The ism in the Arabic Grammatical Tradition : Reflections on Its Origin and Meanings

This article aims to present an overall reconstruction of the debate on the definition and etymology of Arabic ism (‘noun’, ‘substantive’), by discussing and comparing texts from the Arabic linguistic tradition. The first part deals with the definition of the grammatical element and its functions, while the second is fully dedicated to the examination of the etymological issues, focusing on the two assumed roots of derivation for ism (namely s-m-w and w-s-m). The arguments are presented through the collation of the opinions of the relevant Arabic grammarians, examining both the wider debate between the early grammatical schools of Baṣra and Kūfa (2/8–3/9 centuries), and the reports of the arguments as described by later scholars.


The parts of speech
It is common agreement, within the Arabic linguistic tradition, that language consists of three parts of speech (ʾaqsām al-kalām), namely nouns, verbs, and particles.The delineation of the tripartite vision is usually ascribed to Sībawayhi (d.180/796) who opens the Kitāb by defining the partes orationis: "words are noun, verb, and particle." 1 Each category presents differences in status and characteristics, as pointed out also by later grammarians, with distinctions mainly based on the role played within an utterance.Among the three, the noun (ism, pl.ʾasmāʾ) meets those criteria that make it a superior category, being an element "able to both operate as and receive a predicate" (mā yuḫbaru bihī wa-yuḫbaru ʿanhu) as in 'Muḥammad is our Prophet.'In this example, nouns function as both predicate and predicator, acting as mubtadaʾ and ḫabar in the noun clause.
In addition to this, major distinctions rely on the fact that nouns-as well as verbshave a well-defined morphology and range of meanings, while particles are meaningful words in themselves but have no strict forms.2

Definition
Sībawayhi does not provide clear definitions of the ism in the Kitāb, but-relying on a common linguistic practice-reduces his explanations to a tamṯīl, 3 providing few examples of what may be considered a 'noun', as for instance 'man', 'horse ', and 'wall' (fa'l-ism  raǧul wa-faras wa-ḥāʾiṭ). 4his represents the starting point from which grammarians of later periods drew inspiration to formulate their own definitions, examining the ism according to its intrinsic features and grammatical peculiarities.
The formulation of linguistic theories is framed within the conventional grammarians debate as part of the ʾiǧmāʿ ('consensus') tradition.Being a fundamental element of the whole Arabic culture and usually widely exercised in juridical discussions,5 the ʾiǧmāʿ plays a key role also in grammatical disputes, 6 where unanimity is the main criterion to state the correctness of an argument.Ibn Ǧinnī (d.392/1002) in the Ḫaṣāʾiṣ7 -and with regard to a strictly linguistic framework-calls it ʾiǧmāʿ ʾahl al-baladayn8 and by doing so he circumscribes the practice to the agreement among the Baṣran and the Kūfan grammarians.
However, as in other fields, the explicit agreement is not the only possibility to determine a concurrence of ideas, which may also be reached with either an implicit agreement or lack of explicit disagreement.9This might be the case of the tripartite division of the parts of speech that has never been challenged after being stated in the first place.As a matter of fact, the lack of a clear definition in the Kitāb left room to a profound discussion on the subject, so that grammarians after Sībawayhi could define the nature of the category and focus on providing further details to delineate the characteristics of the noun.

What is a 'noun'?
Lane's Lexicon describes the ism as: 1.1).The impossibility to clearly define what the grammarian meant leaves room for further and opposite interpretations.Hence, some later grammarians interpreted the element as something that-not having a meaning of its own-needs to be in combination with either a verb or a noun, while others acknowledged an inherent meaning.
[…][The name of a thing; i.e.] a sign [such as maybe uttered or written] conveying knowledge of a thing […] and a word applied to denote a substance or an accident or attribute, for the purpose of distinction […] [or a substantive in the proper sense of this term, i.e. a real substantive; and a substance in a tropical sense of this term, i.e. an ideal substantive […] is that which denotes a meaning in itself unconnected with any of the three times [past and present and future]: the pl. is As a consequence of the stigmatization of the approach, grammarians were forced to frame several elements within a rather strict scheme and find a way to make their subdivisions fit the rules set for the tripartition.
Thus, elements differing from each other have been classified under the label 'noun', along with a set of sub-categories to indicate less obvious cases.The elements acknowledged as fully matching the definition of ism are: 19 the common noun (ism al-ǧins), the proper noun (ism al-ʿalam), the concrete noun (ism al-ʿayn), the abstract noun (ism al-maʿnà), and the active and passive participles (ism al-fāʿil and ism al-mafʿūl).Likewise, are associated to the same category the adjectives (ṣifa), the infinitives (maṣdar), the nouns referring to space and time (ism al-zamān wa'l-makān), the pronouns (muḍmar), and the numerals (ism al-ʿadad). 20ummaries of the debate on what may or may not be a noun are available in works such as Ibn Fāris' (d.395/1004) al-Ṣāḥibī: As for the noun, Sībawayhi says: "The noun is for instance man and horse" [...], and Abū l-ʿAbbās Muḥammad b.Zayd al-Mubarrad states that, according to Sībawayhi's view, "The noun is what can be a subject," [...] and al-Kisāʾī states: "The noun is what an attribute can be referred to;" al-Farrāʾ says: "The noun is what can exhibit a tanwīn, be in construct state or annexed to the definite article;" al-Aḫfaš says: "You know that 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 it can be in the dual form or take the plural, as al-zaydāni (the two Zayds) and al-zaydūna (the Zayds), and when it exhibits a triptotic inflection.
19 See also WEHR and COWAN 1994, s.v.‫.اﺳﻢ‬ Yet, oftenthese reports are rather partial, as in the case of al-Mubarrad (d.286/900) who, in the al-Muqtaḍab, declares a longer version than the one reported above, and also partly different.After defining the noun as ism mutamakkin, 22 the grammarian proceeds by stating that: A noun is everything that can be preceded by a preposition, and if it is not possible then it is not a noun.(kullu mā daḫala ʿalayhi ḥarf min ḥurūf al-ǧarr fa-huwa ism, wa-ʾin-i 'mtanaʿa ḏālika fa-laysa bi-'sm) 23 Another significant definition is provided by al-Zaǧǧāǧ (d.311/923) and quoted by Ibn Fāris: [The noun is] an articulated and comprehensible sound that expresses 24 a meaning but has no implications 25 of time and space (ṣawt muqaṭṭaʿ mafhūm dāll ʿalà maʿnan ġayr dāll ʿalà zamān wa-lā makān). 26re the argument evolves from logic 27 and concludes with a definition of the noun clearly influenced by the Greek tradition. 28The importance of his contribution lies in the different approach to the topic, more oriented towards the concept of the meaning of the noun.28 As already examined by FLEISCH, it is clearly recognizable here the influence of Greek logic in the way the definition is articulated.The effects of the influence are so massive that we can also state that there is an almost sharp division between grammarians pre-and post al-Zaǧǧāǧ.Cf.FLEISCH, s.v."Ism", EI 2 , iv: 181-82.

Ism: the discussion on etymology
The debate is presented through the examination of two hypothesized roots, namely sm-w and w-s-m, 32 whose inherent meanings should be intended as practical explanations of the term and refer to the function attributed by the Arabic grammarians to the ism: The Kūfan grammarians argue that the ism derives from wasm, which indicates the 'mark', while the Baṣran grammarians argue that it derives from simuww, 33 which indicates the 'elevation' 34 (ḏahaba 'l-kūfiyyūn ʾilà ʾanna 'l-ism muštaqq min-a 'lwasm wa-huwa 'l-ʿalāma wa-ḏahaba 'l-baṣriyyūn ʾilà ʾannahū muštaqq min-a 'ssimuww wa-huwa 'l-ʿuluww). 35cording to these propositions, simuww-meaning 'elevation', 'height'-is coincident with the function of a ism identified with the signifier, whose role is to stay on a higher level than the signified below (al-musammayāt taḥtahā).Conversely, Kūfan grammarians analyze the ism as derived from wasm (used as a synonym of ʿalāma, 'mark'), and sima 'sign'. 36he Kūfan grammatical school presents a reasoning related to the markedness theory: 37 [ism] derives from wasm because in the language [of the Arabs] 38 it indicates the mark, and a noun is a definition mark distinct from the symbol that identifies it [...].
For this reason, we assert that ism derives from wasm, and the same has been stated by Ṯaʿlab who argues that the noun indicates the sign that has been established 39 for 31 ʾInṣāf, 1-6.
32 The passages quoted below shall show that Arabic grammarians do not refer to roots when discussing the process of derivation (ištiqāq), but rather refer to words.34 He is not mentioned in the text, but the latter proposition seems to be shared also by al-ZAǦǦĀǦ.(Cf.al-Ṣāḥibī: 57).
37 Mark is used here to recall the markedness theory, where the marks are grammatical elements (gender, number and case) and not lexical.According to this theory, the noun is not a mark.But, for the purpose of the translation, the word has been adopted under a generic-and not technical-profile.
38 The kalām al-ʿarab is usually one of the main sources for grammatical observations.The reference is to the Arabic variety spoken by the Bedouins of the Arabic peninsula.The Arabs, whose unconscious knowledge of the language prescription is a manifest concept for every grammarian, are presented by early scholars as having an innate wisdom (ḥikma) which makes them choose the correct forms without being really aware of the grammatical reasons.

Grammatical observations on etymology
Despite the evident logic behind the reasoning, it is evident that the opposition between sm-w and w-s-m is merely speculative and part of a rather theoretical discussion.
The grammatical interpretation posing in contrast the two elements may be or may not be acceptable as a reasoning,44 but for the etymology there are many proofs showing that the Kūfan interpretation is erroneous.Medieval Arabic grammarians pointed out the about language.Language in that thinking was entirely invented.That is to say, it owed its existence to a process of deliberate assignment of patterned vocal utterances-or components of such utterances-to meanings, of ʾalfa ̅ ẓ (→lafẓ) to maʿānī (→maʿnà)".WEISS, s.v."Waḍʿ al-Luġa", EALL, iv: 684.uniqueness of the root through a grammar-oriented reasoning, as much as modern comparative studies do. 45aṣrans' confutation of Kūfan thesis is articulated in five passages,46 corresponding to the five modalities of mistake observed in their propositions.These arguments, four of which shall be examined in detail in the next paragraphs,47 conclude by demonstrating that despite the logic behind the Kūfans' grammatical reasoning, there is one only possible root ism may be ascribed to: s-m-w.48

The form of the term
The first analysis reported in the Kitāb al-ʾInṣāf refers to the form of the term ism and its phonetic realization, which results from the drop of the weak radical in the noun formation.
The Kūfans state that the wāw occurring in the first position is dropped and replaced by a hamza 49 which functions as a letter of compensation (taʿwīḍ).If this phenomenon were productive, then the initial hamza would be a systematic result whenever a weak radical 45 The concept of root is a key element shared all throughout Semitic, and, as pointed out by VOIGT: "All Semitic languages have a verbal form and a nominal form (except for functional words and particles), characteristically consisting of a triradical root and a vocalic pattern which may also require the addition of further consonants" (EALL iv: 173-74).For the case study presented in this paper, other Semitic languages further substantiate the correctness of the arguments in favor of the root s-m-w, for they prove that a root, either biliteral or with a weak radical occurring in final position [s(š)-m or s(š)-mw/y], is attested all throughout Semitic.Hence, given the range of meanings: (1) 'noun', 'name', 'substantive' for Arabic ism, and (2) 'sky', 'heaven(s)' for Arabic samāʾ, their cognates in other Semitic languages are, for example: Akk. ( 1  HOFTIJZER et al. 1995: 1265).
occurred in the first position.Hence, in cases such as waʿada the expected realization would be *iʿd (waʿada-*iʿd as in the pair wasama-ism).
Yet, the resulting terms reflect a different state of affairs, as the outputs of this type of roots would consist of terms ending with a tāʾ marbūṭa (in fact the attested form derived from waʿada is ʿida), as rightly stated by the Baṣrans.
Baṣrans consider the final tāʾ marbūṭa as a compensation letter indicating the drop of the first-weak-radical, while the initial hamza would result from the drop of the last radical.Thus, ism derives from an underlying form *simw, shaped on the pattern fiʿl where the first radical is vocalized in kasra and the second is quiescent.50According to the Baṣran proposition, in that instance the root would suffer from the drop of the third radical , and, resulting biliteral, would compensate the elision of the wāw by appending a hamza at the beginning of the word, on the paradigm ifʿ.

The past tense
The second Baṣran proposition is about verbal analysis and how verbs are formed out of this type of root, starting with the māḍī of the fourth form.The Baṣrans argue that the underlying form of the verb is *ʾasmawtu, but the wāw-which comes to be the third radical of the root and occurs here in the fourth position-undergoes a regular process of transformation, turning into a yāʾ,51 resulting then in the form ʾasmaytu.The proposed theoretical form, as well as the resulting one, shows that the weak radical does not occur in first position in the root, which otherwise would result in a fourth verbal form *ʾawsamtu.
The same phenomenon is also attested in the muḍāriʿ, where the vowel shift is usually very regular due to the vocalization pattern of the form (ʾafʿala-yufʿilu), as in yuʿlī, yudʿī, and yusmī, inferred from the underlying forms *yuʿliw, *yuʿdiw, and *yusmiw, and where a quiescent wāw is preceded by a letter vocalized in kasra.The phenomenon is very regular and is recorded whenever a quiescent wāw comes to occur in a position adjacent to a consonant vocalized in kasra, as for instance in mīqāt, mīʿād, and mīzān, whose underlying forms would be *miwqāt, *miwʿād, and *miwzān, inferred from al-waqt, al-waʿd, and alwazn.

The diminutive
Within the debate, also the formation of the diminutive contributes to the definition of the root.Considering the root proposed by the Kūfans, w-s-m, the diminutive form of ism would result in the form *wusaym, where the first weak radical is retained.Yet, the variant attested for ism is sumayy.
The Arabic lexicon does not record a form containing a weak radical occurring at the beginning of the word, whereas it records a form with a last weak radical.
The word is ascribable to an underlying form *sumayw, on the paradigm fuʿayl, but having a wāw as its last radical it shifts into a yāʾ: when a yāʾ and a wāw occur together and the first letter is quiescent, then the wāw turns into yāʾ, resulting in a geminated form.This paradigm is widely attested and productive, as demonstrated by common terms such as sayyid, ǧayyid, and mayyit, whose underlying forms would be *saywid, *ǧaywid, and *maywit.yāʾ being easier to articulate , 52 a shift of the wāw onto a yāʾ is favored.

The plural
The last Baṣran proposition concerns to the formation of the plural form of ism, for which two major forms are registered: ʾasmāʾ and ʾasāmī.Both are derived from the root s-m-w and cannot be ascribed to w-s-m, from which the resulting forms would be *ʾawsām and *ʾawāsīm.
Of the two possible plural forms mentioned above ʾasmāʾ is more correct because of its major attestation and plausibility.Its underlying form is *ʾasmāw, which records a wāw occurring in the final position, and preceded by a ʾalif zāʾida.In such cases, the wāw usually shifts into a hamza, as recorded in other examples such as samāʾ, kasāʾ, and raǧāʾ whose underlying forms would be *samāw, *kisāw, and *raǧāw.
In addition to this, another proposition argues that the wāw undergoes a process of shift into ʾalif, acknowledging the existence of an abstract form where the fatḥa occurring before the ʾalif is triggered by a vocalized wāw.The latter, when occurring in a vocalized form and preceded by a letter carrying a vocalization in fatḥa, must necessarily turn into a ʾalif.
The conclusion would then be the combination of two ʾalif, one added to the root and one deriving from the last radical.But being both quiescent they do not merge; therefore, the second ʾalif turns necessarily into a hamza, both hamza and ʾalif being two hawāʾiyya 53 letters.

Conclusions
The aim of this contribution was to present an overall view on how Arabic grammarians dealt with the issue of defining the root from which the Arabic term ism should be derived.It was inferred from the inventory of arguments discussed in early works and here presented through the words of a 12 th -century scholar, that ism may be derived from two distinct roots: w-s-m and s-m-w.
The history of the Arabic linguistic tradition shows that the speculative activities of grammarians belonging to the Baghdādi tradition, like al-Anbārī, often reiterate previously discussed arguments with a manifest orientation towards Baṣran ones.Nonetheless, it is undoubtedly important that in retracing the development of Arabic linguistic sciences the debate itself becomes as important as the linguistic issue discussed, because of the methodological approaches proposed.Al-Anbārī's way to present the debate is a kind of storytelling, where the Baṣran predominance is justified by the fact that the method of analysis they propose is better theorized and developed.Thus, strict methodological procedures and 52 ʾaḫaff, lit.'lighter'.
a sophisticated attitude to the reasoning show a clear superiority of the Baṣran group over its counterpart, as in the case study presented in this paper.
In fact, as discussed in paragraphs 3.2.1 through 3.2.4,it is evident how the discussion led by Kūfan grammarians-despite its logic-does not have very solid foundations, but rather presents a more speculative approach.Not considering the propositions on semantics, common to both groups and ascribable to a higher level of linguistic abstraction, Baṣrans' arguments on morphological phenomena are more well-grounded, and properly explain why the term ism cannot be derived from the root w-s-m, contrary to what Kūfans argue.Furthermore, cognates evidence that also in other Semitic languages the root is either biliteral or has a weak letter as its last radical, thus adducing an additional proof on a comparative level.© Simona Oliveri, University of Helsinki / Finland ◄ simolivieri@gmail.com► wissenschaft = Handbooks of linguistics and communication science = Manuels de linguistique et des sciences de communication / mitbegr.von Gerold UNGEHEUER.Hrsg.von Herbert Ernst WIEGAND; Bd. 36.Berlin: De Gruyter Mouton.al-ZAǦǦĀǦ.1971.Mā yanṣarif wa-mā lā yanṣarif.Edited by Hudā Maḥmūd QARĀʿA.Cairo: Maṭābiʿ al-Ahrām.ZIMMERMAN, Fritz W. 1981.Al-Farabi's Commentary and Short Treatise on Aristotle's "De interpretatione" (translation, introduction & notes).London: British Academy.
‫ﺎء‬ َ ْ ‫أﲰ‬ [a pl. of pauc.] and ‫ات‬ ‫ﺎو‬ َ ْ ‫أﲰ‬ the latter said by Lḥ to be a pl. of ‫ﻢ‬ ْ ‫ﺳ‬ ِ ‫ا‬ but it is rather a pl. of ‫ﺎء‬ َ ْ ‫,أﲰ‬ for otherwise there is no way of accounting for it […]. 10This definition mostly relies on those provided by Arabic grammarians, 11 for whom a noun is a word 12 which expresses a meaning but 'is neither connected with a time' (ġayr muqtarin bi-zamān) nor is 'time part of it' (wa-laysa az-zamān ǧuzʾan minhā).Besides, it may be defined as a word indicating something 'perceptible' (šayʾ maḥsūs) or 'nonperceptible' (ġayr maḥsūs) which is a means of rising into notice the thing denoted thereby, yet never referring to time.The discussion on what a 'noun' should be is rather extensive, 13 and grammarians themselves record the proposition of a quantity of different positions and statements. 14Part of the discussion is related to the additional features of the noun, which may: 15 Excerpts of the arguments and a summary of the theoretical development are reported in several grammatical works, such as al-Bāqillānī's (d.403/1013), al-ʿUkbarī's (d.616/1219) and al-Anbārī's.The latter presents the issue on the etymology of ism in two major works, ʾAsrār al-ʿArabiyya 30 and the Kitāb al-ʾInṣāf.31 3.1 The preliminary debateBeing the definition of the category controversial also in the eyes of the Arabic grammarians themselves, 29 semantics and etymology contribute to defining what ism is meant for.22According to the Kitāb, the category of the "mutamakkin" nouns refers to those elements that present peculiar features, may occur in various syntactic constrictions or have syntactic functions themselves, and may be inflected both in the determinate and in undetermined state.(Kitāb, ii:33; ii:40) 23 al-Muqtaḍab, i: 141.24 Lit.'it indicates'.25 Lit.'it does not indicate'.26 al-ṢĀḤIBĪ: 51.27 For an extensive and accurate study of the Greek influence on Arabic linguistic thinking, see Versteegh 1977.