Classification of the Parts of Speech in Arabic

In the Arabic linguistic tradition, the classification of the parts of speech (ism ‘noun’, fiʿl ‘verb’, ḥarf ‘particle’) is first introduced by Sībawayhi, who presents the three key elements in his Kitāb (I:1). The section at issue includes the presentation of the elements but does not provide much in terms of grammatical explanation. Nouns are in fact not introduced with their grammatical characteristics, but rather with examples: fa-l-ism: raǧul, wa-faras, wa-ḥāʾiṭ (“and the noun is ‘man’, and ‘horse’, and ‘wall’”) (Kitāb I:1). In addition to nouns, verbs and particles, Arabic grammar further recognized a number of other categories that are not considered as parts of speech but rather fall into the main three. This contribution aims to present relevant classifications of the parts of speech in the Greek and Arabic traditions, with the aim to account for possible external influences on the Arabic formulations, and foster further discussion on the development of the Arabic grammatical disciplines.


Introduction and Preliminary Assumptions1
The Kitāb Sībawayhi (8th century), the earliest known treatise on Arabic grammar, introduces most if not all aspects of the discipline. Across the linguistic 2 tradition, Sībawayhi (d. 180/796?) is considered a most authoritative figure and his legacy is of utmost importance for coeval as well as later grammarians. However, the origins of Sībawayhi's assertions are vague and, in general, it is difficult to assess the diverse influences that may have played a role in the development of his linguistic theorizations.
As we have discussed in another paper, the technical lexicon presented in the Kitāb seems "to represent a fully-fledged system, based on a well-structured schema. Despite the fact that Sībawayhi does not provide clear technical definitions of the terminology used, the technical meaning is instead inferred from the application of the same term in a number of examples provided" (Olivieri 2020, 6).
While Olivieri (2020) addresses a specific case of the specialized lexicon, namely the ʾiʿrāb, in this paper we will discuss the topic of the parts of speech, taking the category of nouns as a case in point,2 more in general terms and as per it is treated in the Arabic linguistic tradition. The ultimate aim is to present a reading of the matter in the light of possible connections with traditions like the Greek that address the topic in similar ways. Finally, the objective is to foster a discussion on the possible extent of such connections, and inquiring if and to what extent Sībawayhi and other Arabic grammarians might have been influenced by other traditions.

1.1
Previous Studies Modern scholarship has dealt with the origins of Arabic grammatical and linguistic studies3 from different perspectives, so that for instance Carter discusses influences from other Islamic sciences, focusing in particular on the legal studies (Carter 1972;2001). Further to this, he also discusses a possible influence of the naḥwiyyūn, 'grammarians' (cf. Carter 1972, 76), which would be more evident in the introductory part to the Kitāb. On the latter, a thorough discussion is in the paper by Talmon "Naḥwiyyūn in Sībawayhi's Kitāb" (1982). The paper deals with the general profile of these grammarians, in the light of the fact that: their comments. I also express my sincere gratitude to Prof. Shabo Talay and to the anonymous reviewers for the valuable remarks. 2 For a comprehensive discussion on the theory of the parts of speech, see Guillaume (1988), in which the author provides a thorough description of the elaboration of the theory as well as of its development in the Arabic linguistic tradition. On this, see also Viain (2017).

3
Scholars concerned with the beginnings of Arabic grammatical science who wish to utilize the Kitāb as the basis for a new theory encounter considerable difficulty due to paucity of reliable and contemporary sources dealing with this subject. Ideally, such sources would throw light on the state of scientific thought on the Arabic language both previous to sībawayhi and outside of his scholarly circle, allowing the modern scholar to gauge more accurately both sībawayhi's originality and the extent to which material found in the Kitāb may have been common to others having a scientific interest in the Arabic language. (Talmon 1982, 13) Talmon, thus, addresses the issue of the -perhaps external -figures that may have played a role in the emergence of the Arabic linguistic and grammatical tradition, reflecting on the scarcity of evidences to assess the matter. However, he ultimately concludes that there must have been an influence on Sībawayhi. Therefore, his interpretation that "Sībawayhi founded his grammatical system on the groundwork of a fairly advanced school of Grammar!" (Talmon 1982, 29) is of utmost relevance to our considerations, inasmuch as the questions brought up in this paper, as well as in the discussion on the grammatical schools preceding Sībawayhi,4 pose a major point in the general approach to Sībawayhi's Kitāb, namely possible previous or coeval influences.5 Ultimately, a reasonable consideration on these questions is that we cannot be absolutely sure of who such figures may have been, and, if they ever played a role in Sībawayhi's theorizations, to what extent this may have contributed to his propositions on the whole. 4 A substantial contribution on this is Talmon's "Eighth-Century Iraqi Grammar: A Critical Exploration of Pre-Halilian Arabic Linguistics" (2003), in which the author tackles some controversial issues of the early stages of the Arabic grammatical tradition, namely the grammatical teaching of al-Ḫalīl and Sībawayhi, largely adopted by later scholars, and thepossibly external -linguistic sources adopted by the grammarians. Accounting for "possibly significant pre-Halilian elements of grammatical theory" (2003, 41), Talmon discusses how the teaching of al-Ḫalīl and Sībawayhi may be distinct from the mainstream of grammatical theory up to their time, namely from the tradition of the 'the Old Iraqi School of Grammar' (cf. also Talmon 2000). Thus, his extensive textual analysis of the contribution of earlier (i.e., the formative period of grammar before al-Ḫalīl and Sībawayhi) as well as coeval grammarians is first aimed at determining how Sībawayhi's Kitāb may represent an innovative approach to grammatical study, but also describes how other actors may have played a role in, or influenced, the formulation of the early grammatical theories.

4
Another viewpoint on the development of the Arabic linguistic disciplines is that expressed by Versteegh, who discusses how the Greek tradition may have played a role in the development of the Arabic linguistic tradition.6 In the long-standing discussion on the origins of Arabic grammar, as a preliminary assumption, this paper will mostly draw on the arguments advanced by Versteegh on the development of grammatical sciences (Versteegh 1977;1993) as possibly influenced by other traditions.7 In our view, both Talmon's conclusions and Versteegh's propositions may well be considered as an ensemble. It is true that Talmon in other papers (1985;2000) mostly discusses of the Medinean and Iraqi centers, which may have contributed to Sībawayhi's theorizations (although in the latter paper, in particular, he also presents possible connections with the Greek tradition). However, the evidence is so scarce that a broader view on the figures may still be considered for further investigation. Albeit such influences may well not be external to the Arabic tradition, we cannot in principle rule out possible external elements that might have played a role in the formulation of Sībawayhi's arguments.
On the other hand, the issue of reconstruction a possible historiographical transmission8 of, specifically, Greek notions is also to be considered in any investigation dealing with concepts transmission. Nevertheless, we intend not to disqualify possible contacts that Arabs might have had with other traditions, especially when attested historically. The kind of contact between the Arabs and the Greeks surely pertains to different domains, so that Versteegh (1977;1993) distinguishes between a direct and an indirect influence, the former linked to the logical and philosophical traditions transmitted to Arabs through the translations of the Greek scholarship; the latter instead exercised through everyday contact between Arabic and Greek civilizations in the Byzantine 6 See also Brockelmann (1937, vol. I Suppl., 156). 7 We wish to remark that advocating this approach does not imply a total rejection of other viewpoints, especially with regard to the propositions, mostly advanced by Carter, of a deep relation between other Islamic sciences and the Arabic linguistic disciplines. The earliest stages of the Islamic era, scene of outstanding intellectual activities, witnessed the raise of a number of disciplines that were profoundly intertwined (e.g., theology, kalām; Qurʾānic exegesis, tafsīr; philology, qirāʾāt; law, fiqh). Given the profound connection among these sciences, scholars would be contextually working in different fields: for instance, scholars from linguistic studies would be working on the qirāʾāt and/or law as well, and this is testified also by the fact that the earliest grammatical treatises share the technical lexicon with a number of other Islamic scholarly disciplines. The significant role played by the Islamic disciplines in the development of the Arabic grammatical tradition is thus surely undisputed. 8 On the history of transmission of the Organon in the Middle East, see among others D' Ancona (2004), and the works by Hugonnard-Roche (1989; provinces.9 Historically, in both pre-Islamic and Islamic times, Arabs have been exposed to foreign erudition and have been in a deeper contact with traditions like the Greek, even before the attested tradition of translation of Greek works of the 4th/10th century. Finally, our overview on the classification of the parts of speech will rely on possible points of convergence between the Arabic and the Greek tradition, as a possible reading of the transmission and development of grammatical concepts in the Arabic linguistic tradition. In fact, if on the one hand, it is surely true that the legal theory must be accounted for in this research,10 on the other hand, "borrowing grammatical terms does not imply a total dependency, Arabic grammar is a linguistic system in its own rights" (Versteegh 1977, 15), and thus the grammatical terminology should be analyzed deeply, also being open to possible connections to other traditions. This holds especially true for those terms that Carter (1972), too, considers as possibly influenced by other traditions.11 9 In the discussion on the subject, Versteegh points, on the one hand, to the geographical proximity as a possible reason for the con-penetration of ideas and concepts, and that thus may be accounted for in debating how "foreign" concepts and ideas might have reached the Arabs; and, on the other hand, to the movement of translation of the Greek works as a more direct type of influence. On this latter aspect, Versteegh addresses more specifically the Stoic linguistic theories. By comparing the resemblances in the formulations, and in line with Versteegh's considerations, in our discussion we hence wish to remark on the possibility that concepts might have circulated, reaching the Arabs. We suggest that the possibility that Greek concepts might have circulated in the early Islamic context and before the time of the translations of the Aristotelian works should not be ruled out entirely. With this, however, we do not aim to deny the Arabicity of the Arabic linguistic tradition, nor the original contributions of the earliest grammarians to the formulation and development of the linguistic science; our approach rather aims at discussing possible influences in the metalinguistic dimension, on which Arabic grammarians might have drawn in their studies. Finally, we do acknowledge again the evident difficulties in reconstructing possible ways of transmission, especially as we lack substantial portions of the Stoic tradition and any respective Arabic translations, nor do we have evident proofs of any Syriac intermediary, as the Greek linguistic texts that were translated into Syriac (e.g, the Technē Grammatikē) do not seem to have been translated into Arabic. 10 Carter (1972;1983) further argues that Sībawayhi's grammatical practice was an application of legal methodology and terminology to linguistic analysis. However, as also pointed out by Owens, "Carter has never shown in any detail, either terminologically or methodologically, how this transference occurred, nor has he entertained (i.e., falsified) the possibility that the direction of influence was in the opposite direction, or that the disciplines developed parallel but largely separately" (Owens 2005, 106).

6
Conversely, the arguments in favor of the external influences have witnessed a number of rather ventured statements. So for instance, Merx (1889) has proposed that Arabic grammar might have somewhat stemmed from Greek logic. The propositions advanced by Merx were built on six notions that had allegedly been imported from the Greek tradition into the Arabic: the use of the three parts of speech; ʾiʿrāb; the notion of gender; ẓarf; the notion of ḥāl; the notion of predicate. In his discussion, Merx argues that such notions had been received in the Arabic tradition via their Aristotelian propositions "ea vero, in quibus Grammatici notionibus ab Aristotele propositis usi sunt" (Merx 1889, 141).12 The arguments mostly draw on terminological similarities and conceptual connections between technical terms.
More specifically on the topic of this paper, Merx argues that the use of the three parts of speech in Arabic (noun, verb, and particle) would be connected to the Aristotelian noun, verb, and conjunction. The distinction is based on Aristotle's proposition and on the additional definitions provided by later Arabic grammarians.
However, Sībawayhi surely does not qualify as an "Aristotelian" scholar: in the Arab-Islamic world, these would rather be figures such as al-Fārābī (d. 339/950) and Ibn Rušd (d. 595/1198), who surely could not have influenced Sībawayhi. However, it still is reasonable not to rule out in advance the possibility that already at the time of Sībawayhi, Arabs might have had access to such notions and therefore grammarians might have partly drawn on these for their linguistic theorizations. Thus, if we assumed an earlier circulation of concepts from the Greek tradition and a possible Arabic reception of the Aristotelian works, perhaps via the Syriac medium,13 this would have made possible a development of the Arabic linguistic disciplines under some more general Hellenistic influence: 12 As reminded by Versteegh, "Rundgren (1976) revived Merx' thesis by pointing out that even before the period of the translations some knowledge of Greek logic and philosophy had reached the Arabs through the Persian translations that had been made at the academy of Gundishapur. These Persian translations had been translated into Syriac and some elements of Greek logic had even become available in Arabic through a treatise on logic by. Ibn al-Muqaffaʿ, either the famous translator of the Kalīla wa-Dimna (d. 142/759) or his son Muḥammad ibn ʿAbdallah Ibn al-Muqaffaʿ (d. ± 200/815)" (Versteegh 1993, 23). 13 The studies on the Syriac reception of Greek works and their transmission into Arabic are extensive. Among these, see in particular the already-cited works by Hugonnard-Roche (1989;and Troupeau (1991), who discusses how "Pour faciliter la tache des traducteurs du grec en syriaque et du syriaque en arabe, les savants syriaques constituerent de veritables lexiques, sur le modele des recueils de mots « ambigus », qu'ils avaient composes pour les besoins de l'exegese et de la grammaire" (Troupeau 1991, 10). See also López-Farjeat (2019), who presents the role played by the Syriac tradition in the transmission of Greek philosophy into the Islamic context, emphasizing the role of Christian Syriac translators and discussing the impact of Aristotle's Categoriae. Finally, see also Brock (2007 Überall, wo es sich nicht um bloßes Aufzählen oder zufälliges Zusammenreihen handelt, sondern nach sachlichen oder logischen Gesichtspunkten eine Anordnung des Mannigfaltigen versucht wird, darf mit Wahrscheinlichkeit auf griechischen Einfluss geschlossen werden. ( de Boer 1901, 17) 2

Parts of Speech14
In general, the major division on which a classification of the parts of speech is founded regards the opposition between nouns and verbs, in their functions as subjects/topics and predicates: There must be something to talk about and something must be said about this subject of discourse… The subject of discourse is a noun… No language wholly fails to distinguish noun and verb, though in particular cases the nature of the distinction may be an elusive one. It is different with the other parts of speech. Not one of them is imperatively required for the life of language. (Sapir 1921, 21) A preliminary distinction is so to be made between these two substantial elements, whereas all others could well be perceived as not representing the fundamentals of the discourse, but rather as subcategories of the major two.15 When assessing the classification of an element as either a part of speech, or as a class within one of the categories, the criteria used to determine its status pertain to the function that such element may assume.16 Fundamentally, the identification of an element as a part of speech accounts for a number of essential characteristics (e.g., semantics) as well as of secondary characteristics 8 (e.g., morphology). The essential characteristics are, of course, of major importance, whereas the secondary features are considered as collateral. So, for instance, if two elements differ only in the way they are inflected but other than that formally correspond to each other, such difference in the inflection will not be enough so as to consider the two as distinct parts of speech. Taking Arabic as an example, common nouns and proper nouns would be in the same category because they would differ, essentially, in the inflection, which would be either triptotic or diptotic. From this, a principle of syntactic interchangeability emerges as an underlying condition, so that something that may replace a word is identical with that word, and therefore belongs to the same category.

2.1
Greek In the Greek tradition, grammar would be treated as a branch of philosophy. The correctness of language usage would be codified in a number of rules, and therefore grammar would essentially represent a set of regulations aiming to prescriptively indicate how to shape the discourse.
[424e] In just this way we, too, shall apply letters to things, using one letter for one thing, when that seems to be required, or many letters together, forming syllables, as they are called, and in turn combining syllables, [425a] and by their combination forming nouns and verbs. And from nouns and verbs again we shall finally construct something great and fair and complete. (

Arabic
In the Arabic linguistic tradition, the first formulation on the parts of speech is in Sībawayhi's Kitāb, with the well-known formulation: fa-l-kalim: ism wa-fiʿl wa-ḥarf ǧāʾa li-maʿnā laysa bi-sm wa-lā-fiʿl Words are noun, verb, or particle with a meaning that is neither that of the noun nor of the verb (Kitāb, I: 1) The basic division, thus, consists of three major elements, ism, fiʿl, and ḥarf, 'noun' , 'verb' , and 'particle' , respectively: The most fundamental classification of words recognized by the medieval Arab grammarians is the well-known three-fold classification into noun (ism), verb (fiʿl) and particle (ḥarf). Said to have been first enunciated in Islam by none other than the Caliph ʿAlī, this classification has remained to the present day a leading principle of Arabic grammar and is regularly set forth in the opening pages of grammar books still used in the Muslim East. Even the Wright-Caspari grammar, taken still to be the classic Arabic grammar in English, treats the three classes of words as «parts of speech», subsuming the parts of speech traditionally recognized in English grammars under them. (Weiss 1976 , 23)25 Weiss, when presenting the "most fundamental classification of words", rightly mentions the type of subdivision that was firstly advanced by Sībawayhi and then -almost -unanimously ratified and perpetuated, and that is anterior to the introduction of logic into the Arabic tradition.26 Later, and although any of these has ever been fully endorsed, there have been few attempts to propose formulations different from that in the Kitāb. So for instance we find a fourth element, ḫālifa, in al-Suyūṭī (d. 911/1505),27 and al-Fārābī, too, mentions the elements ḫawālif.28 Notwithstanding, in his "purely 25 Weiss presents the parts of the speech investigating them as "A study in "ʿilm al-waḍʿ", which "may be translated as the founding of language, [and] represents a concept that is central to classical Muslim scholarly thinking about language" (Weiss 2009, 684). 26 "Die Auffassung, als gingen bei den Arabern die Anfänge der Logik denen der Grammatik voraus ist ein Hysteronproteron, das die Folge der historischen Tatsachen auf den Kopf stellt" (Weiß 1910, 386). 27 ʾAšbāh (iii: 2). 28 Zimmerman (cf. 1991) translates ḫawālif (sg. ḫālifa) as 'substitute' , used by al-Fārābī as the equivalent of 'pronoun' . In addition, it should be noticed that "several philosophical sources classification of the parts of speech in arabic 12 Aristotelian writings",29 al-Fārābī follows the tripartite division (cf. Versteegh 1995, 41, fn. 29). Another example of this is in Aḥmad Makkī al-Anṣārī who, investigating the role of al-Farrāʾ (d. 207/822) in the Arabic linguistic tradition as the leading figure of the grammatical school of Kūfa, further states that al-Farrāʾ was the first to propose a fourth lexical category between the noun and the verb. Due to some inconsistencies in the use, as well as to the diverse propositions advanced, Talmon set forth the thesis that the parts of speech were hence an unstable system (Talmon 1997, 280). In his argument, he mostly referred to the case of the ism, considering the "semi-independent status" of the naʿt in the Kitāb al-ʿAyn, "in its frequent contrast to ism", as an indication that early grammarians may have been inspired by an older division into eight parts of speech.30 Talmon notes that this "is also observable in the Kitāb, but in a smaller degree" (Talmon 1997, 280).
In addition to Talmon's viewpoint on this, Suleiman's assertions on the matter are also rather representative: Sībawaihi relies on a mixture of elements in the delimitation of the three word-classes he establishes in Arabic. Thus, he invokes what is essentially a notional or conceptual consideration in delimiting the 'particle' . Furthermore, this is carried out in a manner which suggests that both the 'noun' and the 'verb' are susceptible to notional delimitation, as is evident from the 'negative mode' adopted by Sībawaihi in delimiting the particle, although no such delimitation is given in the Kitāb. In delimiting the verb, Sībawaihi also advances what is essentially a morphological, or morphosyntactic, consideration, as is clear from his reference to the states of the verb and the types of conjugations subsumed under it. (Suleiman 1990, 248) demonstrate that there was a general awareness of the connection between the grammatical term ḥarf (together with ism and fiʿl) and the philosophical term ribāṭt (together with ism and kalima). Such a connection is established, for instance, by al-Fārābī (Šarh 54.8-9), by al-Ḫwārizmī  and by Ǧābir ibn Ḥayyān (see Kraus 1942:11, 250)" (Versteegh 1995, 36, fn. 16). 29 al-Fārābī also mentions a direct reference to Greek scholars: "min ʾahl al-lisān al-yūnānī" 42). 30 See the classification put forth by Dionysius Thrax mentioned above. olivieri Nouns In Arabic, the classes belonging to the category of nouns include, inter alia: adjectives (ṣifa, naʿt), plus assimilate and comparative adjectives (ṣifa mušabbaha and ʾafʿal al-tafḍīl); adverbs (ẓarf); pronouns (ḍamīr).31 As pointed out by Kasher, this classification by Sībawayhi's presents a few criticalities in its formulation: The first topic Sībawayhī treats in his al-Kitāb is the division of Arabic words (or morphemes) into three parts of speech (Sībawayhī, al-Kitāb I: 1). His classification, albeit (as far as we know) unanimously endorsed by his followers, raises many theoretical problems, not least of which is the abundance of cases in al-Kitāb, as well as in later grammatical treatises, where the phraseology implies that the classes referred to by the terms ṣifa and ism form a mutually exclusive pair, despite the fact that ṣifa refers (in the contexts in question) to a subclass of the part of speech ism. (Kasher 2009, 460)32 The criticalities addressed by Kasher are utterly well founded, especially considering that Sībawayhi disregards any explanations of why the parts of speech in Arabic would be three, and why those classes would be considered as part of said category. Moreover, in light of the functions and usage of some of the classes of the part of speech ism, these may easily be classified as additional parts of speech for themselves.
However, the division surely did not seem to be problematic for Arabic grammarians, who do not accentuate any type of confusion that the traditional classification may bring:33 31 Baalbaki presents the matter in relation to how subdivisions of the same part of the speech "are also assigned as ʿawāmil", including "a. ism fāʿil (active participle), as in hād̠ ā ḍāribun Zaydan ġadan; b. ṣifa mušabbaha (assimilate adjective), since this is said to be assimilated to ism fāʿil (cf. al-ḥasanu l-waǧhi and al-ḍāribu l-raǧuli); c. ism mafʿūl (passive participle), as in a-Zaydan anta mukābarun ʿalayhi; d. afʿal al-tafḍīl (comparative afʿal), as in huwa aḥsanu minka waǧhan; e. nisba (gentilic adjective; referred to by Sībawayhi as iḍāfa), as in a-Qurašiyyun qawmuka; f. maṣdar (verbal noun), as in ʿaǧibtu min ḍarbin Zaydan Bakrun; g. ism fiʿl (lit. proper name of the verb), as in ruwayda Zaydan and ḥayyahala l-ta̠ rīda; and h. ʿadad (number), as in ʿišrīna dirhaman which Sībawayhi likens to ḍāribīna ʿAbdallāhi" (Baalbaki 2008, 89-90). 32 As mentioned by Kasher, "Such a case was noted already by Diem (1974: 320), although the way he resolves it differs from the one which will be proposed below." See on this Diem (1974). 33 The fact that the tripartite division would not be challenged is also set up as a rather straightforward consequence of a common agreement. As the formulation and transmission classification of the parts of speech in arabic 14 The term ʾism, for instance, once introduced by Sībawayhi as a grammatical term denoting a noun, was used as such throughout, causing little confusion-It is rare for any ambiguity between 'noun' and 'name' to arise. Furthermore, the grammarians introduced terms denoting sub-categories of ʾism, such as shayʾ, shakhṣ, juththah and ʿayn, broadly denoting an object, and ḥadath denoting an event. Apart from the latter term, each of these expressions is normally used in medieval metalinguistic discussions as a grammatical term. (Peled 1999, 53) After introducing the parts of speech, Sībawayhi presents the category of nouns by introducing the linguistic arguments by means of examples (tamṯīl), instead of addressing their grammatical functions:

fa-l-ism: raǧul, wa-faras, wa-ḥāʾiṭ
And the noun is 'man' , and 'horse' , and 'wall' (Kitāb, I: 1) Although the Greek definitions discuss more in detail grammatical functions of the elements at issue, rather than relying on examples, formulations by way of examples and similar to that of Sībawayhi are available in Greek works as well.34 Thus, we may encounter definitions such as that by Plato: τί δ᾽ εἰ βουληθεῖμεν εἰδέναι, μὴ μόνον ποῖοι ἄνθρωποί εἰσιν ἢ ποῖοι ἵπποι, ἀλλὰ καὶ τίνες αὐτῶν δρομικοί τε καὶ μή, ἆρ᾽ ἔτι οἱ πολλοὶ τοῦτο ἱκανοὶ διδάξαι; But what if we wished to know not only what men were like or what horses were like, but which of them were good runners or not? Would the many still suffice to teach us this? (Alcibiades, 1, 111d) of the categories stem from a reached consensus, ʾiǧmāʿ, "one of the most frequently cited criteria for the correctness of a linguistic expression or even a grammarian's reasoning and as such it is related to the use of unanimity as an argument by the theologians." Versteegh further continues arguing that "In grammatical discussions we find, for instance, the unanimity of the grammarians (pp. 52, 79) or the unanimity of the Kufans and the Basrans (p. 202 ʾigmāʿ al-Kufiyyīn wa-l-Basriyyīn). In Ibn al-ʾAnbārī's analysis of linguistic criteria we find the statement that unanimity is a decisive argument (huǧǧa qāṭiʿa, Lumaʿ 44.9, 47.2)" (Versteegh 1995, 30, fn. 2). 34 Although the arguments we are advancing here with regard to the classification of the parts of speech and the definition of the category of nouns are in line with the earliest Greek formulations on the matter, namely as per in Plato and Aristotle, we here wish to remark that the Greek tradition has then mostly followed the formulations by Dionysius Thrax. However, as in our discussion we are trying to stress possible connections, we will mostly refer to the earlier works that we believe might have had an influence on the Arabic tradition, and not to the later standard doctrine. olivieri 15 Or that by Aristotle: Οὐσία δέ ἐστιν ἡ κυριώτατά τε καὶ πρώτως καὶ μάλιστα λεγομένη, ἣ μήτε καθ' ὑποκειμένου τινὸς λέγεται μήτε ἐν ὑποκειμένῳ τινί ἐστιν, οἷον ὁ τὶς ἄνθρωπος ἢ ὁ τὶς ἵππος. Substance, in the truest and primary and most definite sense of the word, is that which is neither predicable of a subject nor present in a subject; for instance, the individual man or horse. (Categoriae, 2a 14) Both the excerpts quoted above show some similarities with Sībawayhi's formulation, also with regard to the examples used, man and horse.35 However, what we find in the Greek instances, and that is definitely lacking in the Arabic, is the elucidation of the grammatical characteristics of the noun as a part of speech. In fact, Sībawayhi does not provide a real definition. Thus, to investigate the types of arguments devised for the purpose, we need to move further and examine later works. A first formulation is in the al-Muqtaḍab by al-Mubarrad (d. 286/900), who asserts that: A noun is everything that can be preceded by a preposition, and if it is not possible then it is not a noun. I: 141) Later, in the al-Ṣāhibī by Ibn Fāris (d. 395/1004), we find a report of earlier arguments,36 presented as follows: As for the noun, Sībawayhi says: "The noun is for instance man and horse" […] and states Abū l-ʿAbbās Muḥammad b. Zayd al-Mubarrad that according to Sībawayhi's view: "The noun is what can be a subject," […] and states al-Kisāʾī: "The noun is what an attribute can be referred to," and says al-Farrāʾ: "The noun is what can exhibit a tanwīn, and be in construct state or annexed to the definite article," and says al-Aḫfaš: "You know that 35 Further resemblances are observed by Versteegh who states that "in Sibawayhi's Kitab the primary examples for the noun are raǧul "man" and faras "horse", which correspond with the well-known examples of the Stoic school tradition anthropos and hippos. In both traditions the verb "to hit" (túpto/ḍaraba) is the most popular example for the category of the verbs" (Versteegh 1993, 23 you are dealing with a noun when a verb or an attribute can be referred to it, as for instance in zayd qāma (Zayd stood) or in zayd qāʾim (Zayd is standing), when can be in the dual form or take the plural, as al-zaydāni (the two Zayd(s)) and al-zaydūna (the Zayd(s)), and when exhibits a triptotic inflection. (al-Ṣāhibī, 49) Finally, of all exemplifying definitions, the one by al-Zaǧǧāǧ (d. 311/923) seems to be of particular interest, for it shows a number of more evident resemblances with the Greek, so in that: [The noun is] an articulated and comprehensible sound that expresses a meaning but has no implications of time and space. quoted by Ibn Fāris,51) may well be put in correlation with: A noun is a composite sound with a meaning, not indicative of time, no part of which has a meaning by itself; for in compounds we do not use each part as having a meaning of its own (Poetica, 1457a).37

Discussion
In line with these general approaches, if a preliminary distinction is to be made between verbs and nouns, the two fundamental elements,38 then all others may be associated with each other insomuch as they partake with the concept of being the expression of what lays below.39 It then becomes customary to group the noun and the adjective together because all elements that share characteristics that do not belong to that of the verb, would form a category for 37 On this, Farina notes that in ‛ Amīra's Grammatica Syriaca (16th cent.) "[…] three definitions of the noun attributed to different Syriac grammarians (Elias of Nisibis, John the Stylite and the «Grammar in Arabic language») are quoted, both in their Syriac form and in Latin translation, and then compared. The latter one, that ‛ Amīra considers as the most complete and the clearest, is then discussed and commented upon in detail. Such a definition, in ‛ Amīra's Latin translation, reads: «Nomen est vox significative in perfectione, quae sine tempore est, qua pars cum separatur, non significat aliquid»" (Farina In press,20 themselves as opposed to the verbs.40 Therefore, anything that points at the meaning of its nominatum may be considered as a noun.41 In Greek grammatical literature the ónoma may indicate any word that refers to the nominatum, 'the named one' . Similarly, al-Mubbarad's approach to the matter, distinctly in line with Sībawayhi's proposition, also considers the ʾasmāʾ as elements to be used for all words that may be regarded as the 'name' of its meaning (musammā, 'nominatum'):42 The noun refers to the nominatum below […] And since the noun is raised above the nominatum and stands above its meaning, this means that it is derived from simuww and not from wasm. 2) Similar oppositions based on the meaning conveyed may be found in Aristotle's De Interpretatione, which focuses on the concept that the meaning of the verb is the 'action' , and when being conjugated, the tenses add a time axis to it: αὐτὰ μὲν οὖν καθ' αὑτὰ λεγόμενα τὰ ῥήματα ὀνόματά ἐστι καὶ σημαίνει τι,ἵστησι γὰρ ὁ λέγων τὴν διάνοιαν, καὶ ὁ ἀκούσας ἠρέμησεν, -ἀλλ' εἰ ἔστιν ἢ μή οὔπω σημαίνει· οὐ γὰρ τὸ εἶναι ἢ μὴ εἶναι σημεῖόν ἐστι τοῦ πράγματος, οὐδ' ἐὰν τὸ ὂν εἴπῃς ψιλόν. αὐτὸ μὲν γὰρ οὐδέν ἐστιν, προσσημαίνει δὲ σύνθεσίν τινα, ἣν ἄνευ τῶν συγκειμένων οὐκ ἔστι νοῆσαι. When uttered just by itself a verb is a name and signifies something--the speaker arrests his thought and the hearer pauses--but it does not yet signify whether it is or not [the meaning]. For not even 'to be' or 'not to be' 40 Cf. Lyons (1966), who argues that "from Alexandrian times it has been customary to group the noun and the adjective together (the distinction of 'nomen substantivum' and 'nomen adjectivum' as separate 'parts of speech' being a medieval development" […]) and to distinguish them sharply from the verb" (Lyons 1966, 216). 41 According to this, the ḥarf ('particle') could not be included in this category, as Sībawayhi's statement that the ḥarf has a meaning that is neither that of the noun nor of the verb (Kitāb I:1) is interpreted so that particles convey a meaning only when in combination with other words, and therefore do not have one themselves. For further discussions on this, also in relation with the other parts of speech, see Versteegh (1995, ch. iii, especially fn. 35 is a sign of the actual thing (nor if you say simply 'that which is'); for by itself it is nothing, but it additionally signifies some combination, which cannot be thought of without the components. (De Interpretatione,16b,(19)(20)(21)(22)(23)(24)(25) Thus, if any elements that points to a nominatum may be a noun, all the classes that we have mentioned above would straightforwardly be included in the category of noun as a logical conclusion, in both the Arabic and the Greek traditions. This correlation noun-nominatum, then, would be the principle for considering an element as part of said category. Similarly, but dealing with the matter from another angle, operates the principle of syntactic interchangeability that we have briefly mentioned above (cf. §2). Based on this, elements cannot be considered as separate parts of speech if they may replace a word and remain identical with that. A clear example of this is that of pronouns. So, taking again Arabic as an example, this is the case of the ism muḍmar, ('hidden'), which as a personal pronoun denotes the noun it refers to; or of the ism mubham, ('ambiguous' , because it is not clear what it refers to), which identifies the demonstrative pronouns. From a semantic viewpoint, finally, the ism partakes of a number of similarities with the ónoma, at least as per its Aristotelian formulation. In both traditions, nouns are elements that possess an independent meaning, but do not provide references to time, differently from verbs that instead exhibit temporal (namely tense) features. Furthermore, in the Poetica we do not have those lists of nominal classes that we may instead find in the Technē Grammatikē.
Dionysius Thrax' linguistic formulations seem in fact to have been neglected altogether by Arabic grammarians, if they ever knew about these at all. Furthermore, the Technē Grammatikē "with its paradigms and system of verbal forms was completely foreign to the spirit of a Semitic language, and could in no way appeal to an Arab or Hebrew grammarian" (Fischer 1962, 2). The Aristotelian formulations, if we can go as far as to assume an influence, seem instead to be well received in the Arabic tradition.
If any connection with regard to the classification of the parts of speech is to be assumed, it would well rather be with the outline and grammatical compendium of the state of Greek language studies as presented in Aristotle's Poetica at the end of the fourth century bce, as "whatever variations of the text may have confronted the Syriac and Arab translators, they considered it the final Aristotelian formulation, without any concern about the later, post-Aristotelian, developments in Greek grammatical science" (Fischer 1963, 133).
Surely, any of these arguments or similarities alone may ever prove a direct dependence of Arabic grammar from the Greek tradition, which is not our olivieri intent. However, the resemblances that the approaches show may hint at a contact with the Greek tradition. Similarities between the doctrines, taking as an example the earliest subdivision of the parts of speech advanced by Sībawayhi and the corresponding Aristotelian formulation in the Greek tradition, in fact, could never be explained by influence through translations of Greek writings, due to historical reasons.
However, we believe that albeit these arguments do not prove the influence of Greek logic which would instead come at a later stage, they might point at a possible contact with the grammatical outline set forth by Aristotle.
Sībawayhi's premise with regard to the parts of speech, in fact, is largely relying on formulations that seem to be Greek-inspired, the type of griechischen Geistes suggested by de Boer (cf. 1901, 17, quoted above) and that Fleisch remarks as follows: Des influences grecques sont à signaler: la spéculation grammaticale arabe a emprunté des concepts initiaux à la science grecque, non pas à la grammaire grecque, mais à la logique aristotélicienne. (Fleisch 1961, 23) Furthermore, Fischer's studies in the 1960s, and that we have quoted at length in this paper, have also advanced the idea that Aristotle's Poetica might have been more widely available in the "eastern part of the Mediterranean", and that its transmission and circulation might have been overlooked: This lack of appreciation of the role played by the Syro-Arabic transmission of the Poetics was caused by several factors: i) In the Syro-Arabic transmission the Poetics often does not appear as a separate, independent treatise, but forms a part of the Organon and is joined to the Categories. It was thus placed, together with the Rhetoric, among the logical works of Aristotle. 2) Despite the fruitful translational activities of the Arabs, they were denounced as completely unfamiliar with Greek life and literature, and they were charged with all kinds of mis-conceptions, with the result that the value of Arabic interpretations outside the fields of philosophy and sciences was greatly reduced. This was especially true with regard to the literary and esthetic theme of the Poetics, which because of the obscurities of the text and stylistic difficulties presented additional linguistic problems. 3) The only attempt to confront the Arabic grammatical concepts with earlier Greek formulas was made on the basis of Dionysius Thrax' Technē Grammatikē which was used widely by the Syrian writers and is so well attested in Syriac grammatical science. (Fischer 1963, 4-5) classification of the parts of speech in arabic 20 It is rather evident that the possible points of convergence that may be found across Arabic and Greek works and that we have referred to so to support this case are circumstantial evidences, to say the least. Doubt may be cast on whether these same convergence points could be even read in a way that disqualifies the possibility at all. However, the possibility that Sībawayhi, or scholars either coeval or earlier, may have got in contact with the Greek tradition, is a suggestion that cannot be ruled out completely. This would ultimately point at a possible conceptual influence, rather than at a historical source; however, it is not illegitimate to ask whether Sībawayhi's theorizations on this fundamental issue of grammar may have been inspired by a grammatical framework external to the 'purest' Arabic tradition.