236 £arf Sìbawayhi, Kitàb = ±Abù Bišr ≠Amr ibn ≠Uμmàn ——. 1995. The explanation of linguistic causes: Sìbawayhi, al-Kitàb. 2 vols. Ed. Bulaq, 1316 A.H. Az-Za©©à©ì’s theory of grammar. Amsterdam and (Repr., Baghdad: Al-Muthanna Library, n.d.) Philadelphia: J. Benjamins. Xalìl, ≠Ayn = ±Abù ≠Abd ar-Ra™màn al-Xalìl ibn ±A™mad al-Faràhìdì, Kitàb al-≠ayn. Ed. Mahdì al- Kees Versteegh (University of Nijmegen) Maxzùmì and ±Ibràhìm as-Samarrà±ì. 8 vols. Beirut: Mu±assasat al-±A≠lamì li-l-Ma†bù≠àt, 1988. Xwàrizmì, Mafàtì™ = ±Abù ≠Abdallàh Mu™ammad ibn ±A™mad al-Xwàrizmì, Kitàb mafàtì™ al-≠ulùm. Ed. Gerlof van Vloten. Leiden: E.J. Brill, 1895. £arf Zajjàjì, ±î∂à™ = ±Abù l-Qàsim ≠Abd ar-Ra™màn ibn ±Is™àq az-Zajjàjì, al-±î∂à™ fì ≠ilal an-na™w. Ed. The term ™arf (pl. ™urùf, a™ruf) ‘part, particle, Màzin al-Mubàrak. Cairo: Dàr al-≠Urùba, 1959. edge, end, boundary’ is used in Arabic linguistic terminology to indicate (1) the final segment Secondary sources formed as a result of the linear segmentation Al-Nassir, A.A. [≠Abd al-Mun≠im ≠Abd al-±Amìr an- Nàßir]. 1993. Sibawayh the phonologist: A critical of the Arabic word; (2) a component of the study of the phonetic and phonological theory of prosodic, morphological, and lexical pattern Sibawayh as presented in his treatise Al-Kitab. Lon- of a word; (3) any discrete unit of an Arabic don and New York: Kegan Paul International. Bauer, Gertrud. 1972. Athanasius von Qus, Qilàdat text that has a linguistic function (word, mor- at-ta™rìr fì ≠ilm at-tafsìr: Eine koptische Gramma- pheme); (5) a certain class of linguistic units; tik in arabischer Sprache aus dem 13./14/Jahrhun- or (6) one of the parts of speech, against nouns dert. Freiburg: K. Schwarz. and verbs. Bohas, Georges. 1985. “L’explication en phonolo- gie arabe”. Zeitschrift für Arabische Linguistik As a polyvalent notion, the term ™arf has no 15.45–52. equivalent in the conceptual system of Euro- —— and Jean-Patrick Guillaume. 1984. Etude des pean linguistics (Frolov 1991:57). Its polysemy théories des grammairiens arabes. I. Morpholo- derives from the systematic use of the same term gie et phonologie. Damascus: Institut Français de Damas. at different levels, not only within one scientific Ermers, Robert. 1999. Arabic grammars of Turkic: domain but also across a broad range of subjects, The Arabic linguistic model applied to foreign a characteristic typical of the entire spectrum of languages and translation of ±Abù £ayyàn al- medieval Arabic science (al-≠ulùm al-≠arabiyya). ±Andalusì’s Kitàb al-±idràk li-lisàn al-±atràk. Leiden: E.J. Brill. This is not the consequence of undifferentiated Kraus, Paul. 1942. Jàbir ibn £ayyàn, contribution à functional contents of the denoted units; rather, l’histoire des idées scientifiques dans l’Islam: Jàbir it stems from the conceptual perception of lan- et la science grecque. Cairo. (Repr., Paris: Les Belles guage as a unitary process (Carter 2004:53), Lettres, 1986.) Mehiri, Abdelkader. 1973. Les théories grammati- whereby the model of language structure is cales d’Ibn Jinnì. Tunis: Université de Tunis. linear and one-dimensional and lacks multi- Owens, Jonathan. 1990. Early Arabic grammati- tier organization (Frolov 1991:134). The term cal theory: Heterogeneity and standardization. Amsterdam and Philadelphia: J. Benjamins. ™arf can be applied to elements of any size Revell, E.J. 1975. “The diacritical dots and the devel- and length, as long as it denotes a quantum opment of the Arabic alphabet”. Journal of Semitic of enunciated and hence recorded information Studies 2.178–190. that is small in scale but not strictly limited Schaade, A. 1911. Sìbawaihi’s Lautlehre. Leiden: E.J. Brill. (Weiss 1910:375–379; Fischer 1989:140), its Semaan, Khalil I. 1977. Arabic phonetics: Ibn Sìna’s actual meaning depending entirely on the con- Risàlah on the points of articulation of the speech- text. Building upon the perception of the term sounds translated from Medieval Arabic. 2nd ed. ™arf as an ultimate unit (Ibn Jinnì, Sirr I, 15–19) Lahore: Sh. Muhammad Ashraf. (1st ed., 1963.) Talmon, Rafael. 1997. Kitàb al-≠ayn and its attribu- with its own meaning (Fleisch 1986:204b), it tion to ‡alìl b. A™mad. Leiden: E.J. Brill. may be treated as a segment of speech with Troupeau, Gérard. 1989. “Voyelles et semi-voyelles a semiological value, specified both semanti- dans le Kitàb de Sìbawayhi”. Jerusalem Studies in cally and semiotically. Consequently, two gen- Arabic and Islam 12.31–39. Versteegh, Kees. 1985. “The development of argu- eral meanings of the word ™arf can be viewed mentation in Arabic grammar: The declension of as most prominent in early Arabic texts: a unit the dual and the plural”. Studies in the history of with a syntactic status (semantically specified), Arabic grammar, II, ed. Michael G. Carter and Kees on the one hand, and a unit of a phonological Versteegh, 152–173. Amsterdam: J. Benjamins. ——. 1993. Arabic grammar and Qur±ànic exegesis (scriptural) character (semiotically specified), on in early Islam. Leiden: E.J. Brill. the other (Owens 1990:245). (c) 2011 Koninklijke Brill NV. All Rights Reserved. £arf 237 Both meanings derive from the most compre- as an operational element (≠àmil ‘operator’; ¤ hensive definition of the term ™arf, by Sìbawayhi ≠amal). Since particles are the exclusive instru- (Kitàb I, 1), as a word that “has a meaning, ment for performing linguistic functions, each but is neither a noun nor a verb” (™arf jà±a of them can be defined by its function, for li-ma≠nan laysa bi-smin wa-là fi≠l). The second example ™arf istifhàm ‘interrogative particle’, part of this definition, ‘neither a noun nor a ™arf nidà± ‘vocative particle’, ™arf qasam ‘par- verb’, implies that the notion ™arf may apply to ticle of oath, vow’, and so on (Carter 2004:88). both nouns and verbs. It is used to indicate the These functions form specific subclasses of par- class of words in general, which can be broken ticles: ™arf aš-širàk (Sìbawayhi, Kitàb I, 211.10, down further into three subclasses. The third 264.15, 382.17) ‘coordinate particle’; ™arf jarr subclass is represented by those units from the (Kitàb I, 244.1, 252.12, 329.14) or the posses- ™arf category that are neither nouns nor verbs sive ™arf (Kitàb I, 276.16), the ‘preposition’ that (a particular usage of the term is the subclass controls the -i inflection. ±Inna and the particles of particles lacking a common semantic charac- that govern two nominal complements (¤ ±inna teristic and morphological regulation), because wa-±axawàtuhà) are also called ™arf (Kitàb I, noun words and verb words have as their basis 241.13, 244.14). Other subclasses of particles a precisely defined set of morphological models called ™arf include particles of warning (tanbìh; and a set range of meanings that only vary Kitàb I, 277.18), particles of exception (Kitàb insignificantly. Particles do not have any struc- I, 314.17), and verb operators (Kitàb I, 361.13, tured form or model whatsoever in their basis, 363.11, 406.1), realized at the beginning of and they may have a wide range of meanings the sentence. Other particles include those (Carter 2004:88). that occur sentence-initially (Kitàb I, 244.12, As far as the first part of the definition is 367.15, 391.10, 429.20), those that must be concerned (a word that ‘has a meaning’), this followed by a verb (Kitàb I, 407.16), and con- feature implicitly puts the ™arf in opposition to ditional particles (Kitàb I, 398.8). the eponymous category of units, which even In the phonological sense, the term ™arf is though lacking semantic meaning (cf. Versteegh viewed as a phoneme. This is validated by data 1977:44, 45; Carter 2004:75; but according coming from early philological texts, which to Levin [2000:45], the correct interpretation present the behavior of sounds in phonetic of this phrase in Sìbawayhi is ‘which occurs processes through the term ™urùf, as well as in order to denote a meaning and nothing by research done by later philologists, such as else [except this specific meaning]’), nonethe- az-Zamaxšarì (see Ibn Ya≠ìš, Šar™ X, 120–155, less have a semiotic function. It is the nature of 124.7-8; Carter 2004:120–121). In descriptions that unit that generates the most frequent use of their articulatory characteristics, reference of the term ™arf. By virtue of the ambiguity of is made to the place of articulation (maxraj its content, this use has produced a wide array lit. ‘place from which [the sound] exits’) of of explanations and interpretations in Western the ™urùf. It may be concluded from this that studies of Arabic, because in this connotation Arab philologists view the ™arf as a discrete, it correlates with a whole number of similar phonetically differentiated sound (ßawt). In this (in terms of the system) concepts of theoretical sense, ™arf is a subclass of ßawt, a sound with linguistics, such as phoneme, grapheme, pros- specific and discrete features (Owens 1988:91, odeme, and morpheme (see Fischer 1989 for a 95). The correlation of the notions of ™arf general review). Attempts at viewing any single and ¤ ™araka is explained through the vowel/ one of these aspects independently face consid- consonant contrast (Levin 1986:425). In terms erable obstacles, created by their interconnected of the graphical realization (™arf yatahajjì bihi nature, because such an approach goes against or ™urùf li-l-hijà±; see Weiss 1910:357, 359– the functional universality of a unit postulated 360), this approach leads to the perception in the Arabic system of grammar. of the ™arf as the equivalent of the ‘letter of The syntactic role of the ™arf (in the narrow the alphabet’ concept (Fleisch 1986:204b; Al- sense of ‘particle’) is defined by its instrumental Nassir 1993:9–10). function, a fact supported by the alternative On the other hand, it is noteworthy that in name of this subclass, ±adà lit. ‘tool’ (Carter more recent grammatical treatises (Ibn Sìnà, 2004:74). In many cases, ™arf manifests itself Risàla, Chap. 2), the terms ™arf and ¤ ™araka (c) 2011 Koninklijke Brill NV. All Rights Reserved. 238 £arf are used to indicate consonants and vowels ing unit. Any segment of Arabic speech can when they are viewed functionally as compo- be coded as a sequence of ™urùf. Through the nents of a word, but when the sounds of speech application of minor modifications, the struc- are viewed as independent units of the phono- ture of the Arabic written language is derived logical system, the terms ßawt ßàmit ‘consonant’ from this code system (San∑es 1968:91). Oth- and ßà±it or mußawwat ‘vowel’ (Axvlediani ers have suggested that as a term, ™arf does not 1981:93; Bravmann 1934:7–18, 112–135) are indicate any physical substrate at all but rather preferred. The structure of a word in the Arabic is a diacritic element, capable of transforming linguistic tradition is described and mapped in into one physical substrate or another, acoustic ™urùf, which suggests that ¤ ™araka should be or scriptural (Gabu∑an 1965:120). regarded as a vowel element, integrated into the The analysis of the basic notions of the ™arf, rather than a vowel sound that is added Arabic prosodic system (≠arù∂), in which ™arf to ™arf. Consequently, sound segments, such as occurs as the main operational notion (San∑es /ba/, /bu/, /bi/, /bØ/ (i.e. b with sukùn), should 1968:86), also suggests that ™arf is a unit, be interpreted as different variants of the same largely similar to the notion of ‘mora’ (Tru- ™arf that emerge as a result of a variation in betzkoy 1977:169–179; Al-Nassir 1993:9–10). the vowel component of the ™arf, while the As such, it is used both in morphology, where consonant remains invariant. According to al- it has a quantitative parameter, and in metrics, Xalìl ibn ±A™mad, every ™arf has its own sound where ™arf indicates metric feet and their com- and variation (Sìbawayhi, Kitàb II, 342.21–23: ponents (Frolov 1991:54). This conclusion is wa-za≠ama l-Xalìl ±anna l-fat™a wa-l-kasra wa- supported by the analysis of the terminological ∂-∂amma zawà±id wa-hunna yal™aqna l-™arf pair ™arf muta™arrik and ™arf sàkin. Underly- li-yùßala ±ilà t-takallum bihi wa-l-binà± huwa ing this contrast is the polarity of the prosodic s-sàkin alla≈ì là ziyàdata fìhi ‘al-Xalìl claims functions of two types of ™arf, rather than the that the /a/, the /i/, and the /u/ are additions modality of their phonetic realization. It is no that attach themselves to the consonant so coincidence that despite their reduced conso- as to make it possible to pronounce it; the nantal status, the so-called ™urùf al-madd wa-l- pausal form is the vowelless consonant without lìn are thought to belong to the ™arf sàkin class addition’). Variation is revealed in the ™araka on the basis of their prosodic function. (Gabu∑an 1965:121). In morphological terms, the notion ™arf can Alternatively, it has been noted that, unlike be interpreted as a (morphological) position. the classic definition of the phoneme as a Relevant here is the distinction between basic unit with differentiating features as well as an ™urùf (™urùf ±aßliyya: root positions of the integrative (morpheme-forming) function, the base depicted through the symbols fà±, ≠ayn, consonantal component ™arf typically has only làm, which are open to realization by phono- differentiating features, whereas the implemen- logically perceived segments) and augmented tation of the integrative function is delegated ™urùf (™urùf zà±ida: positions added to the to the element ™araka. This is why the status base, closed to realization, and represented of the phonological segment capable of form- by one segment from a given list, abstracted ing the exponent of a morpheme requires the from its exact phonetic meaning, occurring combination of both elements (Karabekyan only as a quantitatively structuring morpho- 2004:510). With this approach, ™arf, corre- logical model of a diacritic notion). One of sponding not only to a separate consonant the key questions here is the morphological (™arf sàkin), but also to a combination of status of ™urùf ±aßliyya and ™urùf zà±ida. On sounds (™arf muta™arrik), can be correlated the one hand, both the ™urùf ±aßliyya aggregate with the notion of ‘grapheme’ rather than that (as root morphemes) and the separate ™urùf of ‘letter’ (Frolov 1991:56, 57). Conceptually, zà±ida (as analogues of affixes) are customarily this is linked with the general methodological explained as morphemes. At the same time, the premise of the Arab linguistic tradition, which notions ™urùf ±aßliyya and ™urùf zà±ida are a considered the spoken word to be prior to the complex morphological characteristic (fa≠ala; written (Owens 1988:284). On the other hand, tafa≠ Ø≠ala, etc.) of the segment of speech they it should be noted that as an element of the indicate (kataba, jalasa, rasama; tarakØkaba, script system, ™arf is in the first place a cod- tanafØfasa, etc.), rather than the segment of (c) 2011 Koninklijke Brill NV. All Rights Reserved. £arf 239 speech itself. Therefore, it is not the ™urùf Secondary sources ±aßliyya and ™urùf zà±ida that have the status of Al-Nassir, ≠Abd al-Mun≠im ≠Abd al-±Amìr. 1993. Sib- awayh the phonologist: A critical study of the a morpheme, but this status is rather assigned phonetic and phonological theory of Sibawayh to segments that are formed by a certain series as presented in his treatise Al-Kitab. London and of concrete realizations of the open positions New York: Kegan Paul International. (™urùf ±aßliyya), as well as at a certain local- Axvlediani, Vladimer Grigor’evi∑. 1981. “Arabskoe jazykoznanie srednix vekov” [Arabic linguistics ization of the closed positions (™urùf zà±ida). of the Middle Ages]. Istorija lingvisti∑eskix u∑enii: Furthermore, the ™urùf zà±ida of this type Srednevekovyj Vostok [History of linguistics: The (tà±, sìn, hamza, etc.) differ in function from medieval East], ed. Agnija V. Desnickaja and Solo- the ™urùf zà±ida that additionally have their mon D. Kacnel’son, 53–94. Leningrad: Nauka. Bravmann, Max. 1934. Materialien und Untersu- own discrete nominal or nominal-relative value chungen zu den phonetischen Lehren der Araber. (™urùf al-mu∂àra≠a, tanwìn, tà± at-taμniya, etc.). Göttingen: W.F. Kaestner. The latter are considered to be positions dis- Carter, Michael G. 2004. Sìbawayhi. Oxford: Oxford University Press. tributed around the basis that shape the word Fischer, Wolfdietrich. 1989. “Zur Herkunft des gram- form. Unlike the ™urùf zà±ida of the former matischen Terminus ™arf ”. Jerusalem Studies in type, they are partially open to realization by Arabic and Islam 12.135–145. a strictly defined set of segments that modify Fleisch, Henri. 1986. “£arf ”. Encyclopaedia of Islam III, 204. 2nd ed. Leiden: E.J. Brill. the overall meaning of the given position. For Frolov, Dmitrij V. 1991. Klassi∑eskij arabskij stix example, the position tanwìn may be realized [Classical Arabic verse]. Moscow: Nauka. in two ways, the nØ (nùn sàkina) and the null Gabu∑an, Gra∑ija Mikajelovi∑. 1965. “K voprosu one, whereas the position of, for instance, the o strukture semitskogo slova (v svjazi s pro- blemoj ‘vnutrennej fleksii’)” [On the question of ™urùf al-mu∂àra≠a allows for four realizations, the structure of a Semitic word (in relation to the and so forth. issue of the ‘inner inflection’]. Semitskie Jazyki This leads us to believe that the morpho- 2:1.114–127. logical analysis of the Arabic linguistic tradi- Karabekyan, Samvel B. 2004. “Urovni lingvisti∑eskogo analiza i osnovnyje jedinicy opisanija v tradi- tion is based on a range of interrelated factors cionnom arabskom jazykoznanii (k postanovke whereby the variability of characters in the problemy)” [The levels of linguistic analysis and ™urùf ±aßliyya and their invariance in the ™urùf the basic units of description in the Arabic linguis- zà±ida, on the one hand, and the preservation tic tradition]. Countries and Peoples of the Near and Middle East (Yerevan) 23.507–518. of the quantitative integrity of ™arf as a unit Levin, Aryeh. 1986. “The Mediaeval Arabic term of morphological calculation, on the other, kalima and the modern linguistic term morpheme: form a complex system that constitutes the Similarities and differences”. Jerusalem Studies in Arabic and Islam 7.423–446. main mechanism for forming both the expres- ——. 2000. “The meaning of ™arf ©à±a li-ma≠nan in sion and the content plan of a word within, as Sìbawayhi’s al-Kitàb”. Jerusalem Studies in Arabic well as outside, the root base. The functioning and Islam 24.22–48. of this system itself is based on the conceptual Mosel, Ulrike. 1975. Die syntaktische Terminologie bei Sìbawaih. Ph.D. diss., University of Munich. nature of ™arf as a universal operational unit of Owens, Jonathan. 1988. The foundations of gram- linguistic analysis. mar. Amsterdam and Philadelphia: J. Benjamins. ——. 1990. Early Arabic grammatical theory: Het- erogeneity and standardization. Amsterdam and Philadelphia: J. Benjamins. Bibliographical references San∑es [Sánchez], Alcaén A. 1968. “K voprosu o suš∑nosti sistemy arabskoj metriki” [On the ques- Primary sources tion of the essence of the Arabic metrical system]. Ibn Jinnì, Sirr = ±Abù l-Fat™ ≠Uμmàn Ibn Jinnì, Sirr Arabskaja filologija: Sbornik statej, ed. Aleksandr ßinà≠at al-±i≠ràb. 3 vols. Ed. ±A™mad Farìd ±A™mad. Aleksandrovi∑ Kovalev and Gra∑ija Mikajelovi∑ Cairo, n.d. Gabu∑an, 86–95. Moscow: University Press. Ibn Sìnà, Risàla = ±Abù ≠Alì al-£usayn ibn ≠Abdallàh Trubetzkoy, Nikolaj S. 1977. Grundzüge der Phono- Ibn Sìnà, Risàla ±asbàb ™udùμ al-™urùf. Ed. logie. 2nd ed. Repr. Göttingen: Vandenhoeck und Vladimer Grigor’evi∑ Axvlediani, Foneti∑eskij Ruprecht. (1st ed., 1958.) traktat Avicenny. Tbilisi: Mecniereba, 1966. Versteegh, Cornelis [Kees] H.M. 1977. Greek elements Ibn Ya≠ìš, Šar™ = Muwaffaq ad-Dìn Ya≠ìš Ibn Ya≠ìš, in Arabic linguistic thinking. Leiden: E.J. Brill. Šar™ al-Mufaßßal. 10 vols. Beirut and Cairo: ≠âlam Weiss, Josef. 1910. “Die arabische Nationalgramma- al-Kutub and Maktabat al-Mutanabbì, n.d. tik und die Lateiner”. Zeitschrift der Deutschen Sìbawayhi, Kitàb = ±Abù Bišr ≠Amr ibn ≠Uμmàn Sìba- Morgenländischen Gesellschaft 64.349–390. wayhi, al-Kitàb. 2 vols. Ed. Hartwig Derenbourg. Paris: Imprimerie Nationale, 1881. (Repr., Hildes- Samvel Karabekyan and Marat Yavrumyan heim and New York: G. Olms, 1970.) (Yerevan State University) (c) 2011 Koninklijke Brill NV. All Rights Reserved.