The Vocabulary of the Qurʾān and Multilingualism in Arabia

Abstract

This article examines five Qurʾānic lexical items—surādiq (Q 18:29), qiṭṭ (Q 38:16), ḥiṭṭah (Q 2:58; 7:161), fūm (Q 2:61), and yaqṭīn (Q 37:146)—through a theoretical framework that combines multilingualism in Arabia before Islam with muqārana, understood as comparative philological practice. Rather than simply asking whether each word is Arabic or foreign, the article evaluates each case through Qurʾānic context, Arabic morphology and lexicography, phonotactic markedness, comparative Semitic, Iranian, and Mediterranean evidence, variant readings (qirāʾāt), and early exegetical reception (tafsīr). Surādiq illustrates Iranian–Aramaic mediation in eschatological imagery; qiṭṭ and ḥiṭṭah show how documentary and religious-formulaic semantics may preserve older Semitic contact strata; fūm demonstrates how a Qurʾānic food term can be pulled between an archaic Arabic grain/bread meaning, non-canonical reading tradition, and harmonisation via Biblical comparison; and yaqṭīn functions as a control case against the over-identification of borrowings. The article argues that Qurʾānic vocabulary is best studied as multilingual lexical memory: a field in which etymology and exegesis interact without collapsing into a binary opposition between Arabic and foreign vocabulary.

1. Introduction

Because the Qurʾān emerged in an Arabian contact zone, Qurʾānic vocabulary has long been a privileged site for debates about language, religion, and cultural transmission. In the Hijaz and north-west Arabia, the clearest direct traces of multilingualism are epigraphic: Aramaic and Nabataean Aramaic, Ancient Arabian inscriptional traditions—especially Dadanitic and other Ancient North Arabian epigraphic material, as well as Minaic/Old South Arabian material in trade contexts (Rossi 2014)—and Old Arabic written in several scripts (Al-Jallad and Sidky 2022). Hebrew, Greek, and Latin are more occasional in the regional epigraphic record, while Syriac, Geʿez/Ethiopic, and Middle Iranian languages belong above all to the wider contact environment rather than to the primary local written languages of the Hijaz (Dye 2017; Cheung 2017). Dye collects evidence for traces of bilingualism and multilingualism in Qurʾānic Arabic and warns against treating Arabic as isolated from the language practices surrounding it (Dye 2017), and epigraphic research confirms that—before and around the rise of Islam—Arabic was embedded in multilingual environments. Here, multilingualism refers not only to the coexistence of several languages, but to the practical movement of words, scripts, genres, and religious concepts across Arabic, Aramaic/Syriac, Hebrew, Greek, Persian, and other linguistic traditions. It is therefore treated as a historical condition of Qurʾānic lexical formation and reception, not simply as a background fact.

Jeffery’s Foreign Vocabulary of the Qurʾān (Jeffery 1938) remains the last broadly exhaustive catalogue of proposed Qurʾānic loanwords, but it belongs to a longer tradition of Muslim muʿarrab and gharīb al-Qurʾān scholarship and to a modern philological tradition that is being revised through research into Semitic linguistics, Arabian epigraphy, Syriac/Aramaic studies, and contextual semantics. Although the study of Qurʾānic lexis has changed significantly since Jeffery’s work, his catalogue remains indispensable as a collection of proposals, as its 2006 Brill republication indicates. At the same time, its explanatory framework often treats lexical resemblance as evidence for borrowing too quickly and sometimes describes Qurʾānic forms as ‘corrupt’ or ‘garbled’. Pennacchio (2011) has shown why Jeffery’s list requires reassessment, and recent scholarship has increasingly shifted the question from simple origin to semantic function, transmission history, and linguistic ecology. The question is therefore how multilingual contact, lexical transfer, and Arabicisation can be studied without reducing the Qurʾān either to a list of foreign words or to an artificially monolingual text. Instead, when a Qurʾānic word is traced through Aramaic, Persian, or Greek, the question is not only where the word came from but how it functions in a sacred text and how interpreters made it meaningful. The study of the vocabulary of the text therefore contributes to the history of translation, religious discourse, and multilingual knowledge in Islamic intellectual history.

The research gap addressed here is methodological. Existing studies have either catalogued possible foreign vocabulary or examined individual Qurʾānic terms in depth, but they have less often offered a controlled procedure for deciding when lexical resemblance indicates borrowing, shared inheritance, semantic convergence, exegetical normalisation, or no direct lexical relationship at all. The original contribution of this article is therefore not the identification of “foreign words” as such, but a graded model for evaluating Qurʾānic lexical evidence across internal Arabic usage, comparative philology, phonotactics, scriptural context, and reception history.

The five case studies were selected because they represent distinct evidentiary profiles: surādiq tests strong mediated Arabicisation; qiṭṭ and ḥiṭṭah test segmental patterning; fūm tests exegetical and scriptural normalisation; and yaqṭīn tests whether the method can reject an attractive but weak borrowing hypothesis.

The modern discussion of Qurʾānic loanwords has deep roots in the Arabic-Islamic scholarly tradition itself. Muslim philologists, lexicographers, and exegetes did not treat the Qurʾānic lexicon as a linguistically homogeneous corpus and developed several categories for discussing unusual vocabulary. The literature on gharīb al-Qurʾān sought to explain rare, archaic, dialectal, or semantically difficult expressions through poetry, tribal usage, and linguistic comparison. Related discussions concerned muʿarrab (‘Arabicised’) words, understood as words that had entered Arabic from other languages and become integrated into Arabic speech, and dakhīl (‘foreign’ or ‘intrusive’) vocabulary, a category developed more fully in later lexicographical and philological literature (al-Jawālīqī 1990; Jeffery 1938, pp. 1–11). These discussions emerged from broader debates about the linguistic character of the Qurʾān and the relationship between its claim to be an ‘Arabic’ revelation (e.g., Q 12:2; 16:103; 26:195) and the evident circulation of words, concepts, and names across the multilingual environments in which Arabic speakers participated.

Classical scholars adopted different positions. Some rejected the presence of any non-Arabic vocabulary in the Qurʾān, arguing that divine revelation in ‘clear Arabic’ precluded foreign elements. Others, including authorities cited by al-Suyūṭī, accepted that certain words had originally entered Arabic from Persian, Ethiopic, Greek, Aramaic, or other languages but maintained that such forms had become fully Arabic before the Qurʾānic revelation and therefore belonged to the Arabic language of the Qurʾān (Al-Suyūṭī 2006, vol. 2, pp. 339–41; Carter 2006, pp. 121–24). This latter position is especially relevant for the present article because it anticipates the distinction made here between lexical prehistory and Qurʾānic function. A word may preserve traces of contact, borrowing, or mediation while functioning as an ordinary element of Qurʾānic Arabic.

Sinai’s Key Terms of the Qurʾan is especially important for this article because it asks what Qurʾānic terms mean in usage, how they should be translated, and how they contribute to the Qurʾān’s conceptual and theological world (Sinai 2023). This semantic orientation should be combined with work on the history of Qurʾānic Arabic itself. Van Putten’s study of Qurʾānic Arabic uncovers the earliest Hijazi linguistic layer of the text through manuscript and medieval literary evidence and describes its later development into the Classical Arabic reading traditions (van Putten 2022). This matters for lexical study because words should be evaluated within the linguistic history of Qurʾānic Arabic, prior to external comparison with other Semitic or non-Semitic languages as evidence for borrowing.

The present article builds on this scholarship by combining historical premises with an analytical procedure. The historical premise is that Qurʾānic vocabulary emerged within a multilingual Arabian environment shaped by interaction with neighbouring linguistic and religious traditions. The analytical procedure is muqārana, understood here as controlled comparison across form, meaning, context, reception, and historical plausibility (Rashwan 2024b). The following section sets out the sources and criteria used in the analysis, while Section 3 explains the theoretical value of multilingualism and muqārana for Qurʾānic lexical study.

The article therefore asks how Qurʾānic lexical items can be evaluated as elements of a multilingual religious ecology without being prematurely classified either as foreign intrusions or as purely internal Arabic formations. A Qurʾānic word may preserve a non-Arabic prehistory and yet be fully Arabic in Qurʾānic usage; conversely, a word may appear Biblically resonant without being a loanword. By treating Qurʾānic lexical interpretation as a site where Arabic, Aramaic/Syriac, Hebrew, Iranian, Greek, and exegetical traditions intersect, the article contributes to the study of multilingualism, translation, and religious knowledge in Islamic intellectual history.

2. Materials and Methods

The primary Arabic sources used here include early and classical Qurʾānic exegesis (tafsīr) and general and specialist Arabic lexicography, as well as later exegetical or lexicographical witnesses where they preserve relevant earlier traditions. The empirical basis for the phonotactic work is lisan345 (v1.0.0), a freely available GitHub dataset derived from a digitised edition of Ibn Manẓūr’s (d. 1311) Lisān al-ʿArab (cf. Elmaz 2011, pp. 123–25, 303–17). The repository offers complete lists of the 6529 triliteral, 2551 quadriliteral, and 183 quinquiliteral roots in the Lisān al-ʿArab, labelled in Arabic and transliteration1, with each consonant coded for its place and manner of articulation. This work is used because of its comprehensiveness as a repository of the Arabic lexicographical tradition, while its relatively late compilatory date is controlled for through comparison with earlier lexical witnesses recorded in al-Khalīl b. Aḥmad al-Farāhīdī’s (d. c. 786) Kitāb al-ʿAyn, regarded as the earliest extant Arabic dictionary. Comparative evidence is drawn from Semitic, Middle Iranian, and Greek etymological and lexicographical literature.

The analysis considers five types of evidence. First, each word is read in its Qurʾānic context, including its narrative, polemical, eschatological, formulaic, or botanical function. Second, its Arabic morphology and root structure are evaluated in light of Arabic lexicography and phonotactic constraints. Third, proposed cognates or source forms in Aramaic, Hebrew, Akkadian, Persian, Greek, or other languages are assessed for formal similarity, semantic fit, historical plausibility, and likely mediation. Fourth, early exegetical and lexicographical interpretations are treated as evidence for reception and semantic normalisation. Fifth, lexical origin is distinguished from Qurʾānic function, since a word with a non-Arabic prehistory may function as fully Arabic in the Qurʾān, while a word with biblical narrative correspondence need not be borrowed.

Following Saleh’s critique of the etymological fallacy, lexical interpretation begins with Qurʾānic usage, syntax, literary context, and early reception before broader comparative parallels are invoked (Saleh 2010). In this article, a borrowing or Arabicisation hypothesis is considered strong only when several criteria converge: phonological or phonotactic markedness in Arabic, a plausible donor or mediating form, semantic continuity, historical plausibility of contact, and support from early lexicography, exegesis, or parallel textual traditions. Conversely, a borrowing hypothesis is weakened when the Arabic form is morphologically regular, semantically well explained in Arabic, and unproblematic in early Muslim interpretation.

The five case studies were selected according to three criteria (see

Table 1. Evidence matrix for the Qurʾānic case studies.

Case	Context	External Comparison	Arabic Evidence	Evaluation
surādiq	Q 18:29; eschatological enclosure or surrounding forecourt	Mand./Syr. and Middle Iranian comparanda	Marked phonotactic neighbourhood; Arabicised parallels	Strong case for mediated Arabicisation
qiṭṭ	Q 38:16; written share or allotment before reckoning	Semitic comparanda	Short noun with marked -iṭṭ- sequence	Strong older Semitic documentary layer
ḥiṭṭah	Q 2:58; 7:161; formulaic utterance of forgiveness	Aram./Syr. religious semantics; Akk. cognate field	Possible Arabic root ḤṬṬ but religious context supports comparison	Possible Aramaism or semantic convergence
fūm	Q 2:61; produce/food list	Num. 11:5 vs. Ezek. 4:9; non-canonical reading; Semitic garlic terms	Early ḥinṭah/khubz/grain glosses; verb fawwama ‘to bake bread’	Probable grain/bread archaism, secondarily associated with ‘garlic’
yaqṭīn	Q 37:146; Jonah’s plant	Heb. qīqāyōn proposed by Jeffery; Gk./Lat. translation variation	Regular root, rare yafʿīl pattern, coherent tafsīr	Weak borrowing case; methodological control

Abbreviations used for languages: Akk. = Akkadian; Aram. = Aramaic; Gk. = Greek; Lat. = Latin; Mand. = Mandaic; Syr. = Syriac.

). First, each term is Qurʾānic and interpretively marked, either because it is rare, contextually difficult, or historically associated with the foreign-vocabulary paradigm. Second, the set deliberately covers different semantic domains: eschatological architecture, documentary/legal vocabulary, religious formulae, food terminology, and botanical narrative vocabulary. Third, the cases form a graded evidentiary scale, from strong contact evidence to a control case where borrowing should be rejected. Other Qurʾānic lexical items were excluded not because they are less important, but because the aim here is not exhaustive coverage. The aim is to test a portable method on a small set of terms whose evidence varies in strength and type.

The method has three limitations. Later Arabic lexicography is indispensable but must be checked against earlier lexicographical, exegetical, and Qurʾānic usage. External comparanda are likewise probative only when form, meaning, chronology, and contact pathway converge; similarity alone is not enough. Finally, phonotactic markedness can identify suspicious or marginal lexical neighbourhoods, but it cannot by itself prove borrowing. The argument is therefore cumulative and probabilistic: each case is evaluated by convergence among several kinds of evidence, not by any single criterion.

The theoretical framework of the article rests on the claim that the vocabulary of the Qurʾān should be studied as part of the religious and linguistic environment in which it emerged. Arabic occupied a central position in the Qurʾān and in later Islamic scholarship, but recent work has shown that the centrality of Arabic in Islamic societies did not erase multilingual practice, transculturation, or comparative engagement (Rashwan 2024a; Salvaggio 2024). Multilingualism is therefore not a marginal background condition, but part of the lexical environment in which the vocabulary of the text was produced, received, and interpreted.

The Arabian Peninsula was a zone of mobility and contact, in which traders, monks, scribes, poets, tribes, translators, administrators, and exegetes moved words across languages, scripts, and domains. Some words entered Arabic as cultural loans, such as words for ‘silk’, commonly connected with Middle Persian such as sundus and stabrag, or siyarāʾ/*sīrāʾ, ultimately from Old Chinese *səɣ (Elmaz 2021, pp. 44–46); others passed through Greek, Aramaic/Syriac, Persian, or mixed channels of mediation, as in the frequently discussed cases of injīl, firdaws, and ṣirāṭ (Sinai 2023, pp. 103–7, 191, 473–75); others have been interpreted as borrowings because their narrative or theological roles recall scriptural parallels, even when direct borrowing is not demonstrable, as with sakīnah or al-yamm in the Moses tradition (Reynolds 2018, pp. 96, 275); and still others are better explained as inherited Semitic cognates or shared lexical stock, such as Arabic bayt beside Hebrew bayit, Aramaic baytā, and Akkadian bītu, or Arabic kalb beside Hebrew keleb, Aramaic kalbā, and Akkadian kalbu (Kogan 2011). The task is therefore not to label each word as simply Arabic or non-Arabic, but to define its lexical ecology. The term lexical ecology is used here to describe this wider field of relationships: borrowing, mediation, shared inheritance, semantic convergence, scriptural association, and exegetical reception. It avoids reducing each word to a single question of origin and instead asks how a form, meaning, and interpretive history interact.

In this context, muqārana designates the article’s analytical procedure. It is not merely the comparison of Arabic words with non-Arabic candidates, but a controlled comparison of forms, meanings, contexts, and receptions. Its purpose is to determine what kind of relationship best explains the evidence, and it differs from the older foreign-vocabulary paradigm in three ways: it does not assume that resemblance equals borrowing; it treats early Muslim exegesis and lexicography as evidence for reception and semantic stabilisation; and it separates etymological history from Qurʾānic function.

Arabicisation is used more narrowly for cases in which a non-Arabic or mediated form becomes integrated into Arabic phonology, morphology, lexicography, and usage. A word may be Arabicised in this sense even if its earlier history lies outside Arabic; conversely, not every word with a foreign comparandum is necessarily Arabicised from that source.

A related process is semantic transfer: the movement of a meaning, concept, or discourse function from one domain into another. In the cases studied here, what travels is not always a word-form alone. Documentary vocabulary may be recontextualised in eschatological discourse, religious formulae may enter Qurʾānic Arabic through shared Semitic idioms, and plant or food terms may be reshaped by scriptural comparison. Translation is therefore approached not only as the rendering of texts, but also as the movement and reconfiguration of lexical meanings across religious and scholarly traditions.

3. Results

The following case studies apply the method outlined above to five Qurʾānic lexical items with different evidentiary profiles. The order of the cases moves from the strongest external-contact hypothesis to the clearest control case. Surādiq represents the strongest case for mediated Arabicisation, since internal phonotactic evidence converges with Iranian and Aramaic/Mandaic comparanda. Qiṭṭ and ḥiṭṭah occupy a more complex middle ground, where documentary and religious semantics interact with Arabic morphology and broader Semitic parallels. Fūm illustrates how early Arabic glossing, biblical comparison, and variant reading can pull interpretation in different directions. Yaqṭīn, finally, functions as a control case: although it occurs in a biblical narrative context, the Arabic morphological and exegetical evidence weighs against deriving it from the Hebrew plant name qīqāyōn. Together, the five cases test the article’s central claim: Qurʾānic lexis is best analysed through cumulative comparison rather than through a binary opposition between Arabic and foreign.

3.1. Iranian–Aramaic Mediation and Eschatological Enclosure: Surādiq

The first case study is Qurʾānic surādiq in Q 18:29:

إِنَّا أَعْتَدْنَا لِلظَّالِمِينَ نَارًا أَحَاطَ بِهِمْ سُرَادِقُهَا

ʾinnā ʾaʿtadnā li-l-ẓālimīna nāran aḥāṭa bihim surādiquhā ‘We have prepared for the wrongdoers a fire that (lit. the enclosure/walls of which) will surround them’.2 It illustrates how muqārana can identify a strong contact hypothesis without reducing Qurʾānic usage to its prehistory. The question is not only where surādiq came from, but how a lexeme connected with enclosure, canopy, or architectural space becomes part of Qurʾānic eschatological rhetoric. It is not only an architectural term, but it helps define the spatial imagery of punishment. Its possible Iranian–Aramaic prehistory therefore matters for the study of how material, royal, or architectural vocabulary could be resemanticised in religious discourse.

A useful way to begin testing whether a word is a borrowing is to examine its phonotactic neighbourhood within Arabic rather than treating it as an isolated lexical term. This means comparing it with roots that exhibit comparable consonantal configurations: roots with the same radicals, roots containing consonants from the same places of articulation, and roots whose consonants share the same manner of articulation (cf. Elmaz 2011, pp. 75–78). Such comparison helps determine whether the word is phonologically anomalous within Arabic or whether it belongs to a broader pattern already available in the inherited lexicon. This is especially relevant for loanwords, which often appear in Arabic lexicography as semantically isolated items within a formal root environment. They may be assigned to an Arabic root for organisational purposes, but they do not normally generate the full range of derivations, semantic extensions, and idiomatic uses expected of inherited roots. Their meanings instead tend to remain attached to a specific object, plant, material, institution, or technical term.

Table 2. Roots similar to SRDQ by literal configuration.

SRD-B Arabicised ‘an underground room’ (cf. Syr. srdʾb ‘underground cold chamber’ < Pers. sardāb lit. ‘cold water’, a kind of cellar to retreat to in the summer heat; see Hejazi and Saradj 2015, pp. 151–52)	Ǧ-RDQ jardaqa ‘piece of bread’ Arabicised (<Pers.), var. ǦRḎQ (cf. Syr. grdg ‘unleavened bread, sprinkled with oil’ < MPers. girdag ‘disk, round’; see Ciancaglini 2008, p. 146)
SRD-Ǧ ʾahmala ‘to omit’	Ḫ-RDQ khurdīq ‘broth’ < Pers. khūrdīk (< MPers. xwrdyg ‘food’)
SRD-Ḥ sirdāḥ ‘big, tall, huge, etc.’	D-RDQ dardaq ‘small; child’ (<Aram. DRDQ by dissimilation from DQDQ; cf. Akk. daqqu ‘small (child)’
SR-M-Q sarmaq ‘a kind of plant’ (‘garden orache’; cf. Syr. srmg ‘orach’ < MPers *sarmag; see Ciancaglini 2008, p. 227)	Z-RDQ zardaq ‘elongated thread; queue of people; a row of palm trees’ [?] < Pers. zardeh (cf. Syr. zrdq ‘safflower’ < MPers. zardak; see Ciancaglini 2008, pp. 74–75; pace Asbaghi 1988, p. 143 zardag (!) which means ‘yolk’)
S-N-DQ < ṢNDQ ṣundūq ‘box’	Ġ-RDQ ghardaq or gharqad bot. Nitraria, ghardaqa ‘enwrapping [with darkness, dust, veil]’

Abbreviations used for languages: Akk. = Akkadian; Aram. = Aramaic; MPers. = Middle Persian; Pers. = Persian; Syr. = Syriac.

and

Table 3. Phonologically similar roots to SRDQ by place and manner of articulation.

C1	C2	C3	C4	Root	Comment
F	T	P	P	SRDQ	surādiq cf. Mand. srdq ‘curtain; canopy, pavilion’ < MPers. srād or srāy ‘house, hall’ (Schmitt 1986)
F	N	P	P	SNDQ	sundūq, cf. Syr. and Mand. sndwq ‘box’ (from Arabic; CAL n.d., s.v. sndwq); var. of ṢNDQ
F	T	P	P	ZRDQ	zardaq cf. Syr. zrdq ‘safflower’ < (early) Pers. zardak ‘safflower’ (MacKenzie 1986, p. 120)
F	T	N	F	ZRNḪ	zirnīkh, cf. Aram. zrnyk ‘arsenic’ < OP *zarniyaka (cf. Ciancaglini 2008, pp. 74, 178; > Gk. arsenikón; see Beekes 2010, p. 141)
F	T	N	P	ZRNQ	zarnūq cf. Aram. zrnwq ‘water bag; irrigation device’ < Akk. zuruqqu (CAL n.d., s.v. zrnwq; cf. Kaufman 1974, p. 114; Borg 2011, p. 26)
F	N	P	P	ZNDQ	zindīq, cf. Mand. zndyq Manichean (?) < Parth. zandīk (Widengren 1960, p. 104)

Abbreviations used for languages: Akk. = Akkadian; Aram. = Aramaic; Gk. = Greek; Mand. = Mandaic; MPers. = Middle Persian; Parth. = Parthian; OP. = Old Persian; Syr. = Syriac.

, are not meant to prove etymology by statistics. They reproduce the comparative material used to test whether roots shaped like SRDQ belong to inherited Arabic root patterns or whether they cluster with loanwords, technical terms, and marginal lexemes. The results are used only as supporting evidence, alongside semantics and external comparanda. The most similar roots, defined as those matching SRDQ in three consonantal positions, are as follows:

Five of these ten words with consonantal configurations close to surādiq are loanwords or Arabicised forms, a distribution that supports treating the root-neighbourhood of SRDQ as phonotactically marked rather than ordinary.3 The lemmas sarmaq and ghardaq are plant names and as such less probative, since botanical vocabulary is often lexically unstable and may preserve regional or specialised terms. The entry sardaja is also uncertain, since it may result from a misunderstanding of the lexicographical label ʾuhmila ‘left unused/unattested’, as though this label represented a lexical item connected with the verb ʾahmala ‘to omit, neglect’. This leaves sirdāḥ, ṣundūq, and ghardaq as the least clearly foreign comparanda, although ṣundūq/sandūq is related to Syriac forms (CAL n.d., s.v. sndwq/ṣndwq).4 The absence of several of these roots from the Kitāb al-ʿAyn (SRDB, ǦRDQ, and ḪRDQ) may reflect rarity, foreign origin, omission, or limits of early lexicography.

Among these forms, the closest match to surādiq in terms of both place and manner of articulation is zardaq, which differs chiefly in the voicing of the initial sibilant. Like surādiq, zardaq also appears to point beyond inherited Arabic if the gloss ‘safflower’ is taken as primary, since it can be compared with Syriac zrdq ‘safflower’ and Middle Persian/New Persian zardak (MacKenzie 1986, p. 120; Ciancaglini 2008, pp. 74–75). The other meanings recorded in Lisān al-ʿArab—‘elongated thread’, ‘queue of people’, and ‘row of palm trees’—should therefore be treated as separate or secondary semantic developments rather than as evidence against the Iranian comparison (Ibn Manẓūr 1970, s.r. ZRDQ).

Table 3 generalises the analysis from literal consonantal similarity to phonological similarity by place and manner of articulation (Watson 2002, p. 13). It compares roots in which the first consonant is a [DA] dentoalveolar [F] fricative (/s/, /z/); the second is a [DA] dentoalveolar [T] tap (/r/), [L] lateral (/l/), or [N] nasal (/n/); the third is a [DA] dentoalveolar [P] plosive (/d/), [N] nasal (/n/), [L] lateral (/l/), or [T] tap (/r/); and the fourth is an [U] uvular [P] plosive (/q/) or [F] fricative (/ḫ/ [-voice], /ġ/ [+voice]).5 The heatmaps below visualise a stepwise reduction of the 2551 quadriliteral roots in the lisan345 root dataset in R, filtering roots according to the broader place-and-manner articulatory profile represented by SRDQ (see Figure 1).

This second level of comparison is methodologically important because it tests not only whether roots share the same consonants, but whether they occupy the same phonological neighbourhood.6

The tables should be read as comparative evidence rather than mechanical proof. The concentration of loanwords and Arabicised forms among roots with a phonological structure similar to SRDQ strengthens the hypothesis that surādiq is not an inherited Arabic formation. The comparable roots largely belong to contact vocabulary or marginal lexical strata. This internal Arabic evidence is reinforced by the occurrence of cognate or closely related forms in Syriac and Mandaic, where srdqʾ denotes a ‘curtain’, ‘canopy’, or ‘pavilion’ (CAL n.d., s.v. srdq; Drower and Macuch 1963, p. 336).

The Iranian origin of surādiq has long been recognised, although the precise source form has been debated. Earlier proposals connected surādiq with Persian *sar-otāq ‘ante-chamber’ or sarā-pardeh ‘tent’ (Damghani 1999), or sorādeq ‘tent-cover, curtain’, ultimately related to Middle Persian srādag ‘hall’ (Fraenkel 1886, p. 29; Nöldeke 1875, p. XXXI; Jeffery 1938, p. 167; Yarshater 1998, p. 53). These explanations, however, are problematic. The derivation from sar-otāq is unlikely if Persian otāq itself reflects a later Turkic loan, since Proto-Turkic *otag ‘tent, house’ appears to have entered Persian through Chagatai (Doerfer 1963–1975, vol. 2, p. 66; EDAL 2003, vol. 3, pp. 1069–70). The derivation from sarā-pardeh is also difficult, since Persian /p/ is normally represented in Arabic by /f/ or /b/ (Siddiqi 1919, p. 71), whereas surādiq contains neither; this objection had already led Nöldeke to reject this proposal, but Pietruschka still gives sarā-pardeh as the source of Qurʾānic surādiq (Nöldeke 1875, p. XXXI, n. 3; Pietruschka 2006, p. 236).

Cheung’s explanation is therefore preferable. He suggests that Qurʾānic surādiq may have reached Arabic through Mandaic sradqa ‘awning, tent-cover’, from an unattested Parthian form *srāδak/srādag (Cheung 2017, p. 328). This reconstruction fits both the Aramaic/Mandaic evidence and the Iranian semantic field. The form is related to Zoroastrian Middle Persian srād or srāy ‘house, hall’, ultimately deriving from Proto-Iranian srāda- (Schmitt 1986), and may be associated with the verb srāyidan ‘to protect’ (Avestan θrāya-; MacKenzie 1986, p. 76; cf. Proto-Indo-European *k̂ēls ‘(store)room’ and *k̂el- ‘to protect, conceal’; see Volpe et al. 2006, p. 282b). On this basis, the Greek derivation proposed by Corriente, from stratḗgion ‘the general’s tent’, is unnecessary and considerably less plausible (Corriente 1997, p. 248a, s.r. SRDQ). The most economical explanation is that surādiq entered Arabic through Iranian–Aramaic or Iranian–Mandaic mediation and was then fully integrated into Qurʾānic Arabic as an eschatological term for an enclosing structure.

For the present article, the significance of surādiq is not only etymological. It shows that Qurʾānic vocabulary may carry traces of multilingual transmission while functioning fully inside Qurʾānic rhetoric. A term with possible Iranian and Aramaic/Mandaic mediation becomes part of Arabic eschatological language. The case therefore belongs equally to Arabic historical lexicography, Iranian–Aramaic contact, and the religious semantics of the Qurʾān. Surādiq is the strongest positive case in the article, because internal Arabic phonotactics, Iranian comparanda, and Aramaic/Mandaic mediation converge.

3.2. Documentary Semantics and Eschatological Reckoning: Qiṭṭ

The next two cases show why a multilingual framework must consider semantic domains as well as word forms. Qiṭṭ and ḥiṭṭah are not simply possible loans; they show how documentary and religious-formulaic semantics can be recontextualised within Qurʾānic discourse.

The hapax legomenon qiṭṭ appears in Q 38:16 (Elmaz 2011, pp. 145–52), in the demand of the unbelievers:

وَقَالُوا رَبَّنَا عَجِّل لَّنَا قِطَّنَا قَبْلَ يَوْمِ الْحِسَابِ

rabbanā ʿajjil lanā qiṭṭanā qabla yawmi l-ḥisāb ‘Our Lord, hasten for us our share [of punishment] before the Day of Reckoning’. The word is plausibly an old loan related to Jewish Aramaic giṭṭā and Rabbinic Hebrew gēṭ ‘writing, document’, which acquired the specialised meaning ‘bill of divorce’ in Mishnaic usage (Jeffery 1938, p. 241; cf. Biblical Hebrew sēp̄ɛr kərīṯuṯ in Deut 24:1). The Aramaic word itself is connected to Akkadian giṭṭu(m) ‘longish tablet, receipt, document’, a Sumerian loanword used with determinatives indicating writing materials such as clay (IM) or parchment (KUŠ; AHw 1965–1981, vol. 1, p. 294; CAD 1956–2010, vol. G, p. 112a; Maraqten 1998, pp. 307–8).

The relevance of qiṭṭ to this article is that documentary vocabulary enters a Qurʾānic eschatological-polemical context. A term associated with written allotment, document, or legal record becomes part of a religious demand for one’s ‘share’ before the Day of Reckoning. This is an example of semantic transfer: documentary semantics move into a scriptural discourse of accountability.

The phonotactic profile of qiṭṭ strengthens the comparison with older documentary vocabulary. Arabic lexicographical tradition treats certain consonantal combinations as incompatible with genuinely Arabic root structure; in particular, roots in which /ṭ/ co-occurs with /j/, /t/, or /ṣ/ are excluded by the incompatibility rules summarised by al-Jawālīqī (al-Jawālīqī 1990, pp. 22–23). In the Kitāb al-ʿAyn, other triliteral roots with geminated /ṭ/ are attested after C₁ = {ʾ, B, Ṯ, Ḥ, Ḫ, Z, š, ʿ, Ġ, Q, L, M}, but comparable nominal formations are limited. When these roots are generalised by place of articulation, the Lisān al-ʿArab yields 142 phonologically similar roots; when they are compared by manner of articulation, however, only two close parallels remain: ʾṬṬ, an onomatopoeic root meaning ‘to creak’ or ‘to moan’, and BṬṬ, which gives baṭṭa ‘to slit a wound; squeeze a boil’ and baṭṭ ‘duck’ (al-Farāhīdī 1980–1985, vol. 7, pp. 408, 470).7

The latter is unlikely to represent a productive inherited Arabic root in the relevant sense. The noun baṭṭ ‘duck’ is commonly treated as a cultural loanword or Kulturwort, probably related to Persian bat, although wider comparisons with possibly onomatopoeic Indo-European *pad- ‘duck, teal’ have also been proposed (Boeder 1993, p. 92; Greppin 2006, p. 171).8 Its cultural circulation is supported by Iranian material evidence, including pre-Achaemenid duck-shaped wine vessels (‘wine ducks’), and by the later explanation of the short-necked lute barbaṭ, which al-Kh^wārazmī derives from bar ‘chest’ and baṭṭ ‘duck’ because the lute’s profile resembles a duck (Melikian-Chirvani 1995, pp. 47–48; During 1988). Thus, the small set ʾṬṬ, BṬṬ, and QṬṬ does not point to a robust native Arabic pattern combining a plosive with geminated /ṭ/. Rather, it consists of an onomatopoeic root and forms associated with loanwords. This makes Qurʾānic qiṭṭ a strong candidate for older Semitic documentary transmission rather than an ordinary inherited Arabic noun of the root QṬṬ, the base stem verb of which qaṭṭa is equated with qaṭaʿa by al-Farāhīdī, suggesting QṬṬ is a geminate root semantically related to QṬʿ (al-Farāhīdī 1980–1985, vol. 5, p. 15). In qiṭṭ, therefore, the evidence converges around documentary semantics: a term associated with writing, allotment, and legal record is recontextualised in an eschatological demand for one’s share before the Day of Reckoning. It shows that lexical contact may involve documentary and legal semantics, not only material culture or theological vocabulary. A documentary/legal word can be resemanticised within Qurʾānic eschatology, where ‘share’, ‘document’, and ‘reckoning’ overlap.

3.3. Formulaic Remission and Semitic Religious Contact: Ḥiṭṭah

A related but more complex case is ḥiṭṭah, which is the closest match to qiṭṭ in the Qurʾān, but its semantic field is not documentary. It belongs instead to a formulaic context of entry, humility, forgiveness, and remission. Among the other roots in the Qurʾān with a geminated /ṭ/, there are ḤṬṬ, represented by ḥiṭṭah, ḪṬṬ, represented by khaṭṭa, and ŠṬṬ, represented by ʾashaṭṭa and shaṭaṭ. The formal similarity between ḥiṭṭah and qiṭṭ has contributed to the identification of ḥiṭṭah as a possible loanword (cf. Jeffery 1938, p. 110). The case of ḥiṭṭah, however, is more complex, and the Qurʾānic occurrences of ḥiṭṭah and qiṭṭ may be linked intertextually through phonetic association (Elmaz 2011, pp. 151–52). The word ḥiṭṭah occurs in Q 2:58 and Q 7:161 in a formulaic context of entry, humility, and forgiveness, where the Israelites are commanded to enter the gate and say ḥiṭṭah in order to be forgiven:

وَادْخُلُوا الْبَابَ سُجَّدًا وَقُولُوا حِطَّةٌ نَّغْفِرْ لَكُمْ خَطَايَاكُمْ

wa-dkhulū l-bāba sujjadan wa-qūlū ḥiṭṭatun naghfir lakum khaṭāyākum ‘and enter the gate humbly (sujjadan) and say: “ḥiṭṭa^tun” for Us to forgive your sins (khaṭāyā)’ (Q 2:58).

وَقُولُوا حِطَّةٌ وَادْخُلُوا الْبَابَ سُجَّدًا نَّغْفِرْ لَكُمْ خَطِيئَاتِكُمْ

wa-qūlū ḥiṭṭatun wa-dkhulū l-bāba sujjadan naghfir lakum khaṭīʾātikum ‘and say “ḥiṭṭa^tun” and enter the gate humbly (sujjadan) for Us to forgive your sins (khaṭīʾāt)’ (Q 7:161).

An internal Arabic derivation remains possible and should not be dismissed. The form may be connected with Arabic ḤṬṬ ‘to remove, set down, relieve a burden’, in which case ḥiṭṭah could be understood as a nominal expression of remission or removal. Formally, it may be explained as a fiʿlah noun, perhaps as a shortened or analogically reshaped form of faʿīlah or iftiʿāl; al-Farāhīdī’s discussion of ḫiṭṭah in relation to iḫṭiṭāṭ provides a relevant internal model for such analogical analysis (al-Farāhīdī 1980–1985, vol. 4, p. 137). In this reading, ḥiṭṭah would mean something like ‘remission’, ‘relief’, or ‘removal [of sin]’, which fits the Qurʾānic context.

Nevertheless, the word has often been connected with the wider Semitic root ḤṬʾ ‘to sin’. Bellamy, for example, proposed emending Qurʾānic ḥiṭṭah in Q 2:58 and 7:161 to Hijazi Arabic khiṭṭah (<khiṭʾah ‘sin’; see Bellamy 1993, pp. 566–67). This proposal attempts to align the Qurʾānic form with Hebrew and related Aramaic material. The word has also been associated with Nabataean and Syriac forms such as ḥṭīṯā ‘sin, recompense for sin’ and ḥəṭāh ‘sin, sin offering, punishment for sinning’, analysed as back-formation from the emphatic ḥiṭʾā (CAL n.d., s.v. ḥṭh, ḥṭhˀ). Akkadian ḫīṭu/ḫiṭṭu, and ḫiṭītu ‘sin, offence’ (CAD 1956–2010, vol. Ḫ, p. 210a) provide a further Semitic comparandum, although they cannot be treated as the direct source of the Qurʾānic form.

The evidence therefore points in two directions. On the one hand, ḥiṭṭah can be interpreted internally through Arabic ḤṬṬ, especially in view of the Qurʾānic context of forgiveness and relief. On the other hand, the religious semantics of ‘sin’, ‘confession’, and ‘remission’ in Hebrew, Aramaic, Syriac, Nabataean, and Akkadian provide a plausible contact background for the Qurʾānic use. It is therefore preferable not to describe ḥiṭṭah simply as a secure loanword, but as a lexeme whose Qurʾānic meaning may have been shaped by both Arabic morphology and broader Semitic religious terminology.

The only poetic citation for Arabic ḤṬṬ in an explicitly religious context appears to be an anonymous half-verse transmitted in the Kitāb al-ʿAyn as: wa-ḥṭuṭ ilāhī bi-faḍlin minka ʾawzārī ‘and, my God, by Your favour, remove from me my burdens’. The evidentiary value of this citation is limited, since it is both anonymous and incomplete. At the same time, its diction is strikingly close to Qurʾānic phraseology. The expression of bearing or removing ʾawzār ‘burdens [of sin]’ recalls Qurʾānic usage in passages such as Q 6:31, 16:25 and 20:87. This suggests that the half-verse may reflect an early Islamic religious register rather than pre-Islamic poetic usage. If so, it does not provide independent pre-Qurʾānic evidence for ḥiṭṭah but rather shows how ḤṬṬ could be integrated into an Arabic idiom of forgiveness, remission, and the removal of sin in early Islamic or post-Qurʾānic usage. It is best treated as a borderline case: it can be explained internally through Arabic, but its Qurʾānic use also resonates with broader Semitic religious terminology.

3.4. The Shared Phonotactic Sequence -iṭṭ-

The preceding two cases also share a phonotactic problem: the sequence -iṭṭ-. As with other phonotactically marked sequences, for instance, a /r/ following a /n/ (al-Farāhīdī 1980–1985, vol. 5, p. 265; vol. 1, p. 53), the occurrence of -iṭṭ- in nouns may suggest foreign or contact origin.9 Apart from qiṭṭ (al-Farāhīdī 1980–1985, vol. 5, p. 15) and ḥiṭṭah (vol. 3, p. 19), the Kitāb al-ʿAyn records several relevant items containing the sequence -iṭṭ-:

First, biṭṭīkh ‘watermelon’10 (see Behnstedt and Woidich 2011, p. 511) is a clear botanical term with external comparanda. It may be compared with Aramaic baṭṭīḥ and Biblical Hebrew ʾăbaṭṭīḥ in Num 11:5; Paris traces the latter ultimately to Egyptian (Paris 2015, p. 138).11 The metathetic form ṭibbīkh, reported for the Hijaz (al-Farāhīdī 1980–1985, vol. 4, p. 225), shows internal reshaping of the same botanical term. It may also be compared typologically with Greek [mēlo]pépōn, from mêlon ‘apple, tree-fruit’ and pépōn ‘ripe, cooked’. Arabic faqqūṣ ‘unripe melon’ is still known in Egypt (see al-Farāhīdī 1980–1985, vol. 5, p. 67), Western Arabic, and Maltese (see Behnstedt and Woidich 2011, p. 525; Agius 1996, p. 187).

Second, qiṭṭ/qiṭṭah ‘cat’ belongs to a widespread group of similar cat-names across languages (cf. Behnstedt and Woidich 2011, pp. 325–29). Similar names for cats occur in many languages and may be partly onomatopoeic or expressive, rather than straightforward loanwords (Lühr 2014, pp. 434–35).

Third, niṭṭīs and niṭāsī, glossed as ‘physician’, are explicitly treated by al-Khalīl as foreign. Kitāb al-ʿAyn explains the word as al-ʿālim bi-l-ṭibb, ‘one learned in medicine’, and adds that in Greek it is al-nisṭās (al-Farāhīdī 1980–1985, vol. 5, p. 215). This is plausibly connected with Greek gnṓstēs ‘one who knows, expert’, although the phonological route is not entirely transparent. The Arabic representations record lexicographical variation in foreign terms, and the form niṭṭīs may reflect Arabicising reshaping of a foreign cluster, perhaps through assimilation of /sṭ/ to geminate /ṭṭ/ and subsequent vowel reshaping.

Fourth, ghiṭṭah, glossed as manj, is more uncertain. If manj is connected with Persian mang/bang ‘cannabis, hemp, narcotic’, the item would belong to the vocabulary of drugs and materia medica, a field in which borrowing is frequent (Flattery and Schwartz 1989, pp. 14–18, 124–28). Yet the form ghiṭṭah itself is textually insecure and seems to be a textual corruption. Some manuscripts of Kitāb al-ʿAyn spell the word as <ʿd> (al-Farāhīdī 1980–1985, vol. 6, p. 155, n. 2), and it does not appear to be securely continued in later lexicography (Ibn Manẓūr 1970, s.r. ĠṬṬ; al-Azharī 1964–1976, vol. 11, p. 130, s.r. MNJ; al-ʿUbaydī 1975, pp. 48–49, s.r. ĠṬ).

Taken together, the lexical neighbourhood of -iṭṭ- is suggestive but not decisive. The argument is not that -iṭṭ- automatically proves borrowing but, rather, it identifies an environment in which old contact terms, culture words, and lexicographical anomalies cluster. This cluster illustrates the article’s broader method. Several items containing this sequence are demonstrably or plausibly non-native, especially biṭṭīkh and niṭṭīs; others, such as qiṭṭ(ah) (cf. Krebernik 2008, p. 262), are more ambiguous because expressive or onomatopoeic explanations remain possible; and ghiṭṭah is too textually uncertain to carry much weight. The evidence therefore supports a cautious methodological point: nouns with -iṭṭ- in early Arabic deserve external comparison, but the sequence alone cannot prove borrowing. Applied to ḥiṭṭah, this means that its formal similarity to other potentially foreign items strengthens the case for investigating Aramaic and broader Semitic comparanda, but it does not by itself override the possibility of an internal Arabic derivation from ḤṬṬ.

The sequence -iṭṭ- also has broader diagnostic value. A Qurʾānic hapax such as qiṭṭ might be easy to overlook, but when its phonotactic profile, documentary semantics, Aramaic parallels, and Akkadian background are considered together, it becomes evidence for an older layer of written culture that could be reactivated in Qurʾānic religious rhetoric. Ḥiṭṭah extends the same logic to religious formulae: a word or phrase associated with burden-removal, sin, confession, and divine forgiveness. If it is an Aramaism, it illustrates the movement of religious semantics into Qurʾānic Arabic. If it is an internal Arabic formation, it still participates in a broader Semitic semantic field of sin, burden, and relief. In either case, both words belong to a multilingual history of religious speech.

3.5. Lexical Archaism and Exegetical Normalisation: Fūm

Unlike surādiq, qiṭṭ, and ḥiṭṭah, fūm is not primarily a case of a striking non-Arabic comparandum, and it adds a different kind of evidence to the article. It is a case in which Qurʾānic context, Biblical comparison, early Arabic glossing, non-canonical reading tradition, and Semitic etymology pull the interpretation in different directions. It therefore illustrates a central principle of muqārana, namely that comparison must be layered rather than selective.

The case also illustrates exegetical normalisation, understood here as the process by which interpreters align a difficult or archaic Qurʾānic word with more familiar vocabulary, contextual expectations, or scriptural parallels. In Q 2:61, the early grain/bread glosses may preserve an older Arabic sense of fūm. In contrast, the garlic interpretation normalises the word toward thūm, encouraged by the surrounding food list and the non-canonical reading wa-thūmihā:

فَادْعُ لَنَا رَبَّكَ يُخْرِجْ لَنَا مِمَّا تُنبِتُ الْأَرْضُ مِن بَقْلِهَا وَقِثَّائِهَا وَفُومِهَا وَعَدَسِهَا وَبَصَلِهَا

fa-dʿu lanā rabbaka yakhruj lanā mimmā tunbitu l-ʾarḍu min baqlihā wa-qiththāʾihā wa-fūmihā wa-ʿadasihā wa-baṣalihā ‘Call upon your Lord to bring forth for us from what the earth grows, from its beans/greens (baql), cucumbers (qiththāʾ), fūm, lentils (ʿadas), and onions (baṣal)’.

Lexical lists require special caution because readers compare not only individual words but whole scenes. In the case of fūm, the Biblical parallel in Num. 11:5 and the non-canonical reading wa-thūmihā encourage the interpretation ‘garlic’, while early Arabic lexicography preserves a grain or bread meaning. The analysis must therefore distinguish lexical correspondence, narrative correspondence, and exegetical normalisation.

Q 2:61 lists the produce requested by the Israelites: min baqlihā wa-qiththāʾihā wa-fūmihā wa-ʿadasihā wa-baṣalihā, usually translated as ‘its herbs, cucumbers, garlic, lentils, and onions’ (Abdel Haleem 2005, p. 8). The usual modern translation ‘garlic’ is encouraged by the immediate context, especially the neighbouring baṣal ‘onion’, and by comparison with Num. 11:5, where the Israelites remember Egyptian fish, cucumbers, melons, leeks, onions, and garlic. Paret therefore pointed to Num. 11:5 as a relevant Biblical parallel (Paret 1971, p. 220), yet the parallel is not exact as the Biblical list contains—in addition to fish (dāgāh)—cucumbers (qiššuʾim), melons (ʾăḇaṭṭiḥīm), leeks (ḥāṣīr; also understood more generally as herbage or greens), onions (bəṣālīm), and garlic (šūmīm).

Yet, Egypt is also praised in the Bible as a ‘granary in times of famine’ (Gen. 12:10; 41:54; Jos. Ant. XV, 9, 2). In Exod. 16:1–3, two and a half months after the departure from Egypt, remembering the meat-pots and bread, the Israelites wish that they had died by the hand of the Lord. A no less relevant comparison for understanding fūm as grain appears in Ezek. 4:9, which gives a recipe for bread made from wheat, barley, beans, lentils, millet, and spelt (ḥiṭṭîn ûśəʿōrîm ûp̄ôl waʿăḏāšîm wəḏōḥan wəḵussəmîm). The grain in Q 2:61 may correspond to fūm, the beans to baql, and the lentils to ʿadas. The comparison with both lists, Num. 11:5 and Ezek. 4:9, is suggestive, but it does not decide the meaning of fūm.

The oldest Arabic evidence points strongly toward a grain or bread meaning. Al-Khalīl glosses fūm as ḥinṭah ‘wheat’, and Abū ʿUbayda explains it as ḥinṭah or khubz ‘bread’ (al-Farāhīdī 1980–1985, vol. 8, p. 405; Abū ʿUbayda 1954–1962, vol. 1, p. 41). Mujāhid glosses it as khubz, and Zayd b. ʿAlī gives ḥinṭah, with fūmah as the unit noun (Mujāhid b. Jabr 1989, p. 83; Zayd b. ʿAlī 1997, p. 129). The exegetical section of Ṣaḥīḥ al-Bukhārī preserves the broader explanation that ‘all edible grains are fūm’ (al-ḥubūb allatī tuʾkal kulluhā fūm) (al-Bukhārī 2001, kitāb al-tafsīr, sūrat al-Baqara). This early Arabic strand is coherent: fūm belongs to the semantic field of grain, wheat, and bread. Al-Farrāʾ says that fūm is an archaic word (lughah qadīmah) meaning ḥinṭah ‘wheat’ and ḫubz ‘bread’, and he points to the verb fawwama ‘to bake bread’ (al-Farrāʾ 1983, vol. 1, p. 41).

The competing interpretation ‘garlic’ enters through the reading tradition and contextual reasoning for harmonisation. Al-Farrāʾ reports that ʿAbdullāh [b. Masʿūd] read wa-thūmihā, and Ibn Qutayba gives both meanings but favours ‘garlic’ because of the reading attributed to Ibn Masʿūd and the attested substitution (ʾibdāl) of fāʾ and thāʾ, for instance, in ǧadath > ǧadaf ‘grave, tomb’ and maghāthīr > maghāfīr ‘a certain gum/resin’ (Ibn Qutayba 1978, p. 51). Al-Khaṭīb lists Ibn Masʿūd, Ibn ʿAbbās, and Ubayy among those associated with the reading wa-thūmihā (al-Khaṭīb 2002, vol. 1, p. 112). The evidence is therefore not a simple opposition between ‘wheat’ and ‘garlic’; it is a debate about how old Arabic usage, context, and variant reading should be weighted.

The garlic interpretation also faces a comparative difficulty. Arabic thūm ‘garlic’ has the expected cognates with initial š or t: Syriac tūmā, Hebrew šūm, and Akkadian šūmum (Zammit 2002, p. 608). If fūm means ‘garlic’, it is either a dialectal variant of thūm, or a form whose interpretation has been reshaped by the non-canonical wa-thūmihā reading. By contrast, the grain/bread interpretation fits the old Arabic verb fawwama ‘to bake bread’ and may connect, at least tentatively, with the wider Semitic baking root ʾP(Y)/WP(Y), including Akkadian epû(m) ‘to bake/cook’ and Arabic mīfā ‘oven’ (Del Olmo Lete 2004, p. 62).

A cautious but phonologically uncertain comparison may be made with Akkadian upumtum/upuntu, a type of flour also used for offerings (AHw 1965–1981, vol. 3, p. 1426). The comparison is semantically attractive because upumtum/upuntu belongs to the flour/offering lexicon. Kogan and Krebernik’s classification of Akkadian vocabulary into inherited Semitic material, non-trivial retentions, innovations, loanwords, and obscure terms is useful here because fūm need not be forced into a simple binary of ‘Arabic’ or ‘borrowed’; it may instead represent an archaic or obscure cereal term whose later interpretation was reshaped by scriptural context and variant-reading tradition (Kogan and Krebernik 2020, pp. 4–6, 19–21). The final -tu(m) of upumtum/upuntu could in principle have been lost or reanalysed during transmission, since Akkadian nouns borrowed into West Semitic are not necessarily preserved with their Akkadian final morphology intact. A useful analogy is Akkadian bīrtu ‘fort, citadel’, which was borrowed into Hebrew bīrā ‘citadel, acropolis’ and into Aramaic/Syriac forms such as Official Aramaic byrtʾ, Jewish Babylonian Aramaic bīrtā, and Syriac birtā (Kogan and Krebernik 2020, p. 98). This example shows that Akkadian final case/mimation is not retained, and that the feminine -t- may be preserved in some West Semitic forms but is absent in Hebrew bīrā. However, this comparison remains phonologically uncertain: the development of Akkadian u-pu- into Arabic fū- lacks a secure parallel. In this interpretation, fūm could preserve an archaic derivative of the baking/flour lexicon, or it could have been secondarily reanalysed as an Arabic F-W-M root, giving rise to fawwama ‘to bake bread’. A final -m could theoretically reflect fossilised mimation or a nominal augment, but this remains hypothetical for the baking root; unlike the -m in Aramaic pūmā and Arabic fam- ‘mouth’, it is not independently established here (Kogan and Krebernik 2020, p. 76). The safest conclusion is therefore that fūm is probably an archaic Arabic cereal/bread term, while Akkadian upumtum/upuntu and the Semitic baking root ʾ/W-P-Y provide useful comparative context rather than a proven etymology.

Fūm is therefore best described as a potential archaism within a Semitic flour/bread lexical field. Its earliest Arabic glosses make ‘wheat’, ‘bread’, or ‘grain’ the most plausible primary meaning, while ‘garlic’ reflects contextual pressure from the neighbouring baṣal ‘onion’ in Q 2:61, a partial Biblical parallel, and a non-canonical reading tradition. The case is methodologically important because it shows that multilingual comparison may obscure as well as illuminate. If the Biblical list in Num. 11:5 is allowed to dominate the interpretation, fūm is assimilated to garlic. If the early Arabic glosses and the verb fawwama are given their full weight, the word becomes an archaism in the Qurʾānic food lexicon that later exegetes attempted to normalise.

For the theory of muqārana, fūm shows how early Arabic glossing, variant reading, and Biblical comparison can pull interpretation in different directions. Comparison must therefore include disagreement within Arabic tradition itself. Lexical meaning is not established by internal gloss alone, nor by external parallel alone, nor by variant reading alone. It emerges from the controlled comparison of all three. This makes fūm a useful bridge between the strongly contact-oriented cases above and the control case of yaqṭīn below. Like yaqṭīn, fūm resists an over-simple Biblical derivation, but unlike yaqṭīn, it also shows how variant reading and semantic context can generate a genuine interpretive split.

3.6. Narrative Parallelism Without Borrowing: Yaqṭīn

This final case functions as a control case for the method. It tests whether the same comparative procedure that identifies contact in surādiq, qiṭṭ, ḥiṭṭah, and fūm can also reject an attractive borrowing hypothesis when the internal Arabic evidence is stronger.

In the Qurʾānic passage about Jonah in Q 37:139–148, God causes a shajaratan min yaqṭīnin to grow over Jonah after he is cast out of the whale, sick, onto the barren shore:

وَأَنبَتْنَا عَلَيْهِ شَجَرَةً مِّن يَقْطِينٍ

wa-ʾanbatnā ʿalayhi shajaratan min yaqṭīnⁱⁿ ‘and We made a shajara of yaqṭīn grow over him’.

The word shajarah is not botanically precise in this context. It may be translated as ‘tree’, ‘plant’, or ‘vegetation’, depending on how the translator understands the relation between the noun and the following min yaqṭīn. The plant is mentioned only once, but it performs an important narrative function. Jonah has been cast onto barren ground while sick, and God causes a plant to grow over him: wa-ʾanbatnā ʿalayhi shajaratan min yaqṭīn (Q 37:146). Modern translations therefore differ. Abdel Haleem renders the expression as ‘a gourd tree’, while other translators use ‘gourd plant’, ‘gourd vine’, ‘a spreading plant of the gourd kind’, or a more general plant of edible fruit (Abdel Haleem 2005, p. 288). These variations do not reflect modern uncertainty; they reflect the broader problem of translating a scriptural plant whose exact species is less important than its protective function.

The Biblical Jonah story provides the obvious comparandum, but it does not provide a simple lexical solution. Before Jeffery, Torrey tried to link yaqṭīn to Hebrew qīqāyōn in his Jewish Foundations of Islam: ‘It would be interesting to know in what way his curious word yaqṭīn, for Jonah’s gourd (37:146) is related to the Hebrewקִיקָיוֹן , and whether the new creation is in any way his own’ (Torrey 1933, p. 52). Hebrew qīqāyōn, however, occurs only in Jonah 4: 6–10 and has long been debated. The Septuagint renders it as Greek kolókyntha, usually understood as a gourd or colocynth, while the Vulgate uses Latin hedera, ‘ivy’. These translation choices show that the plant was already interpreted differently in ancient traditions. The Hebrew word has attracted scholarly interest throughout the ages. This can be illustrated by several dissertations and disputations from Germany and Sweden in the second half of the seventeenth century alone (Hupka 1653; Weidling 1684; Trautzel 1691). These examples support Moore and Sherwood’s observation about philologists’ obsession with details (Moore and Sherwood 2011, p. 30), while also serving as a warning against over-specification:

Then, there is our obsession with textual minutiae. We have long made our home in the kind of textual details that a traditionally minded literary critic would likely deem incidental or secondary, peripheral or tangential: the etymologies of the personal names in the Mari tablets; the probable geographical location of the land of Nod; the botanical identity of Jonah’s qiqayon plant; fragmentary funerary texts from Ugarit; shopping lists from Oxyrhynchus; Western non-interpolations in the New Testament manuscript tradition; hapax legomena in the Pastoral Epistles; the significance of locusts in the diet followed at Qumran—the list is infinitely long and ever more bizarre.

Modern Biblical scholarship has proposed identifying qīqāyōn as ‘gourd’, ‘ricinus’, ‘ivy’, ‘vine’, ‘bryonia’, or even a nonce-word coined to give the story an exotic flavour (Robinson 1985, pp. 390–403; LaCocque and LaCocque 1990, p. 156; Simon 1999, p. 42). The Hebrew term is therefore itself uncertain and cannot serve as a stable source form from which Arabic yaqṭīn must be explained.

This is a key methodological point: narrative equivalence is not lexical identity. The Qurʾān and the Bible can share a story about Jonah and a plant without sharing the same plant word. The same narrative role—shade, protection, divine provision, and the prophet’s vulnerability—can be fulfilled by different lexical choices in different languages. The evidence from Greek and Latin translation traditions makes this point especially clear, as even within the Biblical tradition, the plant is not lexicalised in a single stable way. It is therefore methodologically weak to infer that Qurʾānic yaqṭīn must be a distorted form of Hebrew qīqāyōn simply because both words occur in the Jonah narrative. While scriptural parallelism may show shared narrative tradition, it does not by itself prove lexical borrowing.

Jeffery’s explanation belongs to an older style of foreign-vocabulary studies in which ‘unusual’ Qurʾānic words were sometimes treated as traces of imperfect borrowing or faulty recollection. His suggestion that yaqṭīn was heard in oral recitation and reproduced from memory as a garbled form of Hebrew qīqāyōn depends on several assumptions: that the Qurʾānic word must derive from the Hebrew word; that the difference between the two forms can be explained as faulty memory; and that early Arabic morphology and exegetical understanding are secondary to that presumed transmission (Jeffery 1938, p. 292). Carter later repeated this characterisation in compressed form in the Blackwell Companion to the Qurʾān and sees yaqṭīn as part of a game with words (Carter 2006, p. 137), while Hebrew is given as the proposed origin of the word in the Dictionary of Qurʾanic Usage (Badawi and Haleem 2008, s.v. yaqṭīn). The problem with this explanation is not simply that it is polemically phrased, but that the Arabic evidence does not support it.

The comparison with Hebrew qīqāyōn remains useful if it is reframed. It shows that Qurʾānic vocabulary participates in a shared scriptural narrative without borrowing the exact lexical material of earlier scriptures. The Qurʾānic story invokes Jonah, illness, divine rescue, and vegetation; the Biblical story invokes Jonah, a booth, a plant, shade, a worm, and divine pedagogy. The lexical choices differ, and those differences are part of each text’s rhetorical economy. The Qurʾānic narrative does not dwell on the plant’s later destruction or on Jonah’s attachment to it. It compresses the plant into a sign of divine care and restoration.

If one begins from Arabic morphology rather than from Biblical comparison, the form is unusual but analysable. Morphologically, yaqṭīn is analysed as a rare yafʿīl form from the root QṬN. It is the only word of this root in the Qurʾān and occurs only once—a hapax. At the same time, it is the only Qurʾānic example of this pattern, and it is the rhyming word in Q 37:146.12 Rare morphology and hapax status can attract etymological suspicion. In the Qurʾān, many roots are represented by only one lexical item or by a very small number of lemmas, and their rarity does not automatically make them borrowings (Elmaz 2020, pp. 162–73; Toorawa 2011).

Patterns with a y-prefix are ancient forms (Barth 1894, 226§154), and Sibawayhi lists yaʿḍīd ‘a type of plant’, while Ibn Durayd adds yaʿqīd ‘thickened honey, a type of thickened dish’ and the place name yabrīn (Sībawayh 1988–1996, vol. 4, p. 265; Ibn Durayd 1987, pp. 659, 1216) as comparable forms.13 Much later, Barth adds yakhḍīr ‘green’ (Barth 1894, p. 232; cf. Fischer 1965, p. 191).

The phonotactic root structure of QṬN does not display any kind of incompatibility or phonotactic markedness that would encourage a borrowing analysis, either. There are numerous roots with this configuration of literals by place of articulation (142 in the Lisān al-ʿArab) and manner of articulation (6 in the Lisān al-ʿArab: {ʾ,B,Q}Ṭ{M,N}; cf. BṬN in the Qurʾān).14

Early Arabic lexicography treats yaqṭīn coherently. Abū ʿUbayda’s definition is particularly important because it gives a category rather than a species. He explains yaqṭīn as ‘trunkless (or: herbaceous) plants like gourds and squashes (dubbāʾ), colocynths (ḥanẓal), and watermelons (biṭṭīkh)’ (Abū ʿUbayda 1954–1962, vol. 2, p. 175). In modern botanical terminology, these examples point broadly to the Cucurbitaceae. Al-Farāhīdī explains qarʿ ‘gourd’ as the fruit of yaqṭīn, and Ibn Sīdah uses yaqṭīn as an umbrella term for cucurbits (al-Farāhīdī 1980–1985, vol. 1, p. 155; Ibn Sīdah 1898–1903, vol. 3, pp. 284–85). These explanations do not treat the word as obscure or foreign.

Early Qurʾānic exegesis supports the internal Arabic explanation found in Arabic lexicography. Mujāhid glosses yaqṭīn as a trunkless tree like a gourd and similar plants (Mujāhid b. Jabr 1989, p. 570). Al-Ḍaḥḥāk identifies it with qarʿ ‘gourd’ (al-Ḍaḥḥāk 1999, p. 712) and Zayd b. ʿAlī adds ‘any plant that does not stand on a trunk’ (kullu shajaratin lā taqūmu ʿalā sāqin; Zayd b. ʿAlī 1997, p. 345). Muqātil b. Sulaymān takes the same general view and gives further examples, defining yaqṭīn as an umbrella term for ‘everything that spreads out like a gourd, vine, cucumber, dodder and the like’ (kullu shayʾin yanbasiṭu mithlu l-qarʿi wa-l-karmi wa-l-qiththāʾi wa-l-kashūtā … wa-naḥwuhā).15 He also transmits the tradition that a waʿla ‘doe, ewe’ used to come to Jonah so that he could drink her milk (wa-kānat takhtalifu ilayhi waʿlatun fa-yashraba min labanihā; Muqātil b. Sulaymān 2002, vol. 3, p. 621). These examples are coherent and do not suggest that early exegetes regarded yaqṭīn as foreign; rather, they understand it as a generic term for spreading, shading, or cucurbit-like plants. The range of identifications is botanical, not etymological, and exegetes specify the plant’s function and type. They do not explain an incomprehensible foreign word, and as such, the early exegetical evidence weighs strongly against Jeffery’s claim. Yaqṭīn shows that narrative parallelism with Biblical material does not automatically imply lexical borrowing. It is essential to the article’s method because it resists a borrowing explanation.

4. Discussion

This article has proposed a multilingual and muqārana-based approach to Qurʾānic vocabulary. It argues that lexical comparison must be cumulative: etymology, morphology, phonotactics, Qurʾānic context, variant readings, translation history, and exegetical reception must be evaluated together. The five case studies show that Qurʾānic Arabic participates in multilingual lexical memory while also organising inherited, borrowed, archaic, and debated forms through Arabic grammar and religious interpretation. Earlier semantic associations, contact histories, scriptural echoes, and exegetical reinterpretations may remain active even after the word has become fully integrated into Qurʾānic Arabic.

The case studies support this conclusion in different ways. Surādiq is a strong case of mediated Arabicisation, probably involving Iranian and Aramaic/Syriac pathways. Qiṭṭ preserves a plausible older documentary Semitic layer, and ḥiṭṭah shows how religious formulae can be interpreted through both Arabic semantics and Semitic parallels. Fūm demonstrates that food vocabulary and Biblical comparison require particular caution: the most plausible primary Arabic meaning is ‘wheat’, ‘bread’, or ‘grain’, while ‘garlic’ reflects a later or parallel interpretive trajectory. Yaqṭīn shows that even a Qurʾānic word in an explicitly Biblical narrative can be morphologically and semantically Arabic without needing derivation from the Biblical lexeme. Without negative examples, such studies risk confirming their own assumptions: every rare word becomes foreign, every Biblical parallel becomes a source, and every semantic correspondence becomes borrowing. A method that can reject an attractive loanword explanation is stronger than one that only accumulates parallels.

The broader contribution of this article is methodological. Qurʾānic vocabulary should not be studied through a simple opposition between Arabic and foreign. It should be examined as part of a multilingual religious ecology in which Arabic, Aramaic/Syriac, Hebrew, Persian, and other traditions intersected in different ways. Muqārana—controlled comparison across forms, meanings, contexts, and receptions—offers a way to describe those intersections without collapsing them into one-dimensional loanword claims. It provides a way to connect modern comparative philology with premodern Islamic scholarly practice. Early exegetes and lexicographers compared words, glossed unusual forms, recorded semantic ranges, and located Qurʾānic expressions within Arabic usage. Modern scholarship can extend this comparative work through Semitic, Iranian, and Greek evidence among others, but it should not bypass the Arabic tradition. The strongest analysis emerges when external comparison and internal Arabic reception are read together.

The method also has limits. In several cases, the evidence is cumulative rather than demonstrative. Later Arabic lexicography may preserve earlier material through compilatory witnesses, but such evidence must be controlled through earlier lexical, exegetical, and contextual sources. External comparanda, in turn, may indicate shared Semitic inheritance, contact, or secondary convergence rather than direct borrowing. The conclusions offered here are therefore probabilistic and philological, not mechanical reconstructions of linear transmission.

The article’s main contribution is a method for evaluating Qurʾānic lexical history without reducing it to a choice between “Arabic” and “foreign”. The five cases show that lexical evidence can point to mediated Arabicisation, older Semitic vocabulary, semantic convergence, exegetical normalisation, or internal Arabic explanation. These outcomes are different, but they are best reached through the same controlled procedure. A wider application of this method to other Qurʾānic lexemes would make it possible to distinguish more clearly between secure Arabicisations, inherited Semitic vocabulary, contact-induced semantic shifts, and cases where external comparison should be rejected. This would move Qurʾānic lexical studies beyond catalogues of proposed loanwords toward a fuller account of multilingual lexical memory in the intellectual traditions of Islam.