The consolation of consonants

This tool maps the constellation of function across the scrolls. Each of the 1,651 roots that the Quranic Arabic Corpus tags in the Qurʾan is set against every root Brown–Driver–Briggs recognizes in the Hebrew and Aramaic of the Bible, and joined wherever the consonants correspond.

Three fixed points

Consonants are primary. A triliteral is three fixed points, and the grammar happens between them. Arabic grammar names the two abstractions whose meeting yields a word: the root, al-jidhr (الجذر), and the measure, al-wazn (الوزن). Laid over the root, the measures generate verbs, participles, verbal nouns, nouns of place and of instrument: a constellation of function. The grammarians weigh every word on the model verb faʿala (فعل) and name the three radicals by its letters: the first is the fāʾ of the word, the second its ʿayn, the third its lām. Every alignment in the tool keeps those names.

Each root carries its own constellation of function: the measures it actually takes in its scroll, with their counts, and the verses where it stands. The glosses indicate a field; the verses show the function.

The consolation

Ibn Janāḥ, in the Kitāb al-Tanqīḥ—the grammar Kitāb al-Lumaʿ and the root book Kitāb al-ʾUṣūl—reduced Hebrew to its triliterals and let Arabic illuminate what Hebrew left obscure. Where the Hebrew grid was ambiguous, Arabic supplied consolation. The tool does the same work across both corpora: it starts from the Qurʾanic root index and asks, root by root, where the same consonantal grid stands in the Hebrew and the Aramaic.

Where the rules come from

The table of regular correspondences—which Arabic letter stands where Hebrew writes which—comes from modern comparative Semitics, the Proto-Semitic reconstructions of Brockelmann, Moscati and Huehnergard. The categories that sort those correspondences and explain their behavior are the Arabic grammarians’ own:

• al-Khalīl ibn ʾAḥmad (d. c. 786–791), Kitāb al-ʿAyn: the alphabet ordered by makhraj from the throat outward, beginning with ʿayn, and the taqālīb, every permutation of a root’s letters.
• Sībawayh (d. c. 796), al-Kitāb: sixteen places of articulation, the qualities of the letters (ʾiṭbāq, ṣafīr), and the chapter on assimilation, ʾidghām.
• Ibn Jinnī (d. 1002): Sirr Ṣināʿat al-ʾIʿrāb, on each letter and its substitutions (ʾibdāl); al-Munṣif, on the weak roots; al-Khaṣāʾiṣ, on al-ishtiqāq al-ʾakbar, the turnings of a root.
• ʾAbū al-Ṭayyib al-Lughawī (d. 962), Kitāb al-ʾIbdāl.
• Judah ibn Quraysh, the Risāla, and Ibn Janāḥ (d. c. 1050), who first carried these categories into Hebrew.

The regular shift

The grids do not line up letter for letter, and they are not expected to. Arabic still holds apart sounds that the Hebrew script folds together under a single letter. Read by makhraj, every correspondence is a seam beneath a letter:

• The throat letters (ḥurūf al-ḥalq). Arabic sounds six, Hebrew writes four. ḥāʾ and khāʾ both stand behind ḥet; ʿayn and ghayn both stand behind ʿayin: ח-מ-ר (ḥet–mem–resh) / ح–م–ر (ḥāʾ–mīm–rāʾ), red; ח-מ-שׁ (ḥet–mem–šin) / خ–م–س (khāʾ–mīm–sīn), five; ע-ר-ב (ʿayin–resh–bet) / غ–ر–ب (ghayn–rāʾ–bāʾ), evening.
• The covered letters (ḥurūf al-ʾiṭbāq). Of the four, ṣād, ḍād and ẓāʾ are gathered under one ṣade; ṭāʾ keeps its ṭet. Hebrew qayiṣ (קַיִץ) stands beside Arabic qayẓ (قَيْظ).
• The gum letters (al-ḥurūf al-lithawiyya, al-Khalīl’s term for thāʾ, dhāl and ẓāʾ). Hebrew has none of them: thāʾ is written šin and dhāl is written zayin: שׁ-ל-שׁ (šin–lamed–šin) / ث–ل–ث (thāʾ–lām–thāʾ), three; ז-כ-ר (zayin–kaf–resh) / ذ–ك–ر (dhāl–kāf–rāʾ), remember.
• The whistling letters (ḥurūf al-ṣafīr) and the spreading shīn (al-tafashshī). sīn answers both šin and samek, while shīn answers sin: שׁ-מ-ע (šin–mem–ʿayin) / س–م–ع (sīn–mīm–ʿayn), hear; ס-ג-ד (samek–gimel–dalet) / س–ج–د (sīn–jīm–dāl), bow down; ע-שׂ-ר (ʿayin–sin–resh) / ع–ش–ر (ʿayn–shīn–rāʾ), ten.

Every other letter stands in its own place: b, t, d, r, z, ṭ, f, q, k, l, m, n, h, and jīm for gimel.

The weak letters

The grammarians call wāw, yāʾ and the seat of the hamza the letters of defect, ḥurūf al-ʿilla. Their rule of ʾiʿlāl accounts for a root whose underlying form, its ʾaṣl, surfaces changed: qāla is weighed as qawala. The same rule governs the correspondences, position by position.

• Mithāl, the weak fāʾ: the Hebrew initial yod answers the Arabic initial wāw by the regular shift—י-ל-ד (yod–lamed–dalet) / و–ل–د (wāw–lām–dāl); י-ר-שׁ (yod–resh–šin) / و–ر–ث (wāw–rāʾ–thāʾ). The wāw of the ʾaṣl surfaces again in the Hebrew hifʿil holid (הוֹלִיד).
• ʾAjwaf, the hollow ʿayn: ק-ו-מ (qof–waw–mem) / ق–و–م (qāf–wāw–mīm), stand.
• Nāqiṣ, the weak lām: where Arabic writes yāʾ or wāw, Hebrew writes he—ב-נ-ה (bet–nun–he) / ب–ن–ي (bāʾ–nūn–yāʾ), build—and the yod returns when the verb is conjugated: banoh baniti (בָּנֹה בָנִיתִי, 1 Kings 8:13).
• Muḍāʿaf, the doubled root: צ-ל-ל (ṣade–lamed–lamed) / ظ–ل–ل (ẓāʾ–lām–lām), shade.
• Mahmūz, the hamzated root: the hamza answers ʾalef—א-כ-ל (ʾalef–kaf–lamed) / ء–ك–ل (hamza–kāf–lām), eat; ק-ר-א (qof–resh–ʾalef) / ق–ر–ء (qāf–rāʾ–hamza), call, recite.
• Maḥdhūf al-lām, the dropped third radical: Hebrew writes yad (יָד), dam (דָּם) and šem (שֵׁם) on two radicals—י-ד (yod–dalet), ד-מ (dalet–mem), שׁ-מ (šin–mem)—where the grammarians weigh yad, dam and ism on three: ي–د–ي (yāʾ–dāl–yāʾ), د–م–و (dāl–mīm–wāw), س–م–و (sīn–mīm–wāw). The plural restores the lost letter: ʾaydin, ʾasmāʾ. The tool pairs these as biradical.

The turning of the letters

Qalb al-ḥurūf—the grammarians’ al-qalb al-makānī—moves a radical and leaves the constellation standing: ע-ב-ר (ʿayin–bet–resh), turned, is ع–ر–ب (ʿayn–rāʾ–bāʾ), ʿibri and ʿarab. The tool tries every turning of every Qurʾanic root; because turnings are many and chance matches are easy, it keeps only those whose glosses stand close (proximity 0.60 or higher).

The Aramaic scroll

The Aramaic of Daniel and Ezra writes the gum letters as stops: thāʾ as taw (ת-ל-ת (taw–lamed–taw), three), dhāl as dalet (ד-כ-ר (dalet–kaf–resh)), ẓāʾ as ṭet (ט-ל-ל (ṭet–lamed–lamed), shade), and ḍād as ʿayin (א-ר-ע (ʾalef–resh–ʿayin), earth). Jeremiah 10:11 writes the earth both ways in one verse, ʾarqaʾ (אַרְקָא) and ʾarʿaʾ (אַרְעָא). The Qurʾan lets the same passage be heard inside Arabic: in muddakir (مُدَّكِر), measure VIII of ذ–ك–ر (dhāl–kāf–rāʾ), the dhāl passes into dāl by assimilation, ʾidghām—fa-hal min muddakir—Qurʾan, Surat al-Qamar سورة القمر “The Moon” 54:15.

How the tool applies the rules

• Weighing both corpora to their radicals. Each Qurʾanic root comes from the Corpus root tags; each biblical root is a BDB root entry, with šin and sin kept apart; common words BDB leaves without a root are reduced to their consonants.
• The regular shift. Each Arabic letter carries a short list of the Hebrew letters that may stand in its place, with a separate list for the Aramaic scroll. The tool writes out every allowed Hebrew spelling of a Qurʾanic root and looks each one up among the biblical roots.
• ʾIʿlāl by position. A weak fāʾ may surface as yod; a weak ʿayn as waw or yod; a weak lām as he (in Aramaic also ʾalef).
• Maḥdhūf al-lām. A weak or doubled radical is set aside to meet the biblical two-radical nouns.
• ʾIbdāl outside the table. Other substitutions are admitted only at a second rank and marked “outside the regular shift.”
• Taqālīb. Every permutation of a Qurʾanic root is tried under the regular shift alone.

The weighing

Every pair is then set on the mīzān three ways: by its correspondence (the regular shift, weak-letter variance, biradical, turned, or outside the laws); by its witness (whether an etymology in Wiktionary names the two words as cognates); and by the proximity of the English glosses, measured from 0 to 1 by a sentence-embedding model. This weighing is the tool’s own; it does not come from the grammarians.

Tier E is kept on purpose. A shared grid whose glosses part ways may be two homographs, or it may be one function heard differently in two scrolls; the verses decide, not the English.

The Septuagint, the lexical spine

The Greek translation of the Hebrew scriptures made in Alexandria is the hinge between the scrolls. The New Testament cites the Hebrew scriptures largely in this Greek and writes in its vocabulary, so the words the Septuagint chose for a Hebrew root are the bridge by which a word of the New Testament reaches back to the Hebrew and, through the regular shift, on to the Qurʾan.

For every biblical root the tool lists the Greek words the Septuagint uses to render it, with the number of times it does so and the first place it does. The alignment is read word by word from the CATSS parallel text of the Hebrew Bible and the Septuagint, across every book. Each Hebrew word is brought to its root through the tagged Hebrew text of Tyndale House; where one spelling belongs to two roots, the meaning of the word decides. Each Greek word is brought to its dictionary form through the CATSS morphology of the Septuagint.

• Two renderings or more make a path. Only a Greek word the Septuagint uses at least twice for a root joins the New Testament to that root and, through it, to the Qurʾan. A root’s own card shows every rendering; the ones used once stand apart, under “Once in the Septuagint.”
• The reach of the map. 2,847 of the 3,081 biblical roots carry Greek renderings: 6,814 Greek words in 27,945 links, 9,936 of them used twice or more. Through them, 3,080 words of the Greek New Testament reach a Hebrew root, 2,308 of them through a rendering the Septuagint uses twice or more.
• The Greek New Testament view starts from the Greek: choose a word of the New Testament and follow it through the Septuagint to the Hebrew roots it renders, and from each root to its Arabic correspondences and their verses in the Qurʾan.
• The whole map as a table. Every link, Hebrew root, Greek word, count, first verse, New Testament count, paired Arabic root with its tier, and every verse of the Qurʾan where that root stands, can be downloaded as one spreadsheet: ocabs-lxx-map.csv.

Around each root

Each card opens outward. In Star Map opens the root in the Constellation of Function, the map of the roots and their correspondences across the scrolls. The lexicons open over the page, each in its own window, with a link back to its source:

• Arabic. The Arabic Lexicon of Ikram Hawramani, which gathers thirteen classical dictionaries at each root, with an English translation of any entry on request; and Lane’s Arabic-English Lexicon, entire.
• Hebrew and Aramaic. Brown–Driver–Briggs; Klein’s Etymological Dictionary and Jastrow’s Dictionary of the Targumim, read from Sefaria; and Gesenius’s Hebrew and Chaldee Lexicon in Tregelles’s translation, with a link to the very page of the printed book.
• Greek. Liddell, Scott and Jones, the full article for each Greek word of the Bible.

Fr. Paul Tarazi on the roots and the verses

Where Fr. Paul Nadim Tarazi speaks to a root in his books or in his recorded lectures, the Orthodox Audio Bible Commentary, the card carries a Tarazi pill. It opens his words on the root, one to four excerpts, each with its citation: the page of the book, or the lecture and the minute in the recording. Every quotation is checked letter for letter against its source. The lectures are copyedited for punctuation, and every Hebrew, Greek and Arabic word is written in the Andalus transliteration; his words are not changed. A second, independent reading keeps only the excerpts in which he speaks to the root itself.

The verses under each Hebrew root are marked the same way. A verse filled in olive is one on which he comments directly; a verse tinted olive stands in a passage or a chapter on which he comments. Choosing the verse opens its Hebrew and English with his words beneath, and a link to all his comments on that chapter. The Hebrew Bible numbers some verses differently from the English Bibles his works cite; the tool maps one onto the other and says so when they differ.

Three further threads run from his work into the map. Where he names the Greek by which the Septuagint renders a Hebrew word, the rendering is marked with a gold dot in the root’s list of Greek words, and his words stand at the foot of the card, on the Greek word’s card as well; the downloadable map carries the citation in its last column. Where he explains a word through its Arabic, his remark stands on the Arabic root. And where he reads a New Testament text together with an Old Testament one, as a quotation, an echo, borrowed wording or a deliberate comparison, the Old Testament verse is marked and opens with his words under “Read with the New Testament”; when the Septuagint joins the two verses through a root and a Greek word, the passage is also listed on that root’s card. Each of these was found by reading his books and lectures, and each quotation is checked against its source.

The New Testament is read the same way, chapter by chapter: his comments on a verse, a section or a chapter open from the New Testament references in the tool, and each Greek word of the New Testament shows what he says on verses where it occurs, first those in which he speaks of the word itself.

Sources and limits

• Qurʾan: Quranic Arabic Corpus morphology (root and lemma tags, 50,269 rooted tokens); English word-by-word glosses from quran.com, with their romanized names brought into Brill/IJMES form.
• Bible: OpenScriptures Brown–Driver–Briggs and Lexical Index; the Westminster Leningrad Codex with OSHB morphology for counts and verses (Masoretic verse numbering; verse links open Sefaria, which follows the same numbering).
• Witnesses: Wiktionary etymologies (Arabic, Hebrew and Aramaic entries) via kaikki.org. Only the words are kept, in their own scripts.
• The digitized BDB replaces its Arabic citations with a placeholder, so the tool can only flag that BDB cites Arabic at a root, not which root.
• BDB homographs (I, II, III…) are kept apart; a witness is attached to the homograph whose glosses stand closest.
• Septuagint: the CATSS parallel aligned Hebrew–Greek text and the CATSS morphology of the Septuagint (CCAT, University of Pennsylvania). Tagged Hebrew, the Greek New Testament and the brief Greek lexicon: STEPBible data, Tyndale House, Cambridge (CC BY). About one Greek form in eleven could not be brought to its dictionary form and is left out of the counts. The Septuagint at times translates a Hebrew text other than the Masoretic; the map records what the Greek has, not what it should have had.
• Lexicons: Lane (Perseus, Tufts University, CC BY-SA, via the laneslexicon project); the Arabic Lexicon (arabiclexicon.hawramani.com); Brown–Driver–Briggs (Open Scriptures, CC BY); Klein and Jastrow (Sefaria); Gesenius–Tregelles (bible-teka.com, with page images of the 1890 printing at the Internet Archive); LSJ (STEPBible, CC BY).
• Fr. Paul Nadim Tarazi: his commentaries published by OCABS Press and the Orthodox Audio Bible Commentary. Excerpts are chosen by machine and checked against their sources; a missing pill means only that no passage was found to speak to that root or verse.
• The proximity score hears English glosses, not the Semitic field. Before a correspondence enters prose, hear it in Lane and in BDB.