Version classiqueVersion mobile

The Tiberian Pronunciation Tradition of Biblical Hebrew. Volume II

Geoffrey Khan

Hidāyat al-Qāriʾ

Commentary on Hidāyat al-Qāriʾ

Texte intégral

Long Version

➢ II.L.0.0.

1The long version opens with an introduction, which presents various aspects of the background of the ensuing work. ʾAbū al-Faraj Hārūn included such authorial paratexts in his other works. His grammatical work al-Kitāb al-Kāfī contains a preface and an introduction (ed. Khan, Gallego and Olszowy-Schlanger 2003, 10- 19). The abridgement of al-Kitāb al-Kāfī known as Kitāb al-ʿUqūd fī Taṣārīf al-Lugha al-ʿIbrāniyya ‘The Book of Rules regarding the Grammatical Inflections of the Hebrew Language’, which was compiled by an anonymous contemporary of ʾAbū al-Faraj Hārūn, likewise contains an introduction (ed. Vidro 2013a, 22– 25). ʾAbū al-Faraj Hārūn’s glossary of difficult words, which is referred to as Tafsīr ʾAlfāẓ al-Miqrā ‘Explanation of the words of Scripture’ and several variant titles, contains a postface added at the end of the work, which served the same purpose as a preface and an introduction (Goldstein 2014).

  • 1 See Freimark (1967) and the discussion and references in Goldstein (2014).

2Authorial paratexts, in the form of prefaces, introductions and postfaces, are a characteristic feature of contemporary medieval Arabic literary compositions.1 The addition of such a paratext, therefore, in the works of ʾAbū al-Faraj Hārūn reflects convergence with the Arabic literary models.

3An introductory paratext is found in some earlier works written by Jewish scholars, such as the Masoretic treatise Seder ha-Simanim (Allony 1965, י־ט ) and some works of Saadya in the tenth century, such as his lexicon ha-ʾEgron (ed. Allony 1969a, 148–63) and his Bible commentaries (Stroumsa 2007).

4The beginning of the introduction to the long version of Hidāyat al-Qāriʾ is missing. If we compare it, however, with the corresponding introduction in the short version, which has been preserved in its entirety, it can be safely assumed that only a short amount of text is missing.

5The introduction can be divided into various components, many of which contain standard themes in such authorial paratexts. §II.L.0.1. discusses the reasons why the principles of biblical reading need to be studied. §II.L.0.2.—§II.L.0.6 concern the historical background of the Tiberian reading and its antiquity. §II.L.0.7.—§II.L.0.8. describe the history of the discipline of fixing rules for the correct reading. §II.L.0.9. explains the author’s motivation to write the work and summarizes its contents.

6ʾAbū al-Faraj states in §II.L.0.9. that the purpose of the work was essentially compilatory, in that it brings together in a comprehensive way specialist works and oral teachings of his predecessors, who remain anonymous. This section contains a ‘request to compose’ without specifying the name of the requester. This is a standard feature of Arabic introductions of the period and it is often no more than a fictional trope (Freimark 1967, 36–40). As is typical, this ‘request to compose’ is combined with a modesty trope in which the author acknowledges his own imperfections.

7In some cases, we know that medieval Karaite works were commissioned by specific individuals. In the postface to his Kitāb ʾAlfāẓ al-Miqrā, ʾAbū al-Faraj indicates the name of the requester, viz. ʾAbū al-Ṭayyib Samuel ibn Manṣūr (Goldstein 2014), who commissioned the work for his children. Another case is the short commentary of ʾAbū al-Faraj Furqān ibn ʾAsad (Yeshuʿa ben Yehuda), in the introduction of which it is indicated that the work was commissioned by the wealthy patron ʾAbū al-Ḥasan Dāʾūd ibn ʿImrān ibn Levi (Khan 1993; Polliack 1997, 47–48). An alternative process is attested in the manuscripts containing the grammatical commentary of the Karaite Yūsuf ibn Nūḥ, known as the Diqduq, in which there is a document indicating that the author dedicated the work as a pious foundation to the Karaite community (Khan 2000b, 153–54).

8The introduction to Hidāyat al-Qāriʾ has a particular focus on the accents rather than the consonants and vowels. This suggests that the main interest of ʾAbū al-Faraj in the work were the accents. Indeed one early source that is apparently referring to Hidāyat al-Qāriʾ calls it Kitāb al-ʾAlḥān ‘The Book of the Accents’ (see § II. Int.0.3.). One of the later European recensions of the work, furthermore, had the title Ṭaʿame ha-Miqra ‘The Accents of the Bible’ (see vol. 1, §I.0.13.1.).

It should be noted, however, that the adducing of examples of different positioning of accents as a means of demonstrating the importance of the knowledge of correct reading in §II.L.0.1. has a close counterpart in the introduction by ʾAbū al-Faraj Hārūn to his grammatical work al-Kitāb al-Kāfī, which demonstrates the importance of the knowledge of grammar. A large proportion of the introduction of al-Kitāb al-Kāfī (ed. Khan, Gallego and Olszowy-Schlanger 2003, 12-19) concerns the positioning of accents and, indeed, several of the examples are the same as those presented in the introduction to Hidāyat al-Qāriʾ. The shared examples, moreover, are used to demonstrate the same points. These include the following. The accent position can distinguish between past tense, e.g. Image (Ruth 2.6) ‘who returned’ and present tense, e.g. Image (Lev. 22.13) ‘and she returns’. The position of the accent can distinguish between lexical meaning, e.g. Image ‘they have turned back’ (Jer. 11.10) but Image ‘they captured’ (Gen. 34.29). Furthermore, the same issues of accent position with overlapping examples also occur in §I.1.1. of al-Kitāb al-Kāfī and in a passage in ʾAbū al-Faraj Hārūn’s reworking of Ibn Nūḥ’s commentary on the Pentateuch (Goldstein 2014, 367).

9It is likely, therefore, that the use by ʾAbū al-Faraj of the arguments relating to accent position in his introduction to Hidāyat al-Qāriʾ was to some extent motivated by the fact that similar argumentation was already at hand in passages in his other works. Distinction of meaning arising from accent position was, moreover, a particularly salient demonstration of how precise knowledge of the language is important for correct interpretation of Scripture.

➢ II.L.0.1.

הגל עוגרלא ... הגל יבסלא
lexical class of ‘returning’… lexical class of ‘capturing’

10In the linguistic thought of Karaite grammarians of the tenth and eleventh centuries, inflections and derivative forms of a verb were said to belong to a particular lexical class. This was expressed by the Arabic term lughah typically followed by an abstract Arabic verbal noun in genitive annexation, as is the case here (al-rujūʿ ‘returning’, al-saby ‘capturing’), or by the Hebrew term lashon followed by a Hebrew, or occasionally Aramaic, abstract noun in annexation, e.g. ןושל הכרב ‘lexical class of blessing’ (§II.L.3.2.7.), ןושל הליחד ‘lexical class of fearing’ (§II.L.3.1.). A lexical class was a class of attested linguistic forms sharing a common kernel of meaning and common letters. Such a lexical class does not include words that are related in meaning but have no letters in common. The common letters that embrace all words belonging to a lexical class are in most cases equivalent to what we would call the root letters. The medieval Karaite grammarians of the tenth and eleventh centuries, however, did not have a fully developed concept of an abstract triliteral root as the base of derivation in their linguistic theory (Khan 2000b, 78–82; 2013a; 2013b).

11The use of the term lashon with this sense of lexical class is found in earlier Hebrew Masoretic treatises, e.g. Diqduqe ha-Ṭeʿamim (ed. Dotan 1967): ןושל הכרב ‘lexical class of blessing’ (§21), ןושל הליכא ‘lexical class of eating’ (§22), ןושל הדירי ‘lexical class of descending’ (§25). The grammatical use of the term lashon ultimately has its origin in the Rabbinic tradition, where it is used broadly in the sense of linguistic form. Such broader usage is still found in the text of Hidāyat al-Qāriʾ in cases such as ןושל םיתב ‘the form ב ָ ת ִּ םי ’ (§II.L.1.9.2.), ןושל דיחי ‘singular form’ (§II.L.3.2.1.).

➢ II.L.0.2.

If one were to say ‘What do you say concerning the formation of these accents?’, the response would be that they originated by convention among the people of the language, by the help of which they fully expressed their purposes, as in the aforementioned examples and others. They established them by convention, just as they established the vowels by convention, as will be explained.

  • 2 The various passages in al-Kitāb al-Kāfī have been gathered together by Gallego (2003).

12This reflects a rationalist and anthropocentric view of the origin of language that ʾAbū al-Faraj expresses elsewhere in Hidāyat al-Qāriʾ and in his grammatical works. According to this view, language did not develop by revelation from God but rather developed among the primeval speech community of humans by convention in order to fulfil their needs of communication. The primeval speech community is referred to as ‘the people of the language’ (ʾahl al-lugha). It appears that this was intended to be the primeval community of Hebrew-speakers. Discussion of the conventional origin of language elsewhere in Hidāyat al-Qāriʾ is found in sections concerning the origin of letters and vowels, e.g. §II.L.1.1.1. §II.L.1.1.2., §II.L.1.4.1., §II.L.1.8.2., §II.L.2.2.1. In §II.L.2.2.1. there is a discussion of the role of the language of God in the development of language by conventional agreement. In his grammar al-Kitāb al-Kāfī, ʾAbū al-Faraj describes how various aspects of grammatical structure, such as verbal inflections and the expression of gender and number, would have arisen by convention within the primeval speech community to fulfil their needs of communication.2

  • 3 See this commentary of al-Fārābī (ed. Beirut 1960, 27, 50–51) and Zimmermann’s translation (1982, x (...)

13The notion of the conventional origin of language was adopted by ʾAbū al-Faraj from the rationalist views of language that were espoused by the Muslim theological movement known as the Muʿtazila. The Muʿtazila, in turn, had received this from the Aristotelian tradition of the conventional origin of meaning of sounds in language (Allen 1948; Kretzmann 1974) through the Arabic translations of Aristotelian texts at the period. Muslim philosophers whose thought was based in the Aristotelian tradition also adopted this concept of the origin of language, e.g. al-Fārābī (d. 950), who wrote a commentary on the Arabic version of Aristotle’s De Interpretatione, where Aristotle expressed his arguments on this question.3

14The leading Karaite intellectuals of the period adopted many aspects of Muʿtalizite thought (Wolfson 1979; Sklare 2017). They followed in particular the so-called Bahshamiyya Muʿtazilite school of Baṣra, which was founded by students and followers of ʾAbū Hāshim al-Jubbāʾī. In the second half of the tenth century and the eleventh century, they were influenced especially by the central figures of this school such as ʾAbū ʿAlī ibn Khallād and ʿAbd al-Jabbār al-Hamadhānī. Yūsuf ibn Ibrāhīm ha-Kohen al-Baṣīr, a prominent scholar in the Karaite Jerusalem school, who was a contemporary of ʾAbū al-Faraj Hārūn, was particularly closely engaged with the Bahshamiyya school led by ʿAbd al-Jabbār and wrote refutations of ʿAbd al-Jabbār’s opponents (Sklare 2017, 159, 163).

15The rationalist notion of the origin of language through conventional agreement among humans can be traced to ʾAbū Hāshim al-Jubbāʾī (d. 933) of the Baṣran Muʿtalizite school and gained ground among the Muʿtazilites through the works of ʿAbd al-Jabbār (d. 1026) and ʾAbū al-Ḥusayn al-Baṣrī (d. 1044) (Weiss 1974, 35). The Muʿtazalite view of the origin of language by convention was adopted by various Karaite scholars of the Jerusalem school other than ʾAbū al-Faraj Hārūn, such as Yūsuf al-Baṣīr (Vajda 1974, 61–62) and ʾAbū al-Faraj Furqān ibn ʾAsad (also known as Yeshuʿa ben Yehudah) (Zwiep 1997, 149–58).

16The key Arabic terms that are used by the Muʿtazilites for convention are ʾiṣṭilāḥ and verbal forms from the root w-ḍ-ʿ, especially muwāḍaʿa and tawāḍuʿ. The latter was the term favoured by ʿAbd al-Jabbār in his discussions of the origin of language (Peters 1976, 304–5). ʾAbū al-Faraj Hārūn uses both of these terms in his references to the conventional origin of language. The terms muwāḍaʿa and tawāḍuʿ appear to be literal translations of the Greek term συνθήκη ‘convention’ in the Aristotelian corpus, all of which have the basic meaning of ‘putting together’.

17In the passage in §II.L.0.2. the focus is on the origin of the accents, which are said to have arisen by convention among the ‘people of the language’, just as they established the vowels by convention, to achieve needs such as the distinctions of meaning described in §II.L.0.1. This would relate to stress position rather than musical cantillation and so would be a feature of natural speech. The passage then goes on to say that the arrangement (tartīb) of the accents may have been based on the practice of the Levites. This is most likely referring to the fixing of the sequence of different pitch accents in the musical cantillation. In a discussion of the origin of the vowels in §II.L.2.18., it is stated that ‘it is possible that the people of the language formed the shapes of the vowels just has they formed the shapes of the letters. It is also possible that the people of the language used to use them in conversation without their knowing any (written) form for them.’ In ʾAbū al-Faraj’s discussion of the origin of the accents in §II.L.0.2., it is not clear whether he is referring to both the written accent signs and the oral reading or only to the oral reading.

18The view of the origin of the accents and vowels by convention was abandoned by later Karaites, after the dispersal of the Karaite Jerusalem school consequent upon the capture of Jerusalem by the Crusaders in 1099. The Karaite Judah Hadassi, who was active in the middle of the twelfth century in Constantinople, for example, adopted a revelationist view of their origin, and argued that the original tablets that were given to Moses on Mount Sinai must have had the vowels and accents:

  • 4 Sefer ʾEshkol ha-Kofer (ed. Eupatoria 1836, 70a): כי בלא חמשת הקולות שהם מלכי הנקוד לא תולד המלה ול (...)

… for without the five vowels, which are [represented by] the vowel signs, a word could not be articulated nor could it be understood without the pronunciation of the vowels and accents.4

➢ II.L.0.3.

  • 5 For the full references see Gil (1996, 165–66).

19This section refers to the devastation of Jerusalem by the Romans, who are referred to figuratively as ‘wicked Edom’, i.e. Esau, the brother and enemy of Jacob. This was a trope that developed early in Rabbinic tradition (Feldman 1992, 47–48). The term here includes also the Byzantine rulers. The reference to pilgrimage to Tiberias and Gaza is evidently referring to the Byzantine period. Pilgrimage to Gaza, Tiberias and Zoar in the Byzantine period are mentioned also by other medieval Karaites.5 One such reference is by Sahl ben Maṣliaḥ ha-Kohen in the preface to his Book of Precepts:

After having left that place (i.e. Jerusalem), it remained for over five [hundred] years as rubble and dens of jackals, and the Jews could not enter. The Jews who resided in the East would come to the city of Maʿazyah (i.e. Tiberias) in order to pray, while those who lived in the West would come to the city of Gaza for that purpose. Those who dwelt in the South would go to the city of Zoar. In the days of the little horn (i.e. the Islamic empire; cf. Dan. 7.8), God opened His gates of mercy to His people, and brought them to His holy city.

20Another reference to such pilgrimage in the Byzantine period is found in the commentary of Daniel al-Qūmisī (end of the ninth century) on Daniel 11.32:

Before his arrival (i.e. of the Arabs), they (i.e. the Jews) could not enter Jerusalem and would, therefore, come from the four corners of the earth to Tiberias and to Gaza in order to catch a glimpse of the Temple.

➢ II.L.0.5.

21The argument here is that the opinion of the community as a whole sanctions the authority of the tradition of the accents. This is similar to the argument of the Karaite al-Qirqisānī that the agreement of the majority of the community (ʾijmāʿ) on the biblical reading tradition is the crucial basis of its authority (Khan 1990).

➢ II.L.0.7.

Surely you see that Muslims, whether they be two or more, cannot read with the same degree of coordination as the Jews read, since each one has his own way (of reading). One makes long a place that another makes it short. One reads melodically a place that another reads flat.

22The crucial difference between the reading of the Hebrew Bible and the reading of the Qurʾān is that in the former the contour and sequence of pitch accents are fixed by tradition whereas in the latter pitch differences are improvised by individual readers (Nelson 2001).

➢ II.L.0.9.

אדה רצתכמלא
‘this short treatise’

23The reference is clearly to the long version of Hidāyat al-Qāriʾ and not to the short version, in connection with which ʾAbū al-Faraj also uses the verb ikhtaṣara ‘to shorten’ (see the introduction to the short version §II.S.0.1.). The term ‘short treatise’ in the introduction to the long version can be interpreted as a case of formulaic modesty.

➢ II.L.1.1.

24The ideas presented in this section correspond closely to the Muʿtazilite views of language, especially those expressed by ʿAbd al-Jabbār in his various works. ʿAbd al-Jabbār’s definitions of speech include ‘the arrangement together of two or more letters’ (mā intaẓama min ḥarfayn fa-ṣāʿidan), ‘what consists of the specific arrangement of these intuitively known letters, occurring in two letters or (more) letters’ (mā ḥaṣala fīh niẓām makhṣūṣ min hādhih al-ḥurūf al-maʿqūla, ḥaṣala fī ḥarfayn ʾaw ḥurūf) and a number of other variant formulations (Peters 1976, 293–94). Likewise ʾAbū al-Faraj states that a communicable utterance must consist of a minimum of two letters. Like ʿAbd al-Jabbār, ʾAbū al-Faraj uses forms of the verb naẓama to refer to the arrangement of letters (e.g. §II.L.1.1.4., §II.L.1.1.5, §II.L.1.1.7.). Similar statements are made by the anonymous Karaite grammarian in the work Kitāb al-ʿUqūd, viz. ‘speech is constituted by articulated sounds arranged in a particular type of arrangement’ (al-kalām huwa al-ḥurūf al-muqaṭṭaʿa al-manẓūma ḍarb min al-niẓām) (ed. Vidro 2013a, 33), and the works of the Karaite Yūsuf al-Baṣīr, e.g. al-Kitāb al-Muḥtawī ‘The Comprehensive Book’: ‘articulated sounds and arranged letters that can by convention convey meaning’ (ʾaṣwāt muqaṭṭaʿa wa-ḥurūf manẓūma yaṣiḥḥ ʾan tufīd bi-lmuwāḍaʿa) (Vajda 1974, 61).

25According to ʿAbd al-Jabbār, sounds are accidents (i.e. properties) and not substances. He uses the term ʿaraḍ to refer to this concept of accident (Peters 1976, 299). This corresponds to the term khāṣṣiyya, which is used by ʾAbū al-Faraj in Hidāyat al-Qāriʾ to denote the property of a letter, i.e. its realization in sound. The term khāṣṣiyya is found in some tajwīd manuals to refer to the distinctive phonetic realization of letters or categories of letters, e.g. ḵhāṣṣiyyat al-ḥarf al-rakhwḵhāṣṣiyyat al-ḥarf alshadīd ‘the distinctive property of a “soft” letter… the distinctive property of a “strong” letter’ in the commentary on al-Dānī’s Kitāb al-Taysīr by al-Mālaqī (d. 705/1305), al-Durr al-Nathīr w-al-ʿAdhb al-Namīr ‘Scattered Pearls and Pure Sweet’ (ed. Beirut 2002, 183). Al-Dānī uses the corresponding term ṣifa ‘attribute’ in his works, e.g. al-Taḥdīd fī al-ʾItqān w-al-Tajwīd ‘The Definition of Precision and Excellent Reading’ (ed. Amman 2001, 105).

26In Hidāyat al-Qāriʾ ʾAbū al-Faraj occasionally uses the Hebrew term םעט (literally: ‘taste’) in the sense of the quality of a letter. In §II.L.1.3.3. this occurs in its Arabicized form םאעט ṭaʿām. The terms are sometimes used together, e.g. םעט ףרחלא התיצאכו ‘the “taste” of the letter and its property’ (§II.L.1.1.2.). The Arabic verb dh-w-q ‘to taste’ is used in connection with the pronunciation of consonants in §II.L.1.3.8.: ּ֩ אדאפ תקד ףרחלא יפ תקד אמנאהלחמ הנמ התיצאכ ‘If you taste a letter (by pronouncing it) in its place of articulation, you will taste its property.’ This suggests that םעט ‘taste’ is used by ʾAbū al-Faraj to refer to the perceived quality of the letter, whereas khāṣṣiyya is its intrinsic property.

27It is stated in §II.L.1.1.1. and §II.L.1.1.2. that ‘the conventional agreement was initially on the property of the letter’, i.e. the conventional agreement in the primeval speech community was on sounds of letters rather than the names of letters. As remarked, according to ʿAbd al-Jabbār sounds were accidental properties and so speech must be an accident too. The conventional agreement was, therefore, initially on speech.

➢ II.L.1.1.2.

28People differ with regard to the number of the letters that are added to the realization of the letter and its property.

29This is referring to the variant forms of the names of the letters that were current during the time of ʾAbū al-Faraj.

➢ II.L.1.1.4.

םהו ךולמלא ידלא אמסת אחנא
namely the vowels, which are called vocalization

30Two terms are used for vowels. The term ‘kings’ (mulūk) expresses its hierarchical relationship viz-à-viz shewa, which is referred to as a ‘servant’ (khādim) in some medieval sources, e.g. the anonymous Masoretic treatise CUL T-S NS 301.84. This would be analogous to the relationship of a conjunctive accent, also referred to in the treatises as a khādim, with a following disjunctive accent. The Arabic term ʾanḥāʾ is the plural of naḥw, the primary meaning of which is the grammatical inflection of a word. In Arabic such inflection is expressed by case vowels and the term has been extended to all vowels.

➢ II.L.1.1.7.

ןדאפ ףרחלא הדחו אל דאפי הנמ ינעמ ילעו אדה הגולא אמס ףרחלא ןייקודקדלא אמדאכ
Now, meaning is not expressed by a letter alone and so the Hebrew grammarians have called a ‘letter particle’ a ‘functional particle’.

31The word ḥarf is the normal term used for ‘grammatical particle’ in the standard Arabic grammatical tradition of the period. ʾAbū al-Faraj states here that the Hebrew grammarians use the term khādim, which literally means ‘serving (particle)’, or ‘subordinate (particle)’, to refer to such grammatical elements. This is the term he himself uses in his grammatical works, e.g. al-Kitāb al-Kāfī (ed. Khan, Gallego and Olszowy 2003, §I.27.1). The designation diqdūqiyyūn is generally used by ʾAbū al-Faraj to refer to the Karaite grammarians of Hebrew who preceded him. They practiced a discipline that was known as diqduq ‘attention to fine details, careful investigation (of Hebrew Scripture)’. This was associated with the Masoretic activity that produced the early Masoretic treatises. Such activity is sometimes referred to as diqduq ha-miqra ‘careful investigation of Hebrew Scripture’ in the sources, e.g. Allony (1964). Note also the anonymous Judaeo-Arabic Masoretic treatise that is preserved in the Genizah fragment CUL T-S D1.2, which states that its source is אממ ב הוני what was explained by‘ ןימלעמלא א ןילואל א דידאנצל פ י ד קודק א ארקמל the early master teachers of the careful investigation of Hebrew Scripture’.

32ʾAbū al-Faraj was sometimes at pains to distinguish his own works from the discipline of diqduq (Khan 1997; 2000b, 1–25). One of the reasons for this appears to be that ʾAbū al-Faraj believed that his grammatical works al-Kitāb al-Mushtamil and al-Kitāb al-Kāfī had a universalist approach that dealt not only with Hebrew but with general issues realting to human language, whereas the diqdūqiyyūn were specifically concerned with the Hebrew language of Scripture.

  • 6 See Khan, Gallego and Olszowy-Schlanger (2003, xiii) for a list of the extant manuscripts.

33After completing al-Kitāb al-Kāfī, ʾAbū al-Faraj wrote a work entitled Kitāb al-Madkhal ʾilā ʿIlm al-Diqdūq fī Ṭuruq al-Lugha al-ʿIbrāniyya ‘Book of Introduction into the Discipline of Careful Investigation of the Ways of the Hebrew Language’, which he states was intended as a guidebook to the terminology of the diqdūqiyyūn. This work is still unpublished.6 In the work ʾAbū al-Faraj states:

  • 7 II Firkovitch, Evr.-Arab. I 4601, fol. 112 (an autograph of ʾAbū al-Faraj الكالم المستعمل ثلثة اقسا (...)

‘Speech that is used (for communication) consists of three components: noun (ism), verb (fiʿl) and particle (ḥarf, literally: letter), which the people of the discipline of diqduq call ‘serving particle’ (khādiman).7

34The manuscript of our passage in Hidāyat al-Qāriʾ has the reading אמס ןייקודקדלא םדאכלא אפרח ‘the diqduq scholars called a “serving particle” (khādim) a “letter” (ḥarf).’ Given the statement in Kitāb al-Madkhal and the following context of the passage in the Hidāya, it is clear that this reading is a scribal error for ּ֩ אמס . ןייקודקדלא ףרחלא אמדאכ

35In the published texts of the early Karaite grammarians no specific technical term is attested for ‘grammatical particle’ (Khan 2000b, 74; 2000a).

דקפ םתכנא יפ כל ל פט̇ה ףרח ףלא והו ןמ ףורח ןיללא
In each of these words a letter ʾalef is hidden, which is one of the soft letters.

36This reflects the theory that long vowels were the result of ‘soft letters’ (ḥurūf al-līn), i.e. vowel letters. Such a theory was borrowed from the Arabic grammatical tradition and developed more systematically by the Hebrew grammarian Ḥayyūj, who was active in Spain in the early eleventh century (Basal 2013). Unlike in Arabic, these vowel letters were sometimes elided in the orthography; cf. the discussion in Hidāyat al-Qāriʾ in §II.L.1.8.

➢ II.L.1.2.

Image ילא לוצאלא הסמכ ףורח יהו יתלא אהנומסי תוטושפ םהו Image
To the basic letters are added five letters, which are called “straight”, namely ץףןםך .

37The Hebrew term פ ָ טוש ‘straight’ for the long final letters has its origin in Rabbinic literature; cf. Jastrow (1903, 1138).

Image ילא לוצאלא Image התס ףורח יהו Image ןיינארבטלא ודיזי הדה ילע התסלא ףרחא ףרח שיר אהנולעגיו Image
To the basic letters are added six letters, namely תפכדגב . The Tiberians add to these six letters the letter resh, making it (i.e. the group of non-basic letters) תרפכדגב .

38The reference here is to the fricative variants of the תפכדגב consonants and the alveolar trill variant of the resh (vol. 1, §I.1.20), all of which were regarded by ʾAbū al-Faraj as secondary variants of the consonants in question. Sefer Yeṣira refers to ּ֩ עבש תרפכדגב תולופכ ‘the seven double letters תרפכדגב ’ (Hayman 2004, 24). According to Morag (1960), this is referring to the Babylonian pronunciation tradition, in which there was a variation in the realization of resh that was different in nature from that of the Tiberian tradition. The fricative variants of the תפכדגב consonants were regarded as secondary also in a Masoretic treatise published by Allony and Yeivin (1985, 97).

יאז ךורכמ

39For the zāy makrūkh see vol. 1, §I.1.7.

➢ II.L.1.3.2.

40The Hebrew term תיב העילבה ‘place of swallowing’ is found in Hebrew Masoretic treatises, e.g. §5 and §6 in Baer and Strack’s (1879) corpus. It is attested already in Rabbinic literature (Jastrow 1903, 173).

41For the four cases of dagesh in ʾalef in the standard Tiberian tradition see vol. 1, §I.1.1.

➢ II.L.1.3.3.

42For the meanings of the terms םעט and khāṣṣiyya see the comments to §II.L.1.1. above.

➢ II.L.1.3.4.



44Cf. the use of the Arabic term ẓāhir ‘appearing’ to designate he with mappiq in the Masoretic treatise published by Allony and Yeivin (1985, 97).

➢ II.L.1.4.1.

יה תאראבע Image אהילע אהבאברא ומהפיל Image גאתחיו ןא םלעי לבק אדה םהנא Image ילע ףורח הצוצכמ יפ הפלתכמ לאחמ
It [i.e. language] consists of expressions that its original speakers established by convention among themselves to make their intentions understood to one another. It needs to be known that before this they established by convention specific letters in various places of articulation.

45The expression ʾarbābuhā ‘its [i.e. language’s] masters’ is synonymous with ʾahl al-lugha ‘the people of the language’.

46The establishment by convention of specific letters (ḥurūf) is referring to the phonetic realization of the letters in the form of the production of sound, as was discussed in §II.L.1.1.

➢ II.L.1.4.2.

לעפ יפ ספנלא
intransitive verb

47This term, which literally means, ‘action on oneself’, is taken from the early Karaite grammatical tradition; cf. the Diqduq of Ibn Nūḥ (Khan 2000b, 108–11). In the standard Baṣran Arabic grammatical tradition an intransitive verb was called fiʿl ghayr mutaʿaddin, which was the term used by ʾAbū al-Faraj in his grammar books (Khan, Gallego, and Olszowy-Schlanger 2003, xliii).

➢ II.L.1.4.5.

םסקלאו עבארלא אמ לדבני ןמ לאחמ הריאגתמ והו םעי אירטמגלאו ףיחצתלא ןאל לאדלא שירלאו ןמ ןילחמ ןיריג אירטמגלאו אבלא ןמ ןישלאו ןילחמ ןירואגתמ
The fourth category is what interchanges from different places of articulation, which includes scribal error and ciphers, for dalet and resh (which are sometimes confused by scribal error) are from two different places of articulation and the cipher of letters (consisting of the interchange of) bet and shin is from two different places of articulation.

48Examples of scribal errors resulting in the interchange of dalet and resh are given in §II.L.1.5.4. They are discussed by ʾAbū al-Faraj in al-Kitāb al-Kāfī, in which he states ‘It is a type of letter interchange arising from a mistake in writing on account of the similarity of two letters’ (ed. Khan, Gallego and Olszowy-Schlanger 2003, §I.28.3.).

49The ciphers referred to are in the encryption code known as ת׳׳א ש׳׳ב , whereby a letter is exchanged for its counterpart in the opposite place in the alphabet, e.g. ʾalef, the first letter in the alphabet, for tav, the last letter, bet, the second letter, for shin, the penultimate, and so forth (Campanini 2013, 638).

➢ II.L.1.4.7.

Image ‘the third is Absalom’ (1 Chron. 3.2).

50This is cited by ʾAbū al-Faraj as an example of a lamed without meaning also in al-Kitāb al-Kāfī (ed. Khan, Gallego and Olszowy-Schlanger 2003, §I.25.45.).

➢ II.L.1.4.8.

לאקי הנא רצתכמ דוי יפ די הרידקתו ידיב
It is said that a yod has been contracted in Image, and its virtual form is Image.

51ʾAbū al-Faraj uses here the notion of taqdīr (‘virtual form’, ‘imagined form’), which he adopted from contemporary Arabic grammatical thought (Levin 1997). This concept is used to explain the existence of various irregularities in grammatical structure. Underlying the actually attested structure of the biblical text, there was a virtual or imagined (muqaddar) structure, which existed in the mind of the author. This imagined structure always conformed to grammatical rules and principles. When the imagined structure deviated from the actually attested structure, it was the imagined structure that constituted the basis of interpretation and translation (Khan 2001, 142–44).

➢ II.L.1.5.

52Some cases of letter interchange that are presented in this section are mentioned by ʾAbū al-Faraj in al-Kitāb al-Kāfī, e.g. ed. Khan, Gallego and Olszowy Schlanger (2013, §I.26.28, §I.27.34, §I.28.3).

➢ II.L.1.5.18.

The intention is that Image would be in place of Image ‘treated violently (and therefore bloody)’.

➢ II.L.1.5.21.

תרכמלאו ןוכי Image
The disjoined form would be ש ָ ת ֻ ם .

ʾAbū al-Faraj uses the Hebrew terms Image ‘disjoined’ and Image ‘conjoined’ alongside the corresponding Arabic terms qaṭʿ ‘disjoined state’ (§II.L.1.4.8.) and muḍāf ‘conjoined’. For Image see §II.L.2.17. These terms are found in Masoretic treatises, e.g. §37 in the corpus of Baer and Strack (1879). In Hidāyat al-Qāriʾ the terms are generally used in the sense of ‘absolute’ and ‘construct’ forms of a noun respectively. In Karaite grammatical literature the terms are used also to denote pausal and context forms (Khan 2007). The term qaṭʿ is used in the sense of pause in Hidāyat al-Qāriʾ in §II.L.2.12.14.

בגיו ןא ןוכי הרמא ש ְ ת ֹ םו וא ש ְ ת ַ ם
Its imperative would have to be ש ְ ת ֹ םו or ש ְ ת ַ ם .

53For the notion of the imperative as the base of morphological derivation see the comments on §II.L.1.10. below.

➢ II.L.1.6.

54The letters היוא are intended here to represent long vowels. This reflects the theory that long vowels were the result of ‘soft letters’ (ḥurūf al-līn), i.e. vowel letters; cf. the comments on §II.L.1.1.7. above and the discussion in Hidāyat al-Qāriʾ in §II.L.1.8.

55Some of the material in this section can be found in other Masoretic treatises, such as §29 in the corpus of Baer and Strack (1879), two of the Judaeo-Arabic treatises published by Allony and Yeivin (1985, 99–101, 112), and various fragments of unpublished texts, e.g. CUL T-S D1.2, Mosseri I, 71.1; Oxford Bod. Heb. d 33, fol. 16.

➢ II.L.1.7.5.

Image ןחל הלבק ירגיל ירגמ Image
so (such cases must be considered) to have a virtual (conjunctive) accent before them in order to conform to (cases such as) Image

56Here ʾAbū al-Faraj uses the notion of taqdīr ‘virtual form’, ‘imagined form’ (see the comment on §II.L.1.4.8. above) to explain an apparent irregularity. The idea is that the conjunctive accent existed in the virtual underlying structure, and so this explained the operation of the ʾathe me-raḥiq.

➢ II.L.1.7.7.

לכ ואו יפ רכא המלכ גרכי ילע יאר ןיימאשלא אבב יפרמ
Every vav at the end of a word is pronounced according to the Palestinians as a bet rafe.

57See vol. 1, §I.1.6. for the pronunciation of the Tiberian vav.

➢ II.L.1.7.9.

58For a possible explanation of the phenomenon described in this section, see vol. 1, §I.3.1.10.

➢ II.L.1.7.11.

59For this feature of the reading of Ben Naftali see Kitāb al-Khilaf (ed. Lipschütz, 1965, 19).

➢ II.L.1.8.2.

60Yaḥyā ibn Dāʾūd the Maghribī

61This is a reference to the Spanish grammarian Yaḥyā ibn Dāʾūd Ḥayyūj (d. c. 1000).

אלו הילכנ ילא ןא וגי להא הגללא םלענפ הדיאפ אמ וחלטצא הילע
and we should not abandon it until the people of the language come and we know the function of what they adopted as their convention.

62Since the ‘people of the language’ are the primeval speech community, this would seem to be tantamount to saying that we should never deviate from the customary reading.

➢ II.L.1.9.2.

63For the various degrees of ‘heaviness’ of tav see vol. 1, §I.

➢ II.L.1.9.3.

64For the different realizations of resh, see vol. 1, §I.1.20.

➢ II.L.1.10.

םלעאו ןא להא הגללא ולעג ףיראצת הגללא ילע העברא םאסקא ןמ דחאו ףרח ילצא לתמ Image הוחנו ןמו ןיפרח לתמ נב ֵ ה הוחנו ןמו Image לתמ Image ןמ הוחנו ד ̇ לתמ לברכ הוחנו
Take note that the people of the language made the conjugations of the language in four categories: from one root letter, such as ה ַ כ ֵ ה ‘hit’ and the like, from two letters, such as Image ‘build’ and the like, from three letters, such as Imageand the like and from four letters, such as Image‘wrap’, and the like.

65According to the medieval Karaite grammarians, the base of the process of the derivation of a word was not an abstract root but rather a concrete structural form consisting of both consonants and vowels. Such a theory of derivational morphology was developed in the early Karaite tradition of grammatical thought in the tenth century, the main extant source for which is the Diqduq of Ibn Nūḥ. According to Ibn Nūḥ the imperative form is the derivational base of most verbal forms. The imperative, therefore, was used as the citation form of verbs. This practice of imperative citation forms was continued by ʾAbū al-Faraj in his grammatical works and Hidāyat al-Qāriʾ, in which he uses them as lemmata to identify and classify verbal conjugations. This was despite the fact that according to his theory of grammar, adopted from the mainstream Baṣran school of Arabic grammar, the infinitive was the base of morphological derivation of verbs; cf. al-Kitāb al-Kāfī (ed. Khan, Gallego, and Olszowy-Schlanger 2003, §II.16.12-15) and Khan (1997).

Although the Karaite grammarians did not work with a notion of an abstract root that functioned as the source of derivation, they did have a concept of an abstract underlying morphological level, which they referred to as the ‘substance’ (Arabic jawhar) of a word (Khan 2000b, 74–78). Unlike the morphological base, the substance is not an actual linguistic form that can be pronounced. It consists of a series of letters that are regarded as the core of the word. This is what is referred to in this passage in Hidāyat al-Qāriʾ as ‘root letter’ (ḥarf ʾaṣlī). Various letters that are present in the morphological base of a word do not belong to the substance and so are not root letters. A final he in a verb is not considered to belong to the substance. The substance of imperative bases such as Image, therefore, are said to consist of two root letters. A prefix such as the he prefix of the hifʿil that does not occur in all inflected forms is not considered to be a root letter. The substance of the verb Image‘hit’ was considered to consist of only one root letter. For further details see Khan (2013a; 2013b).

ףרחלאו דק בלקתת יפ Image ריאגתתו אהינאעמ
A letter may also change position in a word with the result that its meaning changes.

66ʾAbū al-Faraj offers here only one set of examples of such changes of position. In his grammar book al-Kitāb al-Mushtamil (see vol. 1, §I.0.13.4), however, he devotes a long section to this question and presents an inventory of sets of lexical forms that share letters in different sequences (Bacher 1895, 247–49). This is known as an anagrammatical lexical arrangement and is attested in other medieval Jewish sources, in particular Kitāb al-Ḥāwī ‘The Book of Collection’ of Hai Gaon, which is a complete anagrammatical dictionary of Hebrew (Allony 1969b, 1972).

➢ II.L.1.11.

The term ‘enhancement’ (tafkhīm) is used by ʾAbū al-Faraj to refer to the extension of the basic form of a word or affix by the addition of letters without, in principle, bringing about a change in meaning. The list of examples of added letters includes some additions that we would normally interpret as functional affixes, such as directive he, e.g. Image ‘Babylon’—Image ‘to Babylon’ (2 Kings 20.17, etc.), or cohortative he, e.g. בוש ‘return’—Imageּ֩ ‘return’ (Num. 10.36, etc.). ʾAbū al-Faraj uses the term םיכפתלאאה ‘he of enhancement’ to refer to cohortative he in his grammar book al-Kitāb al-Kāfī (ed. Khan, Gallego and Olszowy-Schlanger 2003, §I.27.126), but does not examine there the phenomenon of enhancement systematically. Saadya Gaon, on the other hand, treats the subject extensively in his grammar Kitāb Faṣīḥ Lughat al-ʿIbrāniyyīn ‘The Book of the Eloquence of the Language of the Hebrews’ (Dotan 1997, 142–45). When, however, ʾAbū al-Faraj mentions in Hidāyat al-Qāriʾ ‘somebody who has discussed enhancement’ (§II.L.1.11.5., §II.L.1.12.1.), he does not seem to be referring to Saadya, since the cited examples do not correspond to what is extant in Saadya’s work.

➢ II.L.1.11.5.

רמאלא Image
the imperative forms are Image and Image

67The imperative form was regarded by the early Karaite grammarians as the morphological stem on the basis of which verbal inflections were formed. In his grammar al-Kitāb al-Kāfī, ʾAbū al-Faraj adopts from the mainstream Baṣran tradition of Arabic grammatical thought the notion that the infinitive was the base of derivation (Khan 1997). This Baṣran Arabic tradition was merged with the early Karaite tradition in the grammatical works of ʾAbū al-Faraj, which sometimes results in apparent inconsistency. He, for example, continued to use the imperative as the citation form of verbs (Khan 2013a). In this passage of Hidāyat al-Qāriʾ he appears to be working with the early Karaite notion that the imperative was indeed the stem of derivation.

➢ II.L.1.12.3.

68For the term diqdūqiyyūndiqduq scholars’ see the comments above on §II.L.1.1.7.

69The notion of ‘auxiliary letter’ (ḥarf mustaʿmal) is explained by the anonymous Karaite author of Kitāb al-ʿUqūd (ed. Vidro 2013a, 51):

  • 8 ואלמסתעמל הו (...)

An auxiliary letter is established in the entire conjugation. If it is removed, what remains are the letters of the lexical class standing without a clear meaning Examples are the lamed in the lexical class of ‘taking’, or the nun in the lexical class of ‘going away’ and in the lexical class of ‘approaching’, as in Image etc. Here there is a clear difference between the removal of the auxiliary lamed and that of ḥet or qof. This holds for each auxiliary letter. The difference between its removal and the removal of a root letter is obvious and evident.8

70I.e. the auxiliary letter is an integral component of the conjugation but elided in some inflections, as is the case with nun and lamed in the cited verbs. This would differ from letters of enhancement, which are added to the basic letters of a conjugation, and so are not integral to it.

➢ II.L.1.13.

71The ‘contraction’ (ʾikhtiṣār) is the converse of ‘expansion’, in that it involves the elision of a letter from the basic form of a word or affix without changing meaning. The basic form is often referred to in this section as the ‘virtual form’ (taqdīr), i.e. the expected regular form (see the comments on §II.L.1.4.8. above). ʾAbū al-Faraj uses the concept of contraction of letters in al-Kitāb al-Kāfī and, likewise, refers to the uncontracted form as the taqdīr of the word, e.g. ed. Khan, Gallego and Olszowy-Schlanger (2003, §I.23.7.). This subject is treated systematically by Saadya in his grammar Kitāb Faṣīḥ Lughat al-ʿIbrāniyyīn (Dotan 1997, 140–42).

➢ II.L.2.2.1.

72In this passage relating to the origin of vowels, ʾAbū al-Faraj offers a view of the origin of language that differs from the view that it has a purely human origin through convention among the primeval speech community known as ‘the people of the language’ (ʾahl al-lugha), which he presents elsewhere in Hidāyat al-Qāriʾ and his grammatical works. In the passage he acknowledges that vowels were an integral part of the original conventional establishment of language (‘in the original establishment [of language] by convention they [i.e. the vowels] were indispensable’). He then, however, goes on to associate the origin of language with Adam, according to the biblical account of the origin of language:


‘and He (the Lord) brought them (the animals and birds) to the man to see what he would call them; and whatever the man called every living creature, that was its name. The man gave names to all cattle, and to the birds of the air, and to every beast of the field’ (Gen. 2.19-20)

73ʾAbū al-Faraj suggests various ways of interpreting Adam’s activity. He and the angels may have established language by convention. This would, in effect, be extending the notion of the primeval speech community (ʾahl al-lugha) to a community of Adam and the angels rather than original human speakers alone. Another possibility, says ʾAbū al-Faraj, is that the angels taught Adam language. The angels, moreover, would have been taught language by God. The development of language by convention between God and the angels was not possible since such a process requires limbs with which to point to objects and ʾAbū al-Faraj, like his Muʿtazilite contemporaries, rejects anthropomorphisms attributed to God such as human-like limbs.

The teaching of language to Adam by angels, who in turn learnt it from God, is a revelationist view of the origin of language. This was adopted by some Muslim theologians and grammarians of the period (Weiss 1974; Loucel 1963). The process of revelatory teaching of language was known in the Islamic tradition as tawqīf ‘making known’ or waḥy ‘inspiring’ and was based on the Qurʾanic verse: Image ‘and He (God) taught Adam all the names (of things)’ (2.31).

  • 9 Maimonides followed an Aristotelian view and regarded Adam’s act of naming as indicating by itself (...)

74The fact that ʾAbū al-Faraj presents different versions of the origin of language, conventionalist and revelationist, may have arisen as an attempt to find a synthesis between rationalist Muʿtazilite thought and Jewish biblical tradition.9 Saadya Gaon, likewise, strove to achieve such a synthesis. Unlike ʾAbū al-Faraj, however, Saadya did not attempt this by offering alternative views, but rather presented a hybrid view. According to Saadya a single primeval human, referred to as ‘the establisher of the language’ (wāḍiʿ al-lugha), fixed arbitrary names for entities and this fixing was subsequently accepted by consensus (ʾiṣṭilāḥ) among people and transmitted to later generations. The ‘establisher of the language’ that Saadya had in mind was presumably Adam, but he left the name unspecified, possibly to make his theory more universalist and applicable to all human language. According to Saadya, God did not create language but gave Adam the faculty of language (Dotan 1995; 1997, 96–104).

75It is also possible that ʾAbū al-Faraj was influenced by some streams of Muslim thought in the eleventh century that attempted a synthesis of the revelationist and conventionalist views (associated with the scholar ʾAbū ʾIsḥāq al-ʾAsfarāʾīnī, d. 1027) or were non-committal and held that neither view could be argued conclusively (associated with the scholar ʾAbū al-Bakr al-Bāqillānī, d. 1013) (Weiss 1974).

➢ II.L.2.3.2.

ברעלל יפ םהתגל ןמ בארעאלא התלת יהו עפרלא והו Image ןמ בצנלאו קופ והו החתפלא ןמ קופ Image יהו הרסכלא ןמ לפסא
The Arabs have three inflectional vowels in their language. These are ‘raising’ (rafʿ), i.e. the vowel ḍamma, which is written above; ‘holding level’ (naṣb), i.e. the fatḥa vowel, which is written above; and ‘lowering’ (khafḍ), i.e. the vowel kasra, which is written below.

76The terms rafʿ ‘raising’, naṣb ‘holding level’ and khafḍ ‘lowering’ derive from a theory of the production of vowels originating in Arabic grammatical thought that involves both the position of buccal organs and the direction of the dynamic flow of air. It was adopted also by Saadya (Dotan 1997, 113–26) and is found in a Hebrew Masoretic treatise published by Baer and Strack (1879, §36), see Eldar (1983) and Posegay (2020). In this section of Hidāyat al-Qāriʾ ʾAbū al-Faraj is referring to the vowels of Arabic case inflections, which mark syntactic relations. In §II.L.2.4.– §II.L.2.8. he transfers this terminology to categorize Hebrew vowels in morphological patterns. The seven Hebrew vowels are assigned to the three Arabic categories as follows: rafʿ (ḥolem, shureq), naṣb (pataḥ, segol, qameṣ), khafḍ (ṣere, ḥireq). A similar practice of applying this three-way classification of vowels to Hebrew morphology is found in the Masoretic treatise §36 of the corpus of Baer and Strack (1879).

➢ II.L.2.4.

בצנלאו לכדי התחת ג ̇ תאמגנ החתפלא ירבכלא יהו Image החתפלאו יהו יטסולא Image החתפלאו ירגצלא יהו Image
Three vowels correspond to (the Arabic inflectional vowel) ‘holding level’ (naṣb), namely ‘big fatḥa’, i.e.Image, ‘medium fatḥa’, i.e.Image and ‘small fatḥa’, i.e.Image.

77In early Masoretic terminology a basic distinction was made between pataḥ (open vowel) and qameṣ (vowel with lip-tightening). The vowel segol was associated with pataḥ and referred to as pataḥ qaṭan ‘small pataḥ’. The vowel ṣere was associated with qameṣ and referred to as qameṣ qaṭan ‘small qameṣ’, see vol. 1, §I.2.1.5. Here in Hidāyat al-Qāriʾ a different grouping of the vowels is made based on the perceived proximity of the Hebrew vowels to each of the three Arabic vowels. In the early Masoretic terms pataḥ qaṭan and qameṣ qaṭan, the attribute ‘small’ relates to the notion that these vowels were in some way more attenuated and less open than prototypical pataḥ and qameṣ. This notion can be traced back to Syriac grammatical sources where the Syriac term qaṭṭīn ‘narrow’ is used to describe the higher front vowels (Posegay 2020). The terms ‘big fatḥa’, ‘medium fatḥa’ and ‘small fatḥa’ seem to relate to varying degrees of lip-spreading. The vowel pataḥ was pronounced with the maximal degree of lip-spreading and qameṣ with the lowest degree, with segol exhibiting an intermediate lip position. The same attributes are used in §II.L.2.7. and applied to the term naṣb, viz. ‘big naṣb’ for pataḥ and ‘small naṣb’ for qameṣ.

➢ II.L.2.5.

ילא ריג ךלד אממ דק למעתסא יפ רמאלא רבעלאו לתמ Image אמהנזוו ף למעתסאו יפ רמאלא ןמ ףירצת Image וחנ ס ֹ ב ח ֹ ן למעתסאו רבעלא יפ ןמ ףירצת וריש וחנ Image
To these should be added forms that are used for the imperative and the past such as Image ‘be burnt’, Image ‘was burnt’, and other examples of this pattern; those that are used for the imperative in the conjugaton Image such as ס ֹ ב ‘turn’, Image ‘be gracious’; those that are used for the past in the conjugation Image, such as Image ‘be known’, Image ‘be saved’.

78The Karaite grammarians regarded the imperative form to be the morphological base of most verbal forms (see comments on §II.L.1.10. above). This morphological base enshrined the core structural features of the inflected forms derived from it. Although the imperative base is a concrete linguistic form, it is not necessarily a form that is attested in the corpus of Biblical Hebrew. The Karaite grammarians held that it is crucial that there is a close structural relationship between the base and the inflected form. If there is no attested imperative that is deemed sufficiently close in structure to the inflected form, a common way of resolving this problem is to propose a hypothetical imperative base.

The imperative base always belongs to the same pattern (binyan) as the verbal form derived from it. The notion of pattern (binyan) as a unitary abstract category, however, is not found in the morphological theory of the Karaite grammarians. Imperative bases were classified together in groups not on account of sharing of abstract stems but on account of their sharing of actual structural patterns. The imperative form Image ‘write’, for instance, was said to have the same pattern as imperatives such asImage ‘guard’ and Image ‘harvest’. The imperative בוש ‘return’ was said to have the same pattern as םוק ‘arise’. The forms Image andImage, however, were not classified together on a more abstract structural level, as became the custom in Hebrew grammar following the teachings of Ḥayyūj.

As remarked, the imperative base had to have a maximally close structural resemblance to the form derived from it. In order to achieve this in some cases the imperative form that is posited as the base of an inflected verb does not actually occur in the language, e.g. Image for Image ‘we have given’ (1 Chron. 29.14), Image for Image ‘it walks’ (Psa. 73.9) and Image for Image ‘it has sought’ (Ecc. 7.28). The motivation for positing an imperative base such as Image is to construe the derivation of the form Image, without the dagesh in the qof, as fully regular and not an anomalous inflection of the imperative base Image, which contains dagesh. Such hypothetical imperatives were not used for the writing of creative literature. Indeed, no Karaite author has been found who used these postulated forms in a creative Hebrew text. This reflects the fact that the proposed imperative base of a verb was intended as a purely structural source of derivation. Further evidence for this is that even passive forms expressed by morphological patterns that we now call the puʿal and hufʿal/hofʿal are regarded by the Karaite grammarians as derived from imperative bases, which cannot be naturally functional in the language, e.g. the past form Image‘I was made’ (Psa. 139.15) is said to be derived from the imperative form Image. Here in §II.L.2.5. of Hidayat al-Qāriʾ the past form Image ‘was burnt’ (inferred from Image ‘and behond it was burnt’ Lev. 10.16) is cited alongside the imperative Image ‘be burnt’, which would have been considered to exist hypothetically as its morphological base. Passive imperative bases of puʿal and hufʿal/hofʿal patterns have ṣere in their second syllable by analogy with active piʿel and hifʿil imperatives. For more details see Khan (2000b, 41–45; 2013a; 2013b).

ʾAbū al-Faraj uses mnemonic symbols to arrange the imperative bases of verbal conjugations into classes. Such a classification is presented systematically in §I.22 of al-Kitāb al-Kāfī (ed. Khan, Gallego and Olszowy-Schlanger 2003). These symbols, known in Arabic as ʿalāma or ribāṭ and in Hebrew as ןמיס , consisted of bisyllabic Hebrew words in which the first vowel corresponds to the vowel common to all imperatives and the last vowel to the vowel common to all past forms in the class. The vowels in question may be those of the first syllable of the imperative and past forms or those of the last syllable of these forms (Vidro 2013b, 652). The conjugations of the imperatives Imageand Image, for example, were assigned to a class designated by the mnemonic symbol Image, which was based on the first vowel of the imperatives (Image, Image) and the past forms (Image, Image). In a similar way the symbol ש ִּ י ָ הר included conjugations with imperative bases with the patterns Imageand Image, with ḥireq in their initial syllable and qameṣ in the initial syllable of their past forms, viz. Image and Image. The symbol מ ֵ ס ַ ב included conjugations in which the imperative had a ṣere in the final syllable and the past had a pataḥ in the final syllable, e.g. Image (imperative)— Image (past).

79This system of classification originated in the early Karaite tradition and was developed by ʾAbū al-Faraj. It was subsequently elaborated still further in some other Karaite works, such as the anonymous treatises Kitāb al-ʿUqūd (Vidro 2011; 2013a) and Meʾor ʿAyin (Zislin 1990). The classification of verbal conjugations and the systematization of rules for their derivation that are found in these works are likely to have had a pedagogical motivation (Vidro 2011, 165–79).

Just as hypothetical imperative bases were posited for certain past forms, e.g. Image for Image ‘we have given’, so hypothetical past forms were in some cases posited for attested imperative forms. This applies, for example, to the first in the class of conjugations represented by the symbol Image, which is mentioned in §II.L.2.5. of Hidāyat al-Qāriʾ—in the manuscript used in the edition for this passage the original pataḥ has been replaced by segol due to interference from Arabic phonology on the scribe, viz. Image (cf. vol. 1, §I.4.3.3.). In conjugations represented by the symbol Image the first vowel of the imperative is ḥolem and the first vowel of the past form is pataḥ. The imperatives of the class Image that are given in §II.L.2.5. are ס ֹ ב ‘turn’ and ח ֹ ן ‘be gracious’. The attested 3ms past forms of ס ֹ ב and ח ֹ ן are Image and Image It is assumed, however, that these imperatives have hypothetical 3ms past forms Image and Image, in which the first vowel is pataḥ, in conformity with the symbol Image This is explicitly stated in al-Kitāb al-Kafī (ed. Khan, Gallego and Olszowy-Schlanger 2003, §I.22.19) and Kitāb al-ʿUqūd (ed. Vidro 2013a, 167). The forms ס ַ ב and Imagewere inferred from attested inflected forms such as Image ‘I have turned’ (1 Sam. 22.22), Image ‘he has treated me graciously’ (Gen. 33.11) and ֙Image ‘and I shall be gracious’ (Exod. 22.19).

The other symbol mentioned in §II.L.2.5. is Image which is exemplified by the past forms נ ֹ ו ַ עד and Image, with ḥolem in their first syllable. The imperative forms of these are Image and Image with ḥireq in their first syllable. Full details and vocalization are given in al-Kitāb al-Kafī (ed. Khan, Gallego and Olszowy-Schlanger 2003, §I.22.25) and Kitāb al-ʿUqūd (ed. Vidro 2013a, 193).

דקפ ראצ אדה םסקלא דחאולא למעתסי יפ ןילעאפלא אמכ ברעלא ולמעתסא עפר [ םה פ ] י ןילעאפלא ולמעתסאו ןיינארבעלא יפ ריג אעסות ןילעאפלא אמכ ולמעתסא ברעלא םהעפר יפ ג [ רי ] ןילעאפלא אעסות ןמ אדתבאלא וחנ םאהפתסאלאו ריגו ךלד
This category (of rafʿ) is used for agents, just as the Arabs use rafʿ for agents. The Hebrews use it for forms other than those designating agents by extension, just as the Arabs use their rafʿ for forms other than those designating agents by extension, such as initial positioning (of nominal subjects), interrogative constructions, and so forth.

In this passage ʾAbū al-Faraj identifies a correspondence between the use of the Arabic and Hebrew rafʿ in their marking of agents. In the case of Arabic, the rafʿ vowel -u is a case-vowel marking the agent of the clause and so plays a syntactic role. In Hebrew, by contrast, the use of the rafʿ vowel ḥolem plays a morphological role as an element in the morphological pattern of Hebrew agents, e.g. Image ‘slayer’. According to ʾAbū al-Faraj the use of ḥolem in Hebrew in other morphological patterns is by a process of extension that is analogous to the extension of the rafʿ case-vowel - u in Arabic to items that are not agents, including the subject of nominal (i.e. verbless) clauses and interrogative constructions. The latter is most likely referring to nominal clauses with initial interrogative particles such as hal al-ʾamīru fī al-bayti? ‘Is the commander in the house?’, in which the item with the -u case vowel is not the initial item (mubtadaʾ).

➢ II.L.2.6.

The symbol עוש ָ ל includes conjugations in which the first vowel of the imperative is shureq and the first vowel of the past form is qameṣ, e.g. ImageImage םוק —Image; cf. al-Kitāb al-Kafī (ed. Khan, Gallego and Olszowy-Schlanger 2003, §I.22.18) and Kitāb al-ʿUqūd (ed. Vidro 2013a, 161–65).

➢ II.L.2.7.1.

לועפמלא קלטמלא והו רדצמלא וחנ Image לאתמא ךלד
It is used in the ‘absolute patient’, i.e the infinitive, as in Image ‘guarding’, Image ‘remembering’, Image ‘building’, Image ‘doing’, and the like.

80The Arabic ‘absolute patient’ (mafʿūl muṭlaq) is a verbal noun in the accusative case in constructions such as ḍaraba ḍarban ‘he struck a striking’, sāra sayran ‘he journeyed a journey’. This has the role of an ‘inner object’, i.e. a syntactic object that does not express the role of the undergoer of the action expressed by the verb but rather the action itself. In al-Kitāb al-Kāfī (ed. Khan, Gallego, and Olszowy-Schlanger 2003, §I.21.2.) the Arabic mafʿūl muṭlaq is equated by ʾAbū al-Faraj with the Hebrew infinitive. He cites examples there of both the infinitive absolute and also the infinitive construct, which he did not distinguish as separate forms as we do today. In this passage in Hidāyat al-Qāriʾ the Hebrew examples are infinitive absolutes.

➢ II.L.2.7.2.

The symbol Image represents conjugations that have imperatives with pataḥ in their first syllable and past forms with ḥireq in their first syllable. This clearly applies to the first two examples cited, viz. Image It does not apply, however, to the other two imperatives that are cited, viz.Image and Image, which have segol in the first syllable of their past forms, viz.Image and Image For this reason the author of Kitāb al-ʿUqūd classified the last two verbal conjugations under the symbol Image (ed. Vidro 2013a, 199). In al-Kitāb al-Kāfī ʾAbū al-Faraj does not have Image in his inventory of symbols. In §I.22.10. of al-Kitāb al-Kāfī he includes Image (imperative)—Image (past) in the symbol Image but states that the past form ‘does not have a ḥireq on account of the resh.’ He does not mention the imperatives Image and Image but presumably he would have included them in Image, as here in Hidāyat al-Qāriʾ, in the same way as Image and explain the segol as a change conditioned by the initial guttural.

רדאצמלא הראעתסמלא וחנ Image םולשל
‘transposed infinitives’, such as Image ‘speak’, Image ‘ (we) look for peace’ (Jer. 8.15, etc.)

81‘Transposed infinitives’ (maṣādir mustaʿāra) are infinitives whose form originally belongs to a different category, in this case that of an imperative, but has been extended to the function of an infinitive. This type of infinitive is discussed in Kitāb al-ʿUqūd (ed. Vidro 2011, 303).

➢ II.L.2.8.2.

The inventory that is given here of past forms in the class of conjugations represented by the symbol Image (i.e. conjugations with imperatives with pataḥ in their first syllable and with past forms with ḥireq in their first syllable; cf. the comments on §II.L.2.7.2., §II.L.2.5. above) is longer than the one given by ʾAbū al-Faraj in the section on Image in al-Kitāb al-Kāfī (ed. Khan, Gallego and Olszowy-Schlanger 2003, §I.22.6.–§I.22.13.). It does, however, correspond closely to the fuller inventory of conjugations of Image that is found in Kitāb al-ʿUqūd (ed. Vidro 2103, 96-151). Some of these are hypothetical past forms that require comment.

The list here in Hidāyat al-Qāriʾ includes the two forms Image ‘he threw’ and Image ‘he threw’. Kitāb al-ʿUqūd (ed. Vidro 2013a, 128-130) distinguishes between the two conjugations Image (imperative)—Image (past) and Image (imperative)—Image (past). The first lacks a yod in both imperative and past whereas the second has a yod in both forms. The purpose of distinguishing the two conjugations was to account for the lack of yod in the attested imperative and short imperfect forms such as Image‘and he cast’ (Jud. 9.17) versus the presence of yod in other inflections such as Image (past), Image (active participle) and Image (infinitive). In the first conjugation the past form is hypothetical (Image) whereas in the second conjugation the imperative is hypothetical (Imagewith yod). Our passage in Hidāyat al-Qāriʾ has the two past forms Imageand Image. It appears that this is a scribal error in the extant manuscripts and this should read Image and Image (without yod).

The past form Image and its imperative form Image are inferred from Image ‘stubbornness is iniquity and idolatry’ (1 Sam. 15.23); cf. Kitāb al-ʿUqūd (ed. Vidro 2013a, 131), Ibn Nūḥ, Diqduq (ed. Khan 2000b, 370–71).

The past form Image ‘he sought’ and its imperative form Image are inferred from attested forms such as Image ‘those who seek your face’’ (Psa. 24.6), which does not have dagesh in the qof; cf. Kitāb al-ʿUqūd (ed. Vidro 2013a, 138-141), Ibn Nūḥ, Diqduq (ed. Khan 2000b, 61) and an early anonymous Karaite treatise (ed. Khan 2000a, 106–15).

The past form Image ‘he bent (a bow)’ and its imperative Image are inferred from the attested form ּ֩ImageImage ‘and they bent their tongue like their bow in falsehood’ (Jer. 9.2); cf. Kitāb al-ʿUqūd (ed. Vidro 2013a, 138-139).

Conjugations represented by the symbol Image contain a ḥireq in the first syllable of their imperative form and a qameṣ in the first syllable of their past form (al-Kitāb al-Kāfī, ed. Khan, Gallego and Olszowy-Schlanger 2003, §I.22.20.–§I.22.24.; Kitāb al-ʿUqūd, ed. Vidro 2013a, 172-189, Meʾor ʿAyin ed. Zislin 1990, 129), viz.Image (imperative)—Image (past), Image (imperative)—Image (hypothetical past), Image (imperative)—Image (past), Image (hypothetical imperative)— Image (past) inferred from Image ‘and his tongue will rot’ (Zech. 14.12) and Image ‘my wounds have grown foul and have festered’ (Psa. 38.6), Image (hypothetical imperative)—Image (hypothetical past)—Image (active participle) inferred fromImage‘they are entangled in the land’ (Exod. 14.3). The imperative form Image which is given in this section in Hidāyat al-Qāriʾ as an imperative of the symbol Image, does not appear in the lists of conjugations in the Karaite grammatical texts and it is not clear from what attested forms it could be inferred. The lists offer other imperatives that are close in form, viz. Image inferred from Image ‘they shall be turned back’ (Isa. 42.17) and ּ֩ ה ִּּ נ ֹ ו ַ ע inferred fromImage as it is sieved in a sieve’ (Amos 9.9) (Kitāb al-ʿUqūd ed. Vidro 2013a, 180-183). It is possible that Image is a scribal conflation of these two forms.

The symbol Image included conjugations that had ṣere in the final syllable of the imperative and pataḥ in the final syllable of the past form (al-Kitāb al-Kāfī, ed. Khan, Gallego and Olszowy-Schlanger 2003, §I.22.36; Kitāb al-ʿUqūd, ed. Vidro 2013a, 220- 42; Meʾor ʿAyin, ed. Zislin 1990, 135-36). The pataḥ in the hypothetical past form Image would have been inferred from forms such as Image‘I walked’ (Gen. 24.40).

➢ II.L.2.10.2.

אמהמ התב ךרחתי אל אנכאס אושלא התחת ידלא ףרחלא ראצפ הדעב אממ הלצפו התחת אוה Image הלבק אמ ילא
Whatever letter it occurs under is not mobile at all and it (the quiescent shewa) cuts it off from what is after it and conjoins it to what is before it.

82For this function of silent shewa see vol. 1, §I.2.5.2.

➢ II.L.2.12.

83Some of the material in this section on the mobile shewa has parallels in earlier Masoretic treatises such as §11-§15 of Baer and Strack’s (1879) corpus, Diqduqe ha-Ṭeʿmim (ed. Dotan 1967, §5, §14) and the Treatise on the Shewa (ed. Levy 1936).

➢ II.L.

היעגלא לוכד העמ ןסחי ןא ןאלא Image וחנ אוס ךלמלא Image ואולא תחת אושלא Image היעגלא העמ ןוכת םל אמל אפפכמ ףלאלא תחת אמ לתמ קרחב ואולא
Now, if it is appropriate for gaʿya to occur on it, it is pronounced equal to the vowel (following it), for example, ְ או ִּ כ־ם ָ֣ כ ָ ה ‘and if thus’ (Num. 11.15), in which shewa is under vav and the vav is pronounced with ḥireq like that which is under the ʾalef, but shortened since it it does not have gaʿya.

84The argument here is that a shewa with a gaʿya behaves like a full vowel and so would be long in an open syllable.

➢ II.L.

אושפ Image הדעב אמ ילע רוצקמ ריג דוילא אושו הלבק אמ ילע רוצקמ ףורח
The shewa of the four letters is restricted by what is before it, but the shewa of yod is not restricted by what is after it.

85The intention here seems to be as follows. When a shewa is followed by a guttural consonant, its realization is determined by whether or not it is preceded by a guttural consonant. The realization of shewa before yod, however, is not affected by the vowel that follows the yod.

➢ II.L.

היעגלל בא ןיא
The gaʿya has no principle.

In modern terminology this is tantamount to saying that the gaʿya has no phonological function but is only an optional phonetic phenomenon. This applies to cases where, according to the discussion in this passage, the reader has a choice as to whether to pronounce gaʿya or not. It would not apply to cases such as Image ‘they [do not] sleep’ (Prov. 4.16)—Image ‘and they did it a second time’ (1 Kings 18.34), where the gaʿya is obligatory.

➢ II.L.2.12.5.

תלק אדאפ ראעתסמ ףלא םת ןאפ Image ןכאס אושלא ראצ
For if there is an added ʾalef and you say [ʔƐ∫ˈtʰaːjim] and [ʔƐ∫ˈtʰeː], the shewa becomes quiescent.

86In the available manuscripts, the prosthetic ʾalef is not written. It is explicitly written in discussions of this topic in some other Masoretic treatises, e.g. Treatise on the Shewa and the other sources discussed by Levy (1936, 31–33).

➢ II.L.2.12.6.

87Parallels to this section are found in §14 of Baer and Strack’s (1879) corpus of Masoretic treatises and in Kitāb al-Muṣawwitāt (ed. Allony 1963, 146; 1983, 116).

For the ḥaṭef qameṣ in the closed syllable of Image‘drag her! ’ (Ezek. 32.20) see vol. 1, §I.2.5.1.

➢ II.L.2.12.9.

The point of adducing Image ‘Guard my life and deliver me’ (Psa. 25.20) and Image‘draw her down’ (Ezek. 32.20) as an objection seems to be that in some manuscripts the short qameṣ in words such asּ֩Image and ּ֩Image was represented by ḥaṭef qameṣ, and this could be construed as a mobile shewa, which would be followed by quiescence in the following letter. The response is referring to standard Tiberian vocalization, in which the shortness is not represented by using a ḥaṭef qameṣ in a closed syllable, but there is reliance only on the oral tradition.

➢ II.L.2.12.10.

88The argument of this subsection complements that of the ninth subsection. The claim is that a mobile shewa has to be followed by a vowel and then, on the third letter, a quiescent shewa may occur.

➢ II.L.2.13.1.

89See vol. 1, §I.

➢ II.L.2.13.2.

90For an explanation for the reading of the shewa as silent in the six words that deviate from this rule, see vol. 1, §I.

➢ II.L.2.13.3.

91For the distribution of vocalic shewa after conjunctive vav, see vol. 1, §I.

➢ II.L.2.15.

92For the theory of vowel production described in this section, see vol. 1, §I.2.1.3.

➢ II.L.2.16.

In the Karaite grammatical works the term ʿaraḍ is normally used to designate a noun referring to an abstract quality that is an attribute of an entity, e.g. Ibn Nūḥ, Diqduq (ed. Khan 2000b, 238): Image ‘weeping’ (Psa. 30.6), Image ‘rebellion’ (Num. 17.25), also al-Kitāb al-Kāfī (ed. Khan, Gallego and Olszowy-Schlanger 2003, xliv). Here the concept of ʿaraḍ has been extended to include also finite verbs.

➢ II.L.2.17.

93This section is an analysis of the differences between vowels within the framework of the theory of the production of vowels that is adopted elsewhere in Hidāyat al-Qāriʾ, see comments on §II.L.2.3. above.

➢ II.L.3.2.1.

94Cf. §44 in the Masoretic corpus published by Baer and Strack (1879).

➢ II.L.3.2.3.

95Cf. §56 in the Masoretic corpus published by Baer and Strack (1879).

התלתק עגולא תישכ דבעלא ינקלקא אהטאברו
Their mnemonic combination is ‘The slave has vexed me; I have feared pain; I have killed him’.

96For such mnemonic devices see Vidro (2013b).

➢ II.L.3.2.4.

97Cf. §40 and §72 in the Masoretic corpus published by Baer and Strack (1879); Seder ha-Simanim (Allony 1965, חכ־זכ ) and Kitāb al-Muṣawwitāt (Allony 1963, 148–50; 1983, 106–9). See also vol. 1, §I.2.11.

➢ II.L.3.2.5.

98Cf. Diqduqe ha-Ṭeʿamim (ed. Dotan 1967, §7). For the exceptional cases see the discussion in vol. 1, §I.2.3.2., §I.2.11.

➢ II.L.3.2.6.

99Cf. Diqduqe ha-Ṭeʿamim (ed. Dotan 1967, §8). For the exceptional cases see the discussion in vol. 1, §I.2.3.2., §I.2.11.

➢ II.L.3.2.7.

100Cf. Diqduqe ha-Ṭeʿamim (ed. Dotan 1967, §21). Cf. vol. 1, §I.

Short Version

➢ II.S.0.0.

101The short version begins with an authorial introduction, in which ʾAbū al-Faraj states that he has been requested to compose a short version of the longer work. As remarked in the comments above on the introduction to the long version, such an indication of a ‘request to compose’ was a standard component of introductions to Arabic works of the period and may have been fictitious, especially when, as is the case here, the requester remains anonymous.

➢ II.S.1.7.

102In this section ʾAbū al-Faraj adds a comment on the phenomenon that he discussed a length in §II.L.1.5. of the long version. He indicates that he now prefers not to refer to the phenomenon as ‘interchange’ (tabdīl), since it is possible that the change of letters in the forms in question expresses semantic modification.

➢ II.S.2.9.

The one exception that breaks this rule is Image ‘the Lord in them’ (Psa. 68.18); cf. long version, §II.L.1.7.8.

➢ II.S.4.2.

103For this notion of the vowels being realizations of ‘soft’ letters א , ו , ה , and י see vol. 1, §I.2.1.8.

➢ II.S.5.6.

אושלאב לואלא תכרח לגאל ךרחתי הנאל התב ןוכס הל ינאתלא ףרחלא יפ עקי אל ךרחתמלא אושלא ןא םלעאו
Take note that a mobile shewa never occurs on a second letter (of a word) that has a silent (following letter), because it (this second letter) is mobile due to the mobility of the first letter with shewa.

I.e. if the first letter of a word has shewa, which must be mobile, the second letter cannot have a mobile shewa. This second letter has a vowel and this is followed by a silent letter, which is either a normal consonant, e.g. Image or a ‘soft’ letter inside a long vowel (see short version §II.S.4.2.), e.g.Image.

➢ II.S.5.7.

104Both the qameṣ and the stressed pataḥ of the cited words were pronounced long, so, according to the theory of ‘soft’ letters (cf. short version §II.S.4.2.), they contained an ʾalef, although this does not appear in the orthography.

➢ II.S.6.11.

105The argument here seems to be that a mobile shewa is dependent on a mobile letter with a vowel, which, therefore, obligatorily follows a mobile shewa. A mobile shewa cannot be dependent on another mobile shewa. A quiescent shewa can come after a sequence of mobile shewa and a letter with a vowel, but not immediately after the initial mobile shewa.


1 See Freimark (1967) and the discussion and references in Goldstein (2014).

2 The various passages in al-Kitāb al-Kāfī have been gathered together by Gallego (2003).

3 See this commentary of al-Fārābī (ed. Beirut 1960, 27, 50–51) and Zimmermann’s translation (1982, xli, 12).

4 Sefer ʾEshkol ha-Kofer (ed. Eupatoria 1836, 70a): כי בלא חמשת הקולות שהם מלכי הנקוד לא תולד המלה ולא תעמוד בפה ולא יודע מה היא כי אם בנועם נקודים וטעמים. See further Khan (1992, 173).

5 For the full references see Gil (1996, 165–66).

6 See Khan, Gallego and Olszowy-Schlanger (2003, xiii) for a list of the extant manuscripts.

7 II Firkovitch, Evr.-Arab. I 4601, fol. 112 (an autograph of ʾAbū al-Faraj الكالم المستعمل ثلثة اقسام اسم وفعل وحرف يسميه اهل علم :)written in Arabic script .الدقدوق خادم

8 ואלמסתעמל הו Imageפי אלתצריף כלה Image זאל בקית חרוף אללגה Image מן גיר מעני מסתקר פיהא Imageאללמאד אלד̇י פי לגה Image ואל[נ]ון פי לגה אלרחיל ולגה אלדנו Image לוקחים נוסעים נוגעים ImageImage ממא אלפרק בין זואל אללמאד אלמסתעמל ובין זואל אלקוף או אלחית פרקא ImageImage כל חרף מסתעמלImageאזיל .ואזיל אלחרף אלאצלי כאן אלפרק בינהמא ביינאImage

9 Maimonides followed an Aristotelian view and regarded Adam’s act of naming as indicating by itself that languages are conventional, see Guide of the Perplexed (transl. Pines 1969, 357). By contrast, Jewish exegetes in medieval Europe adopted a revelationist interpretation. Naḥmanides, at the height of the ‘Maimonidean controversy’, explicitly rejected Maimonides’ notion of convention; cf. Saenz-Badillos (2004, 297).


Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search