Originality of the Semantic Approach in Arabic Linguistic Thought, with Particular Reference to Ibn al-Qaṭṭāʿ’s Work

In this study we investigate some aspects of the linguistic thought of Ibn al-Qaṭṭāʿ (d. 515/1121) with the intent of contributing to a better knowledge of this eminent personality of Arab Muslim Sicily. To this aim, we offer a description of the milieu of linguistic thought to which al-Qaṭṭāʿ belonged, with particular reference to some members of that milieu, who are known to modern scholars for efforts distinguished by theoretical and methodological originality. We also clarify some semantically-oriented original traits of Ibn alQaṭṭāʿ’s morphological analysis, as emerging from his treatise Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼlmaṣādir, as precisely such traits make it possible to number him among the infrequent bearers of semantic originality in the context of medieval Arabic linguistic thought.

tic analysis on the level of form rather than of meaning.5In all likelihood, the historical reason that lies behind this attitude is the obscurity that the Arab grammarians and lexicographers might have perceived in the variety of Arabic they wanted to describe, the so-called kalām al-ʿArab. 6It can be hypothesized, in fact, that in transmitting and investigating the kalām al-ʿArab, the Arab grammarians and lexicographers not so infrequently took great pains in understanding it, so they felt somehow forced to access it primarily through its form rather than its meaning, the former being clearer to them than the latter.It is telling in this respect that precisely the study of obscure words (ġarīb) was an important part of the Arab lexicographers' work since the very beginnings of Arabic linguistic thought,7 though further investigation is required to validate such a hypothesis. 8e that as it may, the fact that conservatism tends to tally with a formal approach in the methodology of Arabic linguistic thought implies that the rare traits of originality present take place on the level of meaning.This is illustrated by al-Ǧurǧānī's (d.471/1078) interpretation of the word-order pair ǧumla ismiyya-ǧumla fiʿliyya, e.g., al-zaydūna katabū / kataba al-zaydūna 'the Zayds, they wrote/the Zayds wrote'. 9While Arabic linguistic thought usually derives this syntactic pair from a formal opposition, which consists of the agreement, or lack thereof, between the verb and the noun, 10 al-Ǧurǧānī interprets it as the result of a semantic opposition, in which informational saliency affects either the utteranceinitial noun (i.e, al-zaydūna in al-zaydūna katabū) or the utterance-initial verb (i.e., kataba in kataba al-zaydūna). 118 Outside Arabic, it is well established among linguists that an epistemological connection exists between an obscure language and the resort to a formal approach to analyze it.Lepschy exemplifies this state of affairs by means of the formal approach that American structuralists developed to account for Amerindian languages, which effectively appeared rather puzzling to them (LEPSCHY 1966: 151-2).
11 al-ǦURǦĀNĪ, Dalāʾil al-ʾiʿǧāz: 147.Concretely, al-Ǧurǧānī exemplifies the semantic opposition between ǧumla ismiyya and ǧumla fiʿliyya by means of interrogative utterances (al-istifhām) such as ʾa-faʿalta, ʾa-ʾanta faʿalta.In these utterances, the informational saliency, which consists of the speaker's In the literature, few other examples of semantic traits of originality are seemingly reported, the most notable of which are those developed by al-Astarābāḏī (d.688/1289) 12 and Ibn Hišām (d.761/1359). 13By contrast, it seems that the semantic originality that Ibn al-Qaṭṭāʿ (d.515/1121) brought to Arabic linguistic thought has not yet received scholarly attention.In what follows, we first outline the main aspects of semantic originality of al-Astarābāḏī's and Ibn Hišām's linguistic thought in the form of a review of the literature, then proceed to clarify the contribution of Ibn al-Qaṭṭāʿ in the same respect.
Yet before proceeding further, a caveat is in order: ascertaining the pervasiveness of a formal approach in the conservative transmission of Arabic linguistic thought should not mislead us into oversimplification.In the transmission of such knowledge, the semantic dimension was marginal but not totally absent.Evidence for this assertion comes from the formative stages of Arabic linguistic thought: as Baalbaki points out, 14 Sībawayhi avails himself of "technical terms which refer to formal aspects" and which, at the same time, "have distinct semantic functions", although "[he] does not formulate a semantic theory in the Kitāb" for these terms, relegating them to a marginal role.We can draw an example from morphology to understand this point.In Sībawayhi's view, the construct of affixation (ziyāda) can but must not involve a semantic dimension, contrary to standard assumptions in modern Western linguistics.On the one hand, Sībawayhi explicitly states that affixation may "introduce an element of meaning" (tadḫulu li-maʿnan). 15On the other hand, he also asserts that this function is not quintessential to affixation, the other important function of it being that of ilḥāq, i.e., "reducing one [anomalous] pattern to another [more regular] pattern" (tulḥiqu bināʾan li-bināʾ) regardless of their meaning. 16For instance, the Arab grammarians regard the Quranic hydronym kawṯar as instantiating an unexpected consonant w, which disrupts the regular pattern faʿal, thus yielding the anomalous pattern fawʿal.They also propose to reconceptualize the unexpected consonant w as an affix that, in merely formal terms, occupies the position of a root consonant (ilḥāq), rather than introducing an element of meaning.This analysis allows them to re-interpret the anomalous pattern fawʿal as a regular quadriconsonantal pattern, which is effectively attested to in nouns such as ǧaʿfar. 17urthermore, the formal approach itself was not immune from sporadic traits of originality, in spite of the Arab grammarians' tendency to transmit it conservatively from one generation to the next.An indicative example is the conceptual organization of Arabic grammatical theory devised by Ibn al-Sarrāǧ (d.316/928), the original character of which doubt (šakk), affects either the utterance-initial verb faʿalta (fa-badaʾta bi'l-fiʿli kāna l-šakku fī l-fiʿl) or the utterance-initial (pro)noun ʾanta (fa-badaʾta bi'l-ismi kāna l-šakku fī l-fāʿil).Cp. also VERSTEEGH 1997: 259-260.
Guillaume18 highlights as follows: "The same preoccupation with clarifying the foundations of grammatical theory and with finding new, more explicit ways to formulate it is also perceptible in Ibn al-Sarrāǧ's (d.316/928) ʾuṣūl, a descriptive treatise following an entirely new and systematic order of exposition".Guillaume19 also highlights the isolated nature of this formal originality by observing that Ibn al-Sarrāǧ's successors fossilized his conceptual organization of Arabic grammatical theory into a "canonical mode of exposition for grammatical treatises" so that "no major evolution occurred in subsequent centuries" for such a theory. 20earing this in mind, we can now address the issue of (non-marginal) semantic originality in Arabic linguistic thought.

Al-Astarābāḏī and the Arabic system of case endings
Raḍī l-Dīn Muḥammad b. al-Ḥasan al-Astarābāḏī was an Arab grammarian of Iranian origin.He was born on an unknown date in the city of Astarābāḏ (present-day Iran), which is traditionally described as producing scholars proficient in all the sciences.In al-Astarābāḏī's time, however, the cultural potential of that milieu was probably limited by historical accidents, such as the Mongol invasions, which may explain why his commentary (šarḥ) on the syntactic treatise Kāfiya of Ibn al-Ḥāǧib (d.646/1249) was not circulated or developed by subsequent grammarians in spite of his scholarly prowess.Another possible explanation for the inadequate reception of al-Astarābāḏī's commentary-with the notable exception of al-Suyūṭī (d.911/1505)-was its innovative nature vis-à-vis the predominating conservatism of Arabic linguistic thought at that time.Al-Astarābāḏī may have died in 686/1287 or more likely in 688/1289. 21he conservative methodology of Arabic linguistic thought we have just alluded to devoted considerable attention to the system of case endings (iʿrāb), which constituted a central feature of the variety of Arabic described by the Arab grammarians and lexicographers.The resulting theory stands out for its conceptual simplicity: briefly, 22 what assigns the case ending to the noun is a particle or a verb 23 that precedes the noun in question.As a corol-lary, the phonological realization of the case ending assigned to the noun depends precisely on the nature of the constituent that precedes it.The particle assigns the case ending i to the noun; the verb assigns the case ending u to the first instance of a noun in the utterance, as well as the case ending a to the second instance of it; and a covert constituent, which has a verb-like and/or a particle-like nature, assigns the case ending u to the noun.For instance, in the utterance ḍaraba Zaydun ʿAmran 'Zayd hit ʿAmr', the verb ḍaraba assigns the case ending u to the proper noun Zayd and the case ending a to the proper noun ʿAmr. 24A theory of case along these lines is formal in the sense that no semantic considerations are invoked to explain the phonological realization of the case endings, the position of the utterance constituents only being relevant.Keeping to the example ḍaraba Zaydun ʿAmran, there is a tendency for the Arab grammarians to elaborate only very minimally upon the idea that the case ending u is assigned to the agent of the utterance Zayd, and the case ending a to its object ʿAmran. 25owever, al-Astarābāḏī takes the opposite approach by affirming that the case ending u is assigned to any necessary part of the utterance (ʿumda)26 , such as the subject and the predicate, and the case ending a is assigned to any optional part of it (faḍla), such as the object and the other complements.27A parallel with the modern linguistic notion of minimum clause will be useful to elucidate al-Astarābāḏī's theory of case endings28 , and especially the dialectics between ʿumda and faḍla29 upon which this theory is founded.To begin with, let us consider the utterance John ate an apple, from which we can derive the minimum clause John ate if we omit its object an apple.The relevant fact about this omission is that it deletes a portion of meaning, e.g., an apple, from the utterance, e.g., John ate an apple, without compromising the latter's overall semantics (and grammaticality) and yielding a minimum clause that is made of a subject and a (verbal) predicate, e.g., John ate.The same remarks apply to the utterance John ate yesterday, if we omit its complement of time particle) assign the case ending and the other (noun) receives it.Nonetheless, the ability of the verb to receive the case ending (cp. the imperfective forms yafʿalu, yafʿala) falsifies an interpretation of this sort.
yesterday.This semantic situation is tantamount to saying that in the minimum clause only the subject and the (verbal) predicate qualify as the necessary parts of the utterance, not unlike the ʿumda in al-Astarābāḏī's view, whereas the object and other complements are an optional part of it (cp.their omittability), not unlike the faḍla in his view.By way of illustration, the utterance Zaydun munṭaliqun 'Zayd is leaving' includes two instances of ʿumda (the subject Zaydun and the predicate munṭaliqun), whereas the aforementioned object ʿAmran is an instance of faḍla similarly to complements of time and manner (e.g., masāʾan 'in the evening', al-battata 'surely').This theory of case endings is semantic since it has at its core the notions of ʿumda and faḍla, which ultimately are but two sets of pieces of information one speaker conveys to another, such as substance, attribute (cp. the subject and the predicate that define the ʿumda), time, manner (cp. the complements of time and manner that define the faḍla). 30he mainstream formal theory of case endings and al-Astarābāḏī's semantic theory of case endings seem to be equally capable of explaining the presence of case endings in a simple utterance like ḍaraba Zaydun ʿAmran, where the case endings u and a can be analyzed either as two outcomes of the verb ḍaraba that precedes the nouns bearing them; or as an opposition necessary vs. optional part of the utterance.However, al-Astarābāḏī's semantic theory of case endings is seemingly superior to its formal counterpart when it comes to a more complex instance of utterance, which involves a passive form.Arabists have in the past noticed the difficulties experienced by the mainstream formal theory of case endings with respect to al-Astarābāḏī's theory, but the passive utterances they have taken into consideration belong to a somewhat ad hoc set of utterances often mentioned in the Arab grammarians' treatises, e.g.sīra farsaḫāni 'Two leagues were travelled'. 31Here, we would like to discuss the same theoretical scenario by means of a more concrete instance of passive utterance, drawn from the linguistic data gathered by Sībawayhi.The author of the Kitāb mentions a kind of passive utterance, in which the internal object displays an alternation of case endings u/a, e.g., ḍuriba bi-hi ḍarbun ḍaʿīfun / ḍarban ḍaʿīfan 'a weak blow was hit with it'. 32A certain amount of idealization is undeniable in this linguistic data (cp.the stereotyped example ḍuriba etc.), but the very alternation of case endings u/a in it plausibly points to a real context of dialectal variation. 33s has just been illustrated, the mainstream formal theory predicts that the verb assigns the case ending u to the first instance of a noun in the utterance, so that it accounts for one member of the alternation only, i.e., ḍarbun ḍaʿīfun, leaving the other, i.e., ḍarban ḍaʿīfan, unaccounted for.By contrast, al-Astarābāḏī's semantic theory of case endings provides a straightforward explanation for both members of the u/a alternation by interpreting them as two effects of two different communicative attitudes on the part of the speaker.If the speaker places informational saliency on the piece of information 'weak blow' (cp. the notion of internal object in modern Western linguistics), this element functions as a necessary part of the utterance (ʿumda), and therefore receives the case ending u.If the speaker does not place informational saliency on the piece of information 'weak blow', the same element functions as an optional part of the utterance (faḍla), thereby receiving the case ending a. 34 Insofar as al-Astarābāḏī worked out a semantic theory of case endings, thus departing from the formal theory of case endings that the Arab grammarians conservatively accepted and transmitted from one generation to another, we can credit him as a bearer of semantic originality in Arabic linguistic thought.His semantic originality is particularly remarkable in light of its ability to analyze certain facets of the utterance that Arabic linguistic thought traditionally takes great pains to analyze by means of its formal approach.That said, the disruption that al-Astarābāḏī represents with respect to mainstream Arabic linguistic thought should not prevent us from recognizing his continuity with it. 35Suffice it here to mention two facts.In first place, the notion of faḍla is already found in the work by al-Mubarrad (d.285/898). 36Secondly, and more importantly, al-Astarābāḏī himself presents his semantic theory of case endings as a development of some views held by al-Farrāʾ (d.207/822) 37 , who is well known for his strong interest in the linguistic exegesis of the Koran (cp.his huge work Maʿānī l-Qurʾān). 38The epistemological link between al-Farrāʾ and al-Astarābāḏī therefore provides the crucial indication that the semantic originality revealed by Arabic linguistic thought may possibly find its ultimate origin in the linguistic exegesis of the Koran.
34 This notion merely serves a clarification purpose.The question whether it can be assimilated to the notion of mafʿūl muṭlaq is not relevant here.Consequently, the difference in terms of case-assignment between the Western notion of internal object, as applied here, (alternation of case-endings u/a) and that of mafʿūl muṭlaq (case-ending a only) raises no interpretive difficulties.

Ibn Hišām and the Arabic definite article
The attentive reader will have noticed that the formal theory of case endings, just outlined in the previous section, in turn hinges on a classification of the parts of speech, namely the tripartite classification of Arabic words into noun, verb, particle (ism, fiʿl, ḥarf).One of the tersest formulations of this classification goes back to the incipit of Sībawayhi's Kitāb and has enjoyed great fortune up until recent times, as virtually no modern grammar of literary Arabic discounts the model of classification of Arabic words into ism, fiʿl, ḥarf: "The words are noun, verb and particle" (fa-l-kalimu smun wa-fiʿlun wa-ḥarf). 39The conservatism that pervades the Arab grammarians' classification of parts of speech is self-evident.
To this we could add that the classification in question also entails a certain amount of formalism, as shown by the influential analysis of the particle carried out by Sībawayhi in the aforementioned incipit of his Kitāb.In this passage, in fact, he does not set out a positive semantic definition of the particle (e.g., what denotes time, place, manner etc.), preferring instead to define it negatively as what is semantically neither a noun nor a verb: "the particle that occurs to [convey] a meaning, which is neither nominal nor verbal" (ḥarfun ǧāʾa li-maʿnàn laysa bi-smin wa-lā fiʿl). 40ence, it seems safe to maintain that the formal aspect prevails over the semantic one in the analysis of the particle developed by Arabic linguistic thought from Sībawayhi onward.Concretely, the Arabic definite article is among the particles that receives an analysis of this sort as, according to a recent study by Baalbaki, 41 even definiteness (taʿrīf), which represents its key property, is one of "the technical terms which refer to formal aspects" in the Kitāb (e.g., the position the article fulfills with respect to the noun).Such a formal (positional, etc.) analysis will also become conservative when the subsequent grammarians continue to pursue it, assigning a marginal role to the semantic properties of the Arabic definite article that they could identify, such as the latter's reference to previous knowledge (ʿahdiyya).However, a case can be made for a semantic treatment of the Arabic definite article on the part of Ibn Hišām.
Ǧamāl  (d.808/1406).This is a description of syntax arranged to start from each Arabic ḥarf in alphabetical order.In the Muġnī l-labīb ʿan kutub al-ʾaʿārīb, Ibn Hišām also deals with the Arabic definite article, which he regards as an instance of particle, and provides a more fine-grained account of the aforementioned notion of ʿahdiyya by classifying it into three subnotions, namely, maʿhūd ḏikriyyan, maʿhūd ḏihniyyan, maʿhūd ḥuḍūriyyan.39 SĪBAWAYHI, Kitāb, i: 12. Cp. also VERSTEEGH 1997: 242.40 SĪBAWAYHI, Kitāb, i: 12. Cp. also VERSTEEGH 1997: 242.41 BAALBAKI 2008: 173.Cp. also the beginning of this paper for the interplay between the (prevailing) formal approach and the (marginal) semantic approach in Sībawayhi's work.
They roughly correspond to the modern linguistic constructs of textual anaphora, extratextual anaphora, and deixis, respectively. 43Thus, by means of his tripartite and semantically-oriented classification of the Arabic definite article, Ibn Hišām brings forth a perspective that, because of its uniqueness within Arabic linguistic thought, is undeniably original; although this assertion must be tempered by the acknowledgement that in the same classification Ibn Hišām foregrounds a significant trait of continuity with mainstream Arabic linguistic thought.In fact, as just alluded to, Ibn Hišām takes as the departure point of his tripartite and semantically-oriented classification of the Arabic definite article the traditional (and marginal) notion of ʿahdiyya.A dialectics between originality and continuity therefore emerges in Ibn Hišām's linguistic thought, which constitutes a notable aspect of similarity with al-Astarābāḏī's thought. 44nother aspect of similarity that one grammarian shares with the other is a strong background in the linguistic exegesis of the Koran-as just alluded to, Ibn Hišām was appointed professor of this discipline. 45Ibn al-Qaṭṭāʿ and Arabic prefixation ʿAlī b.Ǧaʿfar b. ʿAlī al-Šantarīnī al-Saʿdī al-Ṣiqillī, also known as Ibn al-Qaṭṭāʿ, was an anthologist, historian, grammarian, lexicographer and poet, who was born in Sicily in 433/1041.In that period the island was first ravaged by civil war, then conquered by the Normans, leading him to leave Sicily in 1061.After a short stay in Andalusia, he finally settled in Egypt, where he died in 515/1121.There he circulated the al-Ṣiḥāḥ dictionary by al-Ǧawharī (d.398/1007-8), of which he is traditionally said to be the greatest transmitter and which he received from his teacher Ibn al-Birr (d.around 493/1100). 46ccording to the Arabic linguistic tradition, Ibn al-Qaṭṭāʿ is the author of two thematic glossaries (mubawwab) devoted to the morphological patterns (ʾabniya) found in the kalām al-ʿArab.While one thematic glossary, the so-called Kitāb al-ʾafʿāl, only deals with verbal patterns, the other, transmitted under the title Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼlmaṣādir, revolves more broadly around the patterns of nouns, verbs and the hybrid category they give rise to: the verbal noun (maṣdar). 4743 See IBN HIŠĀM, Muġnī l-labīb ʿan kutub al-ʾaʿārīb, i: 106, 108.This passage reads as follows: ʾal ʿalà ṯalāṯati ʾawǧuhin […] wa'l-ṯānī ʾan takūna ḥarfa taʿrīfin wa-hya nawʿāni ʿahdiyyatun wa-ǧinsiyyatun wa-kullun min-humā ṯalāṯatu ʾaqsāmin fa'l-ʿahdiyyatu ʾimmā ʾan yakūna maṣḥūbu-hā maʿhūdan ḏikriyyan […] ʾaw maʿhūdan ḏihniyyan […] ʾaw maʿhūdan ḥuḍūriyyan.The parallel between maʿhūd ḏikriyyan, maʿhūd ḏihniyyan, maʿhūd ḥuḍūriyyan and textual anaphora, extra-textual anaphora, deixis is proposed by GULLY 1995: 146-8.Cp. also VERSTEEGH 1997: 265.44 See the end of the previous section.
45 See also the end of the previous section.
The Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir is of particular interest here because of the potential it bears in terms of semantic originality.A good indication of its general tendency to originality is its inclusion of all sorts of Arabic morphological patterns, even those not mentioned by Sībawayhi, in its collection.Moreover, the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir opts to treat the verbal noun as a self-contained object of investigation, in sharp contrast to previous works of the same genre, and in so doing relies upon a definition of verbal noun that is semantic, to the extent that it decomposes this kind of lexeme into a peculiar combination of two semantic primitives, i.e, the nominal and verbal properties (componential analysis). 48From this vantage point, the choice of pinpointing the verbal noun as a self-contained object of investigation is fairly indicative of the particular tendency to semantic originality of the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir and of its author Ibn al-Qaṭṭāʿ.In the remainder of this section, we further corroborate the hypothesis that an original attitude to semantic originality informs the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir by means of a case study of a fundamental ingredient of Arabic morphological patterns-affixation-, and especially in the interpretation of it offered by Ibn al-Qaṭṭāʿ in this treatise.
Within the theoretical framework of the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir, affixation (ziyāda and related terms: zāʾid etc.) in essence has a consonantal nature and performs the function of increasing the length of morphological patterns.The root (aṣl) shares with affixation the same nature and function, as it manifests itself as triconsonantal, quadriconsonantal and so on.Both consonantal affixes and root consonants can co-occur with vowels when increasing the length of morphological patterns.This theoretical framework is apparent in the conceptual structure of the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼlmaṣādir, which organizes the morphological patterns according to a criterion of increasing length of root consonants and consonantal affixes, owing much to Sībawayhi in this regard.By way of illustration, Sībawayhi mentions the morphological patterns fuʿl, fuʿul, ʾafʿul precisely in this order of increasing length, as does Ibn al-Qaṭṭāʿ in his Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir.What is more, the latter grammarian, like the former, makes use of the fundamental terminological pair aṣl/zāʾid. 49ince the criterion of increasing length involves no semantic factor and revives the criterion of increasing length adopted by Sībawayhi, the theoretical framework of the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir is plausibly one of the many instances of formal and conservative approach that characterize Arabic linguistic thought.This observation does not deny the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir the semantically origi-48 From a textual perspective, this choice is reflected in the conceptual structure of the Kitāb ʾabniyat al-ʾasmāʾ wa'l-ʾafʿāl wa'l-maṣādir, which deserves a separate treatment to the verbal noun patterns, contrary to previous works, such as the Kitāb al-Istidrāk authored by al-Zubaydī (d.379/989).See BAAL-BAKI 2014: 285.49 IBN al-QAṬṬĀʿ, Kitāb ʾabniyat al-ʾasmāʾ wa'l-ʾafʿāl wa'l-maṣādir: 135, 140; SĪBAWAYHI, Kitāb, iv:  242-245.For simplicity's sake, the terminological pair aṣl/zāʾid is rendered here as root/affix in the wake of BAALBAKI 2002: 1.This terminological pair is effectively part and parcel of a broader lexical set, which also includes ziyāda (affixation) mazīd (affixed) etc. See, e.g., IBN al-QAṬṬĀʿ, Kitāb ʾabniyat al-ʾasmāʾ wa'l-ʾafʿāl wa'l-maṣādir: 92, 109.But cp. also LARCHER 1995, who brings solid arguments in favor a more accurate translation-and conceptualization-of the terminological pair aṣl/zāʾid.
nal character we have alluded to immediately above and is instead meant to highlight the aspects of continuity that this treatise instantiates along with its aspects of originality.
Returning to the comparison between the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼlmaṣādir and the Kitāb, a closer look at the passages that expound the morphological pattern fuʿul reveals a certain difference between the two treatises.While Sībawayhi exemplifies the morphological pattern fuʿul by means of the word ǧumud without explaining the latter's meaning, Ibn al-Qaṭṭāʿ supplements Sībawayhi's example with the gloss 'name of a mountain' (ism ǧabal). 50nsofar as this gloss helps to elucidate the meaning of the word ǧumud and is not found in Sībawayhi's work, it can qualify as a sort of semantic originality on the part of Ibn al-Qaṭṭāʿ.However, the semantic originality under scrutiny is not as crucial, given that it is not original to Ibn al-Qaṭṭāʿ: the practice of glossing obscure words, the meaning of which Sībawayhi omitted to record, is typical of the genre of thematic glossary to which the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir belongs. 51It is also worth noting that the semantically-oriented practice of glossing obscure words mainly arose and developed in the milieu of the linguistic exegesis of the Koran, as evidenced by the type of thematic glossary traditionally known as ġarīb al-Qurʾān. 52It follows that the original glosses that Ibn al-Qaṭṭāʿ associates with the morphological patterns in the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir lack conceptual originality yet bear testimony, again (cp. the two previous sections), to an epistemological link between semantic originality in Arabic linguistic thought and the background of linguistic exegesis of the Koran.
On the other hand, a major trait of semantic originality that we can in all likelihood fully ascribe to Ibn al-Qaṭṭāʿ alone emerges from a careful examination of a passage of the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir, drawn from its chapter on affixes (Bāb ḥurūf al-zawāʾid). 53The passage in question describes the w-affix as follows: "w can be inserted within a noun or a verb, but not in first position, except for the [expression of] oath; it can be inserted within them in second position, as in kawṯar" (wa'l-wāwu tulḥaqu fī 50 IBN al-QAṬṬĀʿ, Kitāb ʾabniyat al-ʾasmāʾ wa'l-ʾafʿāl wa'l-maṣādir: 135; SĪBAWAYHI, Kitāb, iv: 243. 51 BAALBAKI 2014: 60.In principle, we can hypothesize that Sībawayhi omitted to record the meaning of ǧumud since it was a toponym well-known to him and to the educated people of his time; and that, on the contrary, Ibn al-Qaṭṭāʿ felt the need to expound the meaning of the same word as, centuries later after Sībawayhi, it had become incomprehensible to Ibn al-Qaṭṭāʿ himself and to his educated audience.However, textual research militates against this hypothesis.The early lexicographer Abū ʿUbayda (d.209/824), who died about thirty years after Sībawayhi, glosses precisely the word ǧumud as the name of a mountain located in Najd under the sphere of influence of the Banū Naṣr tribe, which plausibly shows that this word was already obscure in Sībawayhi's time.Abū ʿUbayda's gloss, which had been transmitted by the geographer Yāqūt (d.626/1229), reads as follows: al-ǧumudu bi-ḍammatayni qāla abū ʿubaydata huwa ǧabalun li-banī naṣrin bi-naǧd (cp.YĀQŪT, Muʿǧam al-Buldān, ii: 161).See also BAALBAKI 2014: 19, 165 for further information about Abū ʿUbayda.However, it is also worth pointing out that the different kinds of linguistic analysis carried out by Sībawayhi and Abū ʿUbayda (naḥw and luġa, respectively), might have plausibly influenced the absence vs. the presence of glosses associated with nominal patterns and related words such as fuʿul and ǧumud.

l-ismi waʼl-fiʿli illā anna-hā lā tulḥaqu awwalan illā fī l-qasami l-battata wa-tulḥaqu ṯāniyatan fī kawṯar). 54
The passage of the Kitāb that describes the same affix differs markedly from the previous passage in that it does not admit the occurrence of w in first position, i.e., as an affix that can occur at the beginning of a noun or verb: "regarding w, it can be inserted in second position, as in ḥawqal" (ammā l-wāwu fa-tuzādu ṯāniyatan fī ḥawqal). 55n essence this difference boils down to the interpretation of the expression of oath, which in the variety of Arabic investigated by Sībawayhi and Ibn al-Qaṭṭāʿ frequently takes on the form of a string wa, as in wa-llāhi lā afʿalu. 56On the one hand, Sībawayhi purports that wa is a sort of variant of the particle bi, underscoring two syntactic properties of this expression of oath.First, wa has the ability to co-occur with the name Allāh, just like the particle bi does.Second, wa has the ability to assign genitive, just as the particle bi does.In Sībawayhi's own words: "the bāʾ [that assigns] genitive serves to join and connect [words]  […] and the wa used for the expression of oath fulfills the role of the bāʾ " (wa-bāʾu l-ǧarri inna-mā hiya li-l-ilzāq wa'l-iḫtilāṭi waʼl-wāwu llatī takūnu li-l-qasami bi-manzilati l-bāʾ). 57n sum, due to its focus on two syntactic properties of wa, which involve no semantic factors (co-occurrence, genitive-assignment), Sībawayhi's analysis of wa is formal.
On the other hand, it can be argued that Ibn al-Qaṭṭāʿ's affixal analysis of wa, which we have just illustrated, is semantically-oriented.The argument is built as follows.First, as discussed at the end of the Introduction, from Sībawayhi onward the affix that performs the function of ilḥāq is combined with a pure morphological pattern, as is the case for fawʿal (cp.kawṯar), or faʿwal (cp.ǧadwal).Second, the affixal wa that co-occurs with the name Allāh (e.g., wa-llāhi lā afʿalu) is not combined with a pure morphological pattern, but with a morphological pattern plus the article al (cp. the string Al in Allāh).On these grounds, this instance of wa must perform a function other than ilḥāq.Third, as discussed at the end of the Introduction, from Sībawayhi onward the only other function, besides ilḥāq, assigned to the affix by even the formal approach of Arabic linguistic thought is semantic.Hence, by exclusion, the affixal wa that co-occurs with the name Allāh performs a semantic function: in this case, that of conveying the meaning of oath.
A semantically-oriented analysis along these lines, which is culled from Ibn al-Qaṭṭāʿ's twofold characterization of the w-affix as word-initial and related to oath (i.e., wa), appears to stand as an interesting trait of originality within Arabic linguistic thought.It is very instructive in this regard that three centuries after Ibn al-Qaṭṭāʿ's death and beyond, both the erudite works al-ʾItqān fī ʿulūm al-Qurʾān, authored by al-Suyūṭī (d.911/1505), and Tāǧ al-ʿArūs, authored by al-Zabīdī (d.1205/1790), provide thorough and exhaustive reviews of the several interpretations associated with the string wa in all of its contexts of occurrence, yet neither of them mentions Ibn al-Qaṭṭāʿ's analysis of wa in terms of an affix when they 54 IBN al-QAṬṬĀʿ, Kitāb ʾabniyat al-ʾasmāʾ wa'l-ʾafʿāl wa'l-maṣādir: 101.
55 SĪBAWAYHI, Kitāb, iv: 237.56 SĪBAWAYHI, Kitāb, iv: 217.Cp. also WRIGHT 1896, i: 279.57 SĪBAWAYHI, Kitāb, iv: 217.discuss the instance of wa that expresses oath. 58The only analysis referred to in this connection by al-Suyūṭī and al-Zabīdī is that of Sībawayhi, as is easily gleaned from a simple comparison between his definition of the wa that expresses oath, which we have quoted immediately above, and their definitions of the same instance of wa.Thus, al-Suyūṭī asserts that "the wa that expresses oath is a genitive-assigner" (fa'l-ǧārratu wāwu l-qasam). 59ikewise, al-Zabīdī states that "the wa that expresses oath is an alternant of bi" (wāwu lqasami … badalun min al-bāʾ). 60hat is more, at the beginning of the chapter forty-one of his grammatical treatise al-Muzhir fī ʿulūm al-luġa wa-anwāʿi-hā al-Suyūṭī explicitly mentions the treatise Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir, in which Ibn al-Qaṭṭāʿ analyzes the w-affix as word-initial and related to oath (i.e., wa) 61 and yet in the same work al-Suyūṭī refrains from mentioning this analysis by Ibn al-Qaṭṭāʿ. 62It is of the utmost importance to note at this point that the failure to mention Ibn al-Qaṭṭāʿ's affixal and semantically-oriented analysis of the wa that expresses oath on the part of al-Suyūṭī and al-Zabīdī cannot necessarily be ascribed to their ignorance of the morphological work of the Sicilian grammarian.On the one hand, as we have just observed, in the Muzhir fī ʿulūm al-luġa wa-anwāʿi-hā al-Suyūṭī explicitly and copiously cites Ibn al-Qaṭṭāʿ's Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-58 The lack of an analysis of wa in terms of a word-initial affix in al-Suyūṭī's and al-Zabīdī's work is regarded here as a sort of qualitative evidence of the original nature of such an analysis on the part of Ibn al-Qaṭṭāʿ (in the sense that this kind of evidence focuses on how al-Suyūṭī and al-Zabīdī used to deal with the body of knowledge elaborated on by their predecessors, Ibn al-Qaṭṭāʿ included).It would be also possible to provide quantitative evidence to the same effect.The gist of the proposal is to study the grammatical literature between Sībawayhi's Kitāb and Ibn al-Qaṭṭāʿ's treatise to ascertain whether the Sicilian grammarian really developed an original analysis or took it from one of his predecessors.The scope of this paper prevents a thorough presentation of this kind of quantitative evidence.However, quantitative evidence of this sort is at least in part implied by the qualitative evidence adduced in this study.In fact, the tendency to encyclopedism and erudition on the part of al-Suyūṭī and al-Zabīdī implies that, in order to eruditely enumerate all of the possible analyses of wa (qualitative evidence), they had to check and peruse the grammatical literature between Sībawayhi's Kitāb and Ibn al-Qaṭṭāʿ's treatise (quantitative evidence), included those works that are lost to us.For instance (see BAALBAKI 61 The locus probans is the following: ḏikru ʾabniyati l-ʾasmāʾi wa-ḥaṣri-hā qāla abū l-qāsimi ʿaliyyun-i bnu ǧaʿfara l-saʿdiyyu l-luġawiyyu l-maʿrūfu bi-bni l-qaṭṭāʿi fī kitābi l-ʾabniyah (al-SUYŪṬĪ, al-Muzhir fī ʿulūm al-luġa wa-anwāʿi-hā, ii: 4).In this passage, the Kitāb al-ʾabniya the Egyptian polymath refers to is precisely the Kitāb ʾabniyat al-ʾasmāʾ wa'l-af'āl wa'l-maṣādir, as is inferred from the very phrase ʾabniyat al-ʾasmāʾ in the section heading ḏikr ʾabniyat al-ʾasmāʾ.To summarize the main results of this section, a first examination of the Kitāb ʾabniyat al-ʾasmāʾ waʼl-ʾafʿāl waʼl-maṣādir seemingly reveals an appreciable tendency on the part of Ibn al-Qaṭṭāʿ toward semantically-oriented originality, which is plausibly rooted in the milieu of the linguistic exegesis of the Koran (cp.his practice of glossing obscure words, e.g., ǧumud).The most conspicuous instance of an originality of this kind is his treatment of w as a word-initial affix wa, provided as such with the meaning of oath.This semantic originality is to a certain extent due to Ibn al-Qaṭṭāʿ, as both his predecessors (Sībawayhi) and successors (al-Suyūṭī, al-Zabīdī) reject an interpretation of w as a word-initial affix 63 See the passage quoted in the previous footnote.
and/or subscribe to a formal interpretation of wa, which denies the latter a semantic content in its function as a word-initial affix, instead regarding it as a genitive-assigning particle.

Conclusions
This paper has plausibly substantiated the hypothesis that Ibn al-Qaṭṭāʿ can be considered, along with the perhaps most famous grammarians al-Astarābāḏī and Ibn Hišām, as one of the few bearers of semantic originality in the context of medieval Arabic linguistic thought, as is shown by the construct of a word-initial and meaningful affix w(a).Such a construct is seemingly absent in Sībawayhi's Kitāb, whereas Ibn al-Qaṭṭāʿ posits it and identifies it with the so-called wāw al-qasam.This paper also stresses the point that the traits of semantic originality introduced into Arab linguistic thought by al-Astarābāḏī, Ibn Hišām and Ibn al-Qaṭṭāʿ share a common epistemological aspect: they possibly find their ultimate origin in the milieu of the linguistic exegesis of the Koran.Further research is needed to acquire a better understanding of how, on the whole, the original aspects of the semantic approach pursued by al-Astarābāḏī, Ibn Hišām and Ibn al-Qaṭṭāʿ position themselves within the historical development of Arabic linguistic thought, which witnessed at least three stagesearly, or formative, classical, and late, or post-classical. 68.At the current research stage it seems safer to maintain that the semantically-oriented approach co-existed with the formal approach since the beginnings of Arabic linguistic thought, albeit in an implicit or embryonic form, so the original character of Late grammarians such as al-Astarābāḏī, Ibn Hišām and Ibn al-Qaṭṭāʿ lies mainly in their efforts to make the semantically-oriented approach more explicit and central.