I

Robinson has put forward a single argument, in two variations: it is based on the presence in MQ of a term that supposedly was not part of Samuel Ibn Tibbon's vocabulary at the purported time of the translation. The first version of Robinson's argument is this: “The translation […] uses terms such as ḥippus that Ibn Tibbon, according to his own testimony (see below), only coined much later, after 1213.”Footnote ⁴ Let us consider this argument. It is grounded in the following passage from Ibn Tibbon's “Introduction” to his Commentary on Ecclesiastes:

When did Ibn Tibbon coin the term here so well presented? Robinson observed that there is one occurrence of the word istiqrāʾ in the Arabic text of the Guide of the Perplexed (3:12) and that there Ibn Tibbon did not render it by ḥippus.Footnote ⁶ From this he concluded that when translating the Guide, Ibn Tibbon had not yet coined the term and did so only after completing that translation in 1204 (revised in 1213).Footnote ⁷ He consequently argued that since the term ḥippus occurs in the translation of MQ its ascription to Samuel Ibn Tibbon must be wrong.

This reasoning does not hold water, and for several reasons. First, the key term ḥippus, on which the entire argument rests, does not occur in all the manuscripts of MQ: I did not find the term in the manuscript (Paris, BNF, héb. 1115) which I had transcribed in full (and could therefore computer search); I could not check its occurrence in other manuscripts because Robinson did not specify any reference (manuscript and folio number) for the alleged occurrence of the term. In any event, the question is not crucial. Unsurprisingly, many terms occurring in some manuscripts of MQ do not appear in others. As will repeatedly be seen in the sequel, the text of MQ is not stable: readers and copyists tinkered with the text, replacing one term by another. Therefore, if a term appears in one or more (but not in all) manuscripts of MQ this does not at all imply that it goes back to the original text of MQ. Hence, such an occurrence cannot be sound basis for any conclusion concerning Ibn Tibbon's terminology without a thorough examination of the manuscript tradition. In point of fact, however, assuming the term indeed occurs in at least one manuscript, I would tend to think that the very technical term ḥippus was not interpolated into the text of MQ and that Ibn Tibbon indeed used it already in 1199. We shall however see that this is in no way incompatible with his authorship of this translation.

Second, contrary to Robinson's statement, in the passage from the Commentary on Ecclesiastes quoted above, Ibn Tibbon does not claim that he coined the word when he wrote that work; he says only that he coined it, without any temporal reference. In addition, in the single passage in the Guide in which Maimonides uses the term istiqrā’, Ibn Tibbon's translation is very elegant and precise without using ḥippus; comparison with other translations of the Guide confirms that its use was not at all indispensable.Footnote ⁸ Thus the claim that the term ḥippus was introduced only after the translation of the Guide was completed is unfounded. In point of fact, the term ḥippus does occur in it at least once (Guide 1:50), as a translation, however, of baḥṯ (rather than of istiqrā’). The following can also be added: in the passage quoted from Commentary on Ecclesiastes, Ibn Tibbon says: “I shall explain what I mean by ‘induction’ [ḥippus], when I use it here and elsewhere”; since (as Robinson has correctly noted) the only work preceding the Commentary on Ecclesiastes in which Ibn Tibbon used the term is MQ, it would seem that when he wrote “and elsewhere” he was thinking precisely of this work.

In the 2014 version of the entry to The Stanford Encyclopedia of Philosophy Robinson did not repeat the above argument against the attribution of the translation of MQ to Ibn Tibbon, although he did not explicitly withdraw it. Instead, he brought up a new version of the same argument, now based on the presence in MQ of another term: šem nirdaf (= synonym):

The term nirdaf indeed occurs in some manuscripts MQ,Footnote ¹⁰ and, as we shall see, it goes back to Samuel Ibn Tibbon himself, who introduced it in an early phase of his career as a translator. This argument, too, is therefore invalid.

I conclude that Robinson's two arguments against the ascription of MQ to Ibn Tibbon are flawed. Moreover, in both versions, the argument is also self-contradictory. For the suggested conclusion follows only under the assumption that if Ibn Tibbon translated MQ he did so in 1199, as stated in one of the two colophons, and therefore prior to writing the Commentary on Ecclesiastes. The rejection of one part of the colophon (to the effect that the translation was made by Ibn Tibbon) thus entirely rests on the premise that its other part (concerning the date) is trustworthy. It seems much sounder methodology to assume that either the entire colophon is untrustworthy – and in this case we know nothing about the date of composition of MQ – or that it is trustworthy in its entirety, including both the date and the authorship.

Obviously, however, this refutation of the above arguments does not positively establish Ibn Tibbon's authorship. This task will be undertaken in what follows. There are in fact very good arguments positively corroborating Samuel Ibn Tibbon's authorship of the translation of MQ: first, the testimony of the manuscripts; second, an inquiry into the evolution of Ibn Tibbon's vocabulary; and third, some doctrinal statements of Ibn Tibbon that suggest that he was acquainted with MQ at a very early phase of his career. We shall consider them in turn.

II

The ascription of the translation of MQ to Samuel Ibn Tibbon appears in two (out of six) complete manuscripts:

MS Paris, BNF, héb. 1114 (14th century), fol. 124^r:

London, BL, Add 18689 (15th century), fol. 104^v:

The possibility that a colophon was forged and interpolated in one manuscript, from which it subsequently found its way to the other, seems rather unlikely (there are numerous differences in the texts of these two manuscripts, which thus seem to be independent).Footnote ¹² Unless unequivocal evidence to the contrary is produced, we have no reason to doubt the trustworthiness of the two converging colophons.

III

Ibn Tibbon's authorship is corroborated by considerations having their point of departure in Perush ha-milim ha-zarot (= PMZ; “Explanation of unfamiliar terms”), a small work in which Ibn Tibbon explained a number of mainly philosophical and scientific Hebrew terms, notably (but not exclusively) such that he used in his translation of the Guide, and that he thought might be “unfamiliar” to readers.Footnote ¹³ We will have to content ourselves with a few glimpses at the evolution of Ibn Tibbon's philosophical vocabulary, whose systematic study is a desideratum. In PMZ, Ibn Tibbon repeatedly observes that when he translated the Guide he occasionally changed his mind about the best way to render in Hebrew a given Arabic term or expression; here I will follow some of these indications.

The following methodological point must be stressed. Ibn Tibbon explains that when he decided to replace one term with another in his translation of the Guide, he introduced the changes into his own “master copy.” However, he warns, copies of the translation had already been made and spread “in the country”; accordingly, he asks the readers to correct their copies.Footnote ¹⁴ We indeed observe that in all or almost all copies of the Hebrew version of the Guide (including the printed editions) there is no consistency in the use of terms that had been changed – each copy is the unique sum-total of numerous scribal “corrections” along the transmission process. The same holds also of PMZ, which was written and modified over a number of years.Footnote ¹⁵ The manuscripts of MQ, too, have undergone similar processes of terminological revision followed by “contamination.” Therefore, when studying the terminology used in PMZ, in the Guide, and in MQ we should not expect a clear-cut absence or presence of terms, but rather statistically significant frequencies.Footnote ¹⁶

1. In PMZ,Footnote ¹⁷ Ibn Tibbon writes that he replaced the word על by כפי or לפי, especially in the expression על דעת, which he changed into כפי דעת. A perusal of the Guide in his translation (in the current edition) con-firms that he indeed did so: there are about 70 occurrences of כפי דעת and only some 10 of על דעת. The use of this phrase is thus a reliable criterion by which to determine Ibn Tibbon's authorship of a given text after 1213. In MQ there are nine על דעת of the early period but only one כפי דעת of the later period.

2. In PMZ,Footnote ¹⁸ Ibn Tibbon writes that in a first phase he used the expression על מה שהוא עליו but then replaced it by כפי מה שהוא. This change is confirmed by a perusal of the Guide: of the first expression there is only one occurrence, of the latter over twenty. In MQ there are four occurrences of על מה שהוא עליו and about 10 more of על מה שהוא (also one על הענין אשר ימצאהו עליו) of the early period, as compared with only one occurrence of the late expression כפי מה שהוא עליו and one of לפי מה שעליו. (There are also other locutions, such as הוא מה שעליו and בענין אשר הם עליו.) Thus, despite some “noise” (the result of a still fluid terminology and/or of scribal alterations) MQ displays a preference for the older expression (על מה שהוא (עליו, as was also the case with על/כפי דעת.

3. In PMZ, Ibn Tibbon explains his usage of the verb הרס and its derivatives.Footnote ¹⁹ Drawing on the sense it has in Exodus 19:21 (“Warn the people not to break through to the Lord to gaze”), Ibn Tibbon, probably following Maimonides's advice, ascribes it the meaning of “effrontery,” specifically “investigating that which should not be investigated,”Footnote ²⁰ an innovative usage that he used in various places in his writing.Footnote ²¹ It is therefore noteworthy that in MQ, too, there is one occurrence of it in a related sense:

In Ibn Tibbon's translation, Ibn Riḍwān here says that the physician should treat the patient by employing causes opposite to those active in causing the disease “in orderly fashion,” and not act suddenly and rashly (יהרסו), which could harm the patient.

4. Let us now attend to one terminological usage that at a first blush may seem to contradict my thesis. In PMZ, Ibn Tibbon devotes an entry to the term שמות באים זה אחר זה. He explains that it refers to a situation in which one and the same thing has different names, i.e., the names are synonyms.Footnote ²³ In the last sentence of the entry he adds that he resolved to apply to such names the term שמות תכופים, rather than שמות באים זה אחר זה .Footnote ²⁴ However, in MQ we find neither of these two terms. The notion of synonym is infrequent in MQ (it appears only in comments by Ibn Riḍwān); the various manuscripts use two other terms, namely šemot nirdafim and šemot nimšakim:Footnote ²⁵

and:

How can the occurrence of these two terms in manuscripts of MQ be squared with the fact that they are not mentioned in PMZ? Does not the disagreement between Ibn Tibbon's announcement in PMZ and the use of the terms denoting synonyms that we find in manuscripts of MQ disprove the claim that it was translated by Ibn Tibbon?Footnote ²⁸ Not at all. We will now see that Ibn Tibbon's writings display a continuous evolution in the choice of terms denoting the notion of “synonym”; this will throw light on the presence in manuscripts of MQ of the two (synonymous) terms šemot nirdafim and šemot nimšakim to denote synonyms.

In the Guide, the notion of synonym does not seem to come up; accordingly, no terms denoting it appear in the translation.Footnote ²⁹ Ibn Tibbon apparently used a term for “synonym” for the first time in his “Translator's Preface” to the translation of Maimonides’ Eight Chapters (the introduction to Tractate Avot), written immediately after completing the translation of the Guide. Ibn Tibbon there explains that when the prophet says השכל וידֹע אותי (Jer. 9: 23) the two words השכל and ידֹע (which he takes to be nouns) are not synonyms: they denote two different objects, because, he writes:Footnote ³⁰

Thus, in this Preface, written a little more than a decade after MQ, Ibn Tibbon referred to synonyms with the term šemot rodefim, which is a slight variation on (and derived from the same root as) the term šemot nirdafim used in MQ. (I note en passant that this continuity suggests that in MQ šemot nirdafim was the original term used by Ibn Tibbon and that šemot nimšakim is a later interpolation; this hypothesis is confirmed by other considerations, discussed below).

Some manuscripts of the Preface carry a later gloss by Ibn Tibbon himself (it is not a later interpolation by a reader). There he writes:Footnote ³¹

Thus, in the added gloss Ibn Tibbon denotes the notion of synonym by the term שמות באים זה אחר זה found in PMZ. It seems clear that this term is an innovation, introduced in the gloss and in PMZ to replace šemot nirdafim/rodefim used in the earlier MQ and in the body of the “Translator's preface.”

In the later Commentary on Ecclesiastes, we find both terms:

We see that Ibn Tibbon found yet another way to denote the notion of “synonym,” namely adding to rodefim the words zeh et zeh. The resulting term, šemot rodefim zeh et zeh, is thus situated midway between the early šemot rodefim and the later šemot ba'im zeh aḥar zeh. During the period in which Ibn Tibbon wrote his Commentary on Ecclesiastes he seems to have still wavered on the best terminology for “synonym” and alternated use of the older and newer expressions, perhaps with the unrealized intention of standardizing the usage at some later point. (As already noted, he himself indicates apropos of the Guide that in such cases he occasionally overlooked occurrences of the old term, which thus lingered in the text; and also that the older version of the work had already been copied and the copies could not be corrected.Footnote ³⁴)

Only in Maamar yiqqawu ha-mayim, Ibn Tibbon's last work, do we finally find a stable systematic usage of a single term, namely שמות באים זה אחר זה לענין אחד.Footnote ³⁵ When he wrote this work, Ibn Tibbon consistently used this phrase from the outset.

The inquiry into the terms Ibn Tibbon employed for “synonym” thus suggests (subject to further refinement based on more manuscript evidence) that his terminological choice underwent a continuous evolution. In the early MQ he used šemot nirdafim, a term he soon changed (in the “Translator's Preface” to “Eight Chapters”) to the related šemot rodefim. Somewhat later, he complemented rodefim with the words zeh et zeh, of which a single occurrence is preserved. Next he replaced the rodefim by ba'im, creating the term šemot ba'im zeh aḥar zeh, which first appears in the gloss to “Translator's Preface” and was formally presented in the dedicated entry in PMZ. This remained Ibn Tibbon's definitive choice. We thus have here a natural smooth evolution from a calque translation (mutarādifa → nirdafim) to a rendering that is increasingly removed from the source language. As for šemot teḵufim, also announced in PMZ (at least in the printed version), my limited search did not turn up any evidence that Ibn Tibbon ever used it,Footnote ³⁶ nor does it seem to have been much used by posterity. From this reconstruction of the evolution of Ibn Tibbon's terminology, it clearly follows that the term šemot nimšaḵim, which replaces šemot nirdafim in some manuscripts of MQ, is a later emendation.Footnote ³⁷ The bottom line is that the occurrence in MQ of the term šemot nirdafim is entirely consistent with the attribution of this work to Samuel Ibn Tibbon.

5. Last but not least, we can also attribute to Samuel Ibn Tibbon another terminological innovation, whose occurrence in MQ on the one hand, and in two works whose attribution to Ibn Tibbon is unquestioned on the other hand, corroborates their common authorship. This is the term מתדמה (ה)חלקים, a calque translation of the Arabic mutašābih al-ajzā’, denoting homoeomerous parts. It appears in MQ dozens of times, a few times in Otot ha-shamayim,Footnote ³⁸ and once in the Commentary on Ecclesiastes. Footnote ³⁹ (The term is not mentioned in PMZ, perhaps because it is very specialized and also does not appear in the Guide.Footnote ⁴⁰) It seems that מתדמה (ה)חלקים had not been used by any previous author, notably not by Judah Ibn Tibbon, and the occurrences in MQ and a little later in Otot ha-shamayim and Commentary on Ecclesiastes are apparently the first. The concomitance of this newly coined term in these works lends additional support to the claim that they all issued from the same author.

6. There is another pair of terms about which Samuel Ibn Tibbon changed his mind, so that we might consider taking them as another criterion for verifying his authorship of MQ. In PMZ, Ibn Tibbon informs us that he wavered about the use of the terms ראיה and מופת, but finally chose to use מופת exclusively for demonstrative proofs and ראיה for any kind of proofs, and that he later corrected his copy of the Guide accordingly.Footnote ⁴¹ However, the text of MQ is replete with occurrences of both terms, in a medical sense (indication for an illness) but also with an epistemological meaning (proof): often the two cannot be clearly distinguished, so that such an investigation seems impossible.

IV

A comparison between MQ and PMZ also reveals a number of shared ideas, showing that the author of PMZ was very well acquainted with MQ and borrowed from it. This lends additional support to the claim that Ibn Tibbon is the translator of MQ.

1. In PMZ, Ibn Tibbon writes:Footnote ⁴²

This short text bears strong resemblance to the following two passages in MQ:

What is significant here is above all the substantial proximity in content of the passages in MQ and in PMZ. In the latter, Ibn Tibbon defines geder by saying it denotes its object by indicating its עניינים עצמיים, i.e. the components of its essence. This clearly echoes ‘Ibn Riḍwān's statement: הגדר העצמי הוא מה שיבאר עצם הדבר הגדור בדברים העצמיים.

As to terminology, let us observe only the following. In the passage quoted from Ibn Riḍwān, to'ar and ḥoq are treated as synonyms, both translating the Arabic term rasm (one definiendum is distinguished from another through inessential differences), whereas in the passage quoted from PMZ only ḥoq appears and not to'ar. This discrepancy is again the result of the evolution of Ibn Tibbon's vocabulary. In fact, in PMZ Ibn Tibbon allocated to'ar a dedicated entry in which he defined the term as follows: “a name specifically assigned to [i.e. denoting] a form perceived by the senses, be it natural or artificial.”Footnote ⁴⁶ This meaning obviously has nothing in common with the meaning the term had in MQ. The reason for this semantic shift is not difficult to fathom: Ibn Tibbon now defined to'ar so as to use this term as an equivalent of takṭiṭ in Guide 1:1 (following Maimonides's own usageFootnote ⁴⁷). Thus the differences in meaning of to'ar in MQ on the one hand and in PMZ and the Guide on the other are in no way incompatible with the attribution of MQ to Ibn Tibbon.

2. A further significant instance of Ibn Tibbon following an idea from MQ is the following: In PMZ he writes of the definition of “logic” (higgayon):Footnote ⁴⁸

In MQ Galen writes, in Ibn Tibbon's translation:

a statement that Ibn Riḍwān comments as follows:

In other words, Ibn Riḍwān (in Ibn Tibbon's translation) agrees with Galen that this kind of investigation belongs to manṭiq, but with a proviso: it in fact belongs not to logic tout court, but to one part of it, viz. dialectics (החלק הנצחוני or נצוחי); therefore Galen's statement is true only if by manṭiq we understand dialectics, i.e., we agree to name the part after the whole (lit. genus). In PMZ, Ibn Tibbon makes the converse remark. He states that the science of logic, which is called manṭiq in Arabic, is called by the Christians by the name of one of its parts: “dialectics.” Despite the difference in the specific assertion, Ibn Tibbon's statement that a part of a discipline may be named after the entire discipline or inversely seems to follow MQ.

To this observation concerned with the contents of the two texts, two short observations on terminology can be added. First, the terms מחלוקת הנצחון/הנצוח that we just witnessed in MQ appear in PMZ and in Ibn Tibbon's translation of the Guide in identical forms.Footnote ⁵¹ This is yet another instance for the evolution of Ibn Tibbon's vocabulary. Second, in the manuscripts of MQ, the Arabic term manṭiq and its Hebrew equivalent higgayon both appear pell-mell. For instance:Footnote ⁵²

It would seem that Ibn Tibbon first wrote manṭiq, which he later replaced by higgayon, although, again, the replacement was not entirely systematic and not all manuscripts were corrected. The use of a transcribed Arabic term is characteristic of the early phase of translation. The terminological inconsistency observable in the manuscripts of MQ is an indication that it was written early on in the translator's career.

V

In sum: The two versions of the argument that MQ is a “spurious” translation of Samuel Ibn Tibbon have been shown to be unsubstantiated. Quite on the contrary, a careful investigation of the evidence sustains Ibn Tibbon's authorship of the Hebrew version of MQ as affirmed in the two extant colophons. The weightiest evidence is philological: we have noticed that the terminological evolution of Ibn Tibbon (notably as described in PMZ) entirely corresponds to what we find in MQ. Additional evidence is substantial: we have seen that in some logical matters discussed in PMZ Ibn Tibbon seems to follow MQ, with which he was thus apparently acquainted at an early date. To this must be added the analysis showing that the translation of MQ fits into Samuel Ibn Tibbon's Werdegang as a piece into the jigsaw puzzle (see Freudenthal & Fontaine, “Philosophy and medicine”). All this, I submit, establishes beyond reasonable doubt Samuel Ibn Tibbon's authorship of Melaḵah qeṭanah as his first translation. I am thus happy to follow the traditional more of השבת אבידה לבעליה (returning a lost object to its owner) and reclaim for Samuel Ibn Tibbon, the great translator and scholar, one of his mind's earliest fruits.