The claim: "The Qur'an repeatedly presents itself as clear Arabic — 12:2, 26:195, 16:103 and 41:44 are the best-known places. 41:44 goes further: it says a foreign-language book would have been thrown back as an objection by Arabs. Yet the text's vocabulary contains words widely held to be non-Arabic in origin: istabraq (heavy brocade; 18:31, 44:53, 55:54, 76:21), sundus (fine silk; 18:31, 44:53, 76:21), firdaws (18:107, 23:11), zanjabil (76:17), salsabil (76:18), sijjil (11:82, 15:74, 105:4), qistas (17:35, 26:182)... These are traced to Persian, Aramaic, Ethiopic or Greek. If the book really is clear Arabic, why does it carry words from other languages? Either the claim is overstated, or the word 'Arabic' has lost its meaning here."
What does the Qur'an say?
First the verses the objection rests on, then the words it treats as evidence.
How often does the Qur'an call itself "Arabic"? In this platform's word-and-root data the root ʿ-r-b occurs 22 times across the Qur'an. Ten of those are the adjective ʿarabī describing the revelation itself: 12:2, 13:37, 16:103, 20:113, 26:195, 39:28, 41:3, 42:7, 43:3, 46:12. Ten more are aʿrāb, "the bedouins" (9:90, 9:97, 9:98, 9:99, 9:101, 9:120, 33:20, 48:11, 48:16, 49:14); one is "and an Arab" in 41:44, describing the addressee; one is ʿurub in 56:37. The opposing root ʿ-j-m occurs only 4 times in the whole text: 16:103, 26:198 and 41:44 (twice).
In the first of these the stress falls not on the ancestry of words but on the hearer's understanding:
إِنَّآ أَنزَلْنَـٰهُ قُرْءَٰنًا عَرَبِيًّا لَّعَلَّكُمْ تَعْقِلُونَ
"Indeed, We have sent it down as an Arabic Qur’ān that you might understand." (12:2)
The same description, with the qualifier mubīn, "clear":
بِلِسَانٍ عَرَبِىٍّ مُّبِينٍ
"In a clear Arabic language." (26:195)
16:103 sits inside an accusation. The verse answers the charge "a human being is teaching him." The subject is not the etymology of vocabulary but the tongue of the man being pointed at:
وَلَقَدْ نَعْلَمُ أَنَّهُمْ يَقُولُونَ إِنَّمَا يُعَلِّمُهُۥ بَشَرٌ
"And We certainly know that they say, “It is only a human being who teaches him [i.e., the Prophet (ﷺ)].”" (16:103)
The answer comes in the second half of the same verse:
لِّسَانُ ٱلَّذِى يُلْحِدُونَ إِلَيْهِ أَعْجَمِىٌّ وَهَـٰذَا لِسَانٌ عَرَبِىٌّ مُّبِينٌ
41:44 is a hypothetical. It asks "what if it had been in a foreign tongue?" and reports the reaction that would have followed:
وَلَوْ جَعَلْنَـٰهُ قُرْءَانًا أَعْجَمِيًّا لَّقَالُوا۟ لَوْلَا فُصِّلَتْ ءَايَـٰتُهُۥٓ ۖ ءَا۬عْجَمِىٌّ وَعَرَبِىٌّ
"And if We had made it a foreign [i.e., non-Arabic] Qur’ān, they would have said, “Why are its verses not explained in detail [in our language]? Is it a foreign [recitation] and an Arab [messenger]?”" (41:44)
The rest of the verse moves the axis openly onto comprehension: those who do not believe have "deafness" in their ears, and the book is "blindness" upon them. The dispute is not about purity of language but about opacity versus clarity.
26:198 sets up the same hypothetical from the other side:
وَلَوْ نَزَّلْنَـٰهُ عَلَىٰ بَعْضِ ٱلْأَعْجَمِينَ
"And even if We had revealed it to one among the foreigners" (26:198)
And this is not a rule invented for one book; 14:4 states the general principle:
وَمَآ أَرْسَلْنَا مِن رَّسُولٍ إِلَّا بِلِسَانِ قَوْمِهِۦ لِيُبَيِّنَ لَهُمْ
"And We did not send any messenger except [speaking] in the language of his people to state clearly for them," (14:4)
Now the words the objection treats as evidence. They are genuinely there; nobody hides them. Two fabric names appear in a description of the garden:
وَيَلْبَسُونَ ثِيَابًا خُضْرًا مِّن سُندُسٍ وَإِسْتَبْرَقٍ
"They will be adorned therein with bracelets of gold and will wear green garments of fine silk and brocade," (18:31)
A drink mixture:
وَيُسْقَوْنَ فِيهَا كَأْسًا كَانَ مِزَاجُهَا زَنجَبِيلًا
"And they will be given to drink a cup [of wine] whose mixture is of ginger" (76:17)
A garden name:
ٱلَّذِينَ يَرِثُونَ ٱلْفِرْدَوْسَ هُمْ فِيهَا خَـٰلِدُونَ
"Who will inherit al-Firdaus. They will abide therein eternally." (23:11)
A kind of stone:
وَأَمْطَرْنَا عَلَيْهَا حِجَارَةً مِّن سِجِّيلٍ مَّنضُودٍ
"and rained upon them stones of layered hard clay," (11:82)
Together these seven words (istabraq, sundus, firdaws, zanjabil, salsabil, sijjil, qistas) account for 16 tokens in this platform's word table; the table holds 77,429 word tokens in total.
What do we learn?
(interpretation) No verse says "this book contains not one word of foreign origin." What is said is something else: the book came down in the language its audience spoke. Every time the description appears it is bolted to a purpose clause — "that you might understand" (12:2, 43:3), "for a people who know" (41:3), "that you may warn" (42:7, 46:12). ʿArabī is therefore not a statement of ancestry but a statement of communication: this text speaks in a language its first hearers could actually process.
(interpretation) The logic of 41:44 confirms this. The verse does not defend linguistic purity; it removes the excuse "we cannot understand it." Its premise is that a foreign-language book would have drawn the complaint "why are its verses not explained in detail?" The measure the verse applies is intelligibility. 26:198 tries the same measure in reverse: had it been sent to a non-Arabic speaker, they still would not have believed — so the real barrier is disposition, not language.
(interpretation) 16:103 is a dispute about source, not etymology. The charge is "a human being is teaching him"; the reply contrasts the speech of the accused informant with the text in front of them. In that context aʿjamī cannot mean "containing no word of Persian origin"; it points at the quality of a person's speech.
(interpretation) As for loanwords: no living language is a sealed jar. Once a word enters a language it takes on that language's sounds, patterns and inflections, and it belongs to it. The text shows this happening — istabraq carries Arabic tanwīn after a preposition (إِسْتَبْرَقٍ), zanjabīl takes accusative tanwīn (زَنجَبِيلًا). These words live inside Arabic grammar; they are not foreign bodies bolted onto it. An English speaker saying "algebra", "sugar" or "coffee" has not stopped speaking English; the same measure can be applied here.
(interpretation) Quantity matters too, but should not be treated as proof on its own. The seven examples above occur 16 times in a text of 77,429 word tokens — a very small share. Honesty requires the other half of the sentence: the list of words counted as foreign-origin varies from scholar to scholar and is not limited to seven. The proportion argument shows the scale of the objection; it does not by itself refute it.
Different readings
(interpretation) A root-centred linguistic reading. In this platform's root lexicon, the core sense of ʿ-j-m is not "foreign nation" but the opposite of speaking clearly: obscurity, unintelligibility. The same entry notes that a person whose speech is unclear is called aʿjam whether or not he is an Arab, and that the verb for speaking plainly is aʿraba. On this reading the contrast in 16:103 and 41:44 is not "Arabic tongue versus Persian tongue" but "clear speech versus opaque speech." The continuation of 41:44 — deafness in the ears, blindness upon them — supports it.
(interpretation) A traditional reading, first branch. One view holds that there are no foreign words in the Qur'an at all: such words had already been Arabicised before the revelation, so at the moment they were revealed they were Arabic. On this reading origin and usage are separate matters, and ʿarabī describes usage.
(interpretation) A traditional reading, second branch. Another view accepts that words of foreign origin are present and finds no problem in it, because ʿarabī describes the whole — the idiom, the register, the composition — and not the pedigree of each lexeme. These two branches have stood side by side for centuries; choosing between them comes not from the text but from one's theory of language.
(interpretation) A historical-linguistic reading. Seventh-century Hijazi Arabic was already woven through with Aramaic, Syriac, Persian, Ethiopic and Greek borrowings carried in by trade, migration and religion. On this reading, loanwords in the text are not a defect but a sign that the text really touches the place and time it addresses: not a sealed dictionary dropped from the sky, but an address spoken from inside a living language.
(interpretation) A reading that keeps the objection at full strength. This much should be granted: if someone takes ʿarabī mubīn to mean "lexically unmixed Arabic," then loanwords genuinely damage that understanding. The strength of the counter-reading lies in showing that no verse makes such a purity claim. The argument runs over an expectation placed on the text, not over anything the text asserts.
Honest boundary
Certain in the text. The Qur'an calls itself ʿarabī in 10 places; the root ʿ-j-m occurs 4 times. 41:44 frames a hypothetical, not a language policy. 14:4 lays down as a general principle that a messenger comes in the tongue of his own people. Istabraq, sundus, firdaws, zanjabil, salsabil, sijjil and qistas do occur in the text, 16 times in total. No verse says "there is no word of foreign origin in this book."
Left to interpretation. Whether a given word is "originally" Persian, Aramaic or Ethiopic is a reconstruction of historical linguistics; the Qur'an makes no statement about the origin of any word. That is why origin lists grow or shrink depending on the source. Likewise, root assignment in a morphology database is an editorial decision, not part of the revelation: in this platform's data istabraq, sundus, zanjabil, salsabil and firdaws carry no three-letter root, while sijjil (س ج ل) and qistas (ق س ط س) are attached to one. But the absence of a root is not by itself evidence of foreignness: in the same table 27,161 of 77,429 word tokens — mostly particles, conjunctions and affixes — are recorded without a root.
What this article does not claim. Nothing here settles which word came from which language, and no etymological list is endorsed. The only thing done here is to separate what the verses say from what they do not.
Conclusion: Every time the Qur'an calls itself ʿarabī, the description arrives attached to a purpose clause aimed at being understood; it carries no claim about the ancestry of its vocabulary. 16:103 answers a charge of "someone is teaching him," and 41:44 answers an excuse of "we do not understand"; in both, the axis is clarity against opacity. Loanwords, meanwhile, are the ordinary condition of every living language, and in this text they appear fully inflected inside Arabic grammar. The objection stands only if ʿarabī mubīn is loaded with a "lexical purity" that the verses never assert.
Related articles
- Why is the Qur'an in Arabic? Universality or Arab-centrism?
- Is the Qur'an copied from earlier scriptures?
- Has the Qur'an been preserved? Are variant readings a contradiction?
- Does the Qur'an contradict itself? (Creation: 6 days or 8?)
Source: Qur'anic verses (M. Okuyan's Turkish rendering; Saheeh International in English). Presented with a text/interpretation distinction; this is not a legal ruling.