakes in his Greek and in the Arabic sources.Footnote 6 Moreover, Manfred Ullmann carefully analysed the translation technique employed in the Arabic version of De loc. aff., arguing against Garofalo that the translator must have been Ḥunayn rather than Ḥubayš.Footnote 7 Yet, neither of these studies has yet established how the Greek relates to the Arabic textual tradition.
ANALYSIS OF THE RELATIONSHIP BETWEEN THE GREEK AND THE ARABIC TEXTUAL TRADITIONS
To answer this question, we used a two-pronged approach, taking the source and the target texts, respectively, as starting points. First, we compared the Arabic translation systematically with the Greek text and looked for significant discrepancies. Here we found a number of lacunae as well as other differences. Second, we looked at all the instances in the Greek critical text established by Dr Florian Gärtner where there are significant differences among the various strands of the Greek manuscript tradition; this systematic review had to be limited to the first two books, as they were the only ones available in a critical edition.Footnote 8 In each case, we ascertained whether the Arabic translation would support a particular branch of the stemma. Obviously, there were many cases where it was impossible to tell from the Arabic translation with which readings it would side, as, for instance, in the case of particles or in the presence or absence of words that can be understood from the context.Footnote 9
This twofold survey yielded quite a number of instances where both approaches highlighted discrepancies (for instance examples 2 and 10 below). Yet, in the majority of the cases, one or the other approach led to the discovery of textual differences. In the following, we shall begin with some of the most striking examples, involving lacunae, and then move to other significant variant readings contained in the Arabic textual tradition. We have, however, excluded the majority of ambiguous cases, although we discuss two below (examples 8–9), because they can offer an important methodological caveat. Let us begin with a case where the Arabic translation allows us to fill a substantial lacuna in the Greek.
Example 1 (page 278, lines 11–15 ed. Gärtner):
[1] ἀλλὰ καὶ αὐτῆς τῆς γαστρὸς τὸ στόμα τοῦ κύτους ὡσαύτως διορισθήσεται, [2] εἰ τῶν καταπινομένων δριμέων τὰ μὲν ἐν τῷ στόματι τῆς γαστρὸς ἐργάζοιτο τὴν δῆξιν ἢ κατωτέρω, <lacuna> [6] τὰ δ’ ἐν τῇ κατὰ τὸν θώρακα διόδῳ⋅ [7] τὰ γὰρ τοιαῦτα πάντα τῇ θέσει διορίζεται, [8] καθάπερ ἄλλα τῇ τῆς ὀδύνης ἰδέᾳ. [9] λεχθήσεται δ’ ἐπὶ πλέον ὑπὲρ ἀλγημάτων ἐν τῷ δευτέρῳ γράμματι.
[1] والتفريق بين القرحة التي Footnote 10تكون في فم المعدة أو في تجويفها يكون على هذا المثال .Footnote 11 [2] إذا ازدرد الإنسان شيئا من الأشياء الحريفة الحادة إن كانت القرحة في فم Footnote 12 المعدة، وجد له Footnote 13 وجعا في فم المعدة أو أسفل من ذلك بقليل [3] .وإن كانت في تجويف المعدة،[4] وجد Footnote 14 له وجعا في البطن في مروره ونفوذه هناك .Footnote 15 [5] وإن كانت القرحة في المريء، [6] وجد له وجعا في مروره ونفوذه في الصدر [7].فجميع ما هذا سبيله يميز بوضع العضو، [8] كما أن أشياء أخر يفرق بينها بنوع الوجع [9].وسنذكر أمر الأوجاعFootnote 16 بأكثر مما ذكرناه هٰهنا في المقالة الثانية.
[1] One differentiates between the ulcer in the orifice of the stomach and in its cavity in the following way. [2] If someone has swallowed something sharp and pungent, then if the ulcer is in the orifice of the stomach, he feels a pain in the orifice of the stomach or a little lower. [3] If it is in the cavity of the stomach, [4] then he feels a pain in the belly when it passes and penetrates there. [5] If the ulcer is in the oesophagus, [6] then he feels a pain in the chest when it passes and penetrates there. [7] All such things can be distinguished through the position of the [affected] part, [8] just as one can differentiate between other things through the kind of pain. [9] I shall discuss pain more than I have done here in the second book.
Galen discusses here how pain in certain parts of the body can help us identify the affected part. If one swallows something acrid, it will trigger pain where the ulcer is: at the orifice of the stomach; inside the stomach; or in the oesophagus. Yet, because of the lacuna in the Greek, this point is nearly totally obscured, as only pain at the orifice of the stomach is mentioned. Therefore, the Arabic helps us again to restore the meaning of the Greek source text, even if we should refrain from translating the Arabic back into Greek. The second example is one where even an early modern editor of Galen, John Caius (1510–73), suspected a lacuna; now, thanks to the Arabic translation, we have a better sense of how to fill it.
Example 2 (page 322, lines 1–5 ed. Gärtner):
[1] τὸ γὰρ ἔλαττον ἧττον ὀδυνήσει τοῦ πλείονος, [2] ὥσπερ γε καὶ τὸ λεπτότερον τοῦ παχυτέρου [3] καὶ τὸ μένον τοῦ κινουμένου [4] καὶ τὸ τῷ βιαίῳ τῆς ψύξεως ἀπολειπόμενον <…> [lac. ind. Nickel : in marg. scr. ‘la[cuna] adden[dum]. τοῦ σφοδρου [sic] ψύχοντος’ CaiusFootnote 17]. [5] ἀλλὰ τό γε τῆς ὀδύνης εἶδος ἐντέρου παχέος ἐστίν, [6] ἄν θ᾽ ὡς ὑπὸ σκόλοπος ἐμπεπαρμένου τις ἄν θ᾽ ὡς ὑπὸ τρυπάνου διατιτρᾶσθαι νομίζῃ.
[1] وذلك لأن السبب اليسير المقدار يحدث من الوجع أقل مما يحدثه السبب الكثير المقدار، [2] وكلما كان السبب الكثير المقدار أرقّ وألطف، كان ما يحدثه من الوجع أقل مما يحدثه السبب الذي يكون أغلظ وأثخن، [3] والسبب الساكن يحدث من الوجع أقل مما يحدثه السبب المتحرك، [4] والسبب الذي تكون قوة برودته وشدتها أيسر وأخف يكون ما يحدث عنه أيضا من الوجع أقل مما يحدث عن السبب الذي تكون قوة برودته أشد وأعنف .[5] وهذا النوع من الوجع إنما هو وجع يكون في الأمعاء الغلاظ، [6] إن كان مستمرا بمنزلة الوجع الحادث عن سلاءة أو مسلة مركوزة.
[1] For the cause that is small in quantity causes less pain than the cause that is large in quantity. [2] Whenever the cause that is large in quantity is thinner and subtler, the pain that it causes is less than that caused by the cause that is thicker and denser. [3] The cause that rests causes less pain than the cause that moves. [4] The cause whose coldness’ strength and forcefulness are smaller and lighter also causes less pain than the cause whose coldness’ strength and forcefulness are more extreme and more violent. [5] This kind of pain only occurs in the soft intestines, [6] if it persists, such as the pain caused by a thorn or a needle which has been firmly embedded [in the skin].
In the Greek text, there are a number of comparisons of the potential causes of pain; in §1, it begins with quantity, moving on to density (§2), movement (§3) and coldness (§4). Yet, in §4, the genitive of comparison is missing, a fact already noted by John Caius, who proposed the amendment τοῦ σφοδρου ψύχοντοςFootnote 18 in his copy of the 1538 Basle edition, now held by Eton College.Footnote 19 The Arabic translation confirms Caius's conclusion that something is missing here. The lacuna in the Greek text can be filled by the Arabic, which states: ‘also causes less pain than the cause whose coldness, strength and forcefulness are more extreme and more violent (yakūnu mā yuḥdiṯu ʿanhu ʾayḍan mina l-waǧaʿi ʾaqalla mimmā yuḥdiṯuhū ʿani l-sababi llaḏī takūnu qūwatu burūdatihī ʾašadda wa-ʾaʿnafa).’ This is quite an expansive translation, which may render merely the three words τοῦ σφοδρῶς ψύχοντος, or possibly more text. Again, as in the previous case, it would be unwise to venture into translating the Arabic back into Greek, as this would require a level of confidence that cannot be attained. In the next three examples, the Arabic translation also appears to fill lacunae, although we cannot be quite as certain as in the first two examples.
Example 3 (page 252, lines 7–10 ed. Gärtner):
[1] ζητεῖν δ’ ἐφεξῆς, [2] ὅστις ὁ τρόπος αὐτῷ τῆς βλάβης ἐστὶν, [3] ἆρά γε διάθεσις ἤδη μόνιμος, [4] ἢ γινομένη μὲν ἔτι, τὸ μόνιμον δὲ οὔπω προσειληφυῖα· <lacuna> [8] καὶ εἰ γινομένη, [9] πότερον ἐν αὐτῷ τῷ μορίῳ περιεχομένης τῆς ποιούσης τὸ πάθος αἰτίας, [10] ἢ διοδευούσης αὐτό.
[1] ثم نبحث بعد ذلك عن المضرة الواقعة بالفعل: [2] أيّ ضرب هي من ضروب المضرة، [3] أتغير هو قد استحكم حدوثه وبقي لابثا، [4] أم هو تغير Footnote 20 لم يستحكم بعد وإنما هو في حد ما هو ذا يكون. [5] فإن كان التغير قد استحكم حدوثه [6] والسبب المكون له محتبس في نفس العضو الذي به الآفة، [7] فالمضرة الواقعة بالفعل ثابتة .Footnote 21 [8] وإن كان في حد ما هو ذا يكون، [9] فهل السبب المكون له محتبسFootnote 22 في نفس العضو الذي فيه الآفة، [10] أو Footnote 23 إنما يكون السبب Footnote 24 المكون له يمر وينفذ في ذلك العضو.
[1] Afterwards we need to find out about the damage that actually occurs: [2] what kind of damage is it? [3] is it a change that is firmly established and that will remain stable; [4] or is it a change that has not yet been firmly established, and is still becoming what it will be. [5] If it is a change that has been firmly established, [6] and the cause that brings it [sc. the damage] about is contained in the same part of the body as the illness, [7] then the damage that actually occurs is stable. [8] And if it is still becoming, [9] is the cause that brings it [sc. the damage] about contained in the part of the body where the damage is, [10] or does the cause that brings it about pass and penetrate through this part of the body.
Here, it would appear that Galen sets up a number of dichotomies. Is the condition (διάθεσις, the change [taġayyur] wrought by the ‘damage [βλάβη]’) 1) stable (§3) or 2) still ‘becoming’ in the sense that it has not yet stabilized (§4)?Footnote 25 If stable (§5), and the cause is in the affected part (§6), then the damage is also stable (§7), or if ‘becoming’ (§8), then the cause is either contained in the affected part (§9) or just passing through (§10). In the Greek text, the first of the dichotomies (‘if stable’) is missing and not explained further, which diminishes Galen's point considerably. Yet, the Arabic text which has this dichotomy (§§5–8) is only present in one manuscript, E1. Therefore, one cannot exclude that a later Arabic author simply completed Galen's point. The reading in E1, however, could well go back to a lost Greek source and would remove some of the difficulty of this passage, already identified by Gärtner.Footnote 26
Example 4 (page 268, lines 14–19 ed. Gärtner):
[1] ἀλλὰ διορίσασθαι χρὴ [2] δόντα σιτία τῆς ἐναντιωτάτης φύσεως τῷ τρόπῳ τῆς διαφθορᾶς, [3] ἄρτον μὲν εἰ τύχοι καὶ χόνδρον ἐφ’ ὧν κνισοῦται, [4] μέλι δ’ ἐφ’ ὧν ὀξύνεται. <lacuna> [6] κἄπειτ’ ἐπισκοπεῖσθαι τά τ’ ἐμούμενα καὶ τὰ διαχωρούμενα, [7] πότερον ἅμα χυμῷ τινι, [8] τὰ μὲν χολώδει καὶ θερμῷ, [9] τὰ δὲ φλεγματώδει καὶ ψυχρῷ λαμβάνει τὴν κένωσιν, [10] ἢ χωρὶς χυμοῦ τινος βραχεῖάν τινα μεταβολὴν ἐσχηκότα.
[1] ولكن ينبغي لك Footnote 27 أن تميز ذلك [2] بأن تطعم العليل أطعمة هي في طبعها على غاية المضادة لنوع الفساد، [3] فتطعم في المثل من يتغير الطعام في معدته إلى الدخانية خبزا وخندروسا، [4] Footnote 28 وتطعم من يتغير الطعام في معدته إلى الحموضة عسلا. [5] ثم تأمرهما بالقيء إن كان يسهل ذلك عليهما، [6] Footnote 29 ثم تتفقد بعد ذلك ما يقذفه كل واحد منهما بالقيء، وما يخرج منه Footnote 30 بالغائط، [7] وتنظر ما يستفرغ مع Footnote 31 كل واحد من الطعامين أي Footnote 32 خلط من الأخلاط هو، [8] وهل يخرج Footnote 33 مع الطعام الأول خلط حار مراري، [9] ومع الطعام Footnote 34 الثاني خلط بلغمي بارد، [10] أم يخرج كل واحد من الطعامين خلوا من أن يكون معه Footnote 35خلط من الأخلاط، ويكون قد استحالت الأطعمة استحالة يسيرة.
[1] Yet, you have to differentiate here [2] by giving the patient food to eat that is in its nature extremely opposed to the type of corruption; [3] then accordingly feed bread and gruel to those whose food changes in the stomach into something smokelike, [4] whereas you feed honey to those whose food changes into something acrid. [5] Then order them to vomit, if they can easily bear this. [6] Then afterwards observe what they spit out through vomiting and what they secrete through stool, [7] whilst observing what kind of humour is expelled together with the two [kinds of digested] food, [8] whether with the first [kind of] food, a warm bilious humour is secreted, [9] and with the second [kind of food], a cold phlegmatic humour, [10] or whether the two [kinds of] food are secreted without any humour being present, whilst the food is only changed a little.
The sentence in §5 has no corresponding Greek text. It is possible that the Arabic translator is simply providing further context, and that there is no corresponding Greek sentence in the original source. It is, however, more likely that a portion of text was lost in the subsequent Greek tradition.
Example 5 (page 340, lines 8–9 ed. Gärtner):
[1] ἐγγίγνεται δὲ καὶ ὠτὶ [2] μετά τινος σφυγμοῦ, [3] καὶ τῶν ὀδόντων ταῖς μύλαις, <lacuna> [5] ἔστι δ’ ὅτε καὶ τοῖς ὀφθαλμοῖς, …
[1] وقد يكون هذا النوع من الوجع في الأذن أيضا Footnote 36 [2] ويكون معه شيء من الضربان، [3] ويكون من الأسنان في الأضراس، [4] وربما كان في الفم [5] وربما كان أيضا في العينين.
[1] This kind of pain also occurs in the ear, [2] accompanied by some sort of pulsating, [3] and in the molar teeth, [4] and sometimes in the mouth [5] and sometimes in the eyes.
Galen writes again about a certain type of pain that can occur in different places; here one extra location in addition to the ears, molars and eyes is mentioned, namely the ‘mouth’ (fam). This may merely be an interpolation by the translator or by a later scribe, or it could refer to a Greek text that was subsequently omitted for one reason or another. In the following two examples, the Arabic translation is at odds with the entire Greek tradition as transmitted in the extant manuscripts; they both involve a change from ‘being primarily affected’ (πρωτοπαθεῖν) to ‘being specifically affected’ (ἰδιοπαθεῖν).
Example 6 (page 372, line 28 ed. Gärtner):
ἀλλὰ πρωτοπαθεῖν … αὐτóν,
لكن على طريق أن العلة تخصه في نفسه،
in a way that the disease is specific to it itself
There can be no doubt that the expression ‘the disease is specific to it itself (anna l-ʿillata taḫuṣṣuhū fī nafsihī)’ renders ἰδιοπαθεῖν, as this is a well-established technical term. However, the whole Greek tradition, including F, has πρωτοπαθεῖν. The next example is similar.
Example 7 (8.308, line 4 ed. Kühn):
ὅταν αὐτὸς πρωτοπαθῇ κατά τινα καιρὸν ἕτερον, ἡ ἐντὸς μοῖρα τῶν μεσοπλευρίων συμπάσχει.
كما أنه قد يعرض في وقت آخر أن يكون الجزء الباطن من هذا Footnote 37 العضل الذي فيما بين الأضلاع يعتل مع هذا الغشاء بطريق المشاركة له في علته إذا اعتل علة تخصه في نفسه،
as it may happen at another time that the inner part of these muscles between the ribs is affected with this membrane by way of sympathy, when it is affected by a disease specific to it itself.
The subordinate clause ‘when it is affected by a disease specific to it itself (iḏā ʿtalla ʿillatan taḫuṣṣuhū fī nafsihī)’ renders the initial Greek ὅταν clause, and contains again the typical Arabic equivalent for ἰδιοπαθῇ, which must have been the underlying reading. The next two examples are less clear-cut and allow us to make a methodological point against facile inferences from the target to the sourcetext. The first also involves the term for primary affection again.
Example 8 (page 260, lines 22–4 ed. Gärtner):
οὐ γὰρ δὴ πρωτοπάθειάν [πρωτοπαθ A U: πρωτοπάθειά M: πρωτοπαθεῖν F C] γε αὐτὸ τηνικαῦτα [om. La] φήσομεν, ἀλλὰ δευτεροπάθειάν τε ἅμα καὶ ἰδιοπάθειαν [δευτεροπαθ́… ἰδιοπαθ́ A: δευτεροπάθεια … ἰδιοπάθεια M: δευτεροπαθεῖν … ἰδιοπαθεῖν F C]
فلسنا نقول إن تلك الآفة في ذلك العضو حينئذ آفة أولية، بل نقول إنها آفة نزلت به بعد غيره وخصته في نفسه معا،
If this is so, then we do not say that this ailment in this part of the body at this time is a primary ailment [āfa awwalīya]; rather, we say that it is an ailment that has come down to it [the part of the body] after something else and is, at the same time, specific to it itself.
First, τηνικαῦτα is clearly rendered as ‘ḥīnaʾiḏin (at this time)’. The Arabic appears to confirm the nominal forms πρωτοπάθειάν, etc. rather than the infinitives πρωτοπαθεῖν, etc., as we have the noun ‘āfa (ailment)’ in each case. Yet, this is too premature a conclusion. In order to grasp how the verb is rendered in the Arabic translation of De loc. aff., we provide a survey listing nine occurrences of the verb πρωτοπαθεῖν, translated in quite an astonishing array of ways:
1. Kühn 8.192, lines 9–10
τὸν ἐγκέφαλον ἡγητέον ἐπὶ τούτων πρωτοπαθεῖν
فينبغي لنا أن نعتزم من أمر هؤلاء على أن الدماغ هو المعتل علة حدثت فيه حدوثا أوليا
We need to assume in their case that the brain is diseased by a disease that occurs in it primarily [huwa l-muʿtallu ʿillatan ḥadaṯat fīhi ḥudūṯan awwalīyan].
2. Kühn 8.198, line 10
ἀνωτέρω τοῦ πρωτοπαθοῦντος μορίου
فوق الموضع الذي كان حدوث العلة فيه حدوثا أوليا
above the place in which the occurrence of the diseases was primary [kāna ḥudūṯu l-ʿillati fīhi ḥudūṯan awwalīyan]
3. Kühn 8.200, line 4
ἀπὸ τοῦ πρωτοπαθοῦντος μέρους
من العضو الذي حدثت به الآفة حدوثا أوليا
from the place [of the body] in which the ailment occurs primarily [ḥadaṯat bihi l-āfatu ḥudūṯan awwalīyan]
4. Kühn 8.203, line 12
πρωτοπαθούσης τῆς κεφαλῆς
عندما تحدث الآفة بالدماغ حدوثا أوليا
when the ailment occurs in the brain primarily [ʿindamā taḥduṯu l-āfatu bi-l-dimāġi ḥudūṯan awwalīyan].
5. Kühn 8.203, lines 17–18
ἀπὸ τῆς κεφαλῆς πρωτοπαθούσης
من قبل الرأس إذا حدثت به الآفة حدوثا أوليا
because of the head when an ailment occurs in it primarily [min qibali l-raʾsi iḏā ḥadaṯat bihi l-āfatū ḥudūṯan awwalīyan]
6. Kühn 8.211, line 15
τούτῳ πρωτοπαθεῖν ἐλογιζόμεθα τοῦ νωτιαίου τὸ κατωτέρω …
فعلمنا أن الذي حدثت به الآفة من نخاعه حدوثا أوليا هو الجزء منه الذي أسفل …
We know that the part of the spinal marrow in which the ailment occurs in a primary manner [allaḏī ḥadaṯat bihi l-āfatu min nuḫāʿihī ḥudūṯan awwalīyan] is the part of it that is lower …
7. Kühn 8.227, lines 12–13
οὐκ ἐκ τοῦ αὐτοῦ δὲ πρωτοπαθοῦντος ὁρμᾶται τόπου
إلا أنها ليست تتحرك من Footnote 38 موضع واحد بعينه إذا كانت علته علة أولية بنفسه
but it does not move from the exact same place, when its disease is itself a primary disease [iḏā kānat ʿillatuhū ʿillatan awwalīyatan bi-nafsihī].
8. Kühn 8.263, line 11
τούτῳ πρωτοπαθεῖ
حدثت فيه حدوثاً أولياً،
it occurs in it in a primary manner [ḥadaṯat fīhi ḥudūṯan awwalīyan]
9. Kühn 8.298, line 5
διορίσασθαι χρὴ τίνα μὲν ἰδιοπαθούσης ἢ πρωτοπαθούσης
فتميز الأعراض التي تعرض له على طريق العلل الحادثة فيه خاصةً أو فيه أولاً
then distinguish the symptoms that supervene in it by way of diseases occurring specifically or primarily [ʿalā ṭarīqi l-ʿilali l-ḥādiṯati fīhi ḫāṣṣatan aw fīhi awwalan]
Therefore, to translate πρωτοπαθεῖν, we sometimes simply have the verb ḥadaṯa with an inner accusative of the adjective awwalī (‘to occur primarily’) either with (see no. 9) or without (see no. 8) the verbal noun ḥudūṯan. In the majority of the cases, we find a circumlocution ‘an ailment [āfa] occurs primarily’ (see nos 3–6) or ‘a disease [ʿilla] occurs primarily’ (see no. 2). In one instance (see no. 7), the word disease is directly modified by the adjective ‘primarily’. It is thus possible that the Greek exemplar on which the Arabic version is based still read the infinitives rather than the nouns. Given the inconclusiveness of the evidence provided, it is best to suspend judgement. The same is true in the case of the next example.
Example 9 (page 252, lines 14–16 ed. Gärtner):
τὸ δὲ τηνικαῦτα μόνον ἡγεῖσθαι πεπονθέναι τὸ μόριον, ὅταν τῆς οὐσίας αὐτοῦ μόνιμος ἀλλοίωσις [διάθεσις F Q] γένηται, νομοθετούντων ἐστὶν ἴδια σημαινόμενα τῶν Ἑλληνικῶν ὀνομάτων.
والاعتقاد بأنّ العضو إنما تكون به آفة ومرض، عندما يحدث في نفس جوهره تغير مستحكم لابث فقط، إنما هو من شأن من يستنّ في الأسماء الخاصية بمعاني الأمور سُنة لم تكن.
Thinking that the part [of the body] in which there is pain and disease is only affected, when in its essence an established and stable change occurs, is typical of those who follow their own specific [linguistic] conventions about the meaning of words that do not exist.
One might think that the Arabic clearly supports the reading ἀλλοίωσις (‘change’), as this is the meaning of the corresponding taġayyur. Yet, one must consider the context. In the text just preceding this sentence, the Arabic translator repeatedly rendered διάθεσις with taġayyur, including in example 3 discussed above. Moreover, taġayyur is also employed to render διάθεσις later in the discussion.Footnote 39 In general, διάθεσις is translated in a variety of ways in the Arabic version of De loc. aff., ranging from ʿilla (disease), maraḍ (disease), āfa (harm), ḥāl (condition), alam (pain), and so forth.Footnote 40 Therefore, it is likely that the Greek exemplar could have had the reading διάθεσις, although ἀλλοίωσις is also a possibility. Let us now turn to five examples where the Arabic sides with F against the majority tradition, be it on its own (examples 10–11) or with some other manuscripts (examples 12–14).
Example 10 (page 278, lines 2–3 ed. Gärtner):
τὸ μὲν γὰρ ἐμούμενον [sc. ὑμενώδους χιτῶνος μόριον] ἤτοι τῆς γαστρὸς ἢ τοῦ στομάχου τὸ πάθος εἶναι σημαίνει⋅ τὸ δ’ ἀναβησσόμενον ἤτοι τοῦ λάρυγγος ἢ τοῦ πνεύμονος [om. F C] ἢ τῆς τραχείας ἀρτηρίας [add. ἢ καὶ τοῦ πνεύμονος C]⋅
وذلك أن ما خرج من هذا [جزء من طبقة غشائية] بالقيء يدل على أن العلة والآفة إما في المعدة وإما في المريء، وما خرج منه بالسعال يدل على أن الآفة والعلة إما في الحنجرة وإما في قصبة الرئة.
For when it [sc. part of the membranous layer] is excreted through vomiting, it indicates that the cause and the disease are located either inside the stomach or inside the oesophagus; when it is excreted through coughing, it indicates that the cause and the disease are located either in the voice box or in the wind pipe.
Galen discusses here how to determine the affected part through excretions. All the Greek manuscripts apart from F have the reading ἢ τοῦ πνεύμονος (‘or the lungs’), with only C putting it at the end of the colon. The Arabic translator's Vorlage, however, omits it, thus siding with F against all the other Greek manuscripts. There is good reason to believe that this is Galen's original reading, since it is the lectio difficilior; a later scribe would then have added the reference to the lungs in analogy to the stomach and the oesophagus.
Example 11 (page 324, line 3 ed. Gärtner):
καὶ ναυτία [ναυτίαι F] μέχρι πολλοῦ παραμένουσα [παραμένουσαι F] κενὴ [κεναὶ F : correxit Gärtner : καινὴ (supra lineam corr.) ναὶ μὴν Lb : καὶ μὴν ζ] …
والحركة إلى القيء من غير أن يقذف شيئا، ويدوم ذلك به مدة طويلة …
and an urge to vomit without anything being expectorated, whilst this lasts for a long time …
Here the Arabic translation clearly supports the reading ναυτίαι μέχρι πολλοῦ παραμένουσαι κεναὶ of F, the reading also adopted by Gärtner, although he favoured the singular ναυτία … παραμένουσα κενή. The meaning of ‘empty nausea’ (the feeling of vomiting without actually being able to vomit anything) is clearly correct, and one does not need the Arabic translation to see that the variant καινή makes no sense. Gärtner does not explain why he adopted the singular, but at first the Arabic could be seen to support his choice, as the expression ‘al-qayʾ min ġayri an yaqḏifa šayʾan (lit. the vomiting without his [the patient's] expectorating anything)’ is also in the singular. Yet, in an earlier passage where Galen talks about empty nausea, he also employs the plural αἱ … ναυτίαι … κεναὶ, which is rendered through a different expression, namely the hendiadys ‘al-ġaṯayān wa-l-ḥaraka ilā l-tahawwuʿi (lit. the nausea and the movement towards vomiting)’.Footnote 41 Therefore, the Arabic singular cannot be used to confirm the singular in the Greek.
Example 12 (page 298, line 21 ed. Gärtner):
οὕτως γὰρ ὑπάρξει σοι μόνῳ [μόνως F Q C] διακρίνειν δύνασθαι
فإنك بهذا الطريق وحده تصير في حد ما Footnote 42 يمكنك معه أن تميز وتعلم
For in this way alone [waḥdahū] you come to be able to distinguish and know
The Arabic supports the reading μόνως here, as can be seen from the translation: ‘alone (waḥdahū)’ clearly modifies the manner (Greek οὕτως) rather than ‘you (σοι)’. The next example offers a conundrum difficult to interpret, yet it is clear that the Arabic translator must have had the reading found in F.
Example 13 (Kühn 8.230):Footnote 43
ἡ δ’ οὖν γλῶττα πολλάκις ἑώραται βλαπτομένη τὴν εἰρημένην βλάβην μόνη τῶν κατὰ τὸ πρόσωπον μερῶν, μήτε τῆς ἁπτικῆς αἰσθήσεως ἐν αὐτῇ [αὐτῷ F L] μήτε τῆς γευστικῆς βεβλαμμένων. καὶ ἡ αἰτία πρόδηλος ἡμῖν ἐστιν ἑωρακόσιν …
The tongue is often seen to be the only part of the body in the face to suffer the damage that was previously mentioned, without the sense of touch nor the sense of taste in it [sc. the tongue] having suffered damage. The reason for this is abundantly clear …
وأما اللسان فنجده عياناً مراراً كثيرة تناله المضرة التي ذكرناها وحده، وتكون الأجزاء التي في الوجه لم تنلها مضرة في حسها اللمسي ولا في حركاتها الإرادية، والسبب في ذلك قد تقدمت عندنا معرفته،
We often see with our own eyes that the tongue alone suffers the aforementioned damage, whilst the [other] parts of the body in the face do not suffer any damage to their sense of touch nor their voluntary movement [wa-lā fī ḥarakatihā al-irādīyati]. We have previously come to know the cause for this …
The difficulty in the Greek text partly arises from the interpretation of what is meant by τὴν εἰρημένην βλάβην. As the Greek text stands, it would appear that the damage is to the ability to move, as Galen had previously discussed the causes for the partial or total paralysis of the tongue. The genitive τῶν κατὰ τὸ πρόσωπον μερῶν would then depend on μόνη (‘the only part of the body in the face’), and the subsequent genitives τῆς ἁπτικῆς αἰσθήσεως and τῆς γευστικῆς form a genitive absolute together with the perfect passive participle βεβλαμμένων (‘with neither the sense of touch nor the sense of taste being damaged’). The reading ἐν αὐτῇ then refers to the tongue: the senses of touch and taste in the tongue remain unaffected, even if it is paralysed. Only the tongue suffers paralysis, but when it does it still senses touch and taste.
The interpretation in the Arabic translation, however, is significantly different. Here the damage to the tongue alone is contrasted with a lack of damage in the other parts of the face: the latter do not lose their ability to sense touch and move voluntarily. Therefore, the Arabic translator believed that τὴν εἰρημένην βλάβην referred to the earlier discussion about damage to the ability of the tongue to move voluntarily. This obviously also presupposes a different reading from the one present in the transmitted Greek text τῆς γευστικῆς. Elsewhere in De loc. aff., the Arabic expression ḥaraka irādīya (‘voluntary movement’) translates Greek κίνησις ἡ καθ’ ὁρμὴν (for example ed. Gärtner 374.22–3; ed. Kühn 209.9–10) or κατὰ προαίρεσιν κίνησις (for example ed. Kühn 8.171.8–9); therefore, one could speculate whether the Arabic translator read τῆς καθ’ ὁρμὴν κινήσεως instead of τῆς γευστικῆς, although the two readings are obviously not close from a palaeographical point of view. Finally, the translator clearly read ἐν αὐτῷ with F, and understood the antecedent to be the face (πρόσωπον, waǧh), not the tongue (γλῶττα, lisān). The final example for the Arabic tradition's siding with F is much more straightforward.
Example 14 (page 322, line 17 ed. Gärtner):
χυμὸς δέ τις, οἷον εἶπον ὑαλώδη [om. F Q N η]
بل يخرج من العليل خلط على هذه الصفة التي وصفتها لك قبل
rather, from the patient there comes out a humour which is of the same description that I have given to you beforehand
The Arabic clearly does not have ὑαλώδη, thus siding with F (as well as Q N η). Finally, there is one instance where the Arabic sides against F but with other Greek witnesses:
Example 15 (page 314, lines 18–19 ed. Gärtner):
κατά τε τὰ παχέα καὶ [παχέα καὶ om. ζ] χαλαρὰ μέρη τῶν πλευρῶν
في الأجزاء الرخوة السلسة التي بين الأضلاع
in the limp [raḫwa] and smooth [salisa] parts which are between the ribs
The Arabic clearly sides with ζ against the rest of the Greek manuscript tradition, including F, as the hendiadys ‘limp and smooth [raḫwa salisa]’ renders χαλαρά.Footnote 44
CONCLUSIONS
From these examples we can draw three main conclusions. First, the Arabic translation can clearly help us improve the Greek text, notably by filling lacunae in the Greek manuscripts. Second, we can now offer a suggestion as to where the Arabic version belongs in the Greek stemma. On the one hand, it often sides with F, the oldest extant manuscript, sometimes against all the other Greek witnesses. Yet, in the last example, it sides with ζ against F. Moreover, in the case of examples 1–7, it goes against F (as well as all the other Greek witnesses). This textual situation allows for only one conclusion: the Greek exemplar of the Arabic translation must precede the archetype ω of all the extant Greek manuscripts. This makes eminent sense, given that Ḥunayn and his team of translators, including his nephew Ḥubayš, worked in the ninth century, some three hundred years before F was written. This underscores the importance of consulting the indirect Arabic tradition for the constitution of the Greek text. Therefore, modern editors should be wary of going against the readings of this lost Greek exemplar, unless they have strong reasons to suspect corruption in it; in the examples discussed above, we would have favoured its readings, even when they go against all the extant Greek manuscripts. When editing De loc. aff. (as well as other Greek texts for which there is a strong Arabic secondary tradition), one should adopt a similar method, provided that these readings can be clearly established.
This brings us to our third and last point, which we cannot argue here fully owing to the constrains of space. Examples 8 and 9 demonstrate that one should not be tempted to draw quick inferences from the target to the sourcetext but rather consider them holistically and in the general context of other Graeco-Arabic translations. Too often have scholars relied on facile equivalences that obscure rather than shed light on the state of the Greek tradition, as Aristotle's Poetics amply illustrates.Footnote 45 Despite this caveat, it can only be hoped that more critical editions of Greek texts will take full account of the secondary Arabic evidence when available.
